ment 364 (2006) 45–54www.elsevier.com/locate/scitotenv

Science of the Total Environ

Electrodialytic remediation of CCA-treated waste wood in a 2 m3

pilot plant

Iben V. Christensen a,⁎, Anne J. Pedersen a, Lisbeth M. Ottosen a, Alexandra B. Ribeiro b

a Department of Civil Engineering, building 204, The Technical university of Denmark, DK-2800 Lyngby, Denmarkb DCEA- Faculdade de Ciências e Tecnologia, Universidade Nova de Lisboa, Quinta da Torre, 2829-516 Caparica, Portugal

Received 8 July 2005; received in revised form 14 October 2005; accepted 3 November 2005

Abstract

Waste wood that has been treated with chromated-copper-arsenate (CCA) poses a potential environmental problem due to thecontent of copper, chromium and arsenic.

A pilot plant for electrodialytic remediation of up to 2 m3 wood has been designed and tested and the results are presented here.Several process parameters were investigated, and it was found that the use of collecting units and soaking of the wood prior to theelectrodialytic process had a positive influence on the remediation process. There was a tendency towards higher removal of CCAfrom wood chips <2 cm, compared to larger wood size fractions. The best remediation efficiency was obtained in an experimentwith an electrode distance of 60 cm, and 100 kg wood chips. In this experiment 87% copper, 81% chromium and >95% arsenicwere removed. One other experiment was also analysed for arsenic. In this experiment the distance between the working electrodeswas 1.5 m and here 95% As was removed. The results showed that arsenic may be the easiest removable of the copper, chromiumand arsenic investigated here. This is very encouraging since arsenic is the CCA components of most environmental concern.© 2005 Elsevier B.V. All rights reserved.

Keywords: CCA-treated waste wood; Electrodialytic remediation; Oxalic acid; Phosphoric acid; Pilot plant

1. Introduction

1.1. CCA-treated waste wood

The service life of wood treated with chromated-copper-arsenate (CCA) may be 20–50 years or evenmore (Cooper et al., 2001), implying strong fixation ofthe CCA and subsequently limited leaching from thewood while in service. Limited leaching also means thatthe CCA is still present in the wood when it is removedfrom service.

⁎ Corresponding author. Tel.: +45 45 25 23 97; fax: +45 45 88 59 35.E-mail address: ic@byg.dtu.dk (I.V. Christensen).

0048-9697/$ - see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.scitotenv.2005.11.018

Cooper et al. (2000) found CCA retentions wellabove the toxic threshold for decay in poles removedfrom service after 1–50 years. The average life span forresidential lumber is often much shorter, in partbecause aesthetic reasons also play a role in determin-ing the service life. In a survey conducted by McQueenand Stevens (1998) disposal of residential decks madewith CCA-treated lumber was reported and approxi-mately 30% of the decks were disposed for aestheticreasons. The average service life for these decks was 7years.

Limited leaching of CCA and disposal for aestheticreasons means that the wood still contains high amountsof CCA when becoming waste wood.

Fig. 2. Schematic presentation of an electrodialytic cell. CompartmentI and III is the anode and cathode compartment respectively, the woodchips are placed in compartment II. AN: Anion exchange membrane,CAT: Cation exchange membrane.

46 I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

The amount of treated waste wood is increasing andit is estimated that in Denmark alone the amount oftreated waste wood to be disposed of will increase from17,000 tons in 1992 to 100,000 tons in 2010 (Affald 21,1999). Similar trends are seen in the rest of the world. InCanada the volume of CCA-treated wood to be removedfrom service is expected to grow substantially in thenext 10 years, and stabilize close to 2 million cubicmetres per year (Cooper et al., 2001). In Germany andFrance the total amount of wood waste is around 3–4million tons per year, of which more than 2 million tonsare characterised as dangerous (Helsen and Van denBulck, 1998).

The large amount of wood with a high content ofcopper (Cu), chromium (Cr) and arsenic (As) make anenvironmental safe handling method for the waste wooddesirable. Incineration of CCA-treated wood is prohib-ited in Denmark. This is primarily due to arsenicreleases to the atmosphere, and that the heavy metalswill concentrate in the combustion residues (Affald 21,1999).

A life cycle for CCA-treated wood is presented inFig. 1. The CCA-treated waste wood is chipped and Cu,Cr and As are removed using electrodialytic remedia-tion. Afterwards the wood may be burned, therebyutilizing the energy resource of the wood. To use thewood as bio fuel, the removal of Cu, Cr and As has to be100%, but if As is removed and the concentration of Crand Cu are reduced, then the wood could be burned in amunicipal solid waste incinerator. The removed Cu, Crand As from electrodialytic remediation may be used forproduction of new CCA. However, since the further useof CCA in the life cycle is being restricted in most of theworld, the metals could be reused in other parts of theindustry instead, or stabilized for safe disposal.

Fig. 1. Proposed life cycle for CCA-treated wood.

1.2. Electrodialytic remediation (EDR)

Electrodialytic remediation is a method developed atthe Technical University of Denmark (DTU). Initiallythe method was developed for removing heavy metalsfrom polluted soil (Hansen et al., 1997; Ottosen et al.,1997). The method uses a direct electric current as acleaning agent and combines it with the use of ionexchange membranes to separate the electrolytes fromthe soil. High removal of several heavy metals havebeen obtained, and subsequently the method has beentested on a wider range of materials including fly ash,sludge, harbour sediments and impregnated waste wood(Pedersen, 2003; Jakobsen et al., 2004; Nystroem et al.,2005; Ribeiro et al., 2000; Ottosen et al., 2003).

In laboratory scale experiments CCA-treated wastewood has been remediated both as sawdust and woodchips with encouraging results. In sawdust approxi-mately 95% Cu, 90% Cr and more than 96% As wasremoved (Ribeiro et al., 2000) and in remediation ofwood chips, removal of more than 90% Cu and about85% of both Cr and As has been obtained (Kristensen etal., 2001).

Based on the results obtained in laboratory scaleexperiments, upscaling of the electrodialytic remedia-tion has now been tested. In the experiments presentedhere, between 94 and 469 kg wood chips were used ineach experiment, compared to 50–70 g used inlaboratory scale experiments.

A schematic presentation of an electrodialytic cell isshown in Fig. 2. The laboratory cell consists of threecompartments: an anode compartment (I), a cathodecompartment (III) and a middle compartment (II)containing the wood chips (Fig. 2). The catholyte isseparated from the middle compartment by a cationexchange membrane, a membrane that only allowspositive ions—cations—to pass. The anolyte is sepa-rated from the middle compartment by an anionexchange membrane, a membrane that only allowsnegative ions—anions—to pass.

When an electric potential is applied to the electro-des, the current in the cell is carried by ions in thesolutions. Accordingly cationic species migrates

47I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

towards the cathode and anionic species migratestowards the anode. With ion exchange membranesplaced as described above, no ions can be transported bythe current from the electrode compartments into themiddle compartment, while ions can be transported fromthe middle compartment into the electrode compart-ments. In this system the current is thus prevented fromcarrying highly mobile ions from one electrodecompartment through the middle compartment into theother electrode compartment. Furthermore, competitionbetween such highly mobile ions from the electrodecompartments and the ions in the middle compartment isavoided. The electrodialytic remediation method isdescribed in further details in Ottosen et al. (1997) andHansen et al. (1999).

2. Experimental section

2.1. Wood

The wood used in the present experiments wassupplied by RGS90, the largest recycling company inDenmark. From a large pile of demolition wood wasteapproximately 10 m3 of impregnated wood wascollected. By visual inspection it can be difficult to seeif old (weathered) wood is impregnated or not, thereforeChromazurol S, a colour reagent turning blue in contactwith Cu was used to identify the impregnated wood.After sorting, the wood was chipped and divided intothree size fractions: fine (F) <2 cm, medium (M): 2–4cm and large (L) >4 cm by sieving. The particular sizefractions were chosen based on practical conditionssince 2 and 4 cm were the sift sizes available. The woodchips found in the L fraction were 4–10 cm by visualinspection.

The wood came from the same pile of waste woodand as expected there was no statistical difference of thecontent of Cu, Cr and As in the three fractions (Table 1).The relatively large confidence intervals reflect the largevariation in the waste wood.

Table 1Concentration ±95% CL of Cu, Cr and As in waste wood

Size fraction [Cu] mg/kgwood

[Cr] mg/kgwood

[As] mg/kgwood

<2 cm (F) 1251±68 1256±83 735±1182–4 cm (M) 1349±104 1414±143 807±170>4 cm (L) 1226±157 1308±223 1031±327Mean value 1279±66 1334±91 837±114

Mean values are based on a total of 179 samples for Cu and Cr and 95samples for As.

2.2. Analytical methods

Flame Atomic Absorption Spectroscopy (AAS) wasused for measuring concentrations of Cu, Cr and As.Wood samples were analysed after microwave assistedacid digestion (0.4 g wood to 10 ml nitric acid) anddilution to a final volume of 25.0 ml. Aqueous samples(electrolytes etc.) were preserved with nitric acid (1 :4)prior to AAS.

The concentration of As had to be measured at anexternal laboratory and therefore only selected sampleswere measured for As in the present investigation.

2.3. Pilot plant

A pilot plant was designed for remediation of up to 2m3 treated wood waste. Fig. 3 shows a schematicpresentation of the pilot plant. The plant consisted of aPVC box, approximately 3 m long, 1 m wide and 1 mhigh. Inside the box, there were ribs for every 35 cm andat these ribs it was possible to place an electrode unit(equal to electrode compartments in Fig. 2) or acollecting unit. The use of the two different units isaddressed in the following.

When the full capacity of the plant was in use, theelectrode units were placed in each end of the box, witha distance of 3 m between the electrodes. The electrodeunits could also be placed at the ribs, thereby it waspossible to vary the distance between the electrodesfrom 30 cm to 3 m. The pilot plant was in principle anupscale of the laboratory cell, but some adjustmentswere made in the upscaling. Due to the larger size, thedistance between the electrode units (where the ions arecollected in laboratory scale experiments) may become alimiting factor on the remediation time or efficiency. Tocompensate for this, collecting units were introduced.Collecting units were placed between the electrode unitsand had a cation exchange membrane on one side and ananion exchange membrane on the other side. The setupof the membranes made it possible to trap the ions insidethese units. By introducing the collecting units, thedistance the ions have to travel before being concen-trated could be reduced to 30 cm. The amount of woodto be treated could vary between approximately 0.3 and2 m3. Fig. 4 shows the pilot plant in use.

2.4. Remediation experiments

A total of seven pilot plant remediation experimentswere made, and in Table 2 the experimental conditionsare outlined. The different experimental conditions ofthe seven experiments made it possible to evaluate the

+-

30 cm

1 m

Electrode unit

Collecting unit

Anion exchange membrane

Cation exchange membrane

Middle compartment (wood chips)

3 m

Fig. 3. The pilot plant in principle, as seen from above. Not to scale. The wood chips are placed in the middle compartments between collecting units,where Cu, Cr and As from the wood is collected. The end units contain the electrodes and are named electrode units.

48 I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

influence of different process parameters on theremediation.

In all experiments except exp. 4, the wood wassoaked before remediation. The wood was soaked inlarge open containers. The L:S ratio (soaking solution :wood) was 4 :1 for every soaking. After soaking thewood was placed in the pilot plant and covered with tapwater and the current was applied.

In the electrode units and collecting units approxi-mately 20 l 0.01 M NaNO3 was circulating. During theremediation, pH was adjusted with nitric acid tomaintain pH 2 in the units to prevent precipitations.

Fig. 4. The pilot plant in use. Wood chips in yellow plastic nets areplaced between collecting units, where the Cu, Cr and As from thewood is collected. The end units contain the electrodes, and are namedelectrode units.

After remediation the distribution of Cu and Cr wasmeasured in all seven experiments. The content of Cuand Cr was measured in the soaking solutions, the units,in the middle compartments and in the wood. Thedistribution of As was only measured in two experi-ments (exp. 3 and exp. 6) because the analysis had to bemade at an external laboratory and at high costs.

In all the experiments the aim was to keep the currentstrength as high as possible. In exp. 1 and 2 themaximum current strength that was possible to obtainwas 2 A. For exp. 3–7 the power supply was replacedand the maximum current strength increased to 5 A. InFig. 5 the current strength as a function of time is shownfor all seven experiments. The current strength wasinitially kept constant, but during the remediation thecurrent strength was regulated in all experiments exceptexperiment 2 where it was maintained at 2 A. The reasonfor adjusting the current strength was that the resistancein the pilot plant increased with time and therefore thecurrent strength had to be reduced. In exp. 7 the currentstrength increased during the remediation. In thisexperiment it was decided to start the experiment at amoderate current strength of 1 A since it was expectedthat the large electrode distance would result in a highresistance. This was not the case and during theremediation the current strength was increased to 1.5 A.

3. Results

The influence of collecting units, soaking, wood sizefractions and upscaling on the remediation has beeninvestigated.

Table 2Experimental conditions

Exp. 1 Exp. 2 Exp. 3 Exp. 4 Exp. 5 Exp.6 Exp. 7

Electrode distance(cm)

60 60 60 60 90 150 270

Wood (kg) 94 106 97 99 178 248 469Duration (days) 11 11 21 15 15 21 21Wood fractions M M M M F/M/L M FCollecting units 1 0 1 1 2 4 8Current (A) 1.4–2 2 2–5 0.2–3 2–3 2–3 1–1.5Voltage (V) 24–25 14–18 30–58 14–60 23–29 40–63 23–39Additive Water Water Water 5% oxalic

acidWater Water Water

Soaking solution/duration(hours)

5% oxalicacid/48

5% oxalicacid/48

0.5 M H3PO4/18 5%oxalic acid/24

– 0.5 M H3PO4/18 5%oxalic acid/24

0.5 M H3PO4/18 5%oxalic acid/24

0.5 M H3PO4/24 5%oxalic acid/24

In exp. 4 oxalic acid was used as electrolyte solutions. In all other experiments 0.01MNaNO3 was used. Wood fractions: fine (F) <2 cm, medium (M)2–4 cm, large (L) >4 cm.

49I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

After remediation the content of Cu, Cr and As wasreduced in the wood in all seven experiments (Fig. 6). Inthe following sections the results are discussed inrelation to different process parameters.

3.1. Collecting units

In experiments 1 and 2 the setup was similar exceptfor the use of a collecting unit in exp. 1. The purposewas to illustrate the influence of the collecting unit onthe remediation by comparing the two experiments.

When comparing the final concentrations in exp. 1and exp. 2, it seems that the remediation of Cu wasimproved by the use of a collecting unit to shorten theway the ions had to travel before being collected (Fig.6). On the other hand, no effect was seen for Cr. Tofurther investigate the remediation, the distribution ofremoved Cu and Cr after remediation in the twoexperiments is compared in Fig. 7.

0

1

2

3

4

5

6

0 5 10 15time (days)

I (A

)

Fig. 5. Current strength (ampere) as a functio

The removal of Cr and the amount of Cr present inthe units after remediation was similar in the twoexperiments (Fig. 7a). However, more Cr was found inthe middle compartment in exp. 2 compared to exp. 1.This difference is mostly balanced by the Cr removedduring soaking in the two experiments. Cr removedduring soaking is not related to the use of collectingunits and the seemingly higher removal by soaking inexp. 1 compared to exp. 2 may be because the soakingliquids from exp. 1 were reused in exp. 2.

More Cu was removed in exp. 1 compared to exp. 2(Fig. 7b). This is not only because of the use of acollecting unit. The reuse of the soaking liquids fromexp. 1 in exp. 2 influenced the removal of Cu more thanfor Cr. Four times more Cu was removed during soakingin exp. 1 than soaking in exp. 2.

The possibility of reusing soaking solutions hadpreviously been tested in laboratory scale. The resultsindicated that soaking solutions may be used up to 4

20 25

Exp. 1

Exp. 2

Exp. 3

Exp. 4

Exp. 5

Exp. 6

Exp. 7

n of time (days) for experiments 1–7.

0

200

400

600

800

1000

1200

1400

1600

1 2 3 4 5 6 7

experiment

pp

m

Cu

Cr

As

Fig. 6. The concentration of Cu, Cr and As in the wood after remediation. Concentration of As is only measured in exp. 3 and exp. 6. Horizontal linesindicate initial concentration.

50 I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

times without any influence on the remediationefficiency. Therefore no influence by the reuse of thesoaking solution was anticipated. However, the apparentdecreased removal of Cu and Cr during soaking, whenthe soaking solution were reused contradicts thelaboratory experiments. Further investigation is neededto verify if the reuse of soaking solutions has aninfluence on the remediation efficiency. For the present

Cr

0

10000

20000

30000

40000

50000

60000

70000

80000

soaking units mi

soaking units mi

Cu

0

5000

10000

15000

20000

25000

30000

35000

mg

Cu

a

b

Fig. 7. (a) Distribution of removed Cr in an experiment with collecting unit (excovers both collecting unit and electrode units. (b) Distribution of removed Cunit (exp. 2). The term “units” in the figure covers both collecting unit and

series of experiments the soaking solutions was usedand reused totally three times in all experiments exceptexp. 1 and exp. 3.

3.2. Soaking

In exp. 4 the wood was not soaked beforeremediation, but placed directly in the pilot plant.

ddle comp. total

ddle comp. total

with collecting unit

No collecting unit

with collecting unit

No collecting unit

p. 1) and without collecting unit (exp. 2). The term “units” in the figureu in an experiment with collecting unit (exp. 1) and without collectingelectrode units.

51I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

Oxalic acid was used as an additive in the middlecompartments and in the units. The experiment had to bestopped after 15 days because of technical problems.The resistance increased dramatically, forcing thecurrent strength to a value below 0.2 A (Fig. 5). It waslater discovered that the anode was broken. During theexperiment the current strength had to be decreased dueto extreme gas evolution in the cathode unit, thatresulted in malfunctioning of the circulating pump. Asseen in Fig. 6, the remediation of Cu was greatlyreduced in this experiment compared to exp. 1, wherethe oxalic acid was used in the soaking procedure, andnot during the electrodialytic remediation process. In theremaining experiments, soaking was used.

Two different soaking solutions were investigatedhere. In the first two experiments oxalic acid was used,and in the subsequent experiments dual soaking inphosphoric acid and oxalic acid was used. The change insoaking solution was based on a series of laboratoryexperiments with different acids and combinations ofacids. Phosphoric acid, followed by oxalic acids gavethe best results in laboratory scale. After remediation,only 8% Cu and 18% Cr remained in the wood in thelaboratory experiment. By comparing exp. 1 and exp. 3it seems that the change in soaking solution had aprofound impact on the remediation (Fig. 5). Theremediation increased for both Cu and Cr by the dualsoaking and higher removal was seen during soaking inexp. 3 compared to exp. 1 (Fig. 8).

The main purpose for soaking the wood prior toelectrodialytic remediation was to remove the mostavailable Cu, Cr and As first and use the current toremove the less available fractions. It was possible toremove more than 50% of Cu and Cr by dual soaking inexp. 3 without the use of an electric field (Fig. 8).

0%

20%

40%

60%

80%

100%

Oxalic Phos/Oxalic Oxal

Cu

Fig. 8. Distribution of Cu and Cr after remediation where the wood was initialacid and oxalic acid (exp. 3).

3.3. Wood size fraction

In exp. 5 all three wood size fractions were used. Thedifferent sizes were evenly put in the pilot plant. Afterremediation, 24 wood samples from each of the threefractions were analysed for Cu and Cr. The finalconcentration of both metals was significantly lower inthe fine fraction than in the medium and large fraction(Fig. 9), indicating that it was easier to remediate woodchips with a size <2 cm. The concentration of Cu wasreduced to 127±12 ppm in the fine fraction, 368±67ppm in the medium fraction and 515±176 ppm in thelarge fraction. Overlap of the 95% CL in the mediumand large fraction indicated that the difference betweenthese two fractions was not statistically significant.However, the relatively high 95% CL in the samefractions indicated a high variation. For Cr the variationwas high in the large fraction where the concentrationafter remediation was reduced to 771±294 ppm. In themedium fraction the concentration of Cr was 629±95ppm and in the fine fraction 462±35 ppm. As for Cu, theconcentration of Cr was significantly lower in the finefraction, compared to the medium fraction, indicating aneasier remediation of the fine fraction. The highvariation in the large fraction did not make thedifference between the Cr content of this fraction andeach of the other two fractions statistically significant.

3.4. Upscaling

Evaluating the influence of upscaling on theremediation exp. 3 and exp. 6 are compared. The twoexperiments had the same duration, and the same woodsize fraction and soaking solution were used. In bothexperiments the max. distance the ions had to travel

Phos/Oxalicic

Cr

soaking

Middle comp.

coll. Unit

anolyte

catholyte

wood

ly soaked in oxalic acid soaking (exp. 1) and dual soaked in phosphoric

Cu

0

100

200

300

400

500

600

700

800

pp

m

F

M

L

Cr

0

200

400

600

800

1000

1200

pp

m

F

M

L

Fig. 9. Concentration of Cu and Cr in wood chips in exp. 5 (three different wood size fractions) after remediation (±95 CL). Wood fractions: fine (F)<2 cm, medium (M) 2–4 cm, large (L) >4 cm.

52 I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

before being captured was 30 cm, resulting in onecollecting unit in exp. 3 and four in exp. 6. The electrodedistance was 60 cm in exp. 3 and 150 cm in exp. 6.Increasing the electrode distance by factor 2.5 resultedin a decrease in the remediation of approximately thesame magnitude for both copper and chromium (Fig. 5).The same was not true for the remediation of As. In bothexperiments more than 95% of the As was removed.

The limiting factor when increasing the distancebetween the electrodes was the current density. Whenthe number of collecting units increased from 1 to 4 U,the number of membranes the current had to pass alsoincreased from 4 to 10. In the plant the membranesaccount for the majority of the electric resistance. Tocompensate for the increased resistance in exp. 6 ahigher voltage drop was needed to maintain the samecurrent density. This was not possible due to thelimitations of the power supply used.

In exp. 7 the distance between the electrodes wasfurther increased, resulting in a total of 18 membranes.In exp. 6 only the medium fraction was used and in exp.7 only the fine fraction was used. In Fig. 5 it is seen thatthe remediation in the two experiments seems unaffect-ed by the electrode distance, 150 cm in exp. 6 comparedto (270) cm in exp. 7. The reason for this may have more

to do with the difference in wood size fraction asdiscussed previously than on the difference in electrodedistance and current density.

4. Discussion

A series of seven electrodialytic remediation experi-ments has been presented here and different processparameters have been investigated. Due to the limitedamount of experiments it is important to emphasize thatthe results and interpretations may be viewed astendencies rather than definite conclusions on the matterof remediation in pilot scale.

The use of collecting units to shorten the way the ionshave to travel before being captured proved useful. In allbut one experiment the wood was soaked before theelectrodialytic remediation process. When soaking wasnot used, the remediation efficiency decreased and inaddition to that, major technical problems wereencountered. The purpose of soaking was to removethe most available Cu, Cr and As first and then use theelectric current to remove the less available fractions.The soaking solutions contained acid and/or complexingagents in the form of ions that will move in the electricfield. If the oxalic acid or phosphoric acid was used

53I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

directly in the pilot plant a large proportion of thecurrent would be wasted on removing these ions fromthe middle compartment instead of Cu, Cr and As. Byintroducing soaking, the wood and CCA came intocontact with the acids prior to remediation, and theconcentration of the acid (ions) in the electrodialyticremediation is reduced. Dual soaking in phosphoric acidfollowed by soaking in oxalic acid proved to be mosteffective.

The reuse of the soaking solutions was initiallypresumed to have little influence on the remediation, butwhen comparing exp. 1 and 2 it seems that the removalof especially Cu during soaking was reduced. When thesoaking solution was reused, it is possible that it wassaturated from the previous soaking and that Cuprecipitated as CuOx during the soaking in exp. 2.Thereby the amount of Cu removed during soaking inexp. 2 may be underestimated in Fig. 7b. Even if this isthe case, the measured concentration in the wood afterremediation was still higher in exp. 2 than exp. 1,indicating that the remediation was more successful inexp. 1. It may also be interpreted as an indication ofprecipitation of CuOx in or on the wood chips, andmaybe that the precipitate dissolves to a certain extentduring the remediation when the wood chips are coveredwith tap water.

Three different wood size fractions were remediatedand the results indicate that the remediation efficiencyincreased with decreasing wood size. The remediationof wood chips <2 cm was significantly better than theremediation of the larger wood sizes. Further investiga-tions are needed to locate the reason for this, butinsufficient soaking of the inner parts of the larger woodpieces may be part of it. If the wood is not soaked all theway through, then it is not expected that the current willpass here, since there will be highest electric resistancein this part of the chip. If the soaking solution does notreach the inner part, then the CCAwill not be removed.It is possible that using vacuum soaking of the woodcould insure total soaking prior to the remediation.

The results obtained in the upscaling processindicated that the remediation efficiency decreasedwith increasing electrode distance.

The main problem in estimating the influence ofupscaling the electrode distance is that the currentdensity could not be maintained in the upscaledexperiments.

The use of collecting units was expected to insurethat the remediation time was unaffected by theelectrode distance, since the distance the ions have totravel before being collected would be the same. Thishypothesis has to be verified with another power supply

with a higher voltage range, that can keep the currentdensity constant despite the increase in electric resis-tance due to the increased number of membranes.

The highest removal of both Cu, Cr and As wasobtained in exp. 3, where 87% Cu, 81% Cr and morethan 95% As were removed during remediation. In thisexperiment the electrode distance was 60 cm and ca. 100kg wood chips of medium size fraction (2–4 cm) wereused. When the electrode distance was increased, theremoval of Cu and Cr decreased. In exp. 6 where theelectrode distance was 150 cm and 250 kg wood wereremediated, 58% Cu and 54% Cr were removed. On theother hand the remediation of As did not decrease withincreasing electrode distance. In the same experiment96% As was removed. This is a very encouraging resultsince As is the CCA component of most concern and themain reason why incineration of CCA-treated wood isprohibited in Denmark. If the remediated wood is to beused as a bio fuel, it is assumed that the concentration ofCu, Cr and As must be reduced to practical zero valuefirst. Instead it is possible that wood, where As has beenremoved and the concentration of Cu and Cr is reducedcan be used in a conventional waste incineration plant.Thereby the energy resource of the wood may still beutilised.

5. Conclusion

CCA-treated wood chips was subjected to electrodi-alytic remediation in a series of 7 pilot scale experi-ments. Several process parameters were investigatedand based on the experiments it was concluded that theuse of collecting units was beneficial for the remedia-tion. The units were placed in the pilot plant 30 cm apart.Soaking of the wood prior to remediation proved toenhance the process, and dual soaking by phosphoricand oxalic acid was more efficient than oxalic acidalone. The wood was soaked 18–24 h in each solutionwith a L :S ratio of 4. The possibility of reusing thesoaking solutions has to be investigated further, sincethis resulted in lower removals. The remediation ofwood chips <2 cm was significantly better than theremediation of the larger wood sizes. The removal of Cuand Cr decreased with increasing electrode distance,whereas the remediation of As seemed unaffected by theincrease in electrode distance from 60 to 150 cm.

The highest removal of both Cu, Cr and As wasobtained in exp. 3, where 87% Cu, 81% Cr and morethan 95% As was removed. In this experiment theelectrode distance was 60 cm and the wood was soakedin phosphoric acid and oxalic acid prior to theelectrodialytic remediation.

54 I.V. Christensen et al. / Science of the Total Environment 364 (2006) 45–54

Acknowledgements

This research is part of a LIFE project. The financialcontribution of the LIFE financial instrument of theEuropean Community is greatly appreciated. RGS90 areacknowledged for supplying and chipping the wood.Laboratory technicians: Bo Vendelbo, Ebba Schnell andKirstine Agger are acknowledged for their careful work.

References

Affald 21, Regeringens affaldsplan 1998–2004 (Waste 21) DanishMinistry of the Environment, ISBN 87-7909-297-7 [in Danish].

Cooper P, Jeremic D, Taylor J. Residual CCA levels in CCA-treatedpoles removed from service.The International Research Group onWood Preservation 2000;IRG/WP 00-50152.

Cooper PA and Associates. Analysis of Consumer lumber Wastemanagement Options, Final Report. Prepared for: Wood Preser-vation Strategic Options Process Manufactors/Treaters SteeringCommittee Consumer Waste Lumber Working Group. July 1.2001.

Hansen HK, Ottosen LM, Kliem BK, Villumsen A. Electrodialyticremediation of soils polluted with Cu, Cr, Hg, Pb and Zn. J ChemTechnol Biotechnol 1997;70:67–73.

Hansen HK, Ottosen LM, Hansen L, Kliem BK, Villumsen A, Bech-Nielsen G. Electrodialytic remediation of soil polluted with heavymetals, key parameters for optimization of the process. TransIchemE 1999;77:218–22 [Part A].

Helsen L, Van den Bulck E. The microdistribution of copper,chromium and arsenic in CCA-treated wood and its pyrolysisresidue using energy dispersive X-ray analysis in scanning electronmicroscopy. Holzforschung 1998;52:607–14.

Jakobsen MR, Fritt-Rasmussen J, Nielsen S, Ottosen LM. Electrodi-alytic removal of cadmium from wastewater sludge. J HazardMater 2004;106B:127–32.

Kristensen IV, Ottosen LM, Ribeiro AB, Villumsen A. Electrochem-ical removal of Cu, Cr and As from CCA-treated waste wood.Proceedings from 3rd symposium and status report on electroki-netic remediation, Karlsruhe 18.–20.; 2001. April.

McQueen J, Stevens J. Disposal of CCA-treated wood. Forest Prod J1998;48:86–90.

Nystroem GM, Ottosen LM, Villumsen A. Electrodialytic removal ofCu, Zn, Pb and Cd from harbour sediment—the influence ofchanging experimental conditions. Environ Sci Technol2005;39:2906–11.

Ribeiro AB, Mateus EP, Ottosen LM, Bech-Nielsen G. Electrodialyticremoval of Cu, Cr and As from chromated copper arsenate-treatedtimber waste. Environ Sci Technol 2000;34:784–8.

Ottosen LM, Hansen HK, Laursen S, Villumsen A. Electrodialyticremediation of soil polluted with copper from wood preservationindustry. Environ Sci Technol 1997;31:1711–5.

Ottosen LM, Kristensen IV, Pedersen AJ, Hansen HK, Villumsen A,Ribeiro AB. Electrodialytic removal of heavy metals from differentsolid waste products. Sep Sci Technol 2003;38:1269–89.

Pedersen AJ. Characterization and electrodialytic treatment of woodcombustion fly ash for the removal of cadmium. BiomassBioenergy 2003;25:447–58.