Biodegradation of Metal−[S,S]-EDDS Complexes - American Chemical

Laboratory of Microbial Ecology and Technology, Department of Biological and Applied Sciences, University of Gent,. Coupure L 653, 9000 Gent, Belgium,...
1 downloads 0 Views 86KB Size
Environ. Sci. Technol. 2001, 35, 1765-1770

Biodegradation of Metal-[S,S]-EDDS Complexes P H I L I P P E C . V A N D E V I V E R E , †,§ H A N S S A V E Y N , ‡,§ W I L L Y V E R S T R A E T E , § TOM C. J. FEIJTEL,# AND D I E D E R I K R . S C H O W A N E K * ,# Laboratory of Microbial Ecology and Technology, Department of Biological and Applied Sciences, University of Gent, Coupure L 653, 9000 Gent, Belgium, and Procter & Gamble European Technical Center, Temselaan 100, 1853 Strombeek-Bever, Belgium

The [S,S]-stereoisomer of ethylenediaminedisuccinic acid (EDDS), a biodegradadable strong metal chelant, has substituted traditional chelants in a number of consumer products. However biodegradability of metal-EDDS complexes has remained largely undocumented. In the present study, activated sludge fed with EDDS as sole C and N source, was shown to readily biodegrade 1 mM pulses of Ca-, Cr(III)-, Fe(III)-, Pb-, Al-, Cd-, Mg-, Na-, or ZnEDDS (the latter only after extensive lag phase). On the other hand, the Cu-, Ni-, Co-, and Hg-complexes remained essentially undegraded. Only in the case of HgEDDS was lack of biodegradation due to metal toxicity. Speciation analysis revealed free HEDDS3- concentration was higher than 10-5.4 M for all readily biodegradable metalEDDS complexes and smaller than 10-9.0 M for all recalcitrant complexes at pseudo-steady-state (i.e. after initial rise of aquo metal concentration at onset of biodegradation). The rate of metal-EDDS degradation may be modeled with a Monod expression with HEDDS3as substrate (half-saturation constant ca. 10-6 M). This model explains the drastic effect of additional metal ligands, e.g. phosphate or iron, on biodegradation rate of several recalcitrant metal-EDDS complexes. Continuously fed aerated biofilters removed 10 mM Pb- or ZnEDDS at a rate of ca. 0.4 mM h-1.

Introduction Ethylenediaminedisuccinic acid (EDDS), like its structural isomer EDTA, forms stable hexadentate chelates with transition metals. Unlike EDTA, the [S,S]-stereoisomer of EDDS, is readily degraded in activated sludge systems (1-3). This superior biodegradability of EDDS and an overall favorable environmental profile (4) has prompted its use in consumer products, e.g. washing powders. Reviews on chelant extraction of heavy metals from contaminated soils (5-7) have indicated that though soil washing with solutions of chelating agents is an attractive technology, there remains a lack of consensus concerning the choice of the most appropriate * Corresponding author phone: +32 2 456 2900; fax: +32 2 456 3248; e-mail: [email protected]. † Present address: OWS nv, Dok Noord 4, 9000 Gent, Belgium. ‡ Present address: Particle and Interface Technology Group, Department of Biological and Applied Sciences, University of Gent, Coupure L 653, 9000 Gent, Belgium. § University of Gent. # Procter & Gamble European Technical Center. 10.1021/es0001153 CCC: $20.00 Published on Web 03/23/2001

 2001 American Chemical Society

chelating agent(s). The most commonly used chelating agents are either poorly biodegradable (e.g. EDTA), associated with other safety/regulatory issues (e.g. NTA), or little effective (e.g. citrate). [S,S]-EDDS seems a promising alternative. Though the ready biodegradability of [S,S]-EDDS in domestic sewage has been amply demonstrated, the biodegradability of metal-EDDS complexes has only been examined in cell-free extracts (8). The fate of metal-EDDS complexes is of concern because EDDS used in consumer products or soil wash formulations may end up in surface water bodies. Numerous studies with other chelating agents have shown that the biodegradability of metal complexes can differ drastically from that of the corresponding free chelating agents (9-11). The limiting step to complex biodegradation has generally been identified as complex dissociation because it is thought that predominantly uncomplexed, or sometimes Ca-bound, EDTA or NTA are taken up by bacteria (12-16). It is therefore not unexpected that complexes with a very high stability constant tend to be much less amenable to biodegradation than complexes with a relatively low stability constant (11, 14, 15). The sole consideration of the stability constant as such is however of insufficient accuracy to predict metal complexes biodegradability (14, 17-19). Recent efforts to model the biodegradation kinetics of metal complexes follow two approaches. One is to take into account the rate at which metal complexes dissociate and reassociate, i.e., their lability (11, 18). Another approach is to couple aqueous speciation changes to biodegradation reaction kinetics and stoichiometry (20). Our objective was to examine the biodegradability of metal-EDDS complexes (mM range) in a model activated sludge system, derive a mechanistic model of biodegradation rate, and apply these findings in a semipilot scale biofilter reactor. This study fits in the context of a research project exploring the potential use of [S,S]-EDDS as an environmentally compatible metal decontamination agent.

Experimental Section Biodegradability of Uncomplexed EDDS and Stock Sludge. A 1-L semicontinuously fed activated sludge vessel was started with activated sludge from a municipal wastewater treatment plant and fed with 1 g of milk-COD L-1 d-1. The sludge was acclimated to EDDS by spiking 20 mg of Na3EDDS every other day (from a 1.25 M commercial Na3EDDS solution, 98+% [S,S]-isomer (Assoc. Octel Cy Ltd., South Wirral, U.K.). After 12 days, the mixed liquor was diluted 5 times with tap water and split into five reactors, 1 L each (ca. 0.6 g of MLSS L-1, measured gravimetrically after centrifugation and drying at 105 °C). The reactors were fed with different concentrations of EDDS and/or milk as detailed in the Results section. The NaEDDS-acclimated stock sludge, used in all experiments described below, was maintained in 1-L reactors fed three times weekly. The feeding cycle consisted in settling, decanting the supernatant (700 mL), and adding 500 mL of raw municipal sewage, 195 mL of treated municipal sewage, and 6.2 mmol of [S,S]-Na3EDDS from the stock solution. The sludge concentration fluctuated around 2 g of MLSS L-1. Incubation conditions used for sludge maintenance and all experiments described below were 28 °C in the dark on a rotary shaker at 140 rpm (except for the BAF reactors, kept at 20 °C in daylight). Biodegradability of Metal-EDDS Complexes. Solutions (200 mL) of 1 mM metal-EDDS complexes (equimolar amounts of EDDS and metals) were prepared in distilled water and inoculated with 3 mL of settled EDDS-acclimated activated sludge in duplicate flasks (200 mg of MLSS L-1, VOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1765

FIGURE 1. Downward flow biological aerated filter (BAF) reactor used to treat metal-EDDS feed solutions. after inoculation; pH adjusted to 7.5 ( 0.2 with HNO3 or NaOH). The metal-EDDS solutions had been prepared with the chloride metal salts, except for Al, Cr(III), Cd, Cu(II), Fe(III) for which sulfate salts were used. At different time intervals, total EDDS remaining in solution was determined colorimetrically in the filtrates (8 µm) of 30-mL samples as described below. Effect of Metal Ligands on Metal-EDDS Degradation. Ten-fold dilutions of NaEDDS-adapted sludge in 100 mL of tap water (140 mg of MLSS L-1, final concentration) were spiked with metal-EDDS complexes (1 mM, final concentration). Metal ligands (5 mM K2HPO4, 5 mM FeSO4, or iminodiacetate-functionalized macroporous polystyrene resin beads (Bayer, Leverkusen, Germany)) were added except in the control (all in duplicate flasks) and total EDDS in liquid phase was followed over time. The added Fe(II) is very rapidly oxidized to Fe(III) which forms colloidal FeOOH. This is accompanied by a pH drop which was neutralized via NaOH addition. Initial pH was 7.0 ( 0.2. After 6 days, however, pH had drifted to 8.1-8.7 (except in the iron treatment were the pH was stable). Effect of Colloidal Iron at Various pHs. Solutions (150 mL) of 1 mM metal-EDDS (metal ) Cu, Ni, or Co) were inoculated with 3 mL of settled NaEDDS-adapted activated sludge (200 mg of MLSS L-1, after inoculation). One treatment received no further addition and pH was set at 7.0 (control). Two treatments received 5 mM FeSO4, one being adjusted to pH 6.2 and the other to pH 8.8 (the Fe(II) is readily oxidized and precipitates as FeOOH with either a net positive charge (at pH 6.2) or a net negative charge (pH 8.8)). Another two treatments received iron in the form of the complex Fe(III)EDDS (1 mM), one at pH 6.2 and the other at pH 8.8. The concentration of total soluble EDDS was followed over time with the modified colorimetric assay on filtered samples. Biological Aerated Filter (BAF) Setup. BAF reactors were constructed with plexiglass tubes (5.4 cm inside diameter), filled with 2 L of 5-7 mm porphyre gravel (0.9 m bed height), and inoculated with 1 L of NaEDDS-acclimated activated sludge (Figure 1). The 10 mM EDDS feed solution flowed by gravity downward through the submerged bed, countercurrent with ascending air injected underneath. Liquid volume inside the reactor was 1.3 L, feeding rate 1.3 L d-1, hydraulic retention time 24 h, and volumetric loading rate (Bv) 10 mmol L-1 d-1 or 2.7 g of COD L-1 d-1. The beds were backwashed approximately every other week in order to remove excessive biofilm growth which impaired the gravity flow. Backwash cycle consisted in scrubbing with high-pressure air for one minute and pumping water from underneath to wash out the dislodged biomass flocs. Feeding Cycles of the BAF Reactors. Three BAF reactors were run in parallel, fed with 10 mM EDDS complexes supplemented with different concentrations of either phosphate or FeCl2, which had been shown to aid metal-EDDS biodegradation in the batch experiments. In the first 2-month period, the BAF reactors were fed with solutions of single metal complexes containing (unless otherwise noted) 10 mM 1766

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

FIGURE 2. Molecular structures and acidity constants of ethylenediaminedisuccinic acid (EDDS) and ethylenediaminetetraacetic acid (EDTA) (25 °C, 0.1 M KNO3). metal-EDDS, 2 mM KH2PO4, and 2 mM KHSO4 in tap water (pH 7.0). During the third month, the feed solution was a mixture of all three metals, containing (mM) ZnEDDS (4.25), PbEDDS (0.45), CuEDDS (0.30), KH2PO4 (6.0), NaCl (1.5), and Na2SO4 (4.25) in tap water. EDDS was measured in effluent samples with the colorimetric assay. Total metal concentrations were measured in effluent samples without any pretreatment via flame atomic absorption spectrophotometry (Perkin-Elmer AAS 3110). EDDS Assay. A colorimetric assay of EDDS was developed, based on the 670 nm absorption peak of the blue CuEDDS complex. Samples were filtered on paper, diluted to less than 2 mM EDDS, and acidified to pH 2.1 ( 0.2 with HNO3. Following addition of CuSO4 to 3 mM and 30 min incubation at room temperature, absorbance was measured at 670 nm (this absorption peak was preferred over the one at 290 nm because organic compounds were found to interfere with the absorption in the UV region). Reproducibility was ca. 10% and the detection limit was 0.1 mM EDDS. A disadvantage of this assay is the possibility of interference. While monocarboxylic acids such as acetic acid cause only a negligible interference, aminoethyl aspartic acid, an intermediate degradation product of EDDS, absorbed 25% as much light at 670 nm as an equal amount of EDDS (at pH 2.1). This interference can be neglected in our experiments since aminoethyl aspartic acid does not accumulate during [S,S]-EDDS biodegradation (3). When the presence of Fe(III)EDDS was possible in the sample to be assayed, a modified procedure was used which facilitated the Cu exchange reaction. In the modified procedure, a pH of 5.0 was used instead of 2.1, 50 mM acetate was added as pH buffer, and the incubation period was allowed to proceed overnight in order to allow added Cu to displace Fe(III) completely from EDDS.

Results and Discussion Biodegradation of Uncomplexed EDDS. While EDDS is a close isomer of EDTA (Figure 2), the two molecules display radically distinct biodegradabilities. EDTA may be rapidly degraded as sole carbon source in enrichment cultures (14, 19, 21) but fails to be eliminated in activated sludge systems even after prolonged acclimation (1, 2). On the other hand, the [S,S]-stereoisomer of EDDS (abbreviated EDDS) is rapidly degraded in activated sludge, after a 5-10 days acclimation (2, 3). Municipal sewage sludge, acclimated to NaEDDS in the laboratory, was fed with EDDS only, milk only, or a mixture

TABLE 1. Yield Coefficients of Semicontinuously Fed Activated Sludges Growing on EDDS, Milk, or Mixture Thereofa loading rate Bv (g COD/L.d) milk EDDS milk milk + EDDS milk + EDDS EDDS a

1.0 1.0 1.0 0.0

0.0 0.05 0.5 0.5

sludge concn X yield coeff Y (7 d) (g MLSS/L) (g MLSS/g COD) 2.03 2.13 2.80 1.39

0.30 0.30 0.30 0.34

Calculated using eq 1 for 7 days growth.

of both (Table 1), and the yield coefficient was calculated using the integrated form of eq 1 for a seven day growth period

dX/dt ) YBv - bX

(1)

where X is the sludge concentration (0.6 g MLSS L-1 at time 0), t is time (d), Y is the yield coefficient (g MLSS g COD-1), Bv the organic loading rate (g COD L-1 d-1), and b the decay coefficient (fitted value 0.07 d-1). All vessels displayed a similar yield coefficient (0.30-0.34 g g-1), indicating that EDDS had no adverse effect on growth even upon discontinuous feeding with 3.3 mM pulses every other day. In fact, the sludge could grow very efficiently on EDDS alone, and the high sludge yield obtained with EDDS as sole substrate indicated that the EDDS molecule was most likely fully mineralized. Biodegradation of Metal-EDDS Complexes. Solutions of metal-EDDS complexes, prepared by mixing equimolar amounts of metal salts and Na3EDDS (1 mM each), were inoculated with EDDS-acclimated activated sludge (200 mg of sludge L-1). The concentration of predominant species, calculated with the computer program MINEQL+ (24), is shown in Table 2. According to these speciation calculations, the Fe(III)EDDS complex was not stable and did not form at the pH value 7.5 used in our experiments. This is however contradicted by our experience since we could prepare in deionized water stable 5 mM Fe(III)EDDS solutions which remained transparent for at least 20 h in the pH range 6.89.3. At higher and lower pH values, an FeOOH precipitate formed. The Ca-, Mg-, Cd-, Fe(III)-, Al-, Pb-, and Cr(III)EDDS complexes were readily biodegraded at an average rate of 0.3 mM d-1 (Figure 3). The biological mediation of metal-EDDS disappearance observed here and in all other experiments was demonstrated at the hand of sterile controls showing that no significant EDDS disappearance occurred

in the absence of biological activity, ruling out e.g. sorption effects (data not shown). The Zn-, Cu-, Ni-, Co-, and HgEDDS complexes remained undegraded after 15 days incubation. At that time, 1 mM NaEDDS was added in order to see whether lack of degradation was due to toxicity or recalcitrance. This pulse of EDDS was degraded in all cases except for the Hg complexes, indicating that only in the case of this metal was the lack of biodegradation a consequence of metal toxicity. The general lack of metal toxicity observed here (except for Hg) was not surprising since it is well-known that complexed metals are much less toxic than aquo metals Men+ (14, 17, 22, 23). Related studies with NTA, EDTA, or citrate also showed that biodegradation of metal complexes was strongly dependent on the type of metal (9, 10). At times, an inverse relationship between complex biodegradability and stability constant was noted (11, 14, 15). In this study also, a rough correlation was found between complex stability Kst (defined in Table 2) and biodegradability, as those complexes with a stability constant greater than 1013 were not degraded (Table 2). Kst appears however as a poor predictor of biodegradability. For example, Pb- and ZnEDDS complexes have practically the same stability constant (1012.7 and 1013.5) but drastically distinct biodegradabilities. Ca- and CdEDDS complexes are both readily biodegraded though CdEDDS has a much greater stability constant. Also, addition of phosphate may greatly stimulate the biodegradation of certain metal-EDDS complexes (see below), obviously without affecting these complexes’ Kst values. In fact, many researchers have failed to observe a correlation between the biodegradability of metal complexes of EDTA, NTA, or citrate and their stability constants (14, 17-19). In this study, a good correlation was found between metal-EDDS biodegradability and the concentration of the free HEDDS3- species since the latter was always greater than 10-5.4 M for the readily biodegradable metal-EDDS complexes and always smaller than 10-6.4 M for the recalcitrant complexes (Table 2). This observation suggests that metal-EDDS complexes first need to undergo dissociation to HEDDS3- before cell uptake and biodegradation can occur. This hypothesis has often been stated for metal complexes of EDTA and NTA (12-15, 18, 23). It remained unclear however why ZnEDDS, with 10-6.4 M HEDDS3- in the assay mixture, was not biodegraded since the half-saturation constant Km for HEDDS3- is about 2 × 10-6 M (3, 14, 15). It should be recognized however that the concentration of HEDDS3- often drops by several orders of magnitude as soon as biodegradation is initiated. This is illustrated for CuEDDS in Figure 4. Would bacteria succeed

TABLE 2. Computer Simulation of MeEDDS(4-n)-, Aquo Metal Men+, and Free HEDDS3- Concentrations (log M) Initially Present in the Assay Mixture (1 mM Metal and EDDS Added Each), and after Biodegradation of 0.1 mM EDDS (Pseudo-Steady-State)a after degradation of 10-4 M EDDS

before onset of biodegradation

Kstb Ca Mg Fe(III) Cd Pb Zn + K2HPO4

4.2 5.8 22.0 10.8 12.7 13.5

Zn Cu Co Ni

13.5 18.4 14.1 16.8

solid phase

FeOOH CdCO3 PbCO3 c

Me-EDDS

aquo Me

HEDDS

solid phase

-4.23 -3.27 -6.40 -3.02 -3.00 -3.00

Biodegraded -3.5 -3.8 -24.5 -7.4 -6.8 -8.5

-3.5 -3.6 -3.0 -3.3 -5.4 -4.6

FeOOH CdCO3 PbCO3 c

-3.00 -3.00 -3.00 -3.00

Recalcitrant 6.5 -10.5 -7.7 -8.7

-6.4 -9.1 -7.7 -8.7

ZnCO3 Cu2(OH)2CO3 Co(OH)2 Ni(OH)2

aquo Me

HEDDS

-3.5 -3.7 -24.5 -7.4 -6.8 -8.3

-3.6 -3.7 -3.0 -3.4 -5.4 -4.7

-4.0 -7.4 -6.1 -4.2

-9.0 -12.3 -9.2 -12.1

a Calculations were performed with the computer program MINEQL+ (24) for a system open to the atmosphere (10-6.3 M carbonate) (Al and Cr are omitted as their Kst are not available). b Log value of the complex stability constant (Me-EDDS(4-n)-)/((EDDS4-)(Men+)). c Freed aquo Zn ion is complexed as soluble Zn3(PO4)2 (10-5.1 M).

VOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1767

FIGURE 3. Biodegradation of equimolar metal-EDDS complexes in 10-fold diluted activated sludge, preacclimated to NaEDDS (mean of duplicate flasks).

FIGURE 5. Effect of K2HPO4 and FeOOH on disappearance of equimolar metal-EDDS complexes in 10-fold diluted activated sludge, preacclimated to NaEDDS (mean of duplicate flasks; bars indicate standard deviation). Lines are modeled biodegradation of ZnEDDS, based on the Monod equation wherein the substrate is taken as EDDSH3- at pseudo-steady-state (constant, from Table 2), Km is 2 × 10-6 M, and Vmax is 0.3 mM d-1 (assumed constant, i.e., without growth).

TABLE 3: Effect of Various Metal Ligands on the Biodegradation of Equimolar metal-EDDS Complexes (1 mM) in Diluted Activated Sludge, Preacclimated to NaEDDS (Compilation of Several Experiments)a FIGURE 4. Computer simulation (MINEQL+) of soluble species at onset of biodegradation of 1 mM CuEDDS. Pseudo-steady-state is reached once a copper solid phase forms (malachite, after 0.1% degradation of initial CuEDDS). to degrade 10-6 M CuEDDS (0.1% of initial amount, thus not detectable in our experiment), 10-6 M Cu2+ would be released, most of it hydrolyzed to soluble Cu(OH)2, leaving 10-7.4 M aquo Cu2+ in solution. This 1000-fold increase of Cu2+ concentration (relative to 10-10.5 M initially present) translates into a 1000-fold decrease of HEDDS3- concentration, as required by the Kst equation. Once Cu2+ has reached 10-7.4 M, malachite starts to precipitate and, would further biodegradation of CuEDDS occur, both Cu2+ and HEDDS3- will remain at a pseudo-steady-state concentration (Figure 4). The computed pseudo-steady-state concentration of HEDDS3-, reported in Table 2 (right-most column) for the different metal-EDDS complexes, seems a good predictor of complex biodegradability. The readily degradable complexes had a pseudo-steady-state HEDDS3- greater than 10-5.4 M, whereas the recalcitrant complexes had a pseudo-steadystate HEDDS3- smaller than 10-9.0 M. This is in accordance with a simple Monod model of metal-EDDS biodegradation kinetics where the substrate is HEDDS3- with a Monod constant Km around 10-6 M. A limitation of the proposed model is that is does not yet accurately discriminate between the kinetics of rapidly degraded complexes (cf. Table 2, Figure 3). There must be some additional factors that play a role here. From an environmental and technical perspective it is however important to be able to predict whether a complex will be degraded or persist. Effect of Metal Ligands (Phosphate, Iron, Resin). The addition of metal ligands to metal-EDDS solutions is expected to scavenge nonhydrolyzed aquo metals Men+, which may decrease their pseudo-steady-state concentration. A smaller Men+ concentration would translate into higher 1768

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

control + 5 mM K2HPO4 + 5 mM FeCl2 (or FeSO4) iminodiacetate resin beads

Zn

Cu

Cd

Ni

Co

+ ++ + ++

+ + ++

+ ++ + ++

nd

+ nd

a (-), no degradation; (+), slow or incomplete degradation; (++), degradation complete within a few days; (nd), not determined.

HEDDS3- concentration which, according to the Monod model, should yield a higher rate of metal-EDDS biodegradation. This hypothesis is supported by the work of Klu ¨ ner et al. (14) who observed a more rapid biodegradation of ZnEDTA when phosphate was added to the culture medium, which they ascribed to secondary reactions between Zn2+ and phosphate. In our experiments, the addition of 5 mM K2HPO4 to the sludge suspension containing 1 mM ZnEDDS is expected to decrease the pseudo-steady-state Zn2+ concentration 104.3-fold through the formation of the soluble complex Zn3(PO4)2 (Kst ) 1035.3 (25)). As a result, HEDDS3should increase by the same factor, from 10-9.0 to 10-4.7 M, and therefore be readily biodegraded (Table 2). This prediction was matched by the experimental results (Figure 5). Even in the absence of phosphate, ZnEDDS was rapidly degraded after an 11-day long lag phase (Figure 5). A possible explanation for this lag phase, which in fact can also be noticed in Figure 3, is the progressive accumulation of carbonate during initial biodegration, which would affect the pseudo-steady-state Zn2+ and HEDDS3- concentrations. Phosphate also enhanced Cd- and CuEDDS biodegradation but had no effect on Ni- and CoEDDS biodegradation (Figure 5 and Table 3). In a replicate experiment, CuEDDS complex was fully degraded in the presence of 5 mM phosphate after 11 days incubation (not shown). The interpretation of these results is likely to be similar to that

FIGURE 6. Effect of iron on biodegradation of 1 mM CuEDDS in 10-fold diluted activated sludge, preacclimated to NaEDDS. Iron was added either as 5 mM FeOOH or as 1 mM Fe(III)EDDS, at two pH values. for ZnEDDS since copper will combine with phosphate as soluble CuHPO4 and possibly as solid Cu3(PO4)2 or CuPO4OH. The speciation resulting from the formation of these complexes is not included here because the MINEQL+ version used here lacks the formation constants of most of these complexes. Phosphate was not expected to affect Niand CoEDDS degradation since these metals do not form stable solid phases with phosphate. Other metal ligands, such as colloidal FeOOH and iminodiacetate-functionalized resin beads, also fostered metal-EDDS degradation (Figure 5, Table 3). The strongest metal complexant (iminodiacetate) was also the one which accelerated metal-EDDS biodegradation most drastically. Given the relevance of iron in natural settings, its effect on Cu-, Co-, and NiEDDS degradation was further examined. Iron was added either as 5 mM FeSO4 (rapidly oxidized to FeOOH colloids) or as 1 mM Fe(III)EDDS at pH 6.2 and 8.8. CuEDDS was fully degraded with FeOOH at pH 8.8, while no significant degradation was detected at pH 6.2 (Figure 6). Sterile controls (not shown) ruled out the possibility that sorption of CuEDDS on sludge, rather than biodegradation, may have occurred. While other researchers have ascribed the stimulatory effect of iron on metal complexes’ biodegradation to the removal of toxic metal levels (22), such explanation was ruled out here since CuEDDS had been found innocuous in the first experiment. Another interpretation is that FeOOH exchanges with the complexed Cu or Co to form the biodegradable Fe(III)EDDS complex. Calculations with the relevant stability constants show, however, that such substitutions would not occur in the pH range used in our experiments. There remained therefore only the explanation that added iron raised the concentration of HEDDS3-, through the scavenging of released Cu2+ ions. The fact that iron was effective at pH 8.8 and not pH 6.2 further supports this model, since colloidal iron may bind aquo metals only at high pH values. Interestingly, iron added as Fe(III)EDDS did not foster CuEDDS degradation as FeSO4 did, even though Fe(III)EDDS was rapidly degraded (Figure 6). The reason might be that 5-fold less Fe(III)EDDS was added than FeSO4 or/and that the Fe(III) ions freed during Fe(III)EDDS degradation were complexed onto the bacterial envelope, rather than precipited as FeOOH colloids, resulting in lesser sorption capacity toward copper ions. Similar results were obtained with CoEDDS, with the difference that CoEDDS also disppeared at pH 6.2 (data not shown). Sterile controls indicated however that the disappearance of CoEDDS at pH 6.2 was due to sorption onto colloidal iron rather than biodegradation (not shown). On the other hand, NiEDDS was neither sorbed nor degraded in the presence of iron in the pH range 6-9. Biodegradation of Metal-EDDS Complexes in BAF Reactors. In soil washing operations where EDDS may be used, metal-EDDS complexes would need to be eliminated from wastewater streams. To this end, biological treatment

FIGURE 7. Concentration of total EDDS and total metal (Pb, Zn, or Cu) in the effluent of a BAF reactor fed with 10 mM solutions of metal-EDDS complexes (5 mM for CuEDDS) supplemented with 1-20 mM KH2PO4. of EDDS complexes-laden solutions was examined in biological aerated filter (BAF) reactors, since these retain precipitated metal colloids very effectively and, moreover, generate little waste biomass compared to activated sludge systems. EDDS (added as Na salt) was readily degraded in the BAF reactor, with 0.3 mM EDDS remaining in the effluent (Figure 7) indicating 97% biodegradation (0.4 mM h-1). The EDDS measurements were matched by COD measurements, fluctuating in the range 80-180 mg COD L-1, which indicated that practically all residual compounds in the effluent consisted of EDDS and that no intermediate biodegradation products accumulated. The reactor performance remained high when the EDDS form in the feed was switched to the PbEDDS (Figure 7). The concentration of Pb in the effluent remained consistently below the detection limit (0.5 mg L-1), indicating a removal efficiency of PbEDDS of 99.9+%. EDDS degradation was accompanied by a pH rise to 8.9, which is optimal for the precipitation of freed aquo metals. Continuous removal of ZnEDDS, on the other hand, was less successful. Upon switching to a feed solution containing 10 mM ZnEDDS, the concentration of ZnEDDS in the effluent rose rapidly, indicating a limited capacity for ZnEDDS degradation (Figure 7). However, as observed in the batch experiments (Figure 5), a 1-week acclimation period to ZnEDDS resulted in rapid degradation, with only 1 mM EDDS-Zn remaining in the effluent (ZnEDDS degradation rate of 0.36 mM h-1 or 110 mg L-1 h-1). This good performance was however short-lived as EDDS concentration in the effluent increased steadily to stabilize at 8.5 mM (15% removal). The reason was most likely sludge intoxication resulting from the accumulation within the reactor of the Zn removed from the liquid phase. Parallel to the decreasing removal efficiency of ZnEDDS, the effluent pH dropped from 8.9 to 6.9, further contributing to metal toxicity. Since the batch experiments had indicated that ZnEDDS biodegradation was drastically enhanced when a molar excess of phosphate salt relative to ZnEDDS was added (Figure 5), the concentration of K2HPO4 in the feed solution was raised from 2 to 20 mM (and the reactor was reinoculated with active NaEDDS-acclimated sludge). With such excess phosphate, 90+% ZnEDDS removal was consistently achieved during 1 week (Figure 7). The behavior of CuEDDS in the BAF reactor was similar to that of ZnEDDS. After a 1-week long adaptation period, ca. 80% biodegradation of CuEDDS could be achieved but only for a short time. Build up of Cu within the reactor, probably causing sludge intoxication, lead rapidly to a deterioration of reactor performance. It should be noted that the advent of metal toxicity in our BAF reactors was most likely exacerbated by the fact that the feed solutions contained no other carbon sources than metal-EDDS and by the fact that backwash cycles were carried out only about once every 2 weeks. It is expected that BAF reactors run under more realistic conditions in terms of Me/COD ratio in the VOL. 35, NO. 9, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1769

feed and, in terms of frequency of backwash cycles (e.g. once a day), should be much less sensitive to metal intoxication. To better simulate field conditions, the BAF reactors were fed with a mixture of Zn-, Pb-, and CuEDDS as polluted sites often contain more than one metal. The removal of the metal-EDDS complexes present as a mixture was similar to that observed with single metal complexes, i.e., 100% removal of PbEDDS, 75-95% removal of ZnEDDS, and 0-50% removal of CuEDDS (data not shown). These data show that in cases where poorly degradable metal-EDDS complexes need be biologically eliminated, a fine-tuning of reactor conditions in terms of pH, added ligands, and frequency of backwash is necessary. Another option is a two-stage anaerobic-aerobic treatment, where EDDS-complexed metals precipitate as sulfides in the anaerobic stage (data not yet published), leaving free EDDS readily amenable to aerobic biodegradation.

Acknowledgments The authors wish to thank D. C. McAvoy for his support with the MINEQL+ program.

Nomenclature b

decay coefficient (d-1)

BAF

biological aerated filter (Figure 1)

Bv

volumetric organic loading rate (g COD d-1 L-1 reactor volume)

COD

chemical oxygen demand (mg L-1)

EDDS

[S,S]-stereoisomer of ethylenediaminedisuccinic acid

EDTA

ethylenediaminetetraacetic acid

Km

substrate affinity constant (half-saturation value) (M)

Kst

chelate stability constant (eq 2)

Ka

acidity constant (eq 3)

MLSS

mixed liquor suspended solids

NTA

nitrilotriacetic acid

Vmax

maximum rate of chelate biodegradation (M d-1)

X

sludge concentration (g MLSS L-1)

Y

sludge yield coefficient (g MLSS g COD-1)

Literature Cited (1) No¨rtemann, B. Appl. Microbiol. Biotechnol. 1999, 51, 751-759. (2) Takahashi, R.; Fujimoto, N.; Suzuki, M.; Endo, T. Biosci. Biotech. Biochem. 1997, 61, 1957-1959.

1770

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 9, 2001

(3) Schowanek, D.; Feijtel, T. C. J.; Perkins, C. M.; Hartman, F. A.; Federle, T. W.; Larson, R. J. Chemosphere 1997, 34, 2375-2391. (4) Jaworska, J. S.; Schowanek, D.; Feijtel, T. C. J. Chemosphere 1999, 38, 3597-3625. (5) EPA. Cleaning excavated soil using extraction agents: a stateof-the-art review; EPA/600/2-89/034; U.S. EPA: Cincinnati, OH, 1999. (6) EPA. A literature review summary of metals extraction processes used to remove lead from soils; EPA/600/SR-94/006; U.S. EPA: Cincinnati, OH, 1999. (7) Peters, R. W. J. Hazard. Mat. 1999, 66, 151-210. (8) Witschel, M.; Egli, T. Biodeg. 1998, 8, 419-428. (9) Tiedje, J. M. Appl. Microbiol. 1975, 30, 327-329. (10) Madsen, E. L.; Alexander, M. Appl. Environ. Microbiol. 1985, 50, 342-349. (11) Satroutdinov, A. D.; Dedyukhina, E. G.; Chistyakova, T. I.; Witschel, M.; Minkevich, I. G.; Eroshin, V. K.; Egli, T. Environ. Sci. Technol. 2000, 34, 1715-1720. (12) Xun, L.; Reeder, R. B.; Plymale, A. E.; Girvin, D. C.; Bolton, H., Jr. Environ. Sci. Technol. 1996, 30, 1752-1755. (13) Bolton, H., Jr.; Girvin, D. C.; Xun, L. In Proceedings of the 4th International Symposium on In Situ and On-Site Bioremediation; April 28-May 1, New Orleans, LA, 1997; p 437. (14) Klu ¨ ner, T.; Hempel, D. C.; Nortemann, B. Appl. Microbiol. Biotechnol. 1998, 49, 194-201. (15) Witschel, M. Ph.D. Thesis, Swiss Federal Institute of Technology, Zu ¨ rich, Switzerland, 1999. (16) Witschel, M.; Egli, T.; Zehnder, A. J. B.; Wehrli, E.; Spycher, M. Microbiol. 1999, 145, 973-983. (17) Francis, A. J.; Dodge, C. J.; Gillow, J. B. Nature 1992, 356, 140142. (18) Bolton, H., Jr.; Girvin, D. C.; Plymale, A. E.; Harvey, S. D.; Workman, D. J. Environ. Sci. Technol. 1996, 30, 931-938. (19) Thomas, R. A. P.; Lawlor, K.; Bailey, M.; Macaskie, L. E. Appl. Environ. Microbiol. 1998, 64, 1319-1322. (20) VanBriesen, J. M.; Rittman, B. E. Biodeg. 1999, 10, 315-330. (21) Henneken, L.; No¨rtemann, B.; Hempel, D. C. J. Chem. Technol. Biotechnol. 1998, 73, 144-152. (22) Francis, A. J.; Joshi-Tope, G. A.; Dodge, C. J. Environ. Sci. Technol. 1996, 30, 562-568. (23) Bolton, H.; Girvin, D. C. Environ. Sci. Technol. 1996, 30, 20572065. (24) Schecher, W. D.; McAvoy, D. C. MINEQL+ - A chemical equilibrium program for personal computers. User’s manual, version 3.0; Environ. Res. Software: Hallowell, ME, 1994. (25) Stumm, W.; Morgan, J. J. Aquatic Chemistry, 3rd ed.; John Wiley & Sons: New York, 1999.

Received for review May 31, 2000. Revised manuscript received January 18, 2001. Accepted January 19, 2001.

ES0001153