Dechlorination of chloromethanes on iron and palladium-iron bimetallic surface in aqueous systems

INTRODUCTION

A wide range of chemical wastes are generated by teaching chemistry laboratories. As a part of continuing efforts to reduce the wastes, comprehensive programs are being developed for the in situ remediation of the potentially toxic wastes. Chloromethane solvents are among the chemicals that are generated in substantial quantity and are under consideration for treatment. These wastes are relatively difficult to remediate under ambient conditions partly because they are typically concentrated and frequently contain metals and other substances that might be toxic to microorganisms. Therefore, the use of microbes of dechlorinating competency does not appear to be suitable.

In the early 1970s, Sweeney [1] and Gillham [2] reported the use of zero-valent metals, transition metals in particular, as a practical means to decontaminate organohalides. This approach has attracted a great deal of interest because of its realistic and cost-effective means to degrade organohalides in groundwater. The underlying chemistry of reductive dehalogenation of chlorinated aliphatic hydrocarbons in aqueous solution has since been studied under different reaction conditions [3-8]. For example, Matheson and Tratnyek [3], while studying the kinetics and mechanisms of dechlorination of carbon tetrachloride (CCl₄) for the metal iron–water systems (Fe°-H₂O), found that CCl₄ was dechlorinated to form chloroform (CHCl₃) as a dominant product, which in turn is partly transformed to methylene chloride (CH₂Cl₂), although at a much slower rate. However, the fate of CH₂Cl₂ remained unclear.

Although the Fe°-H₂O approach appears to be promising for dechlorinating CCl₄ and possibly CHCl₃ to CH₂Cl₂ in such environmental matrices as groundwaters and sediments, where the oxygen levels are relatively low, the Fe°-H₂O-mediated treatment of the wastes that come in contact with air for periods of days might be too inefficient, for the half-reaction involving O₂ as electron acceptor, 2 Fe° + O₂ + 2 H₂O ⇌ 2 Fe²⁺ + 4 OH⁻, can compete with the dechlorination reaction of CCl₄, which can be described by the reaction Fe° + RCl + H⁺ → Fe²⁺ + RH + Cl⁻. Recent studies have indicated that the dechlorination rates by the iron–water system can be improved at least in two ways. Wang et al. [9] showed that the dechlorination rates of trichloroethene could be vastly increased by increasing the surface area of iron granules from 10 μm (with surface area of 0.9 m²/g) to 1 to 100 nm (with surface area of 33.5 m²/g). In addition, palladium-treated iron (Pd/Fe°) has been shown to be highly effective in the dechlorination of polychlorinated biphenyls [10] and chloroethenes [11]. A study on the palladium–iron surface indicated that prolonged exposure of the Pd/Fe surface to a saturated solution of aqueous trichloroethene resulted in the accumulation of the hydroxylated iron oxide but could be readily removed by washing with a dilute mineral acid without desorbing palladium [12].

The present study was undertaken with three purposes. First, we wanted to determine the relative effects of both Fe°-H₂O and Pd/Fe°-H₂O on the dechlorination rate of CCl₄. Second, we considered the effect of oxygen on the reductive dechlorination, as the matrices of the solvent wastes are normally in equilibrium with air. Finally, we wanted to further explore the degradation products of chloromethanes for the iron–water system by the use of palladium as a catalyst, an aspect clearly important to the remediation of chloromethanes.

MATERIALS AND METHODS

Iron particles (<10 μ) and potassium hexachloropallidate (K₂PdCl₆; fresh weight 397.32) were obtained from Aldrich Chemical (Milwaukee, WI, USA). Surface oxides of the iron were removed with 2 M HCl. Residual acid was removed by rinsing the acid-treated particles with a bicarbonate solution (0.05 M). The washed iron particles were used in the dechlorination experiments or were palladized by suspension in an aqueous solution of K₂PdCl₆ (7 × 10⁻³ mole of Pd per mole of Fe). The mixture was gently shaken on a shaker-incubator at ambient temperature. The coating process is complete in about 8 min, as indicated by the appearance of a bright yellow solution. The palladized iron was thoroughly washed with 20 to 30 volumes of water under aspiration or until the pH of the rinses rose to about 7.

Quantitation

A calibration curve for each compound was established in the following way. To evaluate the detector response, independent mixtures of CCl₄, CHCl₃, CH₂Cl₂, CH₃Cl, and CH₄ were prepared by injecting variable amounts of the individual components into tightly sealed vials containing 10 ml of distilled water. The solutions were allowed to equilibrate (about 40 min) under the conditions of analysis, followed by manual injection of headspace samples with an injection volume of 25 μL into the gas chromatograph. The detection limits for CCl₄, CHCl₃, and CH₂Cl₂ were 100 to 500 ppb and about 8.0 and 3.0 nmol for CH₃Cl and CH₄, respectively. The FID responses were linear up to three orders of magnitude from their detection limit for all compounds.

RESULTS

To assess the catalytic role of palladium in the dechlorination of CCl₄, control experiments were initially performed with iron alone (Fe°-H₂O). Aqueous solutions of CCl₄ (263 μmol or 40 ppm) were reacted with the incremental amounts of iron particles at room temperature. The reaction was monitored until the CCl₄ reached its detection limit. Figure 1A shows the apparent increasing rate constant (k′) for the disappearance of CCl₄, which increases as a function of iron mass. The rate constants were derived from the plots of natural log concentration of CCl₄ versus time that were fitted to the firstorder reaction model. The coefficient of determination (r²) on the data gives 0.88 for n = 10. Under the experimental conditions, dechlorination of CCl₄ assumed pseudo-first-order kinetics with respect to the mass of iron granules. Parallel experiments were next carried out to investigate the effect of palladium on the dechlorination of CCl₄. For comparative purposes, the iron weights were maintained at 0.5 g, and the amount of palladium used to palladize the iron ranged from 50 to 375 μg (5.2 × 10⁻⁵−3.9 × 10⁻⁴ moles per mole of Fe). Figure 1B shows that the rate constant for the disappearance of CCl₄ increases linearly with the amount of palladium (Pd/Fe°-H²O) used in the study. Linear regression on the data for the disappearance of CCl₄ yields r² = 0.95 for n = 6.

The estimated relative half-reduction potentials of oxygen and CCl₄ in water at pH 7.4 are +0.40 [13] and +0.67 [14], respectively. Thus, Fe° is expected to react with CCl₄ more favorably than with oxygen. Nevertheless, the rate of reductive dechlorination by the Fe°-H₂O might be substantially reduced in the presence of oxygen. The aqueous solution of CCl₄ was deoxygenated as follows. Immediately before the start of the experiment, 10 ml of distilled, deionized water was placed in the vial containing 0.5 g iron or 0.5 g iron treated with 250 μg Pd. The headspace was quickly flushed with nitrogen before crimping the vial. The air in the reaction medium was purged by passing a stream of nitrogen with a slight positive pressure through the septum by means of a microliter-size syringe that was coupled to a compressed nitrogen tank. A second syringe placed on the septum allowed the gas to be evacuated. Microliter aliquots of the aqueous CCl₄ were then injected into the reaction medium. As shown in Table 1, an inhibitory effect of oxygen is clearly evident. It took nearly three to four times as long to dechlorinate the same quantity of CCl₄ in the presence of oxygen. The oxygen effect was even more pronounced when the experiments were performed with the Pd/Fe-H₂O.

To investigate the potential dechlorination of CCl₄ to CH₄, we performed several experiments. For purposes of comparison, dechlorination reactions of CCl₄ and the partially chlorinated products, CHCl₃ and CH₂Cl₂, by the Fe°-H₂O were individually measured in separate experiments. All the reactions were performed under aerobic conditions unless stated otherwise. The chloromethane concentrations were measured over the experimental period of 23 d. As shown in Figure 2, the disappearance of CCl₄ was by far the fastest; it took only a few hours for the compound to fall below the detection limit. On the other hand, it took nearly 23 d for the CHCl₃ to disappear. It is also evident that the dechlorination rates of CH₂Cl₂ were barely measurable over the entire experimental period of 23 d. In the next series of experiments, the transformation of CCl₄ was analyzed on both Fe°-H₂O and Pd/Fe°-H₂ O. For the Fe°-H₂O system, the dechlorination of CCl₄ was accompanied by the formation of CHCl₃ (Fig. 3). In terms of mass balance, CHCl₃ accounts for about 65% of the starting CCl₄. In addition, formation of CH₄ was detected, and its appearance was nearly concurrent with the formation of CHCl₃. The early formation of CH₄ was somewhat surprising in view of the observations that the reactivity of CH₂Cl₂ toward the Fe°-H₂O was extremely low (Fig. 2). Parallel experiments performed with Pd/Fe° yielded additional information (Fig. 4). The dechlorination of CCl₄ on the bimetallic surface proceeded with much greater rate; within 2 h, the dechlorination of CCl₄ was essentially complete, with the level of CHCl₃ already declining from its peak. By the end of the 10-h experimental period, virtually all CHCl₃ was dechlorinated, with the attendant appearance of CH₄ as the major dechlorination product. In addition, the formation of CH₃Cl, although relatively minor, was also evident.

When CHCl₃ was reacted with the Fe°-H₂O as starting substrate, it took nearly 23 d for the CHCl₃ to fall below the detection limit, and its dechlorination was accompanied by the appearance of the putative intermediates: CH₂Cl₂ and CH₃Cl plus CH₄ (Fig. 5). The aggregate amounts of these products account for less than 50% of the starting CHCl₃; however, it should be noted that the formation of all these products continued to rise when the experiments were terminated. When the parallel experiments were carried out with the Pd/Fe°-H₂O, the dechlorination rates of CHCl₃ dramatically increased: its dechlorination took only about 1 h (Fig. 6). Furthermore, the concomitant appearance of CH₄ as the single dominant product was even more evident. Both CH₂Cl₂ and CH₃Cl also appear as minor products, but their formation was clearly delayed.

DISCUSSION

The present data demonstrate that micron-scale iron granules suspended in water (Fe°-H₂O) are highly effective in dechlorinating CCl₄. The kinetics of CCl₄ dehalogenation was related to the amount of iron, suggesting that the rate of dechlorination is set by the availability of iron surface area, as described in the proposed mechanisms of dehalogenation of halogenated hydrocarbons by metals in aqueous systems [3]. A recent study by Wang et al. [9] on the dechlorination of trichloroethene by nanoscale iron powder further illustrates how the reaction rates could be improved by increasing the metal surface area. The present study demonstrates that dechlorination of CCl₄ can also be increased by treating iron granules with palladium (Pd/Fe°-H₂O), verifying the high catalytic effects previously demonstrated in connection with polychlorinated biphenyls [10] and trichloroethenes [11]. However, the catalytic role of palladium for the dehalogenation of organohalides for the Fe°-H₂O system is speculative. Although palladized iron has a clear advantage in the promotion of dechlorination of organohalides, a possibility of desorption of palladium from the Pd/Fe surface as a result of repeated exposure to aqueous organohalides remains a concern because of its potential toxicity. On the other hand, the short-term analysis of the Pd/Fe surface with X-ray photoelectron spectroscopy shows that the Pd/Fe surface is fairly stable [12].

We have examined the effects of oxygen on the reductive dechlorination rates to better estimate treatment of chloromethane wastes that are normally in a state of equilibrium with atmospheric oxygen. In aqueous systems, zero-valent iron is readily oxidized by the presence of H⁺ and H₂O as electron acceptors (corrosion). As anticipated from the reduction potentials of oxygen and organohalides, dissolved oxygen strongly competes with the reaction of CCl₄ with Fe°, as evidenced by the reduced rates of halomethane disappearance. Nevertheless, the application of the Fe° or Pd/Fe° system is practical in the treatment of the concentrated chloromethane wastes under aerobic conditions.

The reaction of CCl₄ by the Fe°-H₂O under the conditions of this study produced CHCl₃ as the dominant product (Fig. 3), which is estimated to be 65 to 70% of CCl₄ consumed. These results are in general agreement with the earlier observations of Matheson and Tratnyek [3], although our experiments were conducted under aerobic conditions. In addition, the degradation of CCl₄ by the Fe°-H₂O system was also accompanied by the formation of CH₄ as a significant minor product, raising doubts on any mechanism in which partially chlorinated chloromethanes are formed as obligatory intermediates in the formation of CH₄. The dechlorination of CCl₄ by the Pd/Fe° system performed under similar conditions also produced CH₄, but its formation as the dominant product accounted for about 50% of the CCl₄ that disappeared.

According to the report of Matheson and Tratnyek [3], CHCl₃ is further degraded to CH₂Cl₂ by the Fe°-H₂O. Our study also demonstrates that when CHCl₃ was employed as parent compound on the Fe°-H₂O system, roughly one-third of the substrate was converted to CH₂Cl₂. In addition, the formation of two other products, CH₃Cl and CH₄, was clearly evident; these products account for 25% of the CHCl₃ consumed. Therefore, formation of these products suggests a possible dechlorination route for the CCl₄ on the Fe°-H₂O:

However, some of the data are inconsistent with this scheme. The independent measurements of the disappearance rates of three chloromethanes, CCl₄, CHCl₃, and CH₂Cl₂, toward the Fe°-H₂O system revealed that the reactivities of the three chloromethanes with the Fe°-H₂O are significantly different (Fig. 2). In relating these data to the results on the dechlorination of CCl₄ by Fe°-H₂O (Fig. 3), it is apparent that neither CHCl₃ nor CH₂Cl₂ could have been obligatory intermediates in the formation of CH₄ because the appearance of CH₄ was nearly concurrent with the degradation of CCl₄ and the formation of CHCl₃, whereas the formation of either CH₂Cl₂ or CH₃Cl was not detected. As discussed earlier, the decomposition of CCl₄ by the Pd/Fe°-H₂O, in comparison with Fe°-H₂O, takes place with a significantly greater rate, so that its application was extremely useful in following the transformation of CCl₄ to other products and in the establishment of precursor–product relationships. On the Pd/Fe°-H₂O system, the decomposition of CCl₄ gave rise to all chloromethanes as well as CH₄ over the period of 10 h (Fig. 4), except for CH₂Cl₂, whose formation could not be detected with consistency. In the presence of palladium, the formation of CHCl₃ decreased to about 20% from 33% of the starting CCl₄, whereas the formation of CH₄ rose from 6 to 44%. Thus, it seems that palladium stimulated the reactions that led to the formation of CH₃Cl and CH₄ at the expense of CHCl₃ formation. These observations, when taken together, suggest that a dechlorination pathway leading to the formation of CH₄ via the formation of a minimum of three intermediates (bracketed) corresponding to the chlorinated products:

It could be that the [CCl₃] is a radical that is formed via the one-electron reduction of CCl₄, as, for example, Criddle et al. [7] described for the electrolytic model system for reductive dehalogenation in aqueous environments. These investigators are supportive of the mechanism that involves oneelectron reduction of CCl₄ to give a trichloromethyl radical. The radical might then be transformed to various products, including CO₂, HCOOH, and CHCl₃, via different routes. In the present system (Fe° or Pd/Fe°), a trichloromethyl radical might be formed as the first step on the iron surface, which is then either hydrogenated to form CHCl₃ or further dechlorinated to form the less chlorinated radical intermediates. Matheson and Tratnyek [3] showed that the kinetics of CCl₄ dehalogenation on the Fe°-H₂O is highly pH dependent. However, whether the H⁺ participates in the protonation of the radical intermediates remains to be investigated.

According to the previously described scheme, for CHCl₃ to degrade to form CH₄, it must first undergo deprotonation to form the trichloromethyl intermediate followed by successive dechlorination while remaining on the iron surface. Likewise, CH₂Cl₂ as well as CH₃Cl could be similarly dechlorinated to form CH₄; however, on the basis of the extremely low rate of degradation of CH₂Cl₂ by the Fe°-H₂O, its degradation to CH₄ should be negligible.

The increased formation of CH₄ by the Pd/Fe°-H₂O (Figs. 4 and 6) also suggests that the palladium accelerates the successive dechlorination reactions with the formation of the intermediates without directly affecting the proton transfer reactions. For example, in the presence of palladium, the formation of CH₄ from CCl₄ increased from 6 to 44% and the formation of CH₄ from 12 to 63% when CHCl₃ was used as the parent compound. That the formation of CH₄ is nearly concurrent with the dechlorination of CHCl₃ indicates that the hydrogenation of [CCl₃] is faster than that for the less chlorinated intermediates, [CCl₂] and [CCl], to form CH₂Cl₂ and CH₃Cl, respectively (as shown by their delayed appearance). It should be noted that the experiments here were shortened to 2 h (vs 23 d without palladium), so that the formation of these two products was not complete when the experiments were terminated.

In terms of the carbon mass balance between the products (CH₄, CH₂Cl₂, and CH₃Cl) and the starting substrate (CHCl₃), the combined mass of the products account for about 70% of the parent compound, implying that perhaps the CCl₄ or the other chlorinated compounds could be partially degraded to form other carbon products, such as CO₂ and formate, when the reactions are performed in the presence of oxygen. For example, the reaction of CCl₄ with pyrite (FeS₂) under aerobic conditions produced CO₂, whereas in an anaerobic pyrite system, about one-half of the CCl₄ was transformed to CHCl₃ [4]. Other experimental conditions that include other metals could also affect the manner in which CCl₄ is decomposed. Boronina and Klabunde [15] reported that decomposition of CCl₄ by the Zn-H₂O produced CH₄ via the partially chlorinated methanes (CHCl₃, CH₂Cl₂, and CH₃Cl), whereas CO₂ and CHCl₃ were the major products in the case of a Sn-H₂O system.

CONCLUSIONS

Of the three chloromethanes examined, iron granules (Fe°-H₂O) were most effective in dechlorinating CCl₄. The Fe°-H₂O is far less effective in the dechlorination of CHCl₃ and is virtually unreactive toward the CH₂Cl₂. Palladized iron (Pd/Fe°-H₂O) dramatically increases the dechlorination reactions. The presence of oxygen significantly reduces the rate of dechlorination of CCl₄ by both the Fe°-H₂O and the Pd/Fe-H₂O. From the mass balance and the time-course appearance of the chloromethanes, it seems that the partially chlorinated products, CHCl₃, CH₂Cl₂, and CH₃Cl, are not obligatory intermediates in the formation of CH₄ from CCl₄. It is proposed that the dechlorination of CCl₄ initially forms a reactive trichloromethyl intermediate, possibly a free radical, that is then successively dechlorinated to produce additional reactive intermediates, hydrogenation of which then results in the formation of the corresponding chloromethanes. Palladium (Pd/Fe°-H₂O) accelerates the decomposition of either CCl₄ or CHCl₃ to form CH₄ as a major product.