the removal of heavy metals from contaminated soil by a combination of sulfidisation and flotation

The Science of the Total Environment 290(2002) 69–80

0048-9697/02/$ - see front matter� 2002 Elsevier Science B.V. All rights reserved.PII: S0048-9697Ž01.01064-6

The removal of heavy metals from contaminated soil by acombination of sulfidisation and flotation

Mathias Vanthuyne, Andre Maes*´

KULeuven, Department of Interphase Chemistry, Laboratory of Colloidchemistry, Kasteelpark Arenberg 23, 3001 Heverlee,Belgium

Received 14 August 2001; accepted 25 September 2001

Abstract

The possibility of removing cadmium, copper, lead and zinc from Belgian loamy soil by a combination ofsulfidisation pre-treatment and Denver flotation was investigated. The potentially available — sulfide convertible —metal content of the metal polluted soil was estimated by EDTA(0.1 M, pH 4.65) extraction and BCR sequentialextraction. EDTA extraction is better at approximating the metal percentage that is expected to be convertible into ametal sulfide phase, in contrast to the sequential extraction procedure of ‘Int. J. Environ. Anal. Chem. 51(1993) pp.135–151’ in which transition metals present as iron oxide co-precipitates are dissolved by hydroxylammoniumchloridein the second extraction step. To compare the surface characteristics of metal sulfides formed by sulfidisation withthose of crystalline metal sulfides, two types of synthetic sediments were prepared and extracted with 0.1 M EDTA(pH 4.65) in anoxic conditions. Separate metal sulfides or co-precipitates with iron sulfide were formed by sulfideconditioning. The Denver flotation of both types of synthetic sediments(kerosene as collector at high backgroundelectrolyte concentrations) resulted in similar concentrating factors for freshly formed metal sulfides as for fine-grained crystalline metal sulfides. The selective flotation of metal sulfides after sulfide conditioning of a pollutedsoil, using kerosene or potassium ethyl xanthate as collectors and MIBC as frother, was studied at high backgroundelectrolyte concentrations. The sulfidisations were made in ambient air and inside an anoxic glove box. Theconcentrating factors corrected by the potentially available metal percentage, determined by 0.1 M EDTA extraction,lie between 2 and 3. The selective flotation of these finely dispersed, amorphous, metal sulfides can possibly beimproved by optimising the bubble–particle interaction.� 2002 Elsevier Science B.V. All rights reserved.

Keywords: Transition metals; Availability; Sequential extraction; Speciation; Sulfidisation; Flotation; Metal sulfide

1. Introduction

Excavation and disposal are no longer consid-ered as a permanent solution for heavy metal

*Corresponding author. Tel.:q32-16-32-15-98; fax:q32-16-32-19-98.

E-mail address: [email protected](A. Maes).

contaminated soils. Consequently, the demand fortreatment techniques such as extraction, hydrocy-clonage, solidification, vitrification and flotation,is growing. In this study, we choose to investigatethe potential use of flotation to remediate a soilpolluted with transition metals which are apt toform highly insoluble metal sulfides withpKsp

values ranging from 25.2 for ZnS to 53.2 for HgS(Framson and Leckie, 1978).

70 M. Vanthuyne, A. Maes / The Science of the Total Environment 290 (2002) 69–80

Flotation is a solid–liquid separation techniquewhich introduces air bubbles in a suspension(Matis, 1995). The selectivity of the separation isbased upon differences in the wettability betweenthe particles. Air bubbles attach preferentially tohydrophobic surfaces and carry these particles toa stable froth layer that is produced by the additionof frothing agents, whereas hydrophilic particlesremain in the suspension. If the natural differencesin wettability between the particles that are to beseparated are not sufficient for their selectiveseparation, the suspension has to be conditionedby flotation reagents. The most important flotationreagents are collectors, which adsorb on mineralsurfaces, rendering them hydrophobic and facili-tating bubble attachment.Flotation has been widely and successfully used

in the mining industries to separate valuable min-eral ores, which are commonly present as metalsulfides, from the tailings. Selective flotation ofthese minerals could be enhanced by sulfidisation(Orwe et al., 1998; Clark et al., 2000).Flotation is also a useful remediation technique

for heavy metal polluted dredged material(Cau-wenberg et al., 1998a,b). Up to 80% of heavymetals which were present as metal sulfides, couldbe concentrated in the froth layer, which repre-sented only 30% of the total mass. Maximumselectivity occurred at in situ pH conditions. Theseresults were obtained from a kerosene-inducedflotation of free and organic matter coated sulfideparticles (Cauwenberg et al., 1998a). Sequentialextractions of the froth and rest fraction afterflotation at pH 8 show an important difference inspeciation between the froth and rest fraction formost of the transition metals; the froth fractionshave a lower transition metal extractability thanthe rest fractions due to the enrichment of metalsulfides in the froth fraction(Cauwenberg et al.,1998b).The decontamination of harbour sediments by

flotation in which the metals are originally not allpresent as metal sulfides, could also be improvedby chemical pre-treatment methods such ashydroxylation and sulfidisation(Eberius, 1989).Without sulfidisation pre-treatment, the total heavymetal content reduction in the sediments afterflotation is only 42%. However, flotation of sulfi-

dised harbour sediments reduces the total heavymetal content by 83%(Eberius, 1989).Flotation of heavy metal contaminated surface

soils differs from ore flotation. One of the mostimportant differences is the fact that heavy metalsin soils are not present in one particular chemicalform like it is mostly the case in ores, but areassociated with the different geochemical phasesof the soil leading to low selective flotation.Transferring the potential available transition met-als to one unique chemical form, e.g. a sulfidephase or an oxide phase, by chemical pre-treatmentand further removal of the heavy metals by select-ing a phase-specific collector is a possibility toovercome this problem.Venghaus and Werther(1998) investigated the

flotation behaviour of a zinc-contaminated soilpre-treated by sulfidisation. A metal recovery ofapproximately 50% and a froth mass recovery of10% were obtained at pH 7.5 in the presence of axanthate collector. Langen et al.(1994) also car-ried out laboratory scale flotation experiments onsoils. Based on the results of a sequential extrac-tion procedure, they used an oxide-specific collec-tor to remove the iron, manganese and aluminiumoxide fraction, which was highly contaminatedwith heavy metals. A concentration factor of 1.7–1.75 at a metal recovery of approximately 60–65% could be achieved.The purpose of this paper is to investigate the

possibility to remove heavy metals from a contam-inated soil by a combination of sulfidisation andflotation. Especially the case of a fine-textured soilis addressed, the treatment of sandy soils andsediments being rather easy(Mosmans and vanMill, 1999). The contaminated soil was pre-treatedwith an Na S solution in ambient air or inside an2

anoxic glove box under NyH atmosphere. The2 2

goal of this chemical pre-treatment method was totransfer the potentially available heavy metals intoa unique metal sulfide chemical phase. Metalsulfides might then easily float with a suitablephase-specific collector, e.g. kerosene for free andorganic coated metal sulfides as shown for dredgedmaterial(Cauwenberg et al., 1998a,b) or xanthatefor valuable metal sulfides known in mining indus-tries (Crozier, 1992).

71M. Vanthuyne, A. Maes / The Science of the Total Environment 290 (2002) 69–80

Table 1General characteristics of Tienen soil

pH-KCl 7.83pH-H O2 7.10

TextureSand(%) 28.22Loam (%) 59.34Clay (%) 12.44

Metal content (mgykg)Cadmium 7.80Copper 68.40Lead 141.90Zinc 356.70

Organic matter (%) 6.54CaCO (%)3 7.77

2. Materials and methods

2.1. Soil samples

The investigated soil samples were taken inTienen, Belgium. The general characteristics ofthis loamy soil are given in Table 1. The organicmatter content was estimated by the ignition lossat 4508C. Texture analysis was done by the pipettemethod (Gee and Bauder, 1986). The CaCO3concentration was measured by the addition ofHCl and back titration with NaOH(Gee andBauder, 1986). The soil fraction smaller than 2mm was used in all the experiments.

2.2. Metal analysis

The soil heavy metal content was measuredafter microwave digestion following the procedureof Cauwenberg et al.(1998a). Digestion was doneby Microwave destruction(Milestone, MLS-1200-MEGA). To 0.5-g soil sample, 3 ml HNO(65%,3

p.a.), 1 ml HClO (70%, p.a.) and 1 ml HCl4

(37%, p.a.) was added. The metal content wasmeasured by AAS measurement(Varian, AAS-20). Metal standard solutions were obtained fromAldrich Chemical Company, Inc. After flotation,the heavy metal content in the freeze-dried frothfraction and rest fraction was determined by thesame procedure.

2.3. Preparation of synthetic sediments

Two types of synthetic sediments were prepared.The first sediment was prepared inside an anoxicglove box wN yH s95:5 (%)x in agreement with2 2

the procedure described by Oakley et al.(1980)using 5-g sand(VEL, p.a.), 0.050 g humic acid(Aldrich), 5-g illite clay (Silver Hill Montana)and 0.1 g of different metal sulfides(Aldrich orCerac) in 200 ml background solution of 0.1 MCa(NO ) . The pH was adjusted to 8.5 with 0.13 2

M NaOH. The synthetic sediment was allowed toequilibrate for 1 week. If necessary, the pH wasdaily readjusted to pH 8.5. The illite clay and thesand were first ground in a ball mill to obtain finematerial (30 min for sand and several times 30min for the clay until all clay passed a 200-mmsieve). To obtain small metal sulfides, 20-g metalsulfides were weighed inside the glove box intothe recipient of the ball mill. The recipient wassealed with sealing tape(Aldrich) and the metalsulfides were ground outside the glove box: ZnS(325 mesh) for 7 min at speed 7; PbS(200 mesh)for 45 min at speed 8; CuS(200 mesh) for 60min at speed 8; and FeS(q40 mesh) for 60 minat speed 7.The second sediment was also prepared inside

an anoxic glove boxwN yH s95:5 (%)x) with 5-2 2

g sand (VEL, p.a.), 5-g illite clay (Silver HillMontana) and nitrate salts of Zn, Pb, Cu and Fein 200 ml background solution of 0.1 MCa(NO ) . Occasionally, 0.050 g humic acid3 2

(Aldrich) was added. Na S was added to form2

insoluble metal sulfides at a dose(molyl) of twicethe total metal concentration. The suspension col-oured black immediately after sulfide additionindicating that metal sulfides were formed. ThepH was adjusted to 8.5 with 1 M HNO . The3

sediment was further allowed to equilibrate for aweek inside the glove box with daily pH controland readjustment to pH 8.5.

2.4. Extraction procedures

The single step EDTA extraction procedure ofCottenie et al.(1979) modified by Maes and


Table 2Experimental conditions used in the adapted Ure et al.(1993) sequential extraction procedure

Assumed phase 1-g soil in centrifuge cupinvolved

Fraction 1 Exchangeable 50 ml 0.1125 N HAcFraction 2 Oxide 50 ml 0.5 M hydroxylammoniumchlorideFraction 3 Organic (1) 12.5-ml H O2 2

(2) 60 ml 1 M NH Ac4

Fraction 4 Residual 30-ml HNOyHCl (1:3) heating3

for 3 h on a sand bath

Cauwenberg(2000), was performed on soil sam-ples of Tienen to determine the heavy metalavailability: 6 g of soil sample was extracted with70 ml of a solution containing 0.5 M NaOAc; and0.1 M EDTA at pH 4.65 in ambient air.The BCR sequential extraction technique

described by Ure et al.(1993) was also performedin ambient air on soil samples to study the asso-ciation of the heavy metals with the differentgeochemical phases of the soil. The extractionconditions and the presumed associated phases ofthe adapted Ure et al.(1993) procedure are brieflygiven in Table 2.To compare the surface characteristics of crys-

talline metal sulfides(synthetic sediment 1) withthe surface characteristics of freshly formed metalsulfides(synthetic sediment 2) in suspension, bothtypes of synthetic sediments were extracted insidean anoxic glove box with 0.1 M EDTA at pH4.65. This procedure was used by Maes andCauwenberg(2000) for a canal sediment. Theyobserved from dissolution experiments at differentpH values that cadmium, copper, lead and zincwere mainly present as crystalline sulfides inagreement with theoretical calculations(CHESS,Van der Lee, 1993). However, the dissolutionbehaviour revealed that the transition metals wereeither present as separate metal sulfides or as co-precipitates with iron sulfide(Maes and Cauwen-berg, 2000).

2.5. Sulfidisation procedure

Sulfidisation was done at room temperature intwo conditions:

1. in ambient air; and

2. inside an anoxic glove box to avoid oxidation.

Anoxic conditions inside the glove box wereobtained by continuously flushing 95:5(%) N y2H gas over a catalyst. The oxygen concentration2

as monitored by the trace oxygen analyser(DeltaF) was 2000mgym . The humidity in the glove3

box was controlled by CaCl pellets. The proce-2

dure described below was used in both conditions.Fifty grams of soil was weighed in a 1-l glass

bottle and suspended in 1 l of 0.1 M NaHCO3(Merck, p.a.) or 0.1 M Ca(NO ) (Merck, p.a.)3 2

using fresh, doubly-distilled water. The sulfidisa-tion was done with flake sodium sulfide(65%Na S, Merck, p.a.). The doses of Na S(molyl)2 2

were based on the total soil heavy metal content(mgyg).One percent(wyw) of fine grained FeS(Cerac,

99.9%) was also added in some sulfidisations. Theglass bottle was closed immediately after thesulfide addition to avoid unnecessary oxidation.The soil suspension was shaken in an orbitalshaker for 1–2 days. Reference samples withoutsulfidising agent were prepared by the same pro-cedure. De-aerated doubly-distilled water was usedfor the sulfidisations inside the glove box. The pHafter sulfidisation was measured by a gel electrode(Xerolyt, Ingold). Both types of sulfide condi-tioned soil suspensions were then transferred tothe flotation tank of a Denver D12 flotationmachine.

2.6. Flotation tests

Flotation experiments were performed in ambi-ent air with sulfide conditioned soil suspensionsand with both types of synthetic sediments. Flota-


Table 3Comparison of zinc, copper and lead extractability(%) of different types of synthetic sediments using EDTAyNaOAc (0.1 M, pH4.65) in anoxic conditions with their concentrating factor after Denver flotation

EDTA extractability Concentrating factor

Cu Zn Pb Cu Zn Pb(%) (%) (%) (%) (%) (%)

Crystalline metal sulfides 0.02 7.95 2.66 2.62 3.23 3.09

Native metal sulfidesWithout humic acid 0.01 51.41 16.14 2.75 2.70 2.46With humic acid 0.01 44.78 16.44 2.63 2.64 2.38

Flotation parameters: 1.5-ml kerosene; 30-ml MIBC(0.1% vyv); 0.1 M Ca(NO ) ; conditioning time, 10 min; and flotation time,3 2

10 min.

tion tests were made with a Denver D12 laboratoryscale flotation machine. The flotation reagentswere added after the sulfidisation pre-treatmentstep. Methylisobutylcarbinol(MIBC, 27–30 ml,vyv 1y1000, Merck) was added to obtain a stablefroth layer. The collectors used in this study were1.5-ml commercial kerosene(Merck) or 25 ml 0.1M Kex (potassium ethyl xanthate, Merck). Thesystem was then allowed to condition for 10 minat 1000 rev.ymin.Subsequently, flotation was started by opening

the air supply of the Denver flotation apparatusfor a period of 10 min and the froth layer wasremoved manually. The obtained froth fraction andrest fraction were freeze-dried and weighed, fol-lowed by microwave digestion and AAS analysisto determine their heavy metal content.The flotation efficiency will be evaluated on the

basis of the following parameters:Ž .Mass recovery% s Mass yMassŽ froth froth

qMass =100.rest

BC w xŽ .Metal recovery % s Mass = MefrothD froth

w x=100 yŽMass = Me. froth froth

w xqMass = Me .rest rest

Concentrating factorsmetal recoveryymass recovery

w x w xs Me y Me .froth total

3. Results and discussion

3.1. Denver flotation of synthetic sediments

In a trial to compare the surface characteristicsof crystalline fine grained metal sulfides with thesurface characteristics of native metal sulfides, twotypes of synthetic sediments were prepared andextracted with EDTA(0.1 M) at pH 4.65 insidean anoxic glovebox. The dissolution of copper,lead, zinc and cadmium sulfides in EDTA solution(pH 4.65) is expected to be low based on theoret-ical calculations(CHESS, Van der Lee, 1993; Maesand Cauwenberg, 2000). Nevertheless, Table 3demonstrates that differences in the extractionpattern exist between crystalline metal sulfides andmetal sulfides formed by sulfidisation, especiallyfor zinc and lead which are more easily extractedfrom native sulfides by EDTA. These observationsindicate that part of the lead and zinc is associatedwith an iron sulfide phase because the stability ofsuch iron co-precipitates is lower than the stabilityof the separate transition metal sulfide phase, butis similar to the stability of iron sulfide. Indeed,Maes and Cauwenberg(2000) have shown thatthe dissolution edge for iron starts at approximatelypH 6. Therefore, upon dissolution of the ironsulfide matrix, lead and zinc which were co-precipitated with iron sulfide, were leached by 0.1M EDTA below pH 6. Table 3 shows that for thecase of copper, no influence of EDTA on coppersolubility could be observed for all the synthetic


Table 4Percentage of total metal extracted with 0.1 M EDTA at pH4.65 for the different metal ions from Tienen soil

Cd Cu Pb Zn(%) (%) (%) (%)

69.85 72.74 71.48 64.11

Table 5Metal distribution over the different extraction steps of the Ure et al.(1993) sequential extraction procedure

Phase Cd Cu Pb Zn(%) (%) (%) (%)

Fraction 1 Exchangeable 64.71 7.18 6.80 36.82

Fraction 2 Oxide 35.29 45.27 66.99 36.31Subfraction 1 EDTA-extractable 5.14 45.27 55.94 19.75Subfraction 2 Not EDTA-extractable 30.15 0.00 11.05 16.56

Fraction 3 Organic 0.00 20.29 8.74 7.54

Fraction 4 Residual 0.00 27.26 17.47 19.33

The extractions were made in ambient air and Tienen soil was used; subfractions 1 and 2 were estimated based on the percentageof different metals extracted by 0.1 M EDTA extraction at pH 4.65(see text).

sediments. Therefore, we conclude that copper ispresent as a separate metal sulfide phase.The former observations prove that heavy metals

could be converted into a separate metal sulfidephase or a co-precipitate with iron sulfide. Theappearance of a dark black colour in the syntheticsediment suspensions immediately after sulfideaddition confirms this conclusion. The flotationbehaviour of such freshly formed metal sulfideswas investigated by Denver flotation using ahydrophobic kerosene collector. The concentratingfactors, which determine the selectivity of theflotation process, are also given in Table 3. Theconcentrating factors for copper, zinc and lead areapproximately the same as for the crystalline sep-arate metal sulfide phases(0–50mm) and are notinfluenced by the presence of humic acid at highbackground electrolyte concentrations. High back-ground electrolyte concentrations were found torender the highest concentrating factor due tominimising the effect of coating of metal sulfideswith soil organic matter(Cauwenberg et al.,1998a). This influence of ionic strength on float-ability of metal sulfides was explained as a ‘saltingout effect’(Yariv and Cross, 1979), which changesthe conformation of humic acids due to efficient

charge neutralisation of the humic macromolecule,causing a reduction in the intramolecular repulsionin the polymer chains and ‘coiling’ of the chains.

3.2. Denver flotation of Tienen soil pre-treatedwith sulfidisation

3.2.1. Metal availabilityIn soil chemistry, a single step EDTA extraction

is commonly used for determining the availablemetal ions (Cottenie et al., 1979). Due to thepossible liberation of high amounts of iron intosolution, we used 0.1 M EDTA in agreement withMaes and Cauwenberg(2000) instead of 0.02 M(Cottenie et al., 1979) in order to complex alliron, zinc, cadmium, copper and lead. This EDTA(0.1 M) extractable metal content at pH 4.65 is agood approximation for the percentage of transitionmetals that can possibly be converted into a metalsulfide phase by sulfidisation.The results of the EDTA extraction(0.1 M, pH

4.65) on the metal polluted soil(Table 4) showthat cadmium, copper, lead and zinc are onlypartially (grossly 70%) available. To explain thismoderate metal availability, we also performed aBCR sequential extraction procedure(Ure et al.,1993). The results of this sequential extraction arepresented in Table 5. Heavy metals ‘included inoxides’ and ‘adsorbed onto oxides’ were bothextracted by hydroxylammoniumchloride. There-fore, we attempted to divide the metals associatedwith the oxide fraction into two subfractions,namely an EDTA extractable subfraction(subfrac-tion 1) and a non-EDTA extractable subfraction


(subfraction 2, metals ‘included in oxides’). Thepercentage of heavy metals associated withsubfraction 1 was estimated by subtracting thepercentage of metals associated with the exchange-able fraction(fraction 1) and the organic fraction(fraction 3) from the EDTA (0.1 M, pH 4.65)extractable metal percentage of Table 4.The metal percentages associated with subfrac-

tion 1 were 5.14(Cd), 45.27 (Cu), 55.94 (Pb)and 19.75%(Zn). Consequently, the sum of thepercentage of metals included in oxides(subfrac-tion 2) and the percentage of metals associatedwith the residual fraction gives an idea of thepercentage of metals that are expected to be non-convertible into a metal sulfide phase by sulfidi-sation pre-treatment. The following estimatednon-sulfide convertible metal percentages wereobtained: Cd, 30.15; Cu, 27.26; Pb, 28.52; and Zn,35.89%.We conclude that the foregoing results indicate

that the BCR sequential extraction technique(Ureet al., 1993) cannot be used to estimate the sulfideconvertible(potentially available) metal content ofa metal polluted soil due to the presence oftransition metals as iron oxide co-precipitates,which are dissolved by hydroxylammoniumchlor-ide in the second step of the sequential extractionprocedure, but only partially by EDTA extraction.Thus, the EDTA(0.1 M, pH 4.65) extraction givesa better estimation of the sulfide convertible metalcontent.

3.2.2. Denver flotation experimentsThe potential use of Denver flotation to remove

freshly formed metal sulfides with a sulfide-spe-cific xanthate collector(Crozier, 1992) or a hydro-phobic kerosene collector(Cauwenberg et al.,1998a) was investigated at high background elec-trolyte concentrations. The flotation conditions andreagents were similar as in Cauwenberg et al.(1998a) for dredged material.Irrespective of the pre-conditioning, all Denver

flotations were done in ambient air. This conditionwas chosen because the use of nitrogen gas torestrict oxidation during flotation did not result inthe expected increase in flotation efficiency ofmetal sulfides from dredged material(Cauwen-berg, 1998). This observation was in agreement

with the generally accepted view that contact withoxygen results in a slightly preferable oxidation ofthe metal sulfide surface to a metal-deficient,sulfur-rich and more hydrophobic(better floatable)surface(Clarke et al., 1995; Glazunov and Cher-nykh, 1994; Muster et al., 1996). In all theexperiments, the pH after sulfidisation lies between8 and 9, which corresponds with the optimum pHfor selective flotation of metal sulfides from riversediments according to Cauwenberg et al.(1998a).

3.2.2.1. Sulfidisation in ambient air. The flotationresults using kerosene as collector for Tienen soiltreated in ambient air with Na S(wNa Sxs2 2

10=wtotal metalx) and in presence of 1% FeS(wyw) are given in Fig. 1. Compared with untreatedsoils, sulfide pre-treatment results in(1) anincrease of the mass recovery, and(2) an increaseof the metal recovery for all the investigated heavymetals(Fig. 1a). Thus, the observed increase ofthe metal recovery after sulfidisation could mainlybe attributed to the increase of flotated mass. Thisis explained by the better froth characteristics ofthe soil suspensions treated with Na S, compared2

with untreated soils. In presence of Na S, we2

observed a very stable froth layer with good frothdepth compared with a more brittle and less stablefroth with very shallow depth for untreated soils.Freeman et al.(2000) also noticed that increasingthe amount of NaHS in ore flotation increased thefroth stability due to changing surface tension atthe bubble–water interface.The concentrating factor, which determines the

selectivity of metal sulfide removal, is not signif-icantly improved by sulfidisation and is smallerthan two(Fig. 1b). In some cases, the concentrat-ing factor after sulfidisation is even lower than theconcentrating factor of the untreated soil. There-fore, we can conclude that the flotation selectivityusing kerosene as collector of metal sulfidesformed by sulfidisation in ambient air is lowcompared with the flotation selectivity of metalsulfides from dredged material. Indeed, the flota-tion of metal sulfides, which are formed in organicmatter rich river sediments by bacterial sulfatereduction in anaerobic conditions(Morse et al.,1987), using the same flotation reagents and flo-tation conditions results in metal recoveries up to


Fig. 1. Influence of the combination of sulfidisation and flo-tation on(a) mass recovery and metal recovery, and(b) con-centrating factor for different metal ions from Tienen soil.Sulfidisation with Na S(Na Ss10=wtotal metalx) and 1%2 2

FeS in ambient air. Flotation parameters: 1.5-ml kerosene; 30-ml MIBC (0.1% vyv); 0.1 M NaHCO ; conditioning time, 103

min; and flotation time, 10 min.

Table 6Concentrating factors of different experiments corrected for potentially available metal content determined by 0.1 M EDTA extractionat pH 4.65(Table 4)

Sulfidisation conditions Flotation reagents Corrected concentrating factor

Na Sa2 FeS Electrolyte Environment Collector Frother Cd Cu Zn Pb(molyl) (%)

10 1 NaHCO3 Ambient air Kerosene MIBC 2.42 1.96 2.37 2.003.5 1 NaHCO3 Ambient air Kex MIBC 2.29 2.03 3.22 1.7210 0 NaHCO3 Ambient air Kerosene MIBC 2.46 2.21 2.40 2.162 1 NaHCO3 Glove box Kerosene MIBC 2.10 2.02 2.76 1.892 0 NaHCO3 Glove box Kerosene MIBC 2.15 2.08 2.54 1.992 0 Ca(NO )3 2 Glove box Kerosene MIBC n.d.b 2.30 3.05 n.d.b

wNa Sxsfactor=wtotal metalx (molyl).a2

n.d.snot determined.b

80% and in concentrating factors from 2.5 to 3 atpH 8–9(Cauwenberg et al., 1998a).However, the presently observed low selective

flotation behaviour could be explained by themoderate potentially available heavy metal con-

tent. Correction of the observed concentrating fac-tors for the potentially available (sulfideconvertible) heavy metal content determined bythe single step 0.1 M EDTA extraction at pH 4.65(Table 4), gives the concentrating factors shownin Table 6. These corrected concentration factorsare similar to the concentration factors observedfor dredged material(Cauwenberg et al., 1998a).It is possible that the rather low concentrating

factors in the foregoing experiments were inducedby sorption of HS on the surface of metaly

sulfides, which renders the surfaces more hydro-philic (Matis, 1995). Such sorption would inhibita good hydrophobic interaction between the metalsulfide surface and the oily kerosene collector.Therefore, an experiment was done with Tienen

soil in which the Na S dose was decreased to 3.52

times the total metal concentration and also asulfide-specific Kex collector was used. The resultsshown in Fig. 2 demonstrate that a sulfide-specificcollector Kex, instead of kerosene, does not resultin an improved metal recovery(Fig. 2a) norimproved flotation selectivity(Fig. 2b). ProbablyHS competes with the xanthate collector for they

metal sulfide surface. Since HS forms muchy

stronger complexes with the metal sulfide surface,it is conceivable that HS can prevent xanthatey

from being sorbed resulting in a more hydrophilic,non-floatable metal sulfide(Matis, 1995).In all foregoing sulfidisation experiments, col-

loidal FeS was added as nuclei to promote metalsulfide formation. Preliminary experiments of


Fig. 2. Influence of the combination of sulfidisation and flo-tation on(a) mass recovery and metal recovery, and(b) con-centrating factor for different metal ions from Tienen soil.Sulfidisation with Na S(Na Ss3.5=wtotal metalx) and 1%2 2

FeS in ambient air. Flotation parameters: 25-ml 0.1 M Kex;27-ml MIBC (0.1% vyv); 0.1 M NaHCO ; conditioning time,3

10 min; and flotation time, 10 min.

Fig. 3. Influence of the combination of sulfidisation and flo-tation on(a) mass recovery and metal recovery, and(b) con-centrating factor for different metal ions from Tienen soil.Sulfidisation with Na S(Na Ss10=wtotal metalx) in ambient2 2

air. Soil sample of Tienen was used. Flotation parameters: 1.5-ml kerosene; 30-ml MIBC(0.1% vyv); 0.1 M NaHCO ; con-3

ditioning time, 10 min; and flotation time, 10 min.

aqueous solutions containing Cd(NO ) showed3 2

that in the presence of fine grained FeS(Cerac,99%) cadmium sulfide was formed faster than inabsence of fine grained FeS. FeS is also responsi-ble for the co-precipitation of transition metals insediments(Arakaki, 1992; Maes and Cauwenberg,2000). Moreover, the Denver flotation experimentswith the synthetic sediments(Table 3) showedthat the flotation selectivity of these mixed precip-itates with FeS is similar as for crystalline, goodfloatable metal sulfides.Nevertheless, we observed that FeS addition

during the sulfidisation step in ambient air doesnot improve the flotation behaviour of freshlyformed metal sulfides compared with metal sul-fides formed in absence of iron sulfide(Fig. 3a,b).Therefore, the role of FeS in the formation ofmetal sulfides will further be investigated in thesulfidisations under anoxic conditions.

3.2.2.2. Sulfidisation in the glove box under N y2

H atmosphere. Some sulfidisations with Na S2 2

were done in anoxic conditions to avoid oxidationduring the sulfidisation step. In addition, the dose

of Na S was further decreased to twice the total2

metal concentration to avoid flotation depressionby excess HS . Compared with the flotationy

results of metal sulfides formed in ambient air, theflotation of metal sulfides formed under anoxicconditions are not more selective(Fig. 4b). More-over, again no effect of iron sulfide addition wasobserved(Fig. 4b). The reason for this behaviouris unclear in the sense that iron sulfide wasexpected to act as a promoter for metal sulfideformation.When 0.1 M Ca(NO ) was used as background3 2

electrolyte instead of 0.1 M NaHCO during one3

sulfidisation experiment in anoxic conditions inorder to improve the flocculation of the finelydispersed metal sulfides and organic matter, it wasnoticed that the metal recovery was rather low, butthat the heavy metals are concentrated in less mass(Fig. 5a,b). The mass flotated is an importantflotation parameter, which influences the re-usablesoil fraction and should be as low as possible.


Fig. 4. Influence of the combination of sulfidisation and flo-tation on(a) mass recovery and metal recovery, and(b) con-centrating factor for different metal ions from Tienen soil.Sulfidisation with Na S(Na Ss2=wtotal metalx) with and2 2

without FeS(1%) in a glove box. Soil sample of Tienen wasused. Flotation parameters: 1.5-ml kerosene; 30 mlMIBC(0.1% vyv); 0.1 M NaHCO ; conditioning time, 10 min; and3

flotation time, 10 min.

Fig. 5. Influence of the combination of sulfidisation and flo-tation on(a) mass recovery and metal recovery, and(b) con-centrating factor for different metal ions from Tienen soil.Sulfidisation with Na S(Na Ss2=wtotal metalx) in a glove2 2

box. Soil sample of Tienen was used. Flotation parameters: 1.5-ml kerosene; 30-ml MIBC(0.1% vyv); 0.1 M Ca(NO ) ; con-3 2

ditioning time, 10 min; and flotation time, 10 min.

However, the metal removal must be sufficientlyhigh to fulfil environmental regulations for the restfraction. Therefore, the decontamination of pollut-ed soils is always a trade-off between metal recov-ery and the concentrating factor. In the presentcase, the lower flotated mass is possibly a conse-quence of using Ca(NO ) instead of NaHCO as3 2 3

background electrolyte. As a result of the interac-tion with Ca , the soil particles flocculate much2q

better resulting in a lower flotated mass.

3.2.3. Explanations for the observed flotationbehaviourPossible explanations for the observed low

selective Denver flotation of metal sulfides formedby sulfidisation pre-treatment of a metal-contami-nated soil might be the surface characteristics(wettability, hydrophobicity, etc.) and the dimen-sions of the freshly formed metal sulfides.

In order to explain this, one should bear in mindthat the probability of collision and probability ofattachment between particle and bubble are themost important parameters which contribute to theprobability of particle flotation(Matis, 1995). Theprobability of collision depends on the dimensionsof both particle and bubble, whereas the probabilityof attachment depends on the surface characteris-tics of the particle.In the size range 20–50mm the selectivity of

Denver flotation reached a maximum because bothcollision probability and attachment probabilitywere favourable(Cauwenberg et al., 1998b). Forparticles with dimensions lower than 5mm, theprobability of attachment remains constant forparticles with the same mineralogy, but the particlediametersyair bubble diameter ratio is less favour-able for collision, resulting in lower Denver flota-tion specificity in agreement with Cauwenberg etal. (1998b). Therefore, it is possible that the finelydispersed, amorphous metal sulfides formed bysulfidisation pre-treatment of a heavy metal pol-


luted soil are too small in comparison with thebubbles generated by the Denver flotation appa-ratus. Moreover, highly turbulent flotation condi-tions (e.g. in a mechanically agitated Denver cell)also increases the aspecific entrainment of non-sulfidic soil particles resulting in a high massrecovery and a low concentration factor.

4. Conclusions

The BCR sequential extraction technique(Ureet al., 1993) is not a good method to estimate thesulfide convertible(potential available) metal con-tent of a metal polluted soil due to the presenceof transition metals as iron oxide co-precipitates.The single step EDTA extraction(0.1 M, pH 4.65)gives a better approximation.The EDTA (0.1 M, pH 4.65) extractability of

transition metals from synthetic sediments in anox-ic conditions showed that transition metals can beconverted into a metal sulfide phase by sulfidisa-tion. Either a separate metal sulfide phase or a co-precipitate with iron sulfide is formed. The Denverflotation of both types of synthetic sediments usingkerosene as collector at high background electro-lyte concentrations, demonstrated that the concen-tration factors for freshly formed metal sulfidesare approximately the same as for fine grainedcrystalline metal sulfides.Notwithstanding the fact that the Denver flota-

tion of freshly formed metal sulfides after sulfidi-sation pre-treatment in ambient air and in anoxicconditions of a metal-contaminated soil is nothighly selective(concentrating factors-2), theobtained results are encouraging for furtherresearch in this field. Up to 50% of the transitionmetals could be removed. The concentrating fac-tors corrected for the potentially available(sulfideconvertible) metal content, determined by EDTA(0.1 M) extraction, lie between 2 and 3(Table 6)and approach the values obtained by Cauwenberget al. (1998a) for dredged material.The selective flotation of these finely dispersed,

amorphous metal sulfides can possibly beimproved by optimising the bubble–particle inter-action. Therefore, the use of other flotation tech-niques (e.g. dissolved air flotation) which morespecifically separate the smallest particles will be

investigated in future research. These flotationtechniques achieve also better hydrodynamic con-ditions compared with the highly turbulent condi-tions in a Denver flotation cell.

Acknowledgments

The authors are grateful to the FWO(G.0306.97N) for financial support. Vanthuyne M.thanks the IWT-Vlaaanderen for awarding aresearch grant.

References

Arakaki T. Interactions of cations with iron monosulfidemackinawite in dilute solutions and seawater. Ph.D. TexasA and M University, 1992:185 p.

Cauwenberg P. Speciation and behavior of transition metals indredged material and their potential removal by flotation.Ph.D. Kuleuven University, 1998:161 p.

Cauwenberg P, Verdonckt F, Maes A. Flotation as a remedia-tion technique for heavily polluted dredged material. 1. Afeasibility study. Sci Total Environ 1998;209:113–119.

Cauwenberg P, Verdonckt F, Maes A. Flotation as a remedia-tion technique for heavily polluted dredged material. 1.Characterisation of flotated fractions. Sci Total Environ1998;209:121–131.

Clark DW, Newell AJH, Chilman GF, Capps PG. Improvingflotation recovery of copper sulfides by nitrogen gas andsulfidisation conditioning. Mineral Eng 2000;13(12):1197–1206.

Clarke P, Fornasiero D, Ralston J, Smart RS. A study of theremoval of oxidation products from sulfide mineral surfaces.Mineral Eng 1995;8(11):1345–1357.

Cottenie A, Camerlynck R, Verloo M, Dhaese A. Fractionationand determination of trace elements in plants, soils andsediments. Pure Appl Chem 1979;52:45–53.

Crozier KA. Flotation: theory, reagents and ore testing. Oxford:Pergamon press, 1992. (329 p).

Eberius E. Verfahren zur dekontaminierung schlammartigersedimente. European patent 0332958, 1989.

Framson PE, Leckie JO. Limits of coprecipitation of cadmiumand ferrous sulfides. Environ Sci Technol 1978;12(4):465–469.

Freeman WA, Newell R, Quast KB. Effect of grinding mediaand NaHS on copper recovery at Northparkes Mines. Min-eral Eng 2000;13(13):1395–1403.

Gee GW, Bauder JW. Particle-size analysis. In: Klute A, editor.Methods of soil analysis, part 1, physical and mineralogicalmethods Madison, Wisconsin, USA: American Society ofAgronomy, 1986:377–382.

Glazunov LA, Chernykh SI. Increase in mineral flotationeffectiveness by creation of an elemental sulfur coating. In:Castro S, Alvarez J, editors. Proceedings of the FourthMeeting of the Southern Hemisphere on Mineral Technologyand Third Latin-American Congress on Froth Flotation, vol.2, Concepcion, Chile 1994:309–318.


Langen M, Hoberg H, Hamacher B. Prospects for separatingheavy metals from contaminated soil. Aufbereitungs-Technic1994;35:1–12.

Maes M, Cauwenberg P. Influence of pH and oxidation kineticson the solubility of metal sulfides in sludge from a canalsediment. Abstracts of the Third International Conferenceon Carpathian Euroregion Ecology, 21–24 May, Miskolc,Hungary 2000:49–50.

Matis KA. Flotation science and engineering. New York, USA:Marcel Dekker Inc, 1995. (558 pp).

Morse JW, Millero FJ, Cornwell JC, Rickard D. The chemistryof the hydrogen sulfide and iron sulfide in natural waters.Earth Sci Rev 1987;24:1–42.

Muster TH, Toikka G, Hayes RA, Prestidge CA, Ralston J.Interactions between zinc sulfide particles under flotation-related conditions. Colloids Surf A 1996;106:203–211.

Mosmans S, van Mill G. Optimising and modelling of flotationtechniques for remediation of contaminated sediments. In:De Schutter G, editor. Proceedings of the Twelfth Interna-tional Harbour Congress, Characterisation and treatment ofsediments, CATS 4, 15–17 September, Antwerpen1999:191–200.

Oakley SM, Delfey CE, Williamson KJ, Nelson PO. Kineticsof trace metal partitioning in model anoxic marine sedi-ments. Water Res 1980;14:1067–1072.

Orwe D, Grano SR, Lauder DW. Increasing fine copperrecovery at the Ok Tedi concentrator, Papua New Guinea.Mineral Eng 1998;11(2):171–187.

Yariv S, Cross H. Geochemistry of colloid systems for earthscientists. Berlin, Germany: Springer-Verlag, 1979. (450 pp).

Ure AM, Quevauviller PH, Muntau H, Griepink B. Speciationof heavy metals in soils and sediments: an account of theimprovement and harmonization of extraction techniquesundertaken under the auspices of the BCR of the commis-sion of the European communities. Intern J Environ AnalChem 1993;51(1–4):135–151.

Venghaus T, Werther J. Flotation of a zinc-contaminated soil.Adv Environ Res 1998;2(1):77–91.

Van der Lee J.CHESS, another speciation and surface com-plexation computer code. Technical report, LHMyRDy93y39. Centre of Mathematical hydrology of the ecole des´

mines de Paris, 1993.

the removal of heavy metals from contaminated soil by a combination of sulfidisation and flotation

Documents