treatment of mine drainage using prb

7212019 treatment of mine drainage using PRB

httpslidepdfcomreaderfulltreatment-of-mine-drainage-using-prb 18

Treatment of Mine Drainage UsingPermeable Reactive BarriersColumn Experiment s

K R W A Y B R A N T dagger C J P T A C E K dagger Dagger A N D

D W B L O W E S dagger

Department of Earth Sciences University of WaterlooWaterloo Ontario N2L 3G1 Canada and National Water Research Institute 867 Lakeshore RoadBurlington Ontario L7R 4A6 Canada

Permeable reactivebarriers designed to enhance bacterialsulfate reduction and metal sulfide precipitation havethe potentialtopreventacid minedrainageandthe associatedrelease of dissolved metals Two column experimentswere conducted using simulated mine-drainage water toassess the performance of organic carbon-based reactivemixtures under controlled groundwater flow conditions

The simulated mine drainage is typical of mine-drainagewater thathasundergone acidneutralization within aquifers

This water is near neutral in pH and contains elevatedconcentrations of Fe(II) and SO4 Minimum rates of SO4

removal averaged between 500 and 800 mmol d-1 m-3 overa 14-month period Iron concentrations decreased frombetween 300 and1200 mgL in the influentto betweenlt001and 220 mgL in the columns Concentrations of Zndecreased from 06-12 mgL in the input to between 001and 015 mgL in the effluent and Ni concentrationsdecreased from between 08 and 128 mgL to lt001 mgL

The pH increased slightly fromtypical input values of 55-60 to effluent values of 65-70 Alkalinity generally lt50mgL (as CaCO3) in the influent increased to between300 and 1300 mgL (as CaCO3) in the effluent from thecolumns As a result of decreased Fe(II) concentrations andincreased alkalinity the acid-generating potential of thesimulatedmine-drainagewaterwas removedanda netacid-consuming potential was observed in the effluent water

In t roduct ionThe oxidation of sulfide minerals within mine-tailingsimpoundments results in the generation of low quality watercharacterized bylow pH andhigh concentrations of dissolvedSO4 Fe(II)and other metals(1)AsthislowpHwatermigratesthrough the tailings and underlying aquifers it undergoes aseries of dissolutionand precipitation reactions that generally

result in acid neutralization (Figure 1 1 2 ) The resulting plume water is characterized by near-neutral pH reducedE h and elevated concentrations of dissolved SO4 Fe(II) andother metals This groundwater ultimately discharges tosurface water flow systems where oxidation of Fe2+ to Fe3+

andthe subsequentprecipitationof ferricoxyhydroxides canresult in the generation of acidity

The use of bacterially mediated sulfate reduction inpermeable reactive barriers is an alternative technique for

the remediation of acid mine drainage (3 -6 ) Permeablereactive barriers are installed in the path of migrating mine-drainage water by excavating the aquifer material andreplacing it with a permeable reactive material (Figure 1)These barriers are designed to remove dissolved Fe2+ andother metals from plumes of flowing tailings-impactedgroundwater through enhanced biological sulfate reduction

and metal sulfide precipitation (5 -

7 )Under favorable conditions sulfate-reducing bacteria(SRB) catalyze the oxidation of organic carbon coupled withthe reduction of sulfate to sulfide through the following general reaction (6 8 )

where (CH2O)x (NH3) y (H3PO4)z represents organic matterundergoing oxidation and x y and z are stoichiometriccoefficients The reduction of SO4 produces H2S releasesHCO3

- and results in an increase in alkalinity and pH Thereaction also releases ammonia and dissolved phosphate which is utilized by the bacteria or released into the

environment (9 ) An increase in H2S concentrations coupled with the low

solubility of metalsulfides resultsin theremoval of dissolvedmetals as metal sulfides

where Me2+ denotes a divalent metal such as Cd Fe Ni CuCo and Zn MeS represents an amorphous or poorly crystalline metal sulfide (10 )

The conditions typically found within a reactive barrierenvironment are well suited to SRB Permeable reactivebarriers provide dissolved C N and P and the plume waterentering thebarrier provides high concentrations of SO4 Feand other metals all of which are necessary for growth and

reproductionResults of columnexperiments designed to assess sulfate

reduction rates and metal removal capacities of selectedreactive mixtures under controlled flow conditions arepresented here These column experiments were conductedunder conditions typical of the environment in whichpermeable reactive barriers would be applied Laboratory column experiments allow the evaluation of the removal of dissolved metals including Fe Ni Zn and SO4 under closely controlled flow conditions and allow the direct evaluationof sulfate and metal removal rates

MethodologyColumn Design and Experimental Setup Two 5 cmdiameter40 cm long acrylic columns were packedwith 5 cm

of silica sand and crushed pyrite on the bottom followed by 335 cm of reactive mixture and topped with 15 cm of silicasand Thebottomand toplayersseparate thereactive mixturefromthe influentand effluentportsThe influentand effluentports wereseparated fromthe packingmaterial witha coarse-meshed nylon screen followed by a fine-meshed NYTEX screen to prevent material from being washed out of thecolumns Crushed pyrite was added to the bottom layer toconsume O2 and Fe(III) in the influent water The reactivemixtures consisted of an organic source a bacterial sourcea neutralizing agent and a nonreactive porous medium(Table 1) The organic carbon source included mixtures of wood chips sawdust composted municipal sewage sludge

Corresponding author phone (519)888-4878 fax (519)746-3882e-mail blowessciborguwaterlooca

dagger University of WaterlooDagger National Water Research Institute

2(CH2O)x (NH3) y (H3PO4)z + x SO42-

f

2x HCO3-+ x H2S + 2 y NH3 + 2z H3PO4 (1)

Me2++ S2-

f MeS (2)

Environ Sci Technol 2002 36 1349-1356

101021es010751g CCC $2200 983209 2002 American Chemical Society VOL 36 NO 6 2002 ENVIRONMENTAL SCIENCE amp TECHNOLOGY 9 1349Published on Web 02142002



and leaf compost from a municipal recycling program Two

different reactive mixtures were used one containing leaf mulch sawdust sewage sludge and wood chips (column 1)and the other containing leaf mulch and sawdust (column2) These mixtures correspond closely to batch mixturespreviously evaluated in bench-scale batch studies (5 )andina full-scale reactive barrier installation (6 7 ) The bacterialsourcewas collectedfromthe anaerobic zone ofa localcreek

The reactive mixtures were soaked in a CaCO3-saturatedsolution containing 1000 mgL SO4 as CaSO4 and5 sodiumlactate The CaCO3-saturated water was made by adding CaCO3 in excess to a known volume of double-deionized water and was bubbled with high-purity CO2(g) for severalhours This solution was left to equilibrate with the atmo-sphere for several days to weeks before use The mixtures were packed into the columns wet to prevent aeration of the

bacterial source The CaSO4 and sodium lactate were addedto help acclimate the bacteria and promote sulfate-reducing conditions Columns were covered with aluminum foil toexclude light and prevent growth of photolithotrophicbacteria Within30-40 days an activepopulation of SRBwasestablished as was evident from the production of H2S gasThe columns were transferred into an anaerobic gloveboxThe anaerobicity of the glovebox was monitored regularly using methylene blue anaerobic strip indicators (GasPak)

The same input solution was used for both columns Theinitial input solution (feed) contained 1000 mgL SO4 asCaSO4 made up using double-deionizedwater saturated withCaCO3 This input solution was used for the first 4 pore

volumes (pv) (feed 0 Table 2) after which simulated mine water was used for the remainder of the study (feeds 1-8Table 2) The chemical composition of the simulated mine water was based upon observations made at a mine-tailingsimpoundment (6 ) Newinputsolutions were prepared every 5-6 weeks This allowed the input solution to be modified

as necessary A conservativetracer test was performed on both columns

by adding NaBr to the feed 1 input solution and measuring effluent Br-concentrationsover timeThe conservativetracertests were used to determine the effective pore volumes of the saturated columns

Influent concentrations of Ni andMn were higher in feed4 as a result of laboratory error The concentration of Fe inthe input solution was decreased in feed 5 after a minorbreakthrough of Fe was observed in both of the columnsThis decrease in Fe was accompanied with decreases inconcentrations of dissolved SO4 and metals After 185 pvNH4Cl was added as a nutrient to determine if the mixture was nitrogen-limited The amount of NH4Cl added to theinfluent corresponded to 7 mgL nitrogen This value was

basedupona CNratio of 251The carboncontentwas baseduponearly DOC measurements in the effluentThe chemicalcomposition and input duration of each input solution aregiven in Table 2 The simulated mine water was stored in anO2-free environment to prevent oxidation of Fe2+ to Fe3+thus keeping the pH relatively constant at 55-65 (Table 2)

A variable speedmultichannel pump(ISMATEC)was usedto deliver the influent to each of the columns to yield anaverage linearvelocity of =10 ma Stainlesssteeltubing wasused between thefeedreservoirand theglovebox (ie whereit was exposed to an oxic environment) and either stainlesssteel or PharMedtubing wasused in anoxicareasThe columneffluentsolutionswere collected in flow-throughsample cells within the glovebox and overflow was directed to waste jugsoutside of the glovebox Flow rates (mLh) were determined

by measuring the overflowSample Collection Column effluent samples were col-

lected within the glovebox to determine pH E h alkalinityand concentrations of dissolved Br Ca Cd Cl Fe H2S KMg Mn Na Ni o-PO4 SO4 Zn dissolved organic carbon(DOC) dissolved inorganic carbon (DIC) δ34S and volatilefatty acids (acetic acid propionic acid butyric acid andformic acid) Measurements of pH E h and alkalinity weredeterminedimmediatelyafter samplingThe pH and E h weremeasured on unfiltered samples in sealed cells to minimizeO2 exposure and a filtered sample was used for measuring alkalinity Electrode calibration and performance werechecked before and after each sample measurement The

FIGURE 1 Schematic diagramof permeable reactive barrier ata mine drainage site

TABLE 1 Composit ion of Reactive Mixtures Used in ColumnExperimentsa

mixturecompositioncolumn1

(drywt)column2

(drywt)

bottom layer (5 cm)crushed pyrite 10 10silica sand 90 90

middle layer (335 cm)composted leaf mulch 195 23wood chips 8sawdust 105 22sewage sludge 10creek sediment 43 44agricultural limestone 2 3silica sand 7 8

top layer (15 cm)silica sand 100 100

a Flow is from bottom to top

1350 9 ENVIRONMENTAL SCIENCE amp TECHNOLOGY VOL 36 NO 6 2002



pH electrode (Orion Ross Sure-Flow Combination pH 81-65) was calibrated using standard 40 and 70 buffers(referenced to NIST SRM) and the E h electrode (Orion 96-78) was checked using Zobellrsquos solution (11) The alkalinity was determined using standardizedH2SO4 and a Hach digitaltitrator Samples were filtered through a 045- microm celluloseacetate filter paper and both acidified and nonacidifiedsamples were collectedand storedat 5 degC until analyseswerecompleted within 60 days unless otherwise specified Ana-lytical procedures were the same as those described previ-ously for batch tests conducted using similar materials (5 )

Mineralogical Characterization After 30 months of flow

through column 2 a subsample of the reactive material wasremoved from column 2 and examined to determine thenature of the reaction products Column material from theinfluent end was examined using a Hitachi S-4500 fieldemission secondaryelectronmicroscope EDX analyseswereobtained using a Noran Instruments light element detectorattached to an ISI DS-130 scanning electron microscope

Results and DiscussionConservative Tracer Test and Modeled Velocity and Dis-persion Coefficient A pulse style conservative tracer test was performed to determine the effective pore volume of each columnunder saturated conditions Calculated effectivepore volumes for columns 1 (leaf mulch sawdust sewagesludge and wood chips) and 2 (leaf mulch and sawdust)

were 420 and 480 mL respectively The transport modelCXTFIT (12 ) was used to determine the velocity and disper-sion coefficient in each column Fittedvelocity estimates forcolumns 1 and 2 were 95 (( 001) and 97 (( 002) myrrespectively calculated dispersion coefficients were 018 ((20times 10-3) m2yr for column 1 and 013 (( 20times 10-3) m2yrfor column 2

Measured Sulfate Reduction Rates and Metal RemovalCapacity The long-term effectiveness of permeable reactivebarriers depends on maintaining the conditions conduciveto biological sulfate reduction Several factors mayaffect therate of sulfate reduction among these the reactivity of thecarbon source and the residence time are perhaps the mostimportant Sulfate reduction rates are used to indicate thereactivity of the mixture Sulfate reduction rates for thecolumn experiments were approximated by sulfate-removalrates (13 ) which were determined by subtracting themeasuredeffluent SO4 concentrationssubjectto conservativetransport and SO4 reduction from modeled effluent SO4

concentrations considering conservative transport only anddividingthis quantity bya time or volumeinterval Modeling the effluent concentration considering dispersion and dis-placement allowed a more accurate calculation of sulfateremovalastheSO4 concentrationof the inputsolution variedover time CXTFIT was used to model effluent SO4 concen-trations using the fitted velocity and dispersion valuespreviously obtained

The geochemical speciation mass-transfer modelMINTEQA2 (14 ) was used to indicate chemical equilibrium

reactions potentially controlling the concentrations of dis-solved constituents in the effluent water The originaldatabase was modified to make it consistent with the WATEQ4F database (15 ) Saturation indices (SI) where SI )log(IAPK sp) for various mineral phases were calculated forthe column effluent solutions assuming that all measuredalkalinity was present as carbonate species pH was enteredandfixed as themeasured valueand E h was calculated basedupon measured concentrations of the redox couple HS-SO4

2- The assumption that all measured alkalinity wascarbonate was based upon earlier observations in batchexperiments (5 ) which showed that DOC had a negligibleaffect on calculated SI values at the observed DOC concen-trations Concentrations of Ni were enteredas the detectionlimit (0006 mgL) because measured concentrations in theeffluent were often at or below detection in both columns

Column experiments were runfor over14 months During this time gt27 pv (11-13 L) of simulated mine water waspumped through each column The input solution variedover time with SO4 input concentrations ranging from 1000to 4000 mgL (10-40 mmolL) (Table 2) consequently SO4

effluent concentrations also varied over time Monitoring of

TABLE 2 Durat ion (Pore Volumes (pv)) and Composi t ion of Input Solut ion Used in Column Experiments 1 and 2

inputsoln

durationcolumn1(pv)

durationcolumn2(pv) pH

E h(mV)

alkalinity(mgLasCaCO3)

SO4

(mgL)Fe

(mgL)Zn

(mgL)Ni

(mgL)Mn

(mgL)Na

(mgL)Ca

(mgL)K

(mgL)Mg

(mgL)Br

(mgL)Cl

(mgL)

feed 0 0-047 0-59 1010 0feed 1 47-74 59-83 64 0 16 1471 3210 694 072 094 975 468 506 274 242 26 142feed 2 74-92 83-101 58 5 74 548 3894 1181 095 140 129 500 99 317 320 0 482feed 3 92-120 101-129 64 1 142 450 3455 876 058 131 128 488 92 323 317 0 516feed 4 120-155 129-161 64 6 179 340 3660 956 097 123 198 488 83 292 288 771feed 5 155-184 161-187 66 0 2 259 2000 444 122 234 141 141 205 118 121 091feed 6 184-215 187-217 61 6 23 112 1623 379 085 160 896 896 214 79 83 154

feed 7 215-249 217-248 66 1 307 278 1353 334 057 336 862 862 220 72 77 176feed 8 249-277 248-275 64 3 180 238 1400 316 144

FIGURE2 AqueousconcentrationsofSO4FeZnandNi andvaluesof pH andE h for the influent and effluent for column 1 Influentconcentrations havebeendisplacedby1pore volume

VOL 36 NO 6 2002 ENVIRONMENTAL SCIENCE amp TECHNOLOGY 9 1351



effluent SO4 H2S pH E h and alkalinity (Figures 2 and 3)indicated that sulfate-reducing conditions were maintainedthroughout the experiments

The initial input solution contained 1000 mgL SO4 asCaSO4 and did not contain dissolved metals This inputsolution was used for the first 47 and 59 pv in columns 1and 2 respectively (Table 2) Sulfate reduction reactionsduring this period resulted in the removal of SO4 from aninput concentration of 1000 mgL (10 mmolL) tolt20 mgL(021 mmolL) in both columns (Figures 2 and3) Thealmostcomplete (98) removalof SO4 suggestedthat the initial SO4

concentration limited the extent of SO4 reduction As a resultof low SO4 concentrations in the effluent and the lack of dissolved metals in the influent it was not possible toaccurately estimate the potential maximum rate of SO4

reduction or the metal removal capacity To quantitatively estimatethe rate ofSO4 reductionand to determine the metalremoval capacity withinthe columns theinputsolutionwasmodified to contain approximately 3000-4000 mgL (30-40mmolL) SO4 700-1200 mgL (12-20 mmolL) Fe and low concentrations of Zn Ni Mn and other dissolved metals(Table 2) This input solution was used between pv 5 and 16after which the SO4 and Fe concentrations were lowered

(Table 2)Subtracting the measured effluent SO4 concentrations

subject to conservative transport and sulfate reduction fromthe modeled effluent SO4 concentrations subject to conser-vative transport only (ie dispersion and displacement) forthis period (5-16 pv Figures 2 and 3) indicates a continuedremoval of =1000 mgL SO4 (700 mmol d-1 m-3) in bothcolumns Calculated sulfate-removalrates for column 1 (leaf mulch sawdust sewage sludge and wood chips) decreasedafter 18 pv from =900 to 500 mgL SO4 (700-400 mmol d-1

m-3) and for column 2 remained relatively constant (leaf mulch andsawdust) at=1000 mgL SO4 (700 mmol d-1 m-3)for the duration of the study The more extensive sulfate

removal at the early part of the experiment may be due toa greater abundanceof labileorganic carbon both as carbonpresent in the sodium lactate added to the column materialand as the most labile fraction of the solid organic carbonmaterial The subsequent decline in sulfate reduction may suggest a depletion of the most labile organic carbon formsThe calculated sulfate reduction rates ranged between 400and 1200 mmol d-1 m-3 with an average of between 500 and750 mmol d-1 m-3 (Tables 3 and 4) These sulfate reductionrates are higher than previously reported rates of between02 and 600 mmol d-1 m-3 (16 -18 ) These previous studieshad shorter residence times and lower input pH values ascompared to the columnstudies presented here Thesulfatereduction rates observed in these experiments however aresimilar to those observed in a field-scale permeable reactivebarrier system (6 10 )

The reduction of SO4 was accompanied by a decrease inthe effluent water E h and an increase in pH alkalinity andH2S concentrations (Figures 2 and3) Insystems characterizedby high rates of SO4 reduction the pH is usually buffered tobetween 6 and 7 (6 19 ) A slight increase in the pH of theeffluent water wasobservedfromtypical inputvalues of 55-65 to typical effluent values in both columns of 65-70(Figures 2 and3) ThispH increaseis attributedto bicarbonatebufferingas a result of the generation of bicarbonate throughsulfate reduction reactions (eq 1) and is supported by increases in alkalinity (Figures 4 and 5)

The alkalinity increased fromlt50 mgL (05 mmolL) (asCaCO3) in the influent to between 900 mgL (9 mmolL) and

FIGURE3 AqueousconcentrationsofSO4FeZnandNiandvaluesof pH andE h for the influent and effluent for column 2 Influentconcentrations havebeendisplaced by 1pore volume

TABLE 3 Calculated Sulfate Reduction Rates for Column 1 a

inputsolution

porevol(pv)

sulfatereduction

rate(mgof SO4 L-1pv-1)

sulfatereduction

rate(mmol of SO4 d-1 m-3)

feed 0 no metals 00-47 862 441feed 1 metals added 47-74 1089 768feed 2 74-92 842 566feed 3 92-120 1116 775feed 4 120-155 749 549

feed 5 155-184 924 700feed 6 NHCl4 added 184-215 638 455feed 7 215-249 592 429feed 8 249-277 498 365average 00-2770 812 510

a Sulfate reduction rates were calculated by subtracting measuredeffluentsulfateconcentrations subjectto dispersiondisplacementandbacterial sulfate reduction from modeled sulfate concentrations con-sidering dispersion and displacement only


inputsolution

porevol(pv)

sulfatereduction

rate(mgof SO4 L-1 pv-1)

sulfatereduction


feed 0 no metals 00-59 863 548feed 1 metals added 59-83 792 512feed 2 83-101 965 631feed 3 101-129 1379 934feed 4 129-161 1786 1209feed 5 161-187 1161 789feed 6 NHCl4 added 187-217 1204 820feed 7 217-248 1156 788feed 8 248-275 898 613average 00-275 1134 761





1100 mgL (11 mmoL) (as CaCO3) in the effluent initially and then slowly began dropping after 10 pv to between 300and 600 mgL (3 and 6 mmolL) (as CaCO3) The high initialalkalinity values suggest that the rate of sulfate reduction in

the early part of the experiment was underestimated or is aresultof alkalinityreleased through otherbacterially mediatedreactions such as methanogenesis Calculated sulfate reduc-tionrateswerebaseduponSO 4 concentrationsin the influentThedissolution of gypsum which wasaddedto thecolumnsas they were setup was not accounted for in the calculationof sulfate reduction rates Calculated SI values for the first15 pv indicate that the effluent water was saturated withrespect to gypsum but became undersaturated over timesuggesting depletion of the initial mass of gypsum (Figures6 and 7) Geochemical analyses of Ca showed an initialincrease in Ca from an average input concentration of 200mgL (5 mmolL) to an effluent concentration between1000and 1500 mgL (25 and 37 mmolL) followed by a gradualdecrease in concentrations over time until effluent and

influent concentrations were approximately the same (Fig-ures 4 and 5) The initial high Ca concentrations may beattributed to gypsum dissolution followed by calcite pre-cipitation Spiro and Aizenshtat (20 ) hypothesized that highrates of sulfate reduction may lead to the precipitation of calcite through the reaction

Decreasing Ca concentrations decreasing SI values forgypsum and positive or near-zero SI values for calciteobserved in the experimental results are consistent with eq3 (Figures 4-7) Effluent concentrations of other major ions

(Mg Na and K) followed influent concentrations closely (Figures 4 and 5)

Dissolved sulfide is produced through sulfate reductionreactions (eq 1) Concentrations of H2S increased fromnondetectable concentrations in the influent to values

FIGURE4 AqueousconcentrationsofCaMgNaKalkalinityandcalculated acid generating potential (AGP) for the influent andeffluentforcolumn1Influentconcentrationshavebeendisplacedby 1pore volume

CaSO4 + 2(CH2O) f CaCO3 + H2S + CO2 + H2O (3)


FIGURE6 Saturationindicesfor selectedcarbonatesulfate andsulfide mineral phases for effluent samples fromcolumn 1




measured as high as 140 mgL (42 mmolL) in the effluent(Figures 2 and 3) The observed increases in H2S wereaccompanied by decreases in concentrations of dissolvedmetals in the effluent water (Figures 2 and 3) The mostnotable were decreases in Fe which decreased from aninfluent concentration of between 700 and 1200 mgL tolt01mgL (125 and215 mmolLtolt0002mmolL) (Figures2 and 3) Concentrations of Zn decreased from 10 to lt015mgL (0015 to lt0002 mmolL) and Ni decreased from 15to lt001 mgL (0025 to 00002 mmolL) (Figures 2 and 3)

Breakthrough of Fe occurred in both columns 9 pv afterthe influent Fe and SO4 concentrations were increased (14and 15 pv in columns 1 and 2 respectively) coinciding with

removal of between 2900 and 3600 mg of Fe After=

16 pvconcentrations of SO4 and Fe in the input solution werelowered by =50

Effluent H2S concentrations showed breakthrough of Fecoincidedwith the almost complete removal of H2S (Figures2 and 3) This suggests that the breakthrough representedthe amount of dissolved Fe in excess of the amount of H 2Sproduced and subsequently utilized in the precipitation of metal sulfides during the reduction of 500-1000 mgL SO4Effluent concentrations of Zn and Ni remained low through-out theexperiment The highconcentrations of H2S observedatpv8and9areattributedtoabuildupofH 2S resulting fromtheabsenceof Fe andother metals fromearlyinput solutions

Stoichiometric calculations basedupon eq 1 indicate thatthe reductionof between 500and1000mgL(5 and 10 mmolL) SO4 produces 180-350 mgL (5 and 10 mmolL) H2SObservations from a field-scale reactive barrierat the NickelRim site in Ontario suggest that the most abundant ironsulfide precipitate is mackinawite with an approximatestoichiometry for FeS of 11 (10 )IfalltheH2S produced wasutilized in theprecipitation of Fe as FeS(eq 2)then 300-600mgL of Fe would be removed from solution Precipitationof all of the influent Fe Zn and Ni as FeS ZnS and NiS(using the chemistry of feed 7 Table 2) 230 mgL or 68mmol of H2S is required This amount corresponds to thereductionof650mgLofSO 4 The calculatedsulfate reductionrate corresponding to this input solution (Tables 3 and 4)exceeds this value in column 2 (leaf mulch andsawdust) butnot in column 1 (leaf mulch sawdust sewage sludge and

wood chips) Consistent with these calculationseffluent fromcolumn 2 showed complete removal of Fe Zn and Ni andelevated concentrations of H2S conversely effluent fromcolumn 1 containing measurable concentrations of Fe andH2S was notdetectedor was just above detection (001 mgLdetection limit)

The water chemistry data suggest that the primary mechanism for metal removal is through the precipitationof sparinglysoluble metalsulfidesand theamount removedis controlled by the H2S concentration which is in turndependent upon the sulfate reduction rate Geochemical

calculations indicate thatthe effluent water in both columnsapproached or attained saturation withrespectto themineralphases rhodochrosite (MnCO3) mackinawite (FeS09) andFeS These calculations suggest that the carbonate phaserhodochrosite or a less crystalline precursor controlled theconcentrationof Mn andthatthe sulfidephases mackinawiteand FeS controlled the concentration of Fe For the first 14pvthe SI valuesfor siderite(FeCO3) werenegativesuggesting a tendency forFeCO3 to dissolve (Figures 6 and7) Theinitialnegative SI values for FeCO3 suggest that SO4 reduction andprecipitation of iron sulfide phases was the mechanismresulting in Fe removal After 14 pv the effluent approachessaturation with respect to FeCO3 suggesting that the pre-cipitation of both the sulfide phases FeS and mackinawiteand the precipitation of the carbonate phase siderite may

affect the concentration of Fe (Figures 6 and 7) Supersatu-ration with respect to siderite has also been observed in afield-scale permeable reactive barrier system (6 )

Negative SI values for smithsonite (ZnCO3) positive SIvalues for amorphous ZnS and sphalerite (ZnS) and low concentrations of Zn in column effluent water samplesindicate that Zn concentrations may be controlled by thesolubility of a zinc sulfide phase (Figures 2 3 6 and 7)

The acid-generating potential (AGP) is a measure of thepotentialfor water to produce acidity The AGP is calculatedby subtracting the acid-producing potential (APP) from theacid-consuming potential (ACP) The APP is a measure of the potential amountof acidity released throughthe oxidationof Fe2+ to Fe(OH)3 (eqs 4 and 5 shown below) and theoxidation of HS- to produce acidity (eq 6) and the ACP isa measure of the potential to consume acidity throughalkalinity represented here as the formation of H2CO3(aq)from protonation of CO3

2-(aq) (eq 7)

A positive AGP value indicates a potential to generate acidityand a negative AGP indicates an acid-neutralizing capacityGiven the close relationship between sulfate reductionreactions and AGP ie the generation of alkalinity (eq 1)andthe precipitationof Fe2+asmetalsulfides(eq 2)it followsthatsufficientrates of sulfatereduction willlead to a decreasein AGP The AGP of the column influent versus that of thecolumn effluent are shown in Figures 4 and 5 The decreaseinAGPfromanetAPPintheinputtoanetACPintheeffluentis a direct consequence of the removal of Fe2+ a potentialacid-producing metal and the generation of alkalinity

The observed decrease in SO4 and dissolved metalsconcentrations and the increase in alkalinity and H2Sconcentrations are attributedto bacterially mediated sulfatereduction The predominant isotopes of sulfur in natural

FIGURE7 Saturationindicesfor selectedcarbonatesulfate and

sulfide mineral phases for effluent samples fromcolumn 2

Fe2+(aq) + 14O2(aq) + H+(aq) f

Fe3+(aq) + 12H2O(l) (4)

Fe3+(aq) + 3H2O(l) f Fe(OH)3(s) + 3H+(aq) (5)

HS-(aq) + 2O2(aq) f SO42-(aq)+ H+(aq) (6)

CO32-(aq) + 2H+(aq) f H2CO3(aq) (7)




systems are 32S and 34S The relative abundance of these twoisotopes is described by their isotopic ratio which for sulfuris defined as

During sulfate reduction bacteria reduce the lighter 32Sisotope in preference to 34S in metabolic functions Asbacterially mediated sulfate reduction proceeds the lighter32S isotope predominates in the reduced sulfide and theheavier 34S isotope accumulates in the residual sulfate (21)Theextentof 34S enrichmentin theresidualsulfate is indicatedby the residual fractionation factor ( f ) Determination of theδ34S andcalculationof theresidual fractionation factorof 34Sin SO4 in the effluent water shows an enrichment of δ34S of between-28permiland-46permilasSO4 concentrations decreased(Table 5) These values are within the range (-155permil to-

60permil) observed in active biological systems (22 ) Further-more the enrichment of 34S increased as the fractionationfactor decreased which also would be expected of SO4

removalby bacteriallymediatedsulfate reduction (22 ) Undersurficial conditions abiotic contributions to isotopic frac-tionation of sulfate are expected to be insignificant The 34Senrichment observed in the column effluent water indicatesthat the sulfate removal observed can be attributed tobacterially mediated sulfate reduction

Mineralogical study of the reactive material following 30months of column operationwas usedto identifythe reactionproducts Small (2-10 microm) spheres were observed on thesurfaces of woodparticles within the reactive mixture (Figure8) Energy-dispersion X-ray analysis indicated that thesespheres are composed primarily of Fe and S with minoramounts of Ca Si Mg and O These spheres are interpretedto be precipitates of ferrous monosulfide or mackinawiteMackinawiteand amorphous FeS were identifiedin samplescollected from a full-scale reactive barrier near SudburyOntario (10 )

Possible Limitations to the Rate of Sulfate ReductionIt was anticipated that the removal of SO4 would increase asthe bacterial population acclimatized to the increase ininfluent SO4 concentration The extent of SO4 removalremained relativelyconstant over time with a slight decreasein column 1 (Tables 3 and 4) This observation suggests thatthe influent SO4 concentration no longer limited the extentof sulfate reduction Other factors that may limit the rate of sulfate reduction include the availability of nutrients prin-

cipally N and P insufficient concentrations of labile Cinadequate retention times or substrate limitations

Concentrations of P N and C were determined on eachorganic substrate and calculated for each reactive mixture(Table 6) and concentrations of DOC organic acids o-PO4and NH4 were determined on columneffluent samples(Table7) Results indicated that there was measurable C (in theform of propionic acid) and inorganic P (as o-PO 4) butinorganic N (as NH4) was not detected suggesting that

inorganic N may have limited the rate of sulfate reductionPrevious batch studies (5 ) indicated that aqueous NH4 wasrapidly depleted in several of the mixtures tested Todetermine whether the sulfate reduction rate was N limited7 mg L of N as NH4Cl was added to the influent atapproximately 185 pv (feed 6) Rates of sulfate reductionremained relatively constant after NH4Cl was added It isunclear whether the addition of N (as NH4Cl) was beneficialmore work is necessaryHowever samples collectedfor δ34Sindicate greater enrichment factors after addition of NH4Cl(Table 5) Inadequate retention times substrate limitationsand lack of sites suitable for microbial attachment remainpossible limitations to the rate of sulfate reduction

Estimating the Longevity of the Carbon Source Thelongevity of the carbon source was estimated by calculating

the mass of C loss from the columns It was assumed thatthe only C loss that occurred in the column was throughleaching and direct consumption in the reduction of SO 4 toH2S generalized by the following equation

where 2 mol of C is oxidized for every mole of SO4 reducedCarbonloss dueto thereductionof SO4 was estimated basedupon the amountof SO4 reduced in each column This massof SO4 reduced was applied to the reaction stoichiometry given above to estimate the mass of C consumed in thereductionof SO4 Theamount of C leached from thecolumns

TABLE 5 Sul fur Isot opic Data for Column 1 Column 2 andInput Solut ionsa

sampleSO4

(mgL)δ34S

(permil CDT) f b Eb

feed 2 3118 17 10 00column 1 2205 138 07 -353column 2 2140 160 07 -385feed 3 3280 20 10 00column 1 2700 102 08 -330column 2 2060 200 06 -387


a Calculated enrichment factors () show anenrichmentof 34S inthecolumn effluentSO 4 b f ) c sc m) residual fraction and) (δs- δm)Inf ) enrichment factor wherec s ) residual concentration of sulfate c m )

initial concentration of sulfateδm) initial valueof 34Sandδs) residualvalue of 34S

δ34Ssample )m (34S32S)sample - m (34S32S)reference

m (34S32S)reference

(8)

FIGURE8 Representative SEM micrograph of spheroidal Fe andS solids present on reactive media at termination of columnexperiment

TABLE 6 Organic Composit ion of Reactive Mixtures andCalculated Total Ni t rogen Phosphorus and Carbon ContentExpressed as Dry Weight (g) and Dry Weight Percent (wt )

columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413

SO42-+ 2CH2O f H2S + 2HCO3

- (9)




was estimated using measured effluent DOC averaged overthe total volume passed through each column

Thetotalamount of SO4 reduced forcolumn 1 containing leaf mulch sawdust sewage sludge and wood chips was 94g (03 mol) and for column 2 containing leaf mulch andsawdust was 146 g (04 mol) These values were calculatedbased upon the difference between the measured effluentSO4 concentrations considering sulfate reduction and con-servative transport and the modeled effluent SO4 concentra-tion considering conservative transport onlyThe calculatedmass ofSO4 reduced does notconsider SO4 which may havebeen released and subsequently reduced through the dis-solution of gypsum Thus the calculated mass representsthe minimum amount of SO4 reduced

Estimatedvalues of C leachedfrom themixtures were 09g (008 mol) for column 1 and 11 g (009 mol) for column2The amountof C availablein each columnwas determinedfromanalyticalcompositionsof theorganicsubstances(Table6) Column 1 contained 63 g (53 mol) of C and column 2contained 64 g (53 mol) of C On the basis of the stoichio-metric reaction between organic C and SO4 (eq 9) 2 mol of C is oxidized for each mole of SO4 reduced Thereforecombining the amount of C consumed through oxidationreduction reactions and C leaching a minimum of 01 molof C (525 of the available C) was consumed in column 1and a minimum 04 mol of C (75 of the available C) wasconsumed in column 2 The difference in C consumedbetween the columns is directly related to the amount of Coxidized through sulfate reduction reactions (01 mol vs 04

mol in columns 1 and 2 respectively) as the amounts of Cleached from the columns were similar (008 mol in column1 and 009 mol in column 2) Tables 3 and 4 show thecalculated rates of SO4 reduction column 2 (leaf mulch andsawdust) averaged 200 mmol d-1 m-3 higher than column1 (leaf mulch sawdust sewage sludge and wood chips)

Studies by Eger and Wagner (16 ) using bioreactorscontaining organic mixtures (yard waste horse manuresawdust and municipal compost) to enhance sulfate reduc-tion found thatlt5of the C present was utilized before thesystem could no longer support SO4 reduction A decreasein the sulfate reduction rates with time was anticipated asthe easily decomposable fraction of C was utilized Bothcolumns utilized gt5 of the C available and although theobservedsulfate reductionrates decreasedslightly in column1 they remain high throughout the duration of the experi-ment The increase in C utilization and the high sulfatereduction rates as compared to previous studies (16 -18 ) areattributed to longer residence times and a higher inputsolution pH the main differences between the studies

The sulfate reduction rates determined from theseexperiments were used in the design of a full-scale reactivebarrier at the Nickel Rim mine site (4 ) A comparison of thesulfatereductionrates observedin these experimentsto thoseobserved at Nickel Rim is presented by Benner et al (23 ) Atthe field installation the temperature and the groundwatervelocity are more variable anduncertain than in the columnexperiments After accounting for these variations Benneret al (23 ) concluded that the reaction rate observed in the

laboratory was approximately two times greater than in thefield This level of agreement is very good in considerationof the uncertainties associated with the field installation

AcknowledgmentsSpecial thanks to L Hinch for assistance with chemicalanalyses anddatamanagementand toM Duchene forhelpfulcomments Funding for this research was provided by theOntario Ministry of Environment and Energy

Literature Cited(1) Dubrovsky N M Morin K A Cherry J A Smyth D J A

Water Pollut Res J Can 1984 19 55-89(2) Morin K A Cherry J A Dave N K Lim T P Vivyurka A

J J Contam Hydrol 1988 2 305-322(3) BlowesD WPtacekC J United States Patent 53623941994(4) Benner S G Blowes D W Ptacek C J Ground Water Monit

Rem 1997 Fall 99-107(5) WaybrantK RBlowes DW PtacekC JEnvironSci Technol

1998 32 1972-1979(6) Benner S G Blowes D W Gould W D Herbert R B Jr

Ptacek C J Environ Sci Technol 1999 33 2793-2799(7) Benner S G Gould W D Blowes D W Chem Geol 2000

169 435-448(8) Richards F A In Chemical Oceanography Riley J P Skirrow

G Eds Academic Press New York 1965 Vol 1 pp 611-645(9) Reeburgh W S Annu Rev Earth Planet Sci 1983 11 269-

298(10) Herbert R B Jr Benner S G Blowes D W Appl Geochem

2000 15 1331-1343(11) Nordstrom D K Geochim Cosmochim Acta 1977 41 1835-

1841

(12) van Genuchten M Th Parker J L Determining Transport Parameters from Laboratory and Field Tracer Tests Virginia

Agricultural Experiment Station Bulletin 84-3 1984(13) Mills A L Bell P E Herlihy A T In Acid Stress and Aquatic

Microbial Interactions Rao S S Ed CRC Press Boca RatonFL 1989 pp 1-19

(14) Allison J D Brown D S Novo-Gradac K J MINTEQA2 PRODEFA2 A Geochemical Assessment Modelfor Environmental Systems Version 30 Userrsquos Manual US Environmental Pro-tection Agency Athens GA 1990 106 pp

(15) Ball J W Jenne E A Nordstrom D K Am Chem Soc SympSer 1979 No 93 815-836

(16) Eger PWagnerJ Proceedings of Sudbury prime95 -Mining and the Environment CANMET Ottawa ON 1995Vol2 pp515-524

(17) Dvorak D H Hedin R S Edenborn H M McIntire P EBiotechnol Bioeng 1992 40 609-616

(18) McIntire P E Edenborn H M Proceedings of the Mining and Reclamation Conference and Exhibition West Virginia Univer-

sity Morgantown WV 1990 Vol 2 pp 409-416(19) Mills A L Soil Reclamation Processes Klein D Tate R L

Eds Marcel Dekker New York 1985 pp 35-81(20) Spiro B Aizenstat Z Nature (London ) 1977 269 235-237(21) Clark I D Fritz P Environmental Isotopes in Hydrogeology

Lewis Publishers Boca Raton FL 1999(22) Nakai N Jensen M L Geochim Cosmochim Acta 1964 28

1893-1912(23) Benner S G Blowes D W Ptacek C J Mayer K U Appl

Geochem 2002 17 301-320

Received for review March 16 2001 Revised manuscript received November 26 2001 Accepted December 6 2001

ES010751G

TABLE 7 Column Eff luent Data Showing Aqueous Concentrations of Phosphate (o-PO4) Ammonium (NH4) Dissolved OrganicCarbon (DOC) and Volati le Fatty Acids (Organic Acids)a

sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)

column 1 18 30 lt5 36 lt5 lt1 98 nacolumn 1 22 65 na na na na 815 lt1column 2 19 110 lt5 340 lt5 lt1 19 nacolumn 2 22 80 na na na na 080 lt1

a pv ) pore volume na ) not analyzed




and leaf compost from a municipal recycling program Two

different reactive mixtures were used one containing leaf mulch sawdust sewage sludge and wood chips (column 1)and the other containing leaf mulch and sawdust (column2) These mixtures correspond closely to batch mixturespreviously evaluated in bench-scale batch studies (5 )andina full-scale reactive barrier installation (6 7 ) The bacterialsourcewas collectedfromthe anaerobic zone ofa localcreek

The reactive mixtures were soaked in a CaCO3-saturatedsolution containing 1000 mgL SO4 as CaSO4 and5 sodiumlactate The CaCO3-saturated water was made by adding CaCO3 in excess to a known volume of double-deionized water and was bubbled with high-purity CO2(g) for severalhours This solution was left to equilibrate with the atmo-sphere for several days to weeks before use The mixtures were packed into the columns wet to prevent aeration of the

bacterial source The CaSO4 and sodium lactate were addedto help acclimate the bacteria and promote sulfate-reducing conditions Columns were covered with aluminum foil toexclude light and prevent growth of photolithotrophicbacteria Within30-40 days an activepopulation of SRBwasestablished as was evident from the production of H2S gasThe columns were transferred into an anaerobic gloveboxThe anaerobicity of the glovebox was monitored regularly using methylene blue anaerobic strip indicators (GasPak)

The same input solution was used for both columns Theinitial input solution (feed) contained 1000 mgL SO4 asCaSO4 made up using double-deionizedwater saturated withCaCO3 This input solution was used for the first 4 pore

volumes (pv) (feed 0 Table 2) after which simulated mine water was used for the remainder of the study (feeds 1-8Table 2) The chemical composition of the simulated mine water was based upon observations made at a mine-tailingsimpoundment (6 ) Newinputsolutions were prepared every 5-6 weeks This allowed the input solution to be modified

as necessary A conservativetracer test was performed on both columns

by adding NaBr to the feed 1 input solution and measuring effluent Br-concentrationsover timeThe conservativetracertests were used to determine the effective pore volumes of the saturated columns

Influent concentrations of Ni andMn were higher in feed4 as a result of laboratory error The concentration of Fe inthe input solution was decreased in feed 5 after a minorbreakthrough of Fe was observed in both of the columnsThis decrease in Fe was accompanied with decreases inconcentrations of dissolved SO4 and metals After 185 pvNH4Cl was added as a nutrient to determine if the mixture was nitrogen-limited The amount of NH4Cl added to theinfluent corresponded to 7 mgL nitrogen This value was

basedupona CNratio of 251The carboncontentwas baseduponearly DOC measurements in the effluentThe chemicalcomposition and input duration of each input solution aregiven in Table 2 The simulated mine water was stored in anO2-free environment to prevent oxidation of Fe2+ to Fe3+thus keeping the pH relatively constant at 55-65 (Table 2)

A variable speedmultichannel pump(ISMATEC)was usedto deliver the influent to each of the columns to yield anaverage linearvelocity of =10 ma Stainlesssteeltubing wasused between thefeedreservoirand theglovebox (ie whereit was exposed to an oxic environment) and either stainlesssteel or PharMedtubing wasused in anoxicareasThe columneffluentsolutionswere collected in flow-throughsample cells within the glovebox and overflow was directed to waste jugsoutside of the glovebox Flow rates (mLh) were determined

by measuring the overflowSample Collection Column effluent samples were col-

lected within the glovebox to determine pH E h alkalinityand concentrations of dissolved Br Ca Cd Cl Fe H2S KMg Mn Na Ni o-PO4 SO4 Zn dissolved organic carbon(DOC) dissolved inorganic carbon (DIC) δ34S and volatilefatty acids (acetic acid propionic acid butyric acid andformic acid) Measurements of pH E h and alkalinity weredeterminedimmediatelyafter samplingThe pH and E h weremeasured on unfiltered samples in sealed cells to minimizeO2 exposure and a filtered sample was used for measuring alkalinity Electrode calibration and performance werechecked before and after each sample measurement The

FIGURE 1 Schematic diagramof permeable reactive barrier ata mine drainage site

TABLE 1 Composit ion of Reactive Mixtures Used in ColumnExperimentsa

mixturecompositioncolumn1

(drywt)column2

(drywt)

bottom layer (5 cm)crushed pyrite 10 10silica sand 90 90

middle layer (335 cm)composted leaf mulch 195 23wood chips 8sawdust 105 22sewage sludge 10creek sediment 43 44agricultural limestone 2 3silica sand 7 8

top layer (15 cm)silica sand 100 100

a Flow is from bottom to top

















inputsoln

durationcolumn1(pv)


E h(mV)


SO4

(mgL)Fe

(mgL)Zn

(mgL)Ni

(mgL)Mn

(mgL)Na

(mgL)Ca

(mgL)K

(mgL)Mg

(mgL)Br

(mgL)Cl

(mgL)


feed 7 215-249 217-248 66 1 307 278 1353 334 057 336 862 862 220 72 77 176feed 8 249-277 248-275 64 3 180 238 1400 316 144

















inputsolution

porevol(pv)

sulfatereduction


sulfatereduction






inputsolution

porevol(pv)

sulfatereduction


sulfatereduction
































2-(aq) (eq 7)





Fe2+(aq) + 14O2(aq) + H+(aq) f

Fe3+(aq) + 12H2O(l) (4)



CO32-(aq) + 2H+(aq) f H2CO3(aq) (7)

















sampleSO4

(mgL)δ34S

(permil CDT) f b Eb






m (34S32S)reference

(8)



columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413


- (9)






















1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)



















inputsoln

durationcolumn1(pv)


E h(mV)


SO4

(mgL)Fe

(mgL)Zn

(mgL)Ni

(mgL)Mn

(mgL)Na

(mgL)Ca

(mgL)K

(mgL)Mg

(mgL)Br

(mgL)Cl

(mgL)


feed 7 215-249 217-248 66 1 307 278 1353 334 057 336 862 862 220 72 77 176feed 8 249-277 248-275 64 3 180 238 1400 316 144

















inputsolution

porevol(pv)

sulfatereduction


sulfatereduction






inputsolution

porevol(pv)

sulfatereduction


sulfatereduction
































2-(aq) (eq 7)





Fe2+(aq) + 14O2(aq) + H+(aq) f

Fe3+(aq) + 12H2O(l) (4)



CO32-(aq) + 2H+(aq) f H2CO3(aq) (7)

















sampleSO4

(mgL)δ34S

(permil CDT) f b Eb






m (34S32S)reference

(8)



columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413


- (9)






















1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)


















inputsolution

porevol(pv)

sulfatereduction


sulfatereduction






inputsolution

porevol(pv)

sulfatereduction


sulfatereduction
































2-(aq) (eq 7)





Fe2+(aq) + 14O2(aq) + H+(aq) f

Fe3+(aq) + 12H2O(l) (4)



CO32-(aq) + 2H+(aq) f H2CO3(aq) (7)

















sampleSO4

(mgL)δ34S

(permil CDT) f b Eb






m (34S32S)reference

(8)



columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413


- (9)






















1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)































2-(aq) (eq 7)





Fe2+(aq) + 14O2(aq) + H+(aq) f

Fe3+(aq) + 12H2O(l) (4)



CO32-(aq) + 2H+(aq) f H2CO3(aq) (7)

















sampleSO4

(mgL)δ34S

(permil CDT) f b Eb






m (34S32S)reference

(8)



columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413


- (9)






















1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)



















sampleSO4

(mgL)δ34S

(permil CDT) f b Eb






m (34S32S)reference

(8)



columnsewagesludge

leaf mulch

woodchips sawdust

totalN

totalP

totalC

1 39 g 75 g 30 g 40 g 15 g 042 g 63 g21 41 16 22 08 023 341

2 77 g 77 g 14 g 021 g 64 g50 50 09 014 413


- (9)






















1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)
























1841













Geochem 2002 17 301-320


ES010751G


sample pvDOC

(mgL)acetic acid

(mgL)propionic acid

(mgL)butyric acid

(mgL)formic acid

(mgL)PO4

(mgL)NH4

(mgL)




treatment of mine drainage using prb

Documents