| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Department of Environmental Science, Kasetsart University, Bangkok, Thailand 10900
b School of Natural Resources, University of Nebraska-Lincoln, 256 Keim Hall, Lincoln, NE 68583-0915
* Corresponding author (scomfort{at}unl.edu).
Received for publication September 19, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: HPLC, high performance liquid chromatography • GR, green rusts • SEM, scanning electron microscopy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.03 µg L-1 (Holden and Graham, 1992). Surface waters have also been affected by metolachlor in at least 14 states, with maximum concentrations of 138 µg L-1 (USEPA, 1987). While attention to agrichemical use and the environment has centered on mitigating nitrate contamination, this focus may shift, or at least broaden as more state and federal regulators become attuned to problems associated with pesticide use and can recommend or endorse reliable remediation treatments for pesticide-contaminated soil and water. With more than 50 permeable reactive barriers (PRBs) installed worldwide (Remediation Technologies Development Forum, 2002), zerovalent iron (Fe0) is becoming recognized as an efficient and cost-effective means of removing a variety of contaminants from ground water. Early studies on Fe0 use focused on chlorinated aliphatics but more recent reports have shown promising results for the treatment of munitions wastes (Agrawal and Tratnyek, 1996; Singh et al., 1998a, 1999), nitrate (Till et al., 1998; Zawaideh and Zhang, 1998; Huang et al., 2003), and metals (Blowes et al., 1997; Pratt et al., 1997; Fiedor et al., 1998). Favorable results with alachlor[2-chloro-N-(2,6-diethylphenyl)-N-(methoxymethyl)acetamide], metolachlor, and atrazine [6-chloro-N-ethyl-N'-(1-methylethyl)-1,3,5-triazine-2,4-diamine] (Eykholt and Davenport, 1998; Singh et al., 1998b; Comfort et al., 2001; Gaber et al., 2002) also support the use of Fe0 as a remedial treatment for pesticide-contaminated soil and water. Though Fe0–related research has elucidated the degradation kinetics and pathways for several contaminants, many aspects of the process are unclear. To effectively manipulate the Fe0–soil–water system for contaminant destruction, Fe0 corrosion and contaminant destruction rates must be characterized in the presence of various electrolytes and under a variety of pH and redox conditions. Moreover, understanding conditions that modify the properties of the iron surface and create secondary reductants, such as surface-bound Fe(II), could lead to more effective treatments.
It is well established that Fe0 oxidation under aerobic conditions results in precipitation of Fe(III) hydroxides, while mixed Fe(II)–Fe(III) phases, including green rust and magnetite (Fe3O4), are formed under anaerobic conditions. Depending on the characteristics of the oxides formed, varying levels of electrical resistance to electron transfer may occur. If the Fe0 surface becomes passivated with Fe(III) (oxy)hydroxides, electron transfer from the iron core will cease while the presence of mixed-valent oxides may allow contaminant destruction to continue. Further compounding effects could result from limitations related to reactive surface area (Alowitz and Scherer, 2002) and contaminant diffusion through the porous oxide coating (Burris et al., 1998).
Under anoxic conditions, green rust compounds have been observed in permeable reactive barriers (Phillips et al., 2000). Green rusts are categorized as layered double hydroxides (Taylor, 1980) with interlayer anions that impart a greenish-blue color. From X-ray diffraction characterization, two types of green rusts have been identified: green rust one (GRI), containing planar interlayer anions to balance the charge of the hydroxide layers (e.g., Cl- and CO2-3), and green rust two (GRII), containing three-dimensional anions 
in the interlayers (Génin et al., 1998). Interest in green rusts has increased because of their ability to promote the reduction of Cr(VI) (Loyaux-Lawniczak et al., 2000; Williams and Scherer, 2001), nitrate (Hansen et al., 1996; Hansen and Koch, 1998), nitrite (Hansen et al., 1994), selenate (Myneni et al., 1997), and some halogenated ethanes and methanes (Erbs et al., 1999; O'Loughlin and Burris, 2000).
Green rusts are known to occur naturally in soils and sediments and can be stable under certain conditions. Génin et al. (1998) identified an OH form of GR in a hydromorphic soil and demonstrated that this mineral was responsible for controlling the Fe concentration of the surrounding soil solution. Green rusts have also been artificially induced in soils. In a recent field-scale demonstration project, we treated metolachlor-contaminated soil with Fe0 in static soil windrows (Comfort et al., 2001). After 90 d, soils in the middle of windrows treated with Fe0 + Al2(SO4)3 exhibited a green rust color and had metolachlor concentrations that were significantly lower (
= 0.05) than soils treated with Fe0 alone (40 vs. 504 mg kg-1; Comfort et al., 2001). Considering green rusts rapidly oxidize when exposed to atmospheric oxygen, our observation indicates conditions favorable to green rust formation were created and maintained under field conditions.
Our objective was to determine the effects of aluminum and iron salts on metolachlor destruction kinetics by zerovalent iron. We also compared two iron sources that differed in surface oxide coatings (annealed and unannealed) and used a pH-stat to determine the optimum pH for iron-mediated metolachlor destruction.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
We used two sources of Fe0 in our batch experiments: annealed and unannealed iron (Peerless Metal Powders, Detroit, MI). Annealing is a manufacturing process that indirectly heats the iron under a hydrogen–nitrogen atmosphere. Specific surface areas of the materials were 0.134 m2 g-1 (annealed Fe0) and 2.55 m2 g-1 (unannealed Fe0) (Micromeritics, Norcross, GA). The surface morphology of the iron sources was compared by mounting with carbon tabs, sputter-coating with gold-palladium, and observing with a Hitachi (Tokyo, Japan) S-3000N scanning electron microscope operated at 15 kV.
Batch Experiments
Batch experiments determined the capacity of Fe0 to transform metolachlor. The initial metolachlor concentration varied between 300 and 400 mg L-1 as determined by comparison with high-purity standards (Syngenta). Unless indicated, batch procedures included treating 100 mL of aqueous metolachlor in 250-mL Erlenmeyer flasks. The metolachlor solutions were treated with either 5 g or 12.5 g of Fe0. Flasks were covered with Parafilm M (American National Can, Chicago, IL) and agitated on an orbital shaker. At preselected times, multiple 1.2-mL aliquots were removed and transferred to 1.5-mL polypropylene microcentrifuge tubes, centrifuged at 13 000 x g for 10 min, and analyzed by high performance liquid chromatography (HPLC).
Effect of Aluminum and Iron Salts and Release of Iron(II)
Aqueous metolachlor (350 mg L-1) was treated with 12.5 g of annealed Fe0 and equal masses (0.5 g) of FeCl3, AlCl3, Al2(SO4)3, or Fe2(SO4)3. Treatments were conducted in triplicate.
An additional batch experiment was performed to measured changes in Al and Fe(II) concentrations following iron treatment of metolachlor [with and without Al2(SO4)3] under aerobic and anaerobic conditions. Metolachlor solutions (100 mL, 400 mg L-1) were treated with 12.5 g unannealed iron and 0.5 g Al2(SO4)3. The anaerobic experiment was conducted in an anaerobic chamber (Coy Laboratory Products, Grass Lake, MI) equipped with a reciprocal shaker. Metolachlor and dechlorinated metolachlor [(2'-ethyl-6'-methyl-N-(methoxyprop-2-yl)acetamide] were measured in samples collected at 2, 4, 6, 8, 10, 12, 24, and 48 h. Concentrations of Al and Fe(II) and pH were also measured.
Color Changes and Scanning Electron Microscopy
To record color changes imposed by the iron treatments, a batch experiment was conducted in 125-mL clear glass tubes using a lower Fe0 concentration (5% annealed, w/v) in combination with FeSO4 (1%, w/v) and Al2(SO4)3 (0.5%, w/v). The lower Fe0 concentration slowed down metolachlor destruction kinetics and allowed color changes to be recorded. Samples were obtained at 2, 4, 6, 8, 10, 12, and 24 h for metolachlor analysis and pH determination. Experimental units were photographed at each sampling. Following the 2 h sampling, samples of the aqueous solution and iron from the Fe0 + FeSO4 + Al2(SO4)3 treatment were quickly vacuum-filtered (Nuclepore 0.22-µm filter; Whatman, Maidstone, UK), sputter-coated as previously described, and immediately scanned with a Cambridge Steroscan (Cambridge, UK) Model 90 scanning electron microscope operated at 15 kV.
Iron Sources and Changes in Redox Potential and pH
We repeated the color changes and scanning electron microscopy (SEM) experiment (above) but expanded the treatment protocol to compare iron sources (annealed vs. unannealed). In addition, we monitored changes in redox potential (Eh) and pH in the annealed iron treatments. A combination redox probe was used to monitor temporal changes in Eh. Redox measurements were converted to a standard hydrogen electrode (SHE) reference by adding 200 mV to observed values (Light, 1972). A redox standard solution [39.21 g Fe(NH4)2(SO4)2·6H2O, 48.22 g FeNH4(SO4)2·12H2O, and 56.2 mL concentrated H2SO4 dissolved in 1 L H2O] was used to periodically check probe performance (Light, 1972). This solution has an Eh of +476 mV when measured with a Ag–AgCl reference electrode and saturated KCl fill solution.
pH-Stat Experiments
Because of fluctuations in pH during Fe0 treatment, we used a pH-stat (Metrohm Titrino 718S; Brinkman Instruments, Westbury, NY) to measure and control pH within the Fe0–metolachlor matrix. A single treatment consisting of 5% (w/v) Fe0 (annealed and unannealed) with FeSO4 (1%, w/v) and Al2(SO4)3 (0.5%, w/v) was used. Metolachlor–Fe0 suspensions were maintained at pH 3, 4, and 5. A similar experiment was conducted using soil slurries with pH maintained at 3, 4, 5, and 6. Soil slurries were prepared by mixing 60 g of metolachlor-contaminated soil (Comfort et al., 2001) with 240 mL H2O and equilibrating 12 h before treating with Fe0.
Chemical Analyses
Metolachlor analysis was performed with HPLC by injecting 20 µL of sample into a 4.6- by 250-mm Keystone Betasil NA column (ThermoHypersil-Keystone, Bellefonte, PA) connected to a Shimadzu (Kyoto, Japan) photodiode array detector with quantification at 220 nm. The mobile phase was 50:50 CH3CN and water at a flow rate of 1 mL min-1. Typical retention times were 12 min for metolachlor and 8 min for dechlorinated metolachlor.
Aluminum concentrations were determined by the method of Barnhisel and Bertsch (1982). In brief, temporal changes in Al concentrations were monitored by removing 0.1-mL samples from the experimental units and adjusting to 6 mL with deionized H2O in 25-mL glass tubes. Two milliliters each of 1 M HCl and 0.5% (v/v) ascorbic acid were added. This solution was heated to 75°C for 30 min and transferred to a 50-mL volumetric flask. Solutions were then brought to volume (50 mL) with 10 mL of aluminon-acetate buffer (120 mL glacial acetic acid, 24 g NaOH, and 0.35 g aluminon diluted to 1 L with H2O) and deionized water. Color intensity was developed for 2 h and quantified spectrophotometrically at 530 nm.
Iron(II) concentrations were determined colorimetrically with a DX120 ion chromatograph (Dionex, Sunnyvale, CA) using a post-column reagent and visible wavelength detection. Samples were injected into an IonPac CG5A 4-mm separating column with a MetPac PDCA (Dionex) eluent at 1.2 mL min-1. Post-column eluent was mixed with a MetPac PAC post column reagent at 0.6 mL min-1 and passed through a Dionex AD25 UV-VIS detector.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
There are several reasons why the properties of an Al-substituted Fe oxide over an underlying Fe0 core may be favorable for transformations of organic contaminants. Because of its smaller ionic radius, isomorphous substitution of Al3+ for Fe3+ in iron oxides disrupts crystallization and results in a larger surface area of the total oxide mineral (Schultz and Schwertmann, 1984), which would increase adsorption. Exchangeable aluminum also increases Brønsted acidity by promoting reaction with water to release H+ ions. Additionally, adsorbed Al can act as a Lewis acid by coordinating the moieties of some organic contaminants, bringing them closer to the iron oxide surface for reductive transformations. Other possible reactions include mineral-catalyzed hydrolysis and oxidation; both of these reactions involve complexation with surface Al(III) (McBride, 1994).
Although the standard protocol for Fe0–batch experiments typically employs anaerobic conditions, we observed faster metolachlor destruction under aerobic (i.e., outside the anaerobic glove box; k = 1.034 h-1) than anaerobic conditions (k = 0.326 h-1) (Fig. 2). This occurred even though the pH was higher (counterintuitive to reductive transformations) and Fe(II) concentrations were lower (Fig. 2). Visual observation of the anaerobic and aerobic treatments showed that the aerobic treatment displayed the greenish-blue color of green rust while the anaerobic treatment remained gray throughout the experiment. Because green rusts are mixed-valent Fe(II)–Fe(III) hydroxides, some Fe(III) must be available for green rust to form (Taylor, 1980; Taylor and McKenzie, 1980). Ferric iron would not readily form within the time frame of our experiments and under the atmospheric and pH conditions imposed by an anaerobic chamber and this probably explains why slower metolachlor destruction occurred under anaerobic conditions.
Color Changes and Scanning Electron Microscopy
Temporal changes in the color of metolachlor solutions treated with Fe0 and salt-amended Fe0 were observed during the first 12 h of treatment (Fig. 3)
. When FeSO4 or Al2(SO4)3 was added with the Fe0, solutions turned green and metolachlor transformation was rapid. After 2 h, the salt-amended Fe0 treatments began turning yellowish brown, coinciding with a decrease in pH (Fig. 3).
|
 | [1] |
 | [2] |
 | [3] |
A color change (green to yellow-brown) in the salt-amended treatments after 2 h (Fig. 3) signified the presence of more Fe(III) ions and formation of ferric oxyhydroxides. A common oxidation product of the sulfate form of GR is 
-lepidocrocite:
 | [4] |
Because our experimental conditions differed from procedures commonly used to synthesize GR, it is likely that a mixture of GR and ferric oxyhydroxides initially formed. With time, lepidocrocite and other mineral phases such as goethite (-FeOOH), akaganeite (ß-FeOOH), hematite (-Fe2O3), and maghemite (-Fe2O3) were also observed by X-ray diffraction (XRD) (Satapanajaru, 2002).
The SEM photos show magnification of the green suspension and crystal growth on the iron surface after 2 h in the Fe0 + FeSO4 + Al2(SO4)3 treatment (Fig. 4) . As mentioned earlier, high salt concentrations can favor the diffusion of reaction products away from the iron surface and promote precipitate formation in the bulk solution (Farrell et al., 2000). We observed oval colloidal particles in the suspension with some displaying hexagonal edges (Fig. 4A). Although green rusts exhibit hexagonal crystals, we do not have XRD data on the colloidal precipitates to confirm its presence. The iron surfaces (Fig. 4B) also displayed crystal growth representative of green rusts (Hansen et al., 1996; Hansen and Koch, 1998; Phillips et al., 2000) but XRD analyses could not positively identify GR(II) in the Fe0 + FeSO4 + Al2(SO4)3 treatment. GR(II) was identified, however, when only Fe0 and FeSO4 were used to treat metolachlor (Satapanajaru, 2002). Roh et al. (2000) also reported difficulty in identifying green rust by XRD when only trace amounts were present or the green rust exhibited a pseudo-hexagonal form.
|
 | [5] |
 | [6] |
Based on HPLC chromatograms from the annealed iron treatment (no dechlorinated metolachlor peak) and previous work using 14C-labeled metolachlor (Gaber et al., 2002), the initial small loss of metolachlor can be attributed to adsorption (Fig. 3). Between 8 and 12 h, however, dechlorinated metolachlor was produced and coincided with the formation of magnetite.
Large differences in destruction kinetics were observed among the three treatments (Fig. 3). The salt-amended iron treatments followed first-order destruction kinetics whereas a 6-h lag was observed in the annealed Fe0 only (control) treatment followed by zero-order kinetics (Fig. 3). Several researchers have observed a similar lag in destruction kinetics at high pH (>8.0), in particular for the treatment of nitrate (Siantar et al., 1996; Huang et al., 1998; Schlicker et al., 2000) and hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX) (Singh et al., 1998a). Alowitz and Scherer (2002) explained this lag period as necessary for: (i) abrasion of the Fe0 surface through mechanical scratching; (ii) equilibration of the iron source with the solution; or (iii) formation of secondary reductants [Fe(II) or Fe(II)-containing oxides and hydroxides] on the Fe0 surface. Our visual observations (i.e., formation of black precipitates at 8 to 12 h, Fig. 3) suggest that the latter mechanism probably contributed to delayed dechlorination of metolachlor in the control (annealed Fe0 only) treatment.
Iron Source and Changes in Redox Potential and pH
Additional batch experiments were conducted to compare iron sources (annealed and unannealed iron) under the same treatments used to record color changes. We also monitored changes in Eh and pH to help explain the effects of Fe or Al salt additions on metolachlor destruction kinetics (Fig. 5)
. Although our initial intent was to determine the effects of annealing, spectral differences in the oxide and carbon regions confirmed that the annealed and unannealed iron probably originated from different sources. That is, the annealed iron was not a heat-treated unannealed iron (Satapanajaru et al., 2001). Nevertheless, SEM photos revealed a sharp contrast between the smooth surface of the annealed iron, which probably resulted from the high temperature treatment of the manufacturing process, and the rough, pitted surface of the unannealed iron (Fig. 6)
. These morphological and mineralogical (Satapanajaru et al., 2001) differences were manifested by differences in metolachlor destruction rates. When used alone, unannealed Fe0 resulted in initial dechlorination of metolachlor whereas, as previously indicated, annealed Fe0 did not dechlorinate metolachlor until magnetite formed (Fig. 5). This difference can be attributed in part to the lower surface area of the annealed iron (0.134 vs. 2.55 m2 g-1 for unannealed Fe0) and its resistance to corrosion in aqueous solutions (Gaber et al., 2002).
|
|
When FeSO4 or Al2(SO4)3 was added, however, the annealed Fe0 produced faster destruction kinetics than unannealed iron (Fig. 5). By adding FeSO4, the Eh of the metolachlor system was lowered further and near-neutral pH was maintained throughout the experiment. Both conditions are conducive to green rust formation. The added Fe(II) also maintains solution charge balance and slows passivation of the Fe0 surface. Adding Al2(SO4)3 with the Fe0 also resulted in green rust color with a lower pH (approximately 6) and slightly higher Eh than the other treatments. As previously indicated, adding aluminum facilitates release of Fe(II) into solution, which could possibly substitute for Fe(III) in the green rust structure (Taylor and Schwertmann, 1978; Taylor and McKenzie, 1980).
pH-Stat Experiments
By monitoring changes in Eh and pH, we observed that the pH for our most favorable treatment [Fe0 + Al2(SO4)3 + FeSO4] fluctuated between 4 and 6. Using a pH-stat, we controlled the pH of the Fe0 + Al2(SO4)3 + FeSO4 treatments. Metolachlor destruction rate increased as pH increased from 3 to 5 in solution and soil slurries for the annealed iron treatment, while the destruction rate decreased with increasing pH when unannealed iron was used (Fig. 7)
. Several factors may explain this contrasting result. Annealed iron is coated with a thin layer of magnetite whereas the coating on unannealed iron is mostly maghemite and hematite (Satapanajaru, 2002). Subjecting the unannealed iron to acidic solution would remove these Fe(III) passivating layers from the Fe0 core and render it free for reductive transformations. When passivating, unreactive layers are not a concern (annealed Fe0); formation of a reactive, electron-conducting green rust, and/or magnetite layer on the annealed Fe0 is favored as pH increases. In addition, maintaining a low pH and Eh would limit the formation of Fe(III), which is necessary for green rust formation. In fact, green rust was not observed at lower pH when salt-amended annealed iron was used. Aluminum substitution for Fe in green rust and speciation may also be important because the predominance of Al3+ at pH 5 would promote Fe2+ release from the oxidizing layer on the Fe0 surface. The slower rate of metolachlor destruction when pH was increased from 5 to 6 may be due to precipitation of Fe(OH)3, Al(OH)3 (gibbsite), and Fe(III)oxyhydroxides, which would passivate the iron surface and limit the amount of Fe(III) and aluminum available for incorporation into the iron coatings.
|
and Cr(VI), Fe0 source did not significantly influence reduction rates because their results could be explained by surface area differences among the irons. By contrast, large differences in destruction kinetics were observed when nitrate 
was treated with the three irons at various pHs. They suggested that classifying iron metal types in terms of reactivity may not only be condition-specific (pH and Eh) but also contaminant-specific. Considering that there are literally hundreds of different cast irons and alloyed steels on the market today, differences in destruction kinetics among iron sources are not surprising. While some of these differences may be attributed to differing surface areas, contrasting results due to iron source, pH, and the effects of surrounding electrolytes indicate that the mineralogy of surface iron oxides must be considered to accurately predict reaction rates. |
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
The manufacturing process used to produce the Fe0 also profoundly affected destruction rates. We used two types of Fe0 that differed in initial surface oxide coatings and observed that metolachlor destruction rates with salt-amended Fe0 were greater with annealed iron (indirectly heated under a reducing atmosphere) than unannealed iron. Moreover, the optimum pH for metolachlor dechlorination in water and soil differed between iron sources (pH 3 for unannealed, pH 5 for annealed). Our results indicate that metolachlor destruction by Fe0 treatment may be enhanced by adding Fe or Al salts and creating pH and redox conditions favoring green rust formation.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  T. Lee and C. H. Benson Sorption and Degradation of Alachlor and Metolachlor in Ground Water Using Green Sands J. Environ. Qual., September 1, 2004; 33(5): 1682 - 1693. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Park, S. D. Comfort, P. J. Shea, and T. A. Machacek Remediating Munitions-Contaminated Soil with Zerovalent Iron and Cationic Surfactants J. Environ. Qual., July 1, 2004; 33(4): 1305 - 1313. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Soil Science Society of America Journal | Journal of Plant Registrations | The Plant Genome | |||
