TREATMENT OF HEAVY METALS USING AN ORGANIC
SULFATE REDUCING PRB
Ralph Ludwig (U.S. EPA, Ada, Oklahoma, USA)
Keith Mountjoy and Rick McGregor (Conor Pacific Environmental
Technologies, Vancouver, B.C., Canada)
David Blowes (University of Waterloo, Waterloo, Canada)
ABSTRACT: A pilot-scale permeable reactive wall consisting of a leaf-rich
compost-pea gravel mixture was installed at a site in the Vancouver area, Canada
to evaluate its potential use for treatment of a large dissolved heavy metal plume.
The compost based permeable reactive wall promotes microbially mediated
sulfate reducing conditions such that dissolved metals are precipitated out as
metal sulfides. The pilot-scale wall, measuring 10 m in length, 5.9 m in depth,
and 2-2.5 m in width, has demonstrated good effectiveness in removing dissolved
copper, cadmium, zinc, and nickel from ground water at the site over a 21-month
period since installation. Performance has been particularly strong within the
lower half of the wall where tidal influences are more limited and sulfate-reducing
conditions are more easily maintained. Dissolved copper concentrations decrease
from concentrations of over 4500 p.g/L in the influent ground water to less than
10 p.g/L within the lower half of the wall. Zinc, cadmium, and nickel
concentrations decrease from average concentrations of over 2300 |ag/L, 15 |ig/L,
and 115 ng/L, respectively to concentrations of less than 30 \xgTL, 0.2 |ng/L, and
10 ug/L, respectively within the lower half of the wall. The activity of sulfate
reducing bacteria is evidenced by a significant increase in sulfide concentrations
within the wall.
INTRODUCTION
As the number of successful permeable reactive barrier (PRB) installations
at contaminated sites continues to increase, permeable reactive barrier technology
is gradually being accepted as a viable alternative to conventional pump and treat.
Much of the focus and field success to date has involved the use of zero valent
iron-based permeable reactive barriers to treat chlorinated hydrocarbons such as
the chlorinated ethenes. With the exception of chromium, limited work to date
has focused on the use of permeable reactive barriers for treatment of heavy
metals. This paper presents results from an organic-based sulfate reducing pilot-
scale permeable reactive barrier installed at an industrial site in British Columbia
to treat heavy metals associated with acid rock drainage.
The concept of using organic-based systems to treat acid rock drainage is
not new. Engineered wetland systems have been used to treat acid rock drainage
impacted surface water runoff at mining sites for many years. The use of organic-
based permeable reactive barrier systems for treatment of acid rock drainage
impacted ground water was first proposed in 1990 (Blowes, 1990). The first full-
scale application of an organic-based sulfate-reducing permeable reactive barrier

-------
for treatment of acid rock drainage was at the Nickel Rim site near Sudbury,
Ontario in 1995 (Benner et al., 1997). Organic-based systems rely primarily on
the microbially mediated conversion of sulfates to sulfides by sulfate-reducing
bacteria residing in the organic media. The simplified reaction involving
reduction of sulfate and oxidation of a typical organic substrate such as lactate is
given below.
3S042" + 2CH3CHOHCOO" + 2H+ -> 3H2S + 6HCO3"	(1)
The reaction involves the production of both sulfide and bicarbonate. The
bicarbonate produced plays an important role in regulating the pH environment of
the sulfate-reducing bacteria. The sulfide produced is available to react with
dissolved metals to form insoluble metal sulfides in accordance with the
following reaction.
H2S + Me2+ -» MeS(s) + 2H+	(2)
where Me2+ denotes a heavy metal such as Cd, Cu, Ni, Pb, Zn, etc. In order to
ensure target metal removal from solution through the process of sulfate
reduction, a sufficient quantity of sulfide must be produced to meet the demand of
the heavy metal flux into the system. In a permeable reactive barrier application,
under an ideal design scenario, the amount of sulfide produced would just equal
the heavy metal flux into the wall. By avoiding excess production of sulfide, the
organic media is not needlessly consumed and the lifetime of the wall is
maximized.
Site Description. The test site is located in the Vancouver area, British Columbia
and has been impacted by acid rock drainage as a result of historical ore
concentrate handling and transfer practices occurring on site. The oxidation of
sulfide minerals on site has resulted in the underlying ground water being
extensively contaminated with heavy metals including dissolved cadmium (Cd),
copper (Cu), nickel (Ni), and zinc (Zn). Copper in ground water at the site has
been measured at some locations at concentrations exceeding 200,000 jig/L.
Impacted ground water at the site discharges into a nearby marine inlet thus
posing a potential threat to the shoreline ecosystem.
The geology at the test site is comprised primarily of deltaic deposits
consisting of sands and gravel with some cobbles. The shallow aquifer, which is
unconfined, begins at approximately 1 m below ground level (bgl) and extends to
at least 20 mbgl. Hydraulic conductivities in the upper 15 m of the aquifer are in
the 102 to 10"3 cm/sec range based on bail tests conducted (McGregor et al.,
1999). The average hydraulic gradient has been calculated at 0.001 based on 71-
hour water level averages. Metal contamination within the ground water is
confined to the upper 15 m of the aquifer with the majority of the contamination
being present in the upper 6 m.

-------
MATERIALS AND METHODS
Reactive Mixture. Selected batch tests were conducted with leaf-rich compost
(obtained from the City of Vancouver municipal composting facility) prior to wall
installation to ensure the compost would support sulfate reduction. The final
reactive mixture utilized in the wall consisted of 15% (by volume) leaf-rich
compost, 84% pea gravel, and 1% limestone and was based on the results of
previous laboratory and field studies (Benner et al., 1997; Waybrant et al., 1998).
The large percentage of pea gravel was required to achieve a minimum desired
hydraulic conductivity of 10"' cm/sec within the wall. The limestone was added
to ensure suitable initial pH conditions for the establishment of a sulfate reducing
bacteria population within the wall. The compost, pea gravel, and limestone were
thoroughly mixed by tossing and turning the materials in batches with a backhoe
bucket. The mixing process for each batch was conducted until a visually-based
homogeneous mixture of the components was obtained.
Pilot Wall Construction, The pilot wall was installed using cut and fill
excavation methods approximately 50 m inland from the shoreline of the site to
avoid ongoing construction activities along the inlet shoreline. As a result, the
wall was installed in a location of known up-gradient and down-gradient soil and
ground water contamination. The wall was constructed using a Komatsu Model
310 excavator to a depth of approximately 5.9 m and a length of 10 m. The width
of the wall is approximately 2.5 m at surface, narrowing to 2 m width at the final
depth. Excavation initially involved benching down approximately one meter to a
depth just above the water table. A guar gum based slurry was used during
trenching to prevent trench collapse and allow emplacement of the reactive media.
The reactive media was placed into the trench using a Manotowc 4500 clam shell
unit and Komatsu Model 310 excavator bucket.
A total of 17 multi-level wells were installed in and around the wall
following construction as shown in Figure 1. Each multi-level well consisted of
seven lengths of 1.27-cm internal diameter (ID) high density polyethylene tubing
with nytex screen affixed to a 1.9 cm (ID) PVC Schedule 40 center stalk at seven
discrete depths. This allowed for sampling of up to 119 sampling points at seven
depths within, up-gradient, and down-gradient of the wall.
Wall Sampling. Six discrete sampling events occurred over an initial 21-month
span following installation of the wall. The initial two sampling events covered
all 17 multi-level wells. Sampling events thereafter were limited to a center-
transect through the wall consisting of wells ML2, ML6, ML 10, ML 13 and
ML 16, as initial results indicated this transect was adequate to monitor wall
performance. Sampling events consisted of ground water level measurements,
and collection and analysis of ground water samples. Ground water samples were
collected using a low-flow peristaltic pump with Teflon tubing and filtered
through 0.45 |am cellulose acetate filters. Field measurements included pH, Eh
(corrected to standard hydrogen electrode), temperature, conductivity, alkalinity,
sulfide, and ferrous iron. Field measurement techniques and equipment including

-------
ML1
ML2
ML3
Bundle
ML5
ML6
ML7
ML4
ML8
ML9
ML10
ML11
Average Groundwater
Flow Direction
/
ML12
ML13
MLI4
Reactive Wall
ML15
ML16
ML17
FIGURE 1. Location of monitoring bundles relative to the reactive wall
(plan view).
QA/QC procedures employed are described in McGregor et al. (1999), Samples
were analyzed for anions by ion chromatography and dissolved metals by ICP-
OES and/or ICP-MS.
Indicators of sulfate reduction within the pilot-scale wall 21 months after
installation included an increase in dissolved sulfide concentrations, a decrease in
the redox potential, and a decrease in metal concentrations relative to the influent
ground water. Other indicators consistent with sulfate reduction included an
increase in alkalinity and increase in pH although the dissolution of limestone
within the reactive wall may have contributed significantly to these observed
increases. Vertically averaged results for the center transect multi-level wells
after a 21-month period are provided in Table 1. Metal concentration profiles
through the center of the wall are shown in Figure 2. Figure 2 shows that
treatment is generally greatest within the lower half of the wall where sulfate-
reducing conditions are likely more easily maintained. The upper half of the wall
shows poorer treatment presumably due to a greater susceptibility to influences
from tidal fluctuations (i.e. wet/drv cycles and back flushing) and perhaps also
oxygen intrusion from the surface. In addition, Figure 2 shows high
concentrations of metals immediately down-gradient of the wall at shallow
depths. This is attributed to the effects of recharge water from the surface that
becomes laden with heavy metals as it infiltrates through the overlying sulfide
impacted soils into the ground water on the down-gradient side of the wall.
As shown in Table 1, field measurements of pH and Eh at well ML2
indicate ground water entering the wall exhibits a relatively high redox potential
(Eh of +430 mV), a pH of 6.36, and an alkalinity of 89 mg/L as CaCOa. The
ground water entering the pilot-scale wall is also characterized by high
concentrations of copper, nickel, and zinc. As ground water passes through
RESULTS AND DISCUSSION

-------
ML2
ML6
ML10
ML13
Reactive Wall
uullULIBI)

^>10000 (ig/L H 1000-10000 ng/L ^200-1000 pg/L
~ 50-200 ng/L U 10-50 (ig/L D<10ng/L
(a)
ML 0
ML 13
Reactive Wall
jj >50 (ig/L
~ 1.0-5.0 ng/L
///////////////////
110-50 (ig/L ^ 5.0-10 ng/L
| 0.2-1.0 pg/L ~ < 0.2 |ig/L
(c)
Groundwater Flow
	~
2m
Groundwater Flow
	~
2m
ML2
ML6
ML 10
ML13
Reactive Wall

atSiiMatefiaftiifoiai
<300 |ig/L
~ 25-50 ng/L
1100-300 (ig/L £2 50-100 |ig/L
110-25 (ig/L ~<10Mg/L
(b)
ML2
Reactive Wall.
ML6
_J		
ML10
_J		
ML13
Zn
	
.		
I1 •
| > 10000 |ig/L g§ 5000-10000 pg/L 1000-5000 (ig/L
~ 100-1000 (ig/L d 30-100 pg/L []<30 (ig/L
(d)
FIGURE 2. Vertical profile for metals through center transect of wall 21 months after installation.
(Vertical scale 1 cm = 1.2 m)

-------
TABLE 1. Chemistry of the ground water entering, within, and exiting
the reactive wall.
Sample Location*
pH
*Eh
Alk.
Sulfide
Cd
Cu
Ni
Pb
Zn
ML-2 Influent
6.36
430
89
<1.0
15.9
4510
118
3.8
2396
ML-6 Wall front
6.76
177
155
1704
2,7
4.4
23.4
4.3
567
ML-10Wall back
6.63
141
202
613
<0.1
10.5
5.4
1.9
82.2
ML-13 Effluent
6.57
175
180
130
<0.1
7.7
6.5
0.7
27.5
* Vertically averaged values for monitoring points. EH values corrected to standard hydrogen
electrode. All units ng/L except pH, Eh (mV) and alkalinity (mg/L as CaC03).
the wall, alkalinity increases to an average of 155 mg/L (as CaCO:,) near the front
end of the wall and an average of 202 mg/L near the back end of the wall. A
slight increase in pH values is also noted, ranging from 6.76 near the front end of
the wall to 6.63 near the back end of the wall. Dissolved sulfide concentrations
within the wall increase to as high as 1704 jig/L and redox potential decreases to
+141 mV.
As ground water flows through the pilot-scale wall, dissolved copper,
nickel, cadmium, and zinc concentrations are significantly reduced. Copper is
reduced from a vertically averaged concentration of 4510 (ig/L in ground water
entering the wall to averages of 4.4 |ag/L and 10.5 p.g/L, at the front and back
ends of the wall, respectively. Figure 3 shows copper removal trends within the
lower half of the wall for six sampling events spanning a 21-month period. As can
Groundwater FIow^-
10000
Months
10D-F
Sampling Point
FIGURE 3. Copper concentration trends with time within lower half of
reactive wall.
be observed, a lower removal efficiency occurs over the first seven months
followed by a significantly higher removal efficiency thereafter. This is
presumably linked to a lag in the establishment of strong sulfate-reducing
conditions within the wall.
Nickel concentrations are reduced from an average of 118 \xgTL to
averages of 23.4 (ig/L and 5.4 p.g/L at the front and back ends of the wall,

-------
respectively. Within the lower half of the wall, nickel concentrations are reduced
to less than 10 jjg/L (Figure 2b). Cadmium concentrations are reduced from an
average of 15.9 jag/L to 2,7 pig/L and <1.0 jig/L, respectively at the front and back
ends of the wail. Within the lower half of the wall, cadmium is reduced to
concentrations of less than 0.2 (jg/L (Figure 2c). Zinc concentrations are reduced
from an average of 2396 ug/L to averages of 567 p.g/L and 82.2 p.g/L at the front
and back ends of the wall. Within the lower half of the wall, zinc concentrations
are reduced to less than 30 jig/L (Figure 2d).
Since the sulfate reduction process involves replacement of less dense
organic substrate (specific gravity 1 to 2) with more dense metal sulfide
precipitate (specific gravity 3 to 5), a decrease in permeability associated with
metal sulfide precipitation within the pilot-scale wall would theoretically not be
expected over time. Clearly, however, other precipitation reactions (e.g.
hydroxides) may occur within the wall depending on the site-specific conditions
in effect and these may ultimately impact the hydraulic conductivity of the wall.
The utility of an organic-based sulfate-reducing permeable reactive barrier
system will depend on site-specific needs. For larger plumes where large barrier
systems may be required, the low cost of using an organic substrate may be
attractive. Organic-based sulfate-reducing permeable reactive barrier systems can
also have the added ecological benefit of helping to restore down-gradient
ecosystems by removing ferrous iron acidity from ground water and
simultaneously generating a carbonate alkalinity plume. Removal of ferrous iron
from the ground water prior to discharge into a surface water body prevents iron
oxidation and the precipitation of ferric iron hydroxides, and production of acid
that would otherwise occur in accordance with the following reaction.
4Fe3+ + 02+10H2O -> 4Fe(OH)3 + 8H+	(3)
The production of a carbonate alkalinity plume associated with the sulfate
reduction process has been observed at the Nickel Rim site (Benner et al, 1997).
There, influent ground water was converted from a net acid producing potential of
7.8 to 46 meq/L to a net acid consuming potential of 16 to 45 meq/L following
passage through the organic substrate based reactive wall.
Wall longevity will be dependent on the reactive material maintaining its
permeability and reactivity properties. Benner et al (1997) calculated that the
organic-based reactive wall at the Nickel Rim Site consisting of 50% organic
substrate (by volume) could be effective for a minimum of 15 years based on
column study results. Metal sulfides precipitated out within the wall can be
expected to remain stable provided they are not subjected to oxidizing conditions.
As long as the metal sulfides remain below the water table, the oxidizing potential
is likely to be limited.
Two additional monitoring events have occurred on the pilot-scale wall
since the 21-month sampling event. Both of these sampling events continue to
demonstrate sulfate reduction and metals removal within the wall. The chemistry
from these sampling events is currently being validated and interpreted to
determine recent wall performance. It is intended that the pilot-scale wall will

-------
continue to be monitored to determine the long-term performance of the wall and
serve as an "early warning system" for "break-through" for full-scale installation.
CONCLUSIONS
The monitoring of geochemical ground water parameters of a compost-
based pilot-scale wall over a 21-month period has indicated that the leaf-rich
compost is providing a suitable organic carbon source for microbially mediated
sulfate-reduction and that dissolved metals (Cd, Cu, Ni, and Zn) are being
effectively attenuated by reactions within the pilot-scale wall. Continued
monitoring is planned to evaluate the long-term performance of the wall.
ACKNOWLEDGEMENTS
The authors would like to thank Environment Canada for approval to write
and submit this paper. The project management oversight of Eric Pringle and
relentless efforts of Mike Choi including his dedication and continuity to quality
field sampling and data analysis are also acknowledged.
Eric Pringle replaces Mick McGregor, who is an author on this paper.
DISCLAIMER
The views expressed in this paper are those of the individual authors and
do not necessarily reflect the views and policies of Environment Canada or the
U.S. Environmental Protection Agency.
REFERENCES
Benner, S.G., D.W. Blowes and C.J. Ptacek. 1997. "A Full-Scale Porous Reactive
Wall for Prevention of Acid Mine Drainage." Ground Water Monitoring and
Remediation. 17(4):99-107
Blowes, D.W. 1990. The Geochemistry, Hydrology and Mineralogy of
Decommissioned Sulfide Tailings: A Comparative Study. Ph.D. Thesis, University
of Waterloo, Waterloo, Ontario, Canada.
McGregor, R„ D. Blowes, R. Ludwig, E. Pringle, and M. Pomeroy (1999)
"Remediation Of A Heavy Metal Plume Using a Reactive Wall." In A. Leeson
and B.C. Alleman (Eds.), Bioremediation of Metals and Inorganic Compounds.
Batelle Press, Columbus, OH, 1999. 190pp.
Waybrant, K.R., D.W. Blowes and C.J. Ptacek. 1998. "Selection of Reactive
Mixtures for Use in Permeable Reactive Walls for Treatment of Mine Drainage."
Environ. Sci. Technol. 32:1972-1979.

-------
NRMRL-ADA-00309
TECHNICAL REPORT DATA

1. REPORT NO,
EPA/600/A-OG/072
2.
3. REC
4. TITLE AND SUBTITLE

5. REPORT DATE
TREATMENT OF HEAVY METALS USIHG AH ORGANIC SULFATE REDUCING PRB

6. PERFORMING ORGANIZATION CODE
7. AUTHOR (S) 1Kalph Lucbrig, JKeith Mount joy and Rick McGregor, and
'David Blowes
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
'USEPA,ORD,NRMRL, SPRD, P.O. Box 1198, Ada, OK 74821
'Conor Pacific tnvironmental Technologies, Vancouver, B.C., Canada
'University of Waterloo, Hatarloo, Canada
10. PROGRAM ELEMENT NO.
TKKY1A
11. CONTRACT/GRANT NO.
In-House
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
13. TYPE OF REPORT AND PERIOD COVERED
Symposium Paper
National Risk Management Research Laboratory
Subsurface Protection & Remediation Division
P.O. Box 1198, Ada, Oklahoma 74821
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
Jerry H. Jones (580)436-8593
16. ABSTRACT


A pilot-scale permeable reactive wall consisting of a leaf-rich compost-pea gravel mixture was installed at a site in the
Vancouver area, Canada to evaluate its potential use for treatment of a large dissolved heavy metal plume. The compost
based permeable reactive wall promotes microbially- medi ated sulfate reducing conditions such that dissolved metals are
precipitated out as metal sulfides. The pilot-scale wall, measuring 10 m in length, 5.9 m in depth, and 2-2.5 m in width,
has demonstrated good effectiveness in removing dissolved copper, cadmium, zinc, and nickel from ground water at the
site over a 21-month period since installation. Performance has been particularly strong within the lower half of the wall
where tidal influences are more limited and sulfate-reducing conditions are more easily maintained. Dissolved copper
concentrations decrease from concentrations of over 4500 ppb in the influent ground water to less than 10 ppb, within the
Iowa half of the wall. Zinc, cadmium, and nickel concentrations decrease from average concentrations of over 2300 ppb,
15 ppb, and 115 ppb, respectively to concentrations of less than 30 ppb, 0.2 ppb, and 10 ppb, respectively within the lower
half of the wall. The activity of sulfate reducing bacteria is evidenced by a significant increase in sulfide concentrations
within the wall.
17.
KEY WORDS AND DOCUMENT ANALYSIS

A. DESCRIPTORS
B. IDENTIFIERS/OPEN ENDED TERMS
C. COSATI FIELD, GROUP



18. DISTRIBUTION STATEMENT
19. SECURITY CLASS
UNCLASSIFIED
21. NO. OF PAGES 8
RELEASED TO TBS PUBLIC
20. SECURITY CLASS
UNCLASSIFIED
22. PRICE
EPA FORM 2220-1 (REV.4-77! PREVIOUS EDITION IS OBSOLETE

-------