See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/250106928

Acid Mine Drainage Treatment with Armored Limestone in Open

Limestone Channels

Article  in  Journal of Environmental Quality · July 1997

DOI: 10.2134/jeq1997.00472425002600040013x

CITATIONS READS

164 2,268

5 authors, including:

Paul Ziemkiewicz Jeff Skousen

West Virginia University West Virginia University
125 PUBLICATIONS   2,743 CITATIONS    211 PUBLICATIONS   5,711 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Land Reclamation View project

Acid Mine Drainage View project

All content following this page was uploaded by Jeff Skousen on 12 May 2015.

The user has requested enhancement of the downloaded file.

Acid Mine Drainage Treatment with Armored Limestone in Open Limestone Channels
P. F. Ziemkiewicz, J. G. Skousen,* D. L. Brant, P. L. Sterner, and R. J. Lovett
ABSTRACT
Much attention has been devoted to developing inexpensive, lime-
stone-based systems for treating acid mine drainage (AMD) with little
A ID MINE DRAINAGE is one of the largest environmen-
tal problems facing the mining industry. Acid mine
drainage originates from active and AML when pyrite
or no maintenance. Treatment of AMD with limestone results in a (FeS2) or other metal sulfides associated with mineral
surface coating of metal hydroxides, a process known as limestone deposits are subjected to oxidizing conditions. Upon
armoring. Once armored, limestone is assumed to cease dissolution exposure to oxygen and water, the sulfide minerals pro-
and acid neutralization. Laboratory and field experiments determined
acidity changes in AMD when contacted by armored and unarmored
gress through a combination of oxidation and microbial
limestone and investigated the implications of armoring on the con- catalyzing reactions to produce large amounts of dis-
struction of open limestone channels for treating AMD. Results of a solved metals, sulfate, and acidity (Singer and Stumm,
laboratory titration study indicated armored limestone was only 2 1970). This acid also dissolves other minerals, releasing
to 45% less effective in neutralizing a hydrochloric acid solution as cations such as iron (Fe), manganese (Mn), and alumi-
unarmored limestone. A laboratory container study showed that ar- num (Al). The resulting AMD is generally characterized
mored limestone was 90% as effective in neutralizing AMD as unar- by low pH (<3.5), high acidity (>500 mg/L as CaCO3;
mored limestone. The field study surveyed 2- to 8-yr-old, rock-lined all acidity values hereafter are in CaCO3 equivalents),
channels constructed of sandstone or limestone, and measured water and high concentrations of total dissolved metals (>50
quality changes down the length of the channel. Open limestone mg/L). Upon reaching a stream, AMD alters the
channels, though armored, reduced more acidity in AMD (4-62%)
stream's chemical balance by consuming alkalinity and
than the sandstone channel (2%). The results from open limestone
channels were compared to an acid neutralization kinetics model that
introducing metal ions, resulting in a degradation of
predicts the rate of acid neutralization for a specified channel size, and biological productivity. If sufficiently severe, AMD will
AMD flow and acidity concentration. The open limestone channels in also render the receiving water unfit for human, agricul-
the field neutralized more acidity than the model predicted. Open tural, industrial, or recreational uses (Atlas and Bar-
limestone channels show promise for neutralizing AMD in watershed tha, 1987).
restoration projects and abandoned mine land (AML) reclamation In order to reduce the concentrations of dissolved
projects where one-time installation costs are incurred, little to no metals and acidity, AMD is treated with alkaline chemi-
maintenance is required, and systems do not have to meet specific cals by most mining operations and the precipitated
water quality standards. metals are collected in settling ponds. Chemical treat-
ment is expensive, however, and may be necessary long
P.P. Ziemkiewicz, D.L. Brant, and R.J. Lovett are Director, Research
after mining has ceased. One alternative to chemical
Associate, and Program Manager, National Mine Land Reclamation treatment is passive treatment, which refers to any zero
Center, West Virginia University; J.G. Skousen and P.L. Sterner are to low maintenance AMD treatment method that does
Associate Professor/Extension Reclamation Specialist and Research not require continual chemical addition and monitoring.
Assistant, Division of Plant and Soil Sciences, West Virginia Univer- Passive systems are of increasing interest as state, indus-
sity, Morgantown, WV 26506-6108. Scientific contribution no. 2592
from the West Virginia Agriculture and Forestry Experiment Station,
try, and federal partnerships are formed to rehabilitate
Morgantown. This research was supported by the U.S. Bureau of watersheds damaged by historic mining (Skousen et al.,
Mines, National Mine Land Reclamation Center under Contract no.
CO388026 and by funds appropriated under the Hatch Act. Received
10 June 1996. *Corresponding author (jskousen@wvu.edu). Abbreviations: ALD, anoxic limestone drain; AMD, acid mine drain-
age; AML, abandoned mine land; PADER, Pennsylvania Department
Published in J. Environ. Qual. 26:1017-1024 (1997). of Environmental Resources.
1018 J. ENVIRON.QUAL., VOL. 26, JULY-AUGUST
1997

1997; Titchenell and Skousen, 1996). Passive systems laboratory container studies. A field study evaluated
offer low maintenance and inexpensive solutions to existing open limestone channels and a sandstone chan-
AMDremediation (Brodie, 1990; Hedin, 1989). Anoxic nel to examine AMDtreatment by armored limestone.
limestone drains (ALDs), wetlands, or a combination The field study acid neutralization results were com-
of both are the most often used passive systems (Faulk- pared to acid neutralization predicted on these sites by
ner and Skousen, 1994). Wetlands are effective in han- the Ziemkiewicz et al. (1994) model.
dling low-acid loadings, but often encounter difficulties
or even fail under high-acid loading (Kleinmannet al., MATERIALS AND METHODS
1991; Wieder, 1989). Problems with ALDsoccur when Laboratory Titration Study
Fe(III) and A1 are present in the water. These cations
precipitate as hydroxides (e.g., Fe(OH)3(s)) and coat Twosamples of limestone were taken from each of two
limestone surfaces (armoring) and can plug the lime- field sites. The field sites were open limestone channelsat
stone void space, thereby reducing limestone dissolution Robinson Run (RR) near Maidsville, WV,and near Dola,
WV.One limestone sample from each site was armored by
and acid neutralization. AMD (consistent metal hydroxide coating) and one sample
Studies by Pearson and McDonnell (1974, 1975a,b) was unarmored.The unarmoredsamples were along the upper
found that armored limestone was only 20%as effective bank of the channel and had not been contacted by flowing
in neutralizing AMD as unarmored limestone. Ziemkie- AMD. The four pieces of limestone were measuredfor weight,
wicz et al. (1994) conducted a preliminary study of open volume, and lengths to determine a shape and elongation
limestone channels on AMLsites that had been re- factor after the methodspresented in Pearson and McDonnell
claimed in West Virginia. Voids in limestone channels (1977). Thesevalues were used to determine surface area
contained metal hydroxides where slopes were <10% each piece of limestone (Table 1). The samples were placed
in six solutions (40 mL)acidified to a knownpHby HCI:
and AMDsimply flowed over the limestone with little 5.0 = 2 mg/Lacidity (these were measuredacidity values),
water contact and treatment. Limestone in open lime- pH4.5 = 4 mg/Lacidity, pH4.0 = 8.5 mg/Lacidity, pH3.5 =
stone channels was also armored, presumably reducing 22 mg/Lacidity, pH3.0 = 70 mg/La~idity, and pH2.5 = 208
the limestone dissolution rate and acid neutralization to mg/Lacidity. As limestone neutralized acid in the solution,
20% of unarmored limestone (Pearson and McDonnell, pHwas kept static by the addition of HC1by a Titralab VIT-
1975a). Ziemkiewicz et al. (1994) concluded that lime- 90 automated titrator (Radiometer-America,Westlake, OH)
stone channels could effectively neutralize AMDif for a period of 20 min. Since limestone dissolution rate is
channels were constructed on steep slopes to reduce constant if pH remains constant, a period of 20 rain was
plugging of limestone pores by metal hydroxides, and deemedadequate for determining limestone dissolution rate.
This time period also minimizedthe amountof surface area
if channels were built five times bigger to account for changeand dissolution of armoringon limestone pieces since
the armoring effect. A model was developed by Ziem- the same pieces were used throughout the experiment. The
kiewicz et al. (1994) to estimate limestone volumes and amountof HC1addedwas determinedin 5-rain intervals (the
channel dimensions for achieving neutralization of HC1added during each 5-min period was very similar) and
AMD based on the following first-order kinetics: summedto give the total amountadded during the 20-min
experiment. At the end of :20 min, each limestone samplewas
In C~F = -kt [1] removedfrom the solution, washedwith distilled deionized
Co water, and ovendried at 90°Cfor 1 h. After drying, the lime-
stone sample was placed in a desiccator until roomtempera-
where ture (20°C)was reached, then the samplewas used in the next
CF = final acidity (mg/L); lower pH solution experiment. Iron hydroxide armoring was
Co --- original acidity (rag/L); reduced very gradually by placing the armoredlimestone in
k = rate constant (l/h); increasingly acidic solutions during the experiments and a
t = time of reaction (h). Table 1. Dimensions of limestone rocks used in the laboratory
titration study. Fromthese dimensions, shape, and elongation
Based on data presented in Pearson and McDonnell factors were calculated from equations in Pearson and McDon-
(1975a), k was solved and equals -2.303/h, which repre- nell (1977) and surface area was calculated.
sents the rate constant of acid neutralization by lime- Robinson Run Dola
stone. To account for armoring, the rate constant was
multiplied by five (the 20%dissolution rate of armored Characteristic Unarmored Armored Unarmored Armored
limestone compared to unarmored limestone) to ac- Weight, g 8.24 7.81 7.97 7.36
~
Volume, cm 3.01 2.84 2.97 2.72
count for the increased contact time needed between ~
Unit wt., g/cm 2.74 2.75 2.68 2.71
armored limestone and AMDfor treatment. Therefore, Max. length, cm 2.80 3.31 2.83 2.70
the residence time of AMD to achieve a target acidity Min. length, cm 1.02 0.83 1.18 1.01
Shapefactor~" 0.72 0.63 0.74 0.72
level (CF) from its initial acidity concentration (Co) Elongationfactor~ 1.73 2.08 1.63 1.71
could be solved. The amount of limestone required for % Armored§(begin) 0 100.00 0, 100.00
% Armored (end) 0 90.00 0 70.00
treatment was based on the flow rate and the measured 2Surfacearea~[, cm 14.07 15.41 13.48 13.08
acidity (in mg/Las CaCO3)of the water. Wetted channel
Shapefactor wascalculated by Eq. [5] in PearsonandMcDonnell (1977).
dimensions were computed to obtain sufficient resi- Elongation factor was calculated by Eq. [4] in Pearson and McDon-
dence time of the water in the limestone channel at nell (1977).
specified water velocities and channel slopes. % Armored wasa visual estimate of the amountof snfface area covered
by iron hydroxidecoating at the beginningof the experimentandafter
This study determined acid neutralization by armored all titrations hadbeen completed.
and unarmored limestone in laboratory titration and Calculated by Eq. [1] in Pearson and McDonnell(1977).
 ZIEMKIEWICZ ET AL.: AMDTREATMENTIN OPEN LIMESTONECHANNELS 1019

Table 2. Solution/limestone combinations used in the laboratory Table 3. Total volume (mL) of acid used (0.01 MHCI) to maintain
container study. Each combination marked had three repli- an initial pH and calculated reaction rates (l~mol/cm2 per s) of
cations. Robinson Run and Dola unarmored and armored limestone.
Robinson Dola Unarmored Armored Reactionrate
Run armored armored Unarmored pH Reaction Reaction decrease due
Water limestone limestone limestone Solution Vol. rate Vol. rate to armoring
Maidsville X X 2mL I~mol/cm 2pmol/cm
mL %~
Shaw mines X X per s per s
Synthetic AMD’~ X X
Deionized water X X X Robinson Run
5.0 0.57 0.94 0.48 0.74 21.3
~" AMD
= acid mine drainage. 4.5 1.15 2.02 0.94 1.60 20.8
4.0 2.66 5.12 1.81 3.12 39.1
3.5 %60 14.83 4.81 8.62 41.9
visual estimate of armoring at the beginning and ending of 3.0 1%55 34.12 11.10 19.47 42.9
the experiments is given in Table 1. Each limestone sample and 2.5 34.65 67.43 18.20 36.95 45.2
pH solution combination was replicated twice. The amount of Dola
HCI added to maintain static pH was used to calculate a 5.0 0.81 1.29 0.99 1.56 - 17.3
reaction rate based on first-order kinetics. This data was used 4.5 1.39 2.53 1.29 2.47 2.4
4.0 2.86 5.79 2.74 5.50 5.0
to compare the two limestones (RR and Dola) and the effect 3.5 8.10 16.89 6.43 13.05 22.7
of armoring on acid neutralization. 3.0 22.55 44.52 16.40 36.69 17.6
2.5 40.70 84.47 28.60 58.01 31.3
Laboratory Container Study Calculated by 1 - (the armoredreaction rate/the unarmoredreaction
rate).
This laboratory study was conducted using 2-L, high-density
polyethylene containers filled with 2.3 kg of 5- to 10-cm sized, concentrated nitric acid (2.5% v/v) prior to analysis. The pa-
armored limestone or unarmored limestone. The two sources rameters tested were: pH (electrode), electrical conductivity
of armored limestone were from the same sites as those used (conductivity bridge), alkalinity, and acidity expressed
in the titration study. The unarmored limestone was from the CaCO3(Brinkman auto-titrator), and concentrations of total
Deer Valley formation in Somerset Co., PA. One of four Fe, At, Mn, Ca, and Mg (Leaman Labs inductively coupled
solutions was added to each of the containers (1.2 L). The plasma spectrometry), and sulfate (BaCt2 precipitation and
four solutions were: AMDfrom the Maidsville Seep near measured on a Milton Roy Spectronic 20) (APHA, 1989).
Morgantown, WV (2080 mg/L acidity), AMD from Shaw water quality results were plotted to show acid neutralization
Mines Run near Meyersdale, PA (518 mg/L acidity), a pH 3.0 with time and to provide a constrast between armored and
HC1 solution (171 mg/L acidity), and deionized water (con- unarmored limestone.
trol). Only 9 of the 12 possible limestone/solution combina-
tions were used in this study (Table 2), and each of the selected
Field Study
combinations had three replications.
Two water samples of 40 mL each were collected with a In the field study, we selected eight existing rock-lined chan-
plastic syringe from each container (one sample for general nels on reclaimed AMLsites treating AMD. These channels
water chemistry and one for metal analysis) at the following were constructed for erosion control or to convey surface
time intervals after water introduction: 0, 1, 2, 4, 6, 12, 18, and water off the site and were not designed to treat AMD. One
24 h. The water samples were filtered (0.45 p.m) and metal channel was constructed of sandstone and seven were made
analysis samples were acidified to pH < 2.0 with 1 mL of of limestone. Two water samples of 250 mL each were col-

Table 4. Changesin acidity loss (mg/L) and the rate of acid loss (mg/L per min) resulting from adding three solutions to Robinson
or Dola armored limestone and unarmored limestone.
Armoredlimestone Unarmoredlimestone
Time Initial acidity Acid loss Initial acidity Acidloss
h mg/L mg/L mg/L per rain mg/L mg/L mg/L per min
Maidsville-Robinson Runarmoredlimestone
0 2080 -~" - 2080 - -
1 1740 376 6.27 1547 533 8.88
2 1692 12 0.20 1337 210 3.50
4 1513 179 1.49 1147 190 1.58
6 1374 139 1.16 1062 85 0.71
12 1210 164 0.46 998 64 0.18
18 1041 169 0.47 874 124 0.34
24 925 116 0.32 753 121 0.34
ShawMines-Dola armoredlimestone
0 518 - - 518 - -
1 174 344 5.73 105 413 6.88
2 106 68 1.13 73 32 0.53
4 60 46 0.38 1 72 0.06
6 4 56 0.47 0 1 0.01
12 0 4 0.01 0 0 -
18 0 0 - 0 0 -
Synthetic AMD’~-Dolaarmoredlimestone
0 171 - 171 - -
1 24 147 2.45 0 171 2.85
2 0 24 0.40 0 0 -
Determinationnot performed.
AMD = acid mine drainage.
1020 j. ENVIRON.
QUAL.,VOL.26, JULY-AUGUST
1997

MaidsviileSeep Table 5. Alkalinity generation with RobinsooRuoand Dola ar-

moredlimestones and unarmored limestone over a 24-h period
whenplaced in deionized water.
,--. 2500 _~ Unarmored RobinsonRun Dola
Time limestone armoredlimestone armoredlimestone
1500 h mg/L
,000t 0
1
2
4
41
29
4
18
19
4
16
20
4 30 28 26
0 5 10 15 20 25 6 32 28 36
Time(~s.) 12 39 34 34
18 47 32 45
~ A~ored
Acidi~ ~ Un~ored
Acidi~ 24 48 30 38
~ A~ored Al~lini~ ~ U~or~ A ~ n ~

Fig. 1. Acidityreduction
andalk~inltygeneration
of ~aidsvilleSeep West channel) both on 20% slopes. The flow of AMDin the
acid mine drainage (A~D) ~tb armored and anarmored West channel was 113 L/min (30 gpm) and the acidity was
limestone. 410 mg/L. The total combined flow in the East channel was
76 L/min (20 gpm) and the acidities were 355 mg/L for the
lected (one for general water chemistry and one for metals first branch and 335 mg/L in the second. The flow rates were
analysis) at identified distances along the channels and ana- equal at the sources and the mouths of each channel.
lyzed as described above at the same laboratory (APHA, The NRCSEichleberger no. 2 open limestone channel is
1989). One sample was filtered in the field (0.45 ixm) located 6.5 km southeast of Coaldale, PA. The channel is 49 m
acidified to pH< 2.0 with 1 mLof nitric acid (0.4% v/v) for long, 2 m wide, and 0.1 m deep (160 by 6 by 0.5 ft) on a 20%
metals analysis. The other sample was not filtered and cooled slope, and the limestone is heavily armored. The flow through
to 4°C to determine pH, acidity, and alkalinity. Flows were the channel was consistent at 378 L/min (100 gpm) and the
measured with a Marsh-McBirney model 2000 Flo-Mate elec- acidity at the source was 510 mg/L. The Pennsylvania Depart-
tromagnetic flow meter for larger flows (>95 L/min) or a cali- ment of Environmental Resources (PADER)open limestone
brated bucket and a stopwatch for smaller flows (<95 L/min). channel is located 1.6 km west of Defiance, PA. This channel
Distances were measured with a 33-m surveying rope. is 11 m long, 1 m wide, and 0.1 m deep (37 by 3 by 0.5 ft)
The results of water quality analyses from field channels a 60% slope with a flow of 95 L/min (25 gpm). The acidity
were plotted against the kinetics model designed to predict was 2600 mg/L at the source and the limestone is heavily
the dimensions required to treat AMDwith open limestone armored. The Pennsylvania Game Commission open lime-
channels (Ziemkiewicz et al., 1994). The model estimates acid stone channel is 11 m long, 1 m wide, and 0.1 m deep (35 by
neutralization for an armored open limestone channel with 3 by 0.5 ft) on a slope of 45%. It is located northeast of
specified dimensions and the AMD’sflow and acidity concen- Vintondale, PA. The flow is 484 L/min (128 gpm) and the
tration. acidity is 330 mg/L at the source. This channel is limestone
The Natural Resources Conservation Service’s (NRCS) and heavily armored.
Coral/Graceton sandstone channel is located adjacent to U.S. The Cottage Townopen limestone channel is located 1.6
Route 119 immediately northeast of the towns of Coral and km west of Cairnbrook, PA. The channel is 137 m long, 1.3 m
Graceton, PA. The channel is 220 m long, 3 m wide and 0.1 m wide, and 0.1 m deep (450 by 4 by 0.5 ft) on a 9%slope with
deep (720 by 9 by 0.5 ft) on a 10% slope. The flow of AMD a flow of 302 L/min (80 gpm) throughout the entire length.
through the channel was 1323 L/min (350 gpm)and the acidity The limestone was armored and the AMDhad an acidity of
was 550 mg/L. 32 mg/L at the source. The Opawskyopen limestone channel
The Morgantown Airport open limestone channels are lo- is located i kmsouth of Mosgrove,PA. This channel is differ-
cated adjacent to U.S. Routes 119/857 east of Morgantown, ent from other open limestone channels because it had a
WV,and consist of two armored limestone channels. The first wetland installed about a third of the way downthe length of
channel (West) is 46 m long, 1.3 m wide, and 0.1 m deep (150 the channel that interrupted the flow. The top portion of the
by 4 by 0.5 ft) on a 14%slope. The East channel is branched, channel (46 m long, 2 m wide, and 0.3 m deep) is limestone
with the first branch being 21 m long (70 ft) and the second and heavily armored on a 9% slope. The water after flowing
branch being 27 m (90 ft) long (same widths and depth as through 46 m of limestone channel enters the wetland that

Table 6. Characteristics and performance(acid loss) of a sandstoneopen channel and seven open limestone channels at field sites
Pennsylvania and West Virginia.
Acidity
Rock Contact
Channel Flow Length type’~ Slope time:~ Initial Final Acidloss
L/min m % min -- mg/L -- % mg/L
Coral/Graceton 1323 220 SS 10 37 550 540 2 10
MorgAirport W 113 46 LS 14 21 410 360 12 50
MorgAirport E 76 27 LS 20 6 355 330 7 25
Eichleberger
no. 2 378 49 LS 20 29 510 325 36 185
PADER§ 95 11 LS 60 4 2600 2500 4 100
PA GameCom. 484 11 LS 45 1 330 125 62 205
Cottage Town 302 137 LS 9 19 32 28 13 4
Opawsky 907 46 LS 9 86 30 15 50 15
SS = sandstone,LS= limestone.
Contacttimewasbasedon waterflows, cross-sectionalareaof limestone(correctedfor voidspace)andlengthof channel(Ziemkiewicz
et al., 1994).
PADER = PennsylvaniaDepartmentof Environmental Resources.
 ZIEMKIEWICZ ET AL.: AMDTREATMENTIN OPEN LIMESTONECHANNELS 1021

Table 7. Rate of acid loss in a sandstone open channel and seven open limestone channels at field sites in Pennsylvania and West
Virginia. The rate of acid loss (%/rain and mg/L per rain) in field channels is comparedto the predicted rate of acid loss based
the kinetics model.
Acidity
Channel Initial Final Acid loss Actualrate of acid loss Predictedrate of acid loss
-- mg/L -- % mg/L %/rain mg/L per rain %/rain mg/Lper min
Coral/Graceton 550 540 2 10 0.05 0.27 0.51 2.83
Morg Airport W 410 360 12 50 0.57 2.38 0.19 0.76
MorgAirport E 355 330 7 25 1.17 4.17 0.16 0.66
Eichlebergerno. 2 510 325 36 185 1.24 6.38 0.55 2.86
PADER’~ 2600 2500 4 100 1.00 25.00 0.08 1.75
PA Game Com. 330 125 62 205 62.00 205.00 1.00 3.00
Cottage Town 32 28 13 4 0.68 0.21 0.66 0.21
Opawsky 30 15 50 15 0.58 0.17 0.70 0.21
? PADER= Pennsylvania Departmentof Environmental Resources.

covers an area of 350 m2. Uponexiting the wetland, the water tion. Based on visual estimates, approximately 10%re-
continues downthe armored channel measuring 137 m long, moval of armoring occurred from the RR limestone
2 m wide, and 0.3 mdeep (450 by 6 by 2 ft). The flow surface area during the experiment, while about 30%
AMD through the entire system was 907 L/min (240 gpm) of the armoring on Dola limestone was removed by the
and the acidity at the source was 30 mg/L. end of the experiment. The greater unarmored surface
area on the Dola armored limestone piece may account
RESULTS AND DISCUSSION for the 30%decrease of limestone dissolution rate at
Laboratory Titration Study pH 2.5 compared to 45% decrease for RR (Table 3).

Robinson Run (RR) armored limestone reaction rate Laboratory Container Study
was slower than unarmored limestone, and therefore
less acid was needed to keep pH static throughout each The initial acidity of the Maidsville seep (2080 mg/
experiment (Table 3). At a pH of 5.0, the unarmored L) was reduced to 925 mg/L (56% reduction) after 24
limestone dissolution rate was 21% greater than the with RR armored limestone (Table 4 and Fig. 1). The
armored limestone rate. Between pH 3.0 and 4.0, the RRunarmored limestone neutralized 65 %of the acidity
rate of limestone dissolution with armored limestone (753 mg/L) after 24 h (a difference of only 9%between
decreased to about 40% compared to unarmored lime- armored and unarmored limestone). The armored lime-
stone. At pH 2.5, there was a 45% decrease in acid stone from Dola completely eliminated the ShawMines’
neutralization due to armoring. initial acidity of 518 mg/Lin the containers after 6 h.
At a pH of 5.0, Dola armored limestone showed more Dola unarmored limestone achieved 100% treatment
acid neutralization than unarmored limestone, which after 4 h.
was the only limestone/solution combination that The unarmored and Dola armored limestones were
showed armored limestone to dissolve faster than unar- treated with a hydrochloric acid solution (0.02 MH2SO4,
mored limestone (Table 3). Weassumed that armoring pH 3.0). The acidity was completely neutralized within
could only reduce limestone dissolution. Both replicates 2 h with Dola armored limestone and within 1 h for
gave the sameresult, so it is not clear whythis happened. unarmored limestone (Table 4). Alkalinity generation
At pH 4.0 and 4.5, unarmored limestone from Dola was leveled off at 67 mg/L after 18 h for Dola armored
only 2 to 5%better at neutralizing acidity than armored limestone and 85 mg/L after 4 h for unarmored lime-
limestone. However with lower pH solutions, armored stone (data not shown).
limestone showed 17 to 31%reductions in the rate of Deionized water was added to unarmored, Dola ar-
acid neutralization vs. unarmored limestone. mored and RR armored limestone to determine alkalin-
These titration results show armored limestone to be
2 to 45% less effective in neutralizing hydrogen ion NRCS- Coral/Graceton
acidity compared to unarmored limestone. As solution
pH decreased, the difference in acid neutralization be-
came greater between armored and unarmored lime-
stone. The data also indicate differences between the 520
two limestone sources. The unarmored limestone at 500
Dola neutralized more acid than unarmored RR lime- ’~480
460
stone (Table 3). The RR armored limestone initially
~440
appeared to be more heavily armored than Dola ar- 420
mored limestone and the data show that the RR ar- 400 ~ 3’0 61 85 1~6 162219
mored limestone dissolution rate was inhibited more
than Dola armored limestone. Reintroducing the same Distance From Top Of Drain (m)
armored limestone piece into increasingly acidic solu- ~_
-’~1"- OBSERVED~ PREDICTED

tions certainly caused armoring to gradually dissolve, Fig. 2. Observedand predicted acidity reductions from a sandstone
thereby increasing the effective surface area for reac- channel at the Coral/Gracetonsite.
1022 J. ENVIRON.QUAL., VOL. 26, JULY-AUGUST
1997

MorgantownAirport West NRCS- Eichleberger #2

410

2500
390 ~ 450

~’380 ~ 400
370 ~350
360 300 --~
0 9 18 27 36 46 0 ~ 118 27 36 49
Distance From Top Of Drain (m) Distance From Top Of Drain (rn)
[’-~1- OBSERVED~ PREDICTED]

Fig. 3. Observedand predicted acidity reductions of an open lime- Fig. 5. Observed

and predictedacidity reductionsof an openlime-
stone channel at MorgantownAirport West site. stone channel at the Natural ResourcesConservationService
(NRCS)Eichlebergerno. 2 site.
ity generation with time amongthese three limestone
samples in the absence of an acid solution. Data indicate ne] performancewas better than predicted (Fig. 3).
that armored limestone required between 4 and 6 h to The actual acidity reduction was 0.57% per rain or
achieve maximumalkalinity, whereas unarmored lime- 2.38 mg/£acidity decreaseper rain (Table 7) compared
stone achieved maximum alkalinity during the first hour to th~ pr¢dicted reduction of 0.]9% per rain from the
(Table 5). Unarmoredlimestone produced about 45 mg/ model. The ~or~antown ~irport East channels also
L alkalinity, RR armored limestone gave 30 mg/L, and performedbetter than Wedictedwith an ~cidity reduc-
Dola armored limestone produced 38 mg/L. Both labo- tion of L ~7 %per rain comparedto a predicted reduction
ratory studies indicate that limestone dissolution was of 0.~6%per rain (Fi~. 4).
reduced between 2 and 50% due to armoring, much Thc NRCSEichlcbcr~er no. 2 open limestone chan-
smaller than the 80%reduction values reported by Pear- nel performedbetter than expectedwith a~ actual acid-
son and McDonnell (1975a). hy v¢ducfiono~ L24%per rain (or 6.38 m~/Eacid reduc-
tion p¢r rain) comparedto a predJctcd reduction
Field Study 055%per m~n (Fi~. 5). Tb¢ P~DERopen limestone
channclis a very short channelwith hJ~ha.cidhy (Tab]¢
Flows, channel dimensions, slopes, contact time, and
6), but r¢moves~%of the acidity per rain or 25 m8/£
initial and final acidity concentrations for the field chan-
acidity per mJn(Table 7). This wasan order of magni-
nels are summarized in Table 6. The NRCSCoral/
tude better than th~ predJct¢dacJdhyr¢ducfion of about
Graceton sandstone channel served as a control to lime- 0A%per m~nwhh such a short channel (FJ~. 6). The
stone channels. The resulting acidity reduction in this
P~ GameCommissionopen limestone channel is a very
sandstone channel was 2% (Table 6) or 0.05% per min
short channel(TaMe6), but ~t showsan JmprcssJveper-
(Table 7). This is much less than the predicted 0.51%
formance whh
per min if it had been constructed with limestone (Fig.
m~/Eacid rcmovedp¢r rain) comparedto a predictcd
2). The prediction line in Fig. 2 was based on the ex-
pcfformance of ~%per m~nacidity ~emova] (Tab]¢
pected acid neutralization (with Pearson and McDon-
and FJ~. 7). Thesteep ~radesof these last two chann¢]s
nell’s data) by an open limestone channel from the
increascd water velocities and splashing, which must
modelof Ziemkiewiczet al. (1994) using the initial acid-
have enhanced]imcstone dissolution and acidity re-
ity, water flow, length of channel, contact time, and the
ductions.
predicted limestone dissolution rate when armored.
The Cotta~c Town open l~mestone channel has a
The Morgantown Airport West open limestone chan-
PADER
MorgantownAirport East

2600

~- 2580 .......
~, 355 ~2560
~ 350 -- ---- ~
~2540

~340 < 2520 -

< 335
2500 ,
330 , 0 11
0 15 18 21 Distance FromTop Of Drain (m)
Distance From Top Of Drain (m)
[ -IB- OBSERVED
-~- PREDICTED]
1 -~-OBSERVED
-~-PREDICTED
1 Fig. 6. Observed and predicted acidity reductions of an open lime-
Fig. 4. Observedand predicted acidity reductions of an open lime- stone channel at the Pennsylvania Department of Environment
stone channel at MorgantownAirport East site. Resources (PADER)site.
 ZIEMKIEWICZ ET AL.: AMDTREATMENTIN OPEN LIMESTONECHANNELS 1023

NRCS- PA GameCommission NRCS- Opawsky

350 35

~, 300
~250

~200

< 150 <15

100 ~ ~ 10 ~---
0 35 0 22 46 68 91 121 152 182 213 234
Distance From Top Of Drain (m) Distance From Top Of Drain (m)
[
I -~"OBSERVED -~-PREDICTED [-I~- OBSERVED -~- PREDICTED ]

Fig. 7. Observed and predicted acidity reductions of an open lime- Fig. 9. Observedandprediced acidity reductions ofanopen limestone
stone channel at the Natural Resources Conservation Service channel at the Natural Resources Conservation Service (NRCS)
(NRCS) PA Game Commission site. Opawskysite.

small amountof acidity (32 mg/L) entering the channel lutions. Acidity removal by limestone (armored or unar-
(Fig. 8) and exhibits an acidity removalbetter than that mored) is proportional to its contact time with AMD,
predicted over the first 110 m of the channel (1.83% but characteristics of the channel like slope and water
per min compared to 0.47% per min predicted). Acidity velocity can increase acidity reductions. Openlimestone
concentration increases over the last 27 mof the channel channels work best on steep slopes. The key factor in
because a small source of AMD enters the base of the designing open limestone channels is preventing iron
channel. This brings the overall acidity removal closer and aluminumprecipitates from settling in voids in the
to the predicted value (0.68% compared to 0.66% per limestone channel. One limestone channel not reported
min). The NRCSOpawsky open limestone channel’s here was found on a nearly flat slope (1-3%). It was
performance was slightly worse than predicted (0.58% filled with metal hydroxides and sediment, and it was
acidity removal per min or 0.17 mg/L acid removal per ineffective in neutralizing acidity. The most successful
min compared to a predicted removal of 0.70% per min) channels had slopes above 40% and used coarse lime-
but the channel still removed 50%of the acidity (Fig. stone (15- to 30-cm sized material). Both slope of chan-
9 and Table 7). nel and size of limestone can minimize the settling of
In the field study, acidity reductions in AMD after precipitates in suspension, and thereby reduce plugging
flowing through open limestone channels varied from of the limestone channel. Evidence of the effect of slope
4 to 62%, and acid reductions per min of contact time on armored limestone dissolution is seen on the
with armored limestone were between 0.5 and 62% per PADERand PA Game Commission open limestone
min or 0.2 to 205 mg/L per min. The steeper channels channels that were constructed on slopes >40%. The
(PADERand PA Game Commission) performed better age of the channels we studied varied from 2 to 8 yr
(25 and 205 mg/L acid removedper rain, respectively) and none of these channels had required maintenance.
than the two channels on flatter (9%) slopes (Cottage If constructed correctly, open limestone channels should
Town and Opawsky with 0.21 and 0.17 mg/L per min, be nearly maintenance free and less expensive to con-
respectively). In the sandstone channel, acidity de- struct than other AMDtreatment systems.
creased by only 2% and by a factor of 0.05% per rain Each of the passive treatment systems (aerobic wet-
of contact time with sandstone. lands, anaerobic wetlands, ALDs,successive alkalinity
Limestone dissolution increases with more acidic so- producing systems, and open limestone channels) have
an area of application (see Faulkner and Skousen, 1995).
NRCS. Cottage Town It is difficult to achieve water quality standards (as re-
quired by NPDESpermits on active mining operations)
34 by passive AMDtreatment with any one method alone.
However, coupling these systems could allow some acid-
ity reduction and metal precipitation in one system,
then introducing the partially treated AMD into another
~8 system for additional acidity and metal removal. The
primary application of most passive treatment systems
~ 24 is on watershed restoration and AMLreclamation proj-
ects, and perhaps for pretreatment of AMDbefore
22
0 27 55 82 110 137 chemical treatment. Open limestone channels are par-
Distance From Top Of Drain (m) ticularly useful in steep terrain where long (300-1000
[-~- OBSERVED -~- PREDICTED m) channels are possible, and they offer a unique AMD
1
Fig.8. Observed
andpredicted
acidity
reductions
of an openlime- treatment opportunity where no other passive system
stonecbanne|at t~e MatBralResources
Conservation Settee is appropriate. Open limestone channels will produce
(NRCS)CottageTo~ site. metal hydroxide precipitates like any effective AMD
 1024 J. ENVIRON. QUAL., VOL. 26, JULY-AUGUST 1997

treatment system, so settlement basins should be incor-

porated into the design. Larger open limestone channels
should have settling ponds or wetlands placed at inter-
mediate points (flat channel segments) to remove the
precipitates and help prevent plugging in limestone
channels.

SUMMARY AND CONCLUSIONS

Treatment of AMD by armored limestone has been
assumed to be minimal. Results of two laboratory stud-
ies indicate that armored limestone was only 2 to 45%
less effective in neutralizing acid than unarmored lime-
stone depending on the pH of the solution. Seven ar-
mored open limestone channels reduced acid concentra-
tions in AMD between 4 and 62%, while a sandstone
channel reduced acid concentration by only 2%. Open
limestone channels work best where the channel is con-
structed on steep slopes (>20%) and where flow veloci-
ties keep metal hydroxides in suspension, thereby lim-
iting their precipitation and plugging of limestone pores
in the channel. Open limestone channels should be de-
signed and constructed on watershed restoration and
AML projects to reduce acid concentrations in AMD
before reaching receiving streams, and in conjunction
with other passive treatment systems to maximize water
treatment and metal removal. Open limestone channels,
if constructed correctly, should be maintenance free and
provide AMD treatment for decades.

ACKNOWLEDGMENTS
The authors thank J.C. Sencindiver of West Virginia Uni-
versity, R.E. Knighton of North Dakota State University, and
several other anonymous reviewers for assistance and help-
ful comments.

View publication stats