Ralph Weiland, Optimized Gas Treating, Inc.

, USA, demonstrates
how cryogenic gas can be produced using only triethylene glycol
alone, with significant cost and operational advantages over
mol sieves, using a case study as an example.

n the LNG production process, before gas can be sufficiently But, if the water-lean solvent is insufficiently dry, no amount
compressed and liquefied, it must: have heavy hydrocarbons of adjusting its flow rate and temperature and adding extra
removed, its carbon dioxide (CO2) content reduced to at least trays or packed depth will change the situation. The regenerated
50 ppmv CO2 to prevent solid CO2 from forming, and its moisture solvent must not contain too much water, a parameter that is
content reduced to below 0.1 ppmv to prevent icing, before controlled by the glycol stripper.
entering the cryogenic liquefaction process itself. There, it is A conventional, reboiled TEG stripper is hopelessly
liquefied at near-atmospheric pressure by cooling to -162˚C. In inadequate for producing truly dry, lean TEG. Removing water
most cases, dehydration is achieved in a sequence of two or from wet glycol by contacting it with steam presents some
three columns of molecular sieve desiccant (mol sieves) while a serious challenges, even conceptually. First amongst them is the
fourth column is being regenerated. The main challenge using fact that one is trying to transfer water into a vapour phase that
glycol is the regeneration of triethylene glycol (TEG) into a is already water saturated. What is the driving force that is
super-dry state. responsible for drawing water from liquid to vapour? The
In any glycol-based process for removing moisture from a answer: really, there is none. Beyond the heat being passed into
gas, the main factors controlling performance are: the liquid boiling in the reboiler, which is causing some of the
� Glycol flow rate. water and a little of the glycol to evaporate, there is no driving
force for mass transfer (water vaporisation) in the column itself
� Dryness of the entering lean solvent. – the vapour is already water-saturated. Therefore, even without
� Temperature of the wet gas. running any simulations or much further thought, it should be
apparent that attaching a column to the reboiler is unlikely to
� Number of trays, or depth and size of packing. be of much benefit. In fact, it is barely dry enough to produce
gas sufficient to meet pipeline specifications (water content of (labelled ‘super drier’ in the image) is a very short, small-diameter
approximately 7 lb/million ft3 or 148 ppmv). The crucial realisation packed or trayed column positioned between the reboiler and
is that there is no way to continually evaporate water into an the accumulator holding the lean solvent. A small slipstream of
already water-saturated gas just by mass transfer alone, without dried gas is diverted from the top of the absorber and enters the
continually supplying energy for further boiling. For the stripper regeneration section either into the base of the super drier (Stahl
shown in Figure 1, this fact is unequivocally demonstrated by the Column), or directly into the reboiler via Stream 21, should an
simulation results shown in Figure 2. actual column not be installed.
Although invisible in Figure 2, there are actually nine TEG No matter where it is injected into the stripping section, the
profiles plotted there, for three different reboiler duties dry gas is a diluent for the stripped water vapour. As the gas and
(1200, 1500, and 2000 Btu/h) and three different tray counts water vapour travel together up the Stahl Column and stripper, the
(6, 8, and 16). All profiles are nearly coincident. Trays 1 – 3 are diluent alleviates the otherwise hopeless struggle the water
reflux trays whose purpose is to recover TEG from the vapour vapour has being drawn into the vapour. The diluent creates a
before it reaches the condenser and is lost with the stripped water. concentration driving force, where before there was none. The
Apart from Tray 4 (which serves as a feed tray), there is no obvious focus of the patent is producing a dried gas suitable for
reason for including Trays 5 – 16 in the column at all; in fact, Tray transportation by pipeline, i.e., with water content no greater than
4 could be eliminated, too, by feeding the hot rich TEG directly to approximately 7 lb/million ft3 (~148 ppmv) – the goal for
the reboiler and allowing the column to function as a three-tray liquefaction is moisture not exceeding 0.1 ppmv.
TEG recovery column. Only the reboiler itself is directly able to Figure 4 shows the TEG dehydration unit using a
transfer water to the vapour phase. The question then is, apart Stahl Column. The vapour feed (Stream 31) to the bottom of the
from boiling harder, what can be done to strip more water from Stahl Column is a small slipstream (0.033%) of the
the solvent? The answer to this enigma is a Stahl Column. well-dehydrated gas (47 ppb water by ProTreat® simulation)
originally destined for liquefaction. Based on what has been
The Stahl Column gleaned about the ineffectiveness of a reboiled stripper when used
The Stahl Column was described in US Patents3,105,748 to dry TEG, this stripper contains only four trays, with wet glycol
issued to Willi Stahl in 1963, and assigned to fed to Tray 4.
The Parkersburg Rig & Reel Company, Houston, Texas, the US. As Visually, the Stahl Column shown in Figure 3 has the
shown in Figure 3 (taken directly from the patent), a Stahl Column appearance of a very short column (or even no column at all, in
which case stripping gas would be fed directly to the reboiler via
Stream 21). The present implementation of a Stahl Column is
perhaps unusual in that it is fairly tall, using 15 m of Mellapak
M250.X structured packing. As will become apparent, the packing
size and bed depth are important parameters in setting system
performance. Similar parameters in the contactor are the packing
type, size (Mellapak-Plus M452.Y), and bed depth (10 m). The water
entering with the wet gas amounts to 827 ppmv. This could vary
quite a bit in actual operation so it would be important to know
something about what control measures could be taken to
accommodate variation in the water content of the wet gas, for
Figure 1. Gas dehydration using TEG with solvent example, or its flow rate. It would also be valuable to identify
regeneration in a single reboiled stripper.
those parameters to which expected performance is most
sensitive. Clues to answering such questions are to be found in
part in absorber and Stahl Column composition and temperature
profiles, and in part in a sensitivity study.

Parameters controlling performance

There are four main equipment items in a dehydration scheme
using a Stahl Column. Each one has a very different function, to
the extent that each deserves a separate discussion. However,
all the equipment items are interconnected, so a change in the
behaviour of one necessarily affects the performance of the others.

Contactor
The contactor is the only vessel where the dry gas is actually
produced by contacting wet inlet gas with the dehydration agent.
The performance of the contactor itself is directly affected by
the moisture content of the wet gas, its flowrate, and the dryness
of the TEG solvent. Performance is measured by the dryness
of the dehydrated gas; however, this metric also affects the
performance of the Stahl Column simply because dried product
Figure 2. How TEG concentration varies with tray number in a gas is responsible for producing the dry solvent in the first place,
reboiled stripper. and this plays back into the contactor’s performance. The Stahl
Column and the contactor are intimately connected through both

Reprinted from December 2022

the dry solvent and the product gas. The contactor cannot really be is not even remotely capable of reaching a dry gas moisture
analysed properly in isolation from the rest of the system. Figure content necessary for gas liquefaction in an LNG plant. Perhaps
5 shows that the dry gas moisture content falls exponentially with for this reason, mol sieves have become the technology of
distance up the column until the gas nears the top of the packing choice. Once the reasons are understood as to why conventional
where it asymptotically reaches a value of approximately 0.04 glycol-based technology fails to work in this application, replacing
ppmv. steam with inert stripping gas becomes an option. In some
LNG plants, however, a conventional TEG dehydration unit is
Reboiler and stripper installed immediately upstream of mol sieve beds to unload the
The reboiler, i.e., boiling, is the primary means for transferring mol sieve unit.
water from the solvent into the vapour. In fact, that is its sole The Stahl Column proposed here is not a short column – it
function – to evaporate water. As the TEG profile in Figure 2 shows, may need to be quite tall for adequate water removal from the
the stripping column does not perform any meaningful removal solvent. A relatively tall Stahl Column can easily reach a TEG
of water from the solvent. In fact, for the conditions of the case content of 99.98 wt% or more and produce dried gas with a
study, a four-tray stripper with TEG fed to Tray 4 produces a dry gas moisture level of well below 0.1 ppmv, quite suitable for
with 0.0483 ppmv water, while a six-tray stripper with TEG fed to liquefaction in an LNG plant. If an even drier gas is needed, there
Tray 6 produces dry gas with 0.0481 ppmv water. The difference is are a number of operating parameters that can be manipulated to
immeasurable. achieve this, including solvent flow rate, reboiler duty, and inert gas
However, the three wash trays allow only 1.1 lb/h of TEG to be flow rate. Operationally, TEG loss can be controlled by stripper
lost with the removed water (Stream 20), while five wash trays reflux flow and higher gas processing rates can be accommodated
permit only a 0.041 lb/h TEG leak, only approximately 4% of the by adjusting these same parameters. In the design phase, the
three-tray case. Calling this a ‘stripper’ is a misnomer. The stripper packed bed depths and reboiler energy capacity set the base or
is actually just a TEG capture column, and when correctly designed,
it is capable of almost completely preventing TEG losses from the
system, while stripping little or no water. The reboiler does bulk
water removal. For the base conditions of this case study, for
example, of the 1455 lb/h of water entering the solvent
regeneration section of the process, approximately 80% is directly
removed by the reboiler, and the rest is removed by the Stahl
Column. There is a small amount of water plus gas recovered
during stripping that is recycled back to the contactor, where the
contactor removes the recycled water.

Stahl Column
The Stahl Column removes the remainder of the water from the
solvent by using a small volumetric flow of dry stripping gas
borrowed from the final product. In fact, in the base case here, the
stripping gas amounts to less than 0.06% of the total product gas,
and 100% of it is recycled from the stripper overhead after being
Figure 3. Stahl Column (super drier) as shown in original
used in the Stahl Column – no hydrocarbons and almost no glycol
patent US 3,105,748.
are lost.
Figure 6 details how
water is stripped in the Stahl
Column, and how TEG is
concentrated there. In the
stripper, and particularly
across the reboiler, water
falls from 3.2 to 0.65 wt% (a
net decrease of 2.55 wt%). In
the Stahl Column, it falls
from 0.65 to 0.0001 wt% (net
decrease of 0.65 wt%). The
reboiler removes 80% of the
water entering the
regeneration section while
the Stahl Column removes
just 20%, but to an extremely
low level.

Technical
summary
A standard contactor and Figure 4. Gas dehydration using TEG and a Stahl stripping column.
reboiled regenerator scheme

Reprinted from December 2022

design capacity of the plant and how these items perform and quotes: a typical TEG plant (US$2.32 million), a TEG plant with
respond to changing process conditions is well mirrored and Stahl Column (US$2.75 million), and a traditional mol sieve
predicted by ProTreat mass transfer rate-based simulation. unit (US$10.89 million). Costing used the same feed as in the
A conventional system with a tall Stahl Column is well ProTreat simulations, and prices are quoted FOB West Texas
within the scope of normal plant operating procedures and Intermediate.
should present absolutely no unique challenges to plant A Stahl Column adds 18.5% to the cost of a conventional
operators. In fact, being fully continuous, this process is simpler TEG unit but leaves one with a more easily operable plant, and
to operate than a mol sieve system where, for example, at only 25% of the cost of the molecular sieve-based unit. The
operator intervention is required to switch and regenerate beds main equipment difference is that the reboiler and overhead
of solid adsorbent. Operationally, a conventional TEG unit with condenser are actually smaller in the Stahl unit. The
a Stahl Column is quite a manageable proposal. The immediate Stahl Column has a slightly larger diameter (725 mm) than the
question is one of economics. stripper (640 mm) but, of course, it is an additional column so it
makes the final price higher than the basic TEG unit. The mol
Economics sieve unit is much more expensive.
The economics of the Stahl Column process were compared
with a typical molecular sieve dehydration unit via an Possible pitfalls
independent cost estimate quote from Reset Energy LP, There are several potential pitfalls with this process. They
Midland, Texas, the US.2 Reset Energy provided three different include the potential for:
zz Co-absorption of oils, and other contaminants (likely to
be rejected into the stripped water).
zz Entrainment of glycol into the product gas (can be
mitigated using mist elimination).
zz Glycol decomposition in reboilers. Temperature never
exceeded 182˚C or 360˚F anywhere in the process, and
lower than usual temperatures seem characteristic
of using a Stahl Column. If 99.98 wt% TEG were to
be produced by boiling the solvent, the temperature
would far exceed 400˚F and decomposition would be
a truly serious issue. In this process, however, boiling
only occurs in the reboiler where the TEG concentration
is fairly low. Water removal in the Stahl Column takes
place via direct contact with an inert stripping gas at
modest temperatures – there is no boiling there so high
temperatures do not exist, and thermal decomposition is
not an issue.

If these pitfalls are recognised and engineered around,

Figure 5. Dry gas water content falls with distance up the there is quite significant upside CAPEX benefit to using TEG
contactor. Dry gas to regeneration is 250 kg/h. dehydration with a Stahl Column. Although not investigated
here, OPEX is likely to be lower because this is an
essentially lossless system so there are virtually no makeup
requirements and, being a completely continuous process,
manpower requirements will be minimal. The most
effective way to get a handle on the technical aspects of
the process is by mass transfer rate-based simulation,
specifically ProTreat, which performs detailed design by
calculating actual absorption and stripping rates, and which
allows reliable sensitivity analyses to be conducted.

References
1. For more information, see ‘Moisture measurement
in the production of Liquefied Natural Gas (LNG)’,
Michell Instruments, (2016), www.michell.com/downloads/
appnotes/Moisture_measurement_in_the_production_of_
Liquefied_Natural_Gas-LNG.pdf
2. Cost figures were independently provided by personal
communication from Chris Villegas, CEO, Reset Energy,
USA.
3. WEILAND, R.H., et al., ‘Stahl Columns – an Alternative to
Molecular Sieves?’, PTQ Gas, pp. 22 – 25, (June 2022).

Figure 6. Water removal in the Stahl Column.

Reprinted from December 2022