Industrial Environmental Chemistry - Waste Minimiz

Industrial
Environmental
Chemistry
Waste Minimization in
Industrial Processes and
Remediation of Hazardous Waste
INDUSTRY-UNIVERSITY COOPERATIVE CHEMISTRY
PROGRAM SYMPOSIA
Published by Texas A&M University Press
ORGANOMETALLIC COMPOUNDS
Edited by Bernard L. Shapiro
HETEROGENEOUS CATALYSIS
NEW DIRECTIONS IN CHEMICAL ANALYSIS
APPLICATIONS OF ENZYME BIOTECHNOLOGY

Edited by Jeffery W. Kelly and Thomas o. Baldwin
CHEMICAL ASPECTS OF ENZYME BIOTECHNOLOGY: Fundamentals

Edited by Thomas o. Baldwin, Frank M. Raushel, and A. Ian Scott
DESIGN OF NEW MATERIALS

Edited by D. L. Cocke and A. Clearfield
FUNCTIONAL POLYMERS
Edited by David E. Bergbreiter and Charles R. Martin
INDUSTRIAL ENVIRONMENTAL CHEMISTRY: Waste Minimization in

Industrial Processes and Remediation of Hazardous Waste
Edited by Donald T. Sawyer and Arthur E. Martell
METAL-METAL BONDS AND CLUSTERS IN CHEMISTRY

AND CATAL YSIS
Edited by John P. Fackler, Jr.
OXYGEN COMPLEXES AND OXYGEN ACTIVATION BY

TRANSITION METALS
Edited by Arthur E. Martell and Donald T. Sawyer
Industrial
Environmental
Chemistry
Waste Minimization in
Industrial Processes and
Remediation of Hazardous Waste
Edited by
Donald T. Sawyer and
Arthur E. Martell
Texas A&M University
College Station, Texas
Springer Science+Business Media, LLC

Llbrary of Congress Cataloging-ln-Publtcatlon Oata
Industrlal env lronmental chemlstry : waste mi n 1m1 z a t i on In industrtal

pro cesses a nd remedlation of hazardous oaste ! ed ited by Oonald T.
Sa"yer and Arthur E. Martell.
p. cn . -- Clndustry-unlvers lty cooperat1ve chemistry progran
sy.postal
·Proceedlngs of t he Texas A&M Unlvers lty. IUCCP tenth annua!
sy.poslu, on Industrlal Envlronmental Che.lstry , Waste Mini.lzation
I n Industrlal Proeesses and Re.ed lat lon of Hazardous Waste. held
March 24-26, 1992, at Texas A&M Uni verslty. College St a t i on , Texas·-
-T .p. verso .
Includes b ib l 1ograpr. ical references and index .
1 . Factory and trade waste--Congresses. 2 . Haste minlmizat 1on-

- Congresses . 3 . Hazardous wastes--Pur1f ication--Congresses .
I . Sa.yer , OonaJd T. 11 . Martell , Arthur Ear l . 1916-
111. Texas A & M Uni verslty . Industr y-Un lverslty Cooperative
Chemist r y Program. IV . Symposiun on Industrial Environmenta l
Chemistry : Haste Min imization 1n In~ustr ial Processes and
Remed latlon of Hazardous Haste (1992 : Texas A & M Unlvers tty)
V. Serles .
T0896.I614
628 .5--oe20 92 -28759
CIP
Proceedings of the Texas A&M University, IUCCP Tenth Annual Symposium

on Industrial Environmental Chemistry; Waste Minimization in
Industrial Processes and Remediation of Hazardous Waste,
held March 24-26 , 1992, at Texas A&M University, College Station, Texas
ISBN 978-1-4899-2322-6 ISBN 978-1-4899-2320-2 (eBook)

DOI 10.1007/978-1-4899-2320-2
© 1992 Springer Science+Business Media New York

Originally published by Plenum Press , New York in 1992.
Softcover reprint of the hardcover 1st edition 1992
All rights reserved

No part of this book may be reproduced , stored in a retrieval system, or transmitted
in any form or by any means, electronic , mechanical, photocopying, rnicrofilming,
recording, or otherwise, without written permission from the Publisher
SPONSORS
BASF Corporation
BF Goodrich
Dow Chemical Company
Hoechs t-Celanese Corporation
Monsanto Chemical Company
Institute of Biosciences & Technology,

Texas A&M University System
v
PREFACE
This monograph consists of manuscripts submitted by invited speakers who

participated in the symposium "Industrial Environmental Chemistry: Waste
Minimization in Industrial Processes and Remediation of Hazardous Waste," held
March 24-26, 1992, at Texas A&M University. This meeting was the tenth annual
international symposium sponsored by the Texas A&M Industry-University
Cooperative Chemistry Program (IUCCP). The program was developed by an
academic-industrial steering committee consisting of the co-chairmen, Professors
Donald T. Sawyer and Arthur E. Martell of the Texas A&M University Chemistry
Department, and members appointed by the sponsoring companies: Bernie A.
Allen, Jr., Dow Chemical USA; Kirk W. Brown, Texas A&M University; Abraham
Clearfield, Texas A&M University; Greg Leyes, Monsanto Company; Jay Warner,
Hoechst-Celanese Corporation; Paul M. Zakriski, BF Goodrich Company; and Emile
A. Schweikert, Texas A&M University (IUCCP Coordinator).
The subject of this conference reflects the interest that has developed in
academic institutions and industry for technological solutions to environmental
contamination by industrial wastes. Progress is most likely with strategies that
minimize waste production from industrial processes. Clearly the key to the
protection and preservation of the environment will be through R&D that
optimizes chemical processes to minimize or eliminate waste streams. Eleven of
the papers are directed to waste minimization. An additional ten papers discuss
chemical and biological remediation strategies for hazardous wastes that
contaminate soils, sludges, and water.
Acknowledgement is made to others whose participation in the Conference is
not reflected in this book. Thanks and appreciation are due to Dr. Michael B. Hall,
Head, Chemistry Department, and to Dr. Herbert H. Richardson, Chancellor, Texas
A&M University System, who gave official welcoming remarks. There were three
keynote lectures, that were very well received: Laurie King, Chief of the Oklahoma-
Texas Permits Section, EPA, on "Upcoming Changes in Regulations that May Have
an Impact on Remediation of Hazardous Waste", Jerry B. Martin, Director,
Environmental Affairs, Dow Chemical, U.S.A. on "No More Room for Dinosaurs"
vii
and Jim Conner, Manager, Environmental Projects, Hoechst-Celanese Corporation
on "WARR - Implementing a Waste Reduction Strategy". All participants at the
Banquet enjoyed the theme lecture by James A. Fava, Roy F. Weston, Inc. on
"Should Improvement to the Environmental Quality of Products Address Product
Life-Cycle Concepts". Also very much appreciated are the services of Chairs of the
Sessions: Sa V. Ho (Monsanto), Paul M. Zakriski (BF Goodrich), Jay Warner
(Hoechst-Celanese), Jams R. Ryffel (BASF Corporation) and Aydin Akgerman (Texas
A&M).
We thank Mary Martell for assistance with logistics of the symposium and
for her help and that of Debbie Shepard with the organization and preparation of
the final manuscript.
Donald T. Sawyer
Arthur E. Martell
College Station, Texas

May, 1992
viii
CONTENTS
Introduction: Strategies for the Protection of the Environment. . . . . . . . . . 1
Donald T. Sawyer and Arthur E. Martell
WASTE MINIMIZATION:
Practical Techniques for Reducing Waste .............................. 3
Kenneth E. Nelson
WARR: Implementing a Waste- and Release-Reduction Strategy. . . . . .. 19
James A. Conner
Waste Minimization by Process Modification ......................... 25
J. R. Hopper, C. L. Yaws, T. C. Ho,

M. Vichailak, and A. Muninnimit
Kinetic Models to Predict and Control Minor Constituents in ........... 45

Process Reactions
Selim Senkan
Kinetic Control of By-Product Reactions with In Situ Distillation. . . . . .. 65

in a Continuous Flow Device
Robert P. Shirtum
Waste Minimization and Remediation in the Metal Finishing .......... 71

Industries
Clifford W. Walton, Kevin S. Briggs,

and Kevin J. Loos
The Role of Catalysis in Industrial Waste Reduction ................... 89
David Allen
Catalysis: The Atom Utilization Concept and Waste ................... 99

Minimization
R. A. Sheldon
ix
Industrial Scale Membrane Processes ................................ 121
James R. Fair
Novel Microporous Membrane-Based Separation Processes ............ 135

for Pollution Control and Waste Minimization
Chang H. Yun, A. K. Guha, Ravi Prasad,
and Kamalesh K. Sirkar
Ion Exchange for Glyphosate Recovery ......... " ................... 147
Lowell R. Smith and Jane L. Barclay
REMEDIATION OF HAZARDOUS WASTES:

Supercritical Extraction in Environmental Remediation ............... 153
and Restoration
Aydin Akgerman
Ozone for Waste Remediation and Waste Water Treatment ........... 171
Henry J. Ledon
Nucleophilic and Reductive Remediation Strategies for .............. 181

Hazardous Halogenated Hydrocarbons (PCBs, HCB, PCP,
TCE, CCI4, DCBP, EDB, DDT, and DOE)
Donald T. Sawyer, Paul K. S. Tsang, Seungwon Jeon,
and Marjorie Nicholson
Process Technology for Hazardous Waste Remediation ................ 213
(KPEGsm and Reductive Technologies)
Robert Hoch
A New Membrane Process for Recovering Organics from. . . . . . . . . . . . .. 229
Aqueous Wastes
S. V.Ho
Application of Immobilized Cell Technology for Biotreatment . . . . . . . .. 247

of Industrial Waste Streams
Daniel E. Edwards and Michael A. Heitkamp
Membrane-Bound Microorganisms for Removing 261

OrganiCS and Metals
S. S. Sofer
Biofiltration: An Air Pollution Control Technology for ............... 273
Hydrogen Sulfide Emissions
Eric R. Allen and Yonghua Yang

Inorganic Ion Exchange Materials for Nuclear Waste .................. 289
Effluent Treatment
Abraham Clearfield
Appendix I
Abstract of Paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 301
Appendix II
Posters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 303
Index ............................................................. 305
xi
INTRODUCfION: STRATEGIES FOR THE PROTECfION
OF THE ENVIRONMENT
Donald T. Sawyer and Arthur E. Martell
Department of Chemistry
College Station, Texas 77843
Although the contamination of our environment by industrial wastes has

been recognized as a major problem since the 1960s, progress towards its resolution
is very slow. Until recently hazardous wastes have been stored or dumped at land-
fills and waste-ponds, and into rivers, lakes, and oceans. There is increasing
evidence that storage and disposal in land-fills is not an effective strategy, and that
incineration has the potential to disperse a liquid/solid waste into the atmosphere.
The Environmental Protection Agency has catalogued thousands of sites in the
United States that are severely contaminated with hazardous materials (PCBs, heavy
metals, pesticides, chlorinated solvents, and wastes from the petroleum, chemical,
pharmaceutical, electronics, and manufacturing industries), and has convinced
Congress to appropriate in excess of $100 billion for the remediation of these
Superfund sites. To date these funds have been used mainly to investigate, to
evaluate, and to litigate Superfund sites, rather than to remediate.
This Symposium was organized to bring together representatives from
Universities, Industry, and Government that are concerned with strategies (a) to
minimize waste production and (b) to remediate existent hazardous wastes with
effective and economic technologies. The "Theme" talk was given by Jerry B.
Martin (Director, Environmental Affairs, Dow Chemical USA), No More Room for
Dinosaurs, which provided compelling evidence that antiquated waste-producing
processes are bad for business economics and will soon make an industry non-
competitive. This, coupled with the presentations from major industries that have
adopted a zero-effluent philosophy, represents a major advance and one that is
driven by economics rather than regulations. We believe that this strategy for
environmental protection is in the economic interest of all manufacturing and
production industries. The leadership that is exemplified by the papers in this
volume is impressive.
Industrial Environmental Chemistry, Edited by D.T. Sawyer

and A.E. Martell, Plenum Press, New York, 1992
An example that illustrates the merit of waste minimization strategies is the
replacement of PVC (polyvinyl chloride) with polybutylene for the manufacture of
plastic pipe. The production of vinyl-chloride monomer has an inevitable waste
stream that includes hexachorobenzene (C6 CI 6 , a human carcinogen), polychloro-
dibenzofurans, and related polychloro-aromatics. Currently these wastes are
processed by high-temperature, oxygen-rich chemical incinerators. Because C 6 Cl 6
does not support combustion, one must have concern as to its fate from pyrolysis in
an oxygen-rich atmosphere. Representatives of the EPA and operators of such an
incinerator believe that C 6 Cl 6 is transformed to C02 and HCl (Federal Courthouse,
Little Rock, Arkansas, June 7, 1991); more chemically reasonable products include
phosgene [Cl 2C(O)]' chlorine oxide (CI20), chlorine (Cl 2 ), and hypochlorous acid
(HOCl). By development of a pure-hydrocarbon polymer (polybutylene from
polymerization of butene-1) as a substitute for PVC, the polymer industry has
eliminated a major hazardous-waste problem through a strategy of waste
minimization. In so doing a superior product is on the market, and one that has
better profitability and zero chlorinated wastes.
Matters are much less encouraging with respect to the remediation of existent
hazardous wastes. Although it is human to want to forget the past, many hazardous
waste sites are persistent and are being dispersed into the ground water and the
atmosphere. Their effective remediation is limited by an adequate match of
technology and economics. If the entire contaminated site (e.g., 30,000 tons of soil
that contains 500 ppm PCBs) is to be remediated by incineration, the costs and the
concerns with the toxicity of (a) the exhaust gases and particulates and (b) the ash
(non-volatile residue) has prompted most to seek restraint in the courts. The
dispersed nature of hazardous materials in soils, sludges, and waste ponds dictates,
in the interests of economics and energy conservation, that separation and
concentration technologies be applied prior to their remediation. Thermal
desorption, distillation, solvent extraction, ion exchange, and supercritical-fluid
extraction (see Chapter by Akgerman) are useful technologies for the separation and
concentration of various contaminated matrices.
Although bioremedation is attractive in terms of its economics and its
potential to transform in situ toxic materials into safe, innocuous products; mo::;t
examples to date are for laboratory situations with moderately toxic substances. As
the papers here confirm, major R&D efforts will be required to achieve practical and
effective systems. Both the microorganisms and their environment must be
optimized in terms of selectivity, efficiency, durability, and remediation rates. At
best bioremediation technologies will be able to reduce the levels of some toxic
materials by one or two orders of magnitude within time periods of 3-25 weeks. The
challenge is to develop microorganisms with sufficient durability and specificity to
achieve these goals with a single inoculation.
2
PRACTICAL TECHNIQUES FOR REDUCING WASTE
Kenneth E. Nelson
Energy Conservation Manager

Dow Chemical U.S.A.
Plaquemine, LA 70765
INTRODUCTION
The future of the chemical industry is closely tied to waste reduction. At the Dow
Chemical Company, reducing waste has been an important objective for many years, and
we have a wealth of knowledge and experience in this important field. This paper is de-
signed to share some of that knowledge. Nearly all the ideas that will be discussed are
based on process changes that have actually been made by Dow plants.
Since waste reduction (rather than treatment) is the ultimate way to lower emissions,
we will focus our attention on techniques that produce fewer pounds of waste. Because
waste reduction and energy conservation often occur simultaneously, many of the pro-
jects discussed will also reduce energy consumption.
WHAT IS WASTE?
Let's start by defining waste. Waste includes streams or materials that are:
• Vented to the air
• Discharged to the water
• Sent to landfill
• Sent to an incinerator
• Sent to a flare
• Sent to a biological treatment facility
Waste reduction can be accomplished by creating less waste initially or by recycling

waste products back to the process. It is also possible to reduce emissions by operating
waste treatment units more efficiently, but that is not the concern of this paper. Also, we
will not discuss the nature of the waste (i.e. hazardous or non-hazardous, toxic or non-
toxic, flammable or non-flammable, etc.). This is not meant to be a comprehensive pa-
per. The author has not made an exhaustive study of all possible ways of reducing

and A.E. Martell, Plenum Press, New York. 1992 3
waste. But it does contain a wealth of practical ideas and is intended to serve as the basis
for discussion and brainstorming. We'll begin with ideas associated with raw materials.
RAW MATERIALS
Raw materials are usually purchased from an outside source or transferred from an
on-site plant. Each raw material needs to be studied to determine how it affects the
amount of waste produced. The specifications for each raw material entering the plant
should be closely examined.
Improve Quality of Feeds
Although the percentage of undesirable impurities in a feed stream may be low, it

can be a major contributor to the total waste produced by a plant. Reducing the level of
impurities may involve working with the supplier of a purchased raw material, working
with on-site plants that supply feed streams or installing new purification equipment.
Sometimes the effects are indirect (e.g. water gradually kills the reactor catalyst causing
formation of by-products, so a drying bed or column is added).
Use Off-spec Material
Occasionally, a process can use off-spec material (that would otherwise be burned or
land filled) because the particular quality that makes the material off-spec is not important
to the process.
Improve Quality of Products
Impurities in your own products may be creating waste in your customers' plants.
Not only may this be costly, it may cause customers to look elsewhere for higher quality
raw materials. Take the initiative in discussing the effects of impurities with your cus-
tomers.
Use Inhibitors
Inhibitors prevent unwanted side reactions or polymer formation. A wide variety of

inhibitors are commercially available. If inhibitors are already being used, check with
suppliers for improved formulations and new products.
Change Shipping Containers
If raw materials are being received in containers that can not be reused and need to
be burned or landfilled, change to reusable containers or bulk shipments. Similarly, con-
sider using alternative containers for shipping products to customers.
Re-examine Need for each Raw Material
Sometimes the need for a particular raw material (one which ultimately ends up as
waste) can be reduced or eliminated by modifying the process. A cooling tower, for ex-
ample, uses inhibitors to control algae growth. In one cross-flow cooling tower, the
need for algae inhibitors was cut in half by shielding the water distribution decks from
sunlight.
4
REACTORS
The reactor is the heart of the process, and can be a primary source of waste prod-
ucts. The quality of mixing in a reactor is crucial. Too often, insufficient design time is
spent on scaling up this key parameter and a new production facility has disappointing
yields when compared to lab or pilot plant data. Several of the ideas listed below are
concerned with quality of mixing.
Distribute Feeds Better
Here is an area that deserves more attention

than it typically gets. The problem is illustrated
in Figure 1 (left). Reactants entering at the top
of a fixed catalyst bed are poorly distributed.
Part of the feed short-circuits down through the
center of the reactor and does not allow adequate
time for conversion to desired products. In con-
trast, feed that is closer to the walls remains in
the reactor too long and "over-reacts" to by-
products that eventually become waste. One
solution is to add some sort of distributor and the
reactor inlet that causes the feed to move uni-
formly through all parts of the reactor as shown
on the right. A special collector at the bottom of Poor Improved
the reactor may also be necessary to prevent the Distribution Distribution
flow from necking down to the outlet. Simi-
Figure 1. Reactor Distribution
larly, if a gas reactant is added to a liquid, the
gas needs to be finely dispersed and evenly dis-
tributed throughout the liquid phase.
Improve Catalyst
Because of the significant effect a catalyst has on reactor conversion and product
mix, searching for better catalysts should be an ongoing activity. Changes in the chemi-
cal makeup of a catalyst, the method by which it is prepared, or its physical characteris-
tics (size, shape, porosity, etc.) can lead to substantial improvements in catalyst life and
effectiveness.
Improve Physical Mixing in the Reactor
Modifications to the reactor such as adding or improving baffles, installing a higher

rpm motor on the agitator(s), or using a different mixer blade design (or multiple impel-
lers) can improve mixing. Pumped recirculation can be added or increased. Two fluids
going through a pump, however, do not necessarily mix well and a static mixer may be
needed to insure good contacting.
Improve Way in Which Reactants are Added
The idea here is to get closer to ideal reactant concentrations before the feed enters
the reactor. This helps avoid secondary reactions which form unwanted by-products.
The way not to add reactants is shown in the upper half of Figure 2 on the following
page. It is doubtful that the "ideal" concentration exists anywhere in this reactor. A
5
consumable catalyst, es-
pecially, should be di-
luted in one of the feed
streams (one which does
not react in the presence Product
of the catalyst). The
lower half of Figure 2
illustrates one approach
to improving the situation
by using three in-line Reactant A
static mixers. Reactant B
Consider Different Re-

actor Design Solvent
The classic stirred-
tank backmix reactor is
not necessarily the best ~~~~~ Product
Reactant A
choice. A plug flow re- Reactant B
actor offers the advantage Solvent
that it can be staged, and
each stage can be run
at different conditions
(especially temperatures),
closely controlling the re-
action for optimum prod- Catalyst
uct mix (and minimum Solvent
waste). Many innovative Figure 2. Adding Reactants
hybrid designs are possi-
ble.
Provide Separate Reactor for
Recycle Streams
Feed
Recycling by-product and Streams Recycle
waste streams is an excellent
technique for reducing waste, but
often the ideal reactor conditions
for converting recycle streams
back to usable products are con-
siderably different from condi-
tions in the primary reactor. One Reactor
solution is to provide a separate, Product
smaller reactor for handling re-
cycle and waste streams as illus- Feed
trated in Figure 3 at the right. Streams
Temperatures, pressures and con-
centrations can then be optimized
in both reactors to take maximum
advantage of reaction kinetics. Figure 3. Separate reactor for recycle streams
6
Improve Control
For a given reactor configuration, there is one set of operating conditions that is op-
timum at any given time. The control system should know that condition and make it oc-
cur, with little fluctuation . Such control may be complex, particularly in the case of
batch reactors, but can yield major improvements. In less sophisticated systems, simply
stabilizing reactor operation frequently reduces the formation of waste products.
Advanced computer control systems are capable of responding to process upsets and
product changes swiftly and smoothly, producing a minimum of unwanted by-products
(see section on PROCESS CONTROL).
Examine Heating and/or Cooling Techniques
The technique for heating or cooling the reactor needs to be examined, especially to
avoid hot-spots or overheated feed streams, both of which cause unwanted by-products.
HEAT EXCHANGERS
Heat exchangers can be a source of waste, especially with products that are tempera-
ture sensitive. There are a number of techniques for minimizing the formation of waste
products in heat exchangers. All reduce the amount of fouling.
Desuperheat Plant Steam
High pressure plant steam may contain several hundred degrees of superheat. Desu-
perheating the steam when it enters a process (or just upstream of an exchanger) reduces
tube wall temperatures and actually in-
creases the effective surface area of the ex- 235 pslg
changer because the heat transfer coefficient Steam
of condensing steam is ten times greater 30 psig
than that of superheated steam. Steam
Use Lower Pressure Steam
When plant steam is available at fixed

pressure levels, a quick expedient may be to
switch to steam at a lower pressure, reduc-
ing tube wall temperatures.
Install a Thermocompressor
Another method of reducing tube wall

temperature is to install a thermocompres-
sor, shown in Figure 4. These relatively
inexpensive units work on an ejector princi-
ple, combining high and low pressure
steams to produce an intermediate pressure
steam. Variable throat models are available 50 pslg
that operate like control valves, automati- Steam
cally mixing the correct amounts of high and Figure 4. Thennocompressor
low pressure steam.
7
Use On-line Cleaning Techniques
On-line cleaning devices such as recirculated sponge balls and reversing brushes
have been on the market for many years. In addition to reducing exchanger mainte-
nance, they also keep tube surfaces clean so that a lower temperature heat source can be
used. Processes benefiting from close temperature approaches (e.g. refrigeration and
condensing steam turbine systems) gain from on-line cleaning.
Use Scraped-wall Exchangers
Scraped-wall heat exchangers consist of a set of rotating blades inside a vertical, cy-
lindrical jacketed column. They can be used to recover saleable products from viscous
streams. A typical application is to recover monomer from polymer tars.
Use Staged Heating
If a heat-sensitive fluid must be heated, staged heating can lessen degradation. Be-
gin, for example, with low level waste heat, then use low pressure steam, and finally
desuperheated intermediate pressure steam. See Figure 5 below.
Low High
Waste Pressure Pressure
Heat Steam Steam
l l l
S~n:~~~e
Fluid
-(,-+------+-,1H,-+-I-----+-,1H,-+-I-----+-,I)-
Figure 5. Staged heating
Monitor Exchanger Fouling
Exchanger fouling does not necessarily occur steadily. Sometimes, an exchanger

fouls because of a change in operating conditions or because of an upset elsewhere in the
process. The cause of fouling can often be found by continuously monitoring the fouling
factor and correlating sudden increases in fouling to operational changes.
Use Non-corroding Tubes
Corroded tube surfaces foul more quickly than non-corroded tube surfaces. Chang-
ing to non-corroding tubes can significantly reduce fouling.
PUMPS
Pumps don't usually contribute to waste except in two areas:
Recover Seal Flushes and Purges
Each pump seal flush and purge needs to be examined as a possible source of waste.
Most can be recycled to the process with little difficulty.
8
Use Seal-less pumps
Leaking pump seals lose product and create environmental problems. Using can-
type seal-less pumps or magnetically driven seal-less pumps eliminates these losses.
FURNACES
Furnace technology is constantly evolving and represent one of the most fertile areas
for efficiency improvement. Furnace manufacturers should be contacted for the latest
techniques in optimizing furnace operation and reducing tube fouling. Although furnaces
are often advertised as having high efficiency (e.g. 80%), it should be recognized that
burning fuel is inherently inefficient, and the true thermodynamic (2nd law) efficiencies
of industrial furnaces are much lower (normally 20-40 %).
Replace Furnace with Intennediate Exchanger
Another option is to elimi-

nate direct heating in the fur-
nace (which necessarily exposes
the heated fluid to high tube
wall temperatures) by using an
intermediate heat transfer me~
dium (e.g. Dowtherm®). This
is illustrated in Figure 6 at the Dowtherm™
right. Heat-
Sensitive
Replace Coil 1+-_ _.,........ \.....:::::....".~- Fluid
In certain applications,
significant improvements can Figure 6. Intermediate heat transfer medium
be made by replacing the exist-
ing furnace coil with one hav-
ing improved design features
(e.g. tubes with low residence
time or designed for split flow). Although it may not be practical to replace an undam-
aged coil, alternative designs should be considered whenever replacement becomes nec-
essary.
Use Existing Steam Superheat
¥=
475 psig steam
Even though the temperature required at 700°F
may be above the saturation temperature of
plant steam, there may be sufficient super- 50 F0
Hea t -
heat available in the steam to totally elimi- Sensitive
nate the need for a furnace. This is illus- Fluid
trated in Figure 7. This may also be an effi-
cient way to desuperheat plant steam as sug-
465 psig steam
gested earlier. Many people are not even at 650°F
aware that their plant steam contains consid-
Figure 7. Steam superheat replaces furnace
erable superheat.
9
DISTILLA TION COLUMNS
Distillation columns typically contribute to waste in three ways:
A. By allowing impurities to remain in a product. The impurities ultimately become

waste. The solution: better separation. In some cases it may be desirable to exceed
the normal product specifications.
B. By forming waste within the column itself, usually because of high reboiler tempera-
tures which cause polymerization. The solution: find ways to lower column tem-
peratures.
C. By inadequate condensing, which results in vented or flared product. The solution:

improve condensing.
In the following paragraphs, we'll look at some column and process modifications
that reduce waste by attacking one or more of these three problems.
Increase Reflux Ratio
The most common way of improving separation is simply to increase the reflux ra-
tio. This raises the pressure drop across the column and increases the reboiler tempera-
ture (using additional energy), but is probably the simplest solution if column capacity is
adequate.
Add Section to Column
If a column is operating close to flood, a new section can be added to increase ca-
pacity and improve separation. The new section can use trays, regular packing or high
efficiency packing. It need not be the same as the original portion of the column.
Retray or Repack Column
Another method of increasing separation is to retray or repack part or all of the col-
umn. High efficiency packing is available and has the added advantage that it also low-
ers pressure drop through the column, decreasing the reboiler temperature.
Change Feed Tray
Don't overlook changing the feed tray. Many columns are built with multiple feed
trays, but the valving is seldom changed. In general, the closer the feed conditions are to
the top of the column (high lights concentration and low temperature), the higher the
feed tray; the closer feed conditions are to the bottom of the column (high heavies con-
centration and high temperature), the lower the feed tray.
Insulate
Good insulation is necessary to prevent heat losses. Poor insulation requires higher
reboiler temperatures and allows column conditions to fluctuate with weather conditions.
During a storm, inadequately insulated columns may cool rapidly, sending light compo-
nents in the column bottoms stream, ultimately producing waste.
10
Improve Feed Distribution
The effectiveness of feed distributors (particularly in packed columns) needs to be

analyzed to be sure that distribution anomalies are not lowering overall column effi-
ciency. As with reactors, poor distribution in columns results in inefficiencies which be-
come accepted with time.
Preheat Column Feed
Preheating the feed should improve column efficiency. Supplying heat in the feed
requires lower temperatures than supplying the same heat to the reboiler, and also re-
duces the reboiler load. It can often be done by cross-exchange with other process
streams.
Remove Overheads Product from Tray Near Top of Column
If the overheads product contains a light impurity, it may be possible to obtain a

higher purity product from one of the trays close to the top of the column. A bleed
stream from the overheads accumulator can be recycled back to the process to purge the
column of lights. Another solution is to install a second column, which may be expen-
sive but well worth doing.
Increase Size of Vapor Line
In low pressure or vacuum columns, pressure drop is especially critical; installing a

larger vapor line reduces pressure drop and decreases reboiler temperature.
Modify Reboiler Design
While thermo syphon reboilers are the most common, they may subject the product
to tube-wall temperatures which cause product degradation. Other designs are available
such as falling film and pumped recirculation reboilers. High flux tubes can sometimes
be retrofitted to an existing reboiler allowing use of lower pressure steam.
Reduce Reboiler Temperature

The same general temperature reduction techniques discussed earlier (in the section
on heat exchangers) such as desuperheating steam, using lower pressure steam, installing
a thermocompressor, using an intermediate heat transfer fluid, etc., also apply to the re-
boiler of a distillation column.
Lower Column Pressure
Reducing the column pressure will also decrease the reboiler temperature and may
favorably load the trays or packing as long as the column stays below flood. The over-
heads temperature, however, will also be reduced, which may create a condensing prob-
lem (see following paragraph).
Improve Overhead Condenser
If overheads are lost because of an undersized condenser, consider retubing, replac-

ing the condenser, or adding a supplementary vent condenser to minimize losses. It may
also be possible to reroute the vent back to the process (if process pressure is stable). If
11
a refrigerated condenser is used, be sure to keep the tubes above 32°F if there is any
moisture in the stream.
Improve Column Control
The comments in the Reactor section about closer process control and controlling at
the right point apply to distillation columns as well as reactors.
Forward Vapor Overheads to Next Column
If the overheads stream is sent to another column for further separation, it may be
possible to use a partial condenser and introduce a vapor stream to the downstream col-
umn.
Before any equipment modifications are undertaken, it is recommended that a com-
puter simulation be done and that a variety of operating conditions be examined. If the
column operating temperature or pressure changes, equipment ratings should also be re-
examined.
PIPING
Something as seemingly innocuous as plant piping can sometimes cause waste, and
simple piping changes can result in a major reduction of waste. Consider the following
piping changes to a process:
Recover Individual Waste Streams
In many plants, various waste streams are combined and sent to a waste treatment
facility as shown in Figure 8. Each waste stream needs to be considered individually.
The nature of the impurities may make it possible to recycle or otherwise reuse a particu-
lar stream before it is
mixed with other waste
streams and becomes un-
recoverable. Stripping,
Waste Waste
filtering, drying or some
B 0
! i ! i·
other type of treatment
may be necessary before Combined
the stream can be reused. Waste - - - - ' L - -....-.....I"---"'T"""---i~
A Waste
Condensate used for seal Stream
flushes are an example of Waste Waste
a stream that does not re- C E
quire a "pure" stream. Figure 8. Recover individual waste streams
Simply filtering a waste
water stream to remove
particulates may be ade-
quate.
Check Pipe Sizes
Undersized pipe, especially in low pressure gas service (e.g. the inlet to a compres-
sor) can be costly. All low pressure gas piping should be checked for optimum flow and
pressure drop. Also look for obstructions, restrictions or undersized valves in lines.
12
Avoid Overheating Lines
If a process stream contains temperature sensitive materials, the temperature level of

line and vessel tracing systems and and/or vessel jacketing needs to be reviewed. If plant
steam is hot enough to cause product degradation, a recirculated warm fluid can be used
to prevent freezing. Tracing systems that use a recirculated fluid heated in a central ves-
sel consume much less energy than typical steam tracing systems. Electric tracing is also
an option.
Avoid Sending Hot Material to Storage
If a temperature sen-
sitive material is sent to storage
it should first be cooled. If this
is uneconomical because the
stream from storage needs to be
heated when it's used, simply Hot
Intermediate
Process
piping the hot stream directly Storage
Stream
into the suction of the storage Tank
tank pump as shown in Figure
9 may solve the problem ade-
quately. Make sure that the
storage tank pump can handle Figure 9. Reducing storage tank.heat losses
hot material without cavitating
and that material in the storage
tank does not degrade.
Eliminate Leaks
Leaks can be a major contributor to a plant's overall waste, especially if the prod-
ucts can not be seen or smelled. A good way to document leaks is to measure the quan-
tity of raw materials that must be purchased to replace "lost" streams (e.g. the amount of
refrigerant purchased).
Change Metallurgy
The type of metal used for vessels or piping may be causing a color problem or be
acting as a catalyst for the formation of unwanted by-products. If this is the case, change
to more inert metals (see next paragraph).
Monitor Vent And Flare Systems
The measurements need not be highly accurate, but should give a reasonable esti-
mate of how much product is lost and when losses occur. It is often worthwhile to install
whatever piping is necessary to recover products that are vented or flared. Additional
purification may be needed before the recovered streams can be reused.
Use Lined Pipes or Vessels
Using lined pipes or vessels is often a cheap alternative to using exotic metallurgy.
A variety of coatings are available for different applications. FRP vessels are also an al-
ternative.
13
PROCESS CONTROL
Modem technology allows us to install highly sophisticated computer control sys-

tems that respond far more quickly and accurately than human beings. We can use that
capability to reduce waste. Some suggestions are summarized on the following pages.
Improve On-line Control
Good process control reduces waste by minimizing cycling and improving a plant's
ability to handle normal changes in flows, temperatures, pressures and compositions.
Statistical quality control techniques help plants analyze process variations and document
improvements. Sometimes additional instrumentation or on-line stream monitors (e.g.
gas chromatographs) are necessary, but good control optimizes process conditions and re-
duces plant trips, often a major source of waste.
Optimize Daily Operation
If a computer is incorporated into the control scheme, it can be programmed to

analyze the process continuously and optimize operating conditions. If the computer is
not an integral part of the control scheme, off-line analyses can be performed and used as
a guide for setting process conditions.
Automate Startups, Shutdowns and Product Changeovers
Huge quantities of waste can be produced during plant startups, shutdowns and
product changeovers, even when such events are well planned. Programming a computer
to control these situations brings the plant to stable operating conditions quickly and
minimizes the time spent generating off-spec product. Further, since minimum time is
spent in undesirable running modes, equipment fouling and damage are also reduced.
This is an excellent place to use input from experienced plant operators.
Program Plant to Handle Unexpected Upsets and Trips
Even with the best control systems, upsets and trips occur. Not all can be antici-
pated, but operators who have lived with the plant for years probably remember most of
the important ones and know the best ways to respond. With computer control, optimum
responses can be pre-programmed. Then, when upsets and trips occur, the computer
takes over, minimizing downtime, spills, equipment damage, product loss and waste.
MISCELLANEOUS
In addition to the ideas already discussed, there are a number of other miscellaneous
improvements that can be made to reduce waste:
Avoid Unexpected Trips and Shutdowns
A good preventative maintenance program and adequate sparing of equipment are

two keys to minimizing trips and unplanned shutdowns. Another key is to provide an
adequate warning system for critical equipment (e.g. vibration monitors). Plant opera-
tors can be extremely helpful by reporting unusual conditions so that minor maintenance
problems get corrected before they become major and cause a plant trip.
14
Use Waste Streams from Other Plants
Within a chemical complex, each plant's waste streams should be clearly identified.
The quantity and quality of these streams should be documented, including the presence
of trace quantities of metals, halides or other impurities that render a stream useless as a
raw material. This list should be reviewed by all plants in the complex to determine if
any are suitable as feedstocks.
Reduce Number and Quantity of Samples
Taking frequent and large samples can generate a surpnsmg amount of· waste.
Many plants find that the quantity can be reduced or that the samples can be returned to
the process after analysis.
Recover Product from Tank Cars and Tank Trucks
Product vented or drained from a tank car or tank truck (especially those dedicated
to a single service) can often be recovered and reused.
Use Removable Insulation
When conventional insulation is removed from equipment it is typically scrapped

and sent to landfill. A number of companies manufacture reusable insulation. Their
products are particularly effective on equipment where the insulation is removed regu-
larly in order to perform maintenance (e.g. heat exchanger heads, manways, valves,
transmitters, etc.)
Maintain External Painted Surfaces
Even in plants handling highly corrosive materials, external corrosion can be a ma-
jor cause of pipe deterioration. Piping and vessels should be painted before being insu-
lated, and all painted surfaces should be well-maintained.
Find a Market for Waste Products

Converting a waste product to a saleable product may require some additional proc-
essing, but can be an effective means of reducing waste. Converted product, however,
should not create a waste problem for the customer.
CONCLUSIONS
Reducing waste is a never-ending activity. Each year, we are faced with new de-
mands, new regulations, new legislation, new challenges and of course, new problems.
Amidst all this activity, we never want to lose sight of one extremely important concept:
The way you reduce waste is by installing

projects that reduce waste.
15
You can state corporate positions, write detailed mission statements, outline ambi-
tious goals, and prepare detailed plans, but:

You can give dynamic speeches, publicize inspiring slogans, develop lofty mottoes,
design award-winning posters, and distribute multi-colored brochures, but:

You can establish high-level committees, appoint technical task forces, organize
environmental teams, and restructure entire departments, but:

You can meet with environmentalists, negotiate with regulators, confer with politi-
cians, talk to community leaders, and address the general public, but:

You can write letters to the newspaper, be interviewed on radio, appear on TV,
speak at technical meetings, lecture at symposia, and even testify before Congress, but:

By now, I hope my message is clear! A successful waste reduction program is more

than talk; it must get results. It must raise employee awareness and concern about envi-
ronmental issues to the point where specific ideas for projects to reduce waste are con-
ceived, designed and implemented.
16
As a closing thought (and additional impetus to action), recognize that waste reduc-
tion is not synonymous with lower profits. You might be surprised to learn that nearly
every idea listed in this paper had a Return on Investment-an ROI-greater than 30%!
Most paid for themselves in less than a year. Some required no capital expenditure.
Reducing waste is not just responsible management, it's good business!
Note: A version of this paper was originally presented at "Pollution Prevention for the 1990's:
A Chemical Engineering Challenge" December, 1989, Washington, D.C., and was subse-
quently published in Hydrocarbons Processing (March, 1990) and Chemtech (August,
1990).
C The Dow Chemical Company

March, 1992
Used with permission
17
W ARR - IMPLEMENTING A WASTE AND RELEASE REDUCTION STRATEGY
James A. Conner
Hoechst Celanese Corporation

Route 202-206
Somerville, NJ 08876
INTRODUCTION
Hoechst Celanese Corporation is embarked upon a Corporate-wide Waste and

Release Reduction program, which we have given the acronym lOWARR". W ARR is a
comprehensive program for achievement of major reductions across the corporation of all
emissions and waste materials.
The WARR program is a part of our Vision of Excellence effort. Vision of

Excellence is a Hoechst Celanese initiative aimed at integration of environmental, health
and safety considerations into everything we do as a Corporation. The Vision of
Excellence is complementary to the CMA Responsible Care" initiative.
We have defined eight end states within Vision of Excellence - one of which is the
Release Reduction/Waste Minimization End State (Table 1).
Table 1. Hoechst Celanese Vision of Excellence End States
HOECHST CELANESE
VISION OF EXCEUENCE END STATES
• EMPI.OVEE AND COMMUNITY 0UTREACt1

• RElEASE REDUC'TlON/WASTE IotlNNlZA11ON
• PROCESS SAFETY
• EMPI.OVEE HEALTH AHD SAFETY
.~SAFETY
• PIIOOUCT SAFETY
• EMPLO'tEE WELLNESS
=-
• INFORMATION SYSTEIoi
L-_____________________________ - - - _ .••
The W ARR program falls within, and is the action plan for, this Release Reduction/Waste
Minimization end state.
01992 Hoechst Celanese Corporation

and A.E. Martell, Plenum Press, New York, 1992 19
OVERVIEW OF HOECHST CELANESE CORPORATION
Hoechst Celanese Corporation was formed in 1987 by the merger of the former
Celanese Corporation and the former American Hoechst Corporation. It is a wholly owned
subsidiary of Hoechst AG of Frankfurt, Germany. The Corporation is a major US
manufacturer, with worldwide sales, of fibers, chemicals, plastics, and pharmaceuticals.
At present, the company has 32 major US locations with approximately 23,000 employees.
Our 1990 sales totalled roughly $6 billion. Major business units are located in Texas
(Chemicals and Plastics), the Carolinas and Virginia (Fibers and Film), and New Jersey
(Pharmaceuticals).
WHAT IS WARR
WARR is a corporate-wide program focused on major reductions of emissions and

waste. It addresses emission to all media - air, water, and land.
There are 3 primary goals of the W ARR effort:
1. An 80% reduction (vs. 1988 base year) of all types of discharges and
releases (excluding combustion products from energy production, treated
process wastewater and non-contact cooling water) by the end of 1996
2. A 70% reduction (vs. 1988 base year) of releases of SARA Title III,
Section 313 chemicals by the end of 1996
3. Discontinue the use of all underground injection wells for waste disposal by
the end of 1996
Hoechst Celanese Corporation was also one of the early signatories to the US
EPA's Industrial Toxics Program (ITP), which is also commonly referred to as the "33/50
Program". The ITP targets 17 priority chemicals (all SARA) for a 33 % reduction by end
of 1992 and a 50% reduction by the end of 1995 (all vs. a base year of 1988).
Hoechst Celanese has reportable emission quantities of 12 of the 17 targeted

chemicals. The goals of the ITP are in concert with, and complementary to, our own
WARR program goals. The projects which are planned to meet our end-of-1996 WARR
goals will also assist us in meeting the 199550% reduction target set by EPA for these 17
priority chemicals.
The initial phase of the WARR effort focuses on nine major facilities (out of 32).
There are presently in excess of 180 projects on the WARR program list.
The anticipated cost of the program is $500 to $600 million. Funding for the
W ARR effort is separate and in addition to the Corporation's normal annual capital budget.
This W ARR spending amounts to roughly 114 - 113 of our Corporation's total capital
spending in the years 1991-1995. This substantial expenditure is viewed by our
Corporation as an investment in the future, as defined by our Vision of Excellence.
PRIORITIES OF THE W ARR PROGRAM
Emission reduction planning and prioritizing of projects is guided by our "Waste

and Release Reduction Hierarchy", a prioritization list for reduction alternatives which
20
SARA EMISSIONS REDUCTION
OTHER SOlR'E AECOIIERV AS

REOUCllON ~ PRODUCT
-~~
~~TREATMENT
:In;
L..._ _ _ _ _ _ _ _ _ - _
Figure 1. Sara Emission Reductions
Table 2. Release Reduction Hierarchy
RELEASE REDUCTION HIERARCHY

• RECOVER AS PRODUCT
• SOURCE REDUCTIOH
• REC'/Ct..E AND REUSE
• RECOVER AS ENeRGY
• TREATMENT
1.-_ _ _ _ _ _ _ _ --:"___
.•
encourages source reduction and recycle/reuse over end-of-pipe treatment (fable 1).
Our organization has been challenged to develop release reduction projects and solutions
which are as high on this hierarchy as can practically be achieved. This challenge has
resulted in considerable success as both source reduction and recycle/reuse figure
prominently in our planned reductions. Roughly half of our total program effort falls into
the source reduction/recycle-reuse categories. Further, nearly 2/3 of our planned SARA
reductions are source reductions or recycle/reuse, and of this total, 1/3 is through recovery
of the emissions as saleable product (Figure 1).
ORIGIN AND DEVEWPMENT OF TIlE PROGRAM
The WARR program is rooted in our Corporate Values, which were established at
the time of the merger. We utilize the Values to help develop visions, and from those
visions we establish plans for implementation.
One of the Values is "Commitment to Safety, Employee Health and Protection of

the Environment". Our vision for environmental excellence was to be recognized as a
leader in environmental stewardship by our toughest critics - our employees, their families,
and the communities in which we operate. Implicit in this vision was recognition that
leadership requires that you commit to the proactive mode of dealing with environmental
issues.
21
The initial step in implementing the Vision was formation in 1989 of the Release
Reduction Task Force. This group developed a comprehensive inventory of emissions and
wastes from each of our locations. They then prioritized the issues to be addressed, and,
using the 80120 principle, the Group recommended an initial focus on nine major facilities.
A Release Reduction Strategy Committee was then appointed, led by the President
of one of our Operating Groups. This committee was chartered to develop a ~
strategy to achieve significant reductions in our emissions and wastes, and to identify the
technology and resource needs required from the organization to achieve that strategy.
The committee challenged each of the targeted facilities to attempt to assemble a

zero-discharge project list - to approach from an "emission elimination" mindset as
opposed to simply "emission reduction". For reasons covered later, the zero-discharge
project list was not achievable, but the value was in the ap,proach of zero-based thinking.
This represented a significant advancement beyond typical thinking, which is normally only
incremental in nature. In retrospect, this approach was of Yiml importance to the
development of our current program for several key reasons:
• It created a mind-set that this was not just an incremental program

• It forced a long-term perspective
• It forced a source-reduction perspective
• It forced creativity and original solutions
The zero-discharge list was not achievable. The rough estimate for the aggregate
initial lists for the targeted facilities was in the billion of $'s, and in many cases, total
elimination of the emission source would require development of unproven or unavailable
technologies. However, judicious selection from these original lists yielded a project list
which accomplished major reductions of emissions, emphasized source reduction over end-
of-pipe treatment, and was achievable in a reasonably rapid time frame. This carefully
selected strategy and project list was approved by the Corporation's Executive Committee
(CEO and Group Presidents) in October 1990.
TECHNOWGIES
To be sure that we are installing the most effective and appropriate technology for
the projects on the WARR list, we have developed mechanisms for technology transfer
from a variety of sources. Interplant and interdivision committees have been set up to
share and coordinate common approaches and technology applications. We have also
benefitted from sharing of technologies with Hoechst in Germany. Other avenues which
we utilize for sharing of technology and solutions include industry groups, such as the
AIChE Center for Waste Reduction Technology, and liaisons set up with universities .•
The primary technologies for waste treatment are biological degradation and
incineration. In the biotreatment arena, the technology has shifted to enclosed above-
ground bioreactors, such as the Hoechst Biohoch design, to maximize energy efficiency
and minimize air emissions. Anaerobic systems will work on some wastes, and offer the
advantage of lower sludge yields and methane recovery. A few technology firms offer
specialized micro-organism cultures which can be utilized to digest specific compounds
which are otherWise toxic or refractory to mixed-bug treatment plants.
22
For incineration, we are seeing a trend toward more exotic stack gas cleanup
systems, and toward fluidized bed or rotary kiln designs which are capable of handling
wastes from a variety of media - solid, liquid, and gaseous. For some applications, we
are developing catalytic incineration technology, which has the advantage of lower
temperatures and reduced NOx generation compared to typical thermal incinerators.
For source reduction, the key is not so much new technologies as it is a

reapplication and refocus of basic chemical process fundamentals and unit operations on
the minimization of wastes. Examples are improvements in reaction selectivity or
conversion, improved separations, higher purity feedstocks, solvent substitution, and in-
process recycling for yield improvement. One notable exception to this is improvements
in separations made possible by the rapidly evolving membrane technologies. Our
Separations Products Division develops and markets membrane systems for phase contact
(e.g. liquid extraction), ultrafiltration, and pervaporation (separation of liquid or vapor
streams via selective permeation and evaporation across a membrane). Separations
Products is working closely with our WARR effort to develop specific applications of these
technologies to enhance our source reduction efforts.
For the longer term, we have increased the focus of the Hoechst Celanese R&D
Community on source prevention. Pollution Prevention begins in the development lab,
where the opportunity exists for toxic use reduction and the development of "green"
chemistry and processes. Another leverage point for source prevention is our project
teams - the process design engineers and the plant layout engineers play a key role in
source prevention through process and plant designs which have fewer leak points, better
separations, and plant layouts which minimize waste during equipment cleaning and
maintenance.
REFINEMENT OF ORIGINAL STRATEGY
As work began on definition and development of the original WARR project list,
the technical staffs at the facilities began to find better alternatives to many of the projects.
In late 1991, we completed a substantial optimization of the project list which increased
our projected reductions of emissions and also shifted more of the projects into the source
reduction category. Many of the original end-of-pipe projects were replaced by multiple
smaller source reduction projects, thereby eliminating the need for treatment. In other
cases, treatment projects dropped from the list after operational changes reduced the source
and made treatment unnecessary. Further optimization of the project list remains a
priority, and we expect to continue the trends of additional emissions eliminations, and
movement toward greater source reduction.
MANAGEMENT OF TIlE PROGRAM
Responsibility for oversight and direction of the program is vested in the Waste and
Release Reduction Implementation Steering Committee. This group is chaired by the
President of one of our largest operating divisions, Technical Fibers. Members are a
cross-sectional group of the Corporation (plant managers, environmental directors, division
VP's, etc.). The Steering Committee has responsibility for:
• Tracking vs. objectives

• Assure best technology
23
• Consistency with long term environmental goals
• Grow program beyond initial focus to encompass major reductions at all
facilities
WARR Coordinators were named at each site. These are predominantly technical
people, many with backgrounds in Operations, Process Engineering or Project
Management. The strengths and skills of these people have direct bearing upon the amount
of source reduction that will be achieved. To catalyze source reduction, you need to
involve these people who best know the process - design engineers, process engineers and
chemists, and operations personnel. The results we have achieved with the optimized
project list proved the value of this approach.
At the group level, additional people were appointed to assist the plantsites to
crossfeed and coordinate shared solutions. A good example is our Film and Fibers Group,
where we have a Vice President of Waste and Release Reduction, a Director of WARR
projects to coordinate plant activities, and a focus person for technology transfer between
divisions and sites. At the Corporate level, the position of Manager, Environmental
Projects was created to provide overall coordination and management across all divisions.
RESULTS AND FORECASTS
The WARR program is on track to achieve all stated objectives by the end of 1996.
Our disposal well elimination projects are on track, and we project a substantial reduction
in all emissions over the next 3-4 years. We have already achieved significant reductions
in SARA releases. As a Corporation, our SARA releases declined roughly 20% from 1988
to 1990. In addition, we are projecting additional SARA release reductions of
approximately 10% annually for the next five years, achieving our 70% reduction target
by the end of 1996.
SUCCESS FACTORS
Unwavering management commitment from the top of the Corporation has played
a major role in our program's success. This commitment held steady in spite of the
economic downturn which beset the industry. In addition, champions for the program
emerged in key positions within the company to help eliminate barriers and assemble the
necessary resources.
Plantsite participation in development of the strategy was also key. Early input to
the project list established an ownership at the plantsites to the project goals, the overall
strategy, and the specific contribution needed from each facility.
Finally, the WARR effort is not perceived as simply a program from the
Environmental organization. Through our Vision of Excellence, we have moved Waste
and Release Reduction into the mainstream of our business objectives and plans - a way
of doing our business, and an integral part of our long term planning.
24
WASTE MINIMIZATION BY PROCESS MODIFICATION
J. R. Hopper, C. L. Yaws, T. C. Ho, M. Vichailak and A. Muninnimit
Chemical Engineering Department

Lamar University
Beaumont, Texas
INTRODUCTION
Waste minimization means the reduction, to the extent feasible of hazardous waste that
is generated prior to treatment, storage or disposal of the waste. The terms waste
minimization and pollution prevention are becoming key initiatives of the United States
Environmental Protection Agency (U.S.EPA) and the term is increasingly being used in all
areas of current applied research and development [15] .
Waste minimization has been defined to consist of two basic techniques: source
reduction and recycling [14]. See Figure 1. Source reduction or avoiding waste generation
is the most desirable goal and should be explored first [16]. In the waste management
hierarchy, source reduction is followed in order by recycling, treatment and, finally,
disposal. Although such processes as incineration, stabilization and/or storage are clearly
preferable to land disposal, they are not included in the definition of waste minimization.
The total concept of pollution prevention is that it is much more desirable not to produce
waste rather than to implement extensive treatment schemes to insure that the quality of the
environment is not damaged. Under the category of source reduction, there are two sub-
categories: source control and product substitution. Source control can be accomplished by:
(1) good housekeeping practices, (2) input material modification, and (3) technology
modification. EPA has reported that, generally, technology modifications have been found
to be the most effective means of reducing waste generation [38]. Technology modifications
have been categorized as follows:
• Process Modifications
• Improved Controls
• Equipment Changes
• Energy Conservation
• Water Conservation
The need for waste minimization studies has become apparent to the Chemical
Processing Industries (CPI) because Congress initially declared waste minimization to be the
national policy of the United States that, wherever feasible, the generation of hazardous
waste is to be reduced or eliminated as expeditiously as possible [35]. The Clean Air Act
Industrial Environmental Chemistry. Edited by D.T. Sawyer

and A.B. Martell. Plenum Press. New York, 1992 25
(CAA) and the Superfund Amendments Reauthorization Act, Title 313 which includes the
Toxics Releases Inventory (TRI) list will clearly impact the drive to implement this concept
which has been given a high priority by Congress.
The objective of this project is to develop examples of waste minimization by process
modifications which can be achieved by modification of chemical process reaction and
separation parameters. These examples can be used as guidelines for process and
environmental engineers to follow in suggesting process modifications which will minimize
waste generation.
I WASTE MINIMIZATION TECHNIQUES

I
RECYCLING SOURCE REDUCTION
ONSITE / OFFSITE
~~ ~D~
USE/REUSE RECLAIM
. Process to recover
• Ingredient in a process usable product
I I
- effective substitute • Regeneration
SOURCE CONTROL PRODUCT

SUBSTITUTION
I~ .. ~:,".jl Ei1~Ei1
GOOD HOUSEKEEPING INPUT MATERIAL TECHNOLOGY

PRACTICES MODIFICATION MODIFICATION
- Waste stream segregation - Input purification · Improved Control
- Inventory control • Input substitution • Proeese Modification
• Employee training • Equipment Chang..
o Spill/leak prevention • Energy Conservation
• Schduling improvement • Water Conservation
Figure 1. Waste Minimization Techniques Diagram
PROCESS MODIFICATION CONCEPTS
Process modifications or changes mean: (1) the use of alternative low-waste process
pathways to obtain the same product, and (2) modification of reaction or separation
parameters.
Multiple chemical reactions can be classified into three general reaction types [10, 12,
20, 24, 30]: (1) series, (2) parallel and (3) combinations of series and parallel reactions. It
is the many possible reactions which generate many of our waste products. Generally, we
have desirable products and undesirable products. The undesirable products may become
waste streams. The ratio of the desirable products to the undesirable products is generally
defined as the selectivity of the reaction. This selectivity can be influenced by the choice
of the reaction parameters or by modification of these reaction parameters.
Waste minimization by modification of reaction parameters is defined as changing the
selectivity of the reaction so that undesirable reactions which produce waste products are
minimized, while at the same time producing the desirable products. This selectivity may
be affected by all of the reaction conditions. These reaction conditions include: temperature,
pressure, catalyst, mixing conditions, ratio of feed concentrations and conditions, reaction
26
type and reactor design options. Although catalyst changes are, without a doubt, the most
effective method of changing a reaction selectively [26, 31], catalyst changes are very
specific for each process and are in most cases highly proprietary. However, changes in the
other reaction parameters have more general applications for many processes and can have
an impact far greater than any of our previous expectations [29, 2].
This work summarizes the results of two studies of modification of the process
parameters for pollution prevention.
CASE STUDY FOR THE ACRYWNITRILE PROCESS
General studies of process modifications have been made in the past [7,25, 34]and an
earlier report was presented for the allyl chloride process on our preliminary results [21].
Recently two specific cases have been reported for the effect of modification of reaction
parameters [29, 2] for industrial processes.
This case study is an example of how alteration of reactor design parameters can
reduce waste generation for the arylonitrile process.
Process Description
Since the commercialization of the Sohio process in the early 1960's, all acrylonitrile
plants built in the world have been based on the reaction between propylene, ammonia, and
air [18, 19, 22, 32, 36, 37, 38] . Acrylonitrile is the product from the gas phase catalytic
air oxidation of propylene and ammonia. The reaction chemistry for the ammoxidation is
given below:
Propylene Ammonia Oxygen Acrylonitrile Water
CH2 =CH-CH3 + NH3 + 3/2 Oz ~ CH2 =CH-CN + 3HzO (1)
with the combination of five others possible side reactions
Acrolein
CH2 =CH-CH3 + Oz ~ CH2 =CH-CHO + H2 0 (2)
Acetonitrile
CH2 =CH-CH3 + NH3 + 9/402 ~ CH3-CN + 1/2 CO2 + 112 CO + 3H20 (3)
CH 2 =CH-CHO + NH3 + 112 Oz ~ CHz=CH-CN + 2H20 (4)

Hydrocyanic acid
CH 2 =CH-CN + 202 ~ CO2 + CO + HCN + H20 (5)
CH3-CN + 3/2 O2 ~ CO 2 + HCN + H20 (6)
The catalyst used for the reaction in the early process version was Bi-Mo-O [38]. The
current commercial catalyst is of unknown composition with antimony-tellurium catalyst
manufactured by Nitto Corporation being one possible catalyst. The flow diagram for this
process is shown in Figure 2. Approximately stoichiometric quantities of propylene,
ammonia, and oxygen (as air) are introduced to a fluidized-bed catalytic reactor at 15 psig
and the contact time in the range of several seconds and a temperature of 350 - 600 "C. The
use of polymerization grade propylene is unnecessary. Once through flow is used since
conversion of propylene is practically complete. The reaction is exothermic, so heat-removal
must be provided. The reactor effluent is neutralized to remove unconverted ammonia. It
is next washed with water to yield an unabsorbed stream of inert gases and the solution of
acetonitrile, acrylonitrile and HCN. This solution is stripped of the dissolved product which
are fractionated to remove pure HCN and then sent to the main purification section. A main
27
fractionator yields an overhead consisting of wet acrylonitrile which is dried and purified
by azeotropic and conventional distillation. The bottoms wet acetonitrile is similarly
concentrated, dried, and converted to high purity product [36, 38].
Acrolein, acetonitrile, and HCN are by products and potential wastes. This study has
evaluated the reaction parameters shown in Table I to determine the conditions which will
minimize the production of by-products while maintaining the production of acrylonitrile.
Reaction System Analysis
The rate equation for the ammoxidation of propylene over Bi-Mo-O has a first order
dependence with respect to propylene and a zero order dependence for both oxygen and
ammonia when they are supplied in at least stoichiometric amounts [4, 5, 6, 9, 12,23,28].
Extro.ctive
HeN distillation Acetonitrile
Reactor Neutralizer Absorber Stripper coluMn cotul"ln pur/fica tion
To waste
Acetonitrile
Acrylonitrile
pur/fiCO tion To Waste
,..-------
Acrylonitrile
To waste
Figure 2. Acrylonitrile Process Flow Diagram
Table 1. The Reaction Parameters.
Reactor Type PFR I CSTR I FBR

Reactor Condition Temperature
Residence Time
Feed Condition Entering Feed Temperature
Kinetic Model Six Reactions
28
The rate equations are:
(-r l ) kICc,H, (7)
(-r2) ~CC,H, (8)
(-r3) k3Cc,H, (9)

(-r4) k4CcH,CHCHO (10)
(-rs) ksCCH,CHCN (11)
(-r6) k6CCH,CN (12)
where the rate constants are expressed in the Arrhenius form by:
(13)
The reaction system constants are shown in Table 2.
Table 2. The Reaction System Constants.
Reaction System Constants

Reaction Activation Energy, Ei Rate constant, ki Heat of Reaction, tJIi
No. (Cal/mol) at 470 "C (sec-I) at 298 K (J/mol)
1 19,000 0.40556 -515,610
2 19,000 0.00973 -333,350
3 7,000 0.01744 -864,990
4 7,000 6.81341 -182,260
5 19,800 0.16222 -800,900
6 7,000 0.07300 -593,120
The heat of reaction at other temperatures can be obtained from:
(14)
using heat capacity values for each of the species j as given by:
(15)
The values for the constants are given in Table 3.
29
Table 3. The Heat Capacities Constants
Heat Capacities Constants
a {3, 'Y, z,
I. Nitrogen, N2 30.7S -1.I7SE-02 2.390E-OS -1.000E-OS
2. Carbon Monoxide, CO 30.S7 -1.2SSE-02 2.7S9E-OS -I. 272E-OS
3. Oxygen, O2 2S.11 -3.6S0E-06 I. 746E-OS -1.06SE-OS
4. Carbon Dioxide, CO2 19.50 7. 344E-02 -S.602E-OS 1.7ISE-OS
S. Propylene, C3H. 3.710 2.34SE-01 -1.I60E-04 2.20SE-OS
6. Ammonia, NH3 27.31 2.3S3E-02 1.707E-OS -I.ISSE-OS
7. Hydrogen Cyanide, HCN 21.S6 6.062E-02 -4.96IE-OS I.SISE-OS
S. Acrolein, C3H.0 11.97 2.106E-Ol -1.071E-04 1.906E-OS
9. Acrylonitrile, C3H3N 10.69 2.20SE-Ol -1.56SE-04 4.60IE-OS
10. Acetonitrile, C2H3N 20.4S 1.196E-Ol -4.492E-OS 3.203E-09
11. Water, H 2O 32.24 1. 924E-03 1.0SSE-OS -3.S96E-09
Cp
T
.. loule/g-mole K
K
'"
Stoichiometric Equation
The stoichiometric equation for each of the species j for the n independent reactions
is:
n
~ = FjO + FAO L /I~i
i-I
(16)
where "u is the stoichiometric coefficient for componentj in reaction i. FjO' Fj, and FAo are
the molar flowrates for component j initially and at any point in time and the primary
reactant A.
Xi is the extent of conversion for reaction i and is defined as:
(17)
where A~i is moles of speciesj consumed (formed) in reaction i.
Component Mass Balance
The steady state component mass balance for the PFR and CSTR is:
llj dV
v
~o - ~ + = 0 (18)
o
For q simultaneous reactions, n of which are independent, 1j is defined as:
30
(19)
where rij is the rate of formation of species j in reaction i. The bubbling bed model [8] has
been used for the fluidized bed reactor. This model has a rising bubble surrounded by a
cloud-wake and enclosed by the emulsion catalyst bed. Mass transfer and reaction proceeds
in each of these phases according to the following component mass balance relations for
reactant A.
(20)
(21)
(22)
Total Energy Balance
The total energy balance for n independent multiple reactions involvingj species are:
For the CSTR and for the FBR.
T m •
Q - lV, - FAO L, j~ 8 j CPj dT - FAO i~ Mlri(T) Xi = 0 (23)
where 8j = ~J FAO : ratio of the initial molar feed rate for each specie j to the initial molar
feed rate of primary reactant.
For the PFR.
where Q = UA(T, - 1) and T. = temperature of the surroundings.

These models have been used to evaluate waste minimization by reaction parameter
modifications for this process.
DISCUSSION AND RESULTS
The major by-products (potential wastes) considered for the acrylonitrile process are
acetonitrile and HCN since the production of acrolein is very much less than either
acetonitrile (ACE) and HCN. An analysis of the reactor type, temperature, and residence
time has been done to establish the reaction conditions which will minimize the production
of these by-products and maintain the production of acrylonitrile (ACN).
A plot of the mole fraction for each of the reactants and products versus residence
time at 400 °C and 2 atm. for the fluidized bed reactor (FBR) is shown in Figure 3. In
Figure 3 for the FBR the reactant oxygen decreases from 18% in the feed to 4.7% after a
residence time of 20 seconds while the reactants ammonia and propylene decrease from 7%
in the feed to 1. 1% after a residence time of 20 seconds. Water is a significant product and
reaches a value of 19 mole %, acrylonitrile reaches 3.6%, acetonitrile reaches 0.3%, HCN
reaches 1.8%, C02 reaches 2. 1% and CO reaches 1.9% for this same residence time. A
similar profile of reactants and products can be obtained for the PFR and CSTR. Similar
plots can also be generated at various temperatures.
31
PLOT OF KOLE FRACTION VERSUS RESIDENCE TIllE
OPERATING CONDmON : ISOTIIERIIAL. .00 ·C. 2 atm
REACTOR TYPE : n.UIDIZED BED REACTOR
0.25
0.20
e
~
0.15
~
51~ 0.10
0.05
0.00 ~:;;:;;~~~~==S~~.
0.00 5.00 10.00 15.00 20.00 25.00
RESIDENCE TIllE (S)
Figure 3. Plots of Mole Fraction versus Residence Time of Fluidized Bed Reactor.
4.5"
4.0
3.5
\
3.0
\
~ 2.5
\
~ 2.0
\
1.5 \
.......
"
1.0
0.5
O·~OO 250 300 350 400 450 500

TEMPERATURE C
r... 2SEC -+o-aSEC ~10SEC I

Figure 4. Plot of the Mole Ratio of ACE!ACN as a function of Temperature for Fluidized Bed Reactor
A comparison of waste generation with the variation in temperature and residence time
can be obtained from the cross section of Figures 3 at various times and temperatures, by
plotting the ratio of HCN/ACN, ACE!ACN, and the combination of both ([HCN +
ACE]/ ACN) with respect to the temperature and residence time for each of the reactor
types.
A plot of the mole ratio of ACE! ACN as a function of temperature for the FBR is
shown in Figure 4. This plot shows that the ratio of ACE! ACN decreases as the temperature
and residence time increase. The change in residence time has no significant effect on this
ratio. A similar profile of the plot shown in Figure 4 can be obtained for the PFR and the
CSTR
A plot of the mole ratio of HCN/ ACN as a function of temperature for the FBR is
shown in Figure 5. The plots for each of the reactor types shows that the ratio HCN/ ACN
32
2.0
1.8 r
1.6
I
1.4
7
I
~ 1.2
I
~ 1.0
0.8
I
0.6 /'
0.4 / /
0.2
77
-~ ~
O·~OO 250 300 3S0 400 4S0 500
TEMPERATURE C
Figure 5. Plot of the Mole Ratio of HCNIACN as a function of Temperature for Fluidized Bed Reactor.
4.S
4.0
3.S
\
~ 3.0
\
~ 2.S
'\
2.0
\
7
+
~ 1.S
\
7
-
1.0 --....
O.S
~ ~ .,...
~ :......,..-~
O'~OO 2S0 300 350 400 4S0 SOO

TEMPERATURE C
I"" 2 SEC -+-6SEC .... 'OSEC I

Figure 6. Plot of the Mole Ratio of (HCN+ACE)/ACN as a function of Temperature for Fluidized Bed
Reactor.
increases as the temperature and residence time increase. The minimum ratio occurs at a
temperature of about 300 °C, however the conversion of propylene at this temperature is
only 12.6% at a residence time of 10 sec. This indicates that a higher temperature and/or
residence time may be necessary.
A plot of total waste ([HCN +ACE]/ ACN) versus temperature and residence time is
shown in Figure 6 for the FBR. This plot shows that there is an optimum temperature for
minimizing the waste and it is in the range of temperature of 400 - 480°C for each type of
reactor. For this temperature range, a high conversion can be obtained by selecting the
correct residence time. A higher conversion can be obtained by choosing a longer residence
time but the ratio of waste to acrylonitrile produced will also increase.
33
Table 4. Reactor Comparison for Waste Generation at 400 °C, 2 atm at Constant Residence
Time
Reactor Residence % Mole Ratio

type time Conversion
(s) HCN/ACN ACE/ACN
2.00 19.97 0.04356 0.09767

PFR 10.00 66.88 0.26389 0.09565
15.00 80.72 0.44471 0.09442
2.00 19.29 0.08907 0.09728
CSTR 10.00 54.43 0.44404 0.09491
15.00 64.18 0.66527 0.09400
2.00 17.14 0.03600 0.09647
FBR 10.00 60.69 0.21180 0.09086
15.00 75.11 0.34930 0.08744
Table 5. Reactor Comparison for Waste Generation at 400 oC, 2 atm at Constant
Conversion
Reactor Residence % Mole Ratio

type time Conversion
(s) HCN/ACN ACE/ACN
0.90 10 0.02050 0.09820

PFR 4.25 40 0.10647 0.09695
9.90 70 0.29294 0.09544
0.95 10 0.04234 0.09789
CSTR 5.60 40 0.24901 0.09599
19.60 70 0.86948 0.09336
1.15 10 0.02016 0.09723
FBR 5.51 40 0.10590 0.09398
12.99 70 0.29020 0.08880
34
An exact comparison of the waste generation for each type of reactor can be made by
comparing the mole ratio of undesired product to the acrylonitrile (HCN/ ACN, ACE! ACN)
at the same residence time (as shown in Table 4.) and also at the same conversion of
propylene (as shown in Table 5.). These two comparisons were made at an operating
temperature of 400 °C and pressure of 2 atm.
A comparison of the mole ratio of HCN/ ACN and the mole ratio of ACE! ACN for
each of the reactors at constant residence times of 2, 10, and 15 seconds in Table 4 shows
that the FBR has the lowest values. The greatest differences in these ratios is for
HCN/ ACN. The highest ratio of HCN/ ACN occurs in the CSTR which also has the lowest
conversion. The mole ratio of both HCN/ ACN and ACE!ACN are higher with the PFR than
the FBR but the conversion in the PFR is also higher.
Table 5. shows the ratio of HCN/ACN and ACE!ACN at three different conversion
for three reactors. The residence time used for each reactor varies with the type of the
reactor. The FBR again generates less waste than other two reactors.
CONCLUSION
To minimize the by-products in the production of acrylonitrile by modification of the

reaction parameters in the process, the following conditions would be recommended:
Reactor Type : Fluidized Bed Reactor

Temperature : 400 - 480°C
Residence Time : 2 - 10 sec.
The reactor effluent from the operation of the reactor at 450 °C and the residence time
of 7 seconds is:
Component
Propylene 1.4
Ammonia 1.4
Oxygen 5.5
Acrylonitrile 3.6
Water 18.1
Acrolein 0.0
Acetonitrile 0.2
Carbon dioxide 1.8
Carbon monoxide 1.7
Hydrogen cyanide 1.6
and the conversion is 81.05%
CASE STUDY FOR ALLYL CHLORIDE PROCESS
This example illustrates how selection of reactor and separation design parameters can
reduce waste generation for the allyl chloride synthesis reaction. The production of allyl
chloride is the first step in the epichlorohydrin manufacture. This example is a classical case
of the combination reactions. Allyl chloride is produced from a non-catalytic reaction of
propylene and chlorine according to the following reaction:
Propylene Allyl Chloride
CH2=CH-CH3 + Cl2 CH2=CH-CH2Cl + HCl (25)
35
The by-products (or wastes) of this reaction include 1,2-dichloropropane and isomers
of 1,3-dichloropropene as shown in equations (26) and (27):
Propylene l,2-dichloropropane
CH2=CH-CH3 + Cl2 CH2CI-CH2CI-CH3 (26)
Allyl Chloride l,3-dichloropropene
and CH2=CH-CH2CI + Cl2 ~ CHCL=CH-CH2CI + HCI (27)
Chlorine is reacting with the hydrocarbons in parallel while the hydrocarbon is reacting
in series_ Thus selectivity to allyl chloride and minimization of 1,2-dichloropropane and 1,3-
dichloropropene are desirable.
Formation of heavy and light ends in the allyl chloride synthesis is governed by these
side reactions. However, the heavy and light ends were neglected because the kinetic data
were not available. In principle, the desired substitution reaction to form allyl chloride is
accompanied by a low temperature addition chlorination reaction, an unsaturated
monochloride isomer formation, an unsaturated dichloride formation by further chlorination
of allyl chloride, and by thermal degradation to tars, carbon and benzene.
In the commercial process, propylene and chlorine gas are continuously fed into the
adiabatic non-catalytic reactor to produce allyl chloride. The reactor effluent is cooled and
sent to a prefractionator to separate HCI and propylene. The remaining organic chloride
fraction is separated in a two-step distillation. The dichloropropane is taken overhead in the
first column and the heavy boiling fraction consisting of the dichloropropene and
degradation products are taken off as bottom products.
This allyl chloride synthesis example is one which demonstrates that modification of
the reaction parameters may have a major effect on the by-product production. An analysis
of the process requires that component mass balance, kinetic rate equations and total energy
balances be solved simultaneously. A complete analysis would also investigate the effects
of mixing on the reaction system.
A study of this process is based on literature data [11, 17] and two other studies [l,
30]. The emphasis of this study in contrast to previous studies is directed toward the concept
of waste minimization. The results are based on evaluating the selectivity to allyl chloride,
the conversion of chlorine and propylene, and the minimization of 1,2-dichloropropane, 1,3-
dichloropropene, and HC!. The evaluated reaction parameters are given in Table 6.
Table 6. The Reaction Parameters
Reactor Types PFRICSTR

Reactor Conditions Isothermal/ Adiabatic/Non-Adiabatic
Residence Time
Feed Conditions Propylene/Chlorine Ratio, e
Entering Feed Temperature
Kinetic Model Three Reactions
The reaction rate equation for the reactions of equation (25) and (26) were given by
Smith [30] as developed from literature data [11]. Biegler and Hughers [1] did further
analysis of this data and developed a rate equation for the reaction of equation (27).
The rate equations are:
36
(28)
(29)
(30)
where the rate constants are expressed in the Arrhenius form by:
. (31)
-E"I
A. e---nr
Table 7. The Reaction System Constants
Frequency Factor, Activation Energy, Heat of Reaction,

Ai Ei MIri (298 K)
Equation (lb-mollhr-ftl -atnr) (Btu/lb-mol) (Btu/lb-mol)
25 206,000 27,200 -48,000
26 11.7 6,860 -79,200
27 4.6E+08 10,650 -91,800
The heat of reaction associated with each reaction i is given by equation (14) and the
specific heat for each specie j is given by equation (15).
Values for the constants for the Arrhenius Equation and the heat of reaction are given
in Table 7 [1] and Table 8 [27, 39].
The stoichiometric equation, component mass balance, and total energy balance are
given by equation (16), (18), (23) and (24), respectively.
Table 8. Constants for Heat Capacity Equations
Component (X) f3} 'Y} z}

Propylene 0.8861 5.601E-02 -2.77IE-05 5.267E-09
Chlorine 6.432 7.278E-03 -9.24IE-06 3.695E-09
Allyl Chloride 0.604 7.278E-02 -5.44IE-05 1. 742E-08
1,3-dichloropropene 2.809 7.700E-02 -6.306E-05 2.147E-08
1,2-dichloropropane 2.496 8.730E-02 -6. 220E-05 1. 849E-08
Hydrogen Chloride 7.325 -1. 720E-03 2.976E-06 -9.310E-IO
Btu/(Ib-mol OF)
Fahrenheit
37
DISCUSSION AND RESULTS
The variation of 1,2 DCP by-product produced in the process is shown in Figure 7
as a function of the propylene/chlorine mole ratio. The figure applies for residence times
of 1.0, 1.5 and 2.0 seconds. These curves show that the production of 1,2 DCP merge
together at propylene/chlorine mole ratio less than 6. At propylene/chlorine mole ratios
above 6, there is slightly more 1,2 DCP at the residence time of 1.0 second than there is
at 1.5 seconds. The amount of 1,2 DCP at 1.5 and 2.0 seconds residence time are almost
identical. This observation suggests that a residence time of about 1.5 seconds is desirable.
Higher residence times corresponds to higher cost (larger reactor) and do not provide
significantly less by-product. Lower residence times produce more by-product.
Table 9. Boiling Point Temperature
Component Boiling Point, K

1. Hydrogen Chloride, HCI 188.1
2. Propylene, C3H6 225.7
3. Chlorine, Cl2 239. 1
4. Allyl Chloride, C3HsCI 318.4
(3-Chloro-l-Propene)
5. 1,2 Dichloropropane, C3H6CI2 370.0
(1,2 DCP)
6. cis 1,3 Dichloropropene, C3H4CI2 377.0
(1,3 DCP)
7. trans 1,3 Dichloropropene, C3H4CI 2 385.2
(1,3 DCP)
1.2 DCP 1.2 Dichloropropone

0.70
c
..
~ 0.60
~
a 0.50 3 Reaction Model
Plug Flow Reactor
j
Non-adiabatic Operation
Inlet Temperature 674°F
0.40
t.>
:s
~ 0.30
~
~0.20 - 1.00 sec.
- 1.50 sec.
H++-t 2.00 sec.
2 LL.U.J./..U.JJ4l.u.LJ.J.J.J.~6..LJ.LL.U..l'-:8.J..UJ..LJ.L'":1~0.1.U.J..LJ.L~12~.1.U.JU;14
0.10 ob..u.J..llJ.ul.
Propylene/Chlorine. mole ratio
Figure 7. The Variation of 1,2 DCP By-Product as a function of Propylene/Chlorine Mole Ratio
38
At the propylene/chlorine mole ratio of 3.0, the production of 1,2 DCP is about 0.62
at the residence time of 1.5 seconds. As the propylene/chlorine mole ratio increases from
0.20 to 6, the production of 1,2 DCP decreases from about 0.62 to 0.30 as shown in the
graph. This decrease is desirable since this corresponds to less production of by-product per
quantity of allyl chloride produced. As the feed mole ratio of propylene to chlorine increases
from 6 to 7, the by-product increases from 0.30 to 0.305 as shown. The quantity of by-
product is larger and continues to increase moderately as the propylene/chlorine mole ratios
increases.
The minimum production of 1,2 DCP appears to occur at a propylene/chlorine mole
ratio of about 6 and at a residence time of about 1.5 seconds.
Thus the minimum production of 1,2 DCP appears to occur at the following reactor
parameter conditions:
Reactor Type Plug Flow Reactor

Reactor Conditions Non-adiabatic Operation
Outlet Temperature, of = 876
Residence time, seconds = 1.5
Feed Condition Propylene/Chlorine, mole ratio =6
Inlet Temperature, of = 674
REACTOR COOLER
0-1
Figure 8. Distillation Flowchart for AUyl Chloride.
The stream exiting from the non-adiabatic plug flow reactor is a mixture containing
hydrogen chloride, trace propylene, trace chlorine, allyl chloride, 1,2 dichloropropane (1,2
DCP) and 1,3 dichloropropene (1,3 DCP). Boiling Point temperatures for these compounds
are provided in Table 8. Separation of allyl chloride form the mixture is required.
Inspection of the boiling points in Table 9 show that the boiling point temperature of
1,3 DCP (both cis- and trans-I,3 dichloropropane) is higher than the boiling point
temperature of 1,2 DCP (377 and 385.2 K vs 370 K). Property data for 1,3 DCP was very
sparse, therefore 1,2 DCP was used to represent 1,3 DCP in the distillation calculations.
This representation corresponds to a conservative view since 1,3 DCP is less volatile.
A distillation flowchart to accomplish the separation and provide a purified allyl
chloride product is shown in Figure 8.
39
1,2 DCP ~ 1,2 Dichloropropone
1.00
i purity limit
] 0.60
~ 0.60
~
~0.40
~
'" 0.20
~ Refill][ Ratio: ccccc R = 6
66666 R = 5
0'08.obO..u...L.LU..u
0 •..l.
1OJ...L..L.L.L.LJU.O.U.2-,-OJ...L..L.L.L.LJ0U..3J..10..u...L.J...U..uO•..l.401..LJ..JL.L.LJu..O.u.5-'-0J..U..L.L.LJO
uJ.60
Distillate/Feed, mole ratio
Figure 9. Variation of 1,2 DCP By-Product.
In the flowchart of Figure 8, the mixture issuing from the non-adiabatic plug flow, 1,
reactor is cooled and sent to distillation for separation of the allyl chloride. In the first
distillation column, D-l, the light boiling fraction (hydrogen chloride, trace propylene, and
trace chlorine) are taken as the overhead distillate, 2. The bottoms from the first distillation,
3, are sent as the feed to the second distillation, D-2. In the second distillation, D-2, the
purified allyl chloride is taken as the overhead distillate, 4. The bottoms from the second
distillation comprise the heavy fraction, 5.
The quantity of 1,2 DCP by-product produced per quantity of purified allyl chloride
product is influenced by the distillate/feed mole ratio for the second distillation. The
variation of 1,2 DCP by-product produced in the process is shown in Figure 9 as a function
of the distillate/feed mole ratio.
In preparing the figure, distillation calculations were performed using the HYSIM
process simulation software and the PRSV property package (Peng-Robinson Stryjek-Vera)
was used for the vapor-liquid calculations. The PRSV Equation of State is a two fold
modification of the PR (Peng-Robinson) equation of state that extends the application of the
original PR method to highly non-ideal systems. For this application, binary interaction
coefficients were assigned a value of zero in the absence of experimental data.
In the Figure 9, the variation of 1,2 DCP by-product is shown at reflux ratios of R =
5 and 6. Calculation at lower reflux ratios did not provide a purified allyl chloride product.
Inspection of the curve indicates that results at R = 5 are equivalent to results at R = 6.
The same finding of equivalent results was ascertained at even higher reflux ratios (R =
8,10,12,14, ... ). Since such higher reflux ratios correspond to higher energy costs (more
boil-up in reboiler and more condensation in condenser), the lower value for reflux ratio of
R = 5 is preferred.
The effect of distillate/feed mole ratio on minimizing the 1,2 DCP by product is
illustrated in the Figure 9. At the low value of distillate/feed mole ratio of 0.18, the amount
of 1,2 DCP by-product per purified allyl chloride is approximately 0.98. As the
distillate/feed mole ratio increases from 0.18 to 0.58, the 1,2 DCP by-product decreases
from about 0.98 to about 0.30. This decrease is desirable since this corresponds to less
production of by-product per quantity of purified allyl chloride product.
The minimum production of 1,2 DCP by-product appears to occur at a distillate/feed
mole ratio of about 0.58 and reflux ratio of 5 or greater.
40
Waste: 1,2 DCP By-product
Product : Allyl Chloride
Plant Size : 210,000 tons/year
io
<.J 0.20 'lib
... 40
~: : ;
II
~
.t::" 20
%~~~~2~~~4~~~6~~~~8~~~10~~~12
Propylene/Chlorine. mole ratio
Figure 10. Cost of Waste Treatment
Thus the minimum production of 1,2 DCP by-product appears to occur at the following
separation parameter conditions:
Distillation column D-2

Pressure, psia 20
Distillate/Feed, mole ratio 0.58
Reflux ratio, R = 5
If the by-product in a process is not salable in the market or valuable in another

process, there is a cost for treatment of the by-product as a waste. The cost of waste
treatment is shown in Figure 10. This figure was developed for a plant size of 210,000
tons/year of allyl chloride (equivalent to large plant producing epichlorohydrin with 210,000
tons/year capacity).
For Figure 10, the cost of waste treatment for the 1,2 DCP by-product is shown as a
function of propylene/chlorine mole ratio at three levels of specific treatment costs of 0.20,
0.10 and 0.05 $/lb. At a low value for the propylene/chlorine mole ratio of about 3, the
waste treatment costs are about 76, 38, and 19 million dollars for the respective specific
treatment costs. As the propylene/chlorine mole ratio is increased from the low value of 3
to about 6, the waste treatment cost are about 36, 18 and 9 million dollars for the respective
curves. This decrease in waste treatment costs is desirable. As the propylene/chlorine mole
ratio increases to higher values (6 to 7, 8, 9, 10 and 11), the waste treatment costs
associated with handling the by-product increases to slightly higher values.
Thus the minimum cost associated with handling the 1,2 DCP by-product occurs at a
propylene/chlorine mole ratio of about 6.
ACKNOWLEDGMENT
This project has been funded in part with Federal Funds as part of the program of
the Gulf Coast Hazardous Substance Research Center which is supported under cooperative
agreement R815197 with the United States Environmental Protection Agency. The contents
41
do not necessarily reflect the views and Policies of the U.S.EPA nor does the mention of
trade names or commercial products constitute endorsement or recommendation for use.
REFERENCES
1. Biegler, L.T. and R.R. Hughes, Comput. Chem. Eng., 7(5), 645-661 (1983).
2. Bundick, H.L., T.B. Kopecky, B.R. Allen, and J.R. Hopper, "Continuous Lube Oil
Additive Process: Waste Reduction By Process Modification and Control," AIChE
1992 Spring National Meeting, New Orleans, LA, March 29 - April 2 (1992).
3. Callahan, H.L., B. Gertisser, and J.J. Szaba, "Process for the Manufacture of
Unsaturated Nitriles from Olefins and Ammonia," U.S. Patent 3,354,197 (1967).
4. Callahan, J.L., R.K. Grasselli, E.C. Milberger, and H.A. Strecker, "Oxidation and
Ammoxidation of Propylene over Bismuth Molybdate Catalyst.," Ind. Eng. Chem.
Prod. Res. Develop., 9(2), p. 134-142 (1970).
5. Callahan, J.L., R.K. GrasselIi, E.C. Milberger, and H.A. Strecker, "Oxidation and
Ammoxidation of Propylene over Bismuth Molybdate Catalyst," Division of
Peteroleum Chemistry Inc. ACS., p. CI3-C27, September (1969).
6. Cathala, M. and J .E. Germin, "Ammoxidation Catalytique des hydrocarbures et reactions
apparentees.," Bulletin de La Societe Chimique de France., No. 11, p. 4114-4119
(1970).
7. CMA Conference Proceedings November, 1987, "Waste Minimization," Chem.
Eng. (NY), 95 (No. 11) 37, August 15 (1988).
8. Daizo, K. and o. Levenspiel, Fluidization Engineering, John Wiley & Sons, New York,
NY (1969).
9. Dalin et al., Azerb Khim. Zhur., No.5, p. 99-102 (1963).
10. Denbigh, K.G., and J.C.R. Turner, Chemical Reactor Theory, 2nd ed., Cambridge
University Press (1971).
11. Fairbairn, A.W., H.A. Cheney, and A.J. Cherniavsky, Chern. Eng. Prog., 43, 280
1947.
12. Farkas, A., "What You Should Know ... Catalytic Hydrocarbon Oxidation.,"
Hydrocarbon Processing, p. 121, July (1970).
13. Fogler, H.S., Elememts of Chemical Reaction Engineering, Prentice-Hall, Engelwood
Cliffs, NJ (1986).
14. Freeman, H.M., "Hazardous Waste Minimization," JAPCA 38, No.1, 59, January
(1988).
15. Freeman, H.M., Hazardous Waste Minimization, McGraw-Hill, New York, NY (1990).
16. Fromm, C.H., M.S. Callahan, H.M. Freeman, and M. Drabkin, "Succeeding at Waste
Minimization," Chern. Eng. (NY), 91-94, September 14 (1987).
17. Groll, H.P.A. and G. Hearne, Ind. Eng. Chem., 31, 1530 (1939).
18. Hadley D.J., in H.G. Hancock, ed., Propylene and its Derivatives, Ernest Benn Ltd.,
London, p. 416-497 (1973).
19. Hanh, A.V., The Petrochemical Industry., McGraw-Hill, New York, NY (1970).
20. Holland, C.D. and R.G. AnthonY,Fundamentals of Chemical Reaction Engineering,
Prentice-Hall, Englewood Cliffs, NJ (1979).
21. Hopper, J.R. and A. Muninnimit, "Waste Minimization by Process Modification.",
Preceedings of the National AIChE Conference, Washington D.C., December
(1989).
22. Kirk-Othmer., Encyclopedia of Chemical Technology, 3rd ed., A Wiley-Inter-science
Publication, John Wiley & Sons, New York, NY (1978).
23. Lankhuyzen, S.P., P.M. Florack, and H.S. van der Baan, "The Catalytic Ammoxidation
of Propylene over Bismuth Molybdate Catalyst.," Journal of Catalyst, 42, p. 20-28
(1976).
42
24. Levenspiel, 0., Chemical Reaction Engineering, John Wiley & Sons, New York,
NY (1962).
25. Overcash, M., Technique for Industrial Pollution Prevention, Lewis Publishing Co.,
Chelsea, MI (1986).
26. Perkinson, G. and E. Johnson, "Designer Catalysts Make Their Debut," Chern. Eng.
(N.Y.), 96 (No.9), 30, September (1989).
27. Reid, R.C., I.M. Pransnitz, and B.E. Poling, The Properties of Gases & Liquids.,
4th ed., McGraw-Hill, New York, NY (1987).
28. Shelstad, K.A. and T.C. Chong, "Kinetics of Ammoxidation of Propylene over
Bistmuth-Molybdate Catalyst.," Can. J. Chern. Eng., 47, p. 597-602 (1969).
29. Shirtum, R.P., "Kinetic Control of By-Product Reaction with In Situ Distillation In A
Continuous Flow Device," AIChE 2nd Topical Pollution Prevention Conference,
Pittsburg, PA, August 20 - 21 (1991).
30. Smith, J.M., Chemical Engineering Kinetics, 3rd ed., McGraw-Hill, New York, NY
(1981).
31. Smyth, L.C., and J.R. Hopper, "Waste Control by Reaction Design and Catalyst
Choice.," Pollution Prevention for the 1990' s, AIChE Conference, WasHington
D.C., December 4-5 (1989).
32. Stobaugh, R.B., S.G. McH. Clark, and G.D. Camirand, "Acrylonitrile: How, Where,
Who-Future.," Hydrocarbon Processing., p.l09-120, January (1971).
33. Stull, D.R., E.F. Westrum, Jr., and G.C. Sinke, The Chemical Thermodynamics of
Organic Compounds., John Wiley & Sons, Inc., New York, NY (1969).
34. Tavalerides, L.L., Process Modification for Industrial Pollution, Industrial Waste
Management Service, J.W. Patterson, Editor, Lewis Publishing Co., Chelsea, MI
(1985).
35. USEPA Report to Congress: Minimization of Hazardous Waste, EPAI530 SW-86-033A
Office of Solid Waste, U.S. EPA, Washington, DC 20460, October (1986).
36. Vectch, F., J.L. Callahan, I.D. Idol, and E.C. Milberger, "New Route to
Acrylonitrile.," Chem. Eng. Prog., 56(10), p. 65-67 (1960).
37. Vectch, F., J.L. Callahan, J.D. Idol, and E.C. Milberger, "New Data on Sohio's
Acrylo Process.," Hydrocarbon Processing & Petroleum Refinery., 41(11), p. 187-
190 (1962).
38. Waste Minimization: Issues and Options Vol I, II USEPA Office of Solid Wastes and
Emerging Response, EPAI530 SW-86-041 October (1986).
39. Yaws, C.L., P.Y. Chiang, "Enthalpy of Formation for 700 Major Organic
Compounds.," Chem. Eng. (NY), p. 81-88, September 26 (1988).
43
KINETIC MODELS TO PREDICT AND CONTROL MINOR CONSTITUENTS
IN PROCESS REACTIONS
Selim Senkam
Department of Chemical Engineering, University of California
Los Angeles, California 90024 U.S.A.
Introduction and Back&round
Detailed Chemical Kinetic Mechanisms (DCKMs) are comprehensive descriptions of

chemical processes expressed in term of irreducible chemical events or elementa[y reactions. That
is, they provide descriptions of chemical processes at the molecular level, and as a consequence
they are of considerable utility to predict and control the formation of both the major and minor
products. In principle, detailed mechanisms comprise large, putatively complete sets of
elementary reactions for which independent rate coefficient parameters, frequently expressed in
the following form:
k = ATftexp(-ElRT)
are available from direct measurements or estimable from theoretical considerations, together
with the thermochemistry of the associated species participating in the reaction. Estimations are
based on the judicious application of theories of rate processes in a conventional manner, as well
as in conjunction with modern quantum chemical calculations, and by using analogies between
similar types of reactions. At present we rely on all of these complementary methods. Although
DCKMs have primarily been used to describe gas phase reaction processes, its application to
condensed systems, e.g. catalysis and chemical vapor deposition, is also possible. Nevertheless,
we shall describe the DCKM approach to gas phase processes in this paper for convenience.
Our ability to develop detailed chemical kinetic models has improved considerably
because of simultaneous developments in a number of interrelated areas. First, with the
availability of fast computers together with efficient numerical algorithms to handle stiff
differential equations, an increasing complexity of problems can now be c'lded and solved rapidly
without making any fundamental assumptions. For example, it is no longer necessary to the
invoke pseudo-steady-state assumption to treat reactive radicals or intermediates in modeling the
kinetics of complex chemical processes in an effort to obtain analytical expressions. In fact,
conclusions based on the assumption of pseudo-steady-state conditions or through the neglect of
reactions presumed to be unimportant in a process should be viewed questionable in the absence
of a .detailed modeling work that proves the validity of these assumptions.
Second, the quality and quantity of data on the thermochemistry of species and on the
kinetics and mechanisms of individual elementary reactions, especially those in the gas-phase,
have improved significantly over the past decade due to advances made in experimental
methods. In particular, the thermochemistry of a large number.of species are now available in
convenient tabulations (see for example Pedley et al. 1985, Chase et al. 1985). Similarly, the
rates and mechanisms of a large number of elementary reactions have also been evaluated and

documented (see for example Demore et a. 1990, Berces and Marta 1988, Tsang and Hampson
1986, Kerr and Drew 1987, Basevich 1987, Westbrook and Dryer 1984, Hanson and Salimian
1984, Warnatz 1984, Kondratiev 1972, Kerr and Moss 1981, Baulch et al. 1981, Westley 1976,
1980, Benson and O'Neal 1970). In additipn, mature empirical methods are now in place to
estimate thermochemistry and reaction rate parameters from theoretical considerations (see for
example Gaffney and Bull 1988, Benson 1976, Senkan 1992). Although considerable progress
also has been made in heterogeneous reactions, the available data are more limited, and the
theoretical framework to estimate thermochemical and rate parameters is less well developed.
Third, with recent advances made in theoretical and computational quantum mechanics,
it is possible to estimate thermochemical information via electronic structure calculations
(Dunning et al. 1988, Senkan 1992). The development of this capability obviously is crucial for
the development of DCKMs to predict the formation of trace by-products in view of obvious
experimental and cost considerations. Electronic structure calculations, together with the
Transition State Theory (TST) also allows the determination of the rate parameters of elementary
reactions from first principles. Our ability to estimate activation energy barriers is particularly
significant because, until the advent of computational quantum mechanics no
fundamentally-based methods were available for the determination of this important variable.
In order to illustrate the utility of quantum mechanical calculations consider the reaction
of C2 H5CI with H atom, which can lead to 3 distinct products: CH3 CH2 + HCl, CH3CHCl + H2 ,
and CH2CICH2 +~. At present the available data on the kinetics of each of these pathways are
quite uncertain. Consequently, we undertook semi-empirical quantum chemical calculations to
assess the relative importance of each channel. In Figure 1 the potential energy diagrams and
geometries of the reactants, products and the associated transition states in each reaction are
shown. As evident from this figure, we expect the ordering of the rates of reactions to be:
CH3CH2+ HCI > CH3CHCl + H2 > CH2CICH2+ H2. Although the results of such calculations may
not be very accurate in an absolute sense, the trends that arise are crucial for the development
of sound DCKMs.
Finally, although DCKMs comprise of a large number of reactions, to preserve the
elementary nature of these mechanisms, only a subset of these reactions are important at any
given time and application. That is, different reactions often dominate the process behavior under
different conditions. Consequently, parts of DCKMs can be developed and tested independently
for the evolution of these models. This is an important consideration in DCKMs as we can start
exploiting the advantages of this type of modeling today.
Since detailed chemical kinetic mechanisms involve the participation of a large number
of species in a large number of elementary reactions, sensitivity and reaction path analyses are
also essential elements of DCKM. Sensitivity analysis provides a means to assess the limits of
confidence we must put on our model predictions in view of uncertainties that exist in reaction
rate parameters, thermochemical and thermophysical data utilized, as well as the initial and
boundary conditions used in the modeling work. Through reaction path analysis, major reaction
pathways responsible for the production and consumption of each- species can be identified.
Powerful formalisms have been developed to undertake sensitivity analysis in detailed chemical
kinetic modeling (see for example Tilden et al. 1981, Siegneur et al. 1982, Rabitz et al. 1983,
Hwang 1983). In addition, efficient and transportable computer codes to implement these
techniques have also been developed (Lutz et al. 1988, Caracotsios and Steward 1985).
In Figure 2 various elements involved with the development of detailed chemical kinetic
mechanisms are illustrated and we shall illustrate the application of DCKM in the following
example.
An Example
We have recently studied the high temperature oxidation and pyrolysis of C2HCI3 , a
prototypical chlorinated hydrocarbon, in order to better understand the identities and
concentrations of its reaction products and by-products (Chang and Senkan 1988, Chang and
46
CH3CHZCl + H -) CH3CHZ + HCl
~ ~
Q
<!)
CH3CHZC 1 + H - > CH3CHCl + HZ
CH 3 CH 2 + HCI -E=10.1 bH =-25.5 0
r
CI1 j CH 2CI + H ----'> CH 3CHCI + H2 [=14.4 bH r =-21.J
+ H2
•
CH 2CICH 2 E=14.7 bH=-176 6
r
50
lANDO
k~ r.~
< • \
40
u
~
'0
..!:.
..... .•• 0000
OJ )0 _.0 00 (!)
~
~
••••• OO~
.;, ~ (HJCII 2C1 + H • > CH2C1CH2 + HZ
~ 20
.
cu
...c 10 6 0
....
0
6· o
... 0
o· . • :Nt f!.1
0 )H ~
-10·
. \
0.000 0.500 1000 1.500 2.000 2.500 3.000 J 500
Reaction Coordinate. r. A (!)

"~
Figure 1. Reaction Pathways and Energy Diagrams for the C2HSC1 + H
.".
-.J
Senkan 1989). The development of better insights for the high temperature reactions of
chlorinated hydrocarbons (CHC) is important because of the growing need to better understand
and control pollutant emissions into the environment from combustion systems burning hazardous
materials that frequently contain CHCs For example, CHCs are associated with the formation
of chlorinated aromatics, di-benzo-dioxins, and di-benzo-furans in incinerators, some of which
are extremely toxic (Junk and Ford 1980, Schaub and Tsang 1983, Oberg et al. 1985). Second,
the manufacture of useful chemicals by the controlled high temperature reactions of CHCs also
is a promising enterprise. This was shown recently by the development of the chlorine-catalyzed
oxidative-pyrolysis (CCOP) process to convert methane, the major component in natural gas and
an important product in the anaerobic digestion of organic matter, into more valuable products
such as ethylene, acetylene and vinyl chloride (Granada et al. 1987).
Reaction Mechanism
An elementary reaction set describing the high-temperature fuel-rich combustion of

C2HCl3 is presented in Table I together with the rate parameters for the forward reaction paths
(Chang and Senkan 1989). This mechanism was constructed by systematically considering all
MEASLfIED DIRECT
VALUES MEASUREMBITS
Figure 2. Elements of Detailed Chemical Kinetic Mechanism Development
48
Table 1. Detailed Chemical Kinetic Mechanism for C2HCl 3 Combustion (k = ATnexp(-EjRT), in cm3 , kcal, mole, s units)
Reaction A n E Reaction A n E
C, Kinetics 23. C,CI. ;=t CI + C,CI, 1.35E36 ·6.48 74430.

24. C,CI. + CI ;=t CI, + C,CI, 6.31E13 O. 18300.
1. C,HCI, ... C,HCI, + CI 3.02E33 ·5.88 88300.
25. C,HCI, + M ... C,HCI + CI + M 8.00E14 O. 28000.
2. C,HCI, + CI ... C,HCI, 3.66E32 ·6.52 5520.
26. C,HCI, + CI;=t C,CI, + HCI 7.26E15 ·0.97 1940.
. 3. C,HCI, ... C,CI, + HCI 1.07E26 ·3.38 61560
27. C,HCI, + CI ... C,HCI + CI, 5.30E05 US 1140.
4. C,HCI, ... C,HCI + CI, 9.55E35 ·7.06 83500.
28. C,HCI, + a, ... CHOCI + COCI 1.00EII O. 5000.
5. C,HCI, + CI ... C,HCI, + CI, 3.16E13 O. 20000.
29. C,CI, + CI ... C,CI, 4.42E33 ·7.21 5710.
6. C,HCI, + CI ... C,CI, + HCI 2.00E12 O. 5000.
30. C,CI, + CI ... C,CI, + CI, 5.26E17 ·1.66 5090.
7. C,HCI, + 0 ... CHOCI + CCI, 1.00E13 O. 2000 .
. 8S0.•• 31. C,CI, + a, ... COCI, + COCI 3.16E11 O. 5000.
8. C,HCI, + OH'" CHCI,COCI + H 3.19EII O.
• 9. 2000 .•• 32. C,CI, + a, "" C,CI,O + cia 3.16ElI O. 5000.
C,HCI, + cia ... CHCI,COCI + CI 1.00E11 O.
8000 .•• 33. C,CI,O + CI ... Co + CCI, 1.00E13 O. O.
10. C,HCI, + CCI, ... CCI, + C,HCI, 3.16ElI O.
34. C,CI, + a ;=t co + CCI, 1.00E13 O. O.
I!. C,HCI, + CHCI, ... CHCI, + C,HCI, 2.51E12 O. 5000.
35. C,CI, + Cia"" co + CCI, l.ooE12 O. O.
12. C,HCI, + C,HCI, p C,HCI, + C,HCI,1.00ElI o. 8000.
36. C,HCI, + 0, ... CHCI,COCI + C/O 1.00ElI O. 5000.
13. C,CI, + OH ... CHCI,COCI + CI 5.66E12 O. 2400 .
• 37. CHCI,COCI + CI ... CCI,COCI + HCI 1.00E13 O. 5000.
14. C,CI, + 0 pO
COCI, + CCI, 1.00E13 O. 5000.
38 CHCI,COCI + CI ... CHCI,CO + CI, 1.00EI4 O. 17600.
15. C,CI, + CIO ... CCI,COCI + CI LOOE11 O. 5000.
• 39. CCI,COCI ... C·CI, + CO 1.00EI2 O. 5000.
16. C,CI, + CI ... C,CI, 2.62E35 ·7.71 5280.
40. CHCI,CO'" CHCI, + CO 3.00E13 O. 13500.
17. C,CI, + CI, p C,CI, + CI 2.51E12 O. 3000.
41. C,HCI. + CI ... CHCI, + CCI, 6.15E20 ·1.75 16510.
18. C,HCI, ... HCI + C,CI, 1.66E33 ·5.79 6S050.
42. C,CI, + CI ... CCI, + CCI, I. 70E27 ·4.01 12120.
19. C,HCI, ... CI + C,HCI, 2.19E29 ·4.14 72000.
43. C,CI. + CI ... C,CI, + CI, 1.27E27 ·4.73 8900.
20. C,HCI, + CI ... HCI + C,CI, 2.00E12 O. 3300.
44. C,CI, + a, ... CCI,COCI + cia 1.00E12 O. 12000.
21. C,HCI, + CI ... CI, + C,HCI, 1.00E13 O. 16600.
45. CCI,COCI + CI ... CCI,CO + CI, 1.00E14 O. 17600.
22. C,CI• ... CI, + C,CI. 1.62E35 ·6.53 63200.
46. CCI,CO" CCI, + Co 3.00E13 O. 8000.
+>
10
til
0 Table I. (continued)
47. C,CI, + 0 ... CCI, + CO 7.00E13 O. O. 70. CCI, + CI, ... CCI, + CI 5.00E12 O. 3000.
•• 48. C,Ch + 0, ... COCI + COCI 1.00E11 O. 5000. 71. CHOCI + M ;:0 CO + HCI + M 1.00E17 O. 40000 .
49. C,CI, + CI, ;:0 C,CI, 2.63E36 ·7.17 40860. 72. CHOCI+CI ... COCI+HCI 2.00E13 O. 3000.
• 50. C,CI, + M ;:0 C,CI, + CI + M 8.00E14 O. 28000. C, - C. Kinetics
51. C,CI, + C/O ;:0 CO + CCI, 1.00E11 O. O. • 73. CCI, + C,CI, ;:0 C,CI. 3.16E11 O. 5300.
52. C,CI, + OH '" CO + CHCI, 1.00E12 O. O. • 74. C,HCI. + CI '" C,CI. + Hel 5.00E12 O. 5000.
C, Kinetics • 75. C,CI. + GI '" C,CI. + CI, 3.16E13 O. 20000.
53. CHCI, ... CHCI, + CI 2.5IE2; ·4.02 ;9iOO. • 76. CCI, + C,HCI, "" C,HCI. 5.77E11 O. 5280.
54. CHGI, + 0 "" COGI, + Hel 1.00El1 O. 4000. 77. C,HCI. + 0, '" C,CI. + HO, 1.00E12 O. 5000.
55. CHCI, + 0 '" CCI, + OH 2.90E12 O. 5000. 78. C,HCI. '" C,HCI. + GI 1.00E13 O. 22000.
56. CHCI, + CI ... CHCI, + CI, 1.00EI4 O. 21000. • 79. C,CI, ... C,CI. + CI 1.00EI3 O. IS600.
• 57. CHCI, + CI ... CCI, + HCI 6.92E12 O. 33~0. 80. CCI, + C,CI, '" C,CI, 3.16El1 O. 4300.
58. CCI, "" CCI, + CI ;.41 E35 ·6.52 '53GO. • 81. C,CI, + C,CI, ;:0 C.CI, 6.31E11 O. 4200.
59. CCI. + 0 ... CCI, + C/O 2.50E10 O. 2270. • S2. C,CI, + C,CI • ... C.CI,· 2.00E12 O. 4000.
• 60. CCI, + 0, ... COCI, + CIO 1.00EI3 O. 28000. • 83. C.CIr ..... C.CI. + CI 1.00EIO O. O.
61. CCI, +0 "" COCI, + CI 1.00EI4 O. O. • 84. C.GI. + CI ... C,CIo + CI, 5.00E12 O. 20000.
• 62. CCI, + CI, ... CCI, + CI 2.51E12 O. 6000. • 85. C,HCI. + CI ... C,CI. + HCI 2.00E13 O. 3000.
63 GCI, + CCI, '" C,CI. 1.41E36 ·7.48 6680. S6. C,GI, + C,GI, ... C,C!, 1.81E11 O. 4100.
64. CCI, + CCI, ... C,CI, + CI, 2.22E26 ·4.43 8980. 87. C,CI, ... C,CI. + CI 1.00E13 O. 18000.
65. CCI, + CHCI, ... C,HCI. 1.66E34 ·6.79 6010. 8S. C,HCI. + 0, "" C,CI. + HO, 1.00EI2 O. 501)0.
66. CCI, + GHCI, ... C,CI, + HCI 2.36E20 ·2.45 6380. • 89. C,GI, + G,HCI, "" C,HCI. 5.00Ell O. 3000.
67. CHCI, + 0, ... CHOCI + CIO 1.00E13 O. 28000. • 90. C,HCI• ... C,HGI. + CI 1.00E13 O. 16000.
68. CHCI, + 0 ... CHOCI + CI 1.00E14 O. O. • 91. C,C/o + C,CI, ... C.CI. + CI 5.00EII O. 1000.
69. CCI, + 0, ... CIO + COCI 1.00EI3 O. 1000. • 92. C.CI, ..... C.CI. + CI, 1. OOE 10 O. O.
CO and H. Kinetics 116. CDC I + OH .. CO + HOCI 1.00E14 O. 0',
93. CO + 0 + M .. CO. + M 5.30E13 O. -4538. 117. CDC I + 0 .. CO + CIO l.ooE14 O. O.

94. CO + OH .. CO. + H 4.40E06 1.5 ·740. 118. COCI + 0 .. CO. + CI 1.00E13 O. O.
, 95. CO + D• ... CO. + 0 1.60E13 O. 41000. 119. COCI + 0, .. CO. + CIO 7.94E10 O. 3300.
96. CO + HO • ... CO. + OH 1.50E14 O. 23600. '120. COCI + CI .. CO + CI, 1.26E13 0.5 500.
97. H. + 0, .. OH + OH 1.70E13 O. 48100. 121. CI. + M '" CI + CI + AI 2.32E13 O. 46960.

98. OH + H, ... H,O + H 1.00E08 1.6 3300. 122. HCI + AI .. H + CI + AI 2.75E13 O. 81760.
99. H+O,"OH+O 1.20E17 ·0.91 16510. 123. HCI+ H .. H, + CI 7.94E12 3400

100. 0 + H, ... OH + H 1.50E07 2.0 7550. 124. H + CI, .. HCI + CI 8.51E13 O. 1170.
101. H+O,+M ... HO,+M 7.ooE17 ·08 O. 125. 0 + HCI .. OH + CI 3.16E13 O. 6700.
102. OH + HO, .. H.O + 0, 2.ooE13 O. O. 126. OH + HCI '" CI + H,O 2.25E12 O. 1020.
103. H + HO, .. OH + OH 1.50E14 O. 1003. 127. 0 + CI, '" C/O + CI 1.26E13 O. 2800.
104. O+HO, ... 0, +OH 2.00E13 O. O. 128. 0 + C/O '" CI + 0, 5.70E13 O. 364.
105. OH + OH .. 0 + H,O 1.50E09 1.14 O. 129. CI + HO, ... HCI + 0, 7.94EI2 O. O.
106. H + H +M ... H, +.11 1.00EI8 ·1.0 O. 130. CI + HO, .. OH + cia 6.31E13 O. 1700.
107. H + H + H, =H, + H, 9.20E16 ·0.6 O. 131. cia + H, .. HOCI + H l.ooE13 O. 13500.
108. H + H + H,O ... H, + H,O 6.00E19 -1.25 O. 132. H + HOCI ... HCI + OH 1.00E13 O. 1000.
109. H + H + co, ... H, + co, 5.49E20 ·2.0 133. CI + HOCI =HCI + cia 1.00E13 O. 2000.
110. H+OH+M=H,O+M 1.60E22 ·2.0 O. 134. CI + HOCI .. CI, + OH 1.26EI3 O. 6000.
III. H+O+M"'OH+Af' 6.20E16 ·0.6 o. 135. 0 + HOCI =OH + CIO 5.00E13 O. 1500.
112. H + HO, H, + 0,
= 2.50E13 O. 693. 136. OH + HOCI .. H,O + CIO 1.80EI2 O. 3000.
'113. cia + CO .. CO, + CI 6.02E11 O. 7400. 137. HOCI+M =OH +CI+ AI 1.00E18 O. 56000.
"114. COCI+AI ... CO+CI+AI 2.00E14 O. 6500 138. ciao + co .. co, + c/O 1.00E14 20000.
lI5. COCI + H .. CO + HCI 1.00E14 O. O. 139. ClOD + AI .. CI + 0, + .II 1.00E15 O. iDOO.
(J1
Table I. (concluded)
Reaction A n E
140. CI + ciao ... CI, + a, 7.94E13 o. SOO.
141. CI + ciao ... cia + cia S.01E12 O. SOO.
142. a + CiaO'" cia + a, 3.16E13 O. SOO.
143. COCI. +M ... COCI + GI + M 1.00E16 O. 76000
144. GOCI, +M ... GO + GI, + M 1.00E16 O. SOOOO.
"145. GOGI, + GI "" GOCI + GI. 3.16E13 0.5 20000.
146. GOCI. + OH ... GOGI + HOGI 1.00E12 O. 10000.
147. COG I. +H "" GOGI + HGl 1.00E13 O. 2000.
plausible elementary reactions of C2 HCl3 an~ O2 , and their daughter species consistent with the
principles of physical organic chemistry and experimental data, and by eliminating those reactions
that did not contribute to reaction rates and were determined to be unimportant by the sensitivity
analysis. The thermochemistry of species were obtained from standard sources when available,
and were estimated by group additivity (Benson 1976) and quantum chemical methods (Senkan
1992). Similarly, reaction rate parameters were obtained from published compilations and other
sources when available or estimated via quantum chemical methods described in Senkan 1992.
The reaction mechanism was then used to simulate the chemical structure of a one-
dimensional flat flame with the following precombustion composition: C2 HCl3 =22.6%,
O2 =33.1 %, and Ar=44.3% (Chang and Senkan 1988).
In Figure 3 calculated mole fraction profiles (indicated by lines) for the major species
C2 HCI3 , O2 , CO, HCl, C1 2 , and CO2 are compared to those determined experimentally (indicated
by symbols). As evident from this figure, the agreement between the model and experimental
data is generally reasonable, suggesting that the major features of the flame chemistry of C2 HCl 3
have been described reasonably well by the model.
However, what is significant is that the DCKMs also provide insights in predicting the
fates of the minor products in C2 HCl 3 combustion. For example in -Figure 4 DCKM predictions
are compared to experiments for COCI 2 , C2C1 2 , C2Cl 4 and CCI 4 • Similar comparisons are
provided for C2 C16 , C3 C1 6 , C4 C16 , in Figure 5, for CHC12 COCI and CCl3 COCI in Figure 6, and
for CHCI3 , C2 HCls and C3 HCls in Figure 7. As can be seen from these figures, the DCKM
developed for C2 HCl3 combustion provide' a reasonable description of the molecular events
leading to the formation and destruction minor constituents under the experimental conditions
investigated. It is also significant to note that because DCKMs are fundamental models, they can
be used to predict product distributions under different conditions, including conditions under
which the conduct of experiments can be difficult or impossible, thus they allow for the a priori
establishment of optimum operating conditions to minimize or maximize the formation of by-
products.
The reaction set presented in Table I can also be subjected to a sensitivity analysis to
determine the rank order of the reactions in the mechanism. For this, normalized first order
sensitivity gradients (Sij) can be calculated along the flame using the following definition:
where C j is the molar concentration for species i, and Aj is the pre-exponential factor for the j'th
52
0.4 r--r----,r-----,----r---.-...,.---,----,---.--,lS00
Symbols: Experimenta!
Lines : Calculated -----
T CEXPT.J /--~~
..... _ / / 4
0.3
///-
1300
""
~
/ ~
/
I w
I
I co a:
/ -;:---6-------.0- ::J
" - "
I-
/ 1100 a:
(V
, a:
HCl
....•..."/" .....................................
w
(L
W
...J + + + :r:
o w
:r: ? .......... _...... - .~.-.. I-
0.1 800
700
0.5 1.0 1.5 2.0 2.5
DIST~NCE nLONG FLnNE (HN)
Figure 3. Comparison of the experimental data with model predictions for
C2HCI3 • O2 • CO. CO2 • HCI, and C12 •
2.0
ru
o
....
X Symbols: Experimental
z 1.5 Lines : Calculated
o
H
I-
U
a:
a:
u. 1.0
w
.J
o
:r:
0.5
1.50 1.75

COCI2 • C2C12 , C1Cl 4 , and CCI4 •
53
reaction at a particular location in the flame. This definition is particularly useful because
influential reactions in the mechanism can be identified even without the explicit presence of a
particular species in that reaction. In addition, the net rates of individual elementary reactions
in the mechanism can be calculated to determine important reaction pathways. From such
considerations the following conclusions can be reached concerning the formation and
destruction of various species in fuel-rich C2 HCI 3 combustion.
to
CHCI 2 CDCI
0
x
Symbol': Experimental
Lines : Calculated
x
z
o
~ 10-3
u x
a
a:
ll.
~ 10- 4
o
I:
,
,
10-£ 0.0 0.5 1.0 1.5 2.0 2.:

DIST~NCE ~LONG FLAHE (HMl
CHCl 2COCI and CCI 3 COC!.
In Figure 8, rate profiles of reactions involving C2 HCl 3 that have the highest rates are
presented. As evident from these profiles the major reaction pathway for C2 HCl 3 in the flame
was by CI radical attack, Le. C2HCI 3 +CI=C 2 CI3 +HCI, which also was the major route for the
formation of HC!. In the earlier, Le. cooler, parts of this fuel-rich flame the addition of C2CI3
and CIO radicals to C1HCl 3 were also significant reactions. In contrast, in fuel-lean flames
C2CI 3 would have been consumed rapidly by its reaction with O2 , i.e. C2 CI3 +02 =
COCI2 +COC!. The addition of C1CIJ to C1HCl3 results in the formation of C4 C16 via the
sequence: C1HCI3+C1CI3=C4HCI6 and C4HCl6+02=C.CI6+H02. The addition of CIO to
C2HCl3 produces CHCI1COCI by C1HCl3 +CIO=CHCI1COCI +C!. Both C4 CI6 and CHCI 2COCI
form in significant levels in fuel-rich flames of C1HCl 3 (see Figures 5 and 6, respectively).
54
From the analysis of net rates, the major reactions responsible for the consumption of
O2 were determined to be: C2CI3+02=COCI 2+COCl, C2CI3+02=C2CI20+ClO and
C2Cl 2 +02 =COCI +COCI. CO forms as a consequence of the unimolecular decomposition of
COCl, i.e. COCl+M=CO+Cl+M, in which COCI forms via.C 2CI3+02=COCI 2+COCI and
C2CI 2 +02 = COCl+COCI. The major reaction responsible for the consumption of CO was
CO+ClO=C0 2 +CI, which also was the primary reaction path for the formation of CO 2,
Because of the hydrogen lean nature of this mixture, the concentration of OH would be low,
thus the CO+OH=C02 +H reaction, which is the major reaction pathway for CO oxidation in
normal hydrocarbon systems, contributes very little to the formation of CO2 in C2HCl 3 flames.
C"C1 S
I
.
f., ..\,
,
Symbols: Experimental
Lines : Calculated
.:
I \
I \
\I
'"o I
~
\
~ 4.0 I
I
I
\
,
I
I \
z
o t, \
H 4: " \
f-
,: CsC1s X.2S
I \
(.) ...-~.--,
a:
a:
,: l
61 I :
\\ -r-·----------·. -----·---------
,
LL
,if ~~C-1C1S
~ 2.0
~ l:f c 'K
I g C '\p
o I • C~6
Cl '\ ,
I: , : .... 0
.6.,;
t 0),1 '"
,'I'~
0
10\
' , .........
~: c \ .... 'Q. ....
oJ: '\ ..
011 I ,
',~
~ -".4
.......
0.0 " ......... .
0.0 1.0 2.0 3.0 4.0 5.0
DISTRNCE RLONG FLAME (MM)
C1 C4. C3C4. and C 4C4.
HCl and Cl 2
HCI forms primarily as a consequence of Cl radical attack on C 2HCI 3• i.e.

C2 HCl 3+Cl =C 2C1 3+ HCl, thus its appearance early in the flame zone is expected (see Figures
3 and 9). In addition, HCl forms by the reaction C 4HCI s +Cl=C 4Cls +HCI, in which C4HCl s
forms by the decomposition of the adduct in reaction C 2Cl3+C 2HCI 3=C 4HCI 6 •
According to the mechanism, major reaction pathways responsible for the formation of
Cl2 were: COCI 2 +Cl=COCl+CI 2 and COCl+Cl=CO+CI2 (Figure 10). Consumption of Cl 2 ,
on the other hand occurs via C2CI3+CI2=C2CI4+CI. Since C 2C1 3 forms very early in the flame
zone, i.e. by Cl attack on C 2HCl 3, it prevents the buildup of CI 2 in the system, thus Cl2 mole
fraction profile exhibits an induction period (see Figures 3 and 10).
55
I I I I
Symbols: Experimental
Lines : Calculated
z
o
H
I-
U
~ 10-4
LL
W
.J
o
I::
10-6
0.0 0.5 1.0 1.5 2.0 2.5
DIST~NCE RLDNG FLRHE (MHl
CHCI3, CzHCIs , C3HCI s·
Phosgene (COCI 2) is an important intermediate in C2 HCI 3 combustion, both under

fuel-rich and fuel-lean conditions, and the model predictions are in reasonable agreement with
the experimental data (see Figure 4). As seen from Figure 11, COCI 2 forms early in the flame
primarily as a consequence of reaction C2CI 3 +02 =COCI 2+COCI, and to a lesser extent by
CHCI2 COCI +CI = CHCI2 +COCI 2. The destruction of COCI 2 occurs essentially due to CI radical
attack, i.e. COCll +CI =COCI +CI 2, which also is one of the major pathways responsible for
the formation of CI 2 and CO.
CzCI4 also is an important chlorocarbon intermediate both in fuel-lean and fuel-rich
flames of CzHCI 3 , and the mechanism predicts its behavior along the flame zone with reasonable
accuracy (see Figure 4). The reaction CzCI3+CI2=C2CI4+CI is majorly responsible for the
formation of CZCI4, with minor contributions from C2CIs =C 2CI4 +CI where C2CIs forms by
CzHCls +CI =C2 Cls + HCI. The destruction of C2CI 4 occurs primarily by the reversal of its major
formation reaction, i.e. C2 CI4+CI =C2CI3+ CI2 later in the flame. Additionally C2 CI4 reacts via
C2CI3 +C 2CI4 = C4C1 7 , which also represents the major path for the formation of C4 C16 •
C2 CI2 forms in substantial quantities in flames of C2HCI3 both under fuel-lean and
fuel-rich conditions. As seen in Figure 4, considerable discrepancy exists between model
predictions and experimental profiles for C2 C1 2 • This, however, is not surprising in view of
uncertainties that exist both in the experiments and in the mechanism. Uncertainties in the
experimental data are substantial because CzCl z mole fraction profiles were determined in the
absence of a calibration standard. The elementary reactions of C2 Cl z are also virtually unknown,
thus its reactions presented in Table I correspond to plausible steps in the absence of other
56
(T1
o
x
""'
(;loO.s
~
lJ
lJ
"-
W
..J
~·1.0 C,HCI, +CI =C,CI, + HCI(6)
'-' C,HCI, +C,CI, =C.HCI,(89)
C,HCI, =C,C1, +HCI(3)
W C,HCI, +CIO =CHCI,COCI +CI(9)
f-
a:
a:
~·1.s CZHC1 3 NET RATE
H
f-
lJ
a:
w
a:
·2·~.O 0.5 1.0 1.5 2.0
DISTRNCE RLDNG FLRME (MM)
Figure 8. Rate profiles for reactions having the highest net rates
that involve ~HC13'
evidence. According to the mechanism, the formation of C2Cl2 is due to the decomposition of
C 4Cls , C 2C13 , and C 2 HCl 3 in an approximate order of importance, i.e. C4Cls=C1CI3+C1CI1'
C 2CI3 +M= C 2CI2+CI +M, and C2HCl 3=C 2CI2+HCI. C2Cl 2 destruction was then proposed to
occur via the complex process C2Cl2+02 = COCI +COCI, although it is unlikely to correspond
to a simple elementary chemical reaction.
Carbon tetrachloride also forms in sl,lbstantial quantities in fuel-rich flames of C1HCI3.
In the proposed mechanism major reaction sequences responsible for the formation of CCl 4
were: C1CI3+02=C2CI20+CIO, C 2CI20+CI=CCI3+CO followed by the chlorination of the
CCl3 radical by CCI3+CI1 =CCI 4+Cl and CCI3+C1HCI3=CCI4+C2HCI2' and directly by the
reaction C 2Cl3+ClO=CCI 4+CO. Minor routes for CCl4 formation also exists, and they include
CCl3COCI +Cl = CCI3CO+CI2 followed by CCl3 CO = CCl 3 +CO, and CHCl3 +Cl = CCl3+ HCl,
again followed by the chlorination of CCI 3 •
For dichloroacetyl chloride (CHCI 2COCI) and trichloroacetyl chloride (CCI3COCI), there
exists discrepancies between the model predictions and the experiment (FIgure 5). The model
correctly predicts the relative ordering and the peak locations in CHCl1COCI and CCl 3 COCI
profiles. However, the model underpredicts the absolute mole fractions for these species by a
factor of 2. In this case, the major uncertainty rests in the reaction mechanism, as the
experimental data have been estimated to be accurate within 20% (Chang and Senkan 1988).
The reactions of acyl chlorides in flames are poorly understood, and the reactions presented in
Table I represent plausible steps in the absence of direct experimental evidence. According to
the proposed mechanism the major reaction paths responsible for the formation of acyl chlorides
57
~Or---~--~--~---'----r---,----r--~
(I)
o
..... HCL NET RRTE
x
'"'~ 1.5 c,HCI,+CI=c,CI,+HCI(6)
(J) C,HCI,+CI-C,CI, + HCI(8S)
"
c,HCI, -c,CI.+ HCI(3)
LJ CHCI, +CI=CCI, + HCI(S7)
LJ
"
W
.J
~ 1.0
....
W
I-
a:
a::
6 0•5
H
I-
LJ
a:
w
a:
O·~.O 0.5 L.O 1.5 2.0
DIST~NCE ~LONG FL~HE CHHJ
that involve HCl.
'"....o
x
NET RIlTE
81.0
en c,CI,+CI.=c,CI,+CI(17)
"
CCI,+CI.=CCI,+CI(62)
LJ COCI+CI=CO+Cl,(l20)
LJ ~ rt4SJ COCl. +Cl =COCl+Cl.(l4S)
"
W
.J
/
\
'
~ 0.5 I \
II '\
\,J
W I \
~ (62) I (120) \
a: " "."'-;'7~
~ 0.0 1---.....,"".....,-:::-.-:.. ;/~/ ··-~-~~-=. G~
H , .
I- :
LJ i
a: I
III I
a: / (17)
\ ,
-O.S.""o--'--1--...L'''''-'-'-o--I..---1S-.l..-...L--..L...--'----'l..--l
-0 1. 1. 2.0 2.5 3.0
DIST~NCE ~LONG FL~HE CHNJ
that involve C12 •
58
III NET RATE
...ox
ro
eso)
U
W
(J)
~ 0.01--::;""
,..........-..
_......- ....
U
"-
W
.J /l'
o I
I: I
I
COCI,+CI-COCI+CI,(14S)
\J I
I
I C,CI,+O,=COCI,+COCI(3I)
W I
I CCI,+O,=COCI,+CIO(60)
I- : CCI,+0-COCI,+CI(61)
~·0.5 I
:
/
z I
o
H
I- (145)
U
ct
W
a:
-l.~.o 3.0
that involve COCI2.
are due to CIO addition to C2HCl3 and C2C14, i.e. C2HCI3 +CIO=CHCI2 COCl+CI and
C2 CI 4+CIO=CCI 3 COCI+CI. Because acyl chlorides are very unstable at high temperatures,
they are expected to decompose, for example by the following sequence of reactions:
CHCl2COCI +CI = CHCl2 CO + Cl 2 and CCl3COCI + CI =CCI3CO +CI 2 , followed by CHCl 2 CO =
CHCI2 +CO and CCI 3CO=CCI3 +CO, respectively. These reactions also constitute pathways
which subsequently lead to the formation of CHCl 3 and CCl 4 by the chlorination of the CHCI 2
and CCI3• respectively.
As seen in Figure 6 major discrepancies exist between DCKM predictions and the
experimental data. Specifically the second maxima in the mole percent profile for CzCl6 was not
reproduced at all by the present model, and this clearly manifests the existence of a deficiency
in the mechanism. According to the model, the earlier formation of C2Cl6 is majorly due to
CCl3 recombination, i.e. CCl3+CCI3 =C2C~, and not via the chlorination of C2C1 4, since the
conditions do not favor the reaction Cl + C2Cl4= C2CIs in the forward direction in the flame. The
destruction of C2Cl6 occurs via its unimolecular decomposition back to CCl3 and by C-Cl bond
fission, i.e. C2CI6=CCI3+CCI3 and C2CI6=C2Cls +CI, and by complex molecular elimination
process, C2CI6=C2CI4+CI2. The increase in C2Cl6 mole fraction for the second time would
likely to be caused by the fragmentation of higher molecular weight chlorohydrocarbons, which
form early in the flame zone. The formation of C3 Cl6 starts by the addition of CCl3 to Cl HCI 3,
C2Cl4 or Cl CI2 , i.e. CCl 3 +C 2 HCI 3=C 3 HCI6, CCl3 +C2C14=C 3 C1 7 and CCl3 +C 2Cl l = C3Cl s, all
of which are chemically activated processes, followed by C3 HCI 6+Cl= C3CI6+HCl,
C3CI 7 =C3CI6+CI and C3Cls+CI2=C3C16+Cl. The destruction of C3Cl6 occurs primarily
through CI radical attack, i.e. C3CI6+CI=C3Cls+CI2. The major route for the formation of
59
C4CI6 begins with the addition of C2CI3 to unsaturated C2 chlorohydrocarbons, i.e.
C2C13+C2HCI3=C4HCI6' C2CI3+C2C14=C4CI7 and C2C13+C2CI2=C4CIs, followed by
C4HC1 6+Cl=C4C16+HCI, C4C17= C4CI6+CI and C4Cls+CI2=C4CI6+Cl. In the current
mechanism, the C4Cl6 formed undergoes further addition reactions via C2C13+ C4C16 =C 6C1 8 +CI
or C4CI6+CI=C4CIs+C12 followed by C2CI2+C4Cls=C6C17' and the C6Cl8 and C6Cl7 formed
subsequently are transformed into C6C16 via C6C1 8- > C6C16+ Cl2 and C6Ch- > C6Cl6+ CI,
respectively.
The major reaction paths for the formation of CHCI3 were CHCl2+C12=CHCI3+CI and
C2HCI3+CHCI2=CHCI3+C2HCI2' in which the CHCI2 radical forms by CHC12COCl+CI=
CHCI2CO+CI2, followed by CHCI 2CO=CHCI2+CO. The destruction of CHCl3 occurs
primarily by CI radical attack, i.e. CHCI3+CI= CCI3+HCl. For C2HCls, its major formation
paths were the radical recombination reactions CCI3+CHCI 2=~HCls and Cl+C2HCl4=
C2HCls, both of which are chemically activated processes. Its major destruction path also
involves CI attack, i.e. C2HCI s +CI=C 2Cls +HCI. The C3HCIs forms via C3HCl6 = C3HCI s
+ CI, in which the C3HCl 6 radical is produced from CCl3+C 2HCI 3=C 3HCI 6, again a
chemically activated process. The destruction of C3HCIs again proceeds through CI radical
attack, i.e. C3HCls +CI =C 3Cls + HCl.
In Figures 12 and 13, the major reaction pathways responsible for the formation and
destruction of species discussed above, as well as for several key radical species, are presented.
~CHCI,
CI
CI
CI
(2HC14 --;-- CzCl" Polycarbon
'---.--~ CCI, CHCs
M
2) M 4/( 2!~!022
CI CI M CzCI"
C CI _ _ (Cl _z-+ C Cl - C Cl Polycarbon
J1
z CHCs
CI 2 COCI2~COCI
C2C 1.0
! ~,~u,eel
! CI O2
C CI ........2 eCI ' C

~2_ CCI 4 CO
1
2 • I
e1 2 • Hel
Inhibiled
CHCI, CO 2
Figure 12. Major reaction pathways for the formation and destruction of C 1 and C 2
species.
60
SOOT
C, CI. Cz CI 6
l~'
C,CI 4
/
Figure 13. Major reaction pathways for the formation and destruction of C3 to C 4
species.
As seen from these figures, the high temperature reactions of C2 HCI 3 are considerable complex
but not intractable. The reaction rate and sensitivity analysis provided crucial information on the
identities of important reactions, which in return can be used to construct simpler, albeit with
reduced range of applicability, reaction mechanisms.
Concluding remarks
As shown in the above discussion and example for C2 HCl 3 combustion, detailed chemical
kinetic mechanisms (DCKM) provides useful insights into reaction processes, especially with
regard to the trace constituents. Since DCKMs involve the participation of a large number of
species in very large number of reactions, to preserve the elementary n<tture of these models,
they represent a convenient framework to describe the formation and destruction of major as well
as minor products in manufacturing processes. In addition, DCKMs, because of their
fundamental bases and predictive features represent numerical tools for the rational design of
clean manufacturing processes, i.e. processes with minimal formation of toxic by-products.
Acknowledgements
This research was supported, in part, by funds from the U.S. Environmental Protection Agency,
Grant No:RSI5136-01-0, and the UCLA Center for Clean Technologies.
61
REFERENCES
Basevich, V. Ya., "Chemical kinetics in the combustion processes: A detailed kinetic mechanism
and its implementation", Prog. Energy. Combust. Sci.,.u, 199 (1987).
Baulch,D.L., Duxbury,J., Grant,S.J., Montague,D.C.," Evaluated Kinetic Data for High

Temperature Reactions: Homogeneous Gas Phase Reactions of Halogen and Cyanide Containing
Species", ~, J.Phys.Chem.Ref.Data, 1O,Supplement No. 1.(1981).
Benson,S.W., Thermochemical Kinetics, John Wiley, N.Y. (1976).
Benson, S.W., and O'Neal, H.E., "Kinetic data on gas-phase unimolecular reactions",
NSRDS-NBS 21, 1970.
Berces, T., and Marta,F., "Activation Energies for Metathesis Reactions of Radicals", in
Chemical Kinetics of Small Organic Molecules: IV Reactions in Special Systems, Alfassi, Z.B.,
Ed., CRC Press, Boca Raton, Florida 1988.
Caracotsios,M., and Steward,W.E., "Sensitivity Analysis of Initial Value Problems Including

ODEs and Algebraic Equations", Compo Chem. Eng., 2,359 (1985).
Chang,W.D., Karra,S.B., and Senkan,S.M.," A Detailed Mechanism of the High Temperature

Oxidation of C2HCL3", Combust.Sci.Tech, v.49. p.107 (1986).
Chang,W.D., and Senkan,S.M.," Chemical Structure of Fuel-Rich, Premixed, Laminar Flames

of Trichloroethylene", 22nd Symposium (Int'!) on Combustion, 1453 (1988).
Chang, W.D., and Senkan, S.M., "Detailed Chemical Kinetic Modeling of the Fuel-Rich Flames
of Trichloroethylene", Enviro. Sci. Tech., n, 442 (1989).
Chase, M.W., Davies, C.A., Downey, J.R., Frurip, D.J., McDonald, R.A., and Syverud,
A.N., "JANAF Thermochemical Tables", J. Phys. Chem. Ref. Data, H, Supplement "No.1,
(1985).
DeMore, W.B., Molina, M.J., Watson, R.T., Golden, D.M., Hampson, R.F., Kurylo, M.I.,
Howard, C.I., and Ravishankara, A.R. "Chemical Kinetics and Photochemical Data for Use in
Stratospheric Modeling", Evaluation No.8, JPL Publication 90-37, 1990.
Dunning, T.H., Harding, L.B., Wagner, A.F., Schatz, G.C., and Bowman, J.M., "Theoretical
Studies of the Energetics and Dynamics of Chemical Reactions·, ~, 240, 453 (1988).
Gaffney,I.S., Bull,K., "Prediction of the Rate Constants for Radical Reactions Using
Correlational Tools· , in Chemical Kinetics of Small Organic Molecules: IV Reactions in Special
Systems, Alfassi, Z.B., Ed., CRC Press, Boca Raton, Florida 1988.
Granada, A., Karra, S.B., and Senkan, S.M., "Conversion of CH4 into C 2 H2 and C2 H4 by the
Chlorine-Catalyzed Oxidative-Pyrolysis (CCOP) process: Oxidative pyrolysis of CH3CI", Ind.
Chern. Eng. Res., 26, p.1901 (1987).
Hanson,R.K., Salimian,S., "Survey of Rate Constants in the N/H/O System", in Combustion

Chemistry, (W.C. Gardiner, Jr., Ed.), Springer-Verlag, New York, 1984.
62
Hwang, J-T., "Sensitivity Analysis in Chemical Kinetics by the Method of Polynomial
Approximations", Int. J. Chem. Kinetics, .u, 959 (1983).
Junk,.G.A., and Ford,C.S., "A Review of Organic Emissions from Selected Combustion
Processes", Chemosphere, v.9, p.187 (1980).
Kee,R.J., Grear,J.F., Smooke,M.D., and Miller,J.A., "A Fortran program for modeling steady
state laminar one-dimensional premixed flames", Sandia Report (1985).
Kerr,A.J., Moss,S.J., Handbook of Bimolecular and Termolecular Gas Reactions, Volumes I

and II, CRC Press, Boca Raton, FL 1981.
Kerr,A.J., Drew,R.M., Handbook of Bimolecular and Termolecular Gas Reactions, Volumes

III and IV, CRC Press, Boca Raton, FL 1987.
Kondratiev, V.N. "Rate Constants of Gas Phase Reactions, COM-72-10014, National Bureau
of Standards, Washington, 1972.
Laidler, K.J., Chemical Kinetics, 3rd Ed., Harper and Row, New York 1987.
Lutz, A.E., Kee,R.J., and Miller,J.A., "SENKlN: A Fortran Program for Predicting
Homogeneous Gas Phase Kinetics with Sensitivity Analysis", Sandia National Laboratories,
Livermore, SAND87-8248, February 1988.
Oberg,T., Aittola,J-P., and Bergstrom,I.G.T., "Chlorinated Aromatics from the Combustion of

Hazardous Waste", Chemosphere, v.14, p.215 (1985).
Pedley, I.B., Naylor, R.D., and Kirby, S.P., Thermochemical Data of Organic Compounds,
Chapman and Hall, London 1986.
Rabitz, H., Kramer, M.A., and Dacol, D., "Sensitivity Analysis in Chemical Kinetics", Ann.
Rev. PhyS. Chern .. ~, 419 (1983).
Senkan, S.M., "Detailed Chemical Kinetic Modeling: Chemical Reaction Engineering of the
Future", Adyances in Chemical Engineering, in press 1992.
Schaub, W.M., and Tsang, W., "Dioxin Formation in Incinerators", Enviro. Sci. Tech .. v.17,
p.721 (1983).
Siegneur, C., Stephanopoulos, G.,and Carr,Jr.,R.W., "Dynamic Sensitivity Analysis of

Chemical Reaction Systems: A Variational Method", Chern. Eng. Sci., n, 845 (1982).
Tilden, J.W., Costanza, V., McRae,G.J., and Seinfeld, I.H., "Sensitivity Analysis of
Chemically Reacting Systems", in Modeling of Chemical Reaction Systems, K.H. Ebert, P.
Deuflhard, and W. Iaeger, Eds., Springer-Verlag, New York, 1981.
Tsang, W., and Hampson, R.F., "Chemical Kinetic Data Base for Combustion Chemistry. Part
I. Methane and Related Compounds", J Phys. Chem. Ref. Data, .u, 1087 (1986).
Warnatz, 1., "Rate coefficients in the CIHIO system", in Combustion Chemistry, (W.C.
Gardiner, Ir., Ed.), Springer-Verlag, New York, 1984.
63
Westbrook, C.K., and Dryer, F.A., "Chemical Kinetic Modeling of Hydrocarbon Combustion",
Prog. Energy Combust. Sci., lQ, 1 (1984).
Westley, F., "Table of Recommended Rate Constants for Chemical Reactions Occurring in
Combustion", NBS Report, NSRDS-NBS 67, 1980.
Westley,F., "Chemical Kinetics of the Gas Phase Combustion of Fuels", NBS Special
Publication No:449, 1976.
64
KINETIC CONTROL OF BY-PRODUCf REACfIONS WITH IN SITU
DISTILLATION IN A CONTINUOUS FLOW DEVICE
Robert P. Shirtum
Dow Plastics
The Dow Chemical Company
Freeport, TX 77541
A detailed kinetic study of a complex, commercially important, organic synthesis

was combined with math models of several reactor configurations to yield a design which
proved viable for large-scale production. The combined characteristics of in situ
distillation, plug-flow behavior, and continuous reagent addition were proven in pilot
plant operations.
INTRODUCTION
A review of the volumes of SRI PEP [1] reports indicates, in general, that many
industrial processes, to produce bulk building block commodity chemicals have yields
to desired products of about 90%. (Certain hydrocarbons, specialties, and plastics
products not included.) Our own experience and informatIOn from other major
chemical. companies supports this average of ten percent yield loss to by-products.
Many of these processes require substantially more capital'jer unit of
production, to destroy the by-products than the allocated capital require to produce
them initially.
The cost of waste destruction has recently escalated so rapidly as environmental
regulations and concerns have received unprecedented emphasis, that many petro
chemical companies are utilizing much of their capital resources to address waste
treatment.
To eliminate a substantial portion of this large volume of by-products by the
research and design of more efficient processes, rather than focus on destroying these
materials, would justify the expenditure of a very large effort.
This work is a successful example of the development of technology which
increases the process yield of a major chemical product from about 90% yield, of one
of the main reactants, to over 99% yield.
Other positive attributes of the new process were increased yield of a secondary
feedstock from 55% to almost 100%, improved ability to recover and recycle spent
catalyst, and several product quality improvements. Slight increases in conversion
energy and capital, due to increased process complexity, were insignificant relative to
waste treatment cost reduction and raw material yield improvement.
Consider the industrial synthesis of an important organic chemical product
where the reaction sequence can be simplified in the following manner:

C,S2
A + B -----> D (1)
D + C -----> Cs + PI + Sj (2)
PI + B -----> E (3)
E + C -----> Cs + P2 + Sj (4)
These main reactions are accompanied by the following by-product reactions:
C
A + S -----> F (5)
A + C -----> G (6)
A + D -----> H (7)
PI + C -----> D t (8)
The effects on product quality of these series of reactions are as follows:
Pl/P2 = constant
(D,E,G,H) < 2500 ppm by weight.
Additional process desirability conditions include:
F = 0 (high yield to products from A) C s = 99.5 + % purity (for ease of
regeneratIOn)
EXPERIMENTAL APPROACH
The development of the kinetic model, as simplified above, began with a
computer model based on stoichiometry and estimates of Arrhenius constants from
previous studies. We used an approach in which the model was used to design
subsequent laboratory reactions. Data from the experiments were then compared to
the model and sensitivity analysis focused on weaknesses in model predictions. The
model was then refined In an iterative fashion until acceptable prediction over a wide
range of conditions were established. We believe that it is significant that a model
preceded the first experiment and we refer to this methodology as model based
experimental design.
The initial laboratory experiments were conducted in a semi-batch fashion with
a conventional glass apparatus as illustrated in Figure 1. The reactor is fitted with a
rectification column, condenser, and reflux reservoir to facilitate the removal of
catalyst solvent and volatile by-products.
RESULTS
Model predictions and semi-batch laboratory data at simulation conditions are
shown for the major reaction components in Figure 2. Products of trace chemistry
and model predictions are illustrated in Figure 3.
The model is then solved numerically usin~ the design equation for a serial
continuous stirred tank reactor (CSTR) configuratIOn and the economically optimal
number of CSTRs is determined [2]. A graphical representation of this approach is
illustrated in Figure 4, where the critical ratio of PI7P2 is plotted as the dependent
variable for the design basis.
66
Figure 1. Semi -batch laboratory reactor.
100
&0
so PI
1II
~
1= III
~
Z 51
~ «I
~ 30
20
10
Q2 ~4 ~6 ~,
DlMENSICNLESS 11 t.IE
Figure 2. Concentration vs. time for semi-batch laboratory
and computer model data.
0000 ~------------------------------~~
e-
o..
.e. 2IlOO
z
0
~
a:
~
w 1000
U
Z
0
U
o 0 .2 0.0 O.C! 0.8
DIMENSIONLESS TIME
Figure 3. Computer model predictions of concentration

vs. time for side reaction trace chemistry.
67
18,----------------------------------,
BATCH
14
12
o 10
~ 8
o+------.------.------.------.-----~
o 10 20 30 40 50
#CSTRs
Figure 4. Isomer ratio Pl\P2 vs. number of CSTR stages.
This comparison between serial CSTRs and batch reactor results is not unlike
the efficiency comparison made by Octave Levenspiel [3], except that the conversion
of the limitmg reactant of interest to us is above 99.9% and the PI/P2 ratio is of
much more pratical interest than volumetric efficiency.
Reaction strategy learned by using a simplified math model:
Maintain a low catalyst concentration, [C], by adding it slowly to
the reaction mixture in concert with the rate of catalyst
depletion in Reaction 2.
Removal of the by-product, S, as its azeotrope with A via in situ
distillation.
Add an unreactive solvent, S2, to insure miscibility of the
catalyst.
Control the product isomer ratio of Pl/P2, using the initial
stoichiometric ratio of AlB.
Later studies were conducted in a horizontal glass pipe with baffled partitions
allowin~a multi-stage CSTR configuration within one apparatus containing a vapor
space In common communication with each stage. Reflux, rectification and
condensation equipment were similar to the batch apparatus.
REACTOR DESIGN
The descriptive model derived from semi-batch experiments was solved on the
computer by Runge-Kutta [4] techniques. The equations for reaction velocities, and
heat and mass balances were then solved by a Quasi-Newton [5] method to estimate
the number of CSTR stages necessary for control of the important ratio of PI to P2 in
the product. Catalyst and solvent were added inter-stage in the continuous reactor in
an optimized profile to aid in achieving the correct PI to P2 ratio with a minimal
number uf c.vmpartments, to minimize reactor cost. Continuous addition of the
catalyst greatly helped control the very fast exothermic reactions so that area
requirements for enthalpy transfer were much easier than in the semi-batch case.
Although a simplified model, illustrated in Equations (1-8) is presented, the model
we used required over 30 chemical reactions, reversible and irreversible, and over 35
anionic and chemical species to accurately describe the trace chemistry (21
equations) along with the seven major reaction pathways. The reactor system WhICh
evolved is illustrated in Figure 5.
68
Figure 5. Continuous horizontal staged CSTR reactor.
This reactor allows for the staged addition of catalyst and co-reactants and the
removal of solvents and volatile products of reaction just as the laboratory semi-batch
reactor did. This is the design that was pilot plant tested to produce 1000 tons per
year of product. The manufacturing scale-up from this pilot plant will be
approximately two orders of magnitude.
The start-up philosophy, which proved successful, was to begin reaction by
introducing catalyst while operatin~ the fully charged reactor on total recycle. At
high conversion of B, catalyst additlOn ceased and the desired steady-state forward
flowrate was begun with catalyst profile initiated sequentially to each CSTR sta~e
after the appropriate residence time had elapsed. The reaction model is a great aId
in implementing start-up and shut down control schemes.
CONCLUSIONS AND DISCUSSION

An environmental solution to the production of a chemical product can be
approached in several ways:
1. Cease production of the product.
2. Add process unit operations to destroy process wastes.
3. Modify the process to reduce or eliminate the production of wastes.
We believe the options are listed in order of increasing desirability. One
exception to this order may be where an alternate product, which fulfills market
needs and is produced by a more environmentally favorable process allows the
replacement of the original product.
Resources spent to add unit operations to destroy process wastes are necessarily
less productive toward the mission of production of the product than resources spent
toward making the process more efficIent. The magnitude of resources may be less;
however, if applied toward the destruction of wastes. This is especially true if short-
term economics are the decision driving forces. As so often occurs, environmental
goals are not so well defined and they are constantly becoming more restrictive.
Dr. W. Edwards Deming [6], the noted statistician and quality pioneer, has said
that there are three things a manufacturer, interested in contmued profitability, must
do to be competitive. They are:
1. Innovation of product
2. Innovation of process
3. Improvement of existing process
69
It is interesting to notice the balance implied between product and process work,
perhaps one-third and two-thirds respectively. A recent report by Suzanne Berger
(et. al.) [71, which tries to compare American and Japanese industry, states that a
marked difference between R&D expenditures exists. American companies devote
one-third of their R&D resources against process technology while product related
activity consumes the other two-thirds. In Japan these roles in R&D are reversed,
this is not surprising as Dr. Deming created the systems model in practice in
industrial Japan today.
We beheve that the age old philosophy of "if it ain't broke, don't fIX it" will no
longer serve well in our new global competitive environment. The new philosophy
should be to ask "Where is our process least perfect?" and then to apply prioritized
resources to judiciously improve continuously.
The cost to adopt a renewed expenditure toward process development may not
prove popular in time of corporate belt tightening but we believe it IS necessary for
long-term business viability.
REFERENCES
1. SRI International, "Process Economics Program (PEP)," Reports: 2-2D,
11-l1C, 22-22B, 33-33B, and 38-38A
2. J.M. Smith, Chemical En~ineerin~ Kinetics, 2nd Ed., McGraw-Hill,
New York, 1970, pgs. 110-111, 184-193, and 236-239.
3. Octave Levenspiel, Chemical Reaction En~ineerin~, John Wiley and
Sons, New York, 1967, pgs. 140-141.
4. B. Carnahan, H.A Luther, and J.O. Wilkes, AI!plied Numerical Methods,
John Wiley and Sons, New York, 1969, pg. 363.
5. Raghu Raman, Chemical Process ComWItations, Elsevier Applied Science
Publishers Ltd., New York, 1985, pgs. 539-540.
6. Dr. W.E. Deming, Out of the Crisis, Massachusetts Institute of
Technology, 1986, pgs. 5-26.
7. S. Berger, M.L. Dertouzos, RK. Lester, RM. Solow, and L.c. Th'lrow,
"Scientific American," Vol. 260, Issue 6, June 1989, pgs. 39-47.
70
WASTE MINIMIZATION AND REMEDIATION
IN THE METAL FINISHING INDUSTRIES
Clifford W. Walton, Kevin S. Briggs, and Kevin J. Loos
University of Nebraska-Lincoln
Department of Chemical Engineering
236 Avery Laboratory
Lincoln, NE 68588-0126
INTRODUCTION
An overview of waste assessment procedures and a knowledge of process options

for remediation and recycle are necessary for a successful waste minimization program.
Factors to consider when constructing a waste minimization program include U.S. EPA
priorities, constructing an efficient assessment program, elements of previously suc-
cessful programs, and opportunities relating to both the process and procedure of the
particular system. This work describes the basic steps necessary to fulfill all the factors
mentioned in order to create a successful and enduring waste minimization program in
the metal finishing industries. Existing and potential technologies are discussed. A vail-
able references are presented that assist in the process of analyzing and constructing
a suitable program for the particular situation, including descriptions of selected case
studies related to surface finishing.
There is a need for more efficient methods for minimization of hazardous waste
brought on by the increasing public concern about waste management. Various federal
and state hazardous and solid waste regulations have made it more difficult and ex-
pensive to dispose of unwanted by-products, prompting prevention techniques. Some
undesirable wastes eventually will be banned from land disposal completely, while other
disposal methods will be restricted considerably. These restrictions have forced a short-
age of approved hazardous waste disposal facilities, causing rapidly rising disposal costs.
Due to the financial and legal incentives of reducing or completely eliminating unwanted
by-products, and that generators of these by-products are required to certify that they
have instituted a waste minimization program, it has become necessary and beneficial
to implement a waste minimization program!.
Figure 1 shows a simplified view of a chemical or product manufacturing process
with special attention to discharges 2 • By developing a waste minimization program
a manufacturer can save money by reducing waste treatment and disposal costs, raw
material purchases, and other operating costs. Also, successful programs are necessary

-.j
'"
Reactants w/impurities
Contaminants
Reaction media Atmospheric

(water, organic emissions
solvents) MANUFACTURING
PROCESS
Recycle
Catalysts
Figure 1. Manufacturing process from the viewpoint of discharges and waste minimization.
Table 1. U.S. EPA Priorities for Waste Management
Rank Procedure/Process
1 Source Reduction
2 Recycling
3 Incineration/Treatment
4 Secure Land Disposal
to meet national, state and local waste minimization policy goals, reduce potential
environmental liabilities, protect public and worker health and safety, and protect the
environment 3 , as well as promote a better corporate image to the public.
An important step in eliminating or reducing undesirable by-products is to first
identify them. Some common by-products generated by metal finishers include indus-
trial treatment sludge containing toxic metals such as cadmium, copper, chromium
and nickel; spent plating and process baths, including various cleaners and contami-
nated solvents; and miscellaneous solid wastes such as absorbents, filters and empty
containers 4.
u.s. EPA PRIORITIES
The former U.S. Environmental Protection Agency (EPA) Administrator, William

Reilly, announced pollution prevention as the environmental goal for the 1990'S5. Ta-
ble 1 shows the ranking of priorities for waste management established by the U.S.
EPAl,6,7, with particular emphasis on source reduction.
Source reduction is the concept of preventing or minimizing the production of mate-
rial which must be treated as waste by altering the actual production process. Measures
include process modifications, feedstock substitutions, improvements in feedstock pu-
rity, improvements in housekeeping and management practices, increases in efficiency
of machinery, recycling within a process, and production of valuable by-products.
Recycling is the use or reuse of undesirable by-products as an effective substitute
for a commercial product or as an ingredient or feedstock in an industrial process. It
can occur on- or off-site and includes the reclamation of useful components from a
waste material or the removal of contaminants from a waste to allow it to be reused.
It is often less expensive to recycle a chemical than it is to purchase new material and
pay for disposal costs. Though a material may no longer meet the specifications for a
process in which it is being used, the material may still be suitable for other uses at the
facility. The ability to reuse is aided by adequate waste separation and concentration.
This also increases possible waste exchange possibilities, where one industry's unwanted
by-products are another's raw material.
Waste treatment is any method, technique, or process that changes the physical,
chemical, or biological character of any hazardous waste in a way that neutralizes the
waste, or renders such waste nonhazardous, less hazardous, safer to manage, able to
be recovered, able to be stored, or reduced in volume. This includes recovery of energy
content by using a waste as a fuel supplement, particularly for organic solvents and
oils. Incineration is a last resort for a material, such as dioxane, that can only be made
non-hazardous by complete oxidation at high temperatures.
Disposal is the discharging, depositing, injecting, or placing of hazardous waste into
or on any land. If a waste has hazardous characteristics, then disposal in a regulated
73
hazardous waste disposal facility is required. This secure land disposal is the least
desirable option. If the waste is treated to eliminate these hazardous characteristics,
it can be disposed of in a non-hazardous waste landfill, which is considerably less
expensive.
The use of the various options depends on extensive federal and state regulations.
Table 2 presents a brief list of the prominent laws and regulations that have an impact on
the surface finishing industry. Note that the list cannot be construed as comprehensive
and any interested party should contact the appropriate federal, state, or local authority
regarding the specifics of these and current regulations.
ESSENTIAL ELEMENTS AND COMMON BARRIERS OF

SUCCESSFUL MINIMIZATION PROJECTS
Various studies of waste minimization programs have identified characteristics of

successful programs. Three key elements found were}
• Management's active involvement and encouragement at all levels,
• Production personnel strongly motivated to implement and maintain necessary

changes, and
• Technologies used were elegant in their simplicity.

At least one of these elements were missing in each minimization project failure l .
People involved in implementing waste reduction programs have noted several in-
hibiting factors which must be identified by those attempting to develop a waste reduc-
tion strategy. This identification is necessary so that the resistance can be recognized
and overcome. Barriers to waste reduction include4
• Lack of information about available waste reduction techniques and the benefits
that can be achieved.
This can be overcome by seeking information on other's successes. Examples can
be found in the case studies published by the U.S. EPA8,9.
• Concern for upsetting product quality.

Both operators and management must be convinced that pollution prevention
changes will not adversely affect quality. This requires a commitment to and
effective application of a new process or technology, including complete training
of operators.
• The "if it ain't broke-don't fix it" attitude.
• A reluctance to develop innovative ideas because of the fear of failure.

Management must take the initiative and absorb all effects of failures, only then
will a commitment from operators and middle managers be made.
• The attitude that a new technology will not succeed because it is outside the
company's normal range of expertise.
This attitude problem can be solved by seeking outside views and help. State
programs, consultants, and professional organizations such as the American Elec-
troplaters and Surface Finishers Society can assist in answering a company's ques-
tions.
74
Table 2. Some Environmental Laws and Regulations Pertinent to the Surface Finishing
Industry4 (1989)
Regulation Description
40 CFR 122, NPDES Federal regulations governing the discharge of wastewaters
to surface waters of the United States.
40 CFR 413, 433 Federal regulations specifying effluent limitations,

pretreatment standards, and new source performance
standards for the electroplating and metal finishing
industries.
40 CFR 268 Federal regulations that restrict the disposal of spent

solvents and solvent-containing wastes.
CCR Title 23 State regulations governing the discharge of wastewaters to

Subchapter 9 surface waters. Includes provisions for issuance of permits
and setting effluent limitations.
Local municipal Discharge requirements set by local POTWs restricting the

codes addressing concentrations of pollutants in wastewaters discharged to
discharges to sanitary sewers.
POTWs
CCR Title 22 Restricts the land disposal of certain liquid and solid
Sections 66900 hazardous wastes and sets time schedules for implementing
and 66905 the restrictions.
CHSC Chapter 6.95 Requires local government agencies to implement hazardous

material management programs requiring local businesses to
submit applications for the storage and handling of
hazardous materials.
CCR Title 22, Division 4 Sets requirements for generators of hazardous wastes,
Chapter 30, Article 6 including restrictions on how long wastes can be
accumulated without the storage facility being permitted.
CHSC 25202.9 Requires certification, by waste generators permitted as TSD

facilities, that a waste minimization program is in operation
and that the treatment, storage, and disposal methods
minimize the threat to human health and the environment
CFR - Code of Federal Regulations

CCR - California Code of Regulations
CHSC - California Health and Safety Code
NPDES - National Pollution Discharge Elimination System
75
When developing and implementing a waste reduction program it is essential that
management be committed to pursuing waste reduction, be willing to experiment with
various ideas, and be prepared to experience failure, as well as success 4 •
WASTE MINIMIZATION ASSESSMENT PROCEDURE
A waste minimization assessment procedure begins with the recognized need to min-
imize waste. The U.S. EPA has published a detailed description of such a procedure3 ,
which can be broken into a
1. Planning and organization phase,
2. Assessment phase,
3. Feasibility analysis phase, and

4. Implementation phase.
The Planning and Organization Phase of developing a successful waste minimization
(WM) program begins with achieving support from the management level. Manage-
ment support of a program will occur only if it is convinced that the benefits obtained,
such as economic advantages, compliance with regulations, reduction in liabilities, im-
proved public image, and reduced environmental impact, will outweigh the costs. A
formal policy is the best way to convey to the company's objectives to the employ-
ees. A policy allows the employees to focus on what they should be achieving and is
a starting point to getting the employees involved. Commitment throughout an entire
organization, not just at the management level, is necessary for the implementation of
successful programs. Since employees are most often responsible for the generation of
wastes, they can be very important to the overall success of a program.
Task forces are one of the first steps a company should take when implementing a
WM program. This group should include employees from every area that the program
will affect. The top priority of a task force is to set goals for the WM program. These
goals should remain consistent with the company's objectives and often it is best to
develop quantitative goals, or goals which can be easily measured. Periodically, the
goals should be reviewed in order to reflect any change in the company's objectives.
Barriers to a specific WM program must also be analyzed in order to understand what
can be achieved by a company's particular program.
After setting goals for the WM program, the Assessment Phase begins with a task
force given the responsibility to identify and characterize the facility waste streams.
Information on the waste streams can come from a variety of sources, including com-
pleted hazardous waste manifests, biennial reports, NPDES (National Pollutant Dis-
charge Elimination System) monitoring reports, and the "right to know" provisions
of the Superfund Amendment and Reauthorization Act (SARA). Once all the waste
streams have been assessed, it is necessary to prioritize these waste streams and/or
operations in order of greatest potential risk. Assessment teams can then be selected
to focus on a particular waste stream or process. These teams should include people
that are knowledgeable about the stream or process, and are often made up of internal
staff and outside people. Outside consultants can add experience about a stream or
process, especially if there are no experts in a particular area within the company.
After visiting the site, a comprehensive set of options can be generated by the
various task forces. Most of these options will come from education and on-the-job ex-
pertise, however technical literature can also be helpful. Once all the options have been
76
generated, screening of these options must occur in order to decide which options are
real possibilities in minimizing waste and reducing cost. A list of particular questions
should be answered in relation to each option. The goal of the screening is to decide
which options are suitable for a technical and economic feasibility analysis.
The next step in developing the WM program is the Feasibility Analysis Phase,
which includes a technical and economic evaluation. Now that an initial set of options
have been selected it is necessary to evaluate them from a technical standpoint, in order
determine whether an option will work in a specific application, and from an economic
standpoint, such as payback period, return on investment, and net present value.
The technical evaluation can include visiting sites to observe in-use equipment and
processes, requesting both vendor and operator comments on a particular process.
Also included in the evaluation are facility constraints and product requirements. If
an option does not appear feasible after the technical evaluation, it should be dropped
from the list.
The economic evaluation includes analysis of capital costs, operating costs, and sav-
ings. Once these have been evaluated a profitability analysis is performed. A project's
profitability is measured using net cash flows for each year of the project's life. If a
project contains no significant capital costs, the profitability of the project can be de-
cided on whether an operating cost savings occurs or not. If it is found that a project
reduces operating cost, it should be implemented as soon as possible.
Implementation of the WM option( s) is the last step of a WM program. If the
option( s) are approved by the management, it is then necessary to obtain funding from
an appropriate level within the organization. A beneficial step is to have an evaluation
team made up of financial and technical personnel. This will assure the project funders
that the sponsor is being objective, and not just enthusiastic. Also, this team will be
able to eliminate funders' attitudes of "it can't be done" and "if it ain't broke, don't fix
it." Once the funding barrier has been overcome, installation of the option can occur.
A demonstration of the option is then performed and its performance is evaluated. A
measure of effectiveness of a particular WM option is its ability to pay for itself through
reduced waste management costs and reduced raw materials costs.
Once a WM program is implemented, it is necessary to continually re-evaluate
the changes, rather than making WM a one-time effort. After the highest priority
streams or processes have been evaluated, a company should continue down the list
of priorities that were fonllulated in the beginning. If necessary, the priorities should
be re-evaluated in order to ensure that each stream or process is prioritized correctly
and is still a significant part of the WM program. An ultimate goal is to maximize the
reduction of waste, and if this has been achieved, companies should evaluate reduction
of other processes, such as industrial wastewater discharges, air emissions, and solid
wastes.
WASTE REDUCTION OPPORTUNITIES
A waste reduction program can be broken down into two broad areas: procedure
and process. The procedure includes employee training, changing methods, preventing
spills, and inventory control, while the process includes optimizing solutions, substitut-
ing products, concentrating waste, and changing equipment.
The most important aspect of the procedure is employee training. Although man-
agement commitment and direction are fundamental to a waste minimization program,
commitment throughout an organization is crucial to resolve conflicts and overcome
barriers to the program. Employees often can increase the generation of waste, so they
77
can be pivotal in contributing to the overall success of a waste minimization program.
Bonuses, awards, plaques, and other forms of recognition often times are the best way
to provide motivation and to boost employee cooperation3 •
Higgins l gives the example of a chemical engineer at a major consulting company
who spent the early part of her career as an environmental coordinator for a major
aircraft manufacturer. Solvents kept turning up in the storm drains and the source was
traced to several floor drains in the shop area where workers were pouring used solvent.
She then posted signs that read "NEXT STOP IS YOUR FAUCET AT HOME." The
violations stopped. When the workers asked her what they could do to reduce waste,
she responded, "Act as though your mother were looking over your shoulderl."
Some control methods that can be implemented include
• Minimizing the number of different raw materials and supplies used, such as
cleaning fluids, oils, etc. This helps clear-up shelf-life problems and reduces the
number of partially filled containers requiring disposal.
• Purchasing container sizes appropriate to the actual use. It can be less expensive
to buy smaller containers of perishable materials than to buy bulk quantities, at
cheaper prices, that require disposal of the expired portion.
• Reducing the inventory of hazardous materials to a minimum. It must be ensured

that containers are being rotated on the shelves so that the oldest materials are
used first. This will reduce the disposal of materials whose shelf-life has expired.
These steps will reduce spending on raw materials and waste disposal, as well as the
investment tied up in inventory. Such areas as preventing spills and inventory control
are simple procedures which can cut significant amounts of waste.
After preparing an inventory of chemicals being used and wastes produced from all
processes, an evaluation is needed to determine if the number and volume of chemicals
being used can be reduced. Optimizing solutions, taking the waste of one process and
adapting it to be used in another, substituting products, less-hazardous for hazardous
materials, and concentrating waste are possible modifications1 •
When dealing with a change of equipment, simplicity is best. Typically, on success-
ful projects, off-the-shelf equipment is adapted to a new application, while complex or
specialized equipment is avoided. The greater number of modifications attempted at
the same time, or the more experimental the equipment is, the greater the chance that
problems will occur and the process will be abandoned for more reliable and proven
methods1 •
In the surface finishing industry, there are numerous simple approaches. For exam-
ple, several improved rinsing techniques that can reduce the amount of process fluid
losses include1 ,4,lO,1l extended drip times to allow complete drainage, workpiece posi-
tioning to minimize holdup of process fluids, spray ffog rinsing to reduce the amount
of rinse water to be treated, countercurrent rinses that can drastically reduce required
flow rates of rinsewater, use of drain boards or drip tanks to catch and return process
fluids to their origin tanks, air knives to remove process fluids without dilution, and
use of turbulence or agitation to aid rinsing effectiveness.
WASTE REDUCTION TECHNOLOGIES
Evaporation, ion exchange (IX), reverse osmosis (RO), electrolytic recovery, and
electrodialysis (ED) have been used to recover chemicals and metals from rinsewaters.
78
Table 3. Summary of Recovery Technology Applications4
Technology
Plating Evaporation Reverse Ion Electrolysis Electrodialysis
Bath Osmosis Exchange
Decorative
Chromium x x
Nickel x x x x
Electroless
Nickel x
Cadmium x x x
Zinc (CN) x x x x
Zinc (CI) x x x
Copper (CN) x x x x
Tin (BF4) x x x
Silver x x x x
General and site-specific factors must be evaluated to determine the best recovery pro-
cess for a specific situation. Factors include the type of metal to be recovered, drag-out
rates, rinsewater concentrations and flows, space requirements, staffing requirements,
availability of utilities (such as steam and electricity), and cost for water and wastewa-
ter treatment and sludge disposaP. Table 3 shows typical recovery technologies related
to specific types of plating baths.
Evaporation is the oldest method used to recover chemicals from electroplating rinse
streams. In this process, enough rinsewater is evaporated to concentrate the solution
sufficiently to allow its return to the particular process. Because of their high-energy
use, evaporators are most cost-effective in concentrating rinsewaters that are to be
returned to hot baths, such as in a chromium plating process l .
Ion exchange (IX) uses charged sites on solid resin to selectively remove either
cations or anions from the solution. The ions removed are then replaced by equivalent
charged ions displaced from the resin. In general, ion exchange systems are suitable
for recovery applications where the rinsewater has a relatively dilute concentration of
plating chemicals. Ion-exchange systems are not cost-effective when drag-out rates are
low and concentrations highl.
Reverse osmosis (RO) is a demineralization process in which water is separated
from dissolved metal salts by forcing the water through a semipermeable membrane
at high pressures (400 to 800 psig). RO use is limited to moderately concentrated
rinsewaters. The cost-effectiveness of RO metal recovery systems varies depending on
several factorsl.
Electrodialysis (ED) concentrates or separates ionic species in a water solution
through use of an electric field and semipermeable ion-selective membranes. ED has
been used effectively to recover cationic metals from plating rinsewaters. The maxi-
79
Table 4. Some Material Recovery Technology Costs (1985)4
Technology Materials Equipment

Recovered Costs
Evaporation Unit: llinse water and $47,000
Capacity of Chromic acid
approx. 20 gph.
Reverse Osmosis Nickel plating $27,000

Unit: Capacity of chemicals
approx. 100 gph.
Ion Exchange Unit: llinse water and $38,000

Capacity of Chromic acid
approx. 20 gph.
Electrolytic Unit: llinse water and $25,000

Capacity of Copper
approx. 15 gph.
mum concentration of an ED unit is limited only by the solubility of the compounds in

solution, which allows ED to produce a more concentrated solution than IX or RO. ED
units are easy to use, economical, require little space, and can be operated continuously
with little maintenance. However, ED cannot selectively remove ions, resulting in the
requirement for periodic plating bath treatments to remove impurities. Along with
evaporation, ED has the quickest payback period of a metal recovery process 1 •
Table 4 lists some typical costs of applying the described technologies to recover
specific materials from a single source wastewater. To achieve savings that justify the
purchase of recovery technology requires the waste stream to be fairly concentrated and
continuous. Information necessary to determine economic feasibility includes waste
stream generation rates, chemical concentrations, and the value of materials to be
recovered4 •
SPECIFIC OPTIONS AND CASES
A Case Study at Pratt and Whitney12
In 1986, Pratt and Whitney's (P&W) North Haven, Connecticut facility began to
plan conceptually for a "zero discharge" metal finishing capability. At that time P&W
discharged about 1 million gallons a day of treated wastewater, of which about 400,000
gaIl day were generated by metal finishing operations. It was very clear that a waste
management plan was needed. The decision was made for a company wide reduction
of hazardous wastes by 40% and toxic air by 50% by 1996. Their program involved:
• identifying waste reduction opportunities which encompass good operating prac-
tices and proven technologies,
• working with the chemical processing industry to investigate emerging waste re-
duction and treatment technologies which could be piloted and evaluated, and
80
• determining new hazardous waste reduction areas for research that can meet
specific corporate needs.
P&W's North Haven facility has already taken a jump and "closed the loop" on such
key processes as their Woods nickel strike, sulfamate nickel plating, hard chromium
plating, cadmium plating, chromating, and cadmium, chromium and nickel stripping.
The 40% of wastewater generated by plating operations of 1986 now accounts for about
4 % of the discharge water.
P&W's program has been organized around the following implementation hierarchy:
Phase One defined good operating practices, including the decision to
• Define minimum water quality standards. All water now used is either deionized
or softened.
• Use countercurrent rinses to reduce water usage.
• Use continuous process purification, as opposed to batch purification, to main-

tain consistent process quality. This includes dummy plating and carbon and
particulate filtration.
• Alter on-line process solutions to control drag-out (for example, reduce concen-
tration and increase temperature).
• Optimize preplate rinsing to control drag-in contaminants.
• Install automatic level controls on all heated processes.
• Train operators to understand proper rinsing and work transfer techniques to

reduce drag-in and drag-out.
• Treat small concentrated batches versus high volume dilute waste streams.
Phase Two implemented procedural changes based upon established good operating
procedures.
Phase Three verified closed loop technology on a single process. For this phase,
a modularized approach has allowed the fine tuning of designs, while disruption of
production has been minimized.
Phase Four incorporated good operating practices and closed-loop technologies in the
design of planned and appropriated new plating lines. Designs for new plating lines
encompassing nickel and chromium plating, cadmium, chromium, and nickel stripping,
and titanium descaling were already under way. Initial plans were revised to incorporate
• Countercurrent rinses,
• Ion exchange for nickel strike, and nickel, cadmium and chromium stripping op-
erations,
• Atmospheric evaporation for hard chromium and sulfamate nickel lines, and
• Deionized water in all critical rinses and softened water in all non-critical rinses
and non-critical evaporation make-up.
Phase Five involved installing new plating lines, which were completed in October
1990.
Phase Six will be the renovation of remaining existing processes, including cadmium
cyanide plating and chromating.
81
Through these waste minimization efforts P&W's raw material costs have been
reduced by 80% and water usage by 95%. The costs of transportation and disposal
have been reduced proportionately and so have the liabilities. In the future P&W
plans on implementation of continuous process purification, implementation of closed-
loop technologies on preplate rinses, and implementation of on-line monitoring of all
systems 12 •
Water Use Reduction Techniques
There are many methods for reducing the overall amount of water used in a plating
process and there are corresponding benefits for these reductions. These may include
reduced water treatment costs, disposal of waste costs, reduced purchasing costs for
chemicals and water, and overall cost reduction for an operating day. Some of these
reduction methods are:
• Increase drip and rinse times on all automated lines by reprogramming computer
times, or changing gears in the hoist system to increase lift time. By increasing
the drip time more of the solution will stay in the tank, thus increasing the life of
the bath and not contaminating the following bath. By increasing the lift time,
better rinsing is accomplished because of the slower rate of breaking the surface
tension. More water is then pulled off the parts.
• Drip shields should be installed on all tanks. This technology has been around for
many years and can drastically reduce the amount of solution lost to drag-out.
The lives of baths are again lengthened due to reduced contamination from drag-
in and chemical loss due to drag-out, and much less waste treatment of diluted
streams has to be performed on the discharge effluent.
• Air agitation should be employed in all process tanks. This is another technique
for breaking the surface tension, which, in turn, allows for better rinsing. This
technique also proIllotes the eliInination of teIllperature gradients, and brings
solids and impurities to the surface where they can be filtered easily.
• A condensate return system may be installed on the plant's boilers which will
reduce overall water consumption. This source of water effluent should not be
mixed with the discharge of the plating operations if at all possible. This stream
will most likely be clean enough to discharge without treatment and will only
dilute the waste treatment stream.
• Atmospheric evaporators can be installed on plating lines. The evaporators take

the rinsewater from the drag-out and concentrate it. The remaining solution
(a concentrated chemical) can be put back into the plating cell to replenish the
bath. Recovering and recycling chemicals from the evaporators can slow the need
to make large additions of chemicals to the process tanks. This provides a savings
not only on the waste treatment of such chemicals, but also on the initial purchase
of the chemicals.
• Continuous filtration should be used on plating cells to clean solids from the
baths. This extends the life of the process solution and fewer chemicals have to
be added because there is a lower level of impurities in the solution. The net
result is less water going to the waste treatment system, and a reduced amount of
82
chemicals having to be purchased. Filtration can be done either in-tank or out-
of-tank. Most current systems are out-of-tank, but the current trend is towards
in-tank methods l3 •
• Continuous in-tank filtration versus out-of-tank. Conventional filter systems typ-

ically recirculate solution through a hose extending from the bath to be filtered
to a chamber containing the filter media. This provides many sites for leaks and
spills. The in-tank filtration method greatly reduces the chances of these leaks or
spills because it uses no hoses, clamps, seals, or any out-of-tank apparatus. The
filter is attached to the pump body and is immersed in the solution to be filtered.
The pump body is secured to the tank lip with an adjustable bracket. Solution
is continuously drawn by a vacuum through the filter cartridge, after which it
travels about three inches before being expelled through a discharge port on the
pump body. The solution is therefore never out of the tank, and is not subject
to the normal leaks or spills that can occur.
In addition to these benefits the in-tank unit creates enough force from it's dis-
charge port into the bath to produce a clean, oil free source of agitation that
suspends particles for easy filtration. This force also gives uniform temperature
and chemical gradients. Frequently, the discharge port is sufficient to replace
mechanical or air agitation. If continuous filtration is not required, a single unit
may be used in multiple tanks. The in-tank filtration units available today are
designed for ease of cleaning, filter changing, and no use of floor space l4 .
• Reverse Osmosis (RO) may be used to reduce the volume of water in waste
streams. RO technology has been used for this purpose in the past but today's
advancements have made smaller, more efficient, units than could have been imag-
ined in the past.
• Ion Exchange may be used to recycle rinsewater from anodization processes. The
ion exchange units can be designed and purchased for use on individual com-
pounds, or the units may be placed in series to remove virtually all of the effluent
metals. The solutions to be used in this type of process must be somewhat seg-
regated and the costs can be high unless sufficient amounts of solution are to be
treated l4 .
Recovery of Metals From Baths
There are many incentives for recovering metals from waste solutions and to reach
these incentives many methods of recovery have been established. By far the most
significant reason for recovery of metals is their intrinsic value. The tremendous savings
in initial purchase costs of most metals should be enough to warrant a recovery system.
The costs of disposal and the tremendous waste of metals are other reasons that recovery
of metals is becoming a process option in more and more plating operations. Some of
these process options include
• Precious metals recovery using cathodic plating. Silver and gold may be plated
out of the rinsewater solution onto a cathode. These cathodes may later be sent
to a precious metals refiner for credit, either being sold for scrap or recycled into
anodes and reused at a reduced cost in the plating lines l3 . Precious metals may
also be recovered using filtration 15.
83
• Electroless copper deactivation by removing formaldehyde from the waste solu-
tion. This is done by adding hydrogen peroxide to the solution. This deactivated
copper will not plate out in drain lines, etc., and can therefore be almost fully
removed. Aluminum is now being used to plate the deactivated copper out of
solution16.
• Electrowinning of metals such as copper, cadmium, and zinc. The advantage

to this method is that only the metals in solution are plated out, the other
contaminants are left unchanged. The factors affecting this method are cathode
surface area, temperature, and agitation. The overall efficiency of this method
is excellent, but the costs can be high unless a large amount of metal is being
recovered15 •
Specific Methods for Reduction of Hazardous Wastes
Hazardous wastes come in all shapes and forms, liquids, solids and sludges. They
can be either by-products of manufacturing processes or maintenance operations or
discarded commercial products. Simply burying the wastes, regardless of their form or
source, is no longer an acceptable means of disposal. Effective treatment, rather than
land disposal is now a necessity. Several methods of hazardous waste minimization are
available, some of the methods available are
• The use of inventory management to reduce the overall amount of wastes gen-
erated as mentioned earlier. All raw materials should be inventoried and only
the minimum amount should be purchased. This can eliminate excess materials
and chemicals from being thrown away if a process change occurs, or if the shelf
life expires. Raw materials should be purchased in a container that most closely
matches the expected use. This can result in less raw materials being thrown
away if a lid is not resealed on a half empty container. Reusable containers
should be purchased if at all possible, thus eliminating triple rinsing of an almost
empty container, or disposing of a container that contains enough residue to be
considered hazardous waste.
• Raw material substitution may be used to reduce the volume of hazardous waste
or its toxicity. The initial raw materials purchased should be as pure as possible
to reduce hazardous by-products. Non-hazardous or less hazardous materials
should be investigated and purchased. Lower volatility solvents and cleaners
can be used to reduce VOC air emissions. Phosphate based corrosion inhibitor
chemicals should be used instead of chromates. The use of ferrous sulfate-sodium
sulfate processes instead of lime for heavy metal precipitation can result in up to
90% weight reductions in waste sludges. A terpine-water emulsion has also now
been developed as a substitute for ozone-damaging chlorofluorocarbons in certain
degreasing operations.
• Process design and operation changes may be implemented to reduce the volume
of wastes generated. Equipment which produces the minimum amount of waste
should be purchased and existing equipment should be modified towards the same
end. Automatic metering pumps should be used to monitor chemical additions.
Consideration of redesigning or reformulating the end products to be less haz-
ardous can also contribute to an overall reduction. Electrostatic paint equipment
may be used to reduce solvent emissions. High solids paint and water-based paint
84
are both alternates to solvent-based liquid paints. If solvent-based liquid paint
must be used, high volume, low pressure guns should be used.
• Overall volume reduction can also be achieved through simple segregation of
streams. Non-hazardous streams should not be mixed with hazardous streams if
at all possible. Streams containing different solvents or chemicals should be kept
separate so that the chemicals may be more easily recycled. The waste streams
which must be treated should be reduced with such techniques as reverse osmosis,
evaporation, ion exchange, precipitation, centrifugation, gravity separation, and
distillation.
• Hazardous waste can also be chemically altered to reduce its toxicity or create
a non-hazardous product. The waste may be encapsulated so that it is sealed
with a material that makes the overall product non-hazardous. Waste oil may
be chemically treated and turned into a salable aromatic distillate. Hydrochloric
acid may be converted into calcium chloride and used in manufacturing polyamide
fibers or into calcium sulfate and used as a soil stabilizer l7.
• The use of liquid cleaners instead of powder. Until recently most platers have
relied on powder cleaners for their pretreatment of the pieces to be plated. The
problems associated with powder cleaners can be safety, handling, mixing and
solubility, control of concentration, useful life, and waste treatment. The use
of liquid cleaners eliminates worker exposure to irritating and corrosive dusts
associated with powder blends, and reduces airborne particles to help meet OSHA
exposure regulations. Powder cleaners may sometimes mix non-uniformly in the
solution or may not mix at all if the temperature and agitation are not adequate.
This can result in powder settling on the bottom of the tank, creating even more
problems. There is no guarantee of the concentration and mixing with a powder
because it is impossible to tell how much has gone into solution. The use of liquid
cleaners can eliminate all of these problems. In addition, liquid cleaners can be fed
by a metering pump into the solution giving excellent concentration control and
requiring minimal worker exposure. The total solids of liquid cleaners are typically
5 to 10 percent, as opposed to 90 to 98 percent for powder cleaners, meaning
that less sludge is generated during waste treatment. Due to manufacturing
improvements, liquid cleaners now out perform most powders and can be designed
for almost any use lS •
• An electrodeposited Pb0 2 anode may be used for the decomposition of low con-
centrations of cyanide. Titanium anodes electroplated with lead dioxide in nitrate
solutions containing a surfactant or an organic additive are more effective than
the traditional graphite anode for decomposing cyanide in solutions containing a
low cyanide concentration and sodium chloride. This is due to the porosity of the
Pb0 2 surface layer and the electrode potential for the reaction 19.
• Microcomputers can be very useful in waste management performing such tasks
as modeling processes, using material balances. The greatest asset of these com-
puters is that they can be put on the desk of the environmental engineer at very
little cost. Many of the microcomputers of today can be purchased very inex-
pensively and can pay for themselves in a short amount of time. FORTRAN
and Basic, as well as many other languages, spreadsheets and equation solving
software, can now be used on these computers. This gives the engineer in charge
of locating, reducing, and treating the effluent waste a very good tool for fast,
effective, and accurate modeling and calculations 2o ,21.
85
SUMMARY
When developing and implementing a waste reduction program it is essential that

management be committed to pursuing waste reduction, be willing to experiment with
various ideas, and be prepared to experience failure, as well as success. Within a
successful waste reduction program three essential elements must be present: active
involvement and encouragement of management at all levels, application of the least
complex technologies, and motivated production personnel to implement and maintain
necessary changes. The second element is especially important because the greater the
number of modifications attempted at the same time, or the more experimental the
equipment is, the greater the chance that problems will occur and the process will be
abandoned for more reliable and proven methods.
Developing a waste minimization program has become a necessity in industry. As
U.S. EPA regulations on waste disposal become stricter, newer, more efficient mini-
mization methods of waste disposal must be developed. Waste management methods
that reduce or eliminate unwanted by-products will prove to be more cost-efficient and
environmentally sound, traits that will be in high demand in the future.
ACKNOWLEDGEMENTS
Although the review described in this paper has been funded in part by the U.S.
EPA under assistance agreement number R-815709, to the University of Nebraska-
Lincoln through the Great Plains/Rocky Mountain Hazardous Substance Research
Center, headquartered at Kansas State University, it has not been subjected to the
Agency's peer and administrative review and, therefore, may not necessarily reflect the
views of the Agency and no official endorsement should be inferred.
The research was partially supported by the U.S. DOE, Office of Environmental
Restoration and Waste Management, Office of Technology Development and the Uni-
versity of Nebraska-Lincoln Water Center.
REFERENCES
1. T.E. Higgins, "Hazardous Waste Minimization Handbook," Lewis Publishers,

Chelsea, Michigan (1989).
2. S.E. Manahan, "Environmental Chemistry," 5th ed, Lewis Publishers, Chelsea,
Michigan, 453 (1991).
3. "Waste Minimization Opportunity Assessment Manual," U.S. EPA, Washington,
DC, EPA/625/7-88/003 (July 1988).
4. PRC Environmental Management, Inc., "Hazardous Waste Reduction in the
Metal Finishing Industry," Noyes Data Corp., Park Ridge, NJ (1989).
5. J.J. Segna and R.K. Raghavan, Approaches to reducing environmental risk through
pollution prevention, in: "The Environmental Challenge of the 1990's," U.S. EPA,
Washington, DC, EPA/600/9-90/039, 613-23 (Sep 1990).
6. R.W. Rittmeyer, Waste minimization-part 1: prepare an effective pollution-
prevention program," Chemical Engineering Progress, 87:5 (May 1991).
7. R.J. Avendt, Waste minimization-approaches and techniques, in: "The Environ-
mental Challenge ofthe 1990's," U.S. EPA, Washington, DC, EPA/600/9-90/039,
32-40 (Sep 1990).
86
8. "Seminar Publication: Meeting Hazardous Waste Requirements for Metal Fin-
ishers," U.S. EPA, Washington, DC, EPA/625/4-87/018 (September 1987).
9. Additional technology and case study information can be obtained from:
U.S. EPA
ORD Publications
P.O. Box 19962
Cincinnati, OH 45219-0962
Request "ORD Publications Announcement" (e.g., EPA/600/M-91/048) and "Tech-

nology Transfer" Newsletter (e.g., EPA/600/M-91/042).
10. "Metal Finishing Guidebook and Directory Issue," Metal Finishing, 89:1A (1991).
11. "Guides to Pollution Prevention: The Printed Circuit Board Manufacturing In-
dustry," U.S. EPA, Washington, DC, EPA/625/7-90/007 (June 1990).
12. P. Gallerani and R. McCarvill, Waste minimization and pollution prevention at
Pratt and Whitney, Plating & Sur/ace Finishing, 78:3, 36 (March 1991).
13. P. Dalton, Water conservation and waste minimization efforts at Hudgins Plating,
Plating & Sur/ace Finishing, 78:7,28 (July 1991).
14. G. Horvath, In-tank filtration/solution purification: a new concept in waste min-
imization, Plating & Sur/ace Finishing, 78:4, 28 (April 1991).
15. F.W. Kursch and G.P. Looby, Case study: Pollution prevention in practice: appli-
cations for electroplating technology, Pollution Prevention Review, 2:1,63 (Win-
ter 1991-92).
16. J.D. Holly, Waste treatment process for electroless copper, Plating & Sur/ace
Finishing, 78:1, 24 (January 1991).
17. R.A. Katlin, Waste minimization-part 3 minimize waste at operating plants,
Chemical Engineering Progress, 87:7, 39 (July 1991).
18. S.F. Rudy and B. Durkin, The use of liquid cleaners in metal finishing, Plating
& Sur/ace Finishing, 77:3, 26 (March 1990).
19. T.-C. Wen, Electrodeposited PbO z anode for the decomposition of low concen-
trations of cyanide, Plating & Sur/ace Finishing, 77:11, 54 (November 1990).
20. C.W. Walton and G.L. Poppe, Applying microcomputers to the analysis of waste
treatment and recovery processes, Plating & Sur/ace Finishing, 77:6, 48 (June
1990).
21. A.C. Hillier and C.W. Walton, Modeling electroplating rinse systems using equation-
solving software, Plating & Sur/ace Finishing, 78:11, 72 (November 1991).
87
THE ROLE OF CATALYSIS IN INDUSTRIAL WASTE REDUCTION
David Allen
Department of Chemical Engineering

University of California, Los Angeles
Los Angeles, Ca. 90024
INTRODUCTION
Millions of tons of hazardous waste are generated each year in the United States,
and the direct costs of managing these materials are in the tens of billions of dollars. As
public pressure for reduced, or even zero discharge grow, waste management costs are
likely to further escalate, and conventional end-of-pipe treatment measures will become
cost prohibitive. A new approach to waste management will be required, which stresses
the elimination or reduction of wastes at their source, rather than capture after their
formation. This new approach to waste management has been called waste reduction,
pollution prevention and source reduction. The approach offers both challenges and
opportunities. The challenge is to control the costs of environmental stewardship so that
products produced in an environmentally benign manner can remain competitive in global
markets. If this challenge is met successfully, then the clean technologies that are
developed have the opportunity to dominate world markets.
Meeting the challenges of waste reduction will require key advances in the field of
engineering design, particularly in catalysis and reaction engineering. The goal of this
paper is to describe the role that catalysis can play in reducing waste. Our starting point
will be a description of the generation and management of industrial hazardous waste.
From this brief overview, it will become apparent that catalysis is currently underused as
method for converting wastes into recyclable material. In addition, a more detailed
examination of the processes generating hazardous wastes will show that chemical
processes that rely on catalysis are responsible for a significant fraction of industrial
hazardous waste generation. Clearly then, improved selectivities in catalytic processes
could reduce waste generation. Our examination of the potential role of catalysis in waste
reduction will conclude with the presentation of a few case studies.
HAZARDOUS WASTE GENERATION AND MANAGEMENT: The role of catalysis
According to recent estimates (Baker and Warren, 1992), more than 750 million
tons of waste, classified as hazardous under the provisions of the Resource Conservation
and Recovery Act, are generated annually in the United States. Over 90% of these wastes
are in the form of wastewaters, and more than 90% of the material is managed at the site
where the waste was generated. Common management methods include various forms of
lllduslrilJl E"viro",,,,mtal CMmistry, Edited by D.T. Sawyer

wastewater treatment, incineration, reuse as fuel, land disposal technologies and recycling.
The flow of material to each of these technologies is given in Figure 1. A detailed
description of the data presented in Figure 1 is beyond the scope of this paper, but is
provided elsewhere (Behmanesh et al., 1992; Baker, et aI., 1992). The industrial sectors
generating these wastes are shown in Figure 2. Chemical manufacturing operations clearly
dominate hazardous waste generation, producing more than 50% of the total waste mass.
While it is clear that the chemical process industries dominate industrial hazardous
waste generation, it is not clear which chemical processes bear primary responsibility.
Data on waste generation and management, of the type shown in Figures 1 and 2, are
generally reported for entire facilities, rather than for individual processes. So it is very
difficult to determine, strictly from waste reporting, which catalytic processes might be
most significant in waste generation. A sense of the relative waste generation of different
chemical processes can be obtained based on industry wide yield data, however. Table 1
reports annual production rates and industry wide average yield data for a number of
commodity chemicals. Within Table 1, the data listed for acrylonitrile, 1,3 butadiene and
vinyl chloride illustrate a number of interesting points. For acrylonitrile, high production
rates and low selectivities indicate potential for high impact improvements. In contrast,
although the data for 1,3 butadiene indicate high production rates and relatively low
conversions, the by-products and unreacted feed are not generally wastes in this process,
but rather have direct fuel or chemical value. Finally, the data for vinyl chloride indicate
moderate production rates and high selectivities, but, as we shall see in a later section,
catalytic technologies can playa role in reducing wastes from this process.
Thus, while waste stream data, production data and yield data can be interesting
starting points in looking for opportunities for catalysis to contribute to waste reduction,
they do not tell the entire story. Until waste generation data ate available on a process by
process basis, it will be difficult for external observers to identify specific opportunities
for waste reduction. Still, the data can be used to reveal some generic opportunities for
catalysis. These include:
1. Catalysis could be used more extensively in waste treatment, either as a stand alone
treatment unit or as a waste pretreatment method. The data presented in Figure 1
clearly show that catalysis currently plays a very minor role in the treatment of
industrial hazardous wastes. There are a number of legal and technological reasons
for this. Among the technological limitations are the wide variations in
composition frequently encountered in waste treatment. Designing robust catalysts
capable of resisting the wide range of deactivating agents found in waste streams
can be difficult. The technological barriers are not, however, impossible to
overcome and a number of treatment technologies employing oxidative and
reductive catalysis have been developed (for example, Aguado et aI., 1991; Kim
et aI., 1990; Hasan et at., 1990). These new technologies can be difficult to
implement, however, due to the administrative and legal hurdles involved in
certifying a new waste treatment technology (NETAC, 1990).
2. Catalysis could be used to convert waste streams into recyclable material. A

number of technologies have been developed which use catalysis to convert waste
streams into recyclable material. Several examples will be presented as case
studies in the next section. All of these case studies will focus on
hydrodechlorination chemistry, but, many other reduction and partial oxidation
chemistries have promise.
3. Catalysis can play a significant role in waste reduction through selectivity

improvement. Maximizing selectivity, i.e. the efficient use of raw materials, has
always been a major factor in the design of chemical processes, and so it would be
90
..,
t
..,
U.S.
..
Induslry
(AlI_'id,
~---------------------------------4~~IUmdT~m.nll
••30
Figure 1. Management of industrial hazardous wastes in the United States; flow rates are in millions
of tons per year; data are for 1986. Reproduced by permission of Mary Ann Liebert
Publishers © 1992; R. D. Baker, J. L. Warren, N. Behmanesh and D. T. Allen, Hazardous
Was:te and Hazardous Materials. 9(1). 32-55 (1992).
91
28 Chemical Products
36 Electronics
29 Petroleum & Coal Products
33 Primary Metals
37 Transponatlon Equipmont
All Other Industries
00/. 25% 50% 75% 100%
llQuantity 01 hazardous waste generated in million tons.

so
Genon1I _ _ rnonufadUring
(2053)
2800 2899 e..._ "'___.. no<:
28.2 Alkalies and chlorine 2911 Petroleum
__ ,.Hring
plastics ",odbcu
28.6 Inotgonlc pigrnem. 3079
so (21'
28.9 100""'01 i"IorganIc _ ' . nee 3312 Bla turMCeS and stMi mil,
282. PlaSlic::a materials end ,.sJns 3315 Steel wire and ,eI.led products
.
21W3 S_adi".~. 3361 Al......... loundri••
2065 C)'dic c:rudes Met ~lermldlal •• 3071 PIaI01v _ poIlhlng
40 2869 tndutriIII cwganfc cn.tricIIs. nee 3079 M""" """tin; _ allled_.

2873 Nitrogenous lrilz.,.. 3662 Radio end TV 00IM'0Un1caOon oquiprnenl
2879 A~.""_ noc 3674 SlfT'IIoonductott end rMaliid dt\liolts
Quantily O. 2892 • ExpIo..... 3711 Motor vlhlc:iH and QW bodies
wule
generated 30
by SIC code
(130)
lmi.ion IOns>
20
(10' (SM)
'0
0
28M 2880 ~ 281' 2182 2111 211133123471 zea 3315 a21 3874 301'1281115 341'f 286$ 2171 :28&SII 2812 3381 ~, 37" 35142111
2MS 2100 2:1'7.1 28C1iV 2MI 2811 2&43 3112 ZII2'1
2IU 2M3 28M 2IeI
SIC code(B) of proce$$ generating wasle

lnurrber 01 wutes generated In paron_> B
Figure 2. a) Generation of industrial hazardous waste in the United States; generation is reported by
two digit Standard Industrial Classification (SIC) codes. b) Generation of industrial
hazardous waste in the United States; generation, in units of millions of tons per year are
reported by four digit SIC codes. Reproduced by permission of Mary Ann Liebert
Publishers © 1992; R. D. Baker and J. L. Warren Hazardous Waste and Hazardous
Materials, 9(1), 13-31 (1992).
92
Table 1
Representative Production and Yield Data for
Commodity Chemical Manufacturing
Chemical & Production Process Annual Yield (%) Yield (%)'.

Produc- calculated'
tion, 1()()() tons
Acrylonitrile 2,182
1. Ammoxidation of propylene 65 53
2. Cyanation/oxidation of ethylene 76 85
1,3-butadiene 2,546
1. Dehydrogenation of n-butylenes 64 70
2. Oxidative dehydrogenation of n-butylenes 87 74-90
3. Dehydrogenation of n-butane 42 57-63
Vinyl chloride 8,439

1. Chlorination & oxychlorination of ethylene 96 86
2. Dehydrochlorination of ethylene dichloride 98 95
·Calculated using data reported in "Petrochemical Technology Assessment" by, D. F. Rudd, S. Fathi-Afshar,
A. A. Trevino and M. A. Stadtherr. lohn Wiley & Sons, 1981.
··Erskine, M. G., "Chemical Conversion Factors and Yields", Chemical Information Services, Stanford
Research Institute, 1969.
93
very naive to assume that a sudden focus on waste reduction will magically
improve selectivity. What will occur is a re-examination of existing process
alternatives. A highly selective process that was not cost effective a decade ago
may , due to escalating waste disposal costs, be the most cost effective alternative
today.
4. Catalysis can play a role in waste reduction by reducing selectivities for highly
toxic waste products. Often, what makes a waste difficult or expensive to dispose
of are highly toxic trace contaminants. For example, if a waste stream contains
only a few parts per billion of 2,3,7,8 tetrachlorodibenzodioxin, it will be much
more difficult and expensive to manage than the same waste without the trace level
of dioxin. Designing catalysts that avoid the formation of highly toxic byproducts
is a selectivity problem, but it is a different selectivity problem than those faced in
the past. Indeed, one may be willing to reduce overall selectivity in a chemical
reaction if the generation of highly toxic contaminants in the waste streams are
reduced.
CASE STUDIES OF CATALYSIS FOR IN-PROCESS RECYCLING
The previous section described four roles that catalysis might play in waste
reduction. Of the four, the only role that currently sees any significant use is catalytic
conversion of wastes into recyclable materials. In this section, several case studies of
catalysis for converting wastes into recyclable materials will be presented. These examples
will draw heavily on the research interests and experience of the author, and as a
consequence, will be devoted exclusively to hydrodechlorination reactions. The use of
catalytic technology for recycling materials involves, of course, many other chemistries.
Some are old (conversion of II:2S to elemental sulfur in petroleum refineries). Some are
new (catalytic conversion of chlorofluorocarbons, CFCs, to partially hydrogenated
chlorofluorocarbons, HCFCs). The group presented here are merely samples from a large
set of technologies.
Catalytic hydroprocessing of chlorinated organics is a relatively new technology
that can be used to recycle waste streams from a number of chemical manufacturing
operations. For example, Kalnes and James (1988) have reported on .the use of catalytic
hydrodechlorination for recycling liquid wastes from vinyl chloride manufacturing and still
bottoms from epichlorohydrin production.
While the use of catalytic hydroprocessing to recycle chlorinated organics is a
relatively new idea, the basic chemistry is merely an adaptation of petroleum
hydroprocessing. Catalytic hydrotreatment has been used for decades in the petroleum
refining industry to remove oxygen, nitrogen and sulfur functionalities from heavy fuels.
The chemistry and kinetics of deoxygenation, denitrogenation and desulfurization have
been studied extensively and several comprehensive reviews of the literature are available
(Rollman, 1977; Gates, et al., 1979; Shah and Krishnamurthy, 1981) Much of this
information is applicable in a general way to the development of catalytic
hydrodechlorination, but, until recently, few data were available on the hydroprocessing
chemistry of chlorinated compounds.
In the author's laboratory, initial work on hydrodechlorination chemistry focussed
on identifying a robust dechlorination catalyst. The goal was to identify a catalyst that was
reasonably active and that would not be deactivated by the wide range of functionalities
found in waste streams. Eventually a commercial petroleum hydroprocessing catalyst was
chosen for further study, and the rates of dechlorination for multiply chlorinated benzenes
were examined. Representative results are shown in Figure 3 (Hagh and Allen, 1990a-c).
94
CI CI
oAo oAo 00
oyo aYo---
a 0
t ~13.4
a
CIr51CI
0
clAcl _~_
CIVCI CIYCI
o / ~
a'
'- 0
45
~c
"'-5.5
~
CI~CI CIA CI~CI _1~'t.CI~CI clAcl _~_
CIY X~YCI CIYCI clY ~.2 CIYCI
CI ""\CI ~'\ / 5~CI ~ ;2.6

~/ CI~ CIAcl 1.1 CIA CIACI
\ cii/l
CI~CI CIY CIY 0.1 CI~20 CIY
\ 3·~;f.9\O
'A CIA ACI CIA ~ ~CI
CIO CI'Y' CI'Y' o'Y' CI'Y/ CI'Y/
'" l / 11.8 ~ 142 2

/ ,110.1 1.8
CI© CI©
t t2~
© A © B
Figure 3. a) Reaction pathways for chlorinated benzenes hydroprocessed over a Ni/Mo on Al20 3
catalyst; observed pathways are shown as solid lines; pathways that were not observed are
shown as dashed lines. b) Pseudo-fll"St order kinetic rate constants for dechlorination of
chlorinated benzenes in units of (liter/{g cat-min»x 103; dashed lines indicate pathways in
which multiple chlorine atoms are removed in a single step. Reproduced by permission
of Pergamon Press @ 1990; B. F. Hagh and D. T. Allen, Chem. Eng. Sci., 45, 695-702
(1990).
95
© ©
Figure 4.
©
Hydroprocessing pathways and pseudo-flISt order rate constants for chlorophenols reacting
over Ni/Mo on AlZ0 3 • Reproduced by pennission of the American Institute of Chemical
Engineers @ 1991 AlChE; S. Chon and D. T; Allen, A1ChE J., 37, 1730-1732 (1991).
96
Dechlorination of singly and multiply chlorinated organics is only one aspect of
dechlorination chemistry. Frequently wastes will contain functionalities other than chlorine
that might compete for catalytic sites. For example, chlorophenols could either
dechlorinate or deoxygenate under hydroprocessing conditions. Over our model catalyst,
dechlorination rates for the chlorophenols were approximately two orders of magnitude
greater than deoxygenation rates. Typical results are shown in Figure 4. The data show
that the location of the oxygen functionality in chlorophenol influences the dechlorination
rate substantially. In the case of orthochlorophenol, the dechlorination rate was reduced
to roughly half that of chlorobenzene dechlorination rate at 275°C. For metachlorophenol
and parachlorophenol, the presence of a hydroxyl group at the meta- and the paraposition
led to approximately twofold and threefold increase in dechlorination reactivity,
respectively, to that of chlorobenzene dechlorination activity. The dechlorination activities
of chlorophenols are therefore para> meta> ortho, with the hydroxyl group activating the
dechlorination in the para- and metaposition and deactivating dechlorination in the
orthoposition. These results can be rationalized based on electron donating properties of
the substituents and steric effects.
The main use of these dechlorination chemistries to date has been the complete
dechlorination of waste streams to generate HCI and a hydrocarbon fuel. A number of
more interesting opportunities may also exist. Referring back to the data on chlorinated
benzenes, it may, for example, be possible to selectively dechlorinate meta-dichlorobenzene
in a mixture of meta and para-dichlorobenzene. This might allow for increased recycle of
by-products in the production of para-dichlorobenzene, a household fumigant.
CONCLUSION
The fact that catalysis can playa major role in waste reduction should come as no
surprise. As the data on industrial waste generation presented in this paper indicate, many
of the materials that become hazardous wastes are initially generated in catalytic reactions.
Since catalysis has played a role in promoting the reactions to form these materials, it can
also play a significant role in reactions to convert these potential wastes back into useful
raw materials. This paper has presented information about one class of reactions,
hydrodechlorination, that can be used to convert waste streams back into raw materials.
Many other chemistries exist. More still are possible. The challenge for catalysis
researchers and practitioners will be to bring these chemistries into widespread use.
REFERENCES
Aguado, M. A., M. A. Anderson and C. G. Hill, 1991 "Ceramic membrane photocatalysts for the
remediation of impaired waters", AlChE Annual Meeting, Paper 97c.
Baker, R. D., and J. L. Warren, 1992 "Generation of hazardous waste in the United States",
Hazardous Waste and Hazardl.!us Materials, 9(1), 13.
Baker, R. D., J. L. Warren, N. Behmanesh and D. T. Allen, 1992 "Management of hazardous waste in the
United States", Hazardous Waste and Hazardous Materials, 9(1), 32.
Behmanesh, N., D. T. Allen and J. L. Warren, 1992 "Flow rates and compositions of incinerated waste
streams in the United States", 1. Air and Waste Management Assoc., 42(4), xxx.
Chon, S., and D. T. Allen, 1991 "Catalytic hydroprocessing of chlorophenols", AIChE 1.,37, 1730.
Erskine, M. G., 1969 "Chemical Conversion Factors and Yields", Stanford Research Institute, Palo Alto, CA.
Gates, B.C., J. R. Katzer, and G. C. A. Schuit, 1979 "Kinetics and reaction chemistry of catalytic
hydrodechlorination of chlorinated benzenes on sulfided NiMo/yAI20 3 ", Chemistry of Catalytic
Processes, McGraw-HilI, New York, p. 390.
Hagh, B. F., and D. T. Allen, 1990a "Catalytic hydrodechlorination", Innovative Hazardous Waste Treatment
Technology, Vol. I, H. M. Freeman, ed., Technomic, Lancaster, PA.
97
Hagh, B. F., and D. T. Allen, 1990b "Catalytic hydroprocessing of chlorobenzene and 1,2-dichlorobenzene",
AIChE J., 36, 773.
Hagh, B. F., and D. T. Allen, 1990c "Catalytic hydroprocessing of chlorinated benzenes", Ch£m. Eng. Sci.,
45,2695.
Hasan, S., P. Cho, K. L. Sublette, D. Pak and A. Maule, 1990 "Porphyrin catalyzed degradation of hazardous
waste", AIChE Annual Meeting, Paper 49g.
Kalnes, T. N., and R. B. James, 1988 "Hydrogenation and recycle of organic waste streams", Environ. Prog.,
7,185.
Kim, S., M. A. Abraham, J. Papap, R. L. Cerro and Y. T. Shah, 1990 "Three phase catalytic oxidation of
phenol in a monolith foam reactor", AIChE Annual Meeting, Paper 49b.
National Environmental Technology Applications Corpomtion (NETAC), 1990, University of Pittsburgh
Applied Research Center, "Incentives and barriers to commercializing environmental technologies",
Pittsburgh, PA.
Rollman, R. D., 1977 "Catalytic hydrogenation of model N-, S-, and O-compounds", J. Catal., 46, 243.
Shah, Y. T., and S. Krishnamurthy, 1981 "Heteroatom removal during liquefaction", Reaction Engineering
in Direct Coal Liquefaction, Y. T. Shah, ed., Addison-Wesley, Reading, MA.
Rudd, D. F., Fathi-Afshar, S., Trevino, A. A., and Stadthem, M. A., 1981 "Petrochemical Technology
Assessment", Wiley, New York.
98
CATALYSIS, THE ATOM UTILIZATION CONCEPT
AND WASTE MINIMIZATION
R.A. Sheldon
Delft University of Technology

The Netherlands
ABSTRACT
Following the advent of the petrochemicals industry in the 1920's, catalysis

was widely applied in the manufacture of bulk chemicals. Traditional, environ-
mentally unacceptable processes have largely been replaced by cleaner catalytic
technologies. Fine chemicals, in contrast, have remained largely the domain of the
synthetic organic chemist who has generally clung to the use of stoichiometric
methods.
But times are rapidly changing. Increasingly stringent environmental
constraints are making the use of classical stoichiometric methods prohibitive.
Consequently there is a general trend towards substitution of such antiquated
technologies by cleaner, catalytic methods that do not generate large amounts of
inorganic salts as waste.
A useful concept for evaluating the environmental acceptability of various
processes for producing a particular substance is atom utilization. The latter is
defined as the ratio of the molecular weight of the desired product to the sum
total of all the materials (excluding solvents) used. Examples of the application of
clean, low-salt technologies are illustrated by reference to alternative routes for
the synthesis of the industrial monomer, methyl methacrylate and the analegesic
drug, ibuprofen.
INTRODUCTION
After an induction period of a few decades we now appear to be in the age

of 'environmentali ty'. This is reflected both in the general trends in society as a
whole and in the chemical industry in particular (see Figure 1).
A particular concern is the amount of waste produced in chemical processes,
largely in the form of inorganic salts. Indeed, intergrated waste management and
zero emission plants are the catch-words in the chemical corridors of power these
Industrial Environmenlal Ch4mistry, Edited by D.T. Sawyer

and A.E. Martell, Plenum Press, New York. 1992 99
days. What goes in must come out, somewhere. Preferably what comes out should
be the desired product. Everything else should be considered as undesirable and
its formation avoided or be kept in the system.
Obviously the key to waste minimization is selectivitv in chemical processing.
In describing the efficiency of chemical transformations organic chemists generally
distinguish between different types of selectivity. The simplest is standard one: the
yield of desired product divided by the conversion of the starting material. In
reactions of complex organic molecules the concepts of chemoselectivitv
(competing reactions at two different functional groups in the same molecule),
regioselectivity (e.g. ortho vs para attack in aromatic substitution) and
stereoselectivity (enantio- or diastereoselectivity) are frequently used. However,
there is another category of selectivity that is generally overlooked by organic
chemists: what we shall call the atom selectivity or atom utilization.
Cleaner, more environmentally acceptable products, e.g. 'green gasoline' (lead-

and aromatics-free)
Products that are more effective, more targeted in their action and
environmentally friendly, i.e. readily recycled or biodegraded
Zero emission plants/integrated waste management
Cleaner technologies with negligible inorganic salt formation (Le. catalytic
processes with optimal atom utilization)
Replacement of toxic and/or hazardous reagents, e.g. COCl 2, (CH3hS04'
H 2CO/HCI, heavy metal salts. Transport and storage of hazardous chemicals,
e.g. halogens, becoming increasingly difficult
Alternatives for chlorinated hydrocarbon solvents, e.g. solvent-free processes,
chemistrY in water
Shorter routes, in some cases via alternate feedstocks, e.g. substitution of
alkanes for alkenes and aromatics
Utilization, where feasible, of renewable raw-materials, e.g. carbohydrates, or
waste materials from other processes
Higher chemo-, regio- and enantioselectivities
Figure 1. General trends in the chemical industry.
THE ATOM UTILIZATION CONCEPT
The atom utilization of a particular process is calculated by dividing the

molecular weight of the desired product by the sum total of all the materials
(excluding solvents) used. It provides a simple way of assessing how 'clean' a
particular process is with regard to the amount of waste that is formed, in the
ideal case of 100% selective conversion of starting material.
An example, to illustrate the point, is the replacement of the traditional
chlorohydrin technology for ethylene epoxidation by catalytic epoxidation:
100
(1)
(2)
\1 atom utilization = 44/173 =25%
catalyst
(3)
atom utilization = 100%
Although the concept of atom utilization is on the one hand trivial, it has largely
been ignored by organic chemists. This is reflected in the enormous amounts of
inorganic salts formed in most organic reactions. Indeed organic chemistry in
general tends to involve far too many neutralization steps, all producing significant
amount of salts. Consequently, in classical organic reactions inorganic salts often
represent a large proportion of the total weight of material formed. Increasingly
stringent environmental measures are, however, stimulating the replacement of
such traditional salt-forming technologies with cleaner catalytic alternatives (Le.
low-salt technologies). One could say that the chemical industry is being subjected
to a 'low-salt diet'.
Similarly, catalytic oxidation of ,9-phenoxy-ethanols provides a potentially
interesting salt-free alternative to the classical process for the production of Q-
phenoxyacetic acid herbicides (see Figure 2).
Classical process:
CH3C0 2H + Cl2 - _..- CICH 2C02H + HCI
1. 2NaOH
---""'r- ArOCH 2C0 2H + 2NaCI + 2H20
2. HCI
Overall stoichiometry:
\1 atom utilization· = 200/353 =57% II

• assuming Ar = 125
METHYL METIlACRYLATE MANUFACTURE
Methyl methacrylate (MMA) is an important polymer raw material the annual

production of which is in excess of one million tons worldwide. Developments in
manufacturing processes for this relatively simple molecule illustrate perfectly the
modem trends in chemical processing. Traditionally, MMA is produced by
reaction of acetone (a byproduct of phenol production) with hydrogen cyanide. (a
byproduct of acrylonitrile production), followed by methanolysis (Figure 3).· As
101
Low-salt alternatjve:
catalyst
atom utilization = 200/218 =92% I

Figure 2. Two routes to phenoxyacetic acids.
o HO CN
A +HCN - X
MeOH
- MMA
Atom utilization = 46"
Figure 3. Classical pToceu fOT MMA manufacture.
such it provides a good example of the use of byproducts from one chemical
process as raw materials in another.
However. the process has one serious flaw: it produces a substantial amount of
ammonium bisulfate as a coproduct. Although the theoretical amount is 1.15 kg of
NH4HS04 per kg MMA, corresponding to an atom utilization of 46%, in practice
a substantial excess of H 2S04 is required resulting in the formation of ca. 2.5 kg
salt per kg MMA. Consequently. much effort has been devoted to the
development of alternative. low-salt routes to MMA. These generally involve
catalytic oxidation or carbonylation of readily available lower olefins (see
Figure 4).
Interestingly. the traditional technology has recently been upgraded by the
development, by Mitsubishi Gas l of an elegant method to circumvent the salt
formation (see Figure 5).
102
A Oz
catalyst - A94" sel-
CHO
(Asahi: Mitaubishi)
Oz
catalyst - A
COzH
96" sel-
-MMA
CHO
~
CO/Hz
- ~CHO
9
+ A
1
byproduct
- -
COzH COzH
Oz
A Oz
catalyst A + HzO
H2 C=CH2 ~.
CO/HzO
alyst /"'-COzH
HzCO
-
COzH
A
CO/Hz ,- HzCO
catalyst /"'- CHO 2- Oz -
(BASF route)
Figure 4. Alternative routes to methylmethacrylate (MMA).
1<03
It should be noted, however, that this leads to a rather circuitous route. In this
context, a devastatingly simple route has recently been reported by Shell workers. 2
MMA is formed in one step, in essentially quantitative yield, by palladium-
catalyzed methoxycarbonylation of methylacetylene (Figure 6). The reaction is
carried out under mild conditions and exhibits a remarkably high catalyst turnover
frequency (up to 100.000 mol methylacetylene per g catalyst per hour).
Methylacetylene is a byproduct of naphtha cracking and is produced in sufficient
amounts to make it an interesting raw material for MMA. In short, the route
illustrated in Figure 6 is the epitomy of modern trends in chemical manufacture: a
one-step, extremely efficient catalytic process with 100% atom utilization and
based on a waste stream as raw material.
HO CN HO CONH2
H2 O
X •
X
j
>=01
CH,OH
CO catalyst
HCN .. -H2 O
HCONH2 +
HO
X
C02 CH 3
I- HO
2
MMA
Figure S. Mitsubishi Gas process.
CATALYSIS IN BULK AND FII\1E CHEMICALS MANUFACTURE
In the drive towards waste minimization in chemical manufacturing catalysis

will play a pivotal role. In the bulk chemicals industry classical environmentally
unacceptable processes have largely (but not completely) been supplanted by
cleaner ~at~~ric ~lternatives. In particu.lar, catalytic oxidation3 and
carbonylatIOn are WIdely used for the converSIOn .of petrochemical feedstocks to
industrial chemicals (see Table 1).
104
CH3 C== CH + CO + MeOH
-
>99% selectivity
60 bar ; 60-80 0 C
TOF = 10-100·000 mol
propyne/g catalyst/hr
Figure 6. Methyl methacrylate by catalytic methoxycarbonylation of methylacetylene.
105
Table 1. Bulk chemicals manufacture USA (1990).
CHEMICAL VOLUME REACTION CATALYST

(106 tons)
Terephthalic 4,0 Oxidation Homogeneous
acid
St)Tene 4,0 Dehydrogenation Heterogeneous
Methanol 4,0 CO + H2 Heterogeneous
Formaldehyde 3,2 Oxidation Heterogeneous
Ethelyne oxide 2,8 Oxidation Heterogeneous
Acetic acid 1,9 Carbonylation Homogeneous
Phenol 1,8 Oxidation Homogeneous
Propylene oxide 1,6 Oxidation Homogeneous
Acrylonitrile 1,5 Ammoxidation Heterogeneous

Vinyl acetate 1,3 Oxidation Homogeneous
Thus, catalytic oxidation and carbonylation are good examples of high atom
utilization, low salt technolgies, e.g.
©
OH
+ H2~
Tl-slllcalite
- &OH ¢
OH
+
OH
+ H2O
OH
ref· 7 Atom utilization = 86X
-
~hl/I ]
ref· 8 Atom utilization = 100X

Figure 7. Catalytic oxidation and carbonylation as low-salt technologies.
106
In the fine chemicals industry, in contrast, catalytic technologies have been only
sporadically applied. 9 One reason for this is the much lower production volumes
involved. :-\evertheless, the number of kilos of byproducts (largely inorganic salts)
per kilo of product are much higher in fine chemicals and specialties as indicated
in Table 2.
Table 2. Byproduct formation in chemicals production.

1!'.'DUSTRY SEGMENT PRODUCT T01'<'NAGE KG BYPRODUCTS/
KG PRODUCT
Oil Refining 106 _108 ca. 0.1
Bulk Chemicals 104 _106 <1-5
Fine Chemicals 102 _10 4 5 -> 50
Pharmaceuticals 10-103 25 -> 100
The figures in Table 2 are based on the author's experience and are intended to
give an indication only. The substantial increase on going downstream is panly
due to the use of stoichiometric vs catalytic processes and partly to the fact that
fine chemicals and specialties are often produced in multi-step syntheses.
Of course comparing processes solely on the basis of the amount of waste is an
oversimplification. Obviously the ~ of the waste is also important; it certainly
makes a difference if it is sodium chloride or, for example, chromium or
manganese salts_ A more sophisticated assessment should, therefore, take into
account both the amount and the nature of waste, e.g.
Environmental = E x Q
acceptability Environmental x Unfriendliness
factor(kg waste/ quotient
kg product)
For example, if innocuous salts such as NaCl or NazS04 were arbitrarily given an
unfriendliness factor of 1, then chromium salts could be assigned a factor of say
100 and toxic metals e_g. Pb, Cd a 1000. Obviously these figures are debatable and
will vary from one company or production site to another, being panly dependent
on the ability to recycle a particular waste stream_
DEVELOPMENT OF ORGANIC SThl'fHESIS AND CATALYSIS
Another reason why catalysis has not been widely applied in the fine
chemicals industry is the more or less separate development of organic chemistry
and catalysis (see Figure 8) since the time of Berzelius, who coined both terms, in
1807 and 1835, respectively.
107
J.J. BERZELIUS
(1779-1848)
ORGANIC CHEMISRY CATALYSIS

(1807) (1835)
Urea Synthesis 1828 ca· 1900 Catalysis definition

(Wohler) (Ostwald)
First synthetic dye, 1856 Catalytic hydrogenation

Aniline purple (Sabatier)
(Peridn) ca· 1920 Petrochemicals
(based on oil)
DYESTUFFS INDUSTRY 1936 Catalytic cracking

(based on coal-tar) 1949 Catalytic reforming
1955 Ziegler-natta catalysts
FlNE CHEMICALS BUll( CHEMICALS

AND POLYMERS
CATALYSIS IN
ORGANIC SYNTHESIS
Figure 8. Development of catalysis and organic synthesis.
In the late nineteenth and early twentieth century catalysis developed largely as a
subdiscipline of physical chemistry. Following the advent of the petrochemicals
industry, catalysis was widely applied in oil refining and bulk chemicals
manufacture. Industrial organic synthesis, on the other hand, really began with
Perkin's serendipitous synthesis of aniline purple (mauveine) in 1856. This marked
the beginning of the synthetic dyestuffs industry based on coal tar· as the raw
material. The present-day fine chemicals and pharmaceuticals industries developed
largely as spinoffs of this activity.
Interestingly, Perkin was attempting to synthesize the pharmaceutical product,
quinine, by oxidation of allyl toluidine with potassium dichromate, according to the
following concept:
108
This, in hindsight, naive attempt was doomed to fail but one may sympathize with
Perkin when one realizes that the structure of the benzene ring was not even
known in 1856.
CLASSICAL vs NON-CLASSICAL AROl\1ATIC CHEMISTRY
Fine chemicals manufacture has, to this day, remained primarily the domain
of the synthetic organic chemist who, generally speaking, clung to the use of
Figure 9. Perkin's attempted synthesis of quinine (1856).
stoichiometric methods, e.g. oxidations with potassium dichromate, the very same
oxidant used by Perkin more than 130 years ago. Indeed, the fine chemicals
industry, with its roots in coal-tar chemistry, is rampant with classical
stoichiometric technologies that generate large quantities of inorganic salts.
Examples include sulfonation, nitration, halogenation, diazotization, Friedel-Crafts
acylations and stoichiometric oxidations and reductions (see Figure 10).
Many of these antiquated technologies are ripe for substitution by catalytic low-
salt technologies (see Figure 11).
109
©rCOR
~EOEL~~ ~
ACYlATION A1CIJ
~S03H
I
I
HALOGENATION' , SULFONATION ,
I SAND7/
x- CI· Br
DIAZOTI,~~~ ~O:;~C
~ON
~NHz ~OH
'REDUCTION'
DIAZOTISATION - IALKALINE FUSION
Figure 10. Classical aromatic chemistry.
110
TWO ROUTES TO HYDROQUINO?l."E
An instructive example is provided by hydro quinone manufacture (Figure 12).

Traditionally it was produced by oxidation of aniline with stoichiometric amounts
of manganese dioxide to give benzoquinone, followed by reduction with iron and
hydrochloric acid. The aniline was derived from benzene via nitration and
reduction and the overall process generated more than 10 kg of inorganic salts
MnS0 4, FeCI2, NaC!, Na2S04) per kg of hydroquinone. In contrast, a more
modern route to hydroquinone involves the (catalytic) autoxidation of p-diiso-
popylbenzene (produced by alkylation of benzene), followed by acid-catalyzed
rearrangement of the bis-hydroperoxide, and produces < 1 kg of inorganic salts
per kg of hydroquinone.
©rCOR
....--__0----, ~OzH R(-HV
I ~©
©r R
CATAlYnC
1CY1.All0N .0 AlCI,
../zEOLITE
CATAlYTIC
Al..l<'l'l-"110N
Ozjl
7 ~ '"
CATAlYnC
OXlDAl10N
OH
©r __ ©rN~
©r CHO
Q;AlYnC I I
HYDROXYlA11ON
I I CATAlYnC
L._AM:.:;;.::INIi.::.;:no.:.;.:,.:.N:.-...J
CATAlYTIC
CARBOlMA1l0N
I
Figure 11. Non-classical aromatic chemistry.
MENADIO?l."E STh'THESIS
Another pertinent example is the production of menadione (vitamin K3 ),

which traditionally involved the stoichiometric oxidation of 2-methylnaphthalene
with chromium trioxide and afforded 18 kg of inorganic salts per kg of product.
Tightening of environmental regulations has made it virtually impossible to carry
out this process in developed countries, which means that less demanding
production locations have to be found or alternative, deaner catalytic technologies
developed.
In the latter context the reported lO oxidation of 2-methylnaphthalene with
H 20 2 in the presence of a Pd II catalyst immobilized on an ion exchange resin
would appear to offer possibilities.
111
a a
N02
Q
NH2 0
Fe
- -
o
Mn02
HCI H2SO4
HN~
H,SO, ICLASSICAL I 0
ICATALYTIC Fej He,
~
C3 He
¢
OH
O2
catalyst - I -
(H+)
-2Me2CO
OH
Figure 12. Two routes to hydroquinone.
o
Oxidant
- «r~
o
Process Oxidant Yield
Classical CrOa 50-80X

(Stoicbdo~etric) in HoAc
Alternative 50-55X
(Catalytic)
Figure 13. Menadione synthesis.
112
IHOECHS~
ACz%
o / AlCl,
~
Hz CATALYTIC
(Pd/C) HYDROGENATION
CO CATALYTIC
(PdD ) CARBONYl.ATION
IBUPROFEN
(ANALGESIC)
Figure 14. Two routes to ibuprofen.
IBUPROFEN MANUFACTURE
Another case-in-point is the manufacture of the drug ibuprofen.H The latter is

an over-the-counter analgesic with annual sales and production volume of ca. $
1400 million and ca. 8000 tons, respectively, in 1990. Two routes for the
production of ibuprofen are compared in Figure 14. Both routes proceed via a
common intermediate, p-isobutylacetophenone. The classical route, used by the
Boots company (the discoverer of ibuprofen) involves a further five steps,
relatively low atom utilization and substantial inorganic salt formation.
The elegant alternative, developed by Hoechst Celanese 12 involves only two
(catalytic) steps from the common intermediate, 100% atom utilization and
negligible salt formation.
113
In both of the routes illustrated in Figure 14 the first step is a Friedel-Crafts
acylation. The latter is also a good example of a 'classical technology' involving
stoichiometric quantities of inorganic reagents, usually aluminium chloride (in the
Hoechst process HF is used). Obviously this is another technology that is ripe for
substitution by a cleaner catalytic method. In this context the recently reported 13
use of zeolites as heterogeneous, regenerable catalysts for Friedel-Crafts acylations
(see Figure 15) would seem to offer possibilities. Interestingly, the catalytic
method utilizes the carbo~:ylic acid as the acylating agent, thus avoiding the
formation of HC!.
ClASSICAL:
+ RCOCI AICI;, 1·' eq· _

+ HCI
r·t·
COR
Stoichiometric A1CI3 • not regeneroble
ZEOUTE CATALYSIS:
©
CH3
+ RCOzH
NoCe-Y
150°C - +HzO
COR
Heterogeneous. regenerable catalyst
REGIOSELECTIVITY:
o ~ ~
NoCe-Y :3 :3 94
A1C13 4 1.6 80
Figure 15. Friedel·Crafts acylations.
CATALYTIC RETROSYNTHESIS
The example of ibuprofen perfectly illustrates the benefits to be gained by

paying attention to the atom utilization in different routes and for being catalysis-
minded. Indeed, organic chemists should be urged to integrate these aspects into
, '4
their retrosynthetic thinking. Thus, in planning an organic synthesis a 'catalytic
retrosynthesis' could be constructed, identifying catalytic pathways to the desired
product.
Such a catalytic retrosynthesis for ibuprofen is shown in Figure 16. We suggest
that catalytic thinking would have easily identified the two routes shown, even
without the benefit of hindsight.
CARBONYlATI~
OH
~
HYDROGENATION ~ ~HYDROFORMYLATION
o
~
!'"RIEDEL-CRAFTS,
ACYlATION
Figure 16. Catalytic retrosynthesis of ibuprofen.
The second catalytic route to ibuprofen, identified via retrosynthetic analysis (see
Figure 16} has also been described in the literature. Thus, Neibecker and
coworkers 4 have described the regioselective hydroformylation of ~-iso
butylstyrene in the presence of a rhodium catalyst and Riley and coworkers the
catalytic autoxidation of the aldehyde product to ibuprofen (see Figure 17).
The synthesis of the starting material p-isobutylstyrene, from isobutylbenzene
(see Figure 18) has also been reported. 16 Interestingly, the first step may also be
amenable to replacement by one involving a zeolite as a regenerable,
hetergeneous catalyst.
115
CO/Hz
~h{eO)zeIJ
-z
eHO
25°C: 20 bar 94% regioselectivity

toluene
eHO
oGe:
de cane
10 bar
- eOzH
85% selectivity
89% conversion
Figure 17. Synthesis of ibuprofen via hydroformylation and oxidation.
92";
Figure 18. Synthesis of p-isobutylstyrene.
116
CATALYSIS, CHIRALITY AND BIOLOGICAL ACTIVITY
Ibuprofen is an interesting synthetic target from yet another point of view:

that of enantioselectivity. Many biologically active substances, e.g. pharmaceuticals,
pesticides, contain one or more asymmetric Centres and consist, therefore, of two
or more optical isomers which should be considered as different chemical entities.
In general, only one of the enantiomers (the so-called eutomer) of a racemic
mixture is responsible for the desired biological activity towards the target
organism. At best the other isomer (the distomer) constitutes unnecessary ballast,
the removal of which would significantly reduce the chemical burden on the
environment (be it the internal environment of the human body or nature at
large).
ICLEAN TECHNOLOGIES I
I
I I
CLEAN PRODUCTS
I I CLEAN PROCESSES I
I I
- LESS BALLAST
- LOWER TOXICITY &: I- WASTE MINIMIZATION
- MORE EFFICIENT PROCESSES
1
CHEMICAL BURDEN
ON THE ENVIRONMENT
- MORE EFFECTIVE.
TARGETED COMPOUDS
I
ENANTIOMERRICALLY I CATALYTIC PROCESSES I
PURE PRODUCTS
1
I
I
I
CATALYTIC ASYIvIMETRIC
SYNTHESIS
I
Figure 19. Clean technologies.
More often than not the distomer actually inhibits the desired effect of the
eutomer and/or exhibits undesirable side effects. Consequentl~, there is an
increasing trend to market such products as single enantiomers. 17,1 Thus, parallel
with the development of cleaner, catalytic processes is a trend towards
enantiomerically pure, i.e. 'cleaner' products. These two trends converge with the
application of catalytic asymmetric synthesis (see Figure 19).
117
Ibuprofen is no exception in this respect. It is the S-enantiomer of ibuprofen that
is solely responsible for the desired therapeutic effects. 11 ,19 The situation is
complicated, however, by the fact that the R-enantiomer undergoes in vivo
inversion to the S-enantiomer.19 This has led many observers to conclude that it is
of no consequence whether ibuprofen is administered as the racemate (as is
presently the case) or as the pure S-eutomer. However, it has been shown that
administering ibuprofen as the racemate (or as the R-enantiomer) leads to
accumulation of ibuprofen in adipose tissue whereas this does not occur with the
S-enantiomer. This could be an argument in favour of administering the pure
eutomer, thus creating a need for a (catalytic) enantioselective synthesis. In
principle, this could be achieved by carrying out the catalytic synthesis described
above in an asymmetric mode, i.e. with the aid of chiral ligands. The reported 20
catalytic asymmetric carbonylation of p-isobutylstyrene (Figure 20) is certainly
interesting in this context.
~ +CO+I\,O
02,THF,L* ,HCI
~dCI2/CUCI2 ]
-
R·T· , 1 bar S-ibuprofen
89" yield
83,. opticol yield
Substrate/L/CuCI2/PdCI2 20 : 1 : 2 : 4
o
II
OP(OH)z
S - BNPA
OP(OH)z
II
o
Figure 20. Synthesis of S-ibuprofen via asymmetric carbonylation.
SUMI\1ARY AND CONCLUSIONS
Summarizing, in the drive towards waste mInImIzation in chemicals

manufacturing, more and more classical, 'high-salt' organic chemistry will be
replaced by cleaner, catalytic methods. Companies that ignore this trend may not
118
survive to see the 21st century. In order to provide for a smooth transition organic
chemists at large should become more catalysis-minded and pay more attention to
the atom utilization of their reactions. This also applies to the teaching of organic
chemistry in academic institutions, which are, after all, the source of future
generations of industrial organic chemists.
An appreciation of the need for greener catalytic processes is often sadly
lacking in these institutions. In short, as we have noted elsewere,9 after more than
150 years in splendid isolation, organic synthesis and catalysis· are being brought
together again, a development that certainly would have appealed to Berzelius.
REFERENCES
1. Japan Chem. Week 30:1517 (1989); Eur. Chem. News 52:1375 (1989).
2. E. Drent, P.H.M. Budzelaar and W.W. Jager, Eur. Pat. AppL 0386833 (1990)
to Shell.
3. R.A Sheldon and J. Kochi, "Metal-Catalyzed Oxidations of Organic
Compounds", Academic Press, New York (1981).
4. RA Sheldon, "Chemicals from Synthesis Gas", Reidel, Dordrecht (1983).
5. J. Falbe, Ed., "New Syntheses with Carbon Monoxide", Springer-Verlag,
Heidelberg (1980).
6. H.M. Colquhoun, D.J. Thompson and M.V. Twigg, "Carbonylation: Direct
Synthesis of Carbonyl Compounds", Plenum Press, London (1991).
7. U. Romano et aJ., Chim. Ind. (Milan) 72:610 (1990) and references cited
therein.
8. D. Forster and T.W. Dekleva,1. Chem. Educ. 63:204-206 (1986).
9. R.A Sheldon, CHEMTECH 566-576 (1991).
10. S. Yamaguchi, M. Inoue and S. Enomoto, Chem. Lett. 827-828 (1985).
11. K.M. Williams, in "Problems and Wonders of Chiral Molecules", M. Simonyi,
Ed., Akademiai Kiado, Budapest (1990), pp. 181-204.
12. V. Elango, M.A Murphy, B.L. Smith, K.G. Davenport, G.N. Mott, and G.L.
Moss, Eur. Pat. Appl. 0284310 (1988) to Hoechst Celanese.
13. B. Chiche, A Finiels, C. Gauthier, P. Geneste, J. Graille and D. Pioch, 1.
Org. Chem. 51:2128 (1986); idem,1. MoL Catal. 42:229 (1987).
14. D. Neibecker, R. Reau and S. Lecolier, J. Org. Chel7l. 54:5208-5210 (1989).
15. D.P. Riley, D.P. Getman, G.R Beck and RM. Heintz, 1. Org. Chem. 52:
287-290 (1987).
16. I. Shimizu and Y. Tolmmoto, Eur. Pat. AppJ., 300,498 (1989) to Nippon
Petrochemicals; CA 111 :57241 u.
17. RA. Sheldon, Cllem. Ind. (London) 212-219 (1990) and references cited
therein.
18. RA. Sheldon, in "Problems and Wonders of Chiral Molecules", M. Simonyi,
Ed., Akademiai Kiado, Budapest (1990), pp. 349-386.
19. K. Williams, R Day, R Knihnicki and A. Duffield, Bioclzem. Phannacol. 35:
3403-3406 (1986).
20. H. Alper and N. Hamel, J. Am. Chem. Soc. 112:2803-2804 (1990).
119
INDUSTRIAL SCALE MEMBRANE PROCESSES
James R. Fair
The University of Texas at Austin
Austin, TX 78712
The use of perm-selective membranes for separating mixtures is far from new, and in
a sense is as old as the human race, since the kidney purification process involves selective
passage of materials through membrane walls. For many decades there has been some
commercial use of membranes for liquid mixture separation in dialysis and reverse osmosis.
More recently, membrane separations have become big actors on the stage of separation
technology, especially in the area of gas recovery and purification. The field has grown quite
large and the literature is voluminous. There is now a North American Membrane Society
and there has been for some years the Journal of Membrane Science. At several academic
institutions throughout the world, including The University of Texas at Austin, there are
major programs in membrane research.
The theme of.this symposium is industrial waste minimization, and it is within this
context that the present paper has been developed. As a further limitation on the scope of the
paper, gas separations will have some emphasis, although liquid separation applications will
be mentioned. Actually, many of the principles discussed will apply to either phase.
WASTE MINIMIZATION SEPARATIONS

It is rare that a process does not generate one or more by-products that must be
discharged, sold, or recycled to the processing operation. If the by-product has little intrinsic
value, and is completely compatible with the environment, then discharge is the course of
action likely to be followed. If it can be recovered and sold at a net profit, and the capital
investment is within the capability of the manufacturer, then good economic sense dictates
that this avenue be taken. If the by-product can be recycled to the process at an economic
advantage, and its recycling is within the technological capability of the manufacturer, then
prudence calls for doing so. These ideas are far from new, and will apply to any
manufacturing operation. The problem is that many by-products do not fall into this neat
categorization, or do not meet the premises just stated. Thus, there is a continuing general
Industrial E1IviroNN!lIlal Chemistry. Edited by D.T. Sawyer
and A.E. Martell, Plenum Press. New York, 1992 121
problem with process effluents that remain to be dealt with, and which contain components
that are not environmentally acceptable.
For chemical manufacturing, which represents one of the greatest opportunities for
waste minimization efforts, the problem by-products are often generated at a reaction step. It
is commonly observed that by taking a lower conversion per pass in the reactor, quantities of
by-products are decreased. A typical chemical process is shown schematically in Figure I,
where it is clear that separation steps can be important in recovering by-products for
recycling, discharge or (hopefully) for sale. It has been stated many times that separations
are a key ingredient of any waste reduction effort
By. prod~ ($) By. prod~ ($)
REAcrJON SEPARATION MAIN PRODUcr
,
WUle WUle
Figure 1. Typical chemical process flows
It is of interest here to consider the typical separation problem that is encountered in

reducing amounts of contaminants in streams leaving the process boundary. The
contaminants are often at low concentration and may be present in discharge gases or in
aqueous outflows. This does not suggest that more concentrated mixtures are not important,
but rather that conventional technologies such as distillation or extraction are likely to have
been applied already to their separation requirements. This leads to the question of whether
membranes can be useful for such separation needs.
MEMBRANE SYSTEMS
When applied to chemical processing, these systems comprise one or more modules
("permeators") that contain the membrane surface in a compact, high surfacelvolume ratio
arrangement. It may be necessary to filter the feed stream in order to avoid plugging
problems. Equipment may be needed to provide the pressure driving force for flow through
the membrane walls. Pre-heating of gases may be necessary to avoid condensed materials
that might damage the membranes. Finally, there may be piping arrangements for recycle of
streams back to the feed or to intermediate points in the process. A schematic flow
arrangement is shown in Figure 2, taken from the paper by Spillman (1989). The application
shown in the figure is for the removal of carbon dioxide from methane gas.
122
Bulk removal Purilicalion
• E
Siream
A e C 0 G
Composition (mole %)
CH, 93.0 98.0 49.2 96.1 56.1 72.1 93.0
CO, 7.0 2.0 50.8 3.9 439 27.9 7.0
Flow rate (MMSCFO) 20.00 17.95 2.05 19.39 1.62 1.44 1.01
Pressure (psig) 850 835 10 840 10 10 850
Methane recovery. 94.6%
Figure 2. Membrane system for treating natural gas.
The permeators used today are of two types. One type embodies a shell and tube type
affair with many hollow fiber membranes bundled into two tube sheets as shown in Figure 3.
Inside diameters of the fibers are usually in the range of 300 to 600 microns. The other type
of permeator is the spirally-wrapped variety as shown in Figure 4. Both types exhibit large
surface-volume ratios and can be used in either gas or liquid service. Commonly, the
membrane surfaces are of a polymeric nature; this suggests possible problems of
compatibility between the fluid mixture and the membrane material.
•Capped
• fibers
Bundle 01
_- '- hollow
fibers
ASME-coded
, •. caltlon·Sleel
shell
Feed -~]::;'
Separator rrodute may be Permeating

staQOd or run In paraUel lias oullel
Figure 3. Shell and tube type permeator utilizing hollow fibers typically 300-600 microns inside diameter.
123
Relect
Permeale ~~=" 1\1 _-.:' -. Permeate
Figure 4. Arrangements of membranes for commercial permeators: upper, hollow fibers; lower, spirally-
wrapped membrane sheets.
THE MEMBRANE SEPARATION BUSINESS

For liquid mixtures, reverse osmosis applications predominate, and these for the
purpose of water purification. An osmotic pressure must be applied to force pure water
through a microporous membrane wall, leaving behind the salts or other impurities entering
with the water. Fouling of the wall by these impurities ("polarization") is a major problem
with reverse osmosis installations and depends, quite naturally, on the nature and
concentration of the impurities.
For gas mixtures there are two large applications of membranes. One is for the
production of inerting nitrogen based on an air feed. Membranes are available for this
oxygen-argon-nitrogen separation, and membrane systems compete with pressure swing
adsorption systems for the inert nitrogen business. The other large application, and the one
that launched the modern gas permeation business, is the recovery of hydrogen from vent
streams or recycle streams in plants such as those producing synthetic ammonia. In addition
to these major uses there are hosts of other applications warranting the attention of membrane
researchers.
MEMBRANE RESEARCH AREAS

The research activity in membrane permeation can be classified as shown in Table 1.
Much of the research could be more broadly categorized as materials research: work directed
toward finding the right membrane material for a specific separation problem. For example,
of the acid gases, carbon dioxide is easily separated from natural gas by polymer membranes;
in fact, one of the "standard" test mixtures is methane/carbon dioxide. Another acid gas,
hydrogen sulfide, has not been as successfully separated with membranes, although work is
in progress and there has been developed a rubbery polymer that that can effect a good
124
Table 1. Research Areas, Membrane Penneation
• Transport through membranes

Rate
Selectivity
Pressure loss
Solubility
Diffusion
• Interaction of solutes with membranes

Enhancement of selectivity
Durability of membrane material
Preparation of membranes
Casting
Spinning
Conditioning
• Design and fabrication of penneators
Configuration of membrane systems

Staging
Counterflow
Integration with overall process
separation without serious deterioration of the membrane material.

For each material under consideration there is a tradeoff between rate of diffusion and
selectivity of the separation, as shown in Figure 5. The rate of permeation is inversely
proportional to the thickness of the membrane wall, and to this end asymmetric membranes
have been developed that have a thin (1000 Angstroms) "skin" to make the separation and
Figure S. Variation of selectivity with rate of penneation for the helium-methane system. Membrane
materials: PSF =polysulfone, PPO =polypropylene oxide. Symbols represent various plasticizers used in
the polymer membranes.
125
a more substantial support (with high porosity) below the skin. Still other studies of
membrane materials involve modeling of the movement through micropores and the
solubility/diffusion characteristics of molecules moving through nonporous membranes.
The interaction of solutes with membrane module materials has been mentioned; some
organics, for example, dissolve the membranes and/or the tube sheet materials that support
hollow fibers or spiral wrapping materials. Fouling of membranes is always a potentially
serious problem, and like the analogous case of heat exchanges, is not easily tractable. For
bioseparation applications, the crossflow filter is used, and the fouling with filter cake is
avoided by using very high stream velocities across the surface of the membrane material.
Preparation of membranes has become a major area of research. Conventional
spinning or casting have their limitations, especially when the asymmetric "skin", mentioned
earlier, is to be used. Should the membrane be formed and then treated to provide the skin?
Or should the skin migrate to the surface during the formation process? Here there are large
possibilities, particularly when co-polymers are being used for the separation.
Design and fabrication of permeators is a significant component of overall commercial
membrane separation technology, yet it is one that involves mechanical know-how that is
closely guarded by the vendors of permeators. One has the problem of arranging hollow
fibers such that there can be good penetration of the shell side fluid into the bundle. There is
the problem of pretreating the fibers or the wrappings such that early failures do not occur.
As alluded to earlier, there is the problem of selecting a potting material for the fiber tube
sheet that can resist such materials as ketones, aromatic solvents and alcohols.
The final entry in Table 1 deals with configuration of membrane systems. Such work
is invariably based on process simulators that can handle a variety of recycle streams and
membrane selectivity/rate characteristics. The fact that this is a mass separating agent type
process means that it is difficult to generalize on the rate/selectivity criterion. At any rate,
research in the area of configuration is carried out primarily by the vendors or, to some
extent, by processors with a need to optimize operating equipment; very little of such work is
being carried out in academia.
WASTE MINIMIZATION APPLICATIONS

The question arises as to how membrane systems can be used in the waste
minimization thrust. It should be clear from the foregoing that membranes must often be
tailor-made to fit an application; this makes it difficult to generalize for waste management
purposes. However, it has often been observed that membranes are much better for
concentration than for purification. And when a relatively pure product such as inerting
nitrogen emerges from a permeator, it is likely that the recovery of this material has not been
high. Thus, for a membrane unit producing high purity nitrogen a good deal of the nitrogen
entering with the air is discarded because of a low recovery, but considering this special case
of a "no-cost" raw material, and the fact that the by-product (enriched air) is not deleterious to
the environment (and might actually be saleable) makes the level of recovery only a concern
126
for the energy expended in compressing the feed stream. Incidentally, when viewed from the
oxygen balance, recovery is very high - but the recovered oxygen is not in high
concentration.
For waste minimization applications, however, recovery must be high, although
"minimization" does not necessarily equate to "elimination". This suggests that if membranes
are to be used, they might best be used in conjunction with one or more other separation
methods; such a combination has come to be known as a hybrid separation system. Some
examples of such systems will be discussed in the following paragraphs.
A representative problem of waste minimization deals with the removal of acid gases
from discharge gas streams. There are several methods available for this removal operation:
condensation, low-temperature absorption, higher temperature physical absorption and
chemical absorption. Figure 6 shows approximate regions of application of these methods
(Blytas, 1986). Thus, it might be economical to make a membrane separation fIrst, and then
follow with a chemical absorption step. Studies sponsored by the Gas Research Institute
have delineated operating conditions where acid gas removal is cheaper using membrane-
absorption hybrids than using straight absorption. (Kellogg, 1991). A flow diagram for
such an operation is shown in Figure 7.
1000
800
600
'in
co. 400
..,;
'"
J!!
200
.5
0'
<..)
100
<> 80
.,.,2!::> 60
I!! 40
co.
;;; (No H2S or other impurities present)
1::
8? 20
/
/ /
/ /
/ // / /
10 / / /
/ / /
8 / / / /
6 Ch~mical 'SOlvents /
4 /
/ / /
/ / /
/ /
/
/ /
/
2 /
/
/
/ /
/ /
/
/ / /
0.1 0.2 0.4 0.6 0.8 1 2 4 6 8 10 20 40 60 80 100

Partial pressure of CO 2 in product, psi
Figure 6. Areas of general applicability of separation methods for the removal of carbon dioxide from gas
streams.
127
Sales gas
Feed Membrane
~~r:::---=::J
__ ~ ___
Acid gas
Acid gas Amine
contactor
Amine stripper
Figure 7. Hybrid system for the removal of acid gases from natural gas.
Another application of hybrid systems deals with hydrogen recovery. As discussed

earlier, this separation may be made by membrane permeation, but classically it has also been
made by cryogenic absorption. Researchers at Air Products, Inc. have found applications
where a combination of these methods can be optimal (Agrawal et al., 1990), as indicated in
Figure 8. As shown, the membrane unit is upstream of the cold box; alternate schemes have
this sequence reversed.
MEMBRANE
UNIT
COLD
BOX 1---_-1
HYDROGEN PRODUCT
WASTE
Figure 8. Sequential membrane/cryogenic hybrid process for hydrogen recovery.
The use of membrane permeation for removing (and recovering) organics from
discharge air has been explored by Peinemann et al. (1986). Their proposed industrial flow
diagram is shown in Figure 9. The flows appear to be similar to those of an adsorptive type
recovery system. The problem is that the membranes may not be as selective as activated
carbon for separating the organics from air. Permeabilities measured by Peinemann et al. for
several organics in nitrogen, using a neoprene rubber membrane, are shown in Figure 10,
and clearly are quite high. If selectivity is defined as a ratio of permeabilities, then removal
128
of acetone, for example, from nitrogen would appear to be quite easy. Here one needs to
consider the basic flux equation for membrane permeation:
(1)
where Ni is the molar flux of component i per unit area, Pi is the permeability of component i
through a thickness z of the membrane, with a driving force of ~Pi partial pressure. The
ratio of permeabilities of two components i and j in the feed mixture may be called a
selectivity ratio:
(2)
Hen-brane
separatton
untt
Solvent
Oven
Condenser
....._ _ _ _ _........._ _ _ _-J ~ ------c~~fn~~=::::

bleed stream
Ltquld
solvent
Figure 9. Membrane system for removing and recovering organic solvents from discharge air.
In Figure 10, the indicated selectivity ratio, acetone/nitrogen, is in the range of 1000.
A convenient relationship for making use of this selectivity has been developed by
Peinemann et at. and is shown in Figure 11. The curves are based on an entering feed
composition of 0.5 vol-% organic (remainder air or nitrogen). The abscissa scale is the ratio
of total pressures of permeate and feed streams. Thus, for the case of acetone removal from
atmospheric air, and a vacuum of 50 mm Hg absolute, the permeate (material passing
through the membrane) would contain about 20 vol-% acetone, remainder air or nitrogen.
This might place a burden on the vacuum-producing equipment for large scale operation, and
would require refrigeration for condensing the acetone from the permeate gas. It is unlikely
that such a process could compete economically with a conventional carbon bed adsorption
process. However, a combination of membrane permeation followed by an activated carbon
system might represent an eC0nomical operation.
129
8/ Trlchloroethene
/
'ii
:c
E
C)
i
U
:: ~
N
E ;-
e e Acetone
I
E
If 103
~
I-
1/1
1"o
......
o
102
NEOPRENE MEMBRANE
101
Nitrogen
0 0
0
100
0 5 10 15 20 25
Pressure, em. Hg.
Figure 10. Permeabilities of several organics compared with the permeability of nitrogen at 40·C.
80
"if!.
C
0
:::
iii
0
D-
E
.
0
C)
0
•>
D-
C)
c
•E!'
0
!
•!§
GI
Do
0.001 0.01 0.1 1.0

Pressure ratio, CI
Figure 11. Penneate vapor composition as a function of permeate/feed pressure ratio and selectivity.
130
The use of membranes for the removal of small amounts of organics from water is a
potential liquid-phase application with waste minimization implications. Here one would
need to use a polymeric membrane in which the organics would selectively absorb and
diffuse to the permeate side of the membrane wall. Microporous membranes with a
hydrophobic treatment would also be a possibility. Work in this area is known to be in
progress, utilizing the process of pervaporation, in which the permeate is taken off as a
vapor. Pervaporation is often used to break organics-water azeotropes. with the water
selectively transferring to the vapor side. Thus. as is the usual case. a materials development
problem emerges: one must find (or design) the proper membrane material that can resist
attack by the organics. yet allow those organics selectively to move through the membrane
material. It would be completely infeasible to use the reverse osmosis approach of
transferring the water through the membrane.
FORWARD RESEARCH IN MEMBRANE TECHNOLOGY

Continuing research in the membrane area will concentrate on finding and developing
better membrane materials. Although there may be some indication in this paper that all
membranes are made from polymers, this is not the case. Membranes of ceramics and
carbon are very much under investigation, especially for higher temperature applications.
The ability to tailor-make membranes for specific separation needs is a key research goal.
There are a number of academic institutions with research concentrations in membranes, and
one of these is at the author's university, as part of the Separations Research Program. Some
of the more recent accomplishments in this particular program are indicative of research
accomplishments in the general field:
The ability to make ceramic-supported polymeric composite

thin film membranes.
A new process for making ultrathin (less than 1000

Angstroms), defect-free gas separation membranes.
Methodology for pre-conditioning membranes to improve
rate without loss of selectivity.
Development of fully aromatic polymers that have high

oxidation and thermal stability.
Improved understanding of how thin structure affects

transport as well as physical properties of membranes.
Ability to fabrication composite membranes.
Ability to form defect-free asymmetric membranes directly as

opposed to using coating techniques.
131
SUMMARY AND CONCLUSIONS
This paper has been directed toward waste minimization applications of membranes,
with the chemical process as the example application. The mixtures involved typically have
small concentrations of contaminants which must be removed almost quantitatively if exit
waste restrictions are to be recognized. Characteristically, membranes serve as
concentrators, which could mean that the majority materials (e.g., water or air) must be
removed from the mixture with minimum amounts of contaminants taken along. Thus, at
fIrst glance one would suspect that membranes might have a minor role to play in waste
minimization efforts. At least this has been the reputation of membrane separations in the
past.
Through continued and intensive research, this picture is changing. Coupling
membranes with other separation techniques to provide hybrid systems is proving to have
economic advantages in some cases; the concentration capability of the membranes provides a
more cheaply-separated mixture for the associated separation technique located downstream.
Special polymeric materials are being developed that can remove selectively organic
contaminants in waste waters. In general, there is a strong thrust to find or design special
membrane materials that can be used for separating particular molecules from mixtures.
While most of the effort seems to be devoted to gas mixtures, backgrounds in reverse
osmosis separations are leading researchers toward the large potential of removing low-
concentration level contaminants from liquid streams.
The research in membrane technology is taking place both in industry and in
academia. Naturally, more is heard from the academicians and reviews such as the present
one must rely largely on published works or in-house research. There are large areas of
research and development that are the special domain of industry: how the membrane
separator is integrated with the rest of the plant; how the membrane devices are constructed
and installed; how the membrane surfaces are formed most economically. Partnerships
between industry and academia are flourishing, and the end result can only mean one thing:
more and better ideas for minimizing wastes, and definite progress toward the ultimate zero-
discharge plant
ACKNOWLEDGMENT
The author wishes to thank Professor W. I. Koros, Separations Research Program,
The University of Texas at Austin, for his many helpful suggestions.
REFERENCES
Agrawal, R., Auvil, S.R., DiMartino, S.P., Choe, 1.S., Hopkins, I.A., 1988, Membrane/
cryogenic hybrid process for hydrogen purifIcation, Gas Sepn. Purif., 2: 9.
Blytas, G.C., 1986, Separation challenges and opportunities in the oil and petrochemical
industry, in "Chemical Separations," Vol. 2, C.I. King and I.D. Navratil, eds.,
Litarvan Literature, Denver.
McKee, R.L., Changela, M.K., Reading, G.I., 1991, C02 removal: membrane plus amine,
Hydrocarb. Proc. 70: (4) 63.
132
Peinemann, K.- V., Mohr, I.M., Baker, R.W., 1986, The separation of organic vapors from
air, AIChE Syrnp. Ser. No. 250, 82:19.
Spillman, R.W., 1989, Economics of gas separation membranes, Chern. Eng. Progr., 85:
(1) 41.
133
NOVEL MICROPOROUS MEMBRANE-BASED SEPARATION PROCESSES FOR
POLLUTION CONTROL AND WASTE MINIMIZATION
Chang H. Yun, Asim K. Guha/ Ravi Prasad,2 and Kamalesh K. Sirkar1•
Department of Chemistry and Chemical Engineering

Stevens Institute of Technology
Hoboken, NJ 07030
lDepartment of Chemical Engineering, Chemistry and Environmental Science
New Jersey Institute of Technology
Newark, NJ 07102
2Hoechst Celanese SPD
Charlotte, NC 28273
·Corresponding Author
INTRODUCfION
Multiphase equilibrium-based processes for separation and purification generally utilize

dispersed systems in which one phase is dispersed in the other as bubbles or drops or thin
films. Using microporous membranes, novel techniques have been developed such that
multi phase processes can now be carried out in a nondispersive fashion for gas-liquid
(Sirkar, 1992) and liquid-liquid (Prasad and Sirkar, 1992) contacting processes. Among such
processes, only nondispersive solvent extraction of pollutants using microporous membranes
will be of concern here. These processes employ immobilized immiscible phase interfaces
at the pore mouths in a microporous membrane. Through such interfaces, solutes are
extracted into the solvent as two immiscible phases flow on two sides of a microporous
membrane. Many advantages of such a technique over conventional extractors have been
summarized (Prasad and Sirkar, 1992).
Solvent extraction selectively extracts the solute/pollutant into the solvent from the feed
aqueous solution. If it is desired to further have the pollutant selectively transferred into
another aqueous phase in a concentrated form, a membrane may be employed between the
two aqueous phases. Liquid membranes are particularly useful for such purification-
concentration.
We have recently developed the concept of the hollow fiber contained liquid membrane
(HFCLM) which, unlike the supported liquid membrane (SLM) (Danesi et aI., 1987), is
highly stable (Sengupta et aI., 1992). Unlike membrane-based solvent extraction which
requires only one immobilized phase interface in a membrane, there are two immobilized
Industrial EnviroNMmal Chemistry, Edited by D.T. Sawyer

phase interfaces in two different sets of microporous hollow fiber membranes in a module
used in HFCLM technology. Feed aqueous solution containing the pollutant flows through
the bores of one set of fibers while a strip aqueous solution flows through the bores of the
other set of fibers. Any loss of the liquid membrane, contained between the two fiber sets
on the shell side, is automatically replenished from an external membrane liquid reselVoir.
The pollutant is first selectively extracted into the liquid membrane from the feed fiber set,
and then it is back extracted and concentrated into the strip aqueous solution flowing
through the other fiber set. Application of this technique to the removal of heavy metals
from wastewater and its simultaneous concentration in the strip aqueous solution is the
second novel process illustrated here.
The hierarchy of strategies utilized for environmental protection from pollutants may
be roughly identified in order of desirability as pollution prevention, waste minimization,
and end-of-pipe control. Although some chemical processes are being modified to approach
pollution prevention, many more are not; they are ideal candidates for waste minimization,
recycling and/or resource recovery. We will illustrate in this paper three types of examples
to that end. The first example will focus on membrane-based solvent extraction for removal
and recovery of toluene from a real aqueous waste stream whose toluene concentration is
reduced to 150-160 ppb without introducing any other pollutant. Membrane-based solvent
extraction is the basis in the second example for treating a synthetic aqueous waste stream
containing five priority pollutants at high concentrations to a level less than 25 ppm of each
pollutant. HFCLM technology is utilized next to remove heavy metals like copper or
chromium from an aqueous waste stream to levels around or less than 1 ppm while
concentrating the heavy metal simultaneously in the strip stream to a level 10-15 times the
feed stream level for recycling.
EXAMPLE I: TOLUENE REMOVAL TO SUB-PPM LEVEL BY MEMBRANE-

BASED SOLVENT EXTRACTION
In late 1989, we were approached by B.p.Goodrich & Co. (Brecksville, OH) to study
how the toluene level of a waste stream could be reduced from around 30 ppm to 30-40
ppb. The stream contained in addition heptane around saturation level besides polymeric
or particulate matter. If one can reduce the toluene level to 140-160 ppb, then dilution with
other plant streams will allow achievement of the targeted discharge level.
Recognize now that the other organic present in the waste stream, heptane, is not a
priority pollutant and its saturation level is low (-50 ppm). Further it is an excellent solvent
for toluene. We therefore decided to carry out membrane-based solvent extraction
removal/recovery of toluene using heptane as the extracting solvent. The extract may then
be separated by distillation or other processes to recover toluene and recycle heptane to the
extractor. The extracting solvent is no longer introduced fresh into the treated stream.
A hollow-fiber module, made by Hoechst Celanese SPD, Charlotte, N.C. was selected.
Fifteen em in length, 1.9 em in diameter, this module had 1,800 Celgard X-20 hydrophobic
microporous polypropylene fibers of 240 JAm I.D. and 290 JAm O.D .. The fiber wall had a
porosity of 0.4 and a pore size of 0.03 Jim. The fiber packing fraction was 0.43 and the
membrane surface area per unit volume was 47.5 em- t • A schematic of this type of module
is provided in Prasad and Sirkar(1989, 1990). General procedures for membrane-based
solvent extraction is available in the above references.
The experimental set up is shown in Figure 1. The extracting solvent heptane kept in
pressure vessel C was driven by nitrogen pressure in cylinder 2 to flow through the shell
side of the microporous hollow fiber(MHF) module. Feed aqueous waste stream from the
pressure vessel B was allowed to flow through a capsule filter (pore size: 10 Jim) to the
tube side of the MHF module countercurrent to the flow of the solvent heptane. The mode
of operation was once through in a continuous fashion.
136
VS C2 V6
R2 R1
V3 C1 V4
MHr
fiLTER MODULE
M2 M1
A B
Vl. ..... V6 ION-orr VALVE

M1 , M2 I F'INE METERING VALVE
A.B,C I PRESSURE VESSEL
C1 , C2 I CHECK VALVE
Rl , R2 I PRESSURE REGULA TOR
Figure 1. Experimental set up for membrane solvent extraction removal of toluene for an
actual plant waste stream sample using hollow fiber module.
137
Feed aqueous waste stream was initially introduced quickly and simultaneously into
pressure vessels A and B. The pressure vessel B was filled completely while the pressure
vessel A had a little space on top to drive the aqueous waste solution by N2 pressure from
cylinder 1. This led to a certain amount of toluene stripping in pressure vessel A but none
in pressure vessel B where a little dilution would occur due to mixing with the solution
from vessel A. Samples of the aqueous waste were withdrawn carefully and quickly before
and after the filter, capped in sample bottles filled to the top and sent to the plant for
toluene analysis. Similarly, samples of the treated aqueous waste, the raffinate, were
collected after Ml (Figure 1) with great care in covered sample bottles till they were full,
capped and sent for analysis.
The flow rates of the aqueous waste stream and the organic solvent (heptane) during
the experiment were 3.0 ml/min. and 7.7 ml/min. respectively. The pressure of the N2
cylinder 1 on the aqueous side was maintained at 9.5 psig while that for the solvent side
was 5.5 psig. This ensured that the solvent heptane was not dispersed in the aqueous phase.
Experimental Results and Discussion
The experimental results are summarized in Table 1. The original waste stream analysis
indicated toluene levels between 30-35 ppm. The toluene concentration in the feed stream
before the filter was between 13-15 ppm. This reduction indicates a certain amount of loss
in transferring from the waste drum to the feed reservoirs and any stripping in vessel A.
The toluene concentration in the feed stream after filtering varied between 9 and 11 ppm
due to any adsorption in the filter. The treated waste stream (raffinate) concentration of
toluene was in the range of 157-200 ppb.
These results demonstrate toluene level reduction by two orders of magnitude to sub-
ppm levels (0.1-0.2 ppm). The raffinate samples were taken only after 45 minutes of
operation; further samples were taken at 20 minute intervals. Because of the steady
behavior of the raffinate, any effect of adsorption-based separation is ruled out.
Reduction of priority pollutant concentration by two orders of magnitude via
continuous countercurrent membrane-based solvent extraction employing a microporous
hollow fiber module has already been achieved (Yun et aI., 1992). However, as we will see
in the next section, the wastewater pollutant concentration levels in that study were far
higher to start with. Thus raffinate pollutant concentrations were around -25 ppm.
The results of Example I of the present study provide opportunities for the following
generalizations on aqueous waste stream treatment/recovery:
(a) Multipollutant containing waste streams may be subjected to solvent extraction based
remediation if the extracting solvent is relatively benign(e.g. heptane) or if the benign
extracting solvent is already present in the waste stream. This assumes, however, that
the pollutants are removable by extraction and the extracting solvent can be recovered
and recycled.
(b) If the multipollutant waste stream contains, amongst others, a priority pollutant (e.g.
toluene) which is also a good solvent for the other pollutants present, then remove and
recover the other pollutants using this priority pollutant as an extracting solvent unless
a relatively benign solvent is available. This method is especially useful if the different
pollutants require different treatments as encountered in specialty chemical industry.
(c) On the basis of the present results and the principle of countercurrent extraction, we
can argue that a membrane solvent extraction-based remediation can reduce pollutants
to sub-ppm levels.
Along with many other advantages of microporous membrane-based solvent extraction,
these features point to many opportunities for microporous membrane-based solvent
extraction in pollution control and waste minimization.
138
Table 1. Toluene concentration in various streams 1•
Toluene concentration
Sample identification
ppb
Original waste stream 30191
35693
Feed #1 before filtering 14499
13250
Feed #1 after filtering 9145
9460
Raffinate #1 209
185
197
Feed #2 before filtering 13418
13454
Feed #2 after filtering 10739
10674
Raffinate #2 189
188
Raffinate #3 184
180
Raffinate #4 163
170
Raffinate #5 157
165
Raffinate #6 157
161
lSamples analyzed at Research Laboratories of B.F. Goodrich, Brecksville, OH.
139
EXAMPLE II: MULTI POLLUTANT REMOVAL FROM HIGH STRENGTH
WASTE WATER BY MEMBRANE-BASED SOLVENT EXTRACTION
One often encounters aqueous waste streams containing a large number of priority
organic pollutants. Further the pollutant concentrations may be high. These pollutants may
be high boiling, low boiling, polar, nonpolar, aromatic, aliphatic, chlorinated or otherwise.
Incineration costs may be too high. Membrane-based solvent extraction may be employed
here. The extract of pollutants and the extracting solvent may then be separated and the
solvent recycled by distillation. Alternately, the extract can be incinerated at a cost order-of-
magnitude lower than that for the wastewater since the solvent volume can be so much
smaller with orders of magnitude higher concentration of pollutants. Only the extracting
solvent is introduced into the treated stream and may have to be removed downstream.
The system studied was a synthetic waste water containing the following priority
pollutants: phenol (500 mgll), 2-chlorophenol (1241 mgll), nitrobenzene (604 mgll), toluene
(216.5 mgll) and acrylonitrile (3627 mgll). The extracting solvents studied were isopropyl
acetate (lPAc), methyl isobutyl ketone (MIBK) and hexane. This problem has been studied
and analyzed; mass transfer characterization of the pollutants being extracted in the
microporous hollow fiber module has been carried out and are available in Yun et ai.
(1992). Here we present some additional results for the cases where the solvent is either
IPAc or hexane.
The hollow fiber module employed has already been described under Example I. The
experimental apparatus employed is quite similar to that in Figure 1 except for the
following: no filter was needed in the aqueous waste line before the membrane module;
further only one aqueous waste stream reservoir was used instead of two shown in Figure
1. Aqueous phase pressure (10 psig) was maintained 5 psig higher than the organic phase
pressure to prevent phase dispersion in the hydrophobic membrane module. Unlike that in
Example I, the phase flow rates were varied over a significant range to study pollutant
removal capabilities. Aqueous phase concentrations were determined by HPLC.
Figure 2 illustrates the aqueous stream concentrations of four pollutants exiting from
the hollow fiber module at different aqueous phase flow rates for a given extracting solvent
(lPAc) flow rate; the figure also provides the percent removal of the pollutants in this
continuous hollow fiber extractor. At low aqueous phase flow rates, we observe from Figure
2(a) that the individual pollutant concentration of each pollutant has been brought down to
10-15 mgll. The concentration of toluene is not being reported here since it was below the
detection levels being employed. We also observe that reduction of acrylonitrile
concentration is the most difficult since its inlet concentration is very high and it has a low
distribution coefficient(-6) in IPAc compared to the other pollutants(phenol-53; 2-
chlorophenol-250; nitrobenzene-300 etc.).
Figure 3 illustrates the pollutant concentrations in the aqueous treated stream at the
hollow fiber module exit for a given flow rate of hexane as an extracting solvent. Since
hexane has a very low affinity for polar pollutants like phenol, 2-chlorophenol and
acrylonitrile, the extents of their removal are much lower. Considerable deliberation is,
therefore, needed on the selection of the extracting solvent.
A polar solvent like IPAc or MIBK is very efficient (Yun et aI., 1992); however, since
their solubility in the treated water is significant, additional treatment will be needed. If,
however, the pollutants can be removed by a nonpolar solvent like heptane or hexane,
additional treatments may be avoided. On the other hand, the treatment plant size or
membrane area required with hexane as the solvent may be quite large since the pollutant
distribution coefficients in the case of heptane or hexane are usually low for polar
pollutants.
An additional feature of membrane solvent extraction is that the maximum level of the
extracting solvent introduced into the treated stream is governed by its solubility since there
140
300
"ea
Qo-0.16 ml/s
lie
250
• phenol(500 mg/I)
'V 2-chlorophenol( 1241 mn/l)
6 nitrobenzene (604 mg/l
...
.2 200 0 acrylonitrile (3627 mg/I)
......c
III
150
Solvent: [PAc 0
(a)
III
U
C 0
0 100
u
...'>(
.. -
r.l 50
o .1iI iI. ~
...c.,
0.00
100 - ...
0.05 0.10 0.15
W'
0
0.20
•
0.25 0.30
~
0.35
95 0
...III
~
"0 90
...
CI.
0 85 Solvent: [PAc (b)
OJ Qo =0.16 ml/s
~
e
0
•
80 0
III
phenol (500 mg/l)
Il:: v 2-chlorophenol (1241 lng/I)
~
75 6 nitrobenzene (604 mg/l)
0 acrylonitrile (3627 mg/l)
70
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Aqueous Phase Flow Rate, ml/s
Figure 2. (a) Concentration and (b) percent removal of the priority organic pollutants in
the exit aqueous phase from the experimental hollow fiber module (solvent: IPAc).
141
3500 0
~ 3000 0
Solvent: IPAc
c:: 2500 0
.
.2 Qn=0.13 ml/s
....
f
c:I
2000 • phenol (500 mg/l)
'\l 2-chlorophenol( 1241 mj/l
t;;. nitrobenzene ~604 mg/l
(a)
~ 1500
0 acrylonitrile ( 627 mg/ )
.=
o
c.l 1000 '\l
'\l
•
'M
~ 500
•
t;;.
•
t;;.
80
60
....
o ffSolvent : Hexane 1 (b)
.. 40 f..f Q o =O.l ml/s 1
~
a
II
0
c:t: 20 0
•
~
'\l
0
0.0
•
0.1 0.2 0.3
0
0.4
Aqueous Phase Flow Rate. ml/s
• :phenol(500 mg/l) '\l :2-chlorophenol(1241 mg/l)
t;;. :nitrobenzene(604 mg/I) 0 :acrylonitrile(3627 mg/l)
Figure 3. (8) Concentration and (b) percent removal of the priority organic pollutants in
the exit aqueous phase from the experimental hollow fiber module (solvent: hexane).
142
is no dispersion. In fact, Yun et a\'(1992) obselVed the level of MIBK in treated water to
be lower than its solubility and decreasing with decreasing residence time. These features
suggest that microporous membrane-based solvent extraction can be a highly efficient
technique for waste minimization and pollution contro\.
EXAMPLE III: REMOVAL OF HEAVY METAL AND ITS SIMULTANEOUS

CONCENTRATION AND RECYCLE BY HOLLOW FIBER CONTAINED LIQUID
MEMBRANEPERMEATOR
We illustrate now the hollow fiber contained liquid membrane (HFCLM) permeator
performance in purifying an aqueous waste stream of a heavy metal like copper or
chromium and simultaneously concentrating it in a strip aqueous solution for recycle. The
structure of HFCLM permeators has been illustrated in Sengupta et a\. (1988) and Basu and
Sirkar (1991) for liquid phase separations. The permeator in this particular study had two
sets of hollow fibers, each set containing 300 hydrophobic microporous polypropylene
fibers (Celgard X-10, 100 pm 1.0., 150 pm 0.0.). The ends of each set of fiber at each
permeator end were set apart. In the middle of the permeator, however, these two sets were
well mixed and bundled into a 0.61 em 1.0. teflon sleeve which was inserted into a 1.05
em 1.0. stainless steel pipe. Each end of each fiber set was potted with epoxy. A general
schematic of such a permeator in a loop with a feed solution and a strip solution in
countercurrent flow is shown in Figure 4.
Figure 5 illustrates the performance of such a permeator when a synthetic aqueous
waste stream containing 208 mgll of copper present as cupric sulphate (feed pH-4.55) was
passed through the feed fiber set while the strip aqueous solution containing 200 gil of
H2S04 flowed cocurrently through the strip fiber set. The liquid membrane present on the
shell-side and the fiber wall pores was 10 v/v % LIX84 diluted in heptane.
At the aqueous-organic interface in the feed fiber set, copper was extracted into the
organic liquid membrane by the forward reaction
Cu ++(aq) + 2RH(org) .. R 2 Cu(org) + 2H+(aq) (1)
whereas at the aqueous-organic interface in the strip fiber set, the reverse reaction takes
place liberating copper into the aqueous strip solution. The liquid ion exchanger LIX84
exchanges H+ with Cu++ to complex copper ion and extract it into the liquid membrane
from the feed solution. The strip solution pH is so low that copper is easily stripped; by
maintaining a low strip solution flow rate, copper is easily concentrated.
As shown in Figure 5, at low aqueous feed flow rates, Cu+· concentration at the outlet
of feed fibers is reduced to around 1 mg/l while the strip outlet concentration is around 900
mg/l. At higher waste stream flow rates, feed outlet concentration is somewhat higher due
to insufficient membrane area in this module while the strip concentration has been
increased to 1250 mg/l. Other data not being reported here indicate strip solution
concentration of copper at a level of 4000 - 6500 mg/l showing simultaneous cleanup of
the waste stream of the toxic heavy metal copper as it is concentrated to a high level in the
strip solution for recycle and recovery. Figure 6 illustrates percent Cu" removed for
different concentrations of LIX84 in the liquid membrane.
Figure 7 illustrates a silI?ilar purification-concentration behavior in the removal of Cr6+
(pH=2.43, obtained by dissolving ~Crp7 in deionized water). The level of chromium is
reduced to less than 1 mg/l whereas the strip stream concentration is raised to 750 mg/l.
The liquid membrane employed was 20 v/v % trioctylamine diluted in xylene and the strip
solution was 0.1 (N) NaOH. We have since studied this system more extensively; Cr6+
concentration in the exiting strip stream has been raised to 4000 mg/l from a 250 mg/l feed
143
Figure 4. Schematic of the experimental set up using HFCLM permeator.
Membrane: 10% v/v LIX 84 in Heptane

50 ~--r---r---r---r---r---r-~~-'r-~~-, 1400
Feed : 208 mg/l Cu
~ Strip: 200 gil H2 S0 4 "-...,

e
40
e
u='
Strip Flow Rate: 0.16 ml/min
• 12008
....o
....0 30
°
I:
I:
0
o
:;J
°• -
- Feed Outlet
:;J
CI
20 '1 ....f
....c
So.
Strip Outlet
r 1000 ~
CI)
o
• °
0 I:
I: o
u
0 10 U
•
o L-----'_--L_--'-_...L_........:O:::..-1---o_-L_--'------l 800
0.00 0.25 0.50 0.75 1.00 1.25
Feed Flow Rate. ml/min
Figure 5•. Effects of feed flow variations on feed and strip outlet concentrations.
144
100
0
o o
90
o o
:::I
U o
.....0
80
OJ
>
0
e
Q) 70
et:: Membrane Liquid: Feed Cu
~ v/v % LlX84 Cone., mg/l
0 10 208.0
60 l:>. 20 200.0
0 30 240.0
50
0.0 0.5 1.0 1.5
Feed Flow Rate, ml/min
Figure 6. Percent Removal of copper for different concentrations of LlX84 in heptane.
Membrane 20% v Iv Trioctyl Amine in Xylene

10 700
I
•
0
- Feed Outlet
- Strip Outletr
0
.:::::.. .:::::..
e 8 600
QII
e
QII
+
'"UI- ..
+
I-
6 500 u
.....
.....0 0 0
I:: I::
...
.2 4 400
0
:;j
...
III
I-
I::
Feed: 100.0 Cr 8 + mg/l (pH=2.43) ...III
I-
I::
Q) Strip Flow Rate: 0.21-0.32 ml/mir Q)
tl
tl
I:: 2 Strip : 0.1 (N) NaOH 300 I::
0
0 U
U 0
o
• • • 200
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Feed Flow Rate. ml/min
Figure 7. Effects of feed flow variations on feed and strip outlet concentrations.
145
solution. Further, on the based of long term studies, we have found such a system to be
quite stable in its behavior (Basu, 1990).
These and other studies being currently investigated suggest that removal of heavy
metals from waste streams and their simultaneous concentration and recycle via HFCLM
permeators are feasible and should be explored further. The strip stream, concentrated in
the heavy metal, may easily be the basis for recycling and resource recovery.
REFERENCES
Basu, R., 1990, "Mass Transfer Enhancement in Phase Barrier Membrane Separators for
Reactive Systems," Ph.D. Diss., Stevens Institute of Technology, Hoboken, NJ.
Basu, R., and Sirkar, KK, 1991, Hollow fiber contained liquid membrane separation of
citric acid, AIChE J. 37:383.
Danesi, P.R., Reichley-Yinger, L., and Rickert, P.G.,1987, Lifetime of supported liquid
membranes: the influence of interfacial properties, chemical composition and water
transport on the long term stability of the membranes, J. Membr. Sci. 31:117.
Majumdar, S., Sirkar, KK, and A. Sengupta, 1992, Hollow fiber contained liquid
membrane, in: "Membrane Handbook," W.S. Winston Ho and KK Sirkar, eds., Van
Nostrand Reinhold, New York.
Prasad, R., and Sirkar, KK, 1989, Hollow fiber solvent extraction of pharmaceutical
products: a case study, J. Membr. Sci. 47:235.
Prasad, R., and Sirkar, KK, 1990, Hollow fiber solvent extraction: performances and
design, J. Membr. Sci. 50:153.
Prasad, R., and Sirkar, KK 1992, Membrane-Based solvent extraction, in: "Membrane
Handbook," W.S. Winston Ho and KK Sirkar, eds., Van Nostrand Reinhold, New
York.
Sengupta, A., Basu, R., and Sirkar, KK, 1988, Separation of solutes from aqueous
solutions by contained liquid membranes, AIChE J. 34:1698.
Sirkar, KK, 1992, Other new membrane processes, in: "Membrane Handbook," W.S.
Winston Ho and K.K. Sirkar, eds., Van Nostrand Reinhold, New York.
Yun, C.H., Prasad, R., and Sirkar, KK., 1992, Membrane solvent extraction removal of
priority organic pollutants from aqueous waste streams, 1& E Chern. Res. (in press).
146
ION EXCHANGE FOR GLYPHOSATE RECOVERY
Lowell R. Smith and Jane L. Barclay

Monsanto Agricultural Company
A Unit of Monsanto Company
800 N. Lindbergh Boulevard
St. Louis, Missouri 63167
INTRODUCTION
The first generation Monsanto glyphosate· (Active ingredient in

Monsanto's Roundup® herbicide, N-phosphonomethylglycine)
manufacturing process, operated since 1976, produced a stream to
waste treatment containing a significant amount of gIyphosate. As a
result of several research and engineering innovations, a new process
was developed and a new plant recently commenced operations which
reduced the glyphosate lost to waste treatment by a very significant
amount. This result was mainly due to the use of an innovative ion
exchange process for glyph os ate recovery from the waste stream
which also contained the process byproducts.
THE GLYPHOSATE WASTE STREAM
The glyphosate process produces a spectrum of impurities as

shown in Table I, some of which are stronger acids than glyphosate
and some of which are weaker acids. In addition, glyphosate and
some of the impurities carry a basic function.
ION EXCHANGE RESIN EVALUATION
A process was conceived which would use these differences in

acidity as the basis for an ion exchange recovery method. For this pur-
pose, ion exchange resins would have to be found which were capable of
separating glyphosate from these impurities and which would concen-
trate it up to useful levels.
Ion exchange resins are of four general types as shown in Figure 1.
In addition to acidity, selectivity in ion exchange depends on several
factors among which are molecular size, charge and polarity. The
'rubutrUJI EllviroNlleltlal cllmWtry, Edited by D.T. Sawyer

and A.E. Martell, Plenum Pres., New Yodt, 1992 147
Table 1. Composition of typical illtrate.
COMPONENf
PHOSPHONOMETHYLIMINODIACETIC ACID 0.01
N-FORMYL GLYPHOSATE 0.01
HYDROXYMETHYLPHOSPHONIC ACID 0.002 STRONG ACIDS
PHOSPHATE 0.02
PHOSPHITE 0.004
IMINOBISMETHYLENE PHOSPHONATE 0.035
GLYPHOSATE 1.50
METHYL GLYPHOSATE 0.06
FORMALDEHYDE 2.85
FORMIC ACID 1.50 WEAK ACIDS
AMINOMETHYLPHOSPHONIC ACID 0.35
P-S03H
Weakly basic anion resin Strongly acidic cation resin
P-COOH
Strongly basic anion resin Weakly acidic cation resin
P = Polymer bead, polystyrene, acrylic, etc.

Figure 1. Types of common ion exchange resins.
ability of a molecule or ion to diffuse into the resin can be as important

as the strength of the bond made with the resins functional group.
There are no ways to predict resin selectivity for organic molecules but
empirical tests can be used to determine selectivity of the resins.
These tests generally are done in a batch or column mode. In the batch
mode, the resin is simply stirred with a solution of an excess and equal
amount the substances which are to be separated, the resin is filtered,
extracted and the extract analysed. This gives a crude idea of the
preference of the resin for the various substrates. Promising resins then
may be tested in the more accurate column mode in which the solution
is passed through a column of the resin and fractions of the effluent are
taken and analysed.
The glyph os ate waste problem mainly required the separation of
148
N-phosphonomethyliminodiacetic acid (PIA) and formic acid from the
glyphosate. This is because formic acid will react with glyphosate
downstream to give N-formylglyphosate and PIA was thought to be
representative of the strong acids chromatographic behavior. Cation
resins were found to have no affinity for any of the materials in the
filtrate. Formaldehyde. being neutral, would not be absorbed. In
general. weakly basic anion resins are preferable to strongly basic
resins because of their greater ease of regeneration after use. Batch
resin tests using commercially available basic anion resins were
performed for the purpose of studying the preference of the resins for
PIA over glyphosate and for glyphosate over formic acid. The resin
selected for these separations need not be the same resin. The results
are shown in Tables 2 and 3.
Table 2. Separation of N-phosphonomethyliminodiacetic acid (PIA)

and glyphosate.
RESIN RATIO PIA/GLYPHOSATE

N1SORR!W
AMBERLITE®*l IRA-68 17.5
DUOLITE®*2 A-392 34.0
AMBERLYST®*l A-21 no glyphosate
IONAC®*3 A-305 12.1
IONAC AFP-329 53.8
IONACA-365 9.7
IONACA-380 17.1
AMBERLITE lRA-400 2l.3
AMBERLITE IRA-900 23.0
AMBERLITE lRA-93 no glyphosate
LEWATIT®*4 MP64 24.1
*1 AMBERLITE and AMBERLYST are registered trademarks of Rohm & Haas Co.
*2 DUOLITE is a registered trademark of Diamond Shamrock Corp.
*3 IONAC is a registered trademark of Sybron Chemicals. Inc.
*4 LEWATIT is a registered trademark of Farbenfabrlken Bayer
Table 3. Separation of glyphosate and formic acid.
RATIO GLYPHOSATEI
FORMIC ACID ABSORBED
AMBERLITE IRA-93 1.79
DUOLITE A-392 4.66
AMBERLITE IRA-68 1l.3
IONAC305 10.0
IONAC380 6.6
IONAC365 5.4
149
In addition to these commercial resins several experimental anion
resins from Dow were tested. Because of their superior adsorption
ratios Amberlite IRA-93 and Amberlite IRA-68 were chosen for
separation of glyphosate from PIA and formic acid respectively.
COLUMN EXPERIMENTS
Experiments were performed using columns of 80 ml. (1" dia-

meter tubes) of IRA-93 and IRA-68 in sequence. Conductivity
measurements were recorded on the effluent from the columns to
determine when the columns were exhausted. After washing the col-
umns with water. the materials were recovered and the resins were
regenerated by passing dilute sodium hydroxide or an amine solution
through the columns. By analysing fractions of the effluent and the
regenerant solution. the capacity of the columns for strong acids and
glyphosate could be determined. It was found that. because of the rel-
atively low concentration of strong acids. the column for the absorption
of glyphosate (main column) was exhausted long before the strong acid
column (precolumn) and that a large excess of glyphosate was needed
to force the formic acid off the main column.
THE ION EXCHANGE SYSTEM
The ion exchange recovery system which was developed based

on the above study consisted of four columns of equal size. one con-
tained IRA-93 and three contained IRA-68. The system was run as
shown in Figure 2.
"CHNtGE "IMfrfCOl..l..lWlWASiANOClVPHOSATEAECO\o'£RY
FClAMM.DEHYDE fQUIC ~o
EJ o
FOfMCACID ETC. TO
ETC. WASTE
II PAECOll. . . RECOVERY NlDREOEHERATIOH .) NTERRUN IPA) COUECllON
HCl~. r;:l
::R
N
ca..- ca. .....
RECYClE
AfTEfWlJN TO MAKE FA 9OLunoN

tuCI ..... PHOSPHATE TQWASTE
Figure 2. Ion exchange system.
During filtrate addition. effluent from the precolumns is sent directly

to the main columns and the strong acids are retained. When the pre-
column is saturated with strong acids. it is washed with water and. if
recovery of the strong acids is desired. a dilute hydrochloric acid
solution is added. This is followed by a water wash and dilute sodium
hydroxide solution which regenerates the resin. An additional water
wash is then provided. During the caustic addition and subsequent
150
water wash the pre column effluent is sent to waste.
The effluent from the precolumns during filtrate addition flows
through three of the main columns until significant glyphosate breaks
through the second column which shows that the first column is
sufficiently loaded with glyphosate. The three columns are washed
with water. Then an isopropylamine solution in an afterrun solution
from a previous run is applied to the first column. When effluent from
the first column contains glyphosate isopropylammonium salt of
sufficient concentration, as indicated by conductivity, columns two and
three are taken off line and the effluent from column one is sent to a
product tank. Water is applied to column one to wash the last portion
of the product off the column. When the effluent reaches a dilution
which is unsatisfactory for product collection an afterrun is collected
which is used to prepare the next isopropylamine solution. Results
are shown in Table 4.
Two pre columns and four main columns are needed for continu-
ous operation.
Table 4. Results of ion exchange. % of total.

N-phosphono- Amino-
Applied Glypho- methylimino Fonn- Fonnic N-Methyl- N-fonnyl- methyl Phos-
Centrate!gl ~ diacetic acid aldehyde acid glyphosate glyphosate phosphonic phate
1098.7 1.89 0.020 1.99 0.85 0.156 0.11 0.33 0.02
1078.8 1.41. 0.005 2.00 0.87 0.074 0.076 0.36 0.07
592.2 1.49 0.005 2.12 0.94 0.081 0.076 0.37 0.02
542.2 1.49 0.005 2.12 0.94 0.081 0.076 0.37 0.03
266.0 1.35 0.003 2.01 0.92 0.069 0.076 0.35 0.02
Recovered
Prpducts!gJ
300 3.66 ND" ND 0.074 0.18 0.01 0.03 0.002
300 4.34 ND ND 0.101 0.31 0.005 0.06 0.005
ND"= not detected
RESULTS
As can be seen, formaldehyde and PIA have been reduced to undetect-
able levels and the other impurities have been reduced to managable
levels. The glyphosate concentration has been increased by a factor of
about 2.5. The ion exchange product can be mixed with the main
product stream and sent to formulation. An accelerated resin life test
in the laboratory which put the resins through 1500 recovery cycles
indicated little or no deterioration of its effectiveness. This would be
more than two years of plant use.
SUMMARY
This ion exchange technology has been transferred to plant use
with a scaleup factor of 90,000 and is very successful. The plant had
151
an easy startup and has operated extremely well. There is no evidence
of resin degradation.
In summary. reliable new ion exchange technology has been devel-
oped which has reduced glyphosate loss to waste treatment by a very
significant amount.
152
SUPERCRITICAL EXTRACTION IN ENVIRONMENTAL REMEDIATION AND
RESTORATION
Aydin Akgennan

College Station, TX 77843
INTRODUCTION
A supercritical fluid (SCF) is a fluid at conditions above its critical temperature and
pressure. Interest in the extraction of solid and liquid media by SCFs have increased during
the last decade due to: (1) environmental problems associated with common solvents
(mostly chlorinated hydrocarbons); (2) the increasing cost of energy intensive separation
processes (for example distillation); and (3) the inability of conventional separation
processes to provide the necessary separations needed in emerging new industries. The
attractive physicochemical properties of SCFs qualify them as a viable alternative to
conventional solvents used in extraction processes.
At temperatures and pressures above its critical point a pure substance exists in a state
that exhibits gas-like and liquid-like properties. The fluid's density would be very close to
that of a liquid, the surface tension is very close to zero, the diffusivity and viscosity have
a value somewhat in between that of a liquid and a gas, and most important the solvent
power of a SCF is related to its density which can be varied over a very wide range by
small variations in temperature and/or pressure in the supercritical region. These properties
result in several advantages in extraction such as ease of solvent recovery, elimination of
residual solvent in the extracted medium, lower pressure drops, and higher mass transfer
rates.
In extraction of a solute from a matrix (such as water) the choice of the solvent
depends on two criteria, its immiscibility with the matrix and the solubility of the solute
in the solvent. The solubility in SCFs is a strong function of density. In the vicinity of the
critical point, l<1'r<1.1 and 1<Pr<2, the density is a very strong function of both the
temperature and the pressure. The solvent characteristics of a SCF can therefore be adjusted
as desired, an important advantage compared to conventional solvents. At extraction
conditions the solvent power would be high so the SCF fluid can remove the solute from
the matrix and at the separation/solvent recovery stage (where the solute is removed from
the solvent) the solvent power would be reduced to close to zero.

and A.E. Martell. Plenum Press. New York, 1992 153
The use of a compressed gas in a separation process was first proposed for oil
deasphalting (Wilson et al., 1936). Although the process was not strictly supercritical, it did
take advantage of the change in solubility associated with a pressure reduction. Later, Elgin
and Weinstock (1959) reported on a phase-splitting technique for recovery of
methyl-ethyl-ketone (MBK) from water using supercritical ethylene. Since then SCFs have
been used as solvents to extract a variety of solid and liquid matrices such as coal
(Wilhelm and Hedden, 1983), caffeine from coffee (Zosel, 1978; Vitzthum and Hubert,
1975), tobacco (Hubert and Vitzthum, 1978), fruit aromas (Schultz and Randall, 1970), and
alcohols from water (McHugh and Krukonis, 1986; Kuk and Montagna, 1983). Various
symposia proceedings and books that are on supercritical extraction phenomena are
available in the literature.
The broad range of organic solutes that can be extracted from aqueous and solid waste
(wastewater, contaminated soils, sludges etc.) by SCFs as well as the aVailability of
inexpensive, readily available, and non-toxic SCF solvents such as CO2 , has directed
attention to this process as a viable method for removing toxic organic compounds. In the
following discussion aqueous waste and solid waste are treated under separate headings
since the fundamentals of the technology as well as the application techniques are slightly
different.
SUPERCRITICAL EXTRACTION OF AQUEOUS WASTE
The advantages of employing SCF extraction to remove organic contaminants from

wastewater can be realized by considering a typical extraction process (Figure 1).
Contaminated water is extracted by a SCF in a countercurrent flow extraction column. The
extract stream goes through a pressure reduction to separate the contaminant from the SCF
which is recompressed and recirculated. The raft-mate (decontaminated water) is also
expanded to recover dissolved SCF solvent. The process would typically be at ambient
Extract
I I
I I
I I
I I
+
Contaminant
I I
I I
I I sCF
I I
I I Compressor Recycle
I I
I I
Make-Up
SCF
De-Contaminated Water
Figure 1. Schematic representation of a typical supercritical extraction process (from Akgerman et al., 1991).
154
temperature and pressures above 5 MPa. If a liquid solvent was used for the process two
additional separation processes would have been necessary, separation of the solute from
the solvent for solvent recovery and separation of residual solvent from the water stream.
Energy shortages of 1970s fueled studies on separation of alcohols from water.

Therefore earlier work on SCF extraction of an aqueous medium concentrated on
alcoho1/water separations. Alcohols of interest were ethanol, n-propanol, i-propyl alcohol,
and n-butanol and the thermodynamics of the ternary systems (alcoho1/water/SCF) are well
studied (paulaitis et al., 1984). SCF extraction of various other organic compounds are also
reported such as acetone from water (Panagiotopoulos and Reid, 1985), aroma constituents
of fruit and other foods (Schultz and Randall, 1970),23 organic compounds typically found
in water for analytical purposes (Ehntholt et al., 1983).
Stringent environmental regulations demand removal of trace organics from water

making SCF extraction a viable alternative to other remediation technologies such as
bioremediation, adsorption, liquid extraction, air-stripping, distillation, incineration, etc. all
of which have their advantages and disadvantages. Supercritical carbon dioxide (SCC02)
is the preferred SCF solvent for environmental applications since it is non-toxic,
inexpensive, and readily available. In addition, it has a conveniently low critical
temperature (304 K) and a moderate critical pressure (7.39 MPa).
Economic evaluation of a separation process necessitates the distribution coefficient

(also called partition coefficient, equilibrium constant, and K value) of the extracted
component between the phases, i.e. for SCF extraction of a solute from water the phases
would be the aqueous phase and the SCF phase. Thus, studies to date concentrated on the
thermodynamics of the SCF/water equilibrium. We have reported on distribution
coefficients of single components such as phenol, benzene, toluene, naphthalene, and
parathion (Roop and Akgerman, 1989; Roop et al. 1989; Yeo and Akgerman, 1990) as well
as mixtures such as a solution containing benzene, toluene, naphthalene and parathion (Yeo
and Akgerman, 1990), a phenolic mixture (Roop and Akgerman, 1990), and petroleum
creosote (Akgerman et al., 1991). Similarly Ghonasgi et al. (1991) and Knopf (1991)
reported on extraction of phenol, m-cresol, p-chlorophenol, and benzene both individually
and as a mixture.
The thermodynamic modeling of the equilibrium starts with equating fugacities of the
component distributed between the two phases
~ WP ~ SCFP
Ii =fi
where the superscripts WP and SCFP refer to the water phase and the supercritical fluid
phase respectively. Expressing the fugacities in terms of the mole fractions and fugacity
coefficients results in the distribution coefficient
~ WP~ SCFP
Ki = Xi SCFPIXi WP = <l>i l<I>i
The fugacity coefficient <l>i can be calculated from an equation of state using standard
rigorous thermodynamic relationships. In our studies we have used the Peng-Robinson
equation of state (Roop and Akgerman, 1989; Yeo and Akgerman, 1990) whereas Ghonasgi
et al (1991) used the Carnahan-Starling-DeS antis- Redlich-Kwong equation of state. Figures
2 & 3 present data on benzene extraction (Yeo and Akgerman, 1990; and Ghonasgi et al.,
1991) and Figure 4 on phenol extraction (Roop and Akgerman, 1989) with predictions. It
is important to note the difference in the magnitude of distribution coefficients for
hydrophobic and hydrophillic compounds.
155
.
.............................. "0............ .
........................................ ·6· __ -- ____ _

..
1(T ----
----~-
-------------------.--
CD
::J
~
~
CD 1<T
c:
CD
N
c:
CD
III
10' • = Ternary data at 318 K

--Ternary model at 318 K
O=Six component data at 318 K
····-Six component model at 318 K
A=Six component data at 330 K
- - - Six component modal at 330 K
7.60 8.32 9.04 9.76 10.48 11.2

Pressure (MPa)
Figure 2. Distribution coefficients of benzene between water and supercritica1 carbon dioxide in ternary and
six component systems (from Akgerman et al., 1991).
• 313 K
....000
323 K
3.000
SIi"
:ooC 2•000
- Calculated by CSDRK EOS
X X
X 323 K. Experimental
1.000 X • 313 K. Experimental
8 9 10 11 12 13 1... 15 16 17 18 19 20
Preau,. (MPa)
Figure 3. Distribution coefficients of benzene between water and supercritical carbon dioxide (from Ghonasgi
et al., 1991).
156
2.0r----------------------,
1.S
• = 298 K
0= 323 K
r:a - = model
~
~ 1.0
~
0.5
•
o ~ __ ~~ __ ____
~ L __ _ _L __ ___J
o 7 14 21 28 35
PRESSURE (MPa)
Figure 4. Distribution coefficients of phenol between water and supercritical C8Ibm dioxide (from Entrainer
Effect for Supercritical Extraction of Phenol from Water, Roop and Akgerman, 1989, American Chemical Society)
EFFECI' OF ENI'RAINERS IN AQUEOUS SYSTEM EXTRACTIONS
Solvent power of SCFs can be increased significantly by addition of small amounts of

co-solvents, called entrainers. Entrainers usually are polar compounds that are added to the
SCF in small amounts (typically <5%) and they increase the solubility of organics in the
SCF by orders of magnitude. Methanol and toluene are the most widely studied entrainers.
In the literature there are significant amount of studies reporting on the solubility increase
of solids in SCFs by entrainers. The increase in solubility is explained using solution
theories. However, studies on the effect of entrainers on SCF extraction, i.e. the distribution
coefficients, are more limited.
Roop and Akgennan (1989; 1990) showed that methanol has no effect on the
distribution coefficients of phenol and a phenolic mixture proving that an entrainer that
increases the solvent power does not necessarily affect the partitioning of a compound
between phases. They have explained the entrainer effect in terms of multicomponent
molecular interactions in calculation of thennodynamic eqUilibrium. Through a quaternary
system equilibrium calculation it was shown that methanol will not have any effect of
phenol distribution coefficients and that benzene would increase the distribution coefficient
of phenol by 50-70% and th$' prediction was verified experimentally. Figures 5 and 6
indicate the entrainer effect of benzene on the distribution coefficient of phenol and a
phenolic mixture. Similarly Yeo and Akgennan (1990) determined the effect of benzene,
toluene, naphthalene and parathion on each other's distribution coefficient. Figure 7
indicates the increase in the distribution coefficient of each component in the mixture due
to the presence of others as compared to single component extraction. A significant
conclusion is that the extraction of all the components are enhanced in mixtures, i.e.
extraction thennodynamics of mixtures is more favorable.
157
2.0.-----------------------,
1.5
Temp. = 298 K
... 17.3 MPa
0 _ 27.6 MPa
- .. model
0.5
OL-_ __ J______-L______J -______ ~ ______ ~
o 2 .. 6 8 10
MOLE PERCENT BENZENE
Ftpre 5. Data aod predicition on the effect of benzene as an enttaiDer on the distribution coefficients of
phenol between warer aod supercritical CO. (from Entrainer Effect for SupercriticaJ Extraction of Phenol from
Water, Roop and Akgerman. 1989. American Chemical Society)
lO~--------------------------------------~
Temp - 298 K
• - 17.2 MPa
2 • - 27.6 MPa
o L-______ ~ ______ ~ ______-L______ ~ _______ J

o 1 ~ 2 , 4
Mole Percent Benzene (based on CO2)
Ftpre 6. Data aod predic:tion ClD tile effect of benzene as an entraiDer ClD the distribution coeflicienlS of a
pbenoIic: mixture between water aod supeIai.tical co.
(from Distribution of a Complex Phenolic Mixture
Between Water and SupercriticaJ CO2, Roop and Agkerman. 1990. American Chemical Society)
158
300r-------------------------------,
1=Benzene
2=Toluene
3=Naphthalene
4~Parathion
oS
CD
~ 100
~
(,,)
c:
o
2 3 4
Organic Compounds
Figure 7. Increase in the distribution coefficients of each component in a six component system comparted
to the ternary system values (from Yeo and Akgennan, 1990).
SUPERCRITICAL EXTRACTION OF SOLID WASTE
Supercritical extraction has been demonstrated in the literature at the bench scale for
extraction of organic contaminants from a variety of solid matrices. Capriel et al. (1986)
used supercritical methanol to extract bound pesticide residues from soil and plant residues.
Hawthrone and Miller (1986) extracted polycyclic aromatic hydrocarbons (PAR) from
diesel soot and Tenax packing for gas chromatographic columns by supercritical carbon
dioxide. Schantz and Chesler (1986), similarly, used supercritical CO2 to extract
polychlorinated biphenyls (PCB) from sediment and PARs from urban particulate matter.
Methanol/N20 mixtures are also used to extract PAHs from river sediments and urban
particulate matter (Hawthrone and Miller, 1987). Supercritical CO2 is used to remove
hexachlorocyclohexane, parathion, PCBs, and PAHs from Tenax packing (Raymer and
Pellizari, 1987) and polyimide based adsorbents (Raymer et al., 1987). Supercritical CO2
is also the solvent of choice for extraction of PARs from various adsorbents selectively by
varying the operating conditions (Wright et al., 1987a; 1987b). A significant amount of
work concentrated on activated carbon regeneration, such as desorption of phenol (Kander
and Paulaitis, 1983), pesticides (DeFilippi et al., 1980), acetic acid and alachlor (Picht et
al., 1982), and ethyl acetate (Tan and Liou, 1988).
The results of these studies have demonstrated at the bench scale that it is possible to
extract compounds with molecular weight as high as 400 at mild conditions and selectively
if desired. This conclusion constitutes the basis of supercritical chromatography. Recently
we have reported on the application of chromatography theory to supercritical extraction
from solid matrices (Erkey and Akgerman, 1990).
159
Concerning environmental applications, Kingsley (1985) applied subcritical and
supercritical CO2 for extraction of oil from metal imes (mill scale) and bleaching clay in
the pilot scale. The process operated on a semi-batch mode and the results indicated that
the recovery of extractable material depended on the solvent flow rate. It was also observed
that, for the bleaching clay, the recovery was improved by a static soaking period before
extraction. Brady et al. (1987) demonstrated the ability of SCFs to extract PCBs and DDT
from contaminated topsoil and subsoils using SCC02. The effect of entrainers on the
solvent power was also determined (Dooley et al., 1987). They demonstrated that 100%
DDT extraction was possible when 5% by weight methanol was used as entrainer whereas
when toluene was used as the entrainer (5% by weight) only 75 % of the DDT was
recovered. Eckert et al. (1986a) studied the removal from soil of chlorinated aromatics such
as trichlorophenol as a model compound for PCBs and dioxins. Using supercritical ethylene
virtually all trichlorophenol was removed from soil.
Supercritical extraction of organic contaminants from dry solids, such as soil,

adsorbents, etc., involves two different phenomena simultaneously. The organic contaminant
on the solid surface can exists in two states, adsorbed state and the deposited state. The
portion of the organic that is deposited as a separate phase on the solid surface is extracted
by simple dissolution in the supercritical phase. On the other hand the extraction of the
portion of the organic that is adsorbed on the solid phase is controlled by the
adsOIption/desOIption equilibrium. Entrainers play a significant role in extraction of the
deposited separate organic phase since the solubility of the organic in the SCF phase is
significantly increased due to the entrainer. In addition, SCFs have an advantage compared
to conventional solvents since they penetrate the pore structure more efficiently due to their
zero surface tension and remove the organic condensed/deposited in the pores of the matrix.
Extraction of organics from a solid matrix that is also wet (contains water such as soil
moisture) is more complicated since the organic now can exist in four different states,
adsorbed on dry soil, dissolved in soil moisture, adsorbed on soil which is covered by a
water layer (partitioning between soil and water), and as a separate organic phase. Again,
the extraction of the separate organic phase is a dissolution phenomenon. However, the
extraction of the other forms involves different types of equilibrium (soil/water/SCF
partitioning) combined with extraction of multiple species (water and the organic
contaminant[s]).
Most important of these extraction processes is the extraction of the adsorbed species
which involve solid/organic/SCF binary and ternary interactions. Thermodynamics of solids
extraction by supercritical fluids is formulated in a similar manner to any adsorption
process. The distribution constant of a species between the solid phase and the SCF phase
is more often referred to as the adsorption equilibrium constant or the partition coefficient.
The partitioning is explained in terms of an adsorption isotherm. The adsorption
equilibrium constant is deimed as
"Kt"" = mass fraction in the SCF phase/mass fraction in the solid phase
which, for dilute systems (mass of contaminant compared to the solid and the SCF mass
negligible) is related to the true thermodynamic distribution coefficient in terms of the mole
fractions by the expression
where the superscript SP refers to the solid phase (Hess et al., 1991).
160
1.20 r-------:=-=-=-"""7~_r_--""""""T~......,.......,
e= 2500 psia
_=
0= 3200 psia
4000 psia
1.15
1.10
1.05
1.00
0.95 L.....~~~~~~':-'-~~.........~~.....J
3.0 3.1 3.2 3.3 3.4
l/T x 103 (1/K)
Figure 8. Isobaric temperatuIe dependence of die partition coefficient of bexachlorobeozeoe between soil
aDd supeteritical co, (from Supercritica/ Extraction of Hexach/orobenzene from Soil. Akgerman et al .• 1992.
American Chemical Society)
1
2.46
E
.j!!
is 1.80
i
.§
I 1.1&
.-III.OK
0-101.4 K
e-11•.OK
0.&0 ,---,~~ .........L-......~~-,-""",~~'--'.~~~"""",

-0.8 -0.8 -0.4 -0.2 0.0
In [p(1O-~. ]
Figure ,. Isotheamal deusity depeDdellce of die partition coefficienta of oapbthaleDe between activated
aIumiDa aDd sapercritical CO2 (from Blkey and AltgeJman, 1990).
161
The adsorption equilibrium constant has important thennodynamic properties. Through
the van't Hoff equation, isobaric reciprocal temperature dependency of the logarithm of
the adsorption constant yields the partial molar enthalpy of adsorption, which is a measure
of the strength of adsorption (Hess et al., 1991; Akgennan et al., 1992). Similarly, the
isothennal density dependence of the adsorption constant yields the partial molar volume
of the solute in the supercritical phase, which is related to its solubility in the supercritical
phase (Erkey and Akgennan, 1990).
Figure 8 shows the isobaric temperature dependence of the partition coefficient of
hexachlorobenzene between soil and SCC02 (K'A is reciprocal of the adsorption
equilibrium constant at inImite dilution). The positive slopes indicate exothermic heats of
adsorption increasing slightly with pressure (Akgennan, et al., 1992). Figure 9 shows the
isothennal density dependence of the partition coefficients of naphthalene between activated
alumina and SCC02 and Figure 10 shows the partial molar volume of naphthalene in
SCC02 compared with data from Eckert et al. (1986b). Both Figures are taken from Erkey
and Akgennan (1990).
A common myth in environmental circles postulates that organic contaminants partition
only to the organic portion of soil (soil humic material) and correlations are developed for
1000
-1000
-2000
-3000
=0
.~E -4000
8- -6000
t>- -8000
'I'
S2
-7000
-8000
-9000
. - 308.4 K (Eckert et al.,1888)
-10000 • - 818.2 K (Eckert et al.,1888)
+- this work
-11000
0.20 0.46 0.70 0.96
10·3p(kg/m 3 )
Figure 10. Partial molar volwnes of naphthalene in supercritical CO2 obtained from data presented in Figure
9, and comparison to experimental values (from Erkey and Akgerman, 1990).
162
partitioning in tenns of soil organic content. Although these correlations are useful for
practical applications they do not have a 'fundamental significance. Figures 11 and 12
compare the partition coefficients for phenol between SCC02 and 8 solid matrices, six soils
of varying organic content, sand and montmorillonite (Hess et al., 1991). These results
indicate that solid matrix surface area correlates the partition coefficient over a wide range.
We have recently measured the adsorption isothenns of phenol from SCC02 on to soil
and activated carbon over the concentration range from zero to saturation solubility limit
\ Sand. K = Infinite
10'
.
Ir-
~
SYMBOL CODES
O=SOIL 1
§ .-SOIL 2
Q) O-SOIL 3
:::J
"'iii e=SOIL4
>
0 f!.=SOIL 5
•
~
0 A.-SOIL 6
10°
••
l.-
I::!.
!I
~
Montmorillonite. K = 0.08
10"
0.0 0.5 1.0 1.5 2.0
Weight loss (%)
Figure 11. 1be partition coefficients of phenol between supercritical CO. and eight solid matrices. correlation
with organic content (from Supercrirical Extracrion of Phenol from Soil. Hess et al .• 1991. American Chemical
Society)
(Figures 13 and 14). Both isothenns are Type I - Langmuir Isothenns indicating monolayer
surface coverage. Similar behavior is observed for adsorption of anthracene, triphenylene,
perylene, and phenanthrene on soil from SCC02 (Andrews et al., 1990). In fact, by
estimating the projected area of the adsorbed species and measuring the surface area of the
adsorbent, it is possible to calculate the maximwn amount that can be adsorbed on the
solid. These results indicate that the total accessible surface area of the solid is the
available area for unifonn adsorption and the amount of solute adsorbed in excess of
unifonn monolayer coverage would be deposited as a separate phase.
163
10'
.t
Sand. K = Infinite
SYMBOL CODES
10' ;... o = SOIL 1
§
... _=SOIL 2
o-SOIL 3
• = SOIL 4
Q)
:J 6-S01L 5
co> 0 A-SOIL 6
:..::
.:>
••
10°
Montmorillonite. K ... 0.08 -.

at 37.3 (m 2 /gm)
10"
4.0 7.0 10.0 13.0 16.0
Surface Area (m 2 /gm)
FIpre 12. 'lbe partition coeflicieDlS of phenol between supercritical CO2 and eight solid matrices. correlation
with solid phase snrl'ac:e uea (from Supercritical Extraction of Phenol from Soil, Hess et al., 1991, American
Chemical Society)
0.014
0.012
~ 0.010
E
~
~ 0.008
0
c(])
.c 0.006
0.
C/)
~ 0.004
0.002
O'OOO¥---,--.---.---r----,-----.--...,.----/
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16
C (moles Phenofl/iter CO2)
Fipre 13. AdsoIption isotherm of phenol from supercritical CO2 onto soil at lOO abn and 3~
164
0.36,-r---------------------,
t:
0.30
~
() 0.24
E
E!1
e~ 0.18
~
~ 0.12
~
c:l
0.06
0.00+'----...-----...-----...-----...-------1
0.00 0.05 0.10 0.15 0.20 0.25
C (moles Phenol/liter C02)
Figure 14. Adsorption isotbellD of phenol from supercritical ~ onto activated carbon at 100 atm and 36"C.
Modeling of a solids extractor involves hydrodynamic and mass transfer parameters in

addition to the adsorption equilibrium and kinetics. The model used is nonnally a
heterogeneous model that employs a dispersed flow equation in the extractor bed combined
with an equation for diffusion into solid particles which includes an accumulation tenn for
adsorption/desorption. The two equations are coupled through the external film mass
transfer boundary condition and the surface accumulation tenn is related to the adsoIption
isothenn. Solution of the set of equations yield the adsoIption or desorption breakthrough
profiles (Sherwood et al., 1983). The parameters involved are the axial dispersion
coefficient, the film mass transfer coefficient, the effective diffusivity, and the adsorption
isothenn parameters. We have shown that the effective diffusivity and the adsorption
parameters are the controlling parameters at nonna! operating conditions (Erkey and
Akgennan, 1990). Nonnally, at least at laboratoIy- conditions, the bIeakthroughs BR; sharp,
indicating control by adsorption equilibrium constant.
APPLICATIONS
The CF Systems Corporation developed a SCF extraction process to separate and

recover oils from refinery sludges and to extract hazardous organic compounds from
wastewater, sludge, sediment and soil (Hall et al., 1990). The process uses supercritical
propane on contaminated solids and SCC02 to treat wastewater. The solid feed materials
are reduced in size and are slurried so that they can be pumped to the extractor.
Wastewater is used directly. The process closely resembles the flow chart presented in
Figure 1. Reportedly, up to 90% of the solvent is recycled in the system; the remaining
10% retains the extracted contaminants. The CF Systems process was demonstrated at pilot
scale for the U.S. EPA's SITE program and shown to be capable of removing PCBs from
sediments.
Recently we proposed an alternate scheme for extraction of solids as shown in Figure

IS. This scheme eliminates the expansion of the SCF for separation of the extracted
165
COMPRESSOR
RECIRCULATION
PUMP
Figure 15. Schematic representation of a soil supercritical extraction process with subsequent adsorption of
the extracted contaminants.
TABLE I. Comparison of Costs of Alternative Technologies for Soil Remediation, from

Carpenter (1986)
Supercritical Water Oxidation $250 - 733 per m3 soil
KPEG $211 - 378 per m3 soil
O. H. M. Methanol Extraction $400 - 514 per m3 soil
Acurex Solvent Wash $196 - 569 per m3 soil
Vitrification $255 - 548 per m 3 soil
Bio-Clean $191 - 370 per m3 soil
Soilex Solvent Extraction $856 - 913 per m3 soil
Chemical Waste Landiill $260 - 490 per m3 soil

---------------------------------------
Supercritical CO2 Extraction $120 per m 3 soil
(Combined with Adsotption)
166
contaminants that necessitates a costly re-compression stage. Instead, the extracted
contaminants are deposited on a suitable adsorbent for destruction by incineration. Costs
compare favorably with other treatment techniques as summarized in Table 1. The cost data
are based on phenol and details of cost estimate are given elsewhere (Akgerman and Erkey,
1992).
CONCLUSIONS
There is no single magic process that will be applicable to remediation and restoration
of all contaminated sites, in fact, more than one technology is usually needed for
remediation of the media at a given site. Thus SCF extraction is a viable alternative for
remediation of wastewater contaminated with hydrophobic compounds and soils, sludges
and sediments contaminated with non-volatile heavy compounds such as polycyclicaromatic
hydrocarbons (PAHs), polychlorinatedbiphenyls (PCBs), pesticides, dioxins, etc.
SCF extraction has been shown to be capable of removing organics from water, soil,
sediments, sludges, and a variety of solid adsorbents. Although the available data are still
very limited, the behavior of SCFs and their solvent powers have only recently begun to
be understood. As the ability to predict the behavior of contaminants in a SCF increases,
the use of SCF extraction will fwd new applications in environmental remediation.
REFERENCES
Akgerman, A., Roop, R. K., Hess, R. K., and Yeo, S.-D., 1991, Supercritical extraction in
environmental control, Chapter in: "Supercritical Fluid Technology: Reviews in
Modem Theory and Application", Bruno, T. J. and Ely, J. F., Eds., CRC Press,
Boca -Raton.
Akgerman, A., Erkey, C., and Ghoreishi, S. M., 1992, Supercritical Extraction of
hexachlorobenzene from soil,Ind. Eng. Chem. Res., 31:333.
Akgerman, A., and Erkey, C., 1992, Supercritical extraction conibined with adsorption for
environmental remediation of contaminated soils, paper presented at AIChE
Meeting, Minneapolis.
Andrews, A. T., Ahlert, R. C., and Kosson, D. S., 1990, Supercritical extraction of
aromatic contaminants from a sandy loam soil, Environ. Progr., 9:204.
Brady, B. 0., Kao, C. C., Dooley, K. M., Knopf, F. C., and Gambrell, R. P., 1987,
Supercritical extraction of toxic organics from soil, Ind. Eng. Chem. Res., 26:261.
Capriel, P., Haisch, A., and Khan, S. U.; 1986, Supercritical methanol: an efficacious
technique for extraction of bound pesticide residues from soil and plant samples,
J. Agric. Food Chem., 34:70.
Carpenter, B. H., 1986, PCB Sediment Decontamination - Technical/Economic Assessment
of Selective Alternative Treatments, EPA/600/2-86/112, Environmental Protection
Agency, Washington, D.C.
DeFilippi, R. P., Krukonis, V. J., Robey, R. J., and Modell, M., 1980, Supercritical fluid
regeneration of activated carbon for adsorption of pesticides, EPA-600/2-80-054,
Environmental Protection Agency, Washington, D. C.
Dooley, K. M., Kao, C., Gambrell, R. P., and Knopf, F. C., 1987, The use of entrainers in
the supercritical extraction of soils contaminated with hazardous organics,lnd. Eng.
Chem. Res., 26:2058.
Eckert, C. A., Van AIsten, J. G., and Sroicos, T., 1986a, Supercritical fluid processing,
Environ. Sci. Techno/., 20:319.
167
Eckert, C. A, Ziger, D. H., Jolmston, K. P., and Kim, S., 1986b, Solute partial molal
volumes in supen:ritical fluids, I. Phys. Chem., 90:2738.
Elmtholt, D. J., Thrun, K., and Eppig, C., 1983, The concentration of model organic
compounds .present in water at parts-per-billion levels using supercritical fluid
c:aIbon dioxide, Int. I. Environ. Anal. Chem., 13:219.
Elgin, J. C., and Weinstock, J. J., 1959, Phase equilibria molecular transport
thermodynamics, I. Chem. Eng. Data, 4:3.
Edcey, C., and Akgerman, A., 1990, Chromatography theory: application to supercritical
extraction, AlChE I., 36: 1715.
Ghonasgi, D., Gupta, S., Dooley, K. M., and Knopf, F. C., 1991, Supercritical CO2
extraction of organic contaminants from aqueous streams, AlChE 1,37:944.
Hall, D. W., Sandrln, J. A., and McBride, R. E., 1990, An overview of solvent extraction
treatment technologies, Environ. Progr., 9(2):98
Hawthrone, S. B., and Miller, D. J., 1986, Extraction and recovery of organic pollutants
from environmental solids and Tenax-GC using supercritical carbon dioxide, I.
Chromatogr. Sci., 24:258
Hawthrone, S. B., and Miller, D. J., 1987, Extraction and recovery of polycylic aromatic
hydrocarbons from environmental solids using supercritical carbon dioxide, Anal.
Chem., 59:1705.
Hess, R. K., &key, C., and Akgennan, A., 1991, Supercritical extraction of phenol from
soil, I. Supercrlt. Fluids, 4:47.
Hubert, P., and Vitzthum, O. G., 1978, Fluid Extractions of hops, spices and tobacco with
supercritical gases, Angew. Chem. Int. Ed., Engl., 17:710.
Kander, R. G., and Paulaitis, M. E., 1983, The adsotption of phenol from dense carbon
dioxide onto activated carbon, in: "Chemical Engineering at Supercritical
Conditions", Paulaitis, M. E., Penninger, J. M. L., Gray, R. D., Jr., and Davidson,
P., Eds., Ann Arbor Science, Ann Arbor, MI.
Kingsley, G. S., 1985, Pilot plant evaluation of critical fluid extractions for environmental
applications, EPA/600/2-85ft)81, Environmental Protection Agency, Washington,
D.C.
Knopf, F. C., 1991, private communication.
Kuk, M. S., and Montagna, I., 1983, Solubility of oxygenated hydrocarbons in supercritical
c:aIbon dioxide, in: "Chemical Engineering at Supercritical Conditions", Paulaitis,
M. E., Penninger, I. M. L., Gray, R. D., Ir., and Davidson, P., Eds., Ann Arbor
Science, Ann Arbor, MI.
McHugh, M. A. and Krukonis, V. I., 1986, "Supercritical Fluid Processing: Principles and
Practice", Butterworth, Stoneham, MA.
Panagiotopoulos, A. Z., and R. C. Reid, 1985, High pres~ure phase equilibria in ternary
fluid mixtures with a supercritical component, ACS Prep. Div. Fuel Chem., 30:46.
Paulaitis, M. E., Kander, R. G., and DiAndreth, I. R., 1984, Phase equilibria related to
supercritical fluid solvent extractions, Ber. Bunsenges. Phys. Chem., 88:869.
Picht, R. D., Dillman, T. R., and Burke, D. I., 1982, AlChE Symp. Ser., 78:136.
Raymer, J. H., and Pellizzari, E. D., 1987, Toxic organic compounds recoveries
from 2,6-diphenyl-p-phenylene oxide porous polymer using supercritical
carbon dioxide and thennal desotption methods, Anal. Chem., 59:1043.
Raymer, I. H., Pellizzari, E. D., and Cooper, S. D., 1987, Desotption charactersitics offour
polyimide sorbent materials using supercritical carbon dioxide and thennal
desotption methods, Anal. Chem., 59:2069.
Roop, R. K., and Akgennan, A., 1989, Entrainer Effect for Supercritical Extraction of
Phenol from Water, Ind. Eng. Chem. Res., 28:1542.
Roop, R. K., and Akgennan, A., 1990, Distribution of a complex phenolic mixture between
water and supercritical carbon dioxide, I. Chem. Eng. Data, 35:257.
168
Roop, R. K., Akgennan, A., Dexter, B. J., and Irvin, T. R, 1989, Extraction of phenol
from water with supercritical carbon dioxide, J. Supercrit. Fluids, 2:51.
Schantz, M. M., and Chesler, S. N., 1986, Supercritical fluid extraction procedure for the
removal of trace organics from soil samples, J. Chromatogr., 363:397.
Schultz, W. G., and Randall, J. M., 1970, liquid carbondioxide for selective aroma
extraction, Food Technol., 24:94.
Sherwood, T. K., Pigford, R L., and Wilke, C. R., 1983, "Mass Transfer", McGraw Hill,
New York.
Tan, C., and liou, D., 1988, Desorption of ethyl acetate from activated carbon by
supercritical carbon dioxide, Ind. Eng. Chern. Res., 27:988.
Vitzthum, O. G., and Hubert, P., 1975, u.S. Patent 3,879,569.
Wilhelm, A., and Hedden, K., 1983, Nonisothennal extraction of coal with solvents in
liquid and supercritical state, Proc. Int. Con/. Coal Sci., p. 6, Pittsburgh.
Wilson, R. E., Keith, P. C., and Haylett, R E., 1936, liquid propane use in dewaxing,
deasphalting, and rerming heavy oils, Ind. Eng. Chern., 28:1065.
Wright, B. W., Wright, C. W., Gale, R W., and Smith, R. D., 1987a, Analytical
supercritical fluid extraction of adsorbent materials, Anal. Chern., 59:38.
Wright, B. W., Frye, S. R, McMinn, D. G., and Smith, R. D., 1987b, On-line supercritical
fluid extraction-capillary gas chromatography, Anal. Chern., 59:640.
Yeo, S.-D., and Akgennan, A., 1990, Supercritical extraction of organic mixtures from
aqueous solutions, AIChE J., 36:1743.
Zosel, K., 1978, Separation with supercritical gases: practical applications, Angew. Chern.
Int. Ed., Engl., 17:702.
169
OZONE FOR WASTE REMEDIATION
AND WASTE WATER TREATMENT
Henry J. Ledon
Chemoxal, Air Liquide Group
75, Quai d'Orsay
75321 Paris Cedex 07, France
INTRODUCTION
Environmental concerns expressed by more and more
stringents regulations provide a strong incentive to design
new technologies for the destruction of hazardous wastes.
Ozone, which powerful oxidizing properties have been
recognized during the second half of the XIXth Century, is a
choice chemical reagent for the treatment of water.
The first large scale application of ozone was
implemented during the 1890's for the disinfection of
drinking water and is established nowadays as a standard
procedure in Western Europe where more than 850 ozonation
plants are in operation.
Improvement of the economics of large scale production of
ozone as well as the development of new methods to enhance
its chemical reactivity (usually referred as "Advanced
Oxidation Processes") offer a large scope of opportunities
for the chemical oxidation of water pollutants.
PHYSICAL AND CHEMICAL PROPERTIES OF OZONE

Ozone is a metastable allotropic form of oxygen which
formation is highly endothermic:
~HO =- 284.5 kJ
Ozone is slightly soluble in water. The value of the

Bunsen coefficient (STP m3 of gas dissolved per STP m3 of
water under a partial pressure of gas of 0,1 MPa) is 0.24 at
20°C compared to 0.031 for oxygen. 1
Industrial Environmel'llal CMmistry, Edited by D.T. Sawyer

The determination of ozone concentration in water
solutions could easely be carried out by using the Indigo
trisulfonate method 2 whereas the direct monitoring of ozone
in the gas phase is possible by UV absorption measurement:
molar absorptivity = 3 000 ± 30 l.M- 1cm- 1 at A = 254 nm 3
water solutions of ozone are moderately stable, the rate
of decomposition being first order with respect to ozone and
hydroxide ion 4.
Ozone is a very strong oxidizing agent as shown in
Table 1.
The reactivity of molecular ozone is governed by a strong
electrophilic character as illustrated by the influence of
substituents on the rate of ozonation of benzene derivatives,
as shown in Table 2.
However in an aqueous basic medium, ozone reacts with the
hydroxide ion in a very complex chain ,f.rocess giving the
hydroxyl radical HO· as an intermediate ' 7-12. As shown in
Table 1 above HO· is an even more powerful oxidant than ozone
and reacts non selectively at a rate close to the diffusion
limits. Comparison of the rate of oxidation of quite
refractory hydrocarbons with molecular ozone or the hydroxyl
radical are presented in the Table 3.
TABLE 1. Oxidation potentials of some usual oxidants.
OXIDANT OXIDATION POTENTIAL

(Volt I NHE)
FLUORINE 3.06
HYDROXYL RADICAL 2.80
OZONE 2.07
PERSULFATE 2.0
HYDROGEN PEROXIDE 1.78
HYDROPEROXYL RADICAL 1.70
PERMANGANATE 1.68
HYPOCHLOROUS ACID 1.49
CHLORINE 1. 36
CHLORINE DIOXYDE 1.27
OXYGEN 1.23
TABLE 2. Second order rate constants for the ozonation

of sUbstituted aromatics in water 5,6
COMPOUND
NITRO BENZENE 0.09 ± 0.02
CHLORO BENZENE 0.75 ± 0.2
BENZENE 2 ± 0.4
TOLUENE 14 ± 3
PARA XYLENE 140 ± 30
ANISOLE 290 ± 50
PHENOL 1 300 ±200
PHENATE ION (1.4±0.4) .10 9
172
TABLE 3. Second order rate constants for the oxidation
of saturated hydrocarbons.
COMPOUNDS k0 3 (M- l sec-l) 7 kHO· (M- l sec-l) 13

TETRACHLORO ETHYLENE < 0.1 1.7 .10 9
TERBUTANOL 0.03 4 .10 8
OXALATE ION < 0.04 1 .10 7
ACETATE ION 3.10- 5 7 .10 7
OZONE GENERATION
Ozone is produced on an industrial scale by passing a dry

gas containing oxygen in a corona discharge.
Air or oxygen can be used as feed stock. The energy
consumption for ozone production is highly dependant on
purity of the feed gas, in particular a very low water
containt is essential (dew point below -50 D C to -80 D C).
Several contaminents like hydrogen, hydrocarbons, fluorinated
hydrocarbons, . . .. have a very detrimental effect on ozone
yield, even at minute concentration. Thus if air is used as
feed gas, a very careful purification is required, whereas
oxygen can be used as delivered.
The theoretical energy required to produce 1kg of ozone
is 0,82 kwh, but in practice, using dry air as feed gas, the
actual consumption for producing ozone (1 to 3% in air) is in
the range of 15 to 20 kwh/kg. Use of oxygen results in a much
better energy efficiency and the most advanced technology
only requires 7 to 10 kwh/kg for a much higher output
concentration (6 to 10% in oxygen). 14
TO PROCESS
DRY.A.lR
I
02
,r
I OZONE
GENERATOR
Il ~
I
I 02 RECYCLE 02/03
I
I
I 02103 SEPARATION UNIT
II I I
I
I
I I
031 N2
I
I
AIR N2 I
Figure 1. Simplified flow-sheet for ozone production with oxygen
recycling.
173
Incentives to benefit from the numerous advantages of
using oxygen as feed gas for generating ozone have led to the
development of a new technology for separation of ozone from
the outlet mixture gas and the recycling of oxygen to the
ozone generator. Figure 1 shows a schematic flowsheet of the
L'AIR LIQUIDE - DEGREMONT loop-process to produce ozone. 15
OZONATXON OF HAZARDOUS ORGANXC COMPOUNDS
OZONE ACTXVATXON
Molecular ozone reacts extremely rapidly wi th electron
rich unsaturated compounds (Table 2) but only slowly with
saturated derivatives (Table 3). In order to extend the scope
of chemical oxidation which could be used in water treatment,
several authors have examined the combination of ozone whith
UV-Irradiation or ozone with hydrogen peroxyde. 16-30
Ozone dissolved in water can be photodecomposed by
irradiation at A = 254 nm with a quantum yield of 0,62 31
hv
0 3 + H20 ---> 02 + H20 2
Hydrogen peroxide is a weak acid (pKa 11,6) which

reacts slowly with ozone whereas its conjugated base H02-
leads to a fast decomposition of ozone:
---> °.
3
The ozonide radical

hydroxyl radical HO·
ion °3 • - decomposes to afford the
whereas the hydroperoxide radical H0 2 • induce the chain

decomposition of ozone:
pKa 4.8
Numerous other free radical pathways have been described

°
for the reaction between species present in solution 0 3 , HO-,
HO· H0 2 ·, 2 • -, HOil-' H202,'" which leads to an extremely
complicated mechan~stic scheme. The main feature of the
activation of ozone either by UV irradiation or hydrogen
peroxide is the rather fast production of the hydroxyl
radical HO·.
The effect of the pH and the addition of hydrogen
peroxide on oxidation of organic compounds is illustrated in
Figure 2 which exhibits the ozonation of methionine monitored
by the chemical oxidation demand "COD" with time.
174
100
80
R
'!b E
M 60
co
OV
OA 40
L
20
0
0 10 20 30 40 50 60
TIME (mn)
Figure 2. oxidation of an aqueous solution of methionine (250mg in 250ml;

03 25mg/min; H202 240mg)
ozonation at low pH gives only a limited removal of COD

whereas ozonation at basic pH lead to almost total oxidation.
Higher rates of oxidation are observed in the presence of
hydrogen peroxide.
INHIBITORS
However, the rate of oxidation of the organic compounds
could be dramatically reduced if inhibitors are present in
the reaction medium. The most common ones in water treatment
are the carbonate/bicarbonate ions which are efficient
scavengers for the hydroxyl radical. 22,32
During the course of the reaction acids are formed which

lower the pH and thus reduce the rate of oxidation by
dep"leting the medium from HO- and H0 2- ions. In order to
obtain a high rate of destruction, controlled addition of a
base is desirable to maintain the pH value in the basic
range. However, when the total destruction of hazardous
organics is required, oxidation in basic medium leads to the
build up of carbonate/bicarbonate ions which inhibit the
reaction.
Thus it may be advisable for full oxidation of chemicals
to have a process including a first stage of ozonation in
neutral or basic medium followed by a second stage which
could be carried out at low pH like Fenton's type reaction (
H20 2 in the presence of ion salts) or UV irradiation of
hydrogen peroxide.
175
OZONE - HYDROGEN PEROXIDE RATIO
The theoretical molar ratio of oxidants in ozone-hydrogen

peroxide system is H2 0 2 :0 3 = 0.5 21. If hydrogen peroxide is
present at high concentration it could react with the HO·
radicals inducing self decomposition reactions without
oxidation of the organic compounds.
A much more efficient use of chemicals is obtained with
continuous addition of H20 2 as shown in Table 4 for the
oxidation of methionine.
TABLE 4 Effect of the mode of H202 addition for the

oxidation of methionine in aqueous solution (lg/l)
SINGLE ADDITION 1 STEPWISE ADDITION 2
REACTION TIME FOR

85% COD REMOVAL (MN) 60 36
03 INTRODUCED (mg) 1 500 900
03 CON SUMMED (mg) 615 615
H2 0 2 INTRODUCED (mg) 240 155
H2 0 2 CON SUMMED (mg) 240 110
1
The total amount of H202 is introduced at t=O
2 35 mg of H202 are introduced at t=O then 40 mg each 10
minutes.
EVOLUTION OF THE TOXICITY OF SOLUTIONS DURING THE OXIDATION
The removal of toxic compounds by total oxidation to

carbon dioxide and water is an expensive process and
frequently only partial reactions are carried out.
However care must be taken that the intermediate products
formed during the oxidation are safe for the environment
Figure 3 shows· the evolution of the toxicity of a 6.7 .10- 3
molar solution of methionine in water with respect to the COD
removal using the 0 3 / H20 2 system.
Toxicity of the solution was estimated following the
concentration which reduce by 50% (EC 50) the light emission
of photobacteria using the MICROTOX® equipment.
Methionine which is an innocuous compound used as
additive for animal nutrients affords toxic intermediates by
oxidation which are themself destroyed by further reaction.
176
In conclusion ozone alone or activated by UV Irradiation
or by association with hydrogen peroxide is a powerful and
clean oxidizing agent as well as a strong desinfectant.
However its uses at the present time are mainly limited to
the treatment of drinking water.
Over the last decade extensive research has led to a much
better understanding of the parameters which control these
very complexes Advanced oxidation Processes. Technical
efficiency has been demonstrated in several pilot scale or
field applications. Reductions of the cost of ozone
production and optimization of the consumption or reagents
will open new area of economically viable processes for the
treatment of effluents.
11001 EC501
10
9
8
T
0
7
x 6
C 5
I
T 4
y
3
2
1230 1012 672 219 90

COD
Figure 3. Evolution of the toxicity of an aqueous solution of
methionine during its oxidation with the 03/H202 system.
ACKNOWLEDGEMENTS
The Author is grateful to "Ministere de l'Industrie et de
l'Amenagement du Territoire" for support through Cooperative
Agreement 90-4-90-124: Technologies Industrielles pour
l'Environnement, to A. DE LADONCHAMPS, D. GOBARD and C.
MESSUTA for technical assistance and to C. COSTE of DEGREMONT
for sharing his expertise in ozone technology through
stimulating discussions.
REFERENCES
General readings
P. S. Bailey. "Ozonation in Organic Chemistry", Academic Press, New York

(1982)
177
R.G. Rice and A. Netzer. "Handbook of Ozone Technology and
Applications". Ann Arbor Science Publishers, Ann Arbor (1982)
S.D. Razumouskii and G.E. Zaikov. "Ozone and its Reactions with organic
Compounds", Elsevier, Amsterdam (1984)
M. Dore "Chimie des Oxydants et Traitement des Eaux", Lavoisier Tech &
Doc, Paris (1989)
H. Roques "Fondements Thi!oriques du Traitement Chimique des Eaux",

Lavoisier Tech & Doc, Paris (1990)
W. J. Masschelein "Ozone et ozonation des Eaux", I .O.A-Lavoisier Tech &

Doc, Paris (1991)
B. Langlais, D.A. Reckhow and D.R. Brink. "Ozone in Water Treatment:

Application and Engineering" Lewis Publishers, Chelsea (1991)
W.W. Eckenfelder, A.R. Bowers and J.A. Roth. "Chemical Oxidations,

Technologies for the Nineties". Technocomic, Lancaster (1992)
1. L'AIR LIQUIDE "Gaz Encyclopedia". Elsevier, Amsterdam, (1976)
2. H. Bader and J. Hoigne, Determination of Ozone in Water by the Indigo

Method: A Submitted Standard Method, Ozone. Sci. Eng. 4:169 (1982)
3. International Ozone Association Standardisation Committee Europe,

002/87
4. J. Staehelin and J. Hoigne, Decomposition of Ozone in Water. Rate of

Initiation by Hydroxide Ions and Hydrogen Peroxide, Envir. Sci.
Technol. 16:676 (1982)
5. J. Hoigne and H. Bader, Rate constants of Reactions of Ozone with

Organic and Inorganic Compounds in Water. I. Non Dissociating organic
compounds, Water Res. 17:2:173 (1983)
6. J. Hoigne and H. Bader, Rate constants of Reactions of Ozone with

Organic and Inorganic compounds in water. II. Dissociating organic
Compounds, water Res. 17:2:185 (1983)
7. J. Hoigne and H. Bader, The Role of Hydroxyl Radical Reactions in

Ozonation Processes in Aqueous Solutions, Water Res. 10:377 (1976)
8. L. Forni, D. Bahnemann and E.J. Hart, Mechanism of the hydroxide Ion

Initiated Decomposition of Ozone in Aqueous Solution, J. Phys. Chem.
86:255 (1982)
9. R.E. Buhler, J. Staehelin and J. Hoine, Ozone Decomposition in Water

Standard by Pulse Radiolysis.1. H02/02 and H03/03 as Intermediates,
.J. Phys. Chem. 88:2560 (1984)
10. H. Tomiyasu, H. Fukutomi and G. Gordon, Kinetics and Mechanisms of

Ozone Decomposition in Basic Aqueous Solutions, Inorg. Chem. 24:2962
(1985)
11. G. Gordon, S. Nakareseisoon and G.E. Pacey, The Very Slow

Decomposition of Aqueous Ozone in Highly Basic Solutions, Proceedings
of the 8th Ozone World Congress. International Ozone Association,
Zurich, E27 (1987)
178
12. C.C.D. Yao, W.R. Haag and T. Mill, Kinetics features of Advanced
oxidation Process for Treating Aqueous Chemical Mixtures, Proceedings
of the Second International Symposium on Chemical Oxidation,
Technology for the Nineties, Vanderbilt University, Nashville,
Tennessee (1992)
13. Farhataziz and A.B. Ross, selected specific Rates of Reactions of

transients from Water in Aqueous Solutions, National Bureau of
Standards NSRDS-NBS 59, Washington, DC (1977)
14. P.E. Erni, Advanced in Ozone Generator Technology, Proceedings of the

10th Ozone World Congress, International Ozone Association, Zurich,
1:13 (1991)
15. P. Petit, P. Gastinne, P. Jan, J.M. Benas, R.S. Bes and J.C. Mora,
Ozone Production with recycling of Non Transformed Oxygen,
Proceedings of the 8th Ozone World congress, International Ozone
Association, Zurich, A72 (1987)
16. R.L. Garrison, H.W. Prengle, Jr and C.E. Mauk, Method of Destroying
Cyanides US Patent 3,920,547 to Houston Research Inc, (1975)
17. S. Nakayama, Y. Taniguchi, K. Namba and N. Tabata, Improved Ozonation

in Aqueous Systems, Ozone Sci. Eng. 1:119 (1979)
18. K. Namba and S. Nakayama, Hydrogen peroxide-Catalyzed Ozonation of

Refractory organics. I. Hydroxyl Radical Formation, Bull. Chem. Soc.
Jpn. 55:3339 (1982)
19. G.R. Peyton, F.Y. Huang, J.L. Burleson and W.H. Glaze, Destruction of
Pollutants in Water with Ozone in Combination with Ultra violet
Radiations 1. General Principles and Oxidation of Tetrachloroethylene
Environ. Sci. Technol. 16:448 (1982); ibid, 2. Natural Trihalomethane
Procursor, Environ. Sci. Technol. 16:454 (1982)
20. G.R. Peyton Modeling Advanced Oxidation Processes for Water Treatment
Proceedings of the 9th Ozone World Congress, International Ozone
Association Zurich, 579 (1989)
21. W.H. Glaze, J.W. Kang and D.H. Chapin, The Chemistry of Water
Treatment Processes Involving Ozone, Hydrogen Peroxide and Ultra
violet Radiation, Ozone Sci. Eng. 9:335 (1987)
22. W.H. Glaze and J.W. Kang, Evaluation of the Ozone-Hydrogen Peroxide
Process in a Semi-Batch Reactor using Tetrachloro ethylene as a model
compound, proceedings of the 9th Ozone World Congress, International
Ozone Association Zurich, 596 (1989)
23. R. Brunet, M.M. Bourbigot and M. Dore, Oxidation of Organic Compounds

through the Combination or Ozone-Hydrogen Peroxide, Ozone Sci. Eng.
6:163 (1984)
24. H. Paillard, R. Brunet and M. Dore, Application of Oxidation by a

combined ozone/Ultraviolet Radiation System to the Treatment of
Natural Water, Ozone Sci. Eng. 9:391 (1987)
25. H. Paillard, R. Brunet and M. Dore, Optimal Conditions for applying

an Ozone - Hydrogen Peroxide oxidizing System, Wat. Res. 2:91 (1988)
26. J.P. Duguet, E. Brodard, B. Dussert, J. Mallevialle, Improvement in

the Effectiveness of Ozonation of Drinking Water Through the Use of
Hydrogen Peroxide, Ozone sci. Eng. 7:241 (1985)
179
27. P.D. Francis, Ozonation by UV and Ozone of Organic contaminants
Dissol ved in Deionized and Raw Mains Water, Ozone Sci. Eng. 9: 369
(1987)
28. J.L. Wallace, B. Vahadi, J.B. Fernandes and W.H. Glaze, Destruction
of Pollutants in Water with Ozone in Combination with UltraViolet
Radiation 1. General Principles and Oxidation of Tetrachloro ethylene
Environ. Sci. Technol. 16:448 (1982); ibid, 2. Natural
Trihalomethane Precursors, Environ. Sci. Technol. 16:454 (1982)
29. N. Nakanishi, Ozonation of several Organic Compounds Having Low

Molecular Weight under Ultra Violet Irriadiation, Ozone Sci. Eng. 2:1
(1990)
30. J.D. Zeff and E. Leitis, Oxidation of Organic Compounds in water, US

Patent 4,792,407 to Ultrox International (1988)
31. H. Taube, Photochemical Reactions of Ozone in Solution Tran. Faraday

Soc. 53:656 (1956)
32. H. Pai11ard, B. Legube and M. Dore, Effects of Alkalinity on the

Reactivity of Ozone towards Humic Subtances and Manganese, Aqua
38:32 (1989)
180
NUCLEOPHILIC AND REDUCTIVE REMEDIATION STRATEGIES FOR
HAZARDOUS HALOGENATED HYDROCARBONS (PCBs, HCB, PCP, TCE,
CCl4- DCBP, EDB, DDT, AND DDE)
Donald T. Sawyer, Paul K. S. Tsang, Seungwon Jeon

and· Marjorie Nicholson
INTRODUCTION
During the past half-century the world has made use of a large group of
halogenated hydrocarbons. The applications include structural polymers [e.g.,
polyvinylchloride (PVC)], pesticides (e.g., DDT, DCBP, EDB, PCP, DOE, and
Lindane), cleaning solvents and degreasing agents (e.g., CCl 4 , TCE,
methylchloroform, and DCE), and non-flammable pump oils, dielectrics, and
heat-exchanger fluids (e.g., PCBs, HCB, and related polychlorinated aromatic
molecules). In the manufacture and utilization of these materials halogenated
-wastes result, -which have been stored, "dumped and flushed", placed in land-
fills, and incinerated. Some of the disposal methods have produced
contaminated soils, sludges, harbor muds, and land-fills. Most of the Super-
Fund sites contain halogenated-hydrocarbon wastes as a major component and
are the most extreme examples of poor "housekeeping" and inappropriate
disposal.
Polyhalogenated alOmatic hydrocarbons (PCBs, PBBs, and hexachloro-
benzene) represent a major environmental problem.1,2 These materials,
which were extensively used as transformer oils and heat-exchanger fluids, are
major components of the hazardous-waste disposal problem of the EPA
Superfund. The chemistry and toxicology of PCBs, and of their impurities and
partially oxygenated products (e.g., dioxins), have been extensively reviewed.3-s
Clearly the materials contain components that are animal carcinogens and can
cause birth defects. 6- 8 Hexachlorobenzene (HCB), which is a by-product of the

polychloroethylene solvent industry, is as environmentally persistent as PCBs
and is a human carcinogen.9 ,lO
Although storage and disposal in landfills have been used for PCBs in
the past, their long environmental life has lead to the contamination of lakes,
rivers, coastal estuaries, and ground water. Currently, incineration is the most
utilized technology for the destruction of halogenated aromatic hydrocarbons.
The best systems are highly efficient in their conversion of PCBs to HCI, C02,
and H20, but appear to produce some dioxins. l1
In addition to the problems of the past, useful materials continue to be
produced from the halogenation of hydrocarbons and waste streams from the
manufacturing processes are inevitable. The ideal remediation strategy for
these wastes should be on-site, on-line processing, which would eliminate the
problems and hazards of storage and transport.
Remediation strategies (chemical and biological) for halogenated
hydrocarbons seek to destroy all of the carbon-halogen bonds (e.g., RX
HO-
----~~- ROH + X-) via formation of inorganic halogen. These can be
subdivided into (a) oxidative and (b) reductive processes. Within the realm of
oxidative chemistry incineration is the most extensively developed technology.
The complete mineralization of heavily chlorinated PCBs (Aroclor 1268,
Cl7-10), dioxins, and PhCl 6 (HCB) requires high temperatures via oxygen-rich
combustion. With incineration the production of oxygenated intermediates
(dioxins and furans) from PCBs and other halogenated aromatics is a major
concern. Also, the volatile products from oxidative combustion and pyrolysis
of PCBs include HC!, Cl2 , HOCI, Cl 2 0, and CI 2C=0, which must be efficiently
removed.
The major chemical oxidants within an oxygen-rich flame are ·0·, HO·,
HOO·, and .02., and their primary chemistry is (a) the abstraction of hydrogen
atoms from substrates (oxidation) and (b) formation of bonds with the resultant
carbon radicals (oxygenation)
r HO·
H3C ----tl_- H 3COH (1)
H 20
Similar oxidants can be produced from 02,03' HOOH, and HN0 3 via (a)
catalytic chemical reactors and aerobic biology. Such oxidants are much less
reactive with carbon-chlorine bonds than with carbon-hydrogen bonds, which
makes oxidative strategies ineffective with fully chlorinated wastes [e.g., CCl 4 ,
PhCl6, and C 12Cl lO (perchloro biphenyl)].
182
In contrast, reductive technologies for halogenated hydrocarbons
invariably transform carbon-bonded halogens to halide ions. This is
accomplished with (a) nucleophiles (e.g, HO-, RO-, O 2-,, H-, and e-),
Mel + HO- -------'1._ MeOH + 1- (2)
(b) hydrogen atoms,
PhCI + 2 [H·] + HO- - - -....- PhH + CI- + H 20 (3)
and (c) strongly reducing metals via electron-transfer, e.g.,
PhCI + 2 Na + H 20 - - -....- PhH + NaCl + NaOH (4)
In the following sections we discuss the current status of various

nucleophilic and reductive strategies for the transformation of hazardous
halocarbon wastes to harmless materials. Most of the studies have used
hexachlorobenzene (PhCI6) and chlorobenzene (PhCl) as models for PCBs, and
CCl 4 and trichloroethene (TCE) as models for cleaning-solvent and de greasing-
agent wastes.
NUCLEOPHILIC DEHALOGENATION OF HALOCARBONS
Most chemical transformations occur via the reaction of a nucleophile

(base, reductant, or electron) with an electrophile (acid or oxidant), which
entails a single-electron-transfer (S.E.T.) mechanism 12
S.B.T.
Nu:- + RX - - - -....- R-Nu + x- (5)
Because the solvent matrix affects the redox potentials of nucleophiles and
electrophiles, it has a major influence on their reactivity. Hence, strong
nucleophiles are "leveled" to a reactivity that is equal to that for the conjugate
base of the solvent. For example,
Nu:- + n H 20 • Nu-H + HO-(H 20)n_l (6)

(Nu:-; H:-, RO-, O 2-,, HOO-)
Although hydroxide ion (HO-) is a highly reactive nucleophile in aprotic

solvents (Me2S0, MeCN, DMF, and ethers), it is ineffective with chlorinated
aromatics. In the primary step, the chloro-aromatic is transformed to a phenol,
183
which in turn is neutralized with a second HO- to give a phenolate anion (a
weak nucleophile) that is inert to further reaction
HO-
- -.....\~..~ CISPhOH --..;..;....;;.......\~...
~ ClsPhQ- (7)
cr H 20
The latter problem is eliminated by the use of alkoxide ions (RO-), which yield
an alkyl ether that remains electrophilic
PhCl6 + RO- - - - - 1 1...

- CISPhOR + Cl- (8)
KPEG Dehalogenation
The utility of alkoxide nucleophiles for the dehalogenation of PCBs via

nucleophilic displacement has been known since the 1980s. In order to avoid
the leveling effects of water and to have an aprotic matrix with a low vapor
pressure that can withstand high reaction temperatures, Prof. L. L. Pytlewski
and co-workers 13 made use of poly(ethyleneglycol) (HPEG) and
poly(ethyleneglycol)-monomethylether (HPEGMe) as solvents. They
discovered that their conjugate-base salts (KPEG, NaPEG, KPEGMe, and
NaPEGMe) are able to dechlorinate PCBs. Subsequent work at General Electric
developed this technology for the remediation of PCB-contamined transformer
oils.l4
More recently (1991) the KPEGMe/HPEGMe technology has been
successfully applied to PCB-contaminated soil at the Superfund site in Wide
Beach, New York. The next paper by Robert Boch discusses the engineering
and technical problems, and the associated economics. Scheme I outlines paths
(a, b) for the nucleophilic remediation of PCBs by the KPEG/HPEG reagent, and
indicates that in its absence KOB and/or carbonates will produce chlorinated
furans and dioxins (paths d-g).
Superoxide Ion
Although superoxide ion is a powerful nucleophile in aprotic solvents,

it does not exhibit such reactivity in water, presumably because of its strong
solvation by that medium (.1Hhydration, 100 kcal) and its rapid hydrolysis and
disproportionation. The reactivity of 02-' with alkyl halides via nucleophilic
substitution was first' reported in 1970. 1S,16 These and subsequent kinetic
studies 17- 19 confirm that the reaction is first order in substrate, that the rates
follow the order primary>secondary»tertiary for alkyl halides and tosylates,
184
Scheme I. Dehalogenation of PCBs by KPEG.
(a)~.KO
(g)
carbonates
t:.
2
H 0,
CO2
1
(d) 2 KOH
(PEG)n-
!
(b) 2 n-l KPEG
~(PEG)n + 2n-IKO
CI,. 0n_l ~ ~
~+K'+Ka
(f)1 ~~o
~~-' + cr ~~?~: + cr
and that the attack by 02-· results in inversion of configuration (SN2). The
stoichiometries and kinetics for the reaction of 02-· with several halogenated
hydrocarbons (alkanes, alkenes, and aromatics) are summarized in Table 1. 20-22
[The normalized first-order rate constants, kl / [S], were determined by the
rotated ring-disk electrode method under pseudo-first-order conditions
([substrate] > [~-.)].20
The nucleophilicity of 02-· toward primary alkyl halides results in an
SN2 displacement of halide ion from the carbon center (Scheme 11). The
normal reactivity order, benzyl>primary>secondary>tertiary, and leaving-
group order, I>Br>OTs>CI, are observed, as are the expected stereoselectivity
and inversion at the carbon center. In dimethylformamide the final product is
the dialkyl peroxide. The peroxy radical (ROO·), which is produced in the
primary step and has been detected by spin trapping,23 is an oxidant that is
readily reduced by 02-· to form the peroxy anion (ROO-). Because the latter can
oxygenate Me2S0 to its sulphone, the main product in this solvent is the
alcohol (ROH) rather than the dialkyl peroxide.
Although formation of the dialkyl peroxide is shown in the prototype
reaction (Scheme II), hydroperoxides, alcohols, aldehydes, and acids also have
been isolated. The extent of these secondary paths depends on the choice of
solvent and reaction conditions. Secondary and tertiary halides also give
substantial quantities of alkenes from dehydrohalogenation by O 2-. /HOO-/HO-.
185
Table I. Stoichiometries and Kinetics for the Reaction of 0.1-5 mM 02-' with
Polyhalogenated Hydrocarbons in Dimethylformamide at 25°C.
Substrate
Concentration, kl/[S]
[S] = 1-10 mM 02-'/S Products/S M- 1s-l
CC4 5 HOC(O)O-, 4 CI-, 3.3 02 3800.

HCCl3 4 HOC(O)O-, 3 Cl-, 2 02 0.4
PhCCl3 4 PhC(0)0-(70%), PhC(0)00-(30%), 50.0

3 Cl-, 2.4 02
MeCCl3 <0.1
HOCH 2CCl3 4 HOCH2C(0)0-, 3 Cl- 47.
(p-CIPh)2CHCCI3(DDT) 1 (p-ClPhhC=CCb, Cl- 100.
PhCHBrCHBrPh 2 2 PhCH(O), 2 Br, 02 1000.
n-BuBr 1 Br- 960.
cis-CHCl=CHCl 4 2 HOC(O)O-, 2CI- 10.0
(p-ClPhhC=CCI2(DDE) 3 HOC(O)O-, (p-CIPhhC=O, 2 Cl-, 02 2.0
The reaction of 02-' with CCl4 and RCCl3 compounds cannot occur via
an SN2 mechanism because the carbon-atom center is inaccessible. Rather,
superoxide ion appears to attack a chlorihe atom with a net result that is
equivalent to an electron transfer from 02-' to chlorine. This step is analogous
to the single-electron-tra:nsfer (SET) mechanism that has been proposed for
many nucleophilic reactions; an initial transfer of an electron followed by the
collapse of a radical pair (eq. 5).24
The rates of reaction for O 2-, with RCCl3 compounds are proportional to
their reduction potentials, which is consistent with the SET mechanism. 20 A
plot of log kIf [S] (Table II) against the reduction potentials of RCCl3 compounds
is approximately linear with a slope of -4.9 decade per volt. Such behavior is
consistent with a mechanism that occurs via simultaneous electron transfer
and nuclear motion. 25 This correlation indicates that in water, where the
02/02-, redox potential is about 0.44 V more positive, the rate of the reaction
for 02-' with CCl4 would be about 100-200 times slower than in aprotic solvents.
186
Scheme II. Nucleophilic Substitution by O 2-' of Halocarbons.
,4_-C-'-H_D__ NR
l--_l....
- -......- ROO· + )(
t:.
IL._.::..:R:.::;O:...:::O;...·_-I..._ [ROOOOR]
l,4-CHD
kD > 108 M-1s-1
ROOR+02
k diss,l 00-1 03 s-1
ROaR + PhH + HOOH

k 101 03 M-1 -1
dehy' -1 s
DPIBF
'----=;;.:..=~--!I- ROaR + DBB + 1/2 O 2
k diox' 104-106 M -1-1
s
The stoichiometric data in Table I for CCl4 are consistent with a l1et
chemical reaction that yields bicarbonate ion and four halide ions in the final
aqueous workup of the reaction products. When 02-' reacts with RCCl3
compounds, the R-C bond is not cleaved. About 30% of the product from the
PhCCl3 /02-' reaction is perbenzoate ion.
Superoxide ion reacts with vicinal dibromoalkanes to form aldehydes
(Table I). The mechanism proposed for these reactions is a nucleophilic attack
on carbon, followed by a one-electron reduction of the peroxy radical and
nucleophilic displacement on the adjacent carbon to form a dioxetane that
subsequently cleaves to form two moles of aldehyde. 22
Both p,p'-DDT and methoxychlor are rapidly deprotonated by HO- in
aprotic solvents with subsequent elimination of Cl- to form the dehydro-
chlorination products; that is, DDT forms DDE. The same products are formed
in their reactions with 02-'. Because the reaction rates that are measured by the
rotated ring-disk electrode method are fairly rapid, the primary step must be a
187
direct reaction with 02-· and not with HO- [from the reaction of 02-· with trace
water in the solvent (2 02-· + H20 ~ 02 + HOO- + HO-)]. Hence, the initial
reaction with 02-· is deprotonation followed by elimination of Cl- to form DOE.
The overall reaction and product stoichiometries for the degradation of
chloroalkene substrates by 02-· in DMF are summarized in Table 1.21 Within
the limits of a reaction time of 10 min or less, chloroethene, trans-1,2-
dichloroethene, Aldrin, and Dieldrin are not oxidized by 02-· in DMF. A
reasonable mechanism for these oxidations is an initial nucleophilic addition
of superoxide to the chloroalkenes. Subsequent loss of chloride ion would give
a vinyl peroxy radical, which can cyclize and decompose to a chloroacyl radical
and phosgene. 26 These would undergo subsequent facile reactions with 02-· to
give the observed bicarbonate and chloride ions.
In the case of (p-CIPhhC=CCI2 (DOE) and 1,1-dichloroethene, addition of
superoxide can only be followed by 13-elimination of chloride if attack occurs at
the carbon with the chlorine atoms.
R
):=C °
/ Cl
+
-
2. ---
~R
:C-C-Cl --- 'j=(
I
- I Cl ]
Cl R
+cr (9)
R '\:1 I
R
I
00· R '00.
The sequence would lead to the ketones [RC(O)R] that are observed as products
(Table I).
Polyhalogenated aromatic hydrocarbons [e.g., hexachlorobenzene (HCB,
C6C16) and polychorobiphenyls (PCBs)] are rapidly degraded by superoxide ion
in dimethylformamide to bicarbonate and halide ions (Table II).2 7 Because
halogen-bearing intermediates are not detected, the initial nucleophilic attack
is the rate-determining step. The rates of reaction exhibit a direct correlation
with the electrophilicity of the substrate (reduction potential) (e.g., C 6Cl 6, £0' =
-1.48 V versus SCE; kl/[S] = 1 x 103 M-I s-1 and 1,2,4-C6H3Cl3, £0'= -2.16 V; kl/[S]
= 2 x 10-2 M-I s-1).27
Although polyhaloaromatics are degraded by 02-· in acetonitrile and
dimethyl sulfoxide, the rates of reaction are about one-tenth as great in MeCN
and 20 times slower in Me2S0. A reasonable initial step for these oxygenations
is nucleophilic addition of 02-· to the polyhalobenzene (e.g., C6Cl6; Scheme III).
Subsequent loss of chloride ion will give a benzoperoxy radical, which will
close on an adjacent aromatic carbon center and add a second 02-· to become a
peroxo nuc1eophile that can attack the adjacent carbo chlorine center with
displacement of chloride ion to give a highly electrophilic tetrachloro center.
188
Table II. Reactions of Superoxide Ion with Polychloro-aromatics in Dimethylformamide.a
Ep,c, b
Substrates(S) Vvs SCE O2--15 Cl- released/S base released/S k1 /[Sl, M-l s-lc
C6Cl6d -1.48, -1.69 12.O±1.0 6.0±0.5 6.0±0.6 1 x 103

C6HCl5 -1.70, -1.98 l1.O±l.0 5.0±0.5 6.0±0.6 8x 101
1,2,3,4-C6H 2CI4 -1.90,-2.18 1O.0±1.0 4.0±0.5 6.0±1.0 2x100
1,2,3,5-C6H 2CI4 -1.95, -2.19 1O.O±1.0 4.0±0.5 6.O±1.0 1 x 100
l,2,4,5-C6H 2CI4 -1.95,-2.19 10.0±1.0 4.0±0.5 6.O±1.0 3x 100
l,2,4-C6H 3Cl3 -2.16, -2.45 2 x 10-2
C12Cl 1O -1.50,-1.78 22.0±4.0 10.O±1.0 12.0±2.0e 2x 102
cCII -1.2 5.0±O.6 4.0±O.4 1.0±0.1 1 x 103
PhCCI:/ -1.47 4.0±0.4 3.0±O.4 4x 101
aOverall reactions: (1) C6CI6 + 12°2-' -73 C2062- + 60- + 3 Oz; (2) C6H204 + 10°2-' -7 2 C2062- + 2 HOC(O)O- + 4 0- + Oz; (3) C12010 + 22 02-'
-76 C2062- + 10 Ct + 4 Oz.
bfjrst two reduction peaks are listed; a separate peak is observed for each chlorine atom. For C606 the other peak potentials are -1.95 V, -2.18 V, -2.44 V, and
-2.70 V (also the reduction potential for PhCI).
CApparent pseudo-first-order rate constants, k(norrnalized to unit substrate concentration [5]), were determined from the ratio (ianodic/icathodic) for the cyclic
voltammogram of 02 in the presence of excess substrate.
dThe values of kl / [S] for C6Ci6 in MeCN and Me2SO solvents are 92 M-1s-l and 47 M-ls-l, respectively.
eritration curve indicates the presence of a weak base, and a white precipitate forms during the last part of the titration; both consistent with the presence of
oxalate ion.
fIne ring-<iisc voltammetric technique for reaction rates gave values for kl/[S] of 3.8 x 103 M-1s-l for CCLl and 50 M-1s-l for PhCCI3'
03
CD
Scheme III. Nucleophilic degradation of hexachlorobenzene (HCB) by O 2-.
ClQ(Cl
Cl
Cl~ I
1 .
Cl
cil Cl
Cla eXT
b
Cl '0"'-
~a
~g
60""
- - -....- 6 CO2 + 6 cr
I 602-.
'---.::....-...·-3 c 0t + 3 O
2 2
I 3 HzO. 6 HOC(O)O- + 1.5 O 2
The latter undergoes facile reactions with 02-· to displace the remaining chloro
atoms. Thus, Scheme III outlines a possible mechanism, but the fragmentation
steps are speculative and not supported by the detection of any intermediate
species. 27
When Aroclor 1268 (a commercial PCB fraction that contains a mixture
of C17, Cis, C19, and ClIO polychlorobiphenyls) is combined with excess 02-·, the
entire mixture is degraded (Figure 1). Samples taken during the course of the
reaction confirm that (a) the most heavily chlorinated members react first (the
initial nucleophilic addition is the rate-determining step) and (b) all
components are completely dehalogenated. Tests with other PCB mixtures
establish that those components with three or more chlorine atoms per phenyl
ring are completed degradE:d by 02-· within several hours.
190
Cig
A. Arochlor 1268
Cis
(ij CI.
c:
Cl
W
... CI •
o
13
Q)
a;
o Cis
CI.
c:
o
~
N
·c
o
I
Q)
E
CIl
u:::
~ C. Solution B + H20 to dissolve sOlids
Jl.---Jv~. ..
o 7.0 9.0 11.0 13.0
GC Retention time, min
Figure 1. Gas chromatograms for Aroclor 1268 PCB mixture (C12C10, C12HCI9, C12H2ClS): (A)
10.2 mg of Aroclor 1268 in 1 mL of DMF; (B) solution A, 12 hr after the addition of 47.2 mg of
(Me4N)02 (-20 equiv per PCB); (C) Solution B after the addition of 1 mL H20 to dissolve any
organic solids.
191
Although superoxide is available as K0 2, the material is expensive and
almost insoluble in most aprotic solvents. The tetramethylammonium salt
[(Me4N)02] is much more soluble, but must be prepared from K0 2 and
(Me4N)OH(s) via a solid-phase metathesis reaction. 28 ,29 A recent report30
describes the in situ generation of O 2-. from the combination of aniline or
hydrazine, O 2, and HO- in a pyridine solvent. However, the most efficient and
convenient means of production is in situ reductive electrolysis of dioxygen.
The cyclic voltammograms for the reduction of 02 to 02-· in the absence and
presence of C6Cl6 are illustrated in Figure 2. The enhancement in the cathodic
peak current and the decrease in the peak current for the reverse scan that
result from the presence of C6Cl6 indicate a facile multistep reaction between
02-· and C6Cl6. The destructive decay of the components of Aroclor 1268 that
results from the electrolytic reduction of oxygen to 02-· in an air-saturated DMF
solution is presented in Figure 3. This illustrates a system whereby heavily
chlorinated PCBs can be degraded to inorganic ions in a continuous electrolytic
reactor. Product assays of numerous related experiments confirm that the
initial step is rate limiting, and that organic intermediates do not accumulate
in the reactor.
Relative Reactivity of Oxo Nuc1eophiles [HOO-, t-BuOO-, O 2 -., MeO-, HO-,

MePEG-, PEG-]
The nuc1eophilicity of oxy anions (YO-) is directly related to their

oxidation potentials (EO'B- lB., Table III) and the bond energies of their products
(YO-R) with electrophilic substrates (RX)
YO- + RX ~ YO-R + X- (10)
Hence, the more negative the oxidation potential and the larger the YO-R bond
energy, the proportionally greater nucleophilic reactivity that will result. The
shift in the oxidation potential of HO- from +1.89 V versus NHE in H20 to
+0.92 V in MeCN reflects the leveling effect of protic solvents on the
nuc1eophilicity of oxy anions. Likewise, the shift in potential for HOO- [+0.20
V (H20) to -0.34 V (MeCN)] is in accord with the exceptional reactivity of HOO-
in aprotic solvents. In aqueous media the reactivity of HOO- is leveled by
extensive anionic solvation,35 but remains significant with many substrates
due to its unique orbital energies 36 and the presence of an unshared pair of
electrons on the atom adjacent to the nucleophilic center (a. effect).37
The reactivity of 02-· with alkyl halides in aprotic solvents occurs via
nucleophilic attack by 02-· on carbon, or on chlorine with a concerted reductive
displacement of chloride ion.16 As with all oxy anions, water suppresses the
192
.,
I
j \
- - 4.8mM O.(1alm)
----- +2mM C.CI.
- ' _. +3mM C.CI.
1.0 o
E,VvsSCE
Figure 2. Cyclic voltammograms for dissolved 02 (1 atm, 4.8 mM) in dimethylformamide (0.1 M
tetraethylammonium perchlorate) at a glassy carbon electrode. The effect of 2 mM C6Cl6 and 3
mM C6Cl6 on the peak is illustrated by the dashed (-) and dash-dot (- . -) curves, respectively.
Electrode area, 0.062 cm2; scan rate, 0.1 V 5- 1.
193
Arochlor 1268
4.8mM 0.(1 atm)
Electrolysis at -1.0V vs SeE
5 Au electrode in DMF
o 1 2 3 4
Electrolysis time, hr
Figure 3. Concentration profiles for the various PCB isomers of Aroclor 1268 during the course of
an in situ electrolytic reduction (-1.0 V vs SCE) of dissolved 02 (1 atm, 4.8 mM) in DMF (0.1 M
TEAP) at a gold-mesh electrode (-20 cm2). Sample concentration, 50 mg Aroclor 1268 in 10 mL
DMF.
194
Table III. Redox Potentials for the Single-Electron Oxidation of HO- and Other Oxy Anion Bases in Water and in
Acetonitrile.
(P'B)H2 0b (P'B)MeCN d
Base (:B-) (pKHB)H20 a (V versus NHE) (p K HB)MeCNa,c (V versus NHE)
Cl- ~ Cl· + e- +2.41 +2.24

HO-~HO· +e:- 15.7 +1.89 30.4 +0.92
PhO- ~ PhO· + e- 9.2 -+0.7 16.0 +0.30
02-' ~ ·02· + e:- 4.9 -0.16e -13 -0.63 e
HOO- ~ HOO· + e:- 11.8 +0.20 -22 -0.34
H-~H +e- -2.20 -3.2
[dH20) == H + Ho-] ~ H20 + e:- -2.93 -3.90
IlpKa of the conjugate acid, Ref. 3l.

hRef.32.
CRef.33.
dRef.34
eStandard state for 02 is 1 M.
(0
U1
nucleophilicity of 02-· (hydration energy, 100 kcaI)38 and promotes its rapid
hydrolysis and disproportionation. The reaction pathways for these
compounds produce peroxy radical and peroxide ion intermediates (ROO· and
ROO-).
Hydroperoxide ion (HOO-) is unstable in most aprotic solvents, but
persists for several minutes in pyridine (kdecomp, 4.6 x 10-3 s-l), which allows
studies of its nucleophilic reactivity. In pyridine HOO- is oxidized in an one-
electron transfer to give HOO·, which is in accord with previous studies in
MeCN.39
HOO- ~ HOO· + e-, Ep,a, -0.34 V versus NHE (MeCN) (11)
In the case of HOO- (and t-BuOO-), the general leveling of nucleophilic

reactivity by protic solvents (water and alcohols) enhances its lifetime such that
the net reactions for HOO- with electrophilic substrates usually are most
efficient and complete in H 20 or t-BuOH. Almost all other solvents react with
HOO- or facilitate its decomposition. The relative lifetime of HOO- in various
solvents is in the order: H20»MeOH>EtOH»diglyme>pyridine-HPEGMe
3S0»MeCN> DMF»Me2S0.
Table IV summarizes the relative reactivities of MeO- /HO- /MePEG-
/HOO-/t-BuOO-/02-· with halocarbons and esters in pyridine and the kinetics
for the reaction of 02-· with the substrates in aprotic solvents. 40 The relative
reactivity with primary halides (n-BuBr and PhCH2Br) in pyridine is HOO-
(3.0»t-BuOO-(2.1»02-·(1.0»MeO-(0.3»HO-(O.2). For CCl4 the relative
reactivities are HOO-(1.8»t-BuOQ-(1.2»02-·(1.0»MeO-(O.6»HO-(H20)(0.4), and
for PhCI6; HOO-(S.0»t-BuOO-(2.9»02-·(1.0»MeO-(O.2»HO-(0.0).
The relative reactivity of HOO-/MeO- with primary halocarbons (n-
BuBr, PhCH2Br, BrCH2CH2Br) is about 10 in pyridine (Table IV), which
compares with a ratio of 13 for their reaction with BrCH2C(0)OH in H2041 and
35 for reaction with PhCH2Br in 50% acetone/water. 42 Although 02-· is a
powerful nucleophile in aprotic media, it does not exhibit such reactivity in
water, presumably because of its strong solvation by that medium and its rapid
hydrolysis and disproportionation. 38 Kinetic and electrochemical studies for
the reaction of 02-· with primary halocarbons confirm that the initial step is
rate limiting and first order with respect to substrate and 02-·. The other oxy
anions of Table IV are believed to react via an analogous nucleophilic
displacement.
In contrast, the reactions of CCl4 with oxy nucleophiles (YO-) cannot
proceed via an SN2 mechanism because the carbon center is inaccessible. The
primary step for the reaction of 02-· with CCl4 appears to involve a
nucleophilic attack of a chlorine to give CI3COO· and Cl-.20 The reactions of
196
Table IV. The Relative Reactivity of Mea- IHo-IMePEG-IHOO-It-BuOO-102-' with Halocarbons and Esters in Pyridine lO.l M (Et4N)Cl04j,
and the Kinetics for the Reaction of 02-' with Substrates.
a
[kX /k02-'] (±20%)
Substrate k02-./[S],b
(1-10 mM) X=Mea- c Ho- MePEG.d HOO-e t-BuOO- 02-' M-l s-l
CC4 0.6±O.3 0.4 0.05 1.8±O.6 1.2 1.0 1.4±D.5 x 1()3
n-BuBr 0.3±O.2 0.2 0.00 3.0±1.0 2.1 1.0 1.0±0.1 x 103

CH2Br2 0.3±O.2 3.1±1.0 1.0 2.3±O.S x 102
BrCH2CH2Br 0.3±O.2 2.7±1.0 1.0 1.8±O.3 xl()3
PhCH2Br 0.3±O.2 3.0±1.0 1.0 >3 x 1()3
Br(Me)CHCN 0.3±O.2 2.8±1.0 1.0 1.4±O.3 x 1031'

PhC16 0.15±O.U 0.0 0.00 5.0±1.5 2.9 1.0 1.0±0.2 x 1()3
MeC(O)OPh 0.2±O.1 4.5±1.5 1.0 1.6±O.5 x 102h
MeC(O)OEt 0.2±D.15 4.7±2.0 1.0 l.l±O.2 x 102h
aOetermined from the impact of X upon the rate of disappearance of 02-' in the presence of excess substrate (ref. 12); rate of disappearance of 02-' monitored by
linear sweep voltammetry.
bDetermined from the ratio of ip,a/iF,,c for the cyclic voltammogram of 02 in OMF in the presence of excess substrate (ref. 40).
CMeO- from (Bu4N)OH in MeOH. MePEC- (anion of monomethylpolyethyleneglycoI3S0). eHOO- prepared either from 02-' plus PhNHNH2 or HOOH plus He-.
(I) fk/[S] = 3.S x 102 M-1s-1 in MeCN. me primary product is ClsPhOMe. hKinetics in pyridine.
....
CCl4 with other oxy anions appears to be similar (e.g., CCl4 + HO- ~ Cl3COH +
Cl-).20 However, chloroform is deprotonated by HO- to give dichlorocarbene
(HCCI3 + HO- ~ :CCl2 + H20 + Cl-).20 Thus, the relative reactivity of oxy
anions with HCCl3 depends on their relative Brensted basisicity (pKa; MeO-
>HO->HOO->02-') rather than their nucleophilicity.
The reactions of hexachlorobenzene (PhCI6) with the oxy anions are
unique. Thus, HOO- is five times as reactive as 02-' and 33 times as reactive as
MeO- (Table IV). Superoxide ion reacts with PhCl6 via an initial nucleophilic
addition followed by displacement of Cl-.27,43 The other oxy anions probably
follow the same pathway. In the case of HOO-, the displacement of Cl- gives
PhCIsOOH, which deprotonated by a second HOO- (or HO-) to give PhCISOO-.
The latter displaces an adjacent chlorine to give the a-quinone.
Hydride Substitution
The hydride ion (H:-) is the penultimate nucleophile (the electron is a

stronger nucleophile and reductant), and has the advantage that it transforms
halocarbons to halide ions and the parent hydrocarbon, e.g.,
n-BuCI + H:- - - - - - -..- n-BuH + Cl- (12)
Although there are numerous approaches for hydro-dehalogenation in

the organic literature44 most are ineffective toward chlorinated and fluorinated
aromatics. Here we describe a new and unique method for the hydrogenolysis
of halogenated aromatics by sodium borohydride (NaBH4) mediated by
alkoxide and a nickel complex. Chlorobenzene (C6HSCl) has been used as the
model substrate for most experiments because it is readily available and has the
largest aromatic carbon-chlorine bond energy; (~HDBE' 106 kcal mol-I); as such
it is the least reactive for this class of molecules.
Figure 4 illustrates the product yields for the reduction of C6H s CI to C6H6
by (a) the combination of NaBH4' MXn, and MePEG- in MePEGH 350 and (b) the
combination of NaBH4 and MX n in monoglyme. In the case of Pd(OAch in
Figure 4a, 1 % of biphenyl also is produced. Curves c and d of Figure 4 indicate
the product yields for the reduction of C6HSCl to C6H6 by the combination of
NaH, MXn , and CH3(OCH2CH2hO- in monoglyme (c) and the reduction of 1-
bromonaphthalene (ClOH7Br) to naphthalene by the combination of NaH,
MXn , and tert-amylate in monoglyme (d).4S With the latter system NiCl2 also
yields 10% naphthalene dimer (R-R), and PdC1 2 yields 1 % R-R.
The reduction of C6HSCI to C6H6 by the mixtures of systems a, b, and c of
Figure 4 also have been evaluated with CrCl2, CrCl3, MnCl2, Mn(OAc)2,
Mn(OAc)3, FeCI2, FeCl3, COCl2, Co(OAch, Ni(OHh, PtC12, CuCl, Cu(OAc),
198
.. ..
100.-----~--------~~~.~~.-------~~~---------r-----,
I
I
I
I
I
I
80 I
I
I
I
I
I
I
I
I
I
I D
I
I
I
I
I
I
I
I
I
I
" ---tt-- (a) MXn/NaBH4/PEGM-
20 " ---<>--(b) MXn/NaBH4
" -_···o·····_(c) MXrlNaH/CH3(OC:H2CH~20-
" '·'·'a'·'·· (d) MXn/NaH/I-AmO-
I
I
I
MXn
Figure 4. Yields of C6H6 with various transition-metal catalysts from the hydrogenolysis of
C6HSCl by the combination of (a) _ _, NaBH4/MXn/PEGM- in PEGM 350; (b) --0-,
NaBH4/MXn in monoglyme; and (c) ..... c ....., NaH/MXn/CH3(OCH2CH2)20- in monoglyme; and
of naphthalene for the reduction of 1-bromonaphthalene by the combination of (d) -·_·4·_·,
NaH/MXn/ t-AmO- in monoglyme45 . In the case of (d), reactions were carried out under argon or
nitrogen for various durations [NiCl2, 17 h; Ni(OAc)2, 5 min; PdCl2 and Pd(OAc)2, 30 min] at
63°c. 45
cuC1 2, Cu(OAc)2, ZnC12, and Zn(OAc)2 as the catalyst; the yield of benzene for
each of these complexes is less than 30%. Control experiments establish that all
three components (transition-metal complex, NaBH4, and alkoxide) are
necessary for effective hydrogenation of C6H5Cl.
The product profiles (gas chromatograms) from the combination of
Aroc1or 1242 (in isooctane) with NiC12, MePEG-, and NaBH4 are presented in
Figure 5. The reactions of this combination with other samples of PCBs are
summarized in Table V.
The results of Figure 4 indicate that the mixture of transition-metal
complex, NaBH4' and alkoxide is effective for the hydrogenation of C6H5Cl.
Substitution of NaBH4 by NaH gives comparable reactivity (Figure 4c);
however, NaH is pyrophoric, more expensive, and has fewer reducing
equivalents than NaBH4. A similar system, MXn/NaH/t-amyl alcoho1 45,
for the hydrogenation of 1-bromonaphthalene (Figure 4d) exhibits similar
reactivity to the MXn/NaH/CH3(OCH2CH2)20- system (Figure 4c), but is
199
completely different from the present system. An optimized ReMED A&M
system consists of MePEG- MePEGH- 350 with a MePEG- : NiCl2 ratio of 12 : 1.
The absence of biphenyl as a by-product from the hydrogenation of
C 6H sCI indicates that the reaction pathway is not a free-radical process. The
counterion for nickel is critical [Ni(OH)2 and Ni(OAc)2 cannot be substituted
for NiCI 2]. The MePEG- alkoxide ion (a) shifts the reduction potential of nickel
Table V. Reactions of NiCI2/MePEG-/NaBH4 with Polychlorobiphenyls (PCBs).

% PCB's Removed on the
Basis of Assays for
~C~0~n~d~it~i~on~s__________________________~P~C~B~s~(+=5~o/,~o)a PhPh (+3%)a ___
Aroelor 1242 in isooctane (500 J.l.g/ g)b 90 72
Aroelor 1242 (500 J.l.g/ g)C 38 23
Aroelor 1242 in Diala AX oil (500 J.l.g/ g )b 35
Aroelor 1254 in untreated condensate (291 J.l.g/ g)b
Aroelor 1254 in pretreated condensate (291 J.l.g/ g)b 51 36
apCBs and PhPh were analyzed by GC/MS using SIM; samples were not
pretreated except as noted. dlO-Biphenyl was used as an internal standard
and surrogate for the quantitative assays of PCBs and PhPh.
bReaction time, 16 hi NiCl2 : MePEG- : NaBf4 = 0.1 : 0.2 : 0.8 mmol equiv.
CReaction time, 3 hi NiCl2 : MePEG- : NaBH4 =0.1: 1.2 : 0.4 mmol equiv.
dMatrix interferences precluded detection of PCBs.
tMatrix interferences precluded detection of PhPh.
to more negative values, (b) neutralizes the released HCI, and (c) acts as a
ligand for nickel.
The results of Figure 5 and Table V indicate that this system reacts with
PCB-contaminated wastes as well as C6HSF and C 6CIsOH. The latter substrates
usually are not subject to hydrogenation under mild conditions, which further
demonstrates the uniqueness of the system and its ability to hydrogenate a
wide range of halogenated aromatic substrates.
200
(II) Arnc:lor 12,,\2 in isooct.lI1c (500 J.l gJg) (ll) Ph Ph lon-chromatogram for (a)
k) RcMt:d A&M+ SolutionUI) (d) PhPh lon-chromatogram for (c)
Illlle
Figure 5. Gas chromatograms for (a) Aroclor 1242 in isooctane (500 f.lg/g); (b) extracted ion-
chromatogram of PhPh (154.1 amu) for solution (a); (c) the product solution from the reaction of
solution (a) with NaBH4/NiCI2/PEGM- in MePEGH350 after 16 h; and (d) extracted ion-
chromatogram of PhPh for solution (c). See Table II for more experimental details.
The Ultimate Nucleophile '" The Electron; Reductive Electrolysis
Although electrical energy is expensive, the stoichiometric introduction

of a universal nucleophile (the electron) whose nucleophilicity is determined
by the potential of the electrolysis cathode.
Figure 6 illustrates the electrolytic reduction of C6Cl6, 1,2,3,4,-C6H2Cl4,
C12Cho, and Arochlor 1268 mixture in DMF at a glassy carbon electrode.
Similar experiments for the other chloro derivatives of benzene and bi-phenyl
confirm that each of the six peaks of Figure 6a is due to the successive
reduction of the six chlorine atoms of C6Cl6. The decay and product profiles for
201
A. C.CI.
! 20IJA
P
B. 1,2,3,4 - C.H.CI.
1100PA
+ --="
C. C\2Cl w
l S0IJA
D. Arachlor 1268
1100IJA
o -1.0 -2.0
E,V vs seE
Figure 6. Cyclic voltammograms for chlorinated aromatic molecules in dimethylformamide

(O,lM TEAP) at a glassy-carbon electrode (area, 0.062 cm2 ); (A) 1.1 mM C6Cl6; (B) 2.3 mM 1,2,3A-
C6H2Cl4; (C) 1.3 mM C12CllO; and (0) 1 mg/mL Aroclor 1268. Scan rate, 0.1 V s-l.
202
the electrolytic reduction of C6Cl6 are illustrated by Figure 7 and indicate that
reduction produces the hydro-derivative of the substrate molecule (e.g., C6HCIs
from C6Cl6)' If the electrolysis of Figure 7 is controlled at -2.8 V vs. SCE
instead of -2.5 V, the final product is benzene rather than chlorobenzene
(PhCl).
Figure 8 illustrates the change in the composition of the Arochlor 1268
PCB mixture (average chlorine content, 68% by weight) during the course of its
controlled potential (-2.5 V vs. SCE) reduction in dimethyl formamide. The
most heavily chlorinated derivatives of biphenyl are reduced first to the lighter
members of the series; all are ultimately reduced to biphenyl. This
experiment and that of Figure 7 demonstrate that halogenated-aromatic-
hydrocarbon wastes (PCBs, PBB's, and halobenzenes) can be completely
dehalogenated to their parent hydrocarbons by electrolytic reduction in
dimethylformamide. Such a system provides an efficient means for the safe
transformation of these toxic materials into useful products.
In the absence of dioxygen, halogenated aromatic hydrocarbons can be
converted to their parent hydrocarbon by electrolytic reductions at a graphite or
gold cathode with Me4NCI as the supporting electrolyte. The data from Figures
6-8 indicate that the overall electrolysis reactions for C6Cl6 and C12CllO yield
C6HSCl and C12HlO, respectively.
-2.5 V
C12CllO + 10 H20 + 20 e- - - - - -...~~ C12HIO + 10 Cl- + 10 HO- (14)
The anodic process involves the oxidation of the chloride ion that is produced
in the cathode compartment
Cl- + HO- - - - - t.._ HOCl + 2 e- (15)
Again the overall process is equivalent to an electrostimulated oxidation of

water (eq 10 plus eq 11)
C12Cho + 10 H20 ---"'~~C12HIO + 10 HOCl (16)
with the anionic products produced at the cathode transfered via diffusion to
the anode and oxidized to HOCI.
203
14mM C.CI. in DMF;
Electrolysis at -2.SV vs SCE
CI
H $ H (18%)
H 0 CI
H
C~CI CI
,,,,WCI Hr$r.CI (4%)
r-" CI H H
H
1.0 \ CI
\
\ H $ H (78%)
H H
CIX~{I Clx~t
+ r--l '",
CI
1
\
0.8 f- \1 CI CI CI CI~ CI CI
'1
c
.Q
t)
...
ell
u..
0.6 f- \
~
/ \
~:~~:
1,_,H, / \ / '\1
:$:
) ....... H
I, \ ,1
\/ \/ \ r/ /\
\ / \, I' /I
'></
!\ ,/\ \, ' fl/ \\I ,:I' \1,
Y Y \' , \ /
0.4
: \ r', Ii
I I I \\
I,\,
I
I , \ ! //
\
/
"" /
0.2 f-
II \\
I \
\\ ! \
' I, \
/
,
\I / "
"
1/ I / 1/ I ' '
V \ / Y I /
/' \ / " \ \ /'
0~----_/--'-'~--L-1_\~:------/----·~)·'~____~""~>'___L-1____________1-1______~
o 1.0 2.0 3.0 4.0
Electrolysis Time, hr
Figure 7. Concentration profiles for the controlled-potential (-2.5 V vs. SCE) electrolysis of 14
mM C6Cl6 in DMF (O.lM TEAP) at a gold-mesh electrode (area, -20 cm2). The relative amounts
of products and reactants were determined by capillary gas chromatographic analysis of samples
from the electrolysis solution.
204
A. Arochlor 1268(50mg in 10ml DMF(O.lM TEAP))
OJ
c
Ol
en
o
l B. After Reductive ElectrolysIs at -2 5V ::rSCE•
L JJJi.o......----...,-
---'--
uQ)
a; C 3hr
o
c
o
~
.!::!
c
~ ,
Q)
E
'"
Ii:
1
D. 5hr
r'-....LL---I_--
o 5 10 15 20
GC Retention time. min
Figure 8. Gas chromatograms for Aroclor 1268 in DMF, and for the product solution during the
course of its electrolytic reduction at -2.5 V vs. SeE with a 20-cm2 gold-mesh electrode. The peak
with a retention time of 2 min is biphenyl, and that at 5 minutes is chlorobiphcnyl.
205
REDUCTIVE HYDROGENATION BY HYDRAZINE
Conventional hydrogenation with H2 under mild conditions is not

effective for chlorinated aromatics. 44 However, anhydrous hydrazine
(H 2NNH 2) and substituted hydrazines are effective for the dechlorination of
polychloro aromatics. Table VI summarizes the reaction efficiencies (expressed
as the number of chloride ions released per C 6C16 molecule) and the products
that result from of hydrazines and hexachlorobenzene (HCB, C6Cl6, PhC16).
The results of Table VI indicate that hydrazine is an effective agent for
the transformation of C6Cl6 to C6H6. The efficiency (expressed as the number
of chloride ions released per C6C16) is less than the theoretical 100% (six Cl-)
due to the evaporative loss of the lower-chlorinated benzenes. The reaction
rate appears to be first-order in H2NNH2 concentration and first-order in
substrate concentration. The relative reaction rates for the various solvents are
in the order Me2S0 - HPEGMe 350 > pyridine - DMF, which is consistent with
their relative polarity. Thus, the reaction intermediates are stabilized by the
more polar medium.
The reaction products from the combination of H2N N H 2 (or
CH3NHNH2) with C 6Cl6 (Table VI) are consistent with an initial elimination of
HC1,
- - - - - -...- ClsC6NHNH2 + HCl (17)
with subsequent para- or or tho-elimination of HCl to give the different

isomers.
-------- 1,2,3,4- and 1,2,4,5-C6H 2CI4 + HCI + N2 (18)
Additional cycles of these steps results in complete reduction via the

overall reaction.
(19)
The proposed mechanism is consistent with that suggested for the reaction of
C6Br6 with hydrazine hydrate. 46 The latter was unreactive with C 6Cl 6 in
boiling ethano1. 46 The N-H bond dissociation energy [-76 kcal mol- 1]47 is the
lowest of the hydrazine derivatives, and probably is the reason for its superior
reactivity. The lack of reactivity for PhNHNH2 and PhNHNHPh is in accord
with this proposition.
The present results provide a basis for a reactor to process
polyhalogenated aromatic wastes, whereby anhydrous hydrazine transforms
206
Table VI. Reactions of Hydrazine and Substituted Hydrazines with C6Cl6·
Cl- Released
Conditionsa Per Substrate (±O.5)b Major Products Minor Products
A. Hydrazine (H2NNH2)C
100% H2NNH:u 25°C 0.0
Me2S0, 4% H2NNH:u 25°C 0.7 1,2,3,4-CGH 2Cl4, 1,2,3,5- and/or

Argon atmosphere ClSC6NHNH2 1,2,4,S-C6H 2CI4d
100% H2NNH:u 70°C 4.9 C6HG, C6HSCl, 1,2-C6H4CI2

1,4-C6H4Cl2
Me2S0, 20% H2NNH:u 70°C 3.7 1,4-C6H4Cl2 C6H6, C6HSCl,

1,2-C6H4Cl:u
1,2,3-C6H3CI3,
1,2,4-C6H 3CI3
Me2S0, 2% H2NNH:u 70°C 4.1 1,4-C6H4Cl2 1,2-C6H4Cl:u

1,2,4-C6H3CI3,
1,2,3,4-C6H2CI4
HPEGMe 350, 20% H2NNH2, 70°C 3.0 1,2,4-C6H3Cl3, 1,2,3-C6H 3ClJ,

1,2,3,4,-C6H2C Li 1,2,3,5- and/or
1,2,4,S-C6H 2Cl i
HPEGMe 350, 20% H2NNH:u 150°C 4.1 1,4-C6R4CI2 1,2-C6H4Cl:u
1,2,4-C6H 3CI3
DMF;20% H2NNH:u 70°C 2.1 1,2,3,4-CGH2CI4 1,2,4-C6H3ClJ,

1,2,3,5- and/or
1,2,4,S-C6H 2Cl4d
Pyridine, 2% H2NNH:u 70°C 2.5 1,2,3,4-C6H2Cl4 1,2,4-C6H3ClJ,

1,2,3,5- and/or
1,2,4,5-C6H 2Cl4d
B. Substituted Hydrazines
100% CH3NHNH2. 70°C 2.9 1.2.4-C6H3Cl3. 1.2.3.4-C6H 2CI4'
ClSC6NCH3NH2 Cl4C6(NCH3NH2)2
100% PhNHNH:u 70°C 0.0
Me2S0, 0.6 M PhNHNHPh, 70°C 0.0
aUnless noted otherwise, reactions were carried out by adding C6Cl6 (3-6 mM in toluene) to 5-10 mL of a
hydrazine/solvent mixture. The reaction cell was heated at ca. 70°C for 18-24 h, with the cap slightly
loosened to prevent pressure buildup.
bThe reactions were quenched with dilute acid and extracted with toluene. The aqueous layer was
analyzed for chloride ion by potentiometric titration with AgN03 at pH 0.4. Because H2NNH2 and its
derivatives are interferences in the AgN03 determination of Cl-, they were removed by making the
solution alkaline (pH 14) in the presence of 02 (0.2 atm). The equivalence point was determined by the
second-derivative method. The organic layer was analyzed by GC/FID and/or GC/MS; evaporative losses
during reaction precluded accurate assays of the products.
CThe reaction of H2NNH2 with either a condensate that contained Aroclor 1254 or a stock solution of
Aroclor 1254 gave erratic analytical results; probably because of inhomogeneities in the condensate or
significant solvent loss during the heated reaction cycle.
dThe GC eqUipment and parameters used for the analyses could not resolve 1,2,3,5- and 1,2,4,S-C6H2C14'
207
toxic organics into N2, Cl- (under alkaline conditions), and the recyclable parent
hydrocarbon.
ATOMIC METALS VIA ELECTROGENERATION
Research has been initiated in our laboratories to develop an

electrocatalytic system that will reduce alkali salts (NaCl, LiCl, or KCl) to their
atomic state (Na·, Li·, or K) within a solution matrix that contains hazardous
halogenated hydrocarbons. Although alkali metals react with water, their
reaction rates are several orders of magnitude slower than with halocarbons.
By an in-situ formation of metal atoms at a carbon-surface electrode, a selective
and efficient dehalogenation process should be achieved, e.g.
RX,HP
2 Na+ + 2 e- - - - -..- 2Na· .. RH+NaX+NaOH (20)
We are evaluating this approach with (a) n-BuCl, CCI4, HCCI3, PhCI, and C 6Cl6
as model substrates, (b) Me2S02, DMF, MeCN, THF, and pyridine as solvents,
(c) NaCI, LiCI, KCI, NiCl2 , and MnCl 2 , as intermediates for the production of
metal atoms, and (d) cyclic voltammetry and controlled-potential electrolysis at
glassy carbon electrodes as the means to generate atoms and anions. The effect
of various levels of water on the efficiency of dehalogenation is being tested.
The advantages of this approach are that (a) the parent hydrocarbon and halide
ions are the products (rather than free halogens and CO 2 from a oxidative
process) and (b) a reductive process does not produce hazardous intermediates.
Although alkali-metal salts are conveniently reduced at platinum and
gold electrodes (reduction potentials, --2.0 V vs SCE), their metal atoms form
metal-metal bonds with Pt and Au, which suppresses their reactivity with
halocarbons. However, at carbon the initial reduction deposits metal atoms
(--2.6 V vs SCE), whose reactivity with C6Cl6 and PCBs is much greater than
with residual water because of their radical character (Na·, Li·, and K). Because
PhCl has the strongest C-Cl bond within the unsaturated-halocarbon family, it
has been used as a model substrate to optimize the conditions and to test the
basic proposition: That in-situ generated alkali-metal atoms will react
preferentially and efficiently with halocarbons (PhCl) relative to residual water.
In acetonitrile at glassy carbon (a) chlorobenzene (PhCI) is reduced to a
phenyl radical,
PhCl + e- MeCN
C .. Ph· + CI- Ered, - 28
. V vs SCE (21)
(b) water to hydrogen atoms,
208
C
H20 + e- MeCN
~
H· + HO- Ered, -4.1 V vs SCE (22)
and (c) sodium chloride to sodium atoms,
Na· + cr E red, -2.6 V vs SCE (23)
IH0
~NaOH+H.
2
which react with residual water. However, with the combination of PhCI,
NaCl, and H 20 the initial reduction is at -2.3 V vs SCE,
C . NaCl, H 20, e-
PhCI + NaCl + e- II' [Na-PhCI] II' PhH (24)
MeCN
+ NaOH
and is specific for PhCl.
Scheme IV outlines the reaction paths that are associated with the
Scheme IV. Reaction of Electrogenerated Atomic Sodium with PhCl.
-2.8 V
vsSCE JI'
MeCN
HP!.-
PhH + HO- + cr
NaONa + H2
0-
2 NaOH
Cl
+ Na+ + e- -2.3 V
- - -....
- :. I Na
' .... _ i
HP~'
PhH + NaCl + HO-
209
process. A clear understanding of the relative bond energies associated with
the substrates, reactants, intermediates, and products is necessary in order to
optimize an efficient and selective dehalogenation. Table VII summarizes the
dissociative bond energies (ilH DBE) for a number of relevant diatomic
molecules, and is the basis of our current research efforts.
Several reductive approaches for the transformation of carbon-halogen

bonds to halide ions have been outlined. No one system is ideal for all
problems, and the environment of the hazardous waste often interferes with
Table VII. Dissociative Bond Energies for M-X Molecules.
ilHDBE, kcal mol- 1
Metal(EA)a M Cl(3.62) H(0.75) 0(1.46) 5(2.08) N(O.O)
Li(0.62) 26 112 57 80
Na(0.55) 17 99 44 61 75
K(0.50) 14 104 42 66
Cu(1.24) 42 92 66 64 66
Ni(1.16) 49 89 60 91 82
Co(0.61) 40 93 54 92 79
Fe(0.15) 24 84 43 93 77
Mn(O.O) 6 86 56 96 72
Cr(0.67) 37 88 67 103 79 90
V(0.52) 58 114 150 115 114
Mo(O.75) 97 134
Ru(1.05) 56 126
C(1.26) 145 95 81 257 171 184
H(0.75) 104 103 104 102 82 81
aEA, electron affinity in electron-volts (eV).
210
effective remediation. Hence, effective treatment will require careful analysis
and intelligent design of the treatment strategy for each problem. For most
contamined soils, sludges, muds and water, some separation process will be
required prior to the chemical destruction of carbon-halogen bonds.
Acknowledgement: This work was supported by the National Science

Foundation under Grant CHE-8516247 and CHE-9106742, the Welch
Foundation under Grant A-I042 with a Robert A. Welch Graduate Fellowship
(S.J.), and a Grant-in-Aid from Captiva Capital, Ltd., Colorado.
References
1. J. Waid. "PCBs and The Environment," CRC Press, Boca Raton, (1986), Col. I-III.
2. F. M. D'Itri and M. A. Kamrin, "PCBs: Human Environmental Hazards," Ann Arbor Science
Book, Boston, (1983).
3. O. Hutzinger, S. Safe and V. Zitko, "The Chemistry of PCBs," CRC Press: Cleveland,
(1974).
4. S. Safe, S. Bandiera, T. Sawyer, L. Robertson, L. Safe, R. Parkinson, P. E. Thomas, D. E.
Ryan, L. M. Reik, W. Levin, M. A. Denomme and T. Fujita, "EHP, Environmental Health
Perspect," 60, 47-56(1985).
5. Safe, S. CRC Crit. Rev. Toxicol,13, 319-396(1984).
6. R. D. Kimbrough. CRC Crit. Rev. Toxicol 2, 445-498(1974).
7. L. Fishbein. Annu. Rev. Pharmacal. 14, 139-156(1974).
8. E. E. McConnell and J. A. Moore, Ann N.Y. Acad. Sci, 320, 138-150(1984).
9. R. S. Takazawa and H. W. Strobel Biochemistry, 25, 4804-4809(1986).
10. J. R. P. Cabral, P. Shubik, T. Mollner and F. Raitano, Nature (London), 269, 510-511(1977).
11. F. W. Karasek and L. C. Dickson, Science(Washington, D.C.), 237, 754-756(1987).
12. D. T. Sawyer and J. L. Roberts, Jr., Acc. Chern. Res., 21, 469-476(1988).
13. L. L. Pytlewski, F. J. Iaconianni, K. Kevitz, and A. B. Smith, U.S. Patent 4,417,977, Nov. 29,
1983.
14. D. J. Brunelle, A. K. Mendiratta, and D. A. Singleton, Environ. Sci. Technology. 19, 740-
746(1985).
15. R. Dietz, A. E. J. Forno, B. E. Larcombe and M. D. Peover, J. Am. Chern. Soc. 102, (1970).
16. M. V. Merritt and D. T. Sawyer J. Org. Chern. 35, 2157(1970).
17. F. Magno, R. Seeber and S. Valcher, J. Electroanal. Chern. 83, 131(1977).
18. J. San Fillipo, Jr., C.-I. Chern and J. S. Valentine, J. Org. Chern. 40, 1678(1975).
19. R. A. Johnson and E. G. Nidy, J. Org. Chern. 40, 1680(1975).
20. J. L. Roberts, Jr., T. S. Calderwood and D. T. Sawyer, J. Am. Chern. Soc. 105, 7691(1983); D. T.
Sawyer and J. L. Roberts, Jr., U.S. Patent 4,410,402, Oct. 18, 1983.
21. T. S. Calderwood, R. C. Newman, Jr. and D. T. Sawyer, J. Am. Chern. Soc. 105, 2337(1983).
22. T. S. Calderwood and D. T. Sawyer, J. Am. Chern. Soc. 106, 7185(1984).
23. M. V. Merritt and R. A. Johnson, J. Am. Chern. Soc., 99, 3713(1977).
211
24. L. Eberson, Adv. Phys. Org. Chem. 18, 79,(1982).
25. C. L. Perrin, J. Phys. Chem. 88, 3611(1984).
26. J. K. Koehl, In Free Radicals <T. K. Kochi, ed.). New York: Wiley, (1973).
27. H. Sugimoto, S. Matsumoto and D. T. Sawyer J. Am. Chern. Soc., 109, 8081(1987).
28. D. T. Sawyer, T. S. Calderwood, K. Yamaguchi and C. T. Angelis, Inorg., Chern, 22, 2577-
2583(1983).
29. K. Yamaguchi, T. S. Calderwood and D. T. Sawyer, Inorg. Chern, 25, 12- 1290(1986).
30. S. Jeon and D. T. Sawyer, Inorg. Chem., 29, 4612-4615(1990); D. T. Sawyer, S. Jeon and
P. K. S. Tsang, U. S. Patent Pending (to be issued, June, 1992).
31. R. G. Pearson, J. Am. Chern. Soc., 108, 6109(1986).
32. A J. Bard, R. Parsons and J. Jordon, "Standard Potentials in Aqueous Solution, Marcel
Dekker, New York(1985).
33. W. C. Barrette, Jr., H. W. Johnson, Jr., and D. T. Sawyer, Anal. Chem., 56, 1890(1984).
34. P. K. S. Tsang, P. Cofn? and D. T. Sawyer, Inorg. Chern., 26, 3604(1987).
35. C. D. Ritchie. Acc. Chern. Res. 5, 348(1972).
36. S. S. Shaik and A. Pross, J. Am. Chem. Soc., 84, 16,1962.
37. J. O. Edwards and R. G. Pearson, J. Am. Chern: Soc., 84, 16(1962).
38. D. T. Sawyer and J. S. Valentine, Acc. Chern. Res., 14, 393(1981).
39. P. Cofn~ and D. T. Sawyer, Inorg. Chem., 25, 2089(1986).
40. J. L. Roberts, Jr. and D. T. Sawyer, J. Am. Chern. Soc., 103, 712(1981).
41. J. E. McIssac, Jr., L. R. Subbaraman, J. Subbaraman, H. A Hulhausen and E. J. Behrman, J.
Org. Chern., 37,1037(1972).
42. R. G. Pearson and D. N. Edgington, J. Am. Chem. Soc. 84,4607(1962).
43. H. Sugimoto, S. Matsumoto and D. T. Sawyer, Environ. Sci. Technol. 22, 1182(1988).
44. J. March, "Advanced Organic Chemistry," 3rd. ed., Wiley & Sons: New York, 1985; and
references therein.
45. J. J. Brunet, R. Vanderesse and P. Caubere, J. Organometal. Chern. 157, 125-133(1978).
46. I. Collins, S. M. Roberts and H. Suschitzky, J. Chem. Soc. (C),167-174(1971).
47. E. W. Schmidt, "Hydrazine and Its Derivatives," Wiley & Sons: New York, 1984; pp 247-
251.
212
PROCESS TECHNOLOGY FOR HAZARDOUS WASTE REMEDIATION -
(KPEG- AND REDUCTIVE TECHNOLOGIFS)
Robert Hoch
SDTX Technologies, Inc.
ABSTRACT
A number of technologies have been proposed and promoted for remediation of media
contaminated with chlorinated hydrocarbons. While some of these approaches are flawed
on first principles, potentially leading, for instance, to production of more hazardous
substances, others lack a realistic engineering assessment of what an overall system
embodying the chemistry would comprise.
Unlike untried technologies, alkaline glycolate dechlorination, commonly called
KPEGIm, has a long and successful history, having been extensively used in the destruction
of PCB contaminated transformer oil. This chemistry has now been extended to solids and
has been applied to the remediation of over 30,000 tons of PCB contaminated soil at the
Wide Beach, NY Superfund site.
This paper will briefly review the history of KPEa-, and present experimental data
on the complete dechlorination of chlorobenzene and trichlorobenzene as model
compounds. The importance of complete dechlorination to project economics will be
discussed and the significance of the Wide Beach Project to technology selection for PCB
contaminated soils will be assessed.
The total cost of the national remediation effort is sufficiently large that economics,
until now quite a minor factor in technology selection, must eventually become a key
driving force. Based on this assumption and first principle arguments about efficacy and
with non-quantifiable, but quite obvious public prejudices in mind, one can define the
design basis of a cost-effective remediation process, where the term process is used in the
usual context of the chemical industry. It is shown that the cost of processing in such a
system is relatively bounded. Furthermore, the regime of applicability of KPEa- can be
identified as can generic critical features of its integration with other technologies, for
remediation of more complex wastes.
OVERVIEW
Chlorinated hydrocarbons present some of the most vexing problems of hazardous

waste contamination. Their environmental persistence coupled with their general and wide

and A.E. Martell. Plenum Press. New York, 1992 213
spread use persisting well into the 1970's leaves us today with many chlorinated
hydrocarbon contaminated media; soil, water, river sediments and various industrial
wastes.
The spectrum of technologies that have been proposed for remediation of these
situations is staggering. In a rational, economics-driven business, this would suggest that
the appropriate combination of remedy and problem is quite specific. In the remediation
industry, however, it rather represents a significant opportunity attracting technology
regardless of relevancy.
This situation is a microcosm of the hazardous waste marketplace where irrationalities
thrive. Among the prominent reasons for this are:
• extensive regulation and attendant red tape

• the relative newness of the industry
• the unwillingness to acknowledge a fundamental base in chemistry
• the tendency to leap from concept to hardware without the intermediacy of
a process engineer
Given that the industry lacks a rational basis for technology selection, the following
criteria are proposed:
• acceptable economics
• technical acceptability
• social acceptability.
It is the thesis of this paper that many of the technologies suggested for remediation of
PCBs fail to meet one or more of the above criteria.
While EPA regulations and procedures, cumbersome as they may be, address the
issues of technical and social acceptability, much entropy has been generated by the
relegation of economics to the bottom of, for instance, the list of criteria for the selection
of superfund remedies and by persistent government support for technology development
without a priori indication of economic superiority to existing technology for any
contaminant in any media. Much more could be written about the unnecessary degree of
confusion surrounding the choice of remediation technology.
This paper, however, will focus specifically on issues related to chlorinated
hydrocarbons and most especially PCBs and will show that many categories of technologies
can be ruled out on first principles. In fact, for sound chemical reasons, reductive systems
of destruction are preferred to oxidative systems. The term dechlorination, in the
remediation art, is usually applied to these reductive systems. Among such reductive
systems, the reaction with sodium or potassium polyethylene glycolate, frequently called
NaPEG, KPEG- or APEG, has a long and successful history and clearly represents the
most fully demonstrated such chemistry.
EARLY HISTORY OF REDUCTIVE DECHWRINATION SYSTEMS
In 1978, Dr. Louis Pytlewski of Drexel University, working at the Franklin Institute
Research Center, discovered that aromatic chlorinated compounds like PCBs could be
completely dechlorinated by reaction with alkaline glycolates (Pytlewski, et al, 1982).
Before this invention, the only practical chemical reaction for destruction of PCBs was the
relatively uneconomic reaction with dispersed metallic sodium originally developed by
Goodyear (Berry, 1981).
At the time of Dr. Pytlewski's invention, government regulations, responding to the
availability of technology, the ease of treatment and the pervasiveness of the problem
214
identified transformer oil as the initial remediation target. The alkaline glycolate treatment
(called NaPeg by the Franklin Institute) lent itself readily to this application. Work in this
area by Franklin, including pilot studies with electric utilities, led to a number of broad
patent filings (Pytlewski, etal 1983 a,b, 1984 a,b,c, 1986). General Electric (Brunelle
1982,1983) and others, eventually fully proved the viability of the glycolate chemistry by
successful commercial application in transformer oil. The focus of commercial interest in
the United States has now shifted away from relatively well defmed systems like
transformer oil to the remediation of PCB and dioxin contaminated soils, sludges and
sediments.
UNDERLYING CHEMISTRY
It is generally accepted that the initial reaction (between, for example, potassium
polyethylene glycolate and a PCB) is a nucleophilic substitution to produce an aromatic
ether, as shown in Reaction 1 of Scheme One.
When this reaction is used for the remediation of transformer oil, further reaction is
unnecessary (and may, in fact, be undesirable) because the ether is insoluble in the
dielectric fluid and thus easily separated. Conventional practice is to landfill or incinerate
the recovered ether.
Although numerous studies commissioned by the EPA and others show various of
these polychloromonoethers to be nontoxic, biodegradable and not to bioaccumulate
(DeMarini and Simmons, 1989), there remains resistance to leaving chlorinated organics
in the soil. The nonhazardous nature of this material is clearly important in soil systems,
since recovery of this reaction product from the contaminated medium is impractical. It
becomes difficult to argue for on-site re-interment of treated soil containing chlorinated
organics.
Early work at the Franklin Institute showed that PCBs can be completely dechlorinated
via the alkaline glycolate process. Reaction 2 of Scheme One shows a likely path for
complete dechlorination. No analytical protocol now exists for analysis of the reaction
products, but complete recovery of the chlorides in inorganic form has been demonstrated
(King and Hoch, 1991).
Complete dechlorination is the key to producing a decontaminated solid phase not
requiring treatment as a hazardous waste. The implied structure of the products of reaction
2 strongly supports such arguments. Not only is polyethylene glycol (pEG) non-toxic, but,
in the biotechnology industry, PEG adducts similar in structure to the products of Reaction
2 are intentionally manufactured to detoxify otherwise dangerous medicinals. (Davis,
etal.,1979) Thus, a substantial body of data, including human clinical trials, have shown
that reaction with PEG can detoxify otherwise toxic compounds.
The sensitivity of these reactions to moisture level is an issue in extension of this
chemistry to soil systems. Scheme One shows chemically why the presence of water is
incompatible with alkaline glycolate dechlorination chemistry. In the presence of water,
reactions 1 and 2 must compete with the quite facile hydrolysis of the glycolate as shown
in reaction 3. Thus, in order to guarantee the desired reaction path, soils, sludges and
sediments should be dried before reaction.
Both theory and experimentation have shown that the hydroxy-chloro-PCBs formed
by reaction 4 are less reactive than the ether products of reaction 1 to further
dechlorination by either KPEG"'" or KOH. (Doubly hydroxy substituted products would
theoretically be essentially completely resistant to further dechlorination.) Any significant
opportunity for reaction path 4 to occur therefore will prevent complete dechlorination.
Perhaps even more troublesome, these hydroxy species are on the reaction path that
converts PCBs to dioxins and polychlorodibenzofurans (PCDFs). Reaction path 6 shows
the formation of PCDF by simple ring closure. Production of dioxins, is shown formally
215
N
01
Scheme 1.
Cl n Cl,. a,. °n-l
·\~PEG +KQ
@-@ .~~ 1 ~
HoO I 3\HZO
carb;nates 212n-lKPEG
/). KOH + HPEG

7
CO 2 412KOH
} I (PEG)n + 2 n-l KCl
(PEG)n I (
0." C n-l
J'\~O- + K'" + KQ
~KOH
6 \ 5~
+ 2KQ
~~~' l§(:®
+ C- DIOXIN
by reaction S. The generation of PCDFs and dioxins will significantly complicate any
remediation.
Another chemical issue relates to the potential in situ formation of the glycolate
reagent. Although frequently practiced, this apparent economy also leads to loss of
certainty as to reaction path, because at practical reaction (or desorption) temperatures,
KOH (or NaOH) will react with a PCB as shown by reaction 4. This reaction is much less
rapid than reaction 1. Furthermore, the products of reaction 4 are particularly undesirable.
Note that alkaline glycolates react as readily with dioxins and PCDFs as with PCBs.
Reactions 5 and 6 become of concern only in the absence of glycolate. As long as the
system has been dried, reaction will proceed along path 2.
Scheme One also shows why thermal desorption of PCB contaminated soils and
sludges without integrated dechlorination is a chemically and economically risky
undertaking. In the presence of normal moisture levels and natural soil bases like
carbonates, at elevated, ie. desorber, temperatures, hydration according to reaction 7 can
lead to the same kind of hydroxy-PCBs as reaction with KOH. Although the simple
hydration is much less favored than reaction 4, theory suggests that at desorber
temperatures in the range of ca. 300"C, the reaction becomes possible.
PROPOSED TECHNOWGIES FOR CHWRINATED HYDROCARBON

REMEDIATION
Based on the arguments in the preceding section, one can understand why essentially
any non-reductive remediation strategy is technically flawed. Two such approaches:
incineration and bioremediation, however, continue to be proposed.
Incineration
The emotion with which the public frequently greets the idea of incineration is well
known. As applied, however, to incineration of chlorinated wastes, there are sound
technical reasons why this hysteria should not be too hastily dismissed.
In 1987, an EPA investigator (Oppelt, 1987) stated:
" ... what appears outwardly to be a straightforward, simple process is actually an

extremely complex one involving thousands of physical and chemical reactions,
reaction kinetics, catalysis, combustion aerodynamics, and heat transfer. This
complexity is further aggravated by the complex and fluctuating nature of the
waste feed to the process. While combustion and incineration devices are
designed to optimize the chances for completion of these reactions, they never
completely attain the ideal. Rather, small quantities of a multitude of other
products may be formed, depending on the chemical composition of the waste
and the combustion conditions encountered. These products along with
potentially unreacted components of the waste become the emissions from the
incinerator. "
Most risk assessment for incineration ignores these truths. Specifically, risks
associated with products of incomplete combustion (PICs) are frequently ignored as is the
startlingly high probability of operation at off-normal conditions. Such mal-operation both
exaggerates the PIC risk and can lead to emission of hazardous uncombusted feed material.
Products Of Incomplete Combustion The potential for incomplete combustion, especially

in systems fed with chlorinated materials, has long been recognized. Studies of PCB
combustion have focused on dioxins and polychlorodibenzofurans (PCDFs) as PICs. The
217
production of dioxins in incineration has been much studied, (Siebert, etal. 1989;
Altwicker, etal. 1989; Shaub and Tsang, 1983), but, like the generation of all products of
incomplete combustion, mechanism is poorly understood. An excellent summary of all the
uncertainties surrounding generation of PICs was recently given by Dellinger and co-
workers. (Dellinger, etal. 1990)
It is well known, however, that the production of dioxins and PCDFs is more serious,
all other things being equal, in municipal solid waste (MSW) incinerators than in industrial
liquid stream incinerators. This is generally attributed to catalysis by fly ash (Vogg and
Stieglitz, 1986; Lane and Hinton, 1990; Altwicker, 1988; Altwicker, etal. 1989).
Although the nature of the catalytic sites of the fly ash has apparently never been
determined, or even sought, there is evidence of a surface catalytic reaction. (Altwicker,
etal. 1989).
In the pyrolysis of soils, sludges and sediments, the volume of solid material per unit
of chlorinated contaminants far exceeds that in the MSW incinerators. Furthermore,
because of the variability of natural soils and the presence in sewage sludge of a variety
of metal residues, the opportunity for adventitious catalysis is ripe. We know of no
literature describing PICs from combustion of PCB contaminated soils, sludges or
sediments. It is significant, however, that test bums of PCB in liquid injection incinerators
(Oppelt, 1987) produced measurable quantities of PCDFs.
Unlike dioxins and PCDFs, some PICs probably remain unidentified. The inability
to close chlorine material balance in even the most careful incinerator analyses better than
70% (Dellinger, etal. 1990) strongly suggests that such unidentified PICs exist. Of
particular concern is the potential for formation of higher molecular weight analogs of
dioxins and PCDFs by sequential occurrence of the same type of condensation reactions
that are known to form dioxin from pentachlorophenol. These products may be themselves
toxic or may be the antecedents of toxic products produced by the action of natural systems
on the discharged soil, for instance, by cleavage back to dioxins and PCDFs.
An alternative mode of exit for PICs is via the gas phase. These materials then enter
either the air or the scrubber water. This risk is frequently understated.
Oppelt (Oppelt 1986) found that the performance of incineration deteriorated when less
than 500 ppm levels of contaminated materials are present. In fact, a survey of many tests
(Oppelt 1987) showed that no compound below 200 ppm met the RCRA 99.99%
destruction limit in any incinerator. This phenomenon is frequently attributed to
recombination reactions in the effluent. An alternative or contributory mechanism,
however, may be by transport of contaminated aerosols. Biswas, (Biswas and Sethi, 1990)
found the formation of aerosols to be exacerbated in incinerators containing metals. We
know of no fundamental work in soil or sludge filled incinerators, but one could anticipate
a similar enhancement of aerosol formation.
Probability of Off-Normal Operation Perhaps, an even more serious criticism of the risk
commonly assigned to incineration concerns the probability of off-normal operation of the
incinerator. Techniques for probabilistic assessment of the significance of such off-normal
operation are only beginning to emerge (Zeng and Okrent, 1989).
The ease with which such off-normal operation can be encountered is startling.
Okrent and co-workers (Zeng and Okrent, 1989) point out that a 100°C change in
operating temperatUre can increase emission of POHC's by a factor of 5 to 500. Such a
temperature anomaly can easily result from thermocouple drift. A recent survey article
(Lewis, 1990) highlights the fact that drift is particularly sensitive to the system
atmosphere. Because of the inherent variability of feed soils, sludges and sediments,
selection of a temperature measurement device resistant to drift cannot be assured. This
drift leads to a slow change in actual operating condition which may not be detected for
some period of time. The effect of such maloperation on a feed containing PCBs, and
metals only heightens the concern over the issues of dioxin, PCDF and PIC generation.
218
That such incidents can occur, even in carefully supervised facilities, is shown
anecdotally by Stumbar and co-workers (Stumbar, etal. 1990) who describe a PCDF
formation excursion in an EPA operated test incinerator.
To recapitulate, there are significant gaps in understanding of both the chemistry and
engineering of incineration of chlorinated compounds. Because of the variability of the
feed, the probability of off-normal operation and the chronic inability to close the chlorine
material balance, these uncertainties remain, even in the face of nominally successful test
bums. Incineration is simply not an acceptable disposal option for chlorinated organics and
will not become acceptable until much more fundamental understanding and analysis has
been amassed.
Bioremediation
Biotechnology, in contrast to incineration, frequently encounters less public hostility.

Furthermore, for reasons unclear, it is, or has been, in the good graces of the EPA. It is
well known, however that chlorinated substrates provide a challenge to bioprocessing. In
fact, many historic applications of PCBs relied upon the resistance of the material to
oxidative and natural attack. The persistence of PCBs in the environment is testimony to
this.
It is important to differentiate between in situ and ex situ bioremediation. In situ
biological systems are frequently an effective option for readily degraded substrates. In its
simplest form, it consists of merely assisting and accelerating naturally occurring
biosystems. Not only is such an approach environmentally friendly, it is cheap.
Ex situ bioremediation, on the other hand, suffers from the same malady that limits
biosynthesis to production of human consumables and extremely complex organics:
processing temperature is limited. When alternative synthetic chemistry is available,
extension of such chemistry to elevated temperatures will result in a rate and cost benefit
that the biologically mediated process simply cannot overcome.
Thus, to be economically attractive, a bioremediation must be an in situ process. This
effectively rules out genetically engineered organisms. Most proposals for bioremediation
of PCBs rely either on engineered organisms or on an undefined combination of
observations involving both anaerobic and aerobic organisms. 1 There is no basis to
believe that a practical bioremediation process for PCB destruction can be developed.
There is, on the other hand, no doubt that by applying the full power of modem
genetic engineering one can, in the fullness of time, create organisms that will consume
PCBs. Even then, however, a process utilizing such organisms will still fail to meet all
three of the criteria for an effective technology. Rates will be too low to produce
acceptable costs. As in all biological systems, very low levels of substrate will adversely
effect population dynamics and therefore very strict "clean levels" will not be achievable.
EARLY EXTENSION TO SOIL SYSTEMS
Extension of the alkaline glycolate technology to solid phase media has been slow to
develop. As early as 1982, workers from Franklin under contract to the EPA applied
1 It is generally accepted that aerobic bacteria do not attack heavily chlorinated PCBs and dechlorinate the
others impractically slowly if at all. Although it has been shown that anaerobic bacteria can partially dechlorinate
fully chlorinated Aroclors, most of these are not, to our knowledge, facultative anaerobes, ruling out in situ use and
mixed cultures with aerobes. Anaerobic processes are also quite slow and cultures are notoriously unstable. Practical
methods for working with mixed cultures of aerobes and anaerobes simply do not exist. Products of reaction are
unknown.
219
NaPEG to soil. At this time, however, it was still hoped that an in situ chemical
dechlorination could be developed. Results were unacceptable, for reasons having to do
with the fundamental irrationality of trying to carry out synthetic stoichiometric chemistry
in unmixed solid phase media exposed to uncontrolled conditions of temperature and
dilution, ie., rain. Nevertheless, the idea that alkaline glycolate does not work in soils
gained momentum.
A series of field trials sponsored by the Environmental Protection Agency through the
70's extended the glycolate chemistry to a wider range of contaminants and contaminant
levels, although reporting of the results frequently confounded site specific and other non-
chemistry factors with the efficacy of the reagent system. Also, because the experimental
approach was evolving from earlier in situ efforts reaction temperature was often limited
by equipment constraints.
Finally, in 1988, EPA sponsored a field test in Guam (Taylor,etal., 1990) which
showed the chemistry to be capable of achieving any contemplated level of remediation in
a soil system. In reporting these results, however, no differentiation was made between the
costs incurred in the demonstration unit and the costs of full scale remediation which could
be reasonably inferred from the successful test.
This history left a residue of skepticism as to the efficacy of alkaline glycolate systems
and the attendant cost. Clearly what is needed to fully understand the chemistry is a body
of data on model systems. Why was this not done long ago? The answer lies in the system
by which technology is selected for Superfund sites which relies on what are called
Treatability tests. These tests are site specific and the emphasis is on quick demonstration
of efficacy, both to rule out totally inappropriate technologies and because the
environmental arena carries an undercurrent of thinking that the chemistry is secondary to
the engineering and to the hardware to be used. Doing research under the Treatability
system is nearly impossible.
Finally, a model system study was carried out at the National Environmental
Technology Applications Corporation (NETAC) (King and Hoch, 1991). Because these
data are not readily available, it is useful to review the highlights here.
NETAC PROGRAM
This privately funded contract program had as its objective identification of process
conditions for economical and total dechlorination by determining the rate of removal of
chlorine from model aromatic compounds. In contrast to many other environmental
studies, this work determined the total amount of chloride ion released as a function of
reaction temperature, time and stoichiometry. This approach differs from the usual
environmental tests which measure the fraction of the chlorinated aromatic that has reacted
- presumably measuring therefore only the first chlorine removal.
Several distinct reactor configurations were employed. At low temperature, i.e. less
than 150" C, and below the boiling point of the solvent, a commercial stirred heater block
using six dram vials was used. At higher temperatures, but below the boiling point of the
reaction solvent, the heater block was placed on a hot plate or a three neck flask with
nitrogen purge was used. For the highest temperature runs a 3/16 stainless steel high
pressure bomb immersed in a fluidized sand bath and agitated by an internal stainless steel
ball served as the reactor.
After reaction, the mixtures were quenched by acidification, diluted with water and
chloride released was measured using a specific ion electrode. These data represent the
only systematic and self-consistent published study of the dechlorination of well
characterized aromatics by alkaline glycolate chemistry.
220
TABLE I. HEXACHLOROBENZENE-KPEG RUNS
Temp Time moles moles Solvent Conversion

degC hr PEG/C1 KOHlCl
25 16 4.01 4.40 Toluene 51.8
60 2 3.91 3.87 Toluene 52.3
60 4 3.89 3.90 Toluene 61.3
60 8 3.90 3.86 Toluene 57.1
60 16 3.89 3.90 Toluene 62.7
80 16 4.00 4.38 Toluene 58.2
80 16 4.01 4.39 Toluene 54.7
80 16 2.01 2.22 Toluene 54.6
80 2 3.99 4.61 Toluene 61.0
100 4 1.61 1.80 Toluene 68.09
100 2 2.53 2.76 Toluene 64.6
147 16 3.92 4.35 Triglyme 93.7
175 2 4.05 4.34 Triglyme 101.6
165 2 4.02 4.36 Triglyme 86.4
175 2 3.90 4.27 Triglyme 82.7
200 2 3.96 4.30 Triglyme 100.7
190-195 2 3.93 4.32 Triglyme 54.01
195-200 1 4.09 4.48 Triglyme 92.0
2 3.98 4.38 Triglyme 104.8
181-185 1 3.84 4.20 Triglyme 74.5
2 3.93 4.30 Triglyme 96.2
185 1 1.35 1.48 Triglyme 81.1
150 16 1.50 1.67 Triglyme 113.4
200 2 3.96 4.51 None 103.9
}<'rom 1I..lng ana lioon ~!Il
Table I summarizes hexachlorobenzene dechlorination data in various solvents. Note

that these data are given in percent of total chlorine removed; as little as 16 percent
conversion may, therefore, correspond to essentially complete disappearance of the starting
material.
A strong correlation can be seen between the achievable degree of dechlorination and
reaction temperature. Loosely speaking, 50 percent dechlorination is found over-night at
room temperature. The fourth chlorine atom can be reacted in two hours at 80"C; the fifth
requires 150°C and complete dechlorination is observed in the 185° to 200"C temperature
range. Data also show that the effect of reagent stoichiometry is not dominant.
Table II presents data on dechlorination of less substituted chlorobenzenes. Generally
speaking these less chlorinated materials require more stringent conditions to achieve the
same degree of dechlorination. A similar effect was reported for PCBs (Brunelle and
Singleton, 1985)
Again relating the conditions required for complete dechlorination to reaction
temperature, 1,2,4 trichlorobenzene requires about 230"C for complete dechlorination, 1,3
dichlorobenzene requires about 250°C and chlorobenzene requires even higher
temperatures.
This data shows clearly why earlier attempts to dechlorinate materials like
pentachlorophenol, which is deactivated to nucleophilic substitution, at moderate
temperature conditions is futile even given extremely long reaction times. On the other
hand, the seemingly more stringent goal of complete dechlorination can be readily achieved
provided that the system is taken to an adequate temperature.
221
TABLE II. REACTION WITII CHWROBENZENES
Reactant Temp Time moles moles Conversion

deg C hr PEG/Cl KOHlCl
TriClB 200 2 4.00 4.58 67.3
TriCIB 225 2 4.00 4.56 82.5
TriClB 230-240 2 3.98 4.57 102.5
TriCIB 230 1 3.97 4.56 89.9
TriC1B 240 1 3.97 4.53 109.4
TriC1B 250 1 3.98 4.47 72.3
TriC1B 240 1 3.73 4.27 122.1
TriC1B 250 1 4.04 4.55 90.9
DiC1B 200 2 4.26 4.46 42.3
DiC1B 225 2 4.49 4.67 86.5
DiC1B 250 2 4.54 4.70 95.8
ChloroB 200 2 7.77 8.31 52.5
ChloroB 225 2 7.47 7.79 63.3
ChloroB 250 2 7.75 8.38 62.3
ChloroB 275 2 7.41 7.68 40.3
f'rom .King and Moch 19S

All runs: triglyme solvent
Furthermore, given the temperatures required, issues surrounding the presence of

water become moot. A back of the envelope calculation easily shows that the energy cost
to dehydrate even the wettest sediments is negligible compared to the normal costs of
remediation. Equipment configured to obtain the temperatures required for dechlorination
is easily capable of simultaneously or sequentially dehydrating the feed.
PROCESS AND ECONOMIC CONSIDERATIONS INTHEDECHWRINATION OF

SOILS, SLUDGES AND SEDIMENTS
The key to acceptable economics in soil systems lies in complete dechlorination.

Anything less leaves a material of unknown toxicity. Despite the previously cited studies
showing these polychloromonoethers to be nontoxic there remains a quite reasonable
resistance to their persistence in the environment. If, therefore, complete dechlorination
is not achieved the chemistry results in a product still requiring treatment as a potentially
hazardous substance. Complete dechlorination, on the other hand, can be coupled with
engineering steps that leave the reaction products in the treated soil. One can argue for
on site reinterment of this organic chloride free material. On site reinterment is absolutely
critical to competitive economics.
Despite its successful history there remained skepticism that alkaline glycolate
chemistry could be successfully extended to soil systems. This skepticism was rooted in
both the traditional difficulties of the remediation industry in translating chemistry directly
into hardware without an intermediate process engineering step as well as in the confusing
results of the EPA studies through the 70's.
All of this was swept away in 1990 and 1991 when the Wide Beach, New York,
Superfund Site was successfully remediated using alkaline glycolate chemistry in a thermal
desorber. It is hard to overstate the significance of the Wide Beach project. Not only did
it provide a full scale and fully successful demonstration of alkaline glycolate chemistry
in a soil system, but it represented the first use of an innovative, which in the parlance of
the remediation industry means non-incineration, technology for PCBs at a superfund site.
222
Also significant is the fact that the chemistry was readily integrated into an existing
thermal desorber. This is significant both because the availability of such desorbers
provides additional systems wherein one might employ alkaline glycolate dechlorinations
and further because operation of a desorber on a chlorinated aromatic carries a significant
risk as explained above with reference to Scheme One.
It should be noted that other reductive technologies for destruction of PCBs have been
proposed. These include the reaction with monoethylene glycol monomethylether
(MEGME), the base catalyzed decomposition (BCD process) developed by the EPA and
reduction by nascent hydrogen developed by Roy F. Weston for the EPA. Beside lacking
the track record of the PEG process, each of these approaches has severe handicaps.
Although, based on published data (Friedman, 1990) and reasoning from first
principles, MEGME appears to be a facile dechlorination agent, it suffers from the serious
drawback of being itself toxic. As a practical matter, this rules out direct application in
soil systems followed by reinterment of the treated soil. As explained above such on site
reinterment is critical to the economics.
Although the chemistry underlying the BCD process has never been made entirely
clear, what has been published suggests that it may suffer from all of the drawbacks
associated with the presence of water that have been attributed in the discussion above to
desorption in the absence of NaPeg. Furthermore, flow sheets for BCD which have been
presented (Kim and Olfenbuttel, 1990) are among the most egregious examples of
technology development without regard to process engineering.
Although only limited data is available (Saha, etal., 1990) it is hard to understand how
the nascent hydrogen reduction system can be successfully applied to soils. This is another
example where elementary process engineering will clearly rule out any practical
application.
ECONOMICS OF REMEDIATION
Having argued that not only are alternative remediation technologies technically or
socially unacceptable and further having shown that the fundamentals of the chemistry as
well as a body of organized data support the choice of alkaline glycolate dechlorination,
it still remains to demonstrate that this technology meets the primary criteria for
remediation, viz. acceptable economics. Table Three below shows illustrative economics
for a model case. The underlying data for this table are not arbitrary; these figures reflect
actual results on a superfund site soil tested at NETAC. Note that these are shovel-to-
shovel economics. Site specific factors are thereby excluded.
TABLE ill. ECONOMICS OF TYPICAL SOIL REMEDIATION USING KPEGsm

25000 TONS DRY WEIGHT
Cost Element $/ton
Reagents and other chemicals 86

Utilities 25
Labor 12
Mobilization/demobilization 7
Fixed costs, including royalty and return 103
Total shovel-to-shovel cost 233
ay mo 1 e urnt;
3 Hr. reaction time;
3-fold stoichiometric excess of KPEG
223
It should be noted that this approach of comparing shovel-to-shovel numbers,
analogous to use in the chemical industry of the battery limits concept, although not
unknown is not stressed in the hazardous waste community. As a result, site specific
factors creep into economic data which is imbued with almost thermodynamic universality.
It is fair to say that differences in accounting methods, project basis and site specific
factors are far more variable than the difference between the economics of ex situ
processes.
Depending upon the site, the cost elements considered in Table Three mayor may not
comprise the bulk of the total project cost. These are process economics, not project
economics. Cost is seen to be dominated by "reagents and other chemicals" and "fixed
costs" .
The reagent consumption is, perhaps surprisingly, not a strong function of the type
or degree of contamination. Rather, it represents uptake of alkali by the soil media itself.
To the extent that representatives samples can be obtained, it can be determined by prior
experimentation in fact, determination of factors such as this is the rational basis for the
Treatability concept. Experience has shown that alkali uptake can vary widely and can in
fact become so large that it forces one to consider other remedies. Many of these,
however, may also be adversely affected by a soil or other media with high alkali uptake
potential. It is important to make such a technology comparison using consistent data.
Fixed costs, which in the above accounting are all related to unit capital, its
amortization and reward, can also vary widely. Here, the reason for inconsistency is not
site specific but vendor specific. Neither EPA rules, nor industry practice, provides any
guidance on how to charge the capital cost of a mobile unit against specific projects. This
reflects the project by project orientation of the remediation industry where, in the extreme
case, a contractor could assign a fixed cost of zero. Unfortunately, unrealistically assigned
fixed costs are not uncommon. Not only does this always lead to disastrous results for the
project at hand, but such numbers find their way into the shoptalk of the industry and by
Gresham's law of economics drive the realistic numbers out of circulation.
The figure of $103.00 in Table Three reflects the assumption that the 250 ton/day
mobile unit stands idle 50 % of the time and is either remediating or being mobilized or
demobilized the other 50 %. Based on this "activity factor" normal capital weights and 25 %
simple return are included.
The utilities cost figure needs little discussion; it reflects fuel at $ 3.oo/MMBTU and
soil moisture of approximately 30 weight percent. It is easy to calculate, then that were the
soil water level as high as 90 weight percent utilities cost would rise by less than $5/ton.
Although one often hears about the adverse effect of soil moisture on dechlorination, this
simple calculation shows that, in systems where the soil is dried before reaction, the
utilities cost penalty is relatively minor.
Both the labor cost and the mobilization-demobilization cost, although minor, are
related to the size of the unit employed. Matching or mismatching of process unit size to
project size can be a further source of strange economic computation. Clearly, the larger
the unit employed, the lower the labor cost becomes and the lower certain elements of the
fixed cost become. On the other hand, mobilization-demobilization rises and assumptions
as to unit utilization become more critical. Assuming the availability of an unreasonably
large unit becomes an alternative path by which the unwise or the unscrupulous can
compute extremely low project costs.
A point should also be made in this section on the comparison between dechlorination
and separation technologies like extraction with various solvents and soil washing. All of
these methods in common concentrate, but do not destroy, the contaminant. Furthermore,
if the process is not very selective, a number of different contaminated media may be
generated. Additionally, driving forces for separation are smaller than for reaction, making
strict clean levels more difficult to reach. Finally, integration of many of these separations
with alkaline glycolate dechlorination is feasible
224
RECAPITULATION
Based on all of the above one can argue that alkaline glycolate dechlorination is the
process of choice for remediation of chlorinated soils, sludges and sediments. The simple
argument is based on the heirarchy of criteria listed in the "Overview" section above and
the technical factors reviewed in this paper. Key points of the logic are the following:
1) Incineration will not be publicly acceptable as well as having serious technical

shortcomings.
2) Biotechnology has the same pair of problems.
3) Other oxidative systems and high temperature anaerobic systems run a significant risk
of producing dioxins or similar compounds.
4) Landfill should be conserved for wastes that cannot be treated or for mixed wastes
that cannot be treated simply and economically.
5) Separation systems merely concentrate the contamination and produce a contaminated

product which still requires treatment and/or disposal.
Finally, in this quite conservative industry, the track record of KPEG""', culminating in the
Wide Beach Project, is simply unmatched.
REFERENCFS
Altwicker, E.R., 1988, Formation of polychlorodibenzo-p-dioxins and

polychlorodibenzofurans during heterogeneous combustion, 3rd Chemical Congress of
North America and 195th ACS National Meeting, Toronto
Altwicker, E.R. etal., 1989, Polychlorinated dioxin/furan formation in incinerators,

Presented at the AlChE Annual Meeting, San Francisco
Berry, R.I., 1981, Chem. Eng., August, 10, p37
Brunelle, D.J.,1982, U.S.Patent 4,351,718
Brunelle, D.J.,1983, Destruction/removal of polychlorinated biphenyls from non-polar

media, Chemosphere,12:183
Brunelle, D.J. and Singleton, D. A., 1985, Chemosphere, 14:173
Davis,F.F.,etal.,1979, U.S.Patent 4,179,337
Dellinger, Barry, etal., 1990, PIC formation - research status and control implications,
Presented at EPA 16th Annual Hazardous Waste Research Symposium, Cincinnati
DeMarini,D.M., and Simmons,J.E., 1989, Chemosphere, 18:2293
Friedman,A., 1990, paper presented at 10th International Coriference on Organohalogen

Compounds, Bayreuth
225
Kim,B.C. and Olfenbuttel,R.F., 1990, Demonstration of the BCDP process at USN PWC
site in Guam, presented at EPA Technology Transfer Conference, Cincinnati
King,A.B. and Hoch,R., 1991, Complete dechlorination of aromatics with alkaline PEG
chemistry, prepared for Air & Waste Management Association Meeting, Vancouver
Lane, A.M. and Hinton, S.W., 1990, Characterization of municipal waste incinerator fly
ash to identify mechanism of dioxin formation, Presented at AIChE Spring National
Meeting, Orlando
Lewis, C.W., 1990, Measure remote temperatures efficiently, Chern. Eng. 97:114
Oppelt, E.T., 1986, Performance Assessment of Incinerators and High Temperature

Industrial Processes for Disposing of Hazardous Waste in the United States, in Lorenzen,
etal. eds. "Hazardous and Industrial Solid Waste Testing and Disposal", ASTM # 933,
Philadelphia
Oppelt, E.T., 1987, Incineration of hazardous waste; a critical review, JAPCA 37:558
Pytlewski,L.L., etal., 1982, U.S.Patent 4,337,368
Pytlewski, L.L., etal.,1983a, U.S.Patent 4,400,552
Pytlewski, L.L., etal.,1983b, U.S.Patent 4,430,208
Pytlewski, L.L., etal.,1984a, U.S.Patent 4,417,977
Pytlewski, L.L., etal.,1984b, U.S.Patent 4,460,797
Pytlewski, L.L., etal.,1984c, U.S.Patent 4,471,143
Pytlewski, L.L., etal., 1986, U .S.Patent 4,602,994
Saha,A.K. etal., 1990, Thermocatalytic hydrodechlorination of chlorinated pesticides left

in soil medium, effect of temperature, catalyst, acid, and solvent on conversion, presented
at AIChE Summer National Meeting, San Diego
Sethi, V. and Biswas, P., 1990, Fundamental studies on particulate emissions from
hazardous waste incinerators, Presented at EPA 16th Annual Hazardous Waste Research
Symposium, Cincinnati
Shaub, W. and Tsang, W., 1983, Dioxin formation in incinerators, Env. Sci. Tech. 17,
721-730
Siebert, P. C., etal., 1989, Toxic trace pollutants from incineration, Presented at AIChE
Summer National Meeting, Philadelphia
Stumbar, J.P. etal., 1990, Factors affecting the reliability of operations of the epa mobile
incineration system, Presented at EPA 16th Annual Hazardous Waste Research Symposium,
Cincinnati
226
Taylor, M.L. etal., 1990, "Comprehensive Report on the KPEG Process for Treating
Chlorinated Wastes", EPA Contract No. 68-03-3413
Vogg, H. and Stieglitz, L., 1986, Thermal behavior of PCDD/PCDF in fly ash from
municipal incinerators, Chemosphere, 15:1373
Zeng, Y. and Okrent, D., 1989, Analysis of off-normal emissions from a hazardous waste
incinerator, Presented at the SRA Annual Meeting, San Francisco
227
A NEW MEMBRANE PROCESS FOR RECOVERING
ORGANICS FROM AQUEOUS WASTES
S. V.Ho
Monsanto Company
800 N. Lindbergh Blvd.
St. Louis, MO 63167
INTRODUCTION
Membmne technology offers great potential for treating industrial wastes. Membmne
methods that could potentially be applied to wastewater treatment have been reviewed
extensively in a recent Department of Energy report on membrane separation systems l and by
various workers in the field2.3. These methods include reverse osmosis, nanofiltration,
ultrafiltration, microfiltration, coupled/facilitated transport Oiquid membmne), pervapomtion,
and electrodialysis. Membranes can also be used as contactors to carry out gas stripping and
liquid-liquid extraction4 •5 •
A common characteristic of most aqueous wastes discharged from chemical plants is
that they are fairly dilute in organic content. These streams typically contain up to several
percent low molecular weight (MW) organic compounds such as carboxylic acids, alcohols,
nitriles, etc. Most streams also contain inorganic salts with concentrations that can reach 30
wt% or so. Examples of aqueous wastes that belong to the high-salt category include one that
contains -5% highly-charged organics and 20% NaCI; another containing 2% various
organics of MW<200 and 6-35% ammonium sulfate; and one stream containing 1%
phenolics and 15-20% potassium chloride. From a separation standpoint, a plausible initial
treatment of these streams is to remove and/or concentrate the organics either for recycle or
to facilitate further treatment. In principle, dilute streams of organics can genemlly be handled
adequately with existing methods such as reverse osmosis or evaporation to reduce the
volume, followed by incineration. However, the presence of high concentrations of salt
greatly complicates treatment options. For instance, reverse osmosis is not practical for
treating these streams due to the very high osmotic pressures involved, and existing
nanofiltration membranes generally do not have sufficient rejection for organic compounds
with MW below 200 or so. As a result, the costs for treating such streams are usually much
higher than for similar ones with no salts present.
Using internal waste streams as models we have developed proprietary composite
membranes with unique properties that enable us to concentrate the low MW organics as well
as separate them from salts. These membranes operate in a dialytic mode where the driving
force is concentration and solubility difference rather than pressure difference. Results
obtained with model compounds as well as with two actual waste streams are presented in
this paper.
Industrial EnWoIlllle1l/al Chemistry. Edited by D.T. Sawyer

and A.B. Martell. Plenum Press, New York, 1992 229
EXPERIMENT AL PROCEDURES
All the chemicals used in this study, other than those in actual waste streams, were
purchased from either Sigma or Aldrich. The actual wastes streams were obtained from
Monsanto plants. Phenol, nitrophenol, nitroaniline, benzoic acid as pure components in
aqueous solutions were quantified using spectrophotometric absorption at wavelengths 287,
400,380, and 224 nm, respectively. Total organic content (TOC) was measured with a TOC
machine (Rosemount Analytical, Dohrmann DC-190. Carboxylic acids present as single
components were analyzed either by titration with O.lN NaOH (phenolphthalein indicator) or
using the TOC machine. As mixtures, they were analyzed using ion chromatography, which
was also the method used for measuring chloride ions in solution. Solution conductivity was
measured with a Cole-Parmer conductivity meter.
Separation experiments were carried out using both flat sheet and hollow fiber
membranes. These are composite membranes consisting of a proprietary separating layer
supported on various microfiltration and ultrafiltration membranes. Due to the patent
situation, we are not allowed to disclose the nature of the membranes at this time. Flat sheet
microporous membranes from Hoechst-Celanese (polypropylene, trade name Ce1gard) and
from W. L. Gore (polytetrafluoroethylene, trade name Gore-Tex) were used in this study.
Some physical characteristics of these membranes are shown below6.7 •
~ Thickness Pore Size % Porosity

Celgard 2400 25 11m 0.050 11m 38
Celgard 2500 25 0.075 45
Gore-Tex 60 0.100 75
Two types of hollow fibers were also used: polypropylene (Celgard) from Hoechst-
Celanese, and polysulfone from Permea (now a subsidiary of Air Products). The
polypropylene hollow fibers are symmetric, i.e., the pore structure is the same throughout
the wall of the fibers. Polysulfone hollow fibers are asymmetric, that is, they have a very
thin, dense skin on the surface, supported by a much thicker and more open structure
underneath. Physical characteristics of these fibers are shown below.
Celgard X20 400: symmetric; OD = 460, ID = 400 IJ-m; dporc = 0.065 IJ-m;
porosity = 40%
Polysulfone : asymmetric; OD = 1100, ID = 866 11m; 160K MWCO;
porosity =76%
These microfiltration hollow fibers were potted into our standard testing modules
(polycarbonate tubes, 1 em inside diameter and 25 cm long with openings on the side of the
tube for shell-side fluid circulation), then modified in situ to introduce the separating layer
forming the proprietary composite membranes. The membrane area available for transport
depends on the number of fibers used; typically the area ranged from 100 to 400 cm Z in our
tests.
Figure 1 shows the schematic diagram of various experimental systems used in this
study. Transport measurements with flat sheet membranes were conducted using either
stirred cells or flow cells. Stirred cells are made up of two glass compartments, each about 30
ml in volume, separated by a sheet of membrane and held together by a clamp. Liquids in the
compartments were magnetically stirred. Membrane area in contact with the liquids was 8
cm z. These stirred cells are convenient for rapid membrane screening as well as for
determining the effects of organic concentration and temperature on membrane transport.
Flow cells are similar to stirred cells except that each compartment has an inlet and an outlet
port for circulating liquid through the cells. Flow cells were used for processing large
volumes of liquid, which were circulated with pumps from external reservoirs through the
cell compartments. With hollow fiber membranes, fluids were circulated through the bore
side and the shell side. Organic concentrations of the feed and strip solutions were measured
230
Stirred Cell
Flow Cell
Figure 1. Schematic diagram for stirred cell. flow cell. and hollow fiber
membrane systems
231
as a function of run time. The data were used to calculate the overall mass transfer coefficient
using the equations derived in the appendix.
RESULTS AND DISCUSSION

Rate, selectivity, and stability are three critical elements in determining the value,
cost/performance and ultimate utility of a membrane for any separation. Permeability and
selectivity are determined by the solubility (interaction) and diffusivity of each component in
a system. A general description of the properties of the composite membranes is given
below, followed by examples of the application of the technology to waste minimization.
PROPERTIES OF COMPOSITE MEMBRANES

The transport through our composite membranes is by chemical interaction, not by size
difference. A key property of these membranes is that they effectively rejects charged species
in solution. This includes both small ions (ions generated by salts, H+ and OH-) and larger
charged molecules. Only neutral compounds can pass through. Thus, the membranes can
be used to recover organics from an aqueous saline waste as illustrated in Figure 2. For
effective separation across such a membrane, a driving force should be provided and the
recovery (strip) solution should provide an environment (a "sink") which limits the back
diffusion of the compound across the membrane into the waste solution. This could be
accomplished in several ways including pH change, reaction, chemical complexation,
biodegradation, pervaporation, etc.
To date small ions which have been shown to be effectively rejected by the membranes
are H+, OH-, Na+, K+, Cl-, NRt+, NOr, S04=. We have not yet found an ion that could
readily pass the membrane. This is true for larger charged molecules as well. Data in Table 1
for PNP and formic acid show the significance of the ionization state of a compound on its
transport rate through the membrane. At solution pH several units above their pKa's,
practically all PNP and formic acid molecules are ionized, resulting in a 20-40x reduction in
transport rates across the membrane compared to the rates obtained when these compounds
exist predominantly in their neutral forms (pH < pKa).
Table 1. Effect of solution pH on transport of organic compounds for composite

Celgard 2400 membranes at room temperature.
Compound pKa Solution pH Initial Rate

(mg/hr-cm2 membrane)
p-Nittophenol 7.0 10 0.15
(at 0.5 wt%) 5 3.5
Formic Acid 3.75 5 0.026

(at4.4wt%) 1.6 1.0
The rate of transport through the composite membrane for non-ionized compounds is
very sensitive to their chemical nature. Table 2 shows the transport rates for various
compounds such as phenol, amines and carboxylic acids. The rates were obtained with the
stirred cell and are reported as the overall mass transfer coefficient (units: em/sec), which
lumps mass transfer resistances of the aqueous liquid films on both sides of the membrane
with the resistance due to the membrane itself. A mathematical model of the stirred cell
developed in the appendix shows that plot of In (concentration change) vs. time yields a
straight line, the slope of which is proportional to the overall mass transfer coefficient, Kf.
Such a linearized plot is presented in Figure 3 for phenol. Due to the intentional vigorous
liquid mixing in the cell compartments, which minimizes the resistance contributions of the
232
Microporous Membrane
with Separating Layer
Waste Solution Strip Solution
Salts
Organics --+7'~~4----1~
e.g. with NaOH as strip solution forming organic salts
Membrane
Organics
_--+_+~ , / NaOH
Na-Organics
Salts
(e.g. CI- Na+
Na+ OH-
K+) H+
Figure 2. Composite membrane for organic/salt separation
233
Table 2. Transport of organic compounds through a composite Ce1gard 2400
membrane at room temperature.
Compound Initial Feed Con. Overall Mass Trans. Coeff.

mg/L cm/sec
Phenol 4000 6.0 x 10-4
p-Nitrophenol 4000 3.4 x 10-4
p-Nitroaniline 500 6.5 X 10-4
Benzoic acid 1700 4.0 X 10-4
Strip solution: O.lN NaOH (O.iN Helfor p-nitroaniline)
Table 3. Phenol permeability through various polymeric membranes <a)
Polymer Film thickness Effective diffusion

11m coefficient, cm 2/sec
Polyethylene 14.1 0.11 x 10-7
Polyvinylidene chloride 10.0 0.06 x 10- 7
Styrene-butadiene copolymer 41.0 0.71 x 10-7
XD-7 (Silane) 68.6 2.()6 x 10-7
Ethyl cellulose 35.2 0.96 x 10-7
Composite Celgard 2400(b) 25.0 15.0 x 10-7

(a)from Klein et al. 8; (b) This study
liquid boundary layers, the values obtained for Kf are essentially the membrane permeability
values. In these experiments, the pH of the feed solution was such that the organic
compounds are un-ionized (pH about 2 units below pKa) and NaOH was used on the strip
side for weak acids, and HCI or HN03 for weak bases.
The practical significance of the overall mass transfer coefficients obtained with the
composite membrane shown in Table 2 can be appreciated by comparison with other
membrane approaches for phenol transport. Table 3 shows the results obtained by Klein et
al. 8 for phenol extraction using various polymeric films as membranes. Transport rates were
reported as effective diffusion coefficient (D in cm 2/sec = permeability x membrane
thickness) for phenol through these membranes. With our composite membrane (251-1m
thick) the permeability of phenol is 6.0 x 10-4 cm/sec (Table 2), which gives an effective
diffusion coefficient of 15 x 10-7 cm 2/sec. This is almost an order of magnitude higher than
the best rate achieved by these workers (D = 2.06 x 10-7 cm 2/sec) with the membrane made
from XD-7 (a polymethylsilane-polycarbonate copolymer). Another interesting comparison is
with the hollow fiber contained liquid membrane approach by Sengupta et a1. 9 . In this
approach, the shell side of a membrane module that contains two sets of intermingled hollow
fiber bundles is filled with an organic solvent that functions as the liquid membrane. The
feed solution flows through the lumen of one set of fibers, and the strip solution through the
other. With methyl isobutyl ketone as the contained liquid membrane, they reported an
overall mass transfer coefficient for phenol of about 3 x 10-4 crn/sec. This is about half of the
rate obtained with our composite membrane, which, incidentally, would utilize all of the
hollow fiber membrane area available for transport rather than only half of it as in the
contained liquid membrane approach.
Transport rates through a composite membrane for a series of carboxylic acids ranging
from C1 to C6 are shown in Table 4. Interestingly, the rate increases two orders of
magnitude from formic acid (C1) to caproic acid (C6). Thus, hydrophobicity and
permeability seem to be closely correlated and the more hydrophobic compounds enjoy faster
transport rates. We also found that in addition to hydrophobicity certain polar characteristics
(e.g. hydrogen-bonding capability) of an organic compound may also contribute to its
enhanced transport rate.
234
10000~-----------------------------------------'
-I
a,
E
'Ti
~
u..
.E 1000
C
o
o
"0
c:
Q)
.c:
Q.
100+-----~--_.----~----r_--~----_r----~--~
o 2 3 4
Run Time, Hr.
Figure 3. Linearized plot for phenol transport through a composite Celgard
membrane. (Feed: 4200 mg/L phenol, pH 5.8; Strip: 0.1N NaOH)
4000 . .
,----------------------------------------~
3800 ppm
3000 "'" """""'" Was,.
',.
-I .............
a,
E ............... . ........ .
cL
2000
l:-.<
....
Z ..................... ' ............
Q.
~~' p
1000
......!.................. i
3
0 . .~~--~--~--~--~--~--~--~~47--A
o 2 4 6 8
Time, hr.
Figure 4. Removal of PNP from an actual phenolics waste stream using
a composite Celgard 2400 membrane
235
Table 4. Transport of carboxylic acids through a composite Celgard membranes
Compound Overall Mass Transfer Coeff.

(em/sec)
Formic Acid (Cl) 0.055 x 10-4

Acetic Acid (Cv 0.10 x 10-4
Propionic Acid (C3) 0.33 x 10-4
Butyric Acid (C4) 0.42 x 10-4
Valerie Acid (Cs) 1.6 x 10-4
Caproic Acid (C(;) 4.0 x 10-4
*Feed side: O.I-5wt% acid in 2Owt% KCl.pH = 2; Strip side: O.IN NaOH; Room Temperature
Effects of the number of hydroxyl groups on phenolic compounds were explored with
resorcinol, hydroquinone, pyrocatechol, and phloroglucinol. More hydroxyl groups attached
to the ring make the compound more hydrophilic, which results in lower transport rates.
Also very interesting is the behavior of amino acids and its analogs. Being zwitterions, which
are always ionized, amino acids even the most hydrophobic ones such as tryptophan,
phenylalanine and tyrosine are totally rejected by the composite membrane. Yet
hydrocinnamic acid, which is phenylalanine without the amino group, can be transported
very fast (comparable to p-nitrophenol) through the composite membrane at low pH where it
is non-ionized.
The above properties of our proprietary composite membranes make them very suitable
for organic/salt separation as well as for separating similar organic compounds that may be
difficult to achieve with other separation methods
APPLICA TIONS TO WASTE TREATMENT

The composite membrane process has been used successfully to remove and concentrate
organics from two actual aqueous waste streams. One contains phenolics and high
concentration of inorganic salts; the other is a dilute stream of carboxylic acids.
Phenolics Waste
This waste stream is generated from one of our plants and typically contains p-
nitrophenol (PNP) as the major organic contaminant at <1 wt%. The waste stream also
contains high salt level (-20wt%, mainly KCI). Existing treatment options for this stream
include PNP removal followed by destruction or biodegradation. We investigated the
feasibility of using our membrane process to remove PNP from the waste solution in a
concentrated form for recycle.
Figure 4 shows the results for a run with the actual waste solution using a flat sheet
composite Celgard membrane_ In this experiment the membrane was in contact with two
circulating solutions: the waste solution at pH 5 on one side, and O.IM potassium hydroxide
(pH 13) on the other side. PNP concentration on the waste side steadily declined and
dropped to about 3 ppm overnight, whereas PNP level on the caustic side rose to about 3800
ppm, representing a concentration ratio of over 1000 on the strip side over the waste side.
Based on conductivity measurements and chloride ion analysis, only about 0.3% of the salts
in the waste passed through the membrane despite the very long contact time allowed. pH
values for both solutions were fairly constant throughout the run suggesting no significant
transport of H+ and OR- ions. Water transport across the membrane after several days of
operation was not noticeable. These results show that the membrane was quite remarkable in
its ability to maintain a large pH differential and significantly block the water transport under
236
Table 5. PNP recovery from a phenolics waste'
PNP (mglL) Chloride Overall Mass Transfer

(at equilibrium) Coeff. (em/sec)
3.4 x 10. 4
Strip side 3788 0.049 gIL

Waste side 3.1 100 gIL
• T = 6()OC 7.0 X 10. 4
Strip side 4042 0.31 gIL

Waste side 3.7 100 gIL
'Membrane: Composite-Celgard 2400; Waste solution: 4000 ppm PNP. 20% KCI. pH 4.5: Strip
side: 0.1 N NaOH
the tremendous osmotic pressure difference of several thousand psi exerted by the 20% salt
solution while efficiently transporting PNP across. Table 5 summarizes the results obtained
at 25 and 60o C. The high temperature doubled the rate without an adverse effect on salt
separation and pH differential across the membrane.
Results for PNP/salt separation using composite hollow fiber membranes prepared with
polypropylene (Ce1gard) and polysulfone (Permea) hollow fibers as the supports are shown
in Figure 5 & 6. The hollow fiber configuration is preferred in industrial applications due to
its more compact form (higher membrane area per unit volume of packing). Hollow fiber
membranes are similar to flat sheet ones in their ability to remove PNP from the saline waste.
Note in Figure 6 that removal and simultaneous concentration of PNP was accomplished in
this batch operation simply by using a strip volume one-tenth that of the feed. In continuous
counter- or co-current operation, the relative flow rates of feed and strip solutions determine
the concentration factor.
Effects of the types of membrane support and their structure on PNP rates of transport
through the composite membranes are shown in Table 6. Of the two different hollow fiber
supports, it was interesting that while the poly sulfone membrane was over 4 times thicker
than the Celgard membrane (140 11m vs. 30 11m), its PNP transport rate was somewhat
comparable to that obtained with the Ce1gard. The asymmetric nature of the polysulfone
hollow fiber support along with its significantly higher porosity probably compensate for the
thicker wall.
Based on the rate data obtained with a polypropylene-based (Celgard X20-400 from
Hoechst-Celanese) composite hollow fiber membrane, we estimated that a membrane area
Table 6. Effects of support characteristics on PNP transport •

Membrane Support NaOHCon. OveralJ Mass Transf.
Coeff. (em/sec)
• Flat Sheet
Celgard 2400 (polypropylene) 0.1 M 3.4 X 10.4
0.5 M 3.5 x 10- 4
• Hollow Fibers
Polysulfone 0.1 M l.7 x 10-4

Celgard X20 400 0.1 M 2.0 x 10- 4
'Feed: 4000 mglL PNP in 20% KCI solution. pH - 4.5; Membrane: composite membranes with
various supports; Strip solution: NaOH; room temperature.
237
5000
4000
3000
-I
g,
E 2000
Il."'
zQ. 1000
Time, Hr.
Figure 5. PNP removal with a composite hollow fiber membrane. (Membrane module:
Celgard X20-400 fibers, area = 122 cm2; Feed: 4500 ppm PNP, 20% KCI; Strip: 0.1N NaOH)
20000~------------------------------------------~
..........
.....
16000
" ............ "
.
E
8: 12000
............
.............. Strip side
c ". ...
8
11.
Z
11.
8000
,. "
...
.... ""oil
4000.' bEEtl-"r-~1:I"" " Feed side

.. D '1:r"'o",o"
~.. .~ ••• £i ••• ~ •••• ~ •• _t3 •• _JJ
O~------------.-------~-----r------~----~
o 5 10 15
Run Time, hr
Figure 6. PNP removal with a composite polysulfone hollow fiber membrane.
(Membrane module: fibers from Permea, area 120 cm2; Feed: 4000 ppm PNP,
20%KCI, lumen side, 1 liter volume; Strip: 0.1 N NaOH, shell side, 0.1 liter volume)
238
about 15,000 ft 2 is required to remove over 99.99% of PNP in a 15 gal/minute waste stream
at 60°C (to get below the 200 ppb needed to meet the regulation discharge). On the basis of 6
to 10 dollars per ft 2 of membrane installed cost10, the required membrane area translates into
a maximum of $150,000 for the entire membrane module. This is contrasted with an alternate
adsorption treatment using activated carbon, which costs from $500,000 to $1 million a year
for the carbon. Since the membrane module is likely to last at least a year or so, the
membrane process looks extremely attractive, especially if the recovered PNP from the waste
could be recycled for some payback value. While PNP is a particularly attractive case, it
should be emphasized that the ability to control rate through control of the membrane
structure can permit the application of this approach to many cases where the rates based on
existing membrane supports are much lower than for PNP.
Carboxylic Acids Waste
This waste stream is generated from one of our manufacturing plants. The organic
content of the waste comprises mainly C2-C6 monobasic carboxylic acids (about 3800 mg/L
organics or 2000 mg/L TOC). The stream also contains 1% nitric acid (pH 1.2) and trace
metals and is at 60-70oC. Our study involved using the membrane process to simultaneously
remove and concentrate the carboxylic acids from the waste solution to facilitate further
treatment. At pH 1.2, all the carboxylic acids in solution are in their neutral (uncharged) state
and would permeate through the membrane, though at different rates as shown in Table 4.
Nitric acid and metal ions also present in the solution exist as charged species, hence can not
go through the membrane. The ability of the membrane to remove the organic acids leading to
significant reduction in the TOe of the solution is demonstrated in Figure 7. In this run, a
10% sodium carbonate solution was used as the strip solution to pick up the organic acids
permeating through the membrane. At the end of the run, TOC in the waste had dropped
from the initial value of 2000 ppm to about 50 ppm, equivalent to over 97% TOC reduction.
As shown in Table 4, the carboxylic acids pass through the membrane at different rates,
with caproic acid (C6) being the fastest followed by smaller acids in the order of size. In
accordance with this, we found that caproic acid, valeric acid and butyric acid in the waste
stream passed through the membrane in this order. Because of this characteristic of the
process, the rate of TOe removal steadily goes down as more TOe is being removed from
the solution. We found, however, that up to 90% Toe reduction the transport process could
be approximated by two average rates. Linearized plot of concentration vs. time gives two
rate constants: one is termed Kf initial and is valid up to 50-60% removal, the other is Kf
final and is good up to about 90% removal. Kf final is roughly half the value of Kf initial.
Figure 8 shows such a plot. This approximation allows simple estimation of the membrane
area required for a particular degree of TOC removal. Since the rate is directly related to the
membrane area required, hence membrane cost, the degree of TOC removal required is a
critical parameter in the economics of this process.
The rates of TOe removal from the waste solution using various commercial
microfiltration membranes as supports (polypropylene, teflon, polysulfone) are shown in
Table 7. Of all the membrane supports tested, the best one in ternlS of rate appears to be the
flat sheet eelgard 2500 (polypropylene). The polysulfone hollow fiber membrane exhibits
comparable rate to the polypropylene fiber membrane. Higher temperatures enhance the rate
of TOe removal significantly: the rate goes up about 2.5 times as temperature increases from
25 to 65 oC (Table 8). This is an advantageous feature of the membrane process since the
waste stream as it exists at the plant is at 65°C.
The organic acids could be recovered in water or a caustic solution. With water as the
receiving solution, no concentration of the acids was achieved. With caustic, simultaneous
removal and concentration of the recovered organic acids could be done. In an experiment
similar to the one shown in Figure 7, concentrating organic acids in the sodium carbonate
strip solution to about 6 wt% organics was accomplished along with TOC reduction in the
waste solution. This is equivalent to IS-fold concentration of the feed. The composite
membrane appears to be stable at the high temperature and in the presence of nitric acid of the
waste solution. Preliminary tests of the stability of the membrane were carried out by treating
the waste solution continuously for over two weeks. Separate characterization (in terms of
239
2000 100
1800 --------------------
1600 80
E
Q.
Q.
1400
Z ...J
«
0 1200 60 >
i= 0
::J :::E
...J
1000 w
0
IJ)
a:
0
w 800 40 0
Iii
« ~
600
.=== 400 20
0
0
I- 200
0 0
0 2 3 4 5 6
Run Time, Hr.
Figure 7. Treatment of a carboxylic acids waste with composite polypropylene hollow fiber
membrane. (Membrane module: Celgard X20-400, area = 256 cm2; Feed: 200 ml waste
solution, lumen side; Strip: 35 ml 10% sodium carbonate, shell side; T = 60 C)
k
-
a;-
UI
C"II
' .... , ......
....... /
Initial Rate
......... :-. ...

==
. . . . . . . . 21
r:::
7
.....
~,
",.~
0
0 ...::.:~ Rate
l-
e--' ............
...........
................
6+-----~--~----~----~----~--~----~--~
o 2 4 6 8
Run Time, hr_
Figure 8. Linearized plot for treatment of a carboxylic acids waste using composite membrane.
240
Table 7. Effects of membrane supports on treatment of a carboxylic acids waste
stream using composite membranes
Membrane Support Temperature Kr initial <a) Kf final (b)

OC X 1()4 cm/sec xl0 4 cm/sec
Flat Sheet
Celgard 2400 25 0.9 0.6
Gore-Tex (O.IJ.Ull pore) 25 1.6
Hollow Fiber
Polysulfone 25 1.0
Celgard X20-400 25 0 .7
la) Kf initial is the overall mass transfer coefficientfor up to 60% TOe removal; Ib) Kf final is
the overall mass transfer coefficient from 60 to 80-90% TOe removal.
Table 8. Effect of temperature on rate of TOC removal using a

composite Celgard membrane.
Temperature Kr initial Kr final

oC (x 1()4 cm/sec) (x 104 cm/sec)
25 2.4 1.4
47 3.1 1.7
65 6.3 3.5
carboxylic acids transport) of the membrane before and after this extended run shows
practically no changes.
Using the rates obtained with the Celgard 2500-based composite membrane at 65°C
(Table 8) we estimated that a membrane area approximately 150,000 ft 2 is needed to reduce
90% TOC of the 400 gal/min waste stream. Of the 150,000 ft2, less than 50,000 ft 2 is
required for reduction of the initial 60% TOC, the rest of the membrane area is just for
removing the additional 30% TOC. This results from the combined effect of membrane
transport (first-order rate) and the chemical makeup of waste solution. On the basis of
installed cost of $1O/ft 2 with allowance for foundations, pumps, piping, etc., the whole
membrane system cost could be about $2 million, which compares very favorably with other
treatment approaches such as distillation, reverse osmosis, and biotreatment. Since a very
large membrane system is involved, the actual membrane cost per square foot could be
lower. Also, a hybrid system in which membrane is used to remove 60% TOC or so,
followed by a less expensive alternative for reducing the remaining TOC (e.g.
biodegradation), may result in a lower overall cost.
ORGANICS SEPARATION
The composite membranes can be used to not only recover but also separate organic
compounds based on differences in their physical properties such hydrophobicity, hydrogen-
bonding capability, and degree of dissociation as indicated by their pKa's. The example here
demonstrates the separation achieved between p-nitrophenol (pKa=7.1) and benzoic acid
(pKa=4.25) based on the difference in the pKa's by varying the solution pH. The results are
shown in Table 9. As pure component at pH several units below their pKa's, these
compounds pass through the membrane at fairly high rates. As a mixture at a solution pH of
5.7, PNP exists in solution mainly in the non-ionized form whereas benzoic acid is
practically all ionized. PNP was found in this case to readily pass through the membrane, as
241
Table 9. Separation ofp-Nitrophenol and Benzoic Acid
Solution pH Membrane Permeability*

Benz. acid PNP
Pure componcnLs <pKa 4.0 4.7
Mixture 2.6 3.3 5.1

(2000 ppm PNP,1700 ppm
Benz. acid, 20% KCI)
Same as above 5.7 0 5.97

.. Membrane permeability unit: 104 em/see
expected, resulting in about 99% PNP removal from the feed solution overnight. Benzoic
acid, however, was virtually blocked by the membrane: overnight only trace amount of
benzoic acid «1 % of feed concentration) was detected in the strip solution. It should be
noted that this remarkable separation was accomplished in the presence of very high ionic
strength (2Owt% KCI), which would render ion-exchange methods practically inoperable for
this separation. To further demonstrate the effect of pH and ionization on the transport rates,
the same experiment was repeated but this time at solution pH of 2.6, which is below both
pKa's. Overnight, both PNP and benzoic acid were virtually gone from the feed, having
been transported completely to the strip side, and the measured transport rates for the two as
a mixture were fairly comparable to those obtained with pure components.
CONCLUSION
We have developed a proprietary membrane process effective for removing and
concentrating low MW organic compounds from dilute aqueous solutions, especially those
that also contain high concentrations of inorganic salts which render treatment approaches
such as reverse osmosis and ion-exchange ineffective. Our membrane process can handle a
broad class of organic compounds including phenolics, aromatic acids and amines,
carboxylic acids, alcohols and esters. The membrane appears to exhibit transport rates for
common organic pollutants significantly higher than existing membranes. Tested with two
actual waste streams, the process appears superior to adsorption using activated carbon,
reverse osmosis, and various alternatives (e.g. oxidation, biodegradation). Attractive features
of this membrane process include the ability to recover the contaminants in concentrated form
for either recycle or more economical disposal, low pressure (ambient) operation, and simple
scale-up using commercial hollow fiber modules. We are currently looking into additional
applications of this technology such as fermentation products recovery.
ACKNOWLEDGEMENTS
The author would like to thank J. J. Burke and J. Long for supplying the poly sulfone hollow
fiber modules; S. Lemp and D. Vinjamoori for analytical support; and E. Krupetsky for his
technical assistance.
REFERENCES
1. "Membrane Separation Systems - A Research and Development Needs Assessment,"
Final Report by U.S. Department of Energy, April 1990.
2. S. B. McCray, R. W. Wytcherley, D. D. Newbold, and R. J. Ray "A Review of

WasteWater Treatment using Membranes," Proceedings of The 1990 International
Congress on Membranes and Membrane Processes, Vol. II, pg 1097-1099.
242
3. P. S. Cartwright "Membranes for Industrial Wastewater Treatment - A
TechnicaVApplication Perspective," Proceedings of The 1990 International Congress on
Membranes and Membrane Processes, Vol. II, pg 1131-1133.
4. M. C. Yang and E. L. Cussler, "Artificial Gills," J. Memb. Sci., 42, 273 (1989).
5. R. Prasad and K. K. Sirkar, "Dispersion-Free Solvent Extraction with Microporous
Hollow-Fiber Modules," A1ChE J., 33, 1057 (1987).
6. Hoechst Celanese Product Literature (1991).
7. W. L. Gore & Associates, Inc,. Publication Literature (1990).
8. E. Kline, J. K. Smith, R. E. C. Weaver, R. P. Wendt, and Shyamkant V. Desai,
"Solute Separations from Water by Dialysis. II. Separation of Phenol by Downstream
Conjugation," Separation Science, 8(5), 585 (1973).
9. A. Sengupta, R. Basu, and K. K. Sirkar, "Separation of Solutes from Aqueous

Solutions by Contained Liquid Membranes," A1ChE J., 34, 1698 (1988).
10. B. Reid, Hoechst Celanese Corporation, Personal Communication (1992).

11. S. Whitaker, "Fundamental Principles of Heat Transfer," Pergamon Press Inc., New
York (1977).
APPENDIX
MATHEMATICAL MODELS FOR MEMBRANE STIRRED CELL AND

HOLLOW FIBER SYSTEM.
Membrane Stirred Cell
Rate of organic transport across the membrane in a stirred cell
-VrdCr/dt= AmKr (Cr-eslKo) (1)
where Am: membrane area, cm2

Kr : overall mass transfer coefficient in cm/sec
Cf: organic concentration on feed side, mg/cm3
Cs : organic concentration on strip side, mg/cm3
Ko: ratio oi equilibrium concentration (mg/L) of the organic compound in
strip solution over that in feed.
Material balance for the organic component, assuming that none was present in the in the
strip solution initially and that the amount taken up by the membrane is not significant:
Vr et = Vrer + Vs Cs (2)
where ct: initial organic concentration in feed
Solving for es:
C s = (Vr Ct - veer )/Vs (3)

then
Ce - CslKo = Cr (1 + VfNs.Ko) - Vf ClI Vs.Ko (4)
243
Let <Xl = 1 + VrNs.Ko (5)
<X2 = - Vr ClNs.Ko (6)
Equation 4 becomes:
(7)
Equation 1 becomes
dCr - AmKr dt
Vr (8)
Substitute in equation 7 and integrate:

(9)
dCr - AmKr dt
Vr
In (<Xl Cr + <X2) = - (<Xl Am KrNr) t + constant (10)
Plotting In (<Xl Cr + <X2) vs. time would yield a straight line with slope equal to
- (<Xl Am KrNr), from which Kr can be computed.
With Vs = Vr and Ko» 1:
<Xl = 1 + VrNs.Ko = 1
<X2 = - Vr Cl / Vs·Ko - Cl / Ko
Equation 10 is simplified to
In (Cr - Cr/Ko) = - (Am Kr/V r) t + constant (11)
The tenn CtlKo on the left hand side of equation 10 becomes significant when the system
gets near equilibrium, i.e. Cr approaching ClIKo, at which time the transport process slows
down significantly. For large values of Ko (<': 100), which is the case for most systems in
this study, the actual numerical value of the term CllKo has little effect on the value of Kr
obtained when the system is far from equilibrium.
Hollow Fiber System

Assuming plug flow of the feed solution through the lumen side of the fibers (strip solution
on the shell side) and no liquid movement through the membrane, mass balance on the
hollow fiber unit for counter-current flow gives the following expression for the
concentration change in the feed during a single pass l1 :
Ln~ Cro - Csi IKo ) ] = (12)

Cfj - CsJKo
Where Cfj & Cro : feed inlet and outlet concentrations

Csi & Cso : strip inlet and outlet concentrations
Qr&Os feed and strip flow rates
Note that for large 1<0, equation 12 can be reduced to
Ln ( Cro / Cfi ) = -AmKr/Qr (13)
244
The parameter Am Kf/Qc is the ratio of the mass of material transported through the
membrane to that passing through the module per single pass. For batch operation in the labs
where the amount of membrane area is low relative to the circulation rate. the change in feed
concentration per pass through the module is very low. and Kf can not be obtained accurately
from equation 12 or 13. Recirculating the feed solution through the module is needed to
build up the concentration change.
Combining equation 12 with the unsteady-state mass balance for the feed reservoir (assuming
that feed and strip solutions in the reservoirs are well mixed and their concentrations in the
reservoirs are the same as those entering the hollow fiber module). the following equation
can be obtained relating the concentration in the feed reservoir to time.
CosJKo - Co fi
Ln [ ] <l>t (14)
(COsi/Ko - Cfi) + (VfIKoVs)(COfi - Cfi)
Where Vf & Vs: volumes of feed and strip reservoirs

COfi & COsi: initial feed and strip concentrations
and <I> = (INf+ INsKo) [ 1 - exp(~)] / [I/Qc- exp(~)/QsKo] (15)
with (16)
Equation 14 can be used to obtain the overall mass transfer coefficient from the data of feed
concentration change vs. time.
For the special case of large Ko and with V f and Vs. and Qr and Qs being somewhat
comparable in values, equation 14 can be significantly simplified to
(17)
And if AmKr/Qf is very small (minute concentration change per pass through the module).
equation 17 can be further reduced to the following
(18)
which is the same as equation 11 that was derived for the membrane stirred cell.
245
APPLICATION OF IMMOBILIZED CELL TECHNOLOGY FOR
BIOTREATMENT OF INDUSTRIAL WASTE STREAMS
Daniel E. Edwards l and Michael A. Heitkampl
lEnvironmental Sciences Center

Monsanto Company
800 North Lindbergh Boulevard
St. Louis, MO 63167
INTRODUCTION
Social and regulatory pressures are forcing industry to further reduce environmental
discharges of hazardous chemicals. Monsanto, like many other companies, is focusing its
waste reduction efforts in three main areas: 1) Improved process efficiency leading to less
waste produced. 2) Recycling of waste chemicals and unreacted intermediates. 3) Innovative
waste treatment technologies. This chapter will discuss applied research designed to evaluate
the fluidized bed reactor (FBR) as one type of innovative technology for biotreatment of
industrial wastes.
Once process modifications and chemical recycling have been maximized, some
compounds are likely to remain in plant effluents. Whenever possible, biological treatment
is the method of choice to remove these waste chemicals, since it is usually less expensive,
often results in more complete destruction of compounds, and is sometimes more socially
acceptable than technologies such as carbon adsorption and incineration. The most common
configurations of biological treatment are activated sludge and trickling filters. In the case of
activated sludge, microorganisms are freely suspended in wastewater and aggregate into
biomass units called flocs. In trickling filters, microorganisms coat a bed of material (such as
gravel or plastic rings) to form a thin biofilm and wastewater is sprinkled over the top of the
bed. Over the past sixty or more years as these processes have been widely used, various
mechanical improvements and innovations have increased the performance of these systems.
However, recent needs for higher levels of chemical removal in wider applications than ever
before has created interest in a new generation of biological waste treatment technologies.
There are two main forms that improvements to conventional biotreatment could take:
1) improved performance; 2) reduced cost. Factors improving performance include removal

of previously non-degradable compounds, more complete removal of compounds, and
tolerance to harsh conditions such as high salt, high temperature, toxic mixtures of chemicals,
surge loadings, etc. Factors reducing cost of the systems include resistance to upsets
(reduced down time), reduced sludge production (lower sludge handling costs), and reduced
reactor size (lower capital and operating expense). One new biotreatment approach which is
promising in many of these areas is immobilized cell technology (len.
Configurations of Immobilized Cell Technology
The basic concept of leT involves bacteria attached to a submerged solid support;
waste water is then passed through the bacteria/support complex. There are a number of
configurations of ICT in use or under development at this time. The main difference between
them is the type of solid support and the degree of motion of the support (i.e. biocarrier).
Biocarrier beds may be fixed (i.e non-moving or "packed"), expanded (i.e. lifted slightly by
upflow of wastewater), or fluidized (i.e. lifted significantly by upflow of wastewater with
each support particle being in motion independent of other particles). The subject of this
paper is the configuration of ICT as an FBR.
One common feature of all FBRs is a reaction vessel containing a volume of sand-sized
granules which act as the bacterial support. A flow of wastewater passed upward through the
biocarrier bed is sufficient to fluidize the granules to a bed height approximately 40% greater
than the unfluidized (i.e. collapsed) bed height (Figure 1).
Following inoculation and operation of the bioreactor, each granule becomes coated
with a relatively low density biofilm which decreases the overall density of the granules and
results in a higher level of fluidization (Figure 1). The biofilm thickness must be
mechanically controlled by some type of shearing system in order to prevent excessive
biofilm development which may decrease reactor performance or cause the bed to be flushed
out of the reactor vessel.
I·.............
•••: •.: ••:.:
.............
............
............
............
.........
.................
.... '"'.. ,"--.....
..... . . ."...
.........
............
.~
............
..............
......... .,. .
..............
, ........ .
• • • • t . . . . . ..
............
............
+ ."
H +O.4H +.............
.:~.:~~~~~
-:::::"-:.:.,::,,
.:~.:~.::.-::
............
H +O.7H
~ ~ ~:..::.-:.:-:.:
o ~:..~.::::.:::.
0
iJ: u:: -::,..::::::::.
.::'::.,::-::
ii,
A B c
Figure 1. Diagram of biocarrier beds (H=bed height): A) before fluidization;
B) following initial fluidization; and e) fluidization after full biofilm development.
248
The primary advantage of the FBR configuration is that small granules provide an
expansive surface area for microbial growth within a small reactor volume. An additional
advantage is that the constantly moving bed is less prone to the biomass clogging experienced
in packed bed reactors. Another significant feature of this system is that oxygen must be
dissolved into the waste water before treatment, since excessive bubbling in the bed creates
turbulance which can disrupt optimum fluidization conditions. This feature may be a
disadvantage for high strength waste streams since the low solubility of oxygen may result in
a system which is oxygen limited at high loadings. However, external oxygenation would be
an advantage for treating waste streams containing volatile compounds since these materials
will not be stripped out of the water as may occur in air-sparged systems. FBRs have been
examined for wastewater treatment since the early 1970's.
Fluidized Bed Reactors

Ecolotrol Inc. developed the Hy-Flo® FBR as a proprietary process to upgrade
municipal wastewater treatment (Jeris et al. 1977). This technology used a fluidized sand bed
as a support for microbial growth which provided very high available surface area ( about
1,000 ft2/ft3 of reactor) leading to a maximum of 40,000 mg L-l of mixed liquor volatile
suspended solids (ML VSS). This level of bacteria is about 10 times higher than normally
found in activated sludge. This high mass of attached cells allowed "".an extremely high rate
system which combines the best features of activated sludge and trickling filtration into one
process". This system also claimed to provide greater stability in handling shock loads with
minimal sloughing of attached biomass. Jeris and co-investigators presented data from three
pilot scale units operated on a municipal waste stream for removal of carbonaceous BOD
(biological oxygen demand), nitrification, and denitrification.
Results for BOD removal showed that greater than 90 percent of the BOD could be
removed with retention times as short as 6 minutes. However, internal recycling of flow was
sometimes needed to dilute the waste so that oxygen would not be limited. Nitrifying bacteria
were [characteristically] slow to establish in this FBR study. Once the proper bacteria became
established, NH4-N was removed from a level of 19.1 mg L-1 in the influent waste to less
than 0.2 mg L -1 in the effluent with a hydraulic retention time of 10.6 minutes. It is
noteworthy that the FBR configuration appeared to be especially favorable for denitrification;
N03-N was reduced from 21.5 mg L-1 in the influent to 0.2 mg L-l in the effluent with a
detention time of 6.5 minutes.
Since sludge production was low from these reactors, the authors concluded that
intermediate clarification would not be necessary and final clarifiers may also be eliminated.
Since retention times were short, the FBRs occupied about 5% of the space required by
conventional technology to treat the same waste. The ability to reduce reactor sizes and
eliminate clarifiers led the authors to project major cost savings through use of this
technology.
Sutton and Mishra (1991) presented a historical review of FBR usage during the
1970's and 1980's. They noted that at an international symposium in 1980 the FBR was
249
declared "the most significant development in the wastewater treatment field in the last fifty
years". However, by 1990 only 65 full-scale FBRs had been installed in North America and
Europe. According to the authors, wider use of the technology was hampered by mechanical
scale-up issues, slow development of economically attractive commercial systems, and
proprietary constraints. The principle mechanical problems affecting scale-up of this
technology were: 1) clogging or uneven flow from the influent wastewater distribution
system; 2) oxygen transfer and control: 3) removal of excess biofilm growth from the
granules.
The authors listed 24 specific cases of full-scale municipal and industrial applications of
FBR technology including ammonia removal from fish hatchery water, denitrification of
nuclear fuel processing wastewater, carbonaceous oxidation of chemical wastewater, and
various other examples of wastewater treatment. The current largest industrial user of this
technology was listed as General Motors, which has installed nine aerobic FBR reactors for
treatment of synthetic metal cutting fluids. The conclusion of this review was that if
mechanical scale-up problems could be resolved economically, FBR technology would be
widely used in many different applications (especially in industry).
OBJECTIVES
Previous research, as documented by Jeris et aI., Sutton and Mishra, and others
(Gardener et al. 1988, Heijen et al. 1990, Hickey et al. 1991, Holladay et al. 1978, Hosaka
et al. 1991, and others), has shown FBR technology to be highly effective and cost-attractive
for a variety of low-strength waste streams. The purpose of research at Monsanto is to
examine this technology as it applies to waste streams commonly found in the chemical
industry. These waste streams are often different from municipal and simpler industrial waste
streams in a variety of ways including (but not limited to) the following features:
• Moderate to high organic concentrations (COD up to 15,000-20,000 mg L-l)

• Recalcitrant chemicals.
• Combinations of diverse chemical structure.
• Compounds which are toxic at elevated concentrations.
• Compounds which are regulated to very low concentrations.
• Volatile compounds.
• Biologically unfavorable conditions (e.g. high salt, metals).
• Broad fluctuations in strength and composition of wastes.
If FBR technology could be shown to be more effective than conventional waste

treatment under many of these conditions, then it would be useful in chemical industry
applications. In addition to evaluating this technology in general, three specific objectives
were identified for study:
• Comparison of sand and granular activated carbon (GAC) as biocarriers.

• Development of biomass estimation techniques.
• Estimation of sludge production.
250
Sand is the biocarrier historically used in FBR's. It has the advantages of low cost
(usually less than $20/ton), high density (therefore smaller granules may be used providing a
high surface area/unit volume), and excellent fluidization properties (e.g. flows well, no
clumping, limited flotation, etc.). GAC is increasingly being used in FBRs because it offers
adsorption of certain chemicals as a second mechanism for chemical removal and has a rough,
porous surface to facilitate microbial attachment. The disadvantages of GAC are higher cost
($0.50-$1.00/pound or more), attrition (10-15% of the GAC may be lost annually due to
grinding, deterioration, etc), and lower density (larger particles must be used and excessive
flotation occurs more readily). The objective of this study was to directly compare
performance of these two biocarriers for treating chemical industry wastes in order to
determine whether improved performance of GAC is sufficient to offset the cost advantages
of using sand.
MLVSS is used as a measure of biomass in design and operation of activated sludge
systems. A similar standard method has not been defined for biomass measurement in
attached growth systems such as the FBR. Since high biomass is claimed as one of the major
advantages of this technology, a reliable biomass estimation technique was desired to measure
this quantity and provide better operational control of the reactor. The second objective of
this study was to develop a technique that could provide data directly comparable to MLVSS.
In order to test this new technique, it was used to track biomass growth trends in
experimental FBR units.
The final main objective of this study was to estimate sludge production from FBRs.
Sludge processing, handling, and disposal are some of the most costly factors in most
biological waste treatment systems. Sludge production was monitored from the experimental
FBR units in this study to determine whether it was lower than expected sludge production
from an activated sludge system treating a similar waste stream. If reduced sludge production
by FBRs could be documented, it would represent significant cost savings for this
technology
EXPERIMENTAL DESIGN
Two laboratory-scale FBRs were set-up and operated side-by-side (Figure 2). One
reactor used river sand (0.6-0.8 mm diameter) as its biocarrier and the other used anthracite
based GAC (1.2-1.4 mm diameter) as its biocarrier. The bed of sand or GAC was contained
in a plexiglass "reactor column" which was 6 ft. in height and 4 in. inside diameter. The
initial, unfluidized, bed heights of both reactors were 22 in. Prior to inoculation, the beds
were fluidized to a height of 31 in. (Le. 40% fluidization) by a recycled flow of 1.2 GPM of
water. The recycle flow passed out of a port 13 in. from the top of the reactor column and
subsequently flowed through a "settling basin" where heavier solids would drop out of the
water. The flow then passed through a recycle pump to a venturi which created a pressure
drop of about 15 PSI across the device. Following the venturi, pure oxygen was metered
into the recycle flow and allowed to bubble through the water in a sealed "oxygenation"
column. Any bubbles not entering solution were pulled back into the flow via a "bubble
251
return line" by the vacuum created in the venturi. Bubbles would continue to recycle in this
way until becoming dissolved. Oxygenated water exited the bottom of the column and
concentrated feed stock solution was metered into the flow by a small feed pump (normal
flow was 5-6 mL min-I) pump. The flow then entered the bottom of the reactor column,
completing the cycle. Excess water would exit the system through an effluent port located 5
t.....,-..+- Senling
Basin
Bed Height
Control Device
Bubble
Return
Line
Auidized Bed
Height
Unfluidized Bed
Height Oxygen
Contact
Column
Reactor Column
PLimp
Figure 2. Diagram of laboratory-scale fluidized bed waste treatment bioreactor.
in. from the top of the reactor column. The total volume of each system was 20 L.
A synthetic waste stream containing four organic components was chosen for this
study. These common industrial waste chemicals were sodium formate (FRM), aniline
(ANL), methyl-ethyl ketone (MEK), and paranitrophenol (PNP). FRM was the largest
252
component of this waste stream by weight, being chosen since it would provide much of the
COD in a highly degradable form which would not be adsorbed to GAC. ANL is a
component of moderate degradability and is highly adsorbed to GAC. MEK is a volatile
compound which was added to the mixture to test the ability of the FBRs to degrade it
without air stripping. PNP was the component which was least degradable of the four and
had the added feature of being highly colored (at pH ~ 7.0). The bright yellow coloration of
PNP (even at concentrations as low as 1 mg L-1) became a useful visual indicator of reactor
performance. The appearance of yellow color in the effluent implied that the reactor was
suffering a correctable malfunction or was overloaded. During operation, it was possible to
eliminate the visible evidence of PNP breakthrough by correcting the malfunction or by
reducing feeding until the excess chemical was degraded. The original feed stock was
prepared using 5 g FRM, 0.5 g PNP, 0.5 rnL MEK, and 0.5 rnL ANL to one liter of well
water containing ammonium chloride (0.1 % w/v) and dipotassium phosphate (0.03% w/v) as
inorganic amendments. Loading was increased in seven equal steps (setpoints) by increasing
the concentration of each of the organic components until at the highest setpoint feedstock
was prepared using 20 g FRM, 5 g PNP, 5 rnL MEK and 5 mL ANL. Approximate chemical
loadings at setpoints 1-7 were 0.2, 0.3, 0.5, 0.6, 0.75, 0.85 and 1.0 lbs chemical oxygen
demand (COD)/ cu ft of fluidized bed/day, respectively. Exact loadings fluctuated slightly
according to variations in bed height, feed pump rate speed, and feed concentration. Both
reactors were fed from the same container throughout the study.
Each FBR was inoculated with 1 L of fresh mixed liquor from a full size activated
sludge plant treating a waste stream containing primarily formate along with various other
chemicals. Feeding began at a low rate and was batched (i.e. feeding temporarily suspended)
until clear effluent and D.O. (dissolved oxygen) accumulation indicated that the chemicals
were removed from solution. As colonization of the granules proceeded, bed height
increased gradually. When the beds reached a height of 42 in., a "bed clipping device" was
installed (Figure 2). This device consisted of an electric stirring motor with a 36 in. shaft and
a horizontal blade placed at the base of the shaft. As the bed reached the level of the blade,
the shearing force was sufficient to knock off loose biomass which would float past the blade
and exit the reactor column through the recycle or effluent port. This device was effective in
maintaining constant bed height at 42 in.
RESULTS
Chemical Loading and Removal
Chemical removal (as evidenced by lack of effluent coloring) began almost immediately
in the GAC reactor, so loading was quickly raised to the first chemical loading setpoint
where it was maintained until about 104 days after start-up (DAS) (Figure 3). The sand
reactor required a longer "start-up period" of slow feeding and batching due to visible
253
breakthrough of PNP. At 30 DAS, when COD sampling began, the average chemical loading
remained less than 0.05lbs COD/cu ft/ day. When consistent chemical removal began about
70 DAS, the loading into the sand reactor was gradually increased to setpoint 1 by 100 DAS.
At 104 DAS, the carbon reactor's loading was increased to setpoint 2. The sand reactor
remained at setpoint 1 for about 20 days to confirm consistent chemical removal and was then
raised to setpoint 2. The two reactors were loaded identically (setpoints 2-4) through 125-
170DAS.
At 171 DAS (setpoint 5) a major upset event occurred in the form of a weekend power
failure. The loss of electrical power caused the recycle pumps to shut down, resulting in bed
collapse and anaerobic conditions. When power was restored, recycle pumps did not re-start,
but feed pumps continued to run, resulting in a surge loading of toxic chemicals. Recycle
1.2,-------------------------,
i:I 0.8
2S
80.6
'"
~
~0.4
i
~0.2
50 100 150 200 250

Time (days after start·up)
Figure 3. Chemical loading of the GAC FBR (0) and the sand FBR(e).
pumps were re-started manually about 18 hours later. Sand reactor performance was severely
damaged by this upset. Loading of the sand reactor was suspended for about 14 days until
significant chemical removal resumed. At that point, loading was quickly raised to the pre-
upset levels, but continued poor performance led to the shut-down of the sand reactor at 193
DAS. It was understood that the reactor could probably have been brought back to good
performance by flushing excess chemical out of the system and restarting under a slower
regime of feed increases. However due to time limitations, another lengthy start-up period
was not feasible. The GAC reactor was not seriously affected by the power outage,
presumably due to adsorption of the toxic compounds in the feed solution. Chemical loading
of the GAC reactor was rapidly raised through the last three set-points until being shut down
at 265 DAS after having successfully treated a maximum loading of about 1.0 lbs/cu ft/day.
254
During the middle of the start-up phase (30-60 DAS) of the sand reactor, effluent COD
varied between 100-300 mg L-l, in response to feedlbatch cycles. In the late start-up phase
(60-100 DAS) effluent COD was consistently less than 120 mg VI. The average effluent
COD from 30-100 DAS was 117 mg VI which represented 97% reduction in COD of the
feed stock concentrate (Table 1) . Average COD removal by the sand FBR during setpoints 1-
4 was 97%-99% (Table 1). Chromatographic analysis of effluent samples revealed no
detectable breakthrough of the four components during setpoint 1 (note that samples were not
taken for specific chemical analysis during setpoint 2). At setpoint 3 there were trace
Table 1. Average feed stock concentrations, average effluent concentrations, and specific
chemical breakthrough during operation ofFBRs using sand and GAC as biocarriers.
Setpoint Time period Feed Stock Effluent Chemical Breakthrough l

Days after Start-up --------COD mg L-L--___ _ Component (mg L-t)
Sand Rca!;;tQr
Start-up 0-100 3728 117 Not Sampled
100-121 4741 118 None Detected
2 122-142 5529 92 Not Sampled
3 143-156 7322 \15 FRM (11.3), MEK (0.46)
4 157-172 9548 191 MEK (0.49)
Post-upset 173-188 468 Not Sampled
QACR~Clor
O-I(J4 3745 74 Not Sampled

2 105-142 5364 62 None Detected
3 143-156 7322 123 FRM (12.3), PNP (0.05)
4 157-191 9815 128 None Detected
5 192-207 12,088 155 FRM (29.7); PNP (0.05)
6 208-222 14,989 315 FRM (71.3); PNP (0.05)
7 223-265 17,086 243 FRM (100.7); PNP (0.05)
1FRM= formate, MEK=methyl ethyl ketone. PNP=paranitrophenol. ANL=aniline. Omission of a specific
component indicates that chemical was not detected by chromatographic analysis.
concentrations of FRM (11.3 mg L-l) and MEK (0.46 mg L-l) (Table 1). At setpoint 4,
FRM was not detectable, but MEK (0.49 mg Vi) was still present in the effluent. COD in
excess of quantities accountable by chromatography is presumed to be due to soluble organic
products of microbial activity. During the post-upset period (173-188 DAS), average effluent
COD remained over 400 mg L-I (Table 1) despite the lack of additional feed input. No
specific chemical analysis was performed during the upset period.
255
Chemical removal by the GAC FBR was superior to the sand FBR during the first 100
DAS due to elimination of the start-up period (Table 1). From 100-172 DAS, when both
FBRs were being loaded similarly, COD removal by each was comparable (98-99%). During
the last three setpoints (after the upset of the sand FBR), COD removal by the GAC FBR
continued at 98-99% through feed stock concentration step-increases from 12,088 mg L-I to
14,989 mg L-I to 17086 mg L-I (Table 1). At the fifth setpoint, FRM was detected in the
effluent (29.7 mg L-I) along with a trace quantity of PNP (0.05 mg L-l). FRM was detected
in increasing concentrations at the final two setpoints (71.3 mg L-I, 100.7 mg L-l
respectively) (Table 1). Trace PNP continued to be detected at the last two setpoints but did
not increase in response to the increasing chemical loading. MEK and ANL (both known to
be highly adsorbed to GAC) were not detected in the effluent of the GAC FBR.
Biomass ~easurement
Biomass was measured gravimetrically from samples (22 mL) of the sand and GAC
beds. Samples from pre-determined bed heights (data shown from 10 in. below top of
fluidized bed) were transferred into a 125 mL flask and rinsed three times with distilled water;
rinsate was filtered through a 1 J..LM glass fiber filter for determination of TSS (total
suspended solids, Cleseri et a!. 1989). This TSS value was used as a measurement of the
non-attached solids (i. e. non-attached biomass or NAB) and expressed in mg L-l of fluidized
bed. The rinsed bed material was dried at lOST, weighed, and digested in concentrated
nitric acid for 24 hours. The digested sample was rinsed and dried again at lOYC. The
change in dry weight following the acid digestion was used to calculate the "attached
biomass" of the fluidized bed (mg L-l). This quantity was statistically corrected for chemical
reaction of the acid with the bed material by treating 10 replicate samples of each virgin bed
material by the same technique and determining a correction factor.
Attached biomass in the GAC FBR was found to be approximately 15,000 mg L-l
when the first biomass samples were taken after 90 DAS (Figure 4). Biomass remained in
equilibrium near that level until the beginning of the fifth setpoint at 190 DAS. When the first
samples were taken from the sand FBR after 120 DAS, attached biomass (4,000-5,000 mg
L_l) was much lower than in the GAC FBR. After 130 DAS when both FBRs were being
loaded at the same level, biomass in the sand FBR increased to the same level as the GAC
FBR (Figure 4). Attached biomass in the sand FBR decreased to about 5,000 mg L-l
following the power failure after 170 DAS (Figure 4), but did recover to near its previous
level before being shut down at 193 DAS. Attached biomass in the GAC FBR was not
reduced by the upset and increased to a new equilibrium level near 25,000 mg L-l during the
fifth and sixth setpoints. Following the final chemical loading increase, attached biomass in
the GAC FBR had reached a level of 35,000 mg L-l to 40,000 mg L-l.
Non-attached biomass (NAB) in both FBRs was consistently much lower than attached
biomass, but was significant (1,000 mg L-l -2,000 mg L-l). NAB in both FBRs remained
near these same levels throughout the study and did not appear to be responsive to chemical
loading increases or system upsets (Figure 4).
256
4~~------------------ ________________________ ~
35000
30000
~25000
!
1 20000
.S 15000
=10000
5000
O~~~~~~~~~~~~~~~
90 110 130 150 170 190 210 230 250 270
Time (days after start-up)
Figure 4. Biomass measured from bed samples taken from GAC FBR and
sand FBR. Symbols e, • represent attached and non-attached biomass from
the GAC reactor, respectively; symbols 0, h. represent attached and non-
attached biomass from the sand reactor, respectively.
Sludge Production
The two major routes for sludge to exit these laboratory-scale units were: 1) solids
suspended in effluent; 2) solids deposited in the settling basins. Throughout most of this
study, effluent was pooled into a large carboy for 2-3 days and was sampled at those intervals
for TSS. Effluent solids were generally less than 600 mg L-t at steady state (i.e. not
immediately after an upset or feed increase) in both FBRs (data not shown). However, since
only the lightest solids exited the systems by this route, effluent solids represent only part of
the sludge production picture. For several periods of time, both effluent solids and trapped
sludge were measured from the systems. Trapped sludge measurement involved allowing the
contents of the settling basins to accumulate for discrete. periods of time and then sampling the
TSS contained within them. Total trapped sludge was determined by calculating the total
mass of solids accumulated in the settling basin during that period of time.
An example of data obtained by intensive sampling of effluent solids and trapped sludge
is shown in Table 2. This data was collected during the final 11 days in operation of the
GAC FBR. Since this was the period of highest loading, it represents the period of highest
steady-state sludge production. Total solids exiting the system through the effluent port were
calculated to be 27.5 g (dry weight); solids deposited in the settling basin were calculated to
be 57.1 g (Table 2). The 2: 1 ratio of trapped:effluent solids was representative of data taken
during other periods of sampling. Average total attached and non-attached biomass in the bed
during this time were calculated to be 315.0 g and 11.2 g respectively (Table 2). By
combining the sludge production and total biomass into a single expression over those 11
days, sludge age was calculated to be 42 days (Table 2). COD loaded into the system during
257
the same period totaled 1490 g, leading to a sludge:COD ratio of 0.056 gig (Table 2).
Experience of Monsanto plants treating formate waste in full-scale activated sludge units
has led us to expect sludge age around 20 days and sludge:COD ratios of 0.15-0.2 Ib/lb. If
sludge production data found in this study is confirmed in pilot-scale testing and full-scale
application, this technology will provide high treatment efficiency using biomass which has a
long retention time and relatively low sludge production. The economic benefits of reduced
sludge production are obvious: reduced sludge handing and disposal costs, downsized or
eliminated clarifiers, etc. Increased sludge age (Le. longer SRT) may allow the use of
microorganisms which have intrinsically slow reproductive rates. Processes which use
slowly reproducing microorganisms include anaerobic/anoxic systems, nitrification, and
recalcitrant chemical degradation.
Table 2. Total solids 1 production by the GAC FBR during the period 254-265 DAS.
Sludge exiting system

Effluent Solids: 27.5 g
Trapped Sludge: Rl...g
Total: 84.6 g
Biomass within fluidized bed
Attached biomass: 315.0 g
Non-attached biomass: l.L2..g
Total: 326.2 g
Total COD treated during period 1490 g
Sludge age: 1). 84.6 g sludge lost + 11 days = 7.7 g day-l
2). 326.2 g biomass in bed + 7.7 g day-l = 42 days
Sludge production: 84.6 g sludge lost + 1490 g COD treated = 0.056 g sludge/g COD
1All sludge and biomass quantities are expressed as grams dry weight.
CONCLUSIONS
Aerobic FBRs containing sand or GAC as biocarriers provided excellent COD removal
from a moderate-high strength synthetic industrial waste stream (COD up a maximum to
17,000 mg L- 1 10aded at 1.0 lb COD/cu ft of fluidized bed/day) used. GAC offered major
advantages in reduced start-up time and resistance to chemical surge loading. Especially for
adsorbable compounds, the advantages of GAC in treating potentially hazardous and/or
regulated industrial wastes appear to outweigh high capital cost of the material. Using a
mixture of unrelated chemicals did not favor or hinder biodegradation of any specific
components.
Biomass in the reactor beds was found at high levels (15,000-40,000 mg L-l); the level
of biomass was proportional to the level of chemical loading. Biomass data corroborated
reports of high MLVSS found in FBRs by Jeris et al. (1977). Sludge production by FBRs
was lower and sludge age was longer than would be expected from activated sludge systems
treating similar waste.
258
These laboratory-scale tests provided evidence that Immobilized Cell Technology (lCT)
in the form of a Fluidized Bed Reactor (FBR) is suitable for treatment of typical chemical
industry waste streams. Due to high treatment efficiency, longer solids retention, and
reduced sludge production, FBR technology appears to offer significant cost and technical
advantages compared to conventional treatment techniques. Based on data and experience
gained from this study, there are no identified biological obstacles to the adoption of this
technology for a wide spectrum of industrial uses.
ACKNOWLEDGEMENTS
This work was funded by the Monsanto Chemical Company. The laboratory-scale FBRs
were professionally custom-manufactured by Monsanto Specialty Shop Services. The
authors thank William Adams for developing the initial concept of this research and support
for this work. We are grateful to the following individuals for their excellent analytical and
technical support: Chi Trang (GC-MS), Jon Wehler (HPLC analysis for PNP and aniline),
Eileen Hahn (HPLC analysis for formate), and William Garvey (COD, TSS). Inoculating
sludge was obtained by Zip Beedle through the coordination of Larry Hallas. Edward
Valines added valuable engineering insight and guidance in the early phases of the study.
REFERENCES
Clesceri, L. S. (Chairman), A. E. Greenberg, R. R. Trussel and M. A. H. Franson (Eds.).
1989. Standard methods for the examination of water and wastewater. 17th
Edition. American Public Health Association, Washington, D. C.
Gardener, D. A., M. T. Suidan, and H. A. Kobayashi. 1988. Role of GAC activity and
particle size during the fluidized-bed anaerobic treatment of refinery sour water
stripper bottoms. Journal WPCF 60(4):505-513.
Heijen, S. J., A. Mulder, R. Weltevrede, P. H. Hols, and H. L. J. M. van Leeuwen. 1990.
Large-scale anaerobic/aerobic treatment of industrial wastewater using immobilized
biomass in fluidized bed and air-lift suspension reactors. Chem. Eng. Technology
13:202-208.
Hickey, R. F., D. Wagner, and G. Mazewski. 1991. Treating contaminated groundwater
using a fluidized-bed reactor. Remediation/ Autumn 1991: 447-460.
Holladay, D. W., C. W. Hancher, C. D. Scott, and D. D. Chilote. 1978. Biodegradation of
phenolic waste liquors in stirred-tank, packed-bed and fluidized-bed bioreactors.
Journal WPCF 50(11): 2573-2589.
Hosaka, Y., T. Minami, and S. Nasuno. 1991. Fluidized-bed biological nitrogen removal.
Water, Environment, and Technology. August: 48-51.
Jeris, J. S., R. O. Owens, R. Hickey, and F. Flood. 1977. Biological fluidized-bed

treatment for BOD and nitrogen removal. Journal WPCF 49:816-831.
Sutton, P. M. and P. N. Mishra. 1991. Biological fluidized beds for water and wastewater
treatment. Water Environment and Technology. August: 52-56.
259
MEMBRANE.BOUND MICROORGANISMS FOR REMOVING
ORGANISMS AND METALS
S. S. Sofer
NJIT Biotechnology Laboratory, Department of

Chemical Engineering, Chemistry, and Environmental Science
New Jersey Institute of Technology
Newark, New Jersey 07102
ABSTRACf
Many contaminated aqueous industrial streams contain mixtures of organics as well as

heavy metals. Bacteria offer the advantage of having the ability to oxidize organics as
well as adsorb heavy metals. Membrane-immobilized bacteria have added advantages
including improved mass transfer for faster reaction and adsorption rates, as well as
the capability to be re-used. Performance
characteristics of immobilized bacteria for 90,-------------------~~_,
oxidation and adsorption are herein

presented. A hybrid process configuration
is proposed for the treatment of heavy me-
tals combined with high concentrations of
organics. Ii 50
i-40
i
IMPORTANCE OF FLOW RATE
830
~20
==:
18 ""'"
Microbial reaction systems are limited by 8'0

their physical surroundings. A major o~--_.----~--_.----~--~
limitation is the availability of substrate at o 2 3 4 5

TIme (hr)
the surface of the microorganism. Con-
Fig. 1 FMC Bioreactor Perfonnance Effect
sider the figure on the right. The curves of flow rate on oxygen consumption
show that the reaction rate increases as a rate in single pass configuration.
Industrial Environmental Chemistry. Edited by D.T. Sawym-

and A.E. Martell. Plenum Pless. New York, 1992 261
function of flow rate - the physical resistance to substrate migration to the cell is dras-
tically reduced by immobilizing the microorganisms and forcing liquid past the reac-
tion surface.
This and other advantages of immobilization are reported by Lakhwala and others
(1988), and Mattiasson (1983). The data for this paper are for bacterial consortia
designed to oxidize organics. The experimental setup (next page) has been described
in detail elsewhere (Lakhwala, Goldberg, and Sofer, 1990 and 1992).
Three runs are illustrated for a single-pass configuration, where contaminated fluid is
pumped past the membrane and exits the system. For each run, an initial rise of
oxygen consumption is observed upon pumping of 50 ppm phenol in water. This is fol-
lowed by a plateau at
which the maximum
steady-state rate is ob-
served. Oxygen consump-
tion rate increases as a
function of flow rate be-
cause the fluid mixing
profile is such that there
exists a thinner static
layer of water over the
bacteria at higher flow, al- JO
lowing phenol to diffuse
more quickly into the ac-
EQ.
.e
tive bacterial layer.
Even though the reaction

rate increases as a func-
tion of flow rate, the
==
single-pass configuration - - flow 3 mllmin 18 PI'!!!
- flow 10 m/min
has its limitations. The fig- - - flow 18 m/min
ure on the right shows the O~----r----'r---~-----r----~
exit phenol concentra- o 2 3 4 5
nme (hr)
tions for the same runs.
Fig. 2 Effect of flow rate on exit concentration
They appear higher be- of phenol in membrane reactor (contnuous mode).
cause the over-all flow
rate is high and carries
much substrate through.
This could be a significant

disadvantage if the bioreactor is in service for water purification.
One way to solve this problem is to operate under a recirculation flow configuration.
Here, the contaminated water is pumped from a reservoir into the bioreactor and
flows back into the well-mixed reservoir. This way, the high flow rates are maintained
262
Air/Hydrogen peroxide
~---'~
Air filter Batch mode Continuous mode
---~
Sampling port
Oxygen
probe
Purge
Reservoir Outlet
Presslft r -,
gauge
I
Membrane Reactor
-I
I
L ..J
Stirrer plate
Inlet
Drain
Oxygen
Recirculation probe
pump
In batch mode the waste stream is recycled between the reservoir and the bioreactor. Oxygen
is added to the reservoir in the form of air or hydrogen peroxide. Dissolved oxygen is monitored before and
after the bioreactor using probes. In continuous mode the waste stream is pumped at a lower flow through the
bioreactor without recycling.
Fig. 3 Experimental Set Up of the Membrane Bioreactor.

N
m
w
to take advantage of the resulting lower mass transfer resistance, and higher as-
sociated reaction rates. The liquid is then recirculated until an acceptable contaminant
level is reached and the product water is clean.
This is demonstrated in the figures below.
1~,------------------------------,
1lIMENSIONL£SS 1 - 50 ppm PHENOl...
""* -x- M- -tIC - ) i f - *"' ~-)f- "'* ~
_ flow 10 ml/min
- 25 mf/min
- 50 mf/min
..... 100 mf/min
_x_control
Fig. 4 Variation in Phenol Concentration with Recirculation Flow

Rate in Membrane Bioreactor. (batch mode)
In this flow regime, increased flow rate increases biodegradation of phenol. For this
membrane bioreactor, a flow rate of 100 ml/min decontaminates the water in less than
three hours.
8,--------------------------, The initial rates for phenol removal

at the high flows are in the order of
40 ppm in less than onehoUT. At
6 these reaction velocities, the exit dis-
solved oxygen (DO) concentrations
can reach very low levels as shown
below. When most of the oxidation is
accomplished, the exit DO ap-
proaches that of the entry level as ex-
2
...... 2 ml/min pected.
-- 8 mlZmin
-20 mli'min
-30 ml/fmin Dissolved oxygen concentration is an
O;---,---,---~--~--~--r_~
o 2 3 456 7
important parameter to keep in
Tima (hr) mind throughout these considera-
Fig. 5 Variation in Exit DO with Flow tions. The bio-oxidations are natural-
in Membrane Reactor (batch mode) ly dependent on oxygen availability.
264
Therefore when increased flow rate increases reaction rate, a sufficient decrease in
DO would have a negative effect on the overall oxidation process due to the lower
availability of oxygen.
A recirculation configuration bioreactor is the ideal way to investigate reaction rate as

a function of oxygen concentration. Under very high recirculation flow conditions, the
change in DO across the reactor lessens, and an accurate assessment of rate versus
DO can then be made because the concentration remains essentially constant across
the length of the membrane.
Biomass Availability
A membrane biosupport with too 100

small an amount of bacteria on the
·····wIth
90 _wllhaut~
surface is not efficient. While the 80

dry ~
.................. (2 .....)
bacteria may remain active, the rela-
tive cost of membrane is too high. 70
0-
As the amount of biomass immobi- K! 80
....,
lized on the membrane increases, !:::II 50
the reaction rate will increase. •
Beyond a certain point, the biomass
layer becomes too thick, and the
I 40
JO
productivity of the bacteria farthest 20
away from the waterlbacterial sur- 10
face decreases. In this case, the cost
of biomass becomes too high. There- 20 40 80 80 100 120
fore, an optimum biomass thickness Flow (ml/mln)
exists for each reaction system. A Fig. 6 Effect of Biomass Loading on
typical membrane biosupport is Pressure-How Relationship.
~,------------------------,
...
_500
5
........ rat.: 25 ~....
wound around itself in a spiral fashion. As

11\400
.s::
bacteria attach to the membrane, they
reduce the spacing between each layer of
~
°300 the spiral, resulting in an increased pres-
5 sure drop across the bioreactor.
~200
2i
~ The figure above demonstrates the effect of
0 100
biomass loading on the pressure!flow
relationship. Pressure at the entrance of the
04
0- - - - -1,..0----20..--------1
30 reactor increases upon loading of biomass.
"'-ure (paig) If the biomass is allowed to grow, then the
Fig. 7 Effect of Biomass Loading on pressure increases accordingly. At left is
Operating Pressure.
265
shown the effect of increased biomass loading on operating pressure at constant flow
rate.
From these two figures it is apparent that growth increases pressure drop, and high
biomass loadings lead quickly to unacceptable pressure drops. Bioreactors for decon-
taminating wastewaters should be run under no-growth conditions where possible to
limit the obstacles posed by this type of rise in inlet pressure.
OXYGEN AND HYDROGEN PEROXIDE

Oxygen requirement can be met by bubbling with air, using pure oxygen, or using
hydrogen peroxide.
Air contains a great deal of nitrogen. This causes the stripping of volatile components
into the atmosphere, converting a water pollution problem into an air pollution prob-
lem. This justifies, in some cases, the use of pure oxygen or hydrogen peroxide.
Hydrogen peroxide is converted by the enzyme, catalase, into water and oxygen. Since
catalase is present in large quantities in bacteria, it becomes a convenient source of en-
zyme. A hydrogen peroxide driven bio-oxidation has the strategic benefit that it can
operate under a
closed system - a par-
ticular advantage for
treatment of certain
toxic volatiles.
z 2.50 - y - - - r . I - - - - - - - - - - - - - - ,
On the right, the inlet o
and outlet DO con- ~ /~ RECIRCULATION FLOW - 25 ml/min
HA INJECTION = 1 ml (301 VjV)
centrations are shown !Z <fiB \
w 2 .oo
<.> I
upon injection of a z I 0e03E> OUTlET D.O.
o
single bolus of H 20 2 <.> I ~INlET D.O.
solution into the r5 1.50 I

I
reservoir of a recir- ~
o
culation mode
membrane bioreac- § 1.00
tor. Oxygen con- z
o
centration increases ~0.5O
w
dramatically, being ::;:
higher at the exit. At o -€I
DIMENSIONLESS 1=8 ppm D.O.
point B, substrate is 0.00 -f---,----.----r.:....:....--,.---.------.---1
o 50 100 150 200 250 300 350
injected, and both TIME (min)
concentration profiles Fig. 8 Release of Oxygen in the Membrane
show an immediate Bioreactor on Addition of Hydrogen
decrease. Within Peroxide. (batch mode)
266
about two hours at point C, the H 20 2 is completely consumed and the outlet con-
centration becomes lower than the inlet concentration as expected.
Hydrogen peroxide is very expensive, and its use should be carefully controlled. If the
oxygen generation rate is too high, oxygen will be lost as the solution goes above
saturation.
SUBSTRATEINHIBTInON
Certain substrates such as phenol have an inhibitory and even toxic effect on the bac-
teria. At high concentrations, rates are drastically reduced. The design of bioreactors
dealing with these and other problems is more formally addressed elsewhere
(Lakbwala and Sofer, 1991).
A membrane biosupport system has a thin layer of bacterial mass attached to it. As
flow rate increases, the effects of substrate inhibition and toxicity increase. Therfore,
membrane bioreactors are not recommended for operation at high concentrations of
inhibitory substrates.
Entrapping the bacteria within a protective gel for this regime offers significant ad-
vantages.
ALGINATE BEAD BIOREACTORS
Immobilization of bacterial biomass within calcium alginate gel beads yields a power-
ful moiety for bio-oxidation.
The figure at the top of the next page summarizes the performance of a gel-immobi-
lized bioreactor. The two critical parameters, dissolved oxygen and phenol concentra-
tion, are presented. Percent degradation is shown for a single pass mode. As expected,
the reaction rates increase as a function of available oxygen. The rates are maximum at
200 ppm phenol- quite high when compared to membrane and especially free-cell sys-
tems. Even at 1,000 ppm the reaction is very robust.
The alginate bead forms a protective sphere around the bacteria. As long as oxygen is
available, the substrate within the sphere is oxidized. Therefore, in the robust oxida-
tive environment of the gel bead, substrate concentrations remain low and the bac-
teria are well protected.
In this environment, a difference between oxidation rate and phenol disappearance

may be observed as shown below. Oxygen is consumed very quickly, and intermediate
products of phenol oxidation build up.
ADSORPTION OF HEAVY METALS
It has long been known that heavy metals are adsorbed by bacteria. Gourdon, Rus,
267
~ 100
~90
~ 80
70
~
z 60
0
~ 50
~ 40
!z
.... 30
_ DO I.E.\'EL AT EXIT: 3 ppm
~
.... 20 ..... 8 ppm
Il. - 15 ppm
10 - 25 ppm
0 200 400 600 800 1000 1200
PHENOL (ppm)
Fig. 9 Effect of D.O. Level and Feed Concentration of
Phenol on Continuous Degradation of Phenol at
Steady State in the Alginate Bead Bioreactor.
-
I
~0
0
:510
j _Im-OXDOION li
.-1IICIIIEIlIWM1I
~ B
~
'U
I r
0 ~
c i
a
s
8
a
§ s0
z 4
§
~ z
~ 2
I
i I
W
~~ 0 iI!i
1 1
srARIING PtENOl. CONCENIRATIOH (ppm)
Fig. 10 Effect of Staning Phenol Concentration on Initial
Bio-Oxidation Rates and Biodegredation Rates in
Calcium Alginate Bead Bioreactor.
268
Bhende, and Sofer (1990 & 1990) have shown that metals such as cadmium,
chromium, lead, and others adsorb very well on immobilized bacteria.
The figure below indicates cadmium uptake as a function of temperature. For lead,
the adsorption is in the order of 30 times higher (next page). The adsorption appears
to follow well defined behavior, and the cells can be desorbed and re-used.
pH 6.6
15 Dry cell conc.= 200 ppm
2 hours contact time
ci-13
Q)
....o
(1)12
Q)
o
.!: 11
"0
()
10+--r~r-'--.--r-.--.--.--.-.r-o
o 5 10 15 20 25 30 35 40 45 50 55
Temperature ( C)
Fig. 11
HYBRID BIOREACTION SYSTEMS
In our laboratory, we have found that immobilized bacteria remove a great deal of
heavy metals before exhibiting any inhibitory characteristics. Heavy metal uptake and
bio-oxidation can take place at the same time.
For a system containing organics as well as heavy metals, a primary bioreactor at the
start is recommended for the removal of heavy metals. This reactor serves as and ad-
sorber, and may be regenerated by a decrease in pH, operating in cycles.
For high concentrations of organic contaminants, an alginate entrapped system is

recommended in recirculation mode. As shown above, gel entrapment offers excellent
economy under these conditions. When the concentration of organics is decreased,
then a single pass membrane bioreactor is ideal for polishing and cleaning the water to
the required specifications.
This difference is even greater amplified for a continous-flow, single pass configura-
tion shown on the next page.
A hybrid gel/membrane system is tht~refore a potent combination for attacking con-

centrated organics and heavy metal contaminated waters.
269
200
-~ pH 1.0
pH 5.5
~450
Ii pH 5.0
u 400 pH 4.5
~
-0
OI J50
"-
01
..5. JOO pH 4.25
g- 250
"!II 200
~ 150 pH 4.0
c
~ 100.
a...
50
pH J.5
O~~~-~OT'-rr~-r~~rr~.-. .,,~TO~
o 50 100 150 200 250 JOO
Pb cone. in solution at eq. (mg/I)
Fig. 12 Isotherm of Lead Biosorption by Free Cells at 30 C
Dry Cell Concentration: 200 mg/l or 100 mg/l.
QJ
.J:J
pH 5.5
~JOO
"U
o
0\
.......... 250· o0 o o o pH 5.0
.J:J
0.. oD
E200
..........
pH 4.0
Reoc,"" .......... : ~ 1111

....'" .. "N' : 70 •••\, l.a-3.7 • d:r
!!o) ...... HId, ...... : 2~-27 • I , •• b ••fI •
..oJ P...... 50d .......... I 100-IOS mI
a ........,
FIe. rat.
: 0.21-0.31
: I ~ ml/min
(I) 100 '."""rotur, un4.II,ood : 2I-"'C
"0
oQJ
pH 3.0
.D
C
.D
0.. o-h'TT'TT'T'T'rT'1rrt"1"T" iii iii i , I ii' iii iii I • i 'T"III
o 500 1000 1500 2000
Pb cone. in solution at eq. (mg/I)
Fig. 13 Isotherm of Lead Biosorption by Bacterial Beads in
Column Reactors Operated in a Recirculation Mode.
270
13.0
.12.0 eooee CAL9J.UM ALGINATE
..... MPS MEMBRANE
~ 11.0 . - - - D.O. 20-25 ppm
~ 10.0
~ 9.0
"0
r B.O
..r: 7.0
::::::.
0 6.0 A
•c
..r: 5.0 -0-....
.......
Q.
4.0 ... ... e...

0>
.§. 3.0 ... ... .......
J ... ...
-- --- ---
"e
~
2.0
1.0 ,,0jJ
---
0.0
200 400 600 BOO 1000 1200
PHENOL (ppm)
Fig. 14 Comparison of Biodegradation Rates in Calcium Alginate
and Membrane Reactors .
.,
......
~40,--------------------------------------,
E
o
:c
E 35
0>
*L. 30 G696E) AlGINATE BEAD

..r: bbbbA MEMBRANE
....... - - ESTIMATED FOR MEMBRANE
"0
~ 25
..r:
Q.
~20
'-'"
100 200 300 400 500 600 700 800 900 10001100
PHENOL (ppm)
Fig. 15 Comparison of Degredation Rates Between Alginate Bead and
Membrane Bioreactor.
271
ACKNOWLEDGEMENTS
This work was done at the NJIT Biotechnology Laboratory in Newark NJ, and was
funded by the State of New Jersey Sponsored Chair in Biotechnology, the New Jersey
Institute of Technology, and the Biosupport Group of FMC Corporation. The excel-
lent technical assistance of Emilia Rus and Fayaz Lakhwala is gratefully acknowledged.
REFERENCES
1. Lakhwala ES., Lodaya M.P., Yang Kai Chung, Lewandowski G.A and Sofer S.S.
"Design of Toxic Waste Treatment Bioreactor: Viability Studies of Microorganisms
Entrapped in Alginate Gel". Presented at International Conference on Physiochemical
and Biological Detoxification of Hazardous Wastes, May 3-5, 1988, Atlantic City, N.J.
2. Mattiasson B., "Immobilization Methods" in Immobilized Cells and Orl:anelles, vol
I,. B. Mattiasson, Ed., CRC Press, Boca Raton, FL. Chap 2, 3-26 (1983)
3. Lakhwala, ES., Sinkar, v., Sofer, S.S., and B. Goldberg "A Polymeric Membrane
Reactor for Biodegradation of Phenol in Wastewater" J Bioactive & Biocompatible
Polymers, vol 5:439-452 (Oct 1990)
4. Lakhwala, ES., Goldberg, B.S., and S. S. Sofer "A comparative study of gel-
entrapped and membrane attached microbial reactors for biodegrading phenol"
Bioprocess Engineering Vol 8 (1992)
5. Lakhwala, ES. and S.S. Sofer "Design Considerations for an Immobilized Cell
Bioreactor Operating in a Batch Recirculation Mode''lournal of Chemical Technology
and Biotechnology Vol 52:499-509 (1991)
6. Gourdon, R, Rus, E., Bhende, S., and S. Sofer "Mechanism of Cadmium Uptake by
Activated Sludge" Applied Microbiology & Biotechnology vol 34, 1990, p 274.
7. Gourdon, R, Rus, E., Bhende, S., and S. Sofer "A Comparative Study of Cadmium
Uptake by Free and Immobilized Cells from Activated Sludge" Journal of Environmen-
tal Science and Health, vol A25:1019 (1990)
272
BIOFILTRATION: AN AIR POLLUTION CONTROL TECHNOLOGY
FOR HYDROGEN SULFIDE EMISSIONS
Eric R. Allen and Yonghua Yang
Environmental Engineering Sciences Department

University of Florida
Gainesville, Florida 32611
INTRODUCTION
Traditional air polllution control technologies for pollutant gases, such as
adsorption, absorption and combustion, were developed to treat high concentration waste
gas streams associated with process emissions from stationary point sources. Although
these technologies rely on established physico-chemical principles to achieve effective
control of gaseous pollutants, in many cases the control technique yields products which
require further treatment before disposal or recycling of treatment materials. In the case
of treatment of dilute waste gas streams, however, these traditional methods are
relatively less effective, more expensive and wasteful in terms of energy consumption;
and identification of alternative control measures is warranted. A suitable alternate air
pollution control technology is biofiltration, which utilizes naturally occurring
microorganisms supported on a stationary bed (filter) to continuously treat contaminants
in a flowing waste gas stream. This application of aerobic biodegradation has received
considerable attention and use in Europe, particularly in Germany and Holland during
the last 15 years. 1-21 In the United States, however, relatively little attention22-29 has been
paid to this emerging, innovative and versatile 'low' technology approach to solving
current environmental problems in air quality.
Biofiltration was originally developed for the control of odorous emissions,3.7.14.24
an application involving low concentration waste gas streams. More recently, however,
biofilters have been shown to have a high potential for successfully controlling volatile
organic compounds (VOCs) and air toxics emissions from both traditional and non-
and A.E. Martell, Plenum Press. New York, 1992 273
traditional sources12,2S,30,31. Vapor phase organic emissions in simple and complex
mixtures have been effectively mitigated using biofilters. Examples of emission control
applications at industrial and commercial sources include chemical waste treatment
plants, smelters, rendering plants, publicly owned treatment works, bakeries, breweries,
etc. Control of VOCS and air toxic emissions is of immediate interest in view of the
importance of reactive VOCs in ozone production in non-attainment areas and the
requirements to limit toxic air pollutants emissions under Titles I and III, respectively,
of the 1990 Amendments to the Clean Air Act. In the latter case, limits of 10 tons/year
for individual air toxics or a total of 25 tons/year in mixtures implies that even dilute
( < 5000 ppm) waste gas stream emissions from process vents and leaks, as well as area
and fugitive emission sources will require controls. In this regard biofiltration appears
to be an ideal air pollution control technology, provided that the appropriate
microorganisms are available in the filter medium. Biofilters have been demonstrated
to have the following attributes for air pollution control of dilute waste gas streams.
They are simple to construct from readily available materials, inexpensive to maintain
and operate, versatile in treatment and flexible in operating conditions, and reliable and
effective in their ability to destroy most air pollutants. Biofiltration may be described
as biologically catalysed oxidation at ambient temperatures yielding products that require
no further treatment, such as CO 2 and H 20 from hydrocarbons, or products requiring
minimal treatment, such as H 2S04 from H 2S. The current status of the development,
applications and use of biofiltration in Europe and the U.S. has been the subject of
several recent reviews. 16,26,27,31
One reason for the limited interest in biofiltration applications in the U.S. is the
lack of detailed information concerning the design, operational variables, maintenance
procedures and microbial processes involved for specific control applications. Also,
systematic compilations of long term operational data and associated problems are
lacking. Such information is necessary considering that the operating lifetime of
biofilters could be as long as 5 years.
Two distinct types ofbiofilter systems have been employed. In-ground (soil beds)
and on-ground compost beds require undeveloped land to be available for installation,
whereas packed biofilter towers have minimal area requirements and are generally more
versatile and controllable under varying gas flow and pollutant concentration conditions.
High pollutant removal efficiencies are obtained through preliminary conditioning of the
biofilter, where the appropriate microbiological popUlations are optimized through
exposure to the pollutants of interest and adjustments to the filter bed physical and
chemical properties.
274
During the last six years the University of Florida has been involved in several
biofilter applications projects27-29,32-34, specifically oriented to the control of hydrogen
sulfide (H2S) emissions. Although these studies were designed to address the problem
of objectionable odor complaints from communities residing near wastewater treatment
plants, considerable quantities of H 2S are released to the atmosphere from other sources.
H 2S is emitted as a by-product of industrial processes, such as petroleum refining,
rendering, paper and pulp manufacturing, food processing, 'sour' natural gas processing,
etc. Human exposure to low concentrations of HzS in air can cause headaches, nausea
and eye irritation and at high concentrations exposure can cause paralysis of the
respiratory system leading to fainting and possibly death. In addition, H 2S is corrosive
to materials and harmful to crops and vegetation. Control of H 2S emissions is essential,
therefore, to protect public health and welfare as well as to mitigate environmental
impacts, such as vegetation and material damage. It is interesting to note that H 2S is
included in the listing of 190 air toxic compounds where emissions are required to be
controlled under Title III of the 1990 Amendments to the Clear Air Act. Thus, the
development and application of efficient and economical control technologies for this
obnoxious gas are both desirable and necessary.
Laboratory studies have been conducted to determine optimal operating
conditions for compost-based biofilters in packed tower systems. Also, field studies have
been conducted in association with engineers at a local wastewater treatment plant on
the design. construction and operation of a full-scale on-ground compost biofilter bed
odor contol system. These projects are described briefly to support the contention that
biofiltration is a viable, competitive but underutilized air pollution control technology.
METHODS
Laboratory Studies
The laboratory-scale experimental biofilter system used in these studies is shown
in Figure 1. The experimental system consists of parallel dual column filters which can
be run simultaneously and controlled separately. The biofilter bed material is enclosed
in transparent rigid plastic (Acrylic) columns having an inner diameter of 0.15 meter (m)
and height of 1.2 m. Each column is packed with the desired compost to a height of 1.0
m. The packed biofilter material is supported by a sieve plate to ensure a homogeneous
distribution of the inlet gas across the face of the bed. Six sampling and measurement
ports are distributed evenly along the column for inlet, intermediate and outlet gas and
compost sampling, as well as temperature and pressure measurements. Room air is
forced into a humidification chamber by an air blower (Gast Regenair Model R3105-1).
275
Orifice Vent
0C============~~~========~ ~
House Air Vent

Supply ~
4. ~-.r.b
I
House Air
......"';w"I'VI._ < } - Supply
09=====Oi::F <}- Water Supply
Drain
Figure 1. Laboratory Scale Biofilter System.
Water is distributed into the chamber using a spray atomizer and wets the Pall rings
packed in the chamber. The latter extend the wetted surface area in contact with the
flowing air. Relative humidities in the range from 95 to 100% were routinely and
continuously achieved with this system.
Purified hydrogen sulfide (Liquid Air, purity 99 + %) was regulated and mixed into
the humid air flow prior to the biofiIter bed. Flow rates of air and H 2S were controlled
by needle valves and monitored by previously calibrated rotameters in order to o\>tain
the desired H 2S concentrations and gas flows through the columns.
Gas samples were taken from various locations on the column by attaching a gas-
tight syringe to a sampling port and extracting an aliquot of gas. These samples were
accurately diluted with pure nitrogen (Liquid Air, Purity 99.99%) after transfer to a
sealed Tedlar bag. Diluted gas samples were analyzed using a commercial gas
chromatograph (Tracor Model 250H) equipped with a sulfur-specific flame photometric
276
detector. The H 2S detection limit of the method was determined to be 0.02 parts per
million by volume (ppm) with 95% confidence.
The filter materials used in the laboratory studies were composts made from yard
trash, pine bark and chips from a variety of sources. Compost samples were stored in
sealed plastic bags at room temperature until needed. Some compost samples were
stored in this way for several months prior to use. When results for stored compost
samples were compared to those for fresh samples, no significant differences in operating
characteristics were observed as a result of extended storage time. The sources and
principal properties of the composts studied are summarized in Table I.
Table I. Sources and Properties of Laboratory Biofilter Materials.

Compost ID#
3 6 13 14 16 17 17A
Bulk Density (gjcc) 0.18 0.20 0.27 0.18 0.22 0.20 0.20
Pore Volume (%) 84.3 89.7 88.4 88.7 88.7
Water Content (wt%) 62.8 64.8 54.9 62.4 56.5 62.7 62.7
Organic Matter (wt%) 72.3 67.2 66.5 64.2 64.3 62.5 62.5
pH 9.22 7.21 1.60 6.44 6.66 8.10 8.10
Description of the Composts:
Compost #3: Yard trash compost from Pompano Beach, FL. one year old.
Compost #6: 25% by volume of sewage sludge compost and 75% of yard trash
mixed and composted; about 18 months old.
Compost #13: Compost #6 mixed with tree bark and sewage sludge, lime was
used to adjust pH before use. Compost obtained from a filter bed
which has been used for H 2S removal for 2.5 years.
Compost #14: Compost from Kanapaha WWTP, Gainesville, FL. Yard trash,
grass and sewage sludge werre mixed and composted; lime was
used to adjust pH; 2.5 years old.
Compost #16: Yard trash compost from Wood resource recovery, Inc.,
Gainesville, FL. 3.5 months old.
Compost #17: From same source as compost #16, 1:1 by volume of yard trash
and grass composted; 3.5 months old.
Compost # 17A: Compost #17 mixed with 2% lime (CaC03), weight percent of dry
weight of compost.
Full Scale Biofiltration Studies

A biofilter system, for control of hydrogen sulfide emissions at a local wastewater
treatment plant, was installed in the summer and made operational in the fall of 1988.
The system consisted of an air collection system, anti-corrosive blower, ductwork,
277
humidifier, central air flow shaft for flow adjustment and a dual-bed Siebo-stone air
distribution system, as shown in Figure 2.
The plant's malodor source was identified to be the grit chamber located in the
headworks building. The building was sealed as tight as possible and a negative static
Waste Gas Inlet

SIEBO-stone Gas
Distribution System
Humidifier
Cinder Block
Retaining Wall
A -+
Plan View
Section A-A
Figure 2. Full Scale Biofilter System.
pressure was maintained by the blower drawing contaminated air at 85 m3/min. through
03 m diameter exhaust ducts. The exhaust air was forced to a humidifier where water
was dispensed by several spray nozzles to saturate the waste gas stream. All pipes in the
waste gas collection and transport system were made from polyvinyl chloride to minimize
corrosion problems.
The biofilter bed gas distribution system was constructed from the German
patented SIEBO stones. These interlocking sinter block base units of the biofilter bed
when installed, were rigid and allowed for heavy vehicles to be driven on the base
without damage. The SIEBO-stone base system provided for an even distribution of the
inlet air to the filter bed as well as functioned as a drainage system. The overall filter
bed area was 100 m2, divided into two equal sections, which could be operated and
278
controlled individuallly. During normal operation each section treated half of the total
contaminated gas flow by balancing the separate input flows. The total flow, however,
could be diverted to one section if necessary, when repairs or refurbishing were needed,
without reducing overall pollutant removal efficiency.
Composts obtained from Pompano Beach and Jacksonville, Florida were mixed
and used as the filter bed material. This mixture consisted of yard waste compost, pine
bark, wood chips and sewage sludge. Lime was applied to the compost material to
buffer the bed acidity (pH) prior to installation. The biofilter material was spread over
the SIEBO-stone air distribution system to a depth of at least 1.3 m. A 1.3 m retaining
wall was used on three sides of the bed to hold the compost. The fourth side of the bed
was left open for access and the compost on this side was overlapped with the edge of
the air distribution system by 1.3 In, through incorporation of an incline of 45° in the bed
from bottom to top.
The waste air flow rate through the biofilter was in the range from 79 to 96 actual
cubic meters per minute, which corresponded to an average gas loading rate of 52
m3 /m2_h and average gas retention time of 88 seconds. The full-scale biofilter design
and operating parameters are summarized in Table II.
To monitor biofilter bed performance, influent and effluent gas samples were
collected and analyzed for H 2S by procedures described in detail elsewhere.32,34
RESULTS AND DISCUSSION

Laboratory Studies
The composts listed in Table I have been studied in the laboratory-scale biofilter
system for periods varying from 1 to 4 months. Typical physical and chemical properties
of these composts are presented in Table III. Most of the composts showed a high
removal efficiency (> 99%) during the first 24 to 48 hours of operation. This high
Table II. Full Scale Biofilter Design and Operating Parameters
Total Flow: 79- 96 acmm

Filter Area: 100 m2
Filter Height: 1.3 meter
Gas Loading Rate: 47 - 57 m3/m2_hr
Retention Time: 82 - 100 sec.
Temperature: 15 - 30°C
Pressure Drop: 150 - 200 mmHp
279
efficiency, however, was probably due to physical and chemical, rather than biological,
interactions between H 2S and the bioftlter materials. During the subsequent 2 to 14
days, the removal efficiency declined to about 90%, presumably as a result of acclimation
and development of the optimum population of sulfur-oxidizing microorganisms. Mer
this conditioning period the bioftlters approached their final stable condition and the
most efficient control (> 99%) of H 2S was achieved. If the compost had been exposed
previously to HzS then the initial physico-chemical and conditioning stages did not occur.
Seeding of sulfur oxidizing bacteria to the compost was not necessary because the
Table III. Physical and Chemical Properties of Typical Laboratory Filters.
Property
Bulk Density 0.18 - 0.30 g/cm3

Particle Density 1.7 - 1.9 g/cm3
Porosity 0.84 - 0.90 cm3/cm3
Water Content 0.45 - 0.63 gig
Organic Matter 0.59 - 0.67 gig
pH 1.6* - 8.1
Total Sulfur 0.74 - 70.4* mg-S/g
Soluble Phosphorus 0.11 - 0.22 mg-P/g
Total Carbon 0.31 - 0.41 g-C/g
Total Nitrogen 0.013 - 0.043 g-N/g
C/N Ratio 7.1 - 32 g-C/g-N
*Extreme values from 'used' biofilter materials previously exposed to H 2S
required bacteria exist naturally in soils and sludges.3S,36 Mixing of small amounts of
sewage sludge with the composts, however, was found to increase the initial bacterial
population and provide for a shorter conditioning period. Although addition of sewage
sludge can provide an additional source of nutrients, unseeded composts were found to
be as efficient as sludge-seeded composts once they were conditioned and stabilized.
An important operating variable in biofilter systems is the energy required for air
movers to transport the waste gas through the ftlter bed at the required flow rates. The
pressure drop across the bed increased markedly as the flow rate was increased. Since
the pressure drop was determined by the bed depth it was necessary to keep the gas
velocity as low as possible. One of the most important properties of the bed material,
which significantly affected the pressure drop, was the compost small particle content
(~< 1.2 mm). It was found that for aIm bed depth if particles with sizes less than 1.2
mm were removed then gas velocities as high as 10 m/min could be obtained for
pressure drops up to 50 cmH20. It should be noted, however, that during operation of
280
the biofilter to remove H 2S, mineralization of bed material and sulfur accumulation
occur which increased the biofilter small particle content. Thus, measures should be
taken to separate the accumulated small partlicles periodically, otherwise a substantial
increase in pressure drop and reduction in gas velocity will be observed during use,
which leads to enhanced energy consumption.
The effect of gas retention time on H 2S reduction has been studied by varying the
gas flow rate through the biofilters. When the H 2S loading rate was kept constant and
below the maximum loading capacity of the bed the removal efficiency remained high
(> 99%) for residence times as low as 23 seconds. In subsequent tests, at very high H 2S
concentrations in the inlet gas the removal efficiency remained high as long as the
residence time of the flowing gas was increased. For example, with an inlet
concentration of 2651 ppm H 2S, a 99.8% removal efficiency was obtained for a residence
time of 197 seconds. In this case the filter bed maximum loading rate was determined
to be 1.2 g-H2S/m2-min for a typical compost bed. However, the latter variable is a
function of H 2S concentration and gas velocity. A more appropriate measure of biofilter
capacity is the concept of H 2S mass loading rate per unit mass of compost (mg-H 2S/kg-
min). The maximum H 2S loading capacity for typical compost biofilters has been
determined to be in the range from 2.3 to 4.8 mg- H 2S/kg-min. The long term influence
of H 2S loading rates on H 2S removal efficiency is shown in Figure 3.
The effect of varying H 2S concentration in the inlet gas stream on biofilter

removal eficiency was investigated at constant waste gas flow rate. No significant
differences in H 2S removal efficiencies (> 99%) were observed for concentrations in the
range from 5.5 to 518 ppm because the H 2S loading rate was always less than the
maximum H 2S loading capacity for the bed. However, if the maximum loading capacity
was exceeded then the H 2S removal efficiency was observed to decrease significantly.
For example, when the H 2S loading rate was 7 mg-H2S/kg-min for a bed with a
maximum loading capacity of 2.3 mg-H2S/kg-min the removal efficiency was reduced to
65% and the biofilter was overloaded. The maximum loading capacity of the bed is the
maximum amount of H 2S that the optimal microbial population can consume, without
inhibition of activity.
The effect of varying compost water content has been studied with respect to H 2S
removal efficiency. The data indicated that for water contents greater than 30% by
weight, high removal efficiencies (>99%) were obtained consistently. However, when
the water content was reduced to less than 30% the efficiency decreased markedly with
decreasing water content. It appears that a liquid water film on the compost is necessary
281
110
\N
100
90
ao
~
70
10
110 ~ •
40 o 20 40 10 ao 100 120 140 110 lao 200
200r-~------------------------------------------~~-,
180 • Inlet ... Outlet

1110
140
120
100
80
10
40 .
20
·1 gc=fO=:;~2;:::0::::4::0:::::~120=::=;:ao::::::;:::::,0;::;0~='::;2~0::!=~14~0~~I:::;I:;::0::;:::=,tao::::::;20:;::0=l
2Sr-~------------------------------------------------,
24
22
20
la
18
14
12
10
8
8
4
2
oL-~0~--2~0~-4~0~~8~0~~80~~10~0~-1~2-0~-14~0~~18~0~-,~ao~--20~0--
Cumulative Operation Time (Day)

Figure 3. Biofilter Removal Efficiency During Long Term
Laboratory Operation.
for the sulfur oxidizing bacteria to survive and propagate in order to establish the
optimal population3? Thus, compost water contents in excess of 30% appeared to be
necessary to both saturate the filter material and produce the desired liquid water film
environment for the bacteria.
The compost was found to deteriorate in terms of accumulated acidity and sulfur
content during operation. These phenomena were a result of the continuous conversion
of H 2S, through several oxidation states, to sulfuric acid. If the bed is not treated then
the 'active' portion of the biofilter becomes acidified and sulfated. Fortunately, this
282
deterioration is not permanent and can be reversed by periodic washing of the bed with
water. Although periodic washings can remove 35% of the accumulated sulfate and
hydrogen ions, they appear to have only a small effect on pH, i.e. from 0.02 to 0.03 pH
units. Nevertheless, periodic washings can maintain or restore the original removal
efficiency, prevent bed overloading and contamination, remove accumulated fine
particles and extend the operating life of the biofilter. Additional details of the
laboratory studies briefly described here are reported elsewhere.33,34
Full-Scale Studies
Influent and effluent gas samples were collected and analyzed during the first 16
days of operation of the full-scale biofilter. During this start-up period H 2S
concentrations in the influent gas stream varied in the range from 156 to 229 ppm and
off-gas concentrations were observed in the range from 0.05 to 0.40 ppm. The efficiency
of H 2S removal was greater than 99% for the sampling period because some of the
compost in the mixture used for preparing the biofilter had been exposed previously to
H 2S, thus, the initial physico-chemical reaction and conditioning stages were not
observed. Subsequently, gas and compost samples were taken every quarter (3 months)
to determine whether the filter bed efficiency and composition had deteriorated during
operation. Physical and chemical properties of the 'aged' on ground biofilter are given
in Table IV.
Table IV. Physical and Chemical Properties of the Aged Full Scale Biofilter.
Bulk Water Organic Total Total Total

Date Density Content Matte pH N C S
(glee) (%) (%) (mg/g) (mg/g) (mg/g)
05/10/88 0.23 52.1 68.1 8.63 2.30 40.9 7.3
11/21/88 0.27 54.6 69.3 4.40 -- -- --
05/16/90 0.30 45.6 66.5 2.62 1.89 34.3 44.5
12/20/90 0.27 54.6 66.5 1.60 1.58 36.6 71.0
02/05/91 -- 57.2 64.7 1.80 -- -- 109
Additional tests were conducted to ensure that off-gas sampling at different

locations were representative of the bed as a whole by comparing results obtained at
four locations on top of the bed. Values obtained were consistently less than 0.2% of
the inlet gas concentration, although individual values varied initially by as much as a
factor of 3. Also, tests were carried out to ensure that the methods employed for
collection and storage of gas samples did not influence the analytical results. These
283
quality control procedures provided asssurance that sampling and analysis data
completeness, accuracy and precision objectives were achieved and that representative
gas samples were collected from the biofilter.
The influent gas humidifier worked effectively to maintain the filter bed moisture
content in the range from 45 to 60%. No other water was introduced to the bed except
that due to rain. As a result the top of the bed dried out significantly during long term
operation. No marked changes in the bulk density and particle size distribution of the
compost were observed during 27 months (11/88 to 02/91) operation of the biofilter.
The organic matter content decreased by about 4.5%, presumably due to mineralization
of the compost. Total nitrogen and carbon contents of the compost decreased at
different rates resulting in an increase of the C/N ratio from 17.8 to 23.2. The most
significant changes observed in the bed material during extended operation were the
total sulfur (S) content and pH. The total sulfur content increased from 7.3 to 109 mg-
S/g-compost on a dry basis. Also, after prolonged operation, parts of the biofilter bed
showed a pronounced color change from dark brown to yellowish-white, which was
accompanied by emission of an obnoxious odor. These observations were similar to
those from laboratory studies, which were obtained during dry-out of the compost or by
polllutant overloading.
The pH of the compost decreased significantly over several years from 8.6 to 2.6
as a result of continuous aerobic biodegradation of H 2S. Although this acidification of
the compost did not appear to affect the overall HzS removal efficiency, considerable
corrosion of the cement blocks in the retaining wall and plugging of the bed inlet-gas
distribution vents in the SIEBO-stone base were observed. As a result of this obvious
system deterioration, it was decided that the SIEBO-stone air distribution system should
be replaced with components made from acid-resistant or anti-corrosive material.
CONCLUSIONS
Both the laboratory and full-scale biofilter systems have been demonstrated to
perform with suitable efficiency during long term operation provided that the systems
are operated within specific ranges for important operating variables. Recommended
ranges for operating variables when using compost biofilters for control of HzS emissions
are presented in Table II. Maintenance procedures for compost beds in this application
are minimal. The biofilters require washing with water from the top of the bed at
weekly intervals to maintain an appropriate water content (> 30% wt/wt) throughout the
bed and to prevent acidity and soluble sulfur from accumulating to levels which are toxic
284
to sulfur oxidizing bacteria. A sprinkler system capable of providing water at a rate of
40 liters/m2-min. for 10 minutes every week should suffice.
Laboratory kinetic studies have shown that H 2S removal is zero order in H 2S at
high concentrations (>400 ppm) and first order in H 2S at low concentrations «200
ppm). Presumably, the bioreaction occurs by diffusion of H 2S into liquid water droplets,
where the sulfur oxidizing bacteria are located?7 At high concentrations of H 2S in the
gas phase the water droplets will be saturated with H 2S and the rate of reaction will
depend primarily on the microbial activity. However, at low H 2S concentrations the
availability of H 2S for the microorganisims will depend on the rate of diffusion of H 2S
to the liquid surface and mass transfer across the gas-liquid interface, which depends on
the H 2S concentrations in the gas phase.
Although certain deficiencies in using compost-based biofilters have been noted,
in general biofiltration has been demonstrated to be a simple, inexpensive, durable and
effective method for H 2S control at wastewater treatment plants. Similar systems could
be designed, constructed and operated effectively in other industrial applications.
Additional research is essential to provide detailed knowledge of the appropriate
biofilter systems, operating conditions and maintenance procedures necessary to
efficiently treat other dilute industrial waste gas streams containing organo-sulfur,
volatile organic and air toxic compounds.
REFERENCES
1. Kneer, F.x. "Device for the Reduction of Gaseous Organic Pollutants from Waste
Gases", German Patent No.: DE 2445315 C2 BOlD 53/54, West Germany, 1976.
2. Kneer, F.X. "Waste Gas Purification with a Biological System", Das Technische
Umweltmagazin, West Germany, 20, October 1978.
3. Jaecklin, F.P. ''The Deodorization of the Sewage Treatment Plant at Staz (St.
Moritz) Using Soil Bed Filters", Stuttgarter Berichte zur
Siedlungswasserwirtschaft, Vol 59, University of Stuttgart, West Germany,
September 1976.
4. Gust, M., Sporenberg and E. Shippert. "Fundamentals of Biological Waste Gas
Purification Part IV: Gas Cleaning by Microorganisms in Bio-Scrubbers", Staub-
Reinhaltung der Luft, West Germany, pp. 308-314, September, 1979.
5. Gust, M., H. Grochowski and S. Sibirz. "Fundamentals of Biological Waste Gas
Purification, Part V: Gas Cleaning by Microorganisms in Biofilters", Staub-
Reinhaltung der Luft, West Germany, pp. 397-438, November 1979.
6. Ottengraf, S.P.P, AH.C. Van Den Oever, and FJ.C.M. Kempenaars. Waste Gas
Purification in a Biofilter Bed," in Innovations in Biotechnology, Houwink, E.H.
and Van Dan Meer, R.R. eds., Elsevier Science Publishers, B.V., Amsterdam,
The Netherlands (1984).
7. Koch, W., H.G. Liebe and B. Striefler. "Experiences with Biofilters for the
Reduction of Odorous Air Emissions", Staub-ReinhaItung der Luft, West
Germany, pp. 488-493, December 1982.
285
8. Ottengraf, S.P.P. and AH.C. Van Den Oever. "Kinetics of Organic Compound
Removal from Waste Gases with a Biological Filter", Biotechnol. Bioen~. ~,
3089-3102, (1983).
9. Bohnke, B. and D. Eitner. "Investigation and Comparsion of Different Kinds of
Compost in a Mobile Biofilter", Gutachten i. A der AG Kompostabsatz NW,
Aachen, West Germany, June 1983.
10. Eitner, D. "Investigations of the Use and Ability of Compost Filters for Biological
Waste Gas Purification with Special Emphasis of Aspects", The Operation Time,
GWA, Band 71, RWTH Aachen, West Germany, September, 1984.
11. Kneer, F.x. "Gas Cleaning using a Biofilter", CAV, West Germany, pp. 112,
October 1985.
12. Gethke, H.G. "Capabilities of a Compost Filter System for Waste Gas
Purification Using, as an Example, Waste Water Treatment Plants with Industrial
Waste Water Loading", Paper presented at Colloquim on Odorants, Baden-
Baden, West Germany, October, 1985.
13. Dragt, AJ. and S.P.P. Ottengraf. "Biofiltration: A New Technology in Air
Pollution Control Experiences in the Netherlands", Paper No. 4-4508-23-10,
Eindhoven University of Technology, Einhoven, The Netherlands. May 1985.
14. Don. J.A and L. Feenstra "Odour Abatement Through Biofiltration", Paper
presented at Symposium in Louvain-La-Neuve, Belgium, April 1984.
15. Don, J.A''Tbe Rapid Development of Biofiltration for the Purification of
Diversified Waste Gas Streams", VOl Berichte 561, VOl Verlag Dusseldorf, West
Germany, p. 63, 1985.
16. V.D.I. "Biological Waste Air Purification-Biofilters", VOI-3477, VOl Handbuch,
Reinhaltung der Luft, Band 6, Dusseldorf, West Germany, December, 1984.
17. V.D.I. "Biological Waste Air Purification - Bioscrubbers", VOl 3478 VOl
Handbuch, Reinhaltung der Luft, Band 6, Germany, July, 1985.
18. Zeisig, H.D. "Biofilter for Agricultural and Industrial Applications, Especially the
Tobacco Industry", Paper presented at Colloquium on Odorants, Baden-Baden,
West Germany, October 1985.
19. Klee, W.K., Lutzke and H. Vollmer, "Reduction of Odorous Emissions by
Biofiltration Demonstrated Using as Examples, a Fiber Plate Production Facility,
a Vulcanization Facility and a Foundry", Paper presented at Colloquium on
Odorants, Baden-Baden, West Germany, October 1985.
20. Ottengraf, S.P.P. "Exhaust Gas Purification", in BiotechnoloilY, Rehm, H.J. and
Reed, G. Eds. Vol. 8; VCH VerlagsgeseIlschaft, Weinheim, FRG (1986).
21. Eitner, D. and H.G. Gethke. "Design, Construction and Operation of Biofilters
for Odor Control in Sewage Treatment Plants", Paper No. 87-95A6. Presented
at 80th Annual Meeting of Air Pollution Control Association, New York, NY.
June 21-26, (1987).
22. Carlson, D.A and c.P. Leiser. "Soil Beds for the Control of Sewage Odors",
J. Water Pollut Control. Fed.. J8, 829 (1966).
23. Rands, M.B., D.E. Cooper, c.P. Woo, G.C. Fletcher, and K.A Rolfe. "Compost
Filters for H 2S. Removal from Anaenrobic Digestion and Rendering Exhausts"
J. Water Pollut. Control Fed .. jl, 185 (1981).
24. Prokop, W.H. and H.L. Bohn. "Soil Bed System for Control of Rendering Plant
Odors", J. Air Pollut. Control AsSOc.,~: 1332 (1985).
25. Kampbell, D.H., J.T. Wilson, H.W. Read, T. Thomas and T.T. Stocksdale.
"Removal of Volatile Aliphatic Hydrocarbons in a Soil Bioreactor", J. Air PoIlut.
Control Assoc.,:Jl: 1236 (1987).
26. Hartenstein, H.U. "Assessment and Redesign of an Existing Biofiltration System"
M.S. Thesis, Univeristy of Florida, Gainesville, Fl., 1987.
286
27. Hartenstein, H.U. and E.R Allen. "Biofiltration an Odor Control Technology for
a Wastewater Treatment Plant", Project Report, Environmental Engineering
Sciences Department, University of Florida, Gainesville, FL., October 1986.
28. Allen, E.R, H.U. Hartenstein and Y. Yang. "Review and Assessment of the
Design and Operation of a Compost Biofilter System for Odor Control", Project
Report, Environmental Engineering Sciences Department, University of Florida,
Gainesville, FL., 1987.
29. Allen, E.R, H.U. Hartenstein and Y. Yang. "Identification and Control of
Industrial Odorous Emissions at a Municipal Wastewater Treatment Facility",
Paper No. 87-95A4, Presented at 80th Annual Meeting of Air Pollution Control
Association, New York, NY, June 21-26, 1987.
30. Ottengraf, S.P.P. "Biological Systems for Waste Gas Elimination," TIBTECH. 5,
132, 1987.
31. Leson, G. and AM. Winer. "Biofiltration: An Innovative Air Pollution Control
Technology for VOC Emissions", J. Air Waste Mana&e. Assoc., 41: 1045, 1991.
32. Yang, Y. and E.R Allen. "Biofiltration Control of Odor Emissions in Wastewater
Treatment Plants", Paper presented at Symposium on Biotechnology for
Wastewater Treatment - II, 201st National Meeting ACS, Atlanta, GA, April 14-
19, 1991.
33. Allen, E.R and Y. Yang. "Biofiltration Control of Hydrogen Sulfide Emissions",
Paper No. 91-103.10, Presented at the 84th Annual Meeting of Air and Waste
Management Association, Vancouver, B.c. Canada, June 16-21, 1991.
34. Yang, Y. "Biofiltration for Control of Hydrogen Sulfide", Ph.D. Dissertation,
University of Florida, Gainesville, FL., 1992.
35. David, M.B., M.J. Mitchell and J.P. Nakas. "Organic and Inorganic Constituents
of a Forest Soil and Their Relationship to Microbial Activity", Soil Sci. Soc.
Amer.1.. 46, 847, 1982.
36. Starkey RL. "Oxidation and Reduction of Sulfur Compounds in Soils", Soil Sci..
i l l 297, 1966.
37. Ottengraf, S.P.P. "Theoretical Model for a Submerged Biological Filter",
Biotechnol. Bioen&. 19, 1411, 1977.
287
INORGANIC ION EXCHANGE MATERIALS FOR
NUCLEAR WASTE EFFLUENT TREATMENT
Abraham Clearfield
STATEMENT OF PROBLEM AND SCOPE

Over the past 48 years, nuclear defense activities have produced large quantities of
nuclear waste that now require safe and permanent disposal. The major high-level
radioactive waste accumulations are stored at Hanford in the southeastern desert of
Washington State and at Savannah River in South Carolina. The magnitude of the problem
is enormous. There are 177 storage tanks of a million-gallon plus size nearly filled with
high level waste at Hanford alone. Much of the waste was placed in the tanks with no
records kept of their content. Over a period of time new additions were made to the tanks
as needed so that the inventory of content is at best approximate. In order to prevent
corrosion of the tanks the pH was raised to above 13 which resulted in the precipitation of
large quantities of hydroxides. Thus, Hanford must not only deal with huge amounts of
liquid waste, but must also develop methods of treating the solid waste.
In all, nine plutonium production reactors were built at Hanford, starting in 1943
and lasting until 1963. The plutonium fuel from these reactors was reprocessed in on-site
processing facilities using a solvent extraction process designated PUREX. In general, the
nuclear fuel was dissolved in nitric acid and a separation of uranium and plutonium effected
by organic phosphate complexing agents dissolved in an organic phase. The extraction of
1 kg of plutonium by the PUREX process produced over 340 gallons of liquid high level
radioactive wastes, more than 55,000 gallons of low to intermediate-level radioactive
wastes and required over 2.5 million gallons of cooling water 1. Since 1948, the U.S. has
produced about lOS kg of plutonium for military use. At Hanford the high level radioactive
wastes, containing most of the fission products, complexing agents, organic solvents and
water were poured into the waste storage tanks and the pH raised to above 13. As a
consequence the solution is 7-8M in NaN03. These tanks are of the single shell type and
after a number of years of storage, several of them were found to leak. Subsequent to this
discovery 28 newer tanks that were built were of the double shell type. The low to
intennediate level radioactive wastes were put into cribs, which are like septic-tank drainage
fields, and the cooling water was pumped into surface ponds2.
The single shell tanks have leaked more than 3/4 of a million gallons of their waste
into the surrounding soil 3. These tanks continue to leak and pose a pressing immediate
problem. There are about 250 million curies of radioactive material in all the holding tanks.
This is five times the amount of radioactivity released into the environment by the
Chernobyl accident. The waste stored in the cribs has overflowed and is moving towards
the aquifer underlying the Hanford site. In addition the movement of radioactivity in the

soil is greater than expected and is flowing towards the Columbia river. This river located
to the north and east of Hanford, and the Yakima river, in the south, are the sources of
drinking and irrigation water for the entire southeastern Washington area
Another problem that must be faced is the danger of explosions in the tank. The
presence of all the organic materials and water, coupled with high radioactivity, results in
radiolytic decomposition to generate significant volumes of H2. Smaller amounts of
nitrogen oxides are also formed. Normally these gases are ventilated out of the tanks.
However, in some tanks the gas is trapped between the sludge at the bottom and salt cake
(largely sodium nitrate) which formed over the sludge. These gases increase in pressure
raising the salt cake and then being violently released as a huge bubble. Another potential
explosion hazard is present in about 20 tanks to which ferrocyanide had been added to
precipitate 137Cs. The ferrocyanide is a reducing agent whereas the nitrates and nitrites are
oxidizing agents. The combination could produce enough heat to cause a violent explosion
of the H2.
ACTION PLAN FOR REMEDIATION
Battelle Pacific Northwest Laboratory (pNL) has been given the responsibility for
long range research on environmental remediation and waste management at Hanford. The
most urgent items are to be dealt with on a short time scale. The greatest danger to human
beings is the threat of a violent explosion in the tanks. Such an eventuality could release
large amounts of radioactive species and heavy metals to the surrounding area. A second
threat stems from the migrations of contaminants to the aquifer or to the surrounding rivers.
PNL, in conjunction with DOE, has developed several strategies to deal with these
problems. Our focus in this paper is to deal with problems connected with the disposal of
the present contents of the holding tanks and will not be concerned with soil remediation.
The general strategy outlined by Battelle is shown in Figure 1. High level waste will be
contained in a borosilicate glass whereas the low level waste will be enclosed in cement.
The salt cake consists largely of sodium nitrate and nitrite and the sludge of insoluble
hydroxides, but may also contain hydroxynitrates. Since the salt cake may contain trapped
radioactive species it is to be dissolved and combined with the liquid waste. The major
high level radioactive species in this combined solution is 137Cs and its removal would
allow the remaining solution to be treated as low level waste (LLW) and formed into
cementitious grout. Removal of the Cs+ is a formidable task because the solutions are
about 8M in Na+ and contain considerable amounts of AI, Cr, K+ and other elements as
shown in Table 1. Thus a highly selective ion exchanger or complexing agent is required.
The sludge and separated 137Cs is to be vitrified directly into a borosilicate glass.
Reference to Table 1 shows that the sludge contains large amounts of boron and silicon4 .
It is estimated that it would be required to prepare about 30,000 canisters of borosilicate
glass and to store them deep underground. The projected cost of this phase of the
operation is $30 billion as compared to $1.3-2 billion for the low level grout. Since the
sludge contains most of the radionuclides, an enormous cost savings would result if the
radionuclides could be removed from the sludge and treated separately. The remaining
LLW would go to much less expensive grout. Given the complex nature of the sludge, a
highly innovative chemical separations technology is required.
The sludge and separated 137Cs is to be vitrified directly into a borosilicate glass.
Reference to Table 1 shows that the sludge contains large amounts of boron and silicon4 .
It is estimated that it would be required to prepare about 30,000 canisters of borosilicate
glass and to store them deep underground. The projected cost of this phase of the
operation is $30 billion as compared to $1.3-2 billion for the low level grout. Since the
sludge contains most of the radionuclides, an enormous cost savings would result if the
radionuclides could be removed from the sludge and treated separately. The remaining
LLW would go to much less expensive grout. Given the complex nature of the sludge, a
highly innovative chemical separations technology is required.
Table 2 provides a summary of elements which would be targeted for removal from
the sludge along with Pu and U. If the noble metals could be recovered efficiently, without
creation of additional waste, then a valuable resource would be gained. Thus, even
thoughplans have been formulated and are proceeding to dispose of defense waste, there
still remains a strong incentive to find a technologically better, more efficient solution to
290
Table 1. Quantities of Chemicals in Single-Shell Tanks.
Uquicl + Sail Cake Sludge Total Total

Element a Mole a Mole a Mole SJ Mole
Ag 281 383,000 383,000

AI 1,880,000 288,000,000 287,810,000 83,000,000
Am 0.01 300 300
As 0 7&,200 7&,200
I 2,480 82,300,000 82,300,000
I .. 103 4,020.000 4,020.000
Ii 1,580 1,040.000 1,040,000 1,250,000
Ca 5,110 3&,800,000 35,800,000 3,200,000
Cd 8112 880,000 880,000 38,000
CI 0 1,010,000 1,080,000 1,130,000
CN 0 3111,000 351,000 8,800,000
Co 2 11,810 5,1110
Co 3 42.800 42,800
CO 20,300,000 27.700.000 48,000.000 27.000.000
Cr 400.000 7.540.000 7.840,000 1.800,000
Cs 4.470 350 4.820
Cu 77 550.000 5150.000
F 0 1.840.000 1,840.000 42.400.000
Fe 477 18.1100.000 88.500.000 11.200.000
Hg 0 247.000 247.000 4.500
I 313 2& 338
K 8,510,000 0 8.510.000
Mg 98 28.800.000 21.800.000
Mn 10.300 '.880.000 •• 870.300 2.200.000
Na 808.000.000 182.000.000 1.000.000.000 2.230.000.000
NI 2.710 3.840.000 3,842.000
NO. 74.100.000 13.000.000 87.100.000 102.000.000
NO. 100.000.000 415.100.000 1411.100.000 1.540.000.000
P 73.200 32.300.000 330.320.000 81.000.000
.... 1.180 1.200.000 1.200.000
Pel 0 80.800 82.800
..
Pu 1 72.100 72.100
...."" 0
0
21.000
88.000
21.000
",000
884 18.300 18.300
II 18.800 288.000.000 288.000.000 15.300.000
SO. 0 8.180.000 8,180,000 5,200,000
S. 1 30,100 30,100
S.
To
•
23,725
177.000
3'.125
177.000
58.400
17,200,000
U 870 7.280,000 7.280.000

Zn 52. 282,000 212,000
Z. ,.,,0 28.000,000 28,000.000 2,700,000
-ED! -ED! m3
Volume 11,1540 47.'10 131,1500
alec aI_
Density 1.37 1.152
291
N
CD
N
QROUI'
.......... ·'.r
..... ,,:~~......_...........................................
~ -= "'6---'"'_......_.1i
Figure 1. Schematic diagram of a disposal option for buried radioactive tank wastes. (Ref. 4).
Table 2. Possible Targeted elements from Old Hanford Waste
• Sr-90, Cs-131: power or irradiation sources(Separation treated in Oth.-

Sessions)
• Noble Metals (Pd, Ru, Rh): Strategic metals, catalysts
• Tc-99: Hazardous waste, catalyst
• Other Specialty Isotopes:
Mo-98 as target for Mo-99 to in tum generate Tc-99m

Zr-90 low neutron cross section isotope
Y-90 from Sr-90 for medical use
Others? (See handout for summary of isotopes)
the problem. It is estimated that the application of advanced chemical separations could
save $6-10 billion just by reducing the high level waste volumes from the sludge alone 4.
NEW SORBANTS AND ION EXCHANGERS FOR USE IN NUCLEAR

WASTE TREATMENT
Specific Exchangers for Cs+ Removal.

In the past, zeolites such as chabazite and mordenite were used to remove 137Cs+
from waste streams even though the uptake of the exchanger for cesium was small in the
presence of high levels of sodium ionS-? We note in Figure 2 that the distribution
coefficients for Cs+ on these zeolites decreases rapidly as the concentration of Na+ or K+
increases8. At the level of 8M Na+ only about 1% of the sites accept Cs+. The cesium is
also difficult to remove from the exchanger. We have prepared a series of
monphenydiphosphonate phosphates of zirconium of general formula
Zr(~PC;;14P03>X<HP04h-2x. A schematic drawing of their structure is shown in Figure
3. These compounds were first prepared by Dines et al9 . In Table 3 we have collected the
distribution coefficients for a series of the phosphonate phosphates in which X varies from
K K
ZEOLON 900 LINDE AII-500
0.001 0.01 O. 1 1 10 0.001 0.01 O. 1 1 10

CONCENTRATION e.ol/1] CONCENTRATION [~ol/l]
Figure 2. Distribution coefficient (Kd) of trace cesium on synthetic zeolites Zeolon 900 and Linde A W-
SOO from a nuclear waste solution ([Na+hnitial = O.S mmol/L) as a function ofNa+ and K+ concentrations.
Equilibration time 20 h. The ratio of solution volume to solid weight 100 mLlg. Zeolite grain size < 0.32
mm. (From Ref. 8 with permission).
293
Table 3. Kd values in 10-3 M solutions of the ions for a series of zirconium
phenylphosphonate phosphates, Zr(03PC61"4P03)x:(HP04h-2x from x=O.6S(IP-l) to
O.lS(IP-S).
Sample Ion Li Na K Rb Cs Mg Ca Sr
No. ~(mVg)
IP-l <1 <1 <1 38.0 65.1 26.8 21.8 208.0
lP-2 <1 2.4 10.9 140.4 276.6 44.0 27.3 139.5
lP-3 5.8 5.0 96.9 288.0 2091 188.0 456.7 594.9
IP-4 1.9 2.4 122.8 552.5 3347 319.8 586.8 613.0
IP-5 <1 <1 133.0 220.0 4491 187.2 361.0 460.8
• Zr
• C
o P (0 ) (b)
o 0
Figure 3. Schematic representation of cross-linked ziIconium arylphosphonate phosphates: (a)
Zr(03PC6H4P03)0.s(HP04); (b) Zr(03PC6H4-C6H4P03)0.5(HP04).
294
E(meq/s) I
O'~--~2----~4----~6~--~8----~1~O~pH
Figure 4. Potentiometric titration curve for Zr(03PC(jH4P03)0.4S(HP04h.l as uptake versus pH.

Titrant: O.lM(KCl + KOH).
0.65 to 0.15. We note that the uptake for Na+ and Li+ ion is negligible but that I«I for Cs+
is quite high. In the case of sample IP-5 Na+ was almost completely rejected and the
affinity for other alkali and alkaline earth ions, except Cs+ is moderate. The ion exchange
I
capacity (IEC) for IP-5 is 3.3 meq/g as seen by the K+ titration curve in Figure 4. These
phosphonate compounds are extremely stable to acid since they can be sulfonated in fuming
sulfuric acid and are thermally stable to at least 350°C in air. At high pH values, phosphate
groups are removed by hydrolysis so these exchangers are best utilized at near neutral pH
values.
The interlayer distance in the monophenyl derivatives is 9.6 A while the layer
thickness is 6.3 AlO. Thus, the free space is 3.3 A which is just sufficient to allow an
unhydrated Cs+ to diffuse into the lattice (rcs+ = 1.7 A). Since the hydrated radius of Na+
is larger than 1.7 A, there must be insufficient energy available at room temperature to
dehydrate the Na+ ions. The distance in the lateral direction between pillars is 5.3 A for
near neighbors, but if spaced by phosphate groups this value is 10.6 A. The ring van der
Waals radius is about 3.4 All so the lateral free space is -7.2 A.
A series of compounds with diphenyl groups has also been prepared. For these
compounds the interlayer distance is 13.6 A, large enough to allow hydrated ions to diffuse
into the interlayer space. The I«I values for M+ and M2+ ions on these compounds are not
remarkable, but in general increase as the size of the ion increases. However, on
sulfonation these compounds show a remarkable specificity for Ba2+ (Table 4). This
specificity must be chemical in nature because the values for the other alkaline earth metals
were all roughly similar and relatively low. The high Ba2+ selectivity may result from a
combination of free space and strong binding of Ba2+ ion to S03-. This is illustrated in
Figure 5B where sulfonate groups from adjacent pillars are in a position to complex a Ba2+
ion between them.
295
Table 4. Kd values in to-3M solutions of the ions for a series of sulfophosphonates,
Zr(03P~S03H~SOJHP03MHP04h-2x, for which x decreases from O.S to 0.5
down the column.
No. Kd(ml/~
2PS-I 6.7 2.3 54.7 96.0 154.5 70.1 111.3 135.2
2PS-2 22.2 31.8 217.0 415.2 586.5 268.4 221.0 380.0
2PS-3 19.3 34.1 401.0 831.5 1915.9 220.0 303.3 289.0
2PS-4 27.5 52.1 521.0 1052.5 2853.5 341.5 430.5 453.8
2PS-5 11.6 19.6 354.1 1052.5 2440 91.2 108.7 117.2
Table 5. Distribution coefficients for alkali and alkaline earth metal ions on
exchangers Zr(03~S03HMHP04h-x (2 samples), amorphous zirconium
phosphate, and AG 50W-XS at 25°C.
Kd(mlJg)
Ion MY-IV-95a MY-VI-2a Amorphous Zrpa,b AG 50W-X8'
x - 0.767 x - 0.43
Li+ 110 7 33
Na+ 205 11 54
K+ 1500 650 120 99
Cs+ 6500 1600 148
~g2+ 21000 9800 790
Ca2+ 89000 37000 1450
Ba2+ 400000 190000 5000
a Kd at pH = 2.00 and a metalloading of 0.1 meq/g.
b Calculated from selectivity coefficients given in Reference [14].
c Kd in 0.1 M RN03. Data from Reference [15] .
• Zr
• C
o P
.0
(a)
Figure 5. Schematic representation of sulfonated zirconium pheny1phosphonates: (a)

Zr(03PC6H4S03H)(HP04)
296
• Zf
• c
o P
.0
(b)
We have also prepared zirconium sulfophenylphosphonate phosphates of the type

shown in Figure 5a. These compounds swell to exfoliation in water but reprecipitate in the
presence of polyvalent cations trapping them between the layers. Some ion exchange data
for these type compounds is shown in Table 5. 12 It is seen that, while these compounds
have a high affinity for Cs+, they might be ideal for Sr2+ sequestration as described in the
next section.
Strontium Specific Ion Exchangers

In the data shown in Table 5 it is clear that Kct increases rapidly from Mg2+ to Ba2+.
Since 10gK versus the ionic radii of the alkaline earth cations are straight lines for the MY
samples, we estimate values of Kct for MY-IV-95 and MY-VI-2 to be 158,500 and 75,800,
respectively for Sr2+. Compounds such as these would be broad spectrum sequestrants.
That is, other polyvalent species would also be sequestered, but if their Kct values are much
lower than those for Sr2+ complete removal of Sr2+ is possible.
Another compound exhibiting very high selectivity for Sr2+ is of the type shown in
Figure 5b but with 3 phenyl rings partially sulfonated. Some preliminary Kd values are
collected in Table 6. it is seen that sample 3PS-2 has an extremely high Kct value for Sr2+
but only moderate values for Ca2+ and Ba2. This exchanger, being cross-linked, is totally
insoluble in strong acid solutions and may also be stable in alkaline solution. Thus, it
might be very suitable as a specific Sr2+exchanger. This supposition is strengthened by the
fact that these terphenyl cross-linked materials are rather indifferent in their affinity for
some first row transition elements (Table 6).
SEQUESTRATION OF COMPLEXES
The aqueous solutions present in the tanks contain large quantities of complexing
agents. Thus, many of the metals in solution may not be present in the hydrated state but
rather as metal complexes. The sulfonates shown in Figure 5B remove a broad spectrum
of such positively charged species from solution. For example, we have precipitated
297
Table 6. K<t values in lo- 3M solutions for the listed ions for zirconium sulfoterphenyl
diphosphonate phosphate with increasing phosphate content
down the column.
No. l<d~mL/g)
3P5-1 17.9 18.3 169.0 270.3 4366.9 1783.8 12680
3PS-2 32.3 31.7 201.8 571.6 8900 1880 63516
3PS~3 33.9 47.7 381.7 851.4 9592.3 1499.9 5523.5
3P5-4 21.8 47.1 308.8 6446 10400 1623.5 3732
species of the type [M(NH3)6]n+from solution with the sulphophosphonates. A second

type of compound that sequesters anionic species has also been prepared. They are
aminophosphonates of general formula Zr(03PCH2(NHCH2CH2)nNH2h- We have
synthesized compounds with n=1-4. On protonation these compounds form colloidal
dispersions which reprecipitate in the presence of polyvalent anionic species. Using these
polyimines [PtC4]2-, Fe(CN)63-, Fe(CN)62-, [PMo12040]3-, W04 2-, polyimines have
been quantitatively removed from solution. Many similar species are present in the tank
solutions as well as anionic technecium species and these could be recovered by application
of the polyimine broad spectrum precipitating agents.
CONCLUSIONS
We have demonstrated the synthesis of new materials with high selectivities for
Cs2+ and Sr2+ and low affinity for Na+. In addition sequestrants for both anions and
cations present in nuclear waste solutions have been prepared. It is now necessary to
demonstrate the utility of these materials in simulated nuclear waste solutions.
Acknowledgment: The author wishes to thank the State of Texas for a grant through its
"Advanced Technology Program" in support of this research. Grateful acknowledgment is
made to the authors scientific collaborators C.-Y. Yang, C. Yolanda Ortiz-Avila, Chhaya
Bhardwaj and especially Guangzhi Peng and the U. S. China Program of NSF through
Grant No. INT 8910902.
REFERENCES
1. Office of Technology Assessment, U. S. Congress, "Long-lived Legacy: Managing
High-Level and Transuranic Waste at the DOE Nuclear Weapons Complex,"
OTA-BP-)-83, U. S. Govt. Printing Office, Washington, D.C. (May 1991).
2. B. G. Levi, Physics Today, March 1992, 17.
3. Govt. Accounting Office "Nuclear Waste: Hanford Single-Shell Tank Leaks Greater
than Estimated," GAO/RCED-91-77, Washington, D.C. (1991).
4. J. R. Morrey and J. L. Swanson, "A Primer on Hanford Defense Tank Wastes and
Prospects for Advanced Chemical Separations," Battelle Pacific Northwest
Laboratories, April 1991.
5. D. o. Campbell, E. D. Collins, L. J. King, J. B. Knaur and R. M. Wallace,
ORNL/fM-7448 (1980).
298
6. E. D. Collins, D. o. Campbell, L. J. King, J. B. Knauer and R. M. Wallace, AIChE.
Symp. Ser. 213, 78, 9 (1982).
7. L. L. Ames and J. L. Nelson, Report HW-74609 (1962) Battelle PNL, Richland,
Wash.
8. R. Harjula and J. Lehto, Nucl. Chern. Waste Mgt. Q, 133 (1986).
9. M. B. Dines, P. M. DiGiacomo, K. D. Callahan, P. C. Griffith, R. H. Lane and R. E.
Cooksey in Chemically Modified Surfaces in Catalysis and Electrocatalysis, J. S.
Miller, Ed., ACS Symp. Ser. 192, ACS, Wash., D.C. 1982, P 223.
10. A. Clearfield and G. D. Smith, Inorg. Chern.~, 431 (1969).
11. L. Pauling, "The Nature of the Chemical Bond," 3rd ed., Cornell University Press,
Ithaca, N.Y. 1960.
12. L. H. Kullberg and A. Clearfield, Solvent Extraction and Ion Ex. 1, 527 (1989).
13. G.-Z. Peng and A. Clearfield, work in progress.
14 L. Kullberg and A. Cleartield,l. Phys. Chern.~, 1578 (1981).
15. F. W. E. Strelow, R. Rothmeyer and C. J. C. Bothma, Anal. Chern.ll, 106
(1965).
299
APPENDIX I
Abstract of a lecture not submitted in chapter torm:
ADVANCES IN BIOREMEDIATION OF ENVIRONMENTAL CONTAMINANTS
K. W. Brown and K. C. Donnelly
Soil and Crop Sciences Department

and Department of Vet. Public Health
ABSTRACT
Many organic chemicals readily undergo biodegradation in the environment.

While higher organisms can degrade some organic chemicals, microorganisms play
the major role in degrading naturally occurring and contaminant organics in surface
waters, soils, and groundwaters. The necessity of developing economical,
environmentally safe methods of decontaminating sites of spills associated with
industrial activities and post waste disposal activities, as well as the need to develop
better methods for the disposal of wastes we are presently producing has resulted in a
great increase in the interest and in the use of biodegradation.
Many of the efforts have involved the use of procedures to alter the
environment to enhance the rate of biodegradation. The date, less success has been
achieved by the introduction of selected microorganisms, although some organisms
have been developed which are capable of degrading xenobiotic chemicals. The
enzymatic mediated pathways of biodegradation of many environmental
contaminants are now understood. The environment contains large number of
diverse microbial species, many of which may be involved simultaneously or
sequentially in the degradation of a given compound. The biodegradation rate
constants are now known for an increasing number of organic chemicals in a variety
of media including surface waters, soils, anaerobic sediments and groundwater.
Information is being developed on the influence of environmental parameters
including temperature, nutrient supply, substrata concentration, the presence of
associated substrates, redox potential, salinity, etc. which can be used to predict rates of
biodegradation in the field and the response of the rates to environmental
manipulation.
Readers wanting more information on the subjects described in this abstract are
invited to communicate directly with Dr. Brown.
301
APPENDIX II
POSTERS
S. N. Ahmadi, Y. C. Huang, S. S. Koseoglu and B. Batchelor, Food Protein Research,

Texas A&M Univeristy
Micellar Enhanced Ultrafiltration of Heavy Metals Using Lecithin
Dr. Ramesh C. Bhardwaj, Department of Chemistry, Texas A&M University
Electrochemical Oxidation of Waste
Dian Chen, Ramunas J. Motekaitis, Derek McManus and Arthur E. Martell,

Department of Chemistry, Texas A&M University, ARI Technologies, Palatine, Illinois
LoCAT Process for the Removal of H2S from Natural Gas with
Fe(llI)-NTA as Catalyst
Tanya Lewis and David H. Russell, Department of Chemistry, Texas A&M University
Characterization of Water Soluble Organics: High Resolution,

Mass Spectrometry
J. B. Shapiro and Emile A. Schweikert, Department of Chemistry, Texas A&M

University
Methodoly for Direct Field Analysis of Organic Contaminants
Wilfredo Delgado-Morales, Mysore S. Mohan, J. Drew Ilger and Ralph A. Zingaro
Identification of AnalYSis of Arsenic Compounds in Natural Gas
Eric J. Munson, Ali A. Kheir, Greg Oliver and James F. Haw, Department of Chemistry,
High- Temperatures In Situ Solid-State NMR Studies of Catalytic Systems

Sharon Taylor-Myers, David B. Ferguson and James F. Haw, Deparment of Chemistry,
Motion of Small Molecules in Polymers by 15N NMR
Jerry L. White, Larry W. Beck and James F. Haw, Department of Chemistry,

Hydrogen-Bonding and Exchange in Zeolites by Mas IH NMR
303
INDEX
Acclimation, 280 Baffles,S

Accumulated acidity, 282 Basic flux equation, 129
Acetate, 140 Batch resin tests, 149
Acetic Acid, 236 Benzoic acid, 234, 241, 242
Acetone removal, 129 Bilfiltration, 273
Acetonitrile, 27, 31 rates, 268
Acrolein, 28 Bio-oxidation rates, 264, 266, 268,
Acrylonitrile process, 27, 31 269,271
Acrylonitrile, 140 Biodegradation rate constants,
Activated sludge, 249 301
Acyl chlorides, 57 Biodegradation, 301
Adsorption equilibrium Biofilm, 248, 250
constant, 160 Biofilter capacity, 281
isotherm, 160 materials, 277
heavy metals, 269 Biological degradation, 22
process, 160 oxygen demand, 249
Adsorption-based separation, treatment, 247
137 Biomass availability, 265
Advanced oxidation processes, Biomass,253,256,258
171 Bioreactor, 262, 269
Aerobic biodegradation, 273 Bioremediation, 2, 219
Air pollution, 273 Bleed stream, 11
toxics, 273 Borosilicate glass, 290
Alginate bead bioreactors, 267, Broad spectrum sequestrants,
268,271 297
Alkali metals, 208 Bulk chemicals, 107
Alkaline, 219 density, 277
Alkoxide nucleophiles, 184 t-BuOO-/02-·, 197
Allyl chloride process, 35 Butyric acid, 236, 239
chloride, 35 By-products, 5
toluidine, 108 C/N ratio, 280
Amberlite, 150 C6Cl& 203, 204, 206, 207, 208
Amino acids, 236
Analysis of Arsenic Cadmium plating, 81
Compounds, 303 uptake, 269
Aqueous waste, 154 Calcium alginate gel beads, 267
Aroclor 1242, 199, 201, 1268, 190, Calcium alginate, 271
191, 194 Caproic acid, 236, 239
Aromatic chemistry, 109 Carbon dioxide removal of, 127
Assessment Phase, 76 Carbon tetrachloride, 57
Asymmetric carbonylation, 118 Carbonylation, 115
membranes, 131 Carboxylic acid waste, 239
Atom utilization concept, 99 Carboxylic acids waste stream,
Atomic metals, 208 240,241
Axial dispersion, 165 Case study at Pratt and
Backmix reactor, 6 Whitney, 80
305
Catalase, 266 Continuous stirred tank
Catalysis, 89, 99 reactor, 66
Catalyst, 5 Control technologies, 273
Catalytic hydroprocessing, Control,7
retrosynthesis, 114 Controlled-potential
Cavitating, 13 electrolysis, 204, 208
Celgard membrane, 237, 241 Copper removal, 143
Chabazite, 293 Copper, 136
Chemical absorption step, 127 Cost of waste treatment, 41
industry, 100 Cost, 20
Chemical Kinetic Mechanism Cross-exchange, 11
for C2HCl3' <,9 Cross-linked zirconium
kinetic mecltanisms, arylphosphonate
46 phosphates, 294
Chemistry, 23 Cryogenic absorption, 128
Chemoselectivity, 100 137CS+, 293
Chlorinated aromatic 137Cs,290
molecules, 202 Cs-137,293
benzenes, 94 Cyclic voltammograms for
hydrocarbons, 48, 213 dissolved O 2, 193
organics, 94 voltammograms, 202
Chlorine, 36 DDE,l88
Chlorine-catalyzed, 48 DDT,187
Chloroalkenes, 188 Dechlorination, 97, 213
Chlorobenzene (C6HSCl), 183, Degradation of Phenol, 268
198,208 Dehydrohalogenation, 185
Chlorobenzenes, 222 Denitrification, 249
Chlorobiphenyl, 205 Design of more efficient
Chloroform, 198 processes, 65
Chlorohydrin technology, 100 Design, 274
2-Chlorophenol, 140 Destruction of C 2Cl4, 56
ChI oro phenols, 97 C3Cl6,59
Chromium plating, 81
removal, 143 Desuperheat, 7
Chromium, 136 Detailed Chemical Kinetic
Co-solvents, 157 Mechanism, 45
Coefficient, 165 Dibromoalkanes, 187
Collection, 283 1,l-Dichloroethene, 188
Column control, 12 Dichloroacetyl chloride, 57
experiments, 150 l,2-Dichloropropane, 36, 38
pressure, 11 1,3-Dichloropropane, 36, 38
Column, 10 Diffusion of H2S, 285
Combustion of C 2HCl 3, 56 Dilute industrial waste gas, 285
Component mass balance, 30 Dioxins, 181, 182, 217
Composite Celgard Disinfection, 171
membrane, 235 Dissociative Bond Energies, 210
Composite celgard 24(, 234 Distillation flowchart, 39
hollow fiber membranes, 238 Distillation, 10
membranes, 131, 229, 232, Distribution coefficient, 155
polysulfone hollow Distribution coefficients for
fiber membrane, 238 alkali and alkaline earth
Composts, 277 metal ions, 296
Compounds, 285 Dow Chemical Company, 3
Computer control, 7 Downtime, 14
Conditioning period, 280 Dowtherm,9
Contaminated soils, 154 Drinking water, 171
306
Dry-out compost, 284 Formic Acid, 232, 236
Drying, 12 Fouling, 7
Economic considerations in Friedel-crafts acylations, 114
Dechlorination, 222 Fugacities, 155
Economics of Remediation, Fugacity coefficients, 155
223 Fugitive emissions, 9
Economics, 213 Full scale biofilter design, 279
Effective diffusion coefficient, Furnace coils, 9
234 Furnaces, 9
Effective diffusivity, 165 Gas chromatograms, 205
Electric tracing, 13 system, 278
Electrochemical oxidadon, 303 retention time, 281
Electrodialysis, 79 samples, 283
Electrogenerated atomic separation membranes, 131
sodium, 209 velocity, 281
Electrolysis, 79 Gel-immobilized bioreactor,
Electrolytic reduction, 194,201, 267,268
203 Glycolate dechlorination, 213,
Electron, 201 215
°
Elementary reactions of
C2HCl3 and 2, 52
Employee awareness, 16
Glycolate, 219
Glyphosate loss, 152
Glyphosate, 147
Energy conservation, 3 Gold,83
diagrams, 47 Granular activated carbon, 250
Entrainer effect, 157 H 2S control, 285
Entrainers, 157 H 2S, 275
Environmental contaminants, Halocarbons, 197
301 Halogenated aromatic
Law and Regulations, 75
remediation, 153 hydrocarbons, 185,208
Hazardous halogenated
EPA Priorities, 73 hydrocarbons, 208
Epichlorohydrin, 35 waste inventories national,
Equilibrium constant, 155 90
Equilibrium-based process, wastes, 2, 171
135 HCB,206
Equipment fouling, 14
Esters, 197 HCN, 28, 31
Heat capacities, 30, 37
Ethyl cellulose, 234
Ethylene epoxidation, 100 exchangers, 7
Evaporati<?n, 79 transfer medium, ~)
Exchange in Zeolites, 303 Heavy metal uptake, 269
External corrosion, 15 Helium-methane system, 125
Extraction, 153 Hexachlorobenzene (HCB), 2,
Fe(III)-NTA as Catalyst, 303 181,188, 198,
Feasibility AnalysiS Hexachlorobenzene-KPEG,221
Feed distribution,S, 11 Hexane, 140
quality,4 HFCLM permeator, 144
tray, 10 High flux tubes, 11
Field studies, 275 Hollow fiber contained liquid
Filtering, 12 membrane (HFCLM)
Fine chemicals, 107 permeator, 135, 143
Flames of C 2HCl3, 56 membranes, 123,230,237
Flat sheet membrane, 230, 237 system, 244
Flat Sheet, 241 Hollow fiber, 241
Flow cell system, 231 Hollow-fiber module, 136, 137,
Fluidized bed reactor, 32, 247 141, 142
307
Hybrid bioreaction system, 269 Light impurity, 11
separation system, 127 Linde Aw-500, 293
Hydrazine, 206 Lined pipes, 13
Hydride ion (H:-), 198 vessels, 13
substitution, 198 Liquid ion exchanger, 143
Hydrodechlorination, 90, 94 membranes, 135,234
Hydroformylation, 115 LLW, 290
Hydrogen peroxide, 174, 266 Loading rate, 281
Hydrogen recovery, 128 Low level waste (LLW), 290
sulfide, 273 Main columns, 151
Hydrogen-Bonding in Maintenance procedures, 284
Zeolites, 303 Management commitment, 24
Hydrogenation, 115, 206 manufacture, 106
Hydrogenolysis, 199 Mass separating agent, 126
Hydroperoxide ion (HOO-), Mass transfer coefficient, 165,
196 234,243
Hydrophobicity, 234 Materials, 13
Hydroquinone, 111 Maximum loading capacity,
Hydroxyl radical, 172 281
Ibuprofen manufacture, 113 Me4N(02), 192
Immobilized cell technology, Measurements, 275
248 Membrane bioreactor, 263, 264
Implementation, 77 biosupport system, 267
Impurities, 4 permeation, 130
Incineration, 1, 23, 182 reactor, 262, 271
Industrial applications, 285 solvent extraction removal
Industrial Toxies Programs, 20 of toluene, 137
Industries, 71 stirred cell, 243
Industry, 75 support, 241
Inhibitors, 4, 175 Membrane technologies, 23, 229
Inorganic ion exchange Membrane-based solvent
materials, 289 extraction, 135, 136, 140
Insulation, 10 Membrane-bound
Interaction of solutes with microorganisms, 261
membranes, 125 Membrane-immobilized
Ion exchange resins, 147 bacteria, 261
system, 150 Menadione synthesis, 111
Ion exchange, 79, 147 Metal finishing, 71
IPAc, 140 Metallurgy, 13
Jacketing, 13 Methoxychlor, 187
K value, 155 Methyl isobutyl ketone, 140, 234
Kanapaha WWTP, methacrylate manufacture,
Gainesville, FL, 277 102
Kinetic model, 66 methacrylate, 105
models, 45 Microbial activity, 285
studies, 285 Microfiltration hollow fibers,
Kinetics, 6 230
K02,192 Microor,ganisms, 247, 301
KPEG dehalogenation, 184 Microporous hollow
KPEG,213 fiber(MHF) module, 136
Laboratory studies, 275 hollow fiber, 140
Land-fills, 1 membrane, 135
Lead adsorption, 269 membrane-based separation
biosorption, 270 processes, 135
Leadership, 21 membrane-based solvent e
Leaks,13 extraction, 138
Lecithin, 303 Microporous membranes, 131,135
polypropylene, 136
308
Mineralization, 281 Overall mass transfer
Mitsubishi gas process, 104 coefficient, 245
Mixed liquor volatile Overhead condenser, 11
suspended solids, 249 Ownership, 24
Mixing,S Oxidative-pyrolysis, 48
MLVSS, 251 Oxo, 192
Monitoring vents, 13 Oxy anion bases, 195
Monitoring, 8 anions, 192, 198
Monphenydiphosphonate Oxygen consumption, 261
phosphates of zirconium, Oxygen, 266
294 Oxygenation, 249
Mordenite, 293 Ozone generation, 173
Multiple feed trays, 10 Ozone, 171
NaBH4,199 Ozone-hydrogen peroxide
NET AC Program, 220 ratio, 176
NiCl 2,199 p-Nitroaniline, 234
p-nitrophenol (PNP), 236
Nitrification, 249 Packed biofilter towers, 274
p-nitrophenol, 241, 242 Packed biofilter, 275
Nitrobenzene, 140 Partial condenser, 12
15N NMR, 303,94
Particle density, 280
Non-corroding Tubes, 8 Partition coefficient, 155
Nondispersive solvent PCB isomers, 194
extraction, 135 PCB, 190,213
Nonporous membranes, 126 PCB-contaminated soil, 184
North American Membrane PCBs, 181, 182, 208
Society, 121 Pentachlorophenol, 221
Nuclear fuel, 289 Periodic washings, 283
waste effluent, 289 perm-selective, 121
Nucleophiles, 192 Permeability, 234
Nucleophilic degradation, 190 Permeator shell and tube
dehalogenation, 183 type, 123
reactions, 186 Permeator, spirally-wrapped,
substitution by O 2 -',187 123
Nucleophilicity of O 2 -,, 185, 196 Permeators, 122
Nucleophilicity, 192 Pervaporation, 131
Nutrients, 280 pH,277
O 2 -',197 Pharmaceuticals, 107
Odorous emissions, 273 Phase, 77
Off-gas sampling, 283 Phase-splitting technique, 154
Off-spec material, 4 Phenol oxidation, 267
Oil refining, 107 transport, 235
On-line cleaning, 8 Phenol, 140,234,264,267
control, 14 Phenolic Waste, 236, 237
stream monitors, 14 Phenoxyacetic acids, 102
Operating parameters, 279 Phosgene, 56
Operational variables, 274 Pipe sizes, 12
Operations, 80 Piping, 12
Optimizing processes, 14 Planning and Organization
Organic contaminants, 303 Phase, 76
matter, 277 Plat use, 151
synthesis, 107 Plug flow reactor, 6, 39
transport, 243 Plutonium production
Organics separation, 241 reactors, 289
Organo-sulfur, 285 Plutonium, 289
PNP recovery, 237
removal, 238
309
PNP, 236,242 Pyrolysis of C2HCl3, 46
Pollutant overloading, 284 Quality control, 284
Pollution prevention, 25 Radical recombination
Polychlorinated aromatic reactions, 60
molecules, 181 Radioactive tank wastes, 292
biphenyls, 159 waste, 289
Polychloro-aromatics, 189 Radionuclides, 290
Polychlorobiphenyls (PCBs), Rate of permeation, 125
188,200 profiles for C2HCl3' 57
Polychlorodibenzofurans, 217 profiles for COCl2, 59
Polycyclic aromatic Raw materials, 4
hydrocarbons, 159 Reaction engineering, 89
Polyethylene, 234 parameters, 36
Polyhaloaromatics, 188 pathways, 47
Polyhalogenated aromatic systems analysis, 28
wastes,206 systems constant, 29, 37
hydrocarbons, 186 Reactor comparison, 34
Polymer membranes, 125 conversion,S
Polymeric membrane, 131, 234 design, 68
Polymethylsilane- Reactors,S
polycarbonate copolymer, Reboilers, 11
234 Recirculated fluid tracing
Polypropylene fiber systems, 13
membrane, 239 Recirculation flow rate, 264
hollow fiber membrane, 240 Recovery of metals, 83
Polysulfone hollow fibers, 230 Recovery technology, 79
Polysulfone, 237 Recycle/reuse, 21
Polyvinylidene chloride, 234 Recycling, 6, 73, 247
Pore volume, 277 Redox potentials, 195
Potassium dichromate, 108 Reduction of hazardous
Potentiometric titration curve wastes,84
for arylphosphonate Reductive dechlorination
phosphate, 295 electrolysis systems, , 201, 214
Precious metals Reflux Ratio, 10
Precolumns, 151 Refrigerated condenser, 12
Preheat, 11 Relative reactivities, 196
Preparation of membranes, Relative reactivity of MeO- /
125 Ho-/MePEG-/HOO- , 192
Pressure drop, 280
Pressure-flow relationship, 265 Release reduction hierarchy,
Priority organic pollutants, 142 21
P:iority pollutant, 136, 138 Release reduction, 19
Process control, 7 ReMED A&M, 200
efficiency 247 Remediation strategies, 182
modifications, 25 Removable insulation, 15
reactions, 45 Removal efficiency, 279
Product changeovers, 14 of carbon dioxide from
degradation, 13 methane gas, 122
loss, 14 of copper, 145
Projects, 16 of H2S from Natural Gas, 303
Properties, 277 of heavy metals, 143
Propionic Acid, 236 Removing metals, 261
Propylene, 27, 35 organics, 261
PRSV equation of state, 40 Repacking column, 10
Pumped recirculation,S Resin life test, 151
PUREX process, 289 Responsible care, 19
Purge, 8 Retraying column, 10
310
Reverse osmosis, 79, 124 Stirred cell system, 231
Reversing brushes, 8 Storage, 13, 283
Rolling film reboilers, 11 Strategy, 22
Sample size, 15 Strip aqueous solution, 143
Sampling, 15, 275 Stripping, 12
Sand, 250 Strontium specific ion
SARA Title III, 20 exchangers, 297
Scaleup factor, 151 Styrene-butadiene copolymer,
science, 121 234
Scraped-wall heat exchangers, Substrate inhibition, 267
8 Sulfonated zirconium
Seal flush, 8 phenylphosphates, 296
Seal-less pumps, 9 Sulfur accumulation, 281
Secondary reactins, 5 Sulfuric acid, 282
Selectivity in ion exchange, Supercritical extraction, 153
148 fluid, 153
ratio, 129 Superfund sites, 1
Sensitivity analysis, 52 Superoxide ion, 184, 187, 192,
Separation of oxygen-argon- 198
nitrogen, 124 Surface area, 163
Sequestration of complexes, finishing
298 System deterioration, 284
Sewage sludge, 280 Taffinate, 138
Shipping Containers, 4 Tank cars, 15
Shutdowns, 14 trucks, 15
SIEBO stones, 278 Technologies for chlorinated
Siebo-stone air distribution hydrocarbon remediation,
system, 278 217
Silver, 83 Technology transfer, 22
Single-electron-transfEr (SET) Technology, 219
mechanism, 186 Temperature sensitive
Single-shell tanks, 291 Thermochemistry, 45
Sludge production, 248, 249 Thermocompressor, 7
Sludge, 154, 251, 257, 280 Thermodynamic (2nd law)
Small ions, 232 efficiencies, 9
Small particle content, 280, 281 Thin film membranes, 131
Sodium atoms, 209 Toluene removal, 136
borohydride~ 198 Toluene, 140
Soil organic content, 163 Total carbon, 280
remediation, 223 energy balance, 31
Soils, 280 nitrogen, 280
Solid waste, 154 sulfur, 280
Solid-State NMR Studies, 303 Toxic use reduction, 23
Soluble phosphorus, 280 Tracing systems, 13
Solvent recovery, 153 Transition state theory, 46
Source reduction, 3, 21., 73 Transport of carboxylic acids,
Sources, 277 236
Specific exchangers for Cs+ of organic compounds, 232,
removal, 293 234
Spills, 14 through membranes, 125
Sponge balls, 8 Trichloroacetyl chloride, 57
Sr-90,293 Trichloroethane, 183
Staged heating, 8 Trips, 14
Startups, 14 Types of common ion
Static mixer,S exchange resins, 148
Steam superheat, 9 Ultimate nucleophile, 201
tracing systems, 13 Ultrafiltration of heavy metals,
303
311
Underground injection wells, Waste, 14
20 Wastewater treatment plants,
Undersized pipe, 12 285
valves, 12 Wastewater, 154
Upsets, 14 Water content, 277
Uranium, 289 soluble organics, 303
UV-irradiation, 174 use reduction techniques, 82
Vacuum columns, 11 Water, 284
Valeric acid, 236, 239 Weakly basic anion resins, 149
Van't Hoff equation, 160 Wide Beach, New York
Vapor line, 11 Superfund Site, 222
Vents, 11 XD-7 (Silane), 234
Vinyl chloride, 94 Xenobiotic chemicals, 301
Volatile organic compounds Zeolite catalysis, 114
(VOCs),273 Zeolites, 293
Washing, 284 Zeolon 900, 293
Waste generation, 24 Zero-discharge, 22
management, 73 Zero-effl uent, 1
Waste minimization, 19, 25, 71, Zirconium
99 phenylphosphonate
Waste reduction phosphates, 294
opportunities, 77 Zirconium
reduction technologies, 78 sulfophenylphosphonate
reduction, 3 phosphates, 296
streams, 15 Zirconium sulfoterphenyl
treatment, 73 diphosphonate phosphate,
water treatment, 171 297
312

Industrial Environmental Chemistry - Waste Minimiz

Uploaded by

Copyright:

Available Formats

Industrial Environmental Chemistry - Waste Minimiz

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Industrial Environmental Chemistry - Waste Minimiz

Uploaded by

Copyright:

Available Formats

Industrial

APPLICATIONS OF ENZYME BIOTECHNOLOGY

CHEMICAL ASPECTS OF ENZYME BIOTECHNOLOGY: Fundamentals

DESIGN OF NEW MATERIALS

INDUSTRIAL ENVIRONMENTAL CHEMISTRY: Waste Minimization in

METAL-METAL BONDS AND CLUSTERS IN CHEMISTRY

OXYGEN COMPLEXES AND OXYGEN ACTIVATION BY

Springer Science+Business Media, LLC

Industrlal env lronmental chemlstry : waste mi n 1m1 z a t i on In industrtal

1 . Factory and trade waste--Congresses. 2 . Haste minlmizat 1on-

Proceedings of the Texas A&M University, IUCCP Tenth Annual Symposium

ISBN 978-1-4899-2322-6 ISBN 978-1-4899-2320-2 (eBook)

© 1992 Springer Science+Business Media New York

All rights reserved

Dow Chemical Company

Hoechs t-Celanese Corporation

Monsanto Chemical Company

Institute of Biosciences & Technology,

Texas A&M University

This monograph consists of manuscripts submitted by invited speakers who

College Station, Texas

Introduction: Strategies for the Protection of the Environment. . . . . . . . . . 1

Donald T. Sawyer and Arthur E. Martell

WARR: Implementing a Waste- and Release-Reduction Strategy. . . . . .. 19

Waste Minimization by Process Modification ......................... 25

J. R. Hopper, C. L. Yaws, T. C. Ho,

Kinetic Models to Predict and Control Minor Constituents in ........... 45

Kinetic Control of By-Product Reactions with In Situ Distillation. . . . . .. 65

Waste Minimization and Remediation in the Metal Finishing .......... 71

Clifford W. Walton, Kevin S. Briggs,

The Role of Catalysis in Industrial Waste Reduction ................... 89

Catalysis: The Atom Utilization Concept and Waste ................... 99

Novel Microporous Membrane-Based Separation Processes ............ 135

Ion Exchange for Glyphosate Recovery ......... " ................... 147

Lowell R. Smith and Jane L. Barclay

REMEDIATION OF HAZARDOUS WASTES:

Nucleophilic and Reductive Remediation Strategies for .............. 181

Application of Immobilized Cell Technology for Biotreatment . . . . . . . .. 247

Daniel E. Edwards and Michael A. Heitkamp

Membrane-Bound Microorganisms for Removing 261

Eric R. Allen and Yonghua Yang

Abstract of Paper. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 301

Index ............................................................. 305

Donald T. Sawyer and Arthur E. Martell

Although the contamination of our environment by industrial wastes has

Industrial Environmental Chemistry, Edited by D.T. Sawyer

Energy Conservation Manager

Waste reduction can be accomplished by creating less waste initially or by recycling

Industrial Environmental Chemistry, Edited by D.T. Sawyer

Improve Quality of Feeds

Although the percentage of undesirable impurities in a feed stream may be low, it

Use Off-spec Material

Improve Quality of Products

Inhibitors prevent unwanted side reactions or polymer formation. A wide variety of

Change Shipping Containers

Re-examine Need for each Raw Material

Distribute Feeds Better

Here is an area that deserves more attention

Improve Physical Mixing in the Reactor