SEDIMENTS, SEC 5 • SEDIMENT MANAGEMENT • RESEARCH ARTICLE

Waterjet injection of powdered activated carbonfor sediment remediation

Chris J. Redell & Andrew Curtis Elmore &

Joel G. Burken & Ryan D. Stringer

Received: 15 March 2011 /Accepted: 13 June 2011 /Published online: 1 July 2011# Springer-Verlag 2011

AbstractPurpose In situ contaminated sediment remediationthrough the addition of activated carbon has been provento be an effective remediation technique. An amendmentdelivery system was developed to accurately place andinject a powdered activated carbon slurry. The system wasset up to deliver a series of discrete injections over arectangular grid with the objective to deliver 3% carbon (C)by dry weight to an inundated saturated sediment at amaximum sediment depth of 30 cm.Materials and methods In situ conditions have beenresearched and small bench-scale models have beendeveloped to provide valuable data for future field-scalemodels. Experiments were performed using kaolinite as asurrogate sediment because the color contrast between theclay and the C permitted the delivered C concentration tobe quantified using a spectroradiometer. The experimentsshowed that a set of operational parameters for the injectionsystem could be identified that provided relatively completemixing of the C and clay at the desired depths. Theexperiments were repeated using soil samples contaminatedwith polyaromatic hydrocarbons (PAHs) and polydimethyl-siloxane coated solid-phase microextraction fiber analysesto quantify pore water concentrations.Results and discussion Several different combinations ofpressure, injection duration, and nozzle diameter wereevaluated during the initial phase of the characterizationof powdered activated C penetration in the surrogate

sediment. Iterative approach tactics were conducted thatconcluded specifically placed, short-duration injectionswere necessary to deliver meaningful amounts of C in thetest sediment. Analysis of these injections found that anaverage amended C concentration of 14% was achieved upto 26.7 cm deep in the surrogate sediment by a 9.5-sinjection, whereas a 7.5-s injection at the same depthyielded an average concentration of 9.3%. The reduction inpore water PAHs concentrations through C injection wasachieved in excess of 90% at all sampled locations withinthe injected perimeter.Conclusions Reduction in contaminant pore water concen-trations within the soil/sediment appeared to be lessdependent on the duration of the individual waterjetinjections, and the effective depth of penetration appearedto be greater than that observed during the surrogatesediment experiments. The empirical nature of the waterjetwork and the expected heterogeneity of contaminated soils/sediments suggest that it is appropriate to conduct site-specific bench-scale treatability testing prior to full-scaleremediation using waterjet-delivered activated C.

Keywords Activated carbon . Bioavailability.

Contaminated sediment . Remediation .Waterjet . Placement

1 Introduction

The remediation of contaminated sediment has beenperformed and studied since the 1970s, and technologiesavailable include removal by dredging as well as in situactions. Typical in situ remediation techniques includecapping and the reduction of bioavailability using anamendment. Olsta (2010) defines capping as the placementof a subaqueous cover over contaminated sediments to aid

Responsible editor: Gijs D. Breedveld

C. J. Redell (*) :A. C. Elmore : J. G. Burken : R. D. StringerMissouri University of Science and Technology,266 McNutt Hall,Rolla, MO 65409, USAe-mail: cjr9b7@mst.edu

J Soils Sediments (2011) 11:1115–1124DOI 10.1007/s11368-011-0392-x

in the stabilization, the minimalization of re-suspension,and the reduction of dissolved contaminant transport intosurface water. Wang et al. (1991) performed research on thematerial make up of an engineered cap to adequately reducethe flux of contaminants through the porous media. Hisresearch found that caps containing high organic contentrestricted chemical emergence the most. This concept notonly pertains to capping but when organics are presentwithin the soil or sediment they have a stronger capabilityof adsorbing the contaminant when compared to othermatter found within the soil matrix. Murphy et al. (2006)also studied the use of sorbent materials placed within anengineered cap to retard the flow of contaminants whilereducing their bioavailability. However, Olsta (2010) statedmany concerns are present when in situ capping isimplemented due to the potential for residual contamina-tion, reduced navigation, bio-intrusion, and geotechnicalstability. The U.S. Environmental Protection Agency (EPA1993) has stated that no current technique can remove,contain, or treat contaminated sediments without causingdisturbances to the local sediment.

The use of activated carbon (AC) as means forremediation has been widely used throughout the U.S.A.for the in situ cleanup of contaminated sediments. Zimmermanet al. (2004) stated that the mixing of activated carbon (C)within contaminated sediments reduces both the chemicaland biological availability of hydrophobic organic contami-nants and the uptake of these contaminants by benthicorganisms. That study showed that the total aqueousconcentrations of polychlorinated biphenyls (PCBs) wasreduced by 87% to 92% with the addition of 3.4% dryweight AC when placed into contaminated sediment exca-vated from Hunter’s Point Naval Shipyard in San FranciscoBay, California, U.S.A. Sun and Ghosh (2007) performedsimilar studies and found that 2.6% by weight ACmechanically mixed in contaminated sediment reducedPCB biouptake by 70% to 92%. The researchers also appliedgranular activated carbon (GAC) as a thin mat layer on top ofcontaminated sediment and allowed natural mixing todistribute the GAC into the sediment which resulted inbioaccumulation reductions up to 70%. Cho et al. (2009)found that after 18 months of field exposure, AC reducedaqueous equilibrium PCB concentrations by about 90%.Similarly, Millward et al. (2005) performed a 28-daysediment exposure test with AC and found that it reducedPCB bioaccumulation by 82% in polyacaete and afterexamining the organism 6 months later the bioaccumulationreduction was 87%. Zimmerman et al. (2005) found that0.34% by dry weight, AC reduced PCB concentrations insediment by 44%, 1.7% AC resulted in 84% reduction, and3.4% AC resulted in 87% reduction. The authors also foundthat sediment mixed with 1.7% dry C reduced the aqueousconcentrations of polyaromatic hydrocarbons (PAHs) by 81%

and that by adding additional C up to 3.4% to the mixturedid not change the aqueous concentration of PAHs.

Typically AC is mixed mechanically into the contami-nated sediment with a rotary device attached to a largermechanical operator. Cho et al. (2007) showed that the useof a construction grade rotovator was an effective way ofamending AC to contaminated sediment and that thismechanical mixing device could place AC into thesediment to depths of approximately 30 cm. Cho et al.(2009) in addition to the rotovator used a land-basedcrawler slurry injector system to apply AC to contaminatedsurface sediments. This method worked well to effectivelyreduce aqueous equilibrium PCB concentrations. However,a major drawback associated with these techniques is thatthe mixing of AC to sediment is only implemented at lowtide or other times when the subject area has beendewatered.

Complete mixing of the activated C and the contaminantis highly desirable; field testing has shown that when theAC is applied at varying concentrations within the area ofinterest, the reductions in bioavailability can be impacted,as stated by Cho et al. (2009). Laboratory testing has alsoshown this phenomenon, as higher reductions in contami-nant bioavailability have been displayed when more intensemixing capabilities are attained. The ability to achievecomplete AC and contaminant mixing and provide thehighest possible C to contaminant contact ratio will producethe most significant reductions in contaminant bioavailability.This direct relationship between the AC–contaminant mixingand the reduction in contaminant bioavailability has providedthe major motivation for this study.

Cable et al. (2005) evaluated the potential use ofwaterjets to deliver liquid amendments to contaminatedaquifers and sediments, and the purpose of this paper is toevaluate the use of waterjets to deliver powdered activatedcarbon (PAC) to a saturated sediment that is inundated witha prescribed height water column at the time of injection.Summers (1995) stated that high pressure waterjets rangingfrom 1.0×105 to 1.0×106 psi have been used for large-scalemining and excavation purposes for more than a century.The more recent industrial applications of waterjets includecleaning and machining applications. The cutting capabil-ities of a waterjet are increased for certain applications bythe addition of a granular abrasive such as garnet or silicasand. Waterjets have the potential to be an effectiveremediation technology by substituting amendments (atgreater concentrations) for the traditional abrasives and thenusing the waterjets to place the amendment in thecontaminated media. The work described in this paperevaluates the use of a series of individual waterjet injectionsto deliver 3% by dry weight PAC to a depth of 30 cm in asediment base. The target concentration and depth wasbased on the results of the literature review. The testing was

1116 J Soils Sediments (2011) 11:1115–1124

performed using a clay surrogate sediment so that thedelivered C percentage could be quantified, and confirma-tion testing was performed by quantifying reduction inbioavailability in a PAH-contaminated sediment.

2 Materials and methods

2.1 Injection system

Awaterjet system consists of several components includinga pressure pump, nozzle, nozzle holder, and connectinghoses as shown in Fig. 1. Variable system parametersinclude pump pressure, pump flow rate, nozzle diameter,and nozzle shape. Abrasive amendments are commonlyused to increase the cutting efficiency of waterjet systems.The abrasives may be added pneumatically to a mixingchamber located between the pressure pump and waterjetnozzle, or a pre-mixed abrasive slurry may be pumpedthrough the pressure pump. The waterjet systems that pumpthe pre-mixed slurry are typically less mechanicallycomplex compared to the mixing chamber systems;however, slurry pumping may require more frequent pumpmaintenance due to the accelerated wear on the pump seals.PAC is significantly softer compared to garnet and otherabrasive materials used for waterjet cutting, so a slurrypumping system was selected for this investigation becauseof the potential advantages associated with a simpler

waterjet system. The pressure and flow rates required todeliver amendments to a saturated sediment are significant-ly lower than those required by a waterjet used to quarrystone. Commercial waterjets operating at pressure ranges of1.5×103 to 5.0×102 psi and flow rates of 5.3 to 3.5 lpmwere evaluated for this sediment project but were found tobe ineffective because: (1) the minimum operating pressureand flow rate were too high and resulted in sedimentexcavation; and (2) the systems were unable to pump aslurry which contained AC at meaningful percentages.Airless paint sprayers were found to be an effective meansfor delivering the PAC slurry because of their relatively lowpressure ranges and pump designed to handle relativelyhigh solid content liquids. A Graco Tradeworks 170 electricpaint sprayer with a 466-W piston pump was modified toeffectively deliver the PAC slurry. The sprayer has amaximum operating pressure of 3.0 x 103 psi and amaximum flowrate of 1.2 lpm. Modifications included anelectric footswitch to control injection pressure, anextended stainless steel waterjet lance, a pressure metermounted on the lance, and paint sprayer nozzles modifiedto serve as waterjet nozzles.

2.2 Amendments and surrogate sediment

The experimental remedial amendment used in this researchproject is WPH® powdered AC distributed by CalgonCarbon Corporation. The amendment is a reagglomeratedcoal-based virgin activated C typically used for potablewater and wastewater treatment. The PAC is relatively finebecause a 325-mesh (0.044 mm) sieve only retains 10% ofthe passing material. Research performed by both Sun andGhosh (2007) and Cho et al. (2009) found that reducingAC-particle size proved to be effective at reducingcontaminant biouptake, so the small grain size of the PACis appropriate for remediation purposes. Various weightpercentages of PAC and tap water solutions were evaluatedto indentify which solution would consistently flow throughthe waterjet system, and the highest usable PAC slurryconcentration was 15% by dry weight. Slurries with greaterPAC percentages were too heavy to be handled by thewaterjet pump and/or resulted in clogging of the waterjetnozzle. Slurries with lesser PAC percentages were not usedto reduce the potential for contaminant mitigation throughdilution.

A surrogate sediment sample was used in experimentsdesigned to characterize the distribution of PAC at depth.Kaolinite clay was chosen as the surrogate because it has auniform white color which provides good visual contrastwith the black AC, and the cohesive nature of the clay isconsistent with cohesive sediments that are expected to beencountered in the field. Dry kaolinite weighing 22.9 kgwas hydrated to a uniform mixture with standard tap waterFig. 1 Schematic showing the waterjet injection system

J Soils Sediments (2011) 11:1115–1124 1117

in a rotary concrete mixer. The moisture content of thismixture was measured at 81% using a modified version ofASTM D2216 (ASTM, 2010) and the unit weight of thismaterial was estimated as 14.4 kN m−3. The resistance ofthe sediment to shear was measured using a mini shear vane(Seekonk Model S0-48 and SL-12) with a 5.1-cm-long by2.5-cm-diameter four-bladed torque head. Measurementswere taken at three different depth intervals and the surfacematerial had the weakest resistance to shear with a torquereading of 2.1 Ncm, the 15.2-cm-deep measurement was2.7 Ncm, and the deepest torque measurement was taken at30.5 cm and yielded a value of 3.5 Ncm.

A sandy-clay soil contaminated with PAHs from aformer manufactured gas plant in central Illinois, U.S.A.,was used to characterize contaminant concentration reduc-tion. The moisture content of the soil was 23% and the unitweight of this material was 22.4 kN m−3 which was heavierthan the previously described surrogate. The surface of thesoil had the weakest resistance to shear with a torquereading of 17 Ncm, the 15.2 cm measurement was 28 Ncm,and the 30.5 cm measurement was 110 Ncm. These resultsshow that the granular sediment had greater strengthrelative to the clay surrogate sediment at all measureddepths. It is likely that these differences are due to themixture of granular and cohesive minerals in the sedimentwhile the surrogate sediment was composed essentially ofcohesive clay minerals.

The present research focuses on characterizing thereduction in the bioavailability of contaminants foundwithin sediment. Actual contaminated sediment was notavailable to be tested for this research. The fine-grained soilused was altered to simulate contaminated sediment foundin field conditions. The fraction of organic C found withinthe soils matrix was not explored due to non-contaminatedsoils being unavailable to be used for comparative testing.

2.3 Testing conditions

A two-step process was used to characterize the distributionof PAC at depth in the surrogate sediment. The purpose ofthe first step was to indentify an approximate range ofwaterjet injection times, pressures, and nozzle diametersthat would be used in the second step. Clear polyvinylchloride (PVC) tubes 5.1 cm in diameter and 91.4 cm inlength were filled with the surrogate sediment. The tubeswere mounted vertically and single AC injections weremade at various pressures, durations, and nozzle diameters.The injections were made adjacent to the side of the tube sothat the depth of penetration could be estimated visually.These initial characterizations were considered to beapproximations because of the potential interference be-tween the side of the tube and the penetrating waterjet. Thesecond step involved the use of 30.5 cm diameter tubes

where the potential for interference from edge effects wasless. Each tube used measured 61 cm in vertical length andincluded a filter cloth, a 10-cm layer of aggregate, and aperforated base cap to permit draining and drying of thesurrogate after each series of PAC injections. The kaolinitewas placed in the tube to a depth of 41 cm and allowed tosettle for 16–20 h before each series of injections topromote the closing of void spaces in the mixture. Shortsettling times were well documented to ensure that the soil/sediment upon each injection was consistent for each seriesof injections. The surrogate sediment was covered with5.1 cm of standing water to create a representativeinundated environment prior to each series of waterjetinjections.

Injections were made possible on a rectangular gridpattern using a two-piece acrylic XY grid table as shown inFig. 1. The 61×61-cm top piece consisted of a grid patternwith nine holes in the X direction and six holes in the Ydirection that were spaced 2.5 cm apart. The bottom piecewas mounted horizontally to the top of the tube andincluded four stationary pins. The waterjet lance was fixedperpendicular to the top plate using mounting collars andthe horizontal position of the waterjet relative to the surfaceof the sediment could be changed by aligning different setsof top plate holes with the fixed bottom plate pins. Thispermitted a regular series of injections with a minimumspacing between injections of 2.5 cm. The mounting collarsalso kept the waterjet nozzle at a fixed height of 1.3 cmabove the top of the surrogate sediment.

The injected sediments were dried using electro osmosisafter each series of injections. Four 1.9-cm-diameter copperpipes 107 cm long were placed around the inside perimeterof the PVC tube and connected to an alternating currentvariable voltage supply which provided a constant currentof 90 A. The injected sediment was dried for a minimum of44 h until the contents became semi-solid and could beextruded from the PVC columns.

2.4 Carbon concentration quantification

Cross sections of 3.8-cm increments were sliced horizon-tally from the surrogate sediment after it had been extrudedfrom the PVC tubes. A 20.3×12.7-cm cardboard stencilwas used to delineate the injection area on each slice, andthe material within that area was sampled and homogenizedwith a stand mixer. A 110-g sample of the homogenizedmixture was placed in a soil moisture dish to be dried in anoven at 110°C for at least 12 h. A FieldSpec Pro® modelspectroradiometer from Analytical Spectral Devices, Inc.was used to quantify the C content by measuring lightreflectance versus wavelength measured for each driedsample. Identification standards consisting of 0%, 1%, 3%,5%, 7%, 9%, 11%, 13%, 15%, 25%, and 35% dry weight C

1118 J Soils Sediments (2011) 11:1115–1124

combined with dry kaolinite were mixed and dried in the lab,and the spectroradiomater was used to create standard curvesfrom those samples. The concentrations of samples from theinjected columns were identified through visual comparisonto standard curves as shown on Fig. 2. Readings fallingbetween standard curves were linearly interpolated.

2.5 Analysis of the PAH-contaminated sediment

The quantification of the bioavailable portion of PAHs wasstudied by Jonker et al. (2007) using different analyticalmethods. The use of solid-phase microextraction (SPME)fibers was of one of the methods they studied, and theSPME fiber-predicted bioaccumulation results were typi-cally within a factor of 10 of the measured bioaccumulationin earthworms. Based on these results, SPME fiber analyseswere used to evaluate the effectiveness of waterjet-delivered PAC in reducing the bioavailability in the porewater of contaminated soils/sediments.

Two series of injections were repeated using bulk soilsamples from a former manufactured gas plant whichcontained elevated levels of PAHs. The contaminated soil

was loaded into the 30.5-cm-diameter tubes used for thesurrogate soil testing described above, and two sets ofexperiments were performed. Each set of experimentsincluded a control column and one or more amendedcolumns. The concentrations of PAHs in the pore water ofeach column was quantified using polydimethylsiloxane(PDMS) coated SPME fibers from Polymicro Technologies.The fibers are constructed with a 30-μm PDMS coating ona 1.0-mm fused silica core. SPME fiber holders werefabricated from 0.76-cm-diameter thick-walled stainlesssteel tubing and cut to 40.6 cm lengths. The SPME fiberholders were perforated on four sides with 0.094-cm-diameter holes spaced 1.09 cm apart. The perforationspermitted pore water from the soil column to come incontact with the SPME fibers while reducing the potentialfor direct contact between the fiber and the soil. The fiberswere held in the center of the holders using Tefloncentralizers, and three fiber holders were inserted verticallyin each soil column after completion of the injection series.

The SPME fibers were left in place for 7 days toequilibrate in both the control columns and the columnsinjected with PAC. The fiber holders were then removed

Fig. 2 Standard curves used toquantify C concentration usingthe spectroradiometer

J Soils Sediments (2011) 11:1115–1124 1119

from the columns, the SPME fibers were removed fromthe holders, and the fibers were broken into 2.0-cmsections. The samples were placed into 1-ml shell vialsfilled with 0.5 ml of acetonitrile (ACN). The sampleswere then evaluated using a Waters high-pressure liquidchromatograph (HPLC) that included a Supelco LC-PAH column coupled with fluorescence detection. Themethod used for PAH identification is a modifiedversion of Method 8310 (EPA, 1986) by using SPMEfibers analysis for detection. Peak areas are converted toPAH concentrations using calibration curves that werecreated using varying concentrations of a standardobtained from Sigma–Aldrich mixed with ACN. TheACN concentrations were converted to PAH concentra-tions in the PDMS coating by assuming the PDMS hasextruded all of the available PAHs. The pore waterconcentrations were calculated using a log PDMS towater partitioning coefficient of 3.71 for phenanthrenecalculated for the specific type of SMPE fiber usedthroughout this study.

3 Results

Several different combinations of pressure, injection dura-tion, and nozzle diameter were evaluated during the initialphase of the characterization of PAC penetration in thesurrogate sediment. The visual observation of the penetra-tion depth in the clear PVC tubes indicated that PACinjected through a nozzle of diameter of 0.058 cm had thebest potential to deliver C to a depth of 30 cm in thesurrogate sediment. A standard paint sprayer nozzle wasmodified by grinding off the fan portion so the resultingcylindrical orifice provided a straight waterjet stream. Thepressure pump was operated so that each injection startedwith a 2.5-s burst of 2.5×103 psi at a flow rate of 2.0 lpm

followed by a longer-duration period of injection at a lowerpressure. The initial high pressure burst was an artifact of thepressure pump operation, but it was found to aid in thepenetration of the waterjet stream in the surrogate. The pumpmaintained nozzle pressures ranging between 5.5×102 and7×102 psi at a flow rate of 1.6 lpm after the initial burst.Penetration depths of approximately 30 cm were achieved inthe clear PVC tubes at this pressure combination for twodifferent injection times—7.5 s (2.5-s high pressure plus 5-slower pressure) and 9.5 s (2.5–s high pressure plus 7-s lowerpressure).

The second phase of testing started with the identifica-tion of the appropriate spacing between injections using theXY table. Three preliminary injection series were per-formed using a spacing of 2.5 cm between injections,5.1 cm, and 2.5 cm alternated with 5.1 cm; the 2.5 cmspacing provided the best results. The injections weredistributed in six rows of nine injections each (for a total of54 injections) over a 20.3×12.7-cm rectangular areacentered on the larger PVC tubes.

Three tubes were injected with a series of 7.5-s shots andthree tubes were injected with a series of 9.5-s shots. Thedurations of these injections had a significant impact on thedistribution of PAC within each column as shown on Fig. 3.The 7.5-s injections generated relatively distinct columns ofPAC within the clay matrix of the surrogate sediment whichresulted in a domino-like distribution pattern. The series oflonger injections resulted in more complete mixing of thePAC and the clay matrix over the rectangular target area.This was a highly desirable feature to obtain because themore complete mixing of the PAC aids in a higher contactpercentage between the PAC and contaminant whichultimately increases the reduction in the bioavailability ofthe contaminated sediment.

Figure 4 shows the results of the percentage of Cmeasured in the surrogate sediment columns. The series of

The average C concentration over the rectangular area was 3.5% for this sample collected at 15.2 cm for a series of 7.5 s duration injections.

The average C concentration over the rectangular area was 13% for this sample collected at 30.5 cm for a series of 9.5 s duration injections.

Fig. 3 Typical PAC distribu-tions for each of the injectionseries duration periods

1120 J Soils Sediments (2011) 11:1115–1124

7.5-s injections produced a range of penetration depthsfrom 0 to 26.7 cm, and the target concentration of 3% wasreached at a maximum depth of 19.0 cm. The highestconcentrations of C were found at depths of 7.6 cm andshallower. The 9.5-s injections had penetration depths up to34.3 cm, and the target concentration of 3% was measuredat a maximum depth of 30.5 cm. The highest C concen-trations were measured above 26.6 cm. The results shownin Fig. 4 are consistent with the Fig. 3c distribution patternsbecause the 9.5-s injections resulted in a more consistentdistribution of C horizontally (see Fig. 3) and vertically (seeFig. 4). The goal of delivering 3% C at 30 cm was achievedfor one of the three 9.5-s series, but it was not achieved forany of the shorter-duration injection series. The overallaverage C concentration for all 7.5-s samples from 0 to

26.7 cm was 9.3% and the 9.5-s overall average for thesame depth range was 14%. However, the C concentrationsmeasured for the 9.5-s series had a coefficient of variation(COV) of 0.14 which indicated that there was much lessvariability between the measured concentrations comparedto the 7.5-s COV which was 0.49.

The same waterjet parameters were used to inject PAC incolumns of contaminated soil. One control column and oneinvestigative column were used to evaluate reduction inbioavailability after a series of 7.5-s injections, and onecontrol column and two investigative columns were used tocharacterize the effectiveness of 9.5-s injections. For eachcolumn, one SPME holder was placed in the center of thecolumn, the second was placed 6.4 cm from the centerbetween the first and second row of injections, and the last

Fig. 4 Carbon concentrationsmeasured in homogenized sur-rogate sediment samples

J Soils Sediments (2011) 11:1115–1124 1121

holder was placed 10.2 cm from the center which wasoutside the first row of injections. The PAH phenanthrenewas used to evaluate the effectiveness of the injectionsbecause the HPLC peaks were best-defined for thiscompound. Figure 5 shows the results of both the 7.5-sinjections and the 9.5-s injections. The effective depth ofPAC placement was greater than that observed during thesurrogate sediment investigation, and SPME fiber sampleswere collected below 30 cm after the fibers from the first9.5-s investigative column was analyzed. All of the experi-ments produced concentration reductions over 90% in all ofthe samples. The vertical average concentration reduction

was greater than 98% for all of the locations except for thesamples measured at 10.2 cm from the center of the first9.5-s injection series where the average percent reductionwas over 95%. The vertical variability of C content wassignificantly less compared to the surrogate sedimentresults. The COV calculated for each vertical SPMEfiber location was always less than 0.05. These resultssupport the concept that the heterogeneous nature of thecontaminated soil promoted additional mixing and distri-bution of the PAC. The 10.2-cm results also indicate thatremediation is potentially occurring outside the rectan-gular area. The concentration reductions are greater than

Fig. 5 Percent reduction of bio-availability measured in con-taminated sediment

1122 J Soils Sediments (2011) 11:1115–1124

the 81% reported by Zimmerman et al. (2004) formechanically mixed PAC, and the larger reductions maybe a result of greater C concentrations. Carbon concen-trations exceeding 3% in the contaminated soil columnswould be consistent with the general results of thesurrogate sediment experiments.

4 Discussion and conclusions

The use of waterjets to minimize the impact ofcontaminated sediments has been shown to be a viablealternative to mechanical mixing at the bench scale for insitu remedial action. Specifically, this research has shownthat discrete waterjet injections over a rectangular gridpattern can place PAC in saturated sediment at prescribedconcentrations and depths. Experiments with a sedimentsurrogate showed that the degree of PAC mixing is afunction of the time of each discrete injection. If theinjection time is too short, then discrete domino-likepatterns of higher concentrations are observed at depthwhere longer injections resulted in more complete mixingof the soil matrix and the PAC. Contaminated soilexperiment results indicated that the heterogeneity ofthe soil and/or other factors improved the penetration andmixing of the waterjet-delivered PAC. The differences inPAC placement with different injection times character-ized in the surrogate sediment were not observed in thecontaminated soil experiments. However, the concentra-tion of amended C significantly exceeded target concen-trations at most depths in the surrogate sediment, and thepore water contaminant concentration reduction datafrom the contaminated soil experiments supported theassumption that more C was being delivered thandesired. The objective of our research was to inject thehighest C concentration slurry to reduce the potential forcontaminant attenuation through dilution. Initially, theamount of C that was added to the remedial slurry toreach meaningful concentrations within the sedimentthrough waterjet injection was unknown but through thisresearch, it was proven that too high of a C concentrationwas obtained and a diluted slurry could technically befeasible to achieve the same meaningful results. Thesurrogate sediment experiments showed that the longerinjection times resulted in liquefaction of the sediment,and this implies that the delivered PAC concentration canbe managed by modifying the C concentration in theslurry before it is injected in to the sediment.

The goal of this work was to measure a dry weightconcentration of 3% PAC in the discrete depth interval from26.7 to 30.5 cm for comparison with the mechanical mixingresults presented in the literature. That goal was accom-plished in one test: however, the average concentration of

amended PAC in the entire 0 to 30.5-cm interval wasapproximately five times higher than the concentrationmeasured in the bottom 3.8 cm. It is difficult to compareour results to field-scale mechanical mixing results pre-sented in the literature because the literature results arebased on a single sampling interval of 0 to 30.5 cm. Thedefinition of the depth interval(s) used to calculate thetarget PAC concentration had significant implications interms of the mass and corresponding cost of the C amendedin the sediment.

The potential for the waterjet system to encounterlogistical problems throughout field applications is highlypossible. Fresh water will always be needed to ensureunnecessary clogging within the internal orifices of theinjection system. The lack of fresh water present to mixwith the PAC could potentially cause problematic mechan-ical failures. The development of a customizable flotationdevice will need to be achieved for the remediation ofsediments in an inundated environment. The additionalstorage of the PAC slurry will place added weight on thefloatation device and cause valuable floor space to besacrificed to accommodate a series of large holding tanks.These drawbacks provide only minor disadvantages andcan easily be overcome through innovative constructionand design of a customizable floatation device. The use ofoff-the-shelf components with minor modifications wasrelatively successful for this bench-scale study. Similarequipment with greater capacity is commercially available,and pilot-scale waterjet injection systems could be readilyfabricated. However, the empirical results presented in thispaper suggest that pilot-scale applications should bepreceded by site-specific bench-scale treatability studies tocharacterize the impact of site soils on the penetration anddispersion of the waterjet.

Acknowledgments This project was funded by the US NationalInstitute of Environmental Health Sciences (NIEHS) through theSuperfund Research Program SRP (Projects 5R01ES016158 and3R01ES016158).

References

ASTM Standard D2216-10 (2010) Standard test methods forlaboratory determination of water (moisture) content of soil androck by mass. ASTM International, West Conshohocken, PA,2010. doi:10.1520/D2216-10, www.astm.org

Cable JW, Duchene ML, Elmore AC, Hill D, Summers DA, Dilly R(2005) Initial investigation on the use of waterjets to placeamendments in the subsurface. Remediation J 16:43–54

Cho YM, Smithenry DW, Ghosh U, Kennedy AJ, Millward RN,Bridges TS, Luthy RG (2007) Field methods for amendingmarine sediment with activated carbon and assessing treatmenteffectiveness. Marine Environ Res 64:541–555

Cho YM, Ghosh U, Kennedy AJ, Grossman A, Ray G, TomaszewskiJE, Smithenry DW, Bridges TS, Luthy RG (2009) Field

J Soils Sediments (2011) 11:1115–1124 1123

application of activated carbon amendment for in situ stabiliza-tion of polychlorinated biphenyls in marine sediment. EnvironSci Technol 43:3815–3823

U.S. Environmental Protection Agency (1986) Method 8310 polynu-clear aromatic hydrocarbons. EPA 8310, Washington, DC

U.S. Environmental Protection Agency (1993) Selecting remedia-tion techniques for contaminated soil. EPA 823/B93-C81:Washington, DC

Jonker MTO, Van Der Heijden SA, Kreitinger JP, Hawthorne SB(2007) Predicting PAH bioaccumulation and toxicity in earth-worms exposed to manufactured gas plant soils with solid-phasemicroextraction. Environ Sci Technol 41:7472–7478

Millward RN, Bridges TS, Ghosh U, Zimmerman JR, Luthy RG(2005) Addition of activated carbon to sediments to reduce PCBbioaccumulation by a polychaete (Neanthes arenaceodentata)and an amphipod (Leptocheirus plumulosus). Environ SciTechnol 39:2880–2887

Murphy P, Marquette A, Reible D, Lowry GV (2006) Predicting theperformance of activated carbon-, coke-, and soil-amended thinlayer sediment caps. J Environ Eng 132:787–794

Olsta J (2010) In situ capping of contaminated sediments withorganophilic clay. Proceedings of Twelfth Triannual InternationalConference. ASCE, Jacksonville

Summers DA (1995) Waterjetting technology. Alden press, Oxford, UKSun X, Ghosh U (2007) PCB bioavailibility control in lumbriculus

variegatus through different modes of activated carbon additionsto sediments. Environ Sci Technol 41:4774–4780

Wang XQ, Thibodeaux LJ, Valsaraj KT, Reible DD (1991) Efficiencyof capping contaminated bed sediments in situ. 1. Laboratory-scale experiments on diffusion-adsorption in the capping layer.Environ Sci Technol 25:1578–1584

Zimmerman JR, Ghosh U, Millward RN, Bridges TS, Luthy RG(2004) Addition of carbon sorbents to reduce PCB and PAHbioavailability in marine sediments: physicohemical tests. EvironSci Technol 38:5458–5464

Zimmerman JR, Werner D, Ghosh U, Millward RN, Bridges TS,Luthy RG (2005) Effects of dose and particle size on activatedcarbon treatment to sequester polychlorinated biphenyls andpolycyclic aromatic hydrocarbons in marine sediments. EnvironToxicol Chem 24:1594–1601

1124 J Soils Sediments (2011) 11:1115–1124