Environ. Sci. Technol. 1999, 33, 3812-3820
Metal Removal from Contaminated Soil and Sediments by the Biosurfactant Surfactin C A T H E R I N E N . M U L L I G A N , * ,† RAYMOND N. YONG,‡ BERNARD F. GIBBS,§ SUSAN JAMES,| AND H. P. J. BENNETT| Department of Building, Civil and Environmental Engineering, Concordia University, 1455 de Maisonneuve Boulevard W, Montreal, Quebec Canada H3G 1M8, Geoenvironmental Engineering Research Centre, Geoenvironmental Engineering Research Centre, Cardiff School of Engineering, University of Wales, P.O. Box 917, Newport Road, Cardiff, CF2 1XH, United Kingdom, Department of Food Science and Agricultural Chemistry, McGill University, St. Anne de Bellevue, Quebec, Canada H9X 1C0, and Department of Medicine, McGill University, 687 Pine Avenue W, Montreal, Quebec, Canada H3A 1A1
Batch soil washing experiments were performed to evaluate the feasibility of using surfactin from Bacillus subtilis, a lipopeptide biosurfactant, for the removal of heavy metals from a contaminated soil and sediments. The soil contained high levels of metals and hydrocarbons (890 mg/ kg of zinc, 420 mg/kg of copper, and 12.6% oil and grease), and the sediments contained 110 mg/kg of copper and 3300 mg/kg of zinc. The contaminated soil was spiked to increase the levels of copper, zinc, and cadmium to 550, 1200, and 2000 mg/kg, respectively. Water alone removed minimal amounts of copper and zinc (less than 1%). Results showed that 0.25% surfactin/1% NaOH could remove 25% of the copper and 6% of the zinc from the soil and 15% of the copper and 6% of the zinc from the sediments. A series of five washings of the soil with 0.25% surfactin (1% NaOH) was able to remove 70% of the copper and 22% of the zinc. The technique of ultrafiltration and the measurement of octanol-water partitioning and ζ-potential were used to determine the mechanism of metal removal by surfactin. It was indicated that surfactin was able to remove the metals by sorption at the soil interphase and metal complexation, followed by desorption of the metal through interfacial tension lowering and fluid forces and finally complexation of the metal with the micelles.
Introduction In the United States, 1200 sites are on the National Priority (Superfund) List for the treatment of contaminated soils, indicating the extent of this problem. Toxic heavy metals are also frequently encountered at hazardous waste sites (1). For example, lead was found at 15% of the sites, followed by * Corresponding author telephone: (514)848-7925; fax: (514)8482809; e-mail: [email protected]
† Concordia University. ‡ University of Wales. § Department of Food Science and Agricultural Chemistry, McGill University. | Department of Medicine, McGill University. 3812
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 21, 1999
chromium, cadmium, and copper at 11, 8, and 7% of the sites, respectively. At Superfund sites with signed Records of Decision (ROD), metals are the only contaminants at 16% of the sites, whereas metals and organic compounds (volatile or semivolatile) are found together at 49% of the sites (2). Therefore, organic and metal contamination is a major concern. Surfactants can be added to washing water to assist in the solubilization, dispersal, and desorption of contaminants from excavated soils in a washing unit for subsequent return of the cleaned soils to their original site. They can also be used during pump-and-treat procedures where the groundwater is recovered for ground-level treatment after in situ flushing of the contaminated soil with a surfactant-containing solution. Several synthetic surfactants have been evaluated in soil decontamination tests (1, 3, 4). Although mechanisms for organic removal by surfactants have been investigated (5, 6), little work has been done regarding metal removal. For organics, solubilization can be represented by the term, Km, which relates the partitioning of the organic compound in the micelle phase to the aqueous phase by the following equation:
where Km is the micelle-water partitioning coefficient that includes Xmic ) MSR/(1 + MSR), which is the mole fraction of total organic compound in the micelle and where MSR is the mole of organic compound in the micelle/mole of surfactant or molar solubilization ratio, and Xaq is the mole fraction in the aqueous phase (i.e., the concentration of the organic in the permeate times the molar volume of water, 0.01805 L/mol) (6). In this work, this approach was applied to metal associated with the micelle. This research focuses on the feasibility of using the anionic biosurfactant, surfactin, to enhance the removal of heavy metals (zinc and copper) from soil and sediments. Surfactin produced by Bacillus subtilis is one of the most effective biosurfactants known. It reduces the surface tension of water from 72 to 27 mN/m at a concentration as low as 0.005% (7). It is a lipopeptide containing seven amino acids ( four leucine, one aspartic acid, one glutamic acid, and one valine residues) bonded to the carboxy and hydroxyl group of a 14-carbon acid (8). The potential advantages of using surfactin include the presence of two charges due to glutamic and aspartic amino acids as part of its peptide structure (9), its biodegradability, its effectiveness as a surfactant (low surface tension and critical micelle concentration (cmc) values), and its potential for in situ production. The technique of ultrafiltration and measurements of octanol-water partitioning (Kow) and ζ-potential were used to determine the mechanism of metal removal by surfactin. These results are presented in this paper in addition to the relationship between Km and Kow.
Experimental Methods Materials. Surfactin from B. subtilis ATCC 21332 was produced in a 1-L fermentor containing 4% glucose and mineral salts medium and isolated using previously described techniques (9). A temperature of 37 °C, an aeration rate of 0.5 vvm, and pH control at 6.7 were used as the bacterial growth conditions. Surfactin was collected in the foam overflow and then purified by removing the bacterial cells by centrifugation (12000g, 10 min), adding concentrated 10.1021/es9813055 CCC: $18.00
1999 American Chemical Society Published on Web 09/28/1999
TABLE 1. Sequential Extraction Characterization of Metal Contaminants in Spiked Soil and Sediment Samples fraction (% of total) sample and metal exchangeable oxide carbonate organic residual copper zinc lead cadmium copper zinc lead a
NDa ND 5 10 ND 5 2
Spiked Soil 5 70 55 60
10 10 4 25
Sediment 2 ND 50 25 70 3
75 10 20 5
15 5 20 ND
88 18 10
10 2 15
ND, not detected.
hydrochloric acid to adjust the pH to 2, and precipitating the surfactin. Dichloromethane was subsequently added three times, and the top organic layer was collected each time, pooled, and evaporated. The residue was then redissolved in basic water (pH 8) and filtered to remove impurities (Whatman No. 1 paper). Its purity and composition were determined by amino acid analysis (9). It was stored as a powder at 4 °C. All other chemicals were laboratory grade. The contaminated matrixes, used in this study, have been previously characterized using standard methods (10) and a modified procedure for cation exchange capacity (11). Selective sequential extraction procedures were also performed according to the procedures described previously (12) on the contaminated matrixes to determine the speciation of the metal contaminants. These values are indicated in Table 1. The contaminated soil matrix was obtained from a harbor area containing storage tanks, refineries, oil terminals, and coal storage and processing industries that released fly ash and those involved in metal refinishing. This matrix was chosen to determine the feasibility of using this biosurfactant for simultaneous oil and metal removal. The organic content was 20.7%, and the oil and grease content was 12.6%. The particle size distribution was 10% silt and 90% sand, and the cation exchange capacity was 7.1 mequiv/ 100 g at pH 7. The concentrations of copper, zinc, lead, and cadmium were 420, 890, 320, and 3 mg/kg, respectively. The sediment sample was obtained from a canal area that had been contaminated with metals and some hydrocarbons by many years of industrial discharges into the canal. The organic matter content was 13.4%, the cation exchange capacity was 17.1 mequiv/100 g (pH 7), and the particle size distribution was 10% sand, 70% silt, and 20% clay. This sediment with high organic content was chosen for study since the organic fraction is a very important component in metal retention for sediments and topsoils. The concentrations of copper, zinc, lead, and cadmium were 110, 3300, 410, and 5 mg/kg, respectively. Soil Spiking Procedure. The soil sample was spiked with a solution containing 1 mM each of Cu(SO4)2, CdCl2, Pb(NO3)2, and Zn(SO4)2. The spiking was done to increase the concentration of cadmium in the soil, which was very low, in addition to further increasing the concentrations of copper and zinc. The sediment sample was not spiked. A ratio of 1 g of soil/16 mL of solution was used. The soil was shaken for 3 days on the rotary shaker then removed by centrifugation (5000g, 20 min) and air-dried. The final spiked soil sample used for extraction tests contained a high concentration of cadmium (2000 mg/kg), followed by zinc (1200 mg/kg) and copper (550 mg/kg). Surface Tension, Critical Micelle Concentration, and Conductivity Determinations. A Fisher Tensiomat model 21 was used to measure surface tension and interfacial tensions by the du Nouy method. The cmc was determined
by measuring the surface tension at various dilutions (13). The cmc was taken as the point at which the surface tension abruptly increased when the logarithm of the dilution is plotted as a function of surface tension. The reciprocal of cmc was used as an indication of relative concentrations. The conductivity of the surfactin solutions was measured using a Cardy pocket conductivity meter model C-173 calibrated with 0.1 N KCl. Interfacial tensions were measured by the tensiometer by submerging the ring in the surfactant solution and then adding a 1 cm depth of oil extracted with hexane from the soil sample. The ring was then pulled through the oil-water interphase until the ring broke the surface (ASTM D971 method). Surfactin Concentration Determination. A method based on amino acid analysis was used for determining surfactin concentration (13). Briefly, an aliquot of surfactin solutions was dried and acid hydrolyzed for 1.5 h at 150 °C in a PICOTAG amino acid analysis system. The residue was redissolved in sodium buffer and injected on a Beckman System 6300 high-performance analyzer equipped with a Beckman model 7000 data station (Palo Alto, CA). All buffer and ninhydrin reagents were purchased from Beckman. The concentration of surfactin was calculated by multiplying the lipopeptide concentration by the molecular weight (1036 amu). Ultrafiltration Procedures. Ultrafiltration experiments were performed using a 50-mL Amicon magnetically stirred ultrafiltration cell containing an XM 50 membrane as described in Mulligan and Gibbs (14). Copper (10 mg/L), lead (0.3 mg/L), cadmium (20 mg/L), and zinc (10 mg/L) were added in the form of (PbNO3)2, (CuSO4)2, Zn(SO4)2, and CdCl2 to the various concentrations of surfactant prior to the ultrafiltration procedure. Volumes of 25 or 50 mL of solution were added to the cell and pressurized to 60 psi. The retentates and permeates were collected; the volumes of each were recorded and then subjected to metal analysis. Rejection ratio (R) is defined as
Cf Co R) × 100% Vo ln Vf ln
where Cf is the final metal concentration in the retentate, Co is the initial metal concentration, Vo is the initial sample volume, and Vf is the final retentate volume (Amicon Laboratory Separation Guide, 1988). Octanol-Water Partitioning Studies. Octanol studies were performed using a modified procedure of Friedel et al. (15) by adding 3 mL of octanol/10 mL of metal/surfactant solution in centrifuge tubes that were shaken overnight. The metals [in the form of (PbNO3)2, (CuSO4)2, Zn(SO4)2, and CdCl2] were at concentrations of copper (5 mg/L), zinc (4.8 mg/L), lead (0.6 mg/L), and cadmium (11.0 mg/L). The octanol and water fractions were separated by centrifugation, followed by atomic absorption spectrometric analysis of the water-soluble fraction using a Perkin-Elmer model 3110 atomic absorption spectrometer according to standard methods (10). ζ-Potential Measurements. The ζ-potential was measured with the aid of a Zeta-meter model ZM-75 (Zeta-Meter Inc., New York, NY). Approximately 0.02 g (presieved through a 200 mesh sieve) of soil or sediment was added to 25 mL of solution prior to measurement. This ratio was used to enable tracking of the particles during measurement. Particle concentrations higher than this level causes interference between the particles and makes tracking of the particles impossible. The ζ-potential was calculated from the electrophoretic mobility, Be ) ur/E, which is the ratio of the migration velocity of the particles to the field intensity applied, VOL. 33, NO. 21, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
according to the equation ζ ) ur/E × η/or, where r is the absolute dielectric constant, o is the influence constant, and η is the viscosity. The dielectric constant and viscosity for water at 20 °C were used for all measurements. Procedure for Batch Soil Washing Studies. All batch soil washing studies were performed by varying surfactant concentrations and pH values in centrifuge tubes while maintaining a constant 10:1 wt/wt solution to soil ratios (15 g/1.5 g) (1). Distilled water alone was used to account for removal of contaminants by physical mixing. Controls included the same additives (either hydrochloric acid or sodium hydroxide) as for the biosurfactant studies. Samples were taken after 24 h shaking on a reciprocating shaker (60 oscillations/min) and then centrifuged (5000g, 10 min). The supernatant was then analyzed for metal concentration by atomic absorption spectrometry. The percentage of metal removal was determined based on the initial metal content in the soil or sediment, and all results are presented as percentage of metal removal. All experiments were performed in triplicate, and the average of the results is presented.
Results and Discussions Characterization of Surfactin. The purity of the surfactin powder was determined to be 62%. In addition, since the ratio of 4:1:1:1 for leucine:aspartic acid:valine:glutamic acid was determined by amino acid analysis, this confirms the structure of surfactin determined by Kakinuma et al. (8). The surface active properties of surfactin were evaluated. A minimum surface tension of 31 mN/m was obtained for surfactin concentrations greater than 0.05 g/L, which is slightly higher than the value of 27 mN/m previously reported (14). The cmc value was 0.02 g/L, which is between the values of 0.011 (14) and 0.025 g/L previously reported (13). Equivalent conductivity determinations decreased significantly at the cmc and reached a plateau at a surfactin concentration of approximately 0.1 g/L. This decrease is due to the hidden charged sites in the larger micelles as compared to the monomers. Addition of 1% NaOH to the surfactin solution increased the surface tension to 38 mN/m but did not affect the cmc significantly (0.02 g/L). The minimum interfacial value measured by the surfactin solutions against the oil isolated from the soil was reduced to approximately 1 mN/m from 10 mN/m for water against the same oil. Batch Soil Washing Studies. To solubilize the surfactin powder, the pH was increased to 8.0 by addition of NaOH (to give a final concentration of approximately 0.2% NaOH). The effect of varying the surfactin concentration on metal removal efficiency during batch washing was determined to be between 0.13 and 4.0% surfactin. Without any other additives, only 2% and 4% surfactin showed any removal of copper (0.8 and 4.1%, respectively) or zinc (1.3 and 2.9%, respectively) from the soil. However, the effect of the addition of 1% NaOH to surfactin at concentrations between 0.13% and 4% (pH 10 after addition) on copper and zinc removal was also determined (Table 2). Copper removal significantly increased (Table 2) in comparison to 1% NaOH alone (where no copper removal was noted) even though the pH was the same. Copper removal increased as the surfactin concentration increased to 0.25%, decreased until 1% concentration, and then increased again as the concentration increased to 4%. Zinc removal followed the same trend with the exception of a decrease at 4% surfactin as compared to the experiment with 2%. Surfactin, thus, is responsible for enhanced metal removal. Triton X-100, a nonionic surfactant (not shown), was also evaluated at concentrations up to 2% and showed slightly better removal than water or NaOH alone (3.5% copper removal and insignificant zinc removal). For the sediment (Table 2), 0.25% and 2% surfactin were evaluated with and without 1% NaOH addition for their abilities to remove heavy metals from this matrix. Controls 3814
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 21, 1999
TABLE 2. Batch Washing Experiments with Contaminated Soil and Sediments To Determine Copper and Zinc Removal by Surfactin surfactin concn for different matrixes
copper removal (%) with SDa
zinc removal (%) with SDa
1% NaOH 0.13% + 1% NaOH 0.25% +1% NaOH 0.50% + 1% NaOH 1.0% + 1% NaOH 2.0% + 1% NaOH 4.0% + 1% NaOH
Contaminated Soil 0.0 ( 0 1.4 ( 0.2 24.8 ( 0.2 5.9 ( 0.4 3.6 ( 0.2 11.2 ( 0.9 15.6 ( 1.1
0.9 ( 0.1 1.1 ( 0.3 6.0 ( 0.1 3.7 ( 0.2 1.8 ( 0.3 4.8 ( 0.4 3.1 ( 0.2
control 1% NaOH 0.25% 0.25% + 1% NaOH 2.0% 2.0% + 1% NaOH
Sediment 0.2 ( 0.1 10.0 ( 0.9 0.5 ( 0.1 15.0 ( 1.2 0.3 ( 0.1 15.0 ( 0.9
1.1 ( 0.2 3.0 ( 0.2 0.9 ( 0.1 6.0 ( 0.3 15.5 ( 0.7 1.0 ( 0.6
Experiments were performed in triplicate.
of water (pH 8) and 1% NaOH were used. Significant zinc removal was obtained at 2% surfactin (pH 8.0) while water removed minimal amounts. Concentrations of 0.25% and 2% surfactin both with 1% NaOH (pH 10.0) removed equal and significantly higher amounts of copper than 1% NaOH alone. Higher levels of lead were removed in the presence of surfactin and 1% NaOH than the control while 2% surfactin gave better results than 0.25% surfactin. A series of five washes was performed on the spiked soil using 0.25% surfactin and 1% NaOH by removing the washing solution each day and replacing it with a fresh solution. Copper, zinc, and cadmium removal (Figure 1) were followed each day for 5 days. The control was 1% NaOH. For copper, the control showed a final cumulative removal of 20%, while approximately 70% was removed by the surfactin. For zinc, removal rates were lower than for copper, 10% for the control and 25% for surfactin. The control showed minimal cadmium removal, while surfactin removed approximately a total of 15%. Multiple washes were thus highly beneficial for copper with less benefit for zinc and cadmium. In addition, highest removal rates were shown before 3 days for all three metals. Surfactin concentrations were analyzed in the supernatant after the soil washing experiments as shown in Table 3 to determine the losses of the surfactin by adsorption. The concentration of surfactin after washing decreased in all cases. The co-addition of NaOH decreased adsorption at 0.25% surfactin but not at the higher concentration. For the studies with multiple extractions, the decrease in surfactin concentration was most significant after the first wash and was negligible thereafter. This is probably due to initial adsorption and then saturation of surfactant on the surface of the soil. Precipitation is not likely since the pH of the soil washing experiments was maintained above 5, below which noticeable precipitation can occur. In addition, in the case of precipitation of surfactin, a white solid on the soil after centrifugation following soil washing experiments can be noticed. This was not the case in any of these experiments. Metal retention is controlled by various phenomenon. They include ion exchange and association with iron and manganese oxides, carbonates, and organic matter. Metals can be immobilized by precipitation and incorporation into the crystalline structure of the soil. From the characterization of the matrixes by selective sequential extraction, it was determined that the main fractions to retain zinc, lead, and cadmium in both the soil and sediment samples were oxides. Cameron (16) has indicated that zinc, in particular, is often associated with iron and manganese oxides. Retention by
TABLE 3. Surfactin Concentrations in the Supernatant before and after Washing the Soil or Sediment initial surfactin concn (%) Soil
final surfactin concn (%) after washing with SDa
2.0 0.25 2.0 0.25