Environ. Sci. Technol. 2002, 36, 762-768
Catalyzed Destructive Adsorption of Environmental Toxins with Nanocrystalline Metal Oxides. Fluoro-, Chloro-, Bromocarbons, Sulfur, and Organophosophorus Compounds SHAWN P. DECKER, JOHN S. KLABUNDE, ABBAS KHALEEL, AND KENNETH J. KLABUNDE* Department of Chemistry, Kansas State University, Manhattan, Kansas 66506
In the temperature range of 300-500 °C, solid nanocrystalline oxides react nearly stoichiometrically with numerous halocarbons, sulfur, and organophosphorus compounds. In some cases, the reaction efficiencies can be improved by the presence of a small amount of transition-metal oxide as catalyst; for example, Fe2O3 on CaO and mobile intermediate species such as FeCl3 or Fe(SO3)x are important in the catalytic process. Herein, a series of environmentally problematic compounds are discussed, including CCl4, COS, CS2, C2Cl4, CHCl3, CH2Cl2, CH3Cl, and (CH3O)2P(O)CH3. Nanocrystals of CaO coated with a thin layer of Fe2O3 (or other transition metals) ≡[Fe2O3]CaO, or intimately mixed ≡Fe2O3/CaO were compared with pure CaO. It was found that (a) the presence of a small amount of surface [Fe2O3] or other transition-metal oxide can have a marked effect on the destructive adsorption activity, (b) for some reagents, such as CCl4, C2Cl4, SO2 and others, the nanocrystalline CaO can react in stoichiometric amounts, especially if a transition-metal oxide catalyst is present, (c) although the reaction with dimethylmethylphosphonate is surface-limited, the nanocrystalline calcium oxide performed well and in high capacity, (d) nanocrystalline calcium oxide exhibits near stoichiometric activity with several interesting sulfur-containing compounds, such as COS and CS2, (e) unfortunately, most fluorocarbons were not destructively adsorbed at 500 °C under the conditions employed; however, some of these can be effectively mineralized over the calcium oxide at higher temperatures. These compounds include C2F6, C3F6, C2ClF3, and CHF3, and (f) upon reaction, surface areas decreased considerably, from about 100 to about 10 m2/g. The results of these experiments further demonstrate that, with the proper choice of catalytic material, some solid-gas reactions can be engineered to be rapid and essentially stoichiometric.
I. Introduction Recent concern about environmental hazards has prompted investigation into benign methods for the treatment of hazardous waste materials. A wide variety of chlorinated * Corresponding author phone: (785) 532-6849; fax: (785) 5326666; e-mail:
[email protected]. 762
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002
compounds are used by society in many beneficial functions. Examples are cleaning solvents, plasticizers, lubricants, and refrigerants. While some of these chlorinated compounds are being replaced by less harmful chemicals, many continue to be used because of the lack of suitable replacements or because of economic considerations (1). Therefore, considerable interest exists in developing methods for the safe disposal of chlorinated and other problematic wastes. The work reported herein deals with the use of high surface area calcium oxides coated with small amounts of transition-metal oxides, such as iron or nickel oxide, to destroy such compounds. Synthetic methods have been developed for generating high surface area magnesium and calcium oxides (2). The syntheses produce very small crystallites of calcium oxide that are in the range of 4-7 nm (one billionth of one meter) and are thus called nanosized particles. Bulk amounts of the nanosized crystallites take the form of a finely divided white powder. The high surface area of the calcium oxide (which is about 60 times that of normal calcium oxide available from commercial chemical manufacturers) is due to the small size of the crystallites. The higher surface area means that the nanoparticles will react to a greater extent than normal CaO because there are more molecules of calcium oxide available for reaction with chemicals adsorbed at the surface of the particles. However, nanocrystals also exhibit intrinsically higher chemical reactivities, and this is due to unusual crystal shapes (3) and lattice disorder (4). Early research showed that both high surface area calcium oxide and magnesium oxide react with carbon tetrachloride to produce calcium chloride (or magnesium chloride) and carbon dioxide as products (2, 5). The term “destructive adsorption” is often applied to these reactions because the carbon tetrachloride is “destroyed” during the reaction with calcium oxide by becoming mineralized as environmentally benign calcium chloride and carbon dioxide gas. Thus, the term is used to differentiate the nanocrystalline calcium oxide from materials such as activated carbon or zeolites that simply adsorb the hazardous materials but do not chemically alter them. Often, the nanosized calcium and magnesium oxide particles are referred to as first-generation destructive adsorbents because a new series of nanosized particles have been synthesized (5). The second-generation particles are best described as “layered” metal oxides because small amounts of transition-metal oxides, such as nickel or iron oxide, are placed on the surface of the particles in an attempt to increase the reactivity of the particles. It is believed that these coatings are less than monolayers, on the basis of the small amount added and the absence of three-dimensional islands in electron microscope images. The thought behind adding a small amount of transition-metal oxide is that they may be able to catalyze the reaction and enhance the activity of the CaO nanoparticles. It is thermodynamically feasible that the iron oxide can react with the carbon tetrachloride first to form iron chloride. The iron chloride can then react with the CaO substrate to regenerate the iron oxide and mineralize the chlorine (5). The first part of this report will survey a small series of [MxOy]CaO samples reacting with CCl4. The purpose of this survey was to compare the effect of several transition-metal oxides and then to select one for further study. In addition, the aspect of catalyst morphology is explored (coating or intimately mixed). This study is followed by sections on mixed halocarbons, sulfur compounds, and phosphorus compounds. 10.1021/es010733z CCC: $22.00
2002 American Chemical Society Published on Web 12/29/2001
II. Experimental Section Nanocrystalline CaO (AP-CaO), microcrystalline CaO (CPCaO), and transition-metal oxide coated [MxOy]CaO such as [Fe2O3]CaO were prepared as previously described (5, 6). Commercially available CaO (CM-CaO) was purchased from Fisher Chemical (Pittsburgh, PA). Typical surface areas were AP-CaO (120 m2/g), CP-CaO (80 m2/g), and CM-CaO (2 m2/g). Reactions of these powderlike materials with a variety of environmentally problematic compounds are described in the following paragraphs. A. Carbon Tetrachloride. The experiments were conducted via a pulse injection method using gas chromatography (GC). In these experiments, a small amount (either 50 or 100 mg) of [MxOy]CaO was placed in a quartz reaction U tube (QRT). The QRT was made of 1/4 in. quartz tubing that was bent into the shape of the letter “U”. The sample was held in place with Saffil matting (alumina) on either side of the sample. The Saffil has the density of a cotton ball and easily allows the passage of gas. Also, the matting is relatively inert. The sample was loaded into the QRT, was weighed, and was then fitted to the side attachment of the gas chromatograph. The gas chromatograph was a Gow-Mac 580 with a set of thermal conductivity detectors (TCD). Purified (7) helium was used as the carrier gas at a flow rate of 30 mL/min. At this flow rate, the carbon tetrachloride sample has a residency time of between 0.72 and 0.84 s over the sample. The GC was fitted with a 12 ft × 1/8 in. stainless steel Carbosphere W-HP (80/100 mesh) column with 10% SE-30 stationary phase. The column temperature was set at 120 °C, injector temperature was 140 °C, and the detector temperature was 140 °C. The sample was heated with a ceramic oven that was set at 425 °C. Heating tape was used to warm all of the exposed stainless steel tubing and was insulated with Saffil matting and wrapped with aluminum foil. Once all of the set temperatures were reached, the flow rate of the helium was rechecked, and 1 µL pulses of CCl4 were injected at 6 min intervals (8). The total amount of CCl4 to be injected was determined by calculating the moles of calcium oxide that were present in the sample and then using a slight excess of the molar amount needed to react stoichiometrically with the sample by the following equation:
2CaO + CCl4 w 2CaCl2 + CO2
(1)
Thus, for a 100 mg sample of calcium oxide, 86.3 µL of CCl4 would be required, and a slight excess was used so 90 × 1.0 µL injections of CCl4 were made. Note that the reaction is assumed to be stoichiometric rather than just with the surface of the calcium oxide. The gaseous products and residual CCl4 were then passed over the column of the GC for separation and were recorded on a chart recorder. The calibration curves were set up by injecting known amounts of each effluent and measuring the peak areas detected so that a relationship between the detector signal (peak area) and the amount eluted could be made. The calibration was performed under the same conditions except without the presence of CaO (Saffil plugs were present). This method works quite well and gives a very linear response over the range of reagent used. Correlation coefficients of 0.95-0.99 were common. The calibration curves were regenerated on a regular basis to take into account wear and tear on the tungsten filaments of the TCD detector. The gaseous products for the reaction were CO2 and a very small amount of C2Cl4. Also, once the breakthrough of CCl4 was reached, unreacted CCl4 was also detected in the effluent. The breakthrough number is defined as the number of CCl4 injections that occur before unreacted carbon tetrachloride is detected. The entire amount of CCl4 in an injection, before this breakthrough number, reacts with the
calcium oxide sample. After this number, the amount of excess CCl4 slowly grows until it reaches a maximum, which is equivalent to the amount of a single injection of CCl4. After the quantities of gaseous effluent for each peak are quantitated, they are summed over all of the injections in order to determine the total amount of CCl4 mineralized and to make analysis of the data easier. Although the reaction ratio can be monitored by many different criteria, the easiest was to simply determine the amount of CCl4 decomposed relative to the amount calcium oxide in the sample. The breakthrough number was also important in determining which samples were the most active. The reaction ratio is defined as the number of moles of CCl4 that were decomposed per mole of CaO reagent.
reaction ratio )
moles of CCl4 destroyed moles of CaO
(2)
The ideal reaction ratio can be determined from a balanced equation of the reaction. If the reaction ratio is taken as a fraction of the ideal reaction ratio, this number can be thought of as the percentage of the CaO that is interacting with the CCl4 during the reaction. The solid products of the reaction were determined by powder X-ray diffraction. B. Dimethylmethylphosphonate (DMMP). The DMMP experiments were conducted at 200, 400, and 500 °C. The DMMP was purified by drying over Na2CO3 to remove acids, followed by three freeze-pump-thaw cycles. The purified DMMP was then stored in a sealed vial over 5A zeolite with an argon atmosphere until use. Approximately 50 mg of CaO (8.92 × 10-4 mol) were used per experiment. A total of 50 × 1.0 µL injections of DMMP (4.61 × 10-4 mol) were made at 6 min intervals. The amount of DMMP used was such that there would be a slight excess per surface CaO because it was known from the literature (9) that the reaction was surface limited. The chromatographic settings were as follows: injection temperature reactor temperature column temperature detector temperature detector current column type flow rate
225 °C 200, 400, 500 °C 200 °C 210 °C 125 mA carbowax 20 on carbograph 30 mL/min
The chromatogram indicated two elution peaks for the reaction. One was a mixture of formic acid and methanol while the other was residual or unreacted dimethyl methyl phosphonate (DMMP). The DMMP peak was used to quantitate the reaction by subtracting the amount that was eluted from the amount injected to give a number that represents the amount of DMMP that has either reacted with the CaO or has adsorbed to its surface. C. Carbonyl Sulfide and Carbon Disulfide. The same procedure used in the previous destructive adsorption experiments was utilized with this investigation, namely, that pulsed GC injections were made over a small bed of CaO. The sulfur-containing reagents were used as received without further purification. The chromatographic settings were as follows: injection temperature reactor temperature column temperature detector temperature column type detector current flow rate
120 °C 500 °C 85 °C 140 °C Poropak Q 80/100 mesh 110 mA 30 mL/min
In the case of COS and CS2, CO2 was the major gaseous product for the reaction with CaO. The sequestering abilities VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
763
TABLE 2. CCl4 Destructive Adsorption at 425 °C for Several Different Loadings of [Fe2O3]
TABLE 1. CCl4 Destructive Adsorption over a Series of [MxOy]CaO Nanocrystals at 425 °Ca sample
reaction ratio SSA breakthrough mol of CCl4/ (mol of CCl4/ (m2/g) no. g of CaO mol of CaO)
CM-CaO
2-3
1
∼0
CP-CaO 2.1%[V2O5]CP-CaO 2.4%[Fe2O3]CP-CaO 2.3%[NiO]CP-CaO 2.1%[CuO]CP-CaO 1.7%[ZnO]CP-CaO
107 92.4 92.0 100 65.5 98.0
1 2 2 6 1 2
3.91 × 10-3 2.92 × 10-3 6.08 × 10-3 3.76 × 10-3 2.91 × 10-3 3.89 × 10-3
∼0 0.22 0.16 0.34 0.21 0.16 0.22
AP-CaO 3.1%[V2O5]AP-CaO 3.3%[Fe2O3]AP-CaO 2.5%[NiO]AP-CaO 2.2%[CuO]AP-CaO 1.7%[ZnO]AP-CaO
130 124 114 125 78.3 108
1 2 9 26 11 5
5.34 × 10-3 6.08 × 10-3 8.55 × 10-3 5.76 × 10-3 3.60 × 10-3 1.78 × 10-3
0.30 0.34 0.48 0.32 0.20 0.10
a Loading of transition-metal oxide ranged from 1.7% to 3.3% by mass as determined by elemental analysis.
of the calcium oxide were determined by the amount of COS or CS2 that were removed from the carrier gas. The amount of COS or CS2 eluted from the reaction was subtracted from the amount injected to determine the amount adsorbed by the CaO. D. Chloromethanes. The experiments were performed in an analogous manner to those of CCl4 decomposition except that the reaction temperature was 500 °C instead of 425 °C. Also, only the AP-CaO and [Fe2O3]AP-CaO nanocrystals were examined in these experiments. E. Chlorinated Ethanes and Ethenes. The destructive adsorption of these compounds with AP calcium oxides were made with the same experimental apparatus that was used for the previous studies. The chromatographic settings were as follows: injection temperature reactor temperature column temperature detector temperature filament temperature column type flow rate
250 °C 500 °C 230 °C 250 °C 270 °C Poropak Q 80/100 mesh 30 mL/min
All of the reagents were used as received. The amounts injected were such that a slight stoichiometric excess of chlorocarbon was used. The reaction ratios were determined in the manner previously described. F. Fluorocarbons, Chlorofluorocarbons, and Bromocarbons. These experiments utilized the same pulse gas chromatographic apparatus that was used in the previous experiments. The chromatographic settings were as follows: injection temperature reactor temperature column temperature detector temperature filament temperature column type flow rate
250 °C 500 °C 230 °C 250 °C 270 °C Poropak Q 80/100 mesh 30 mL/min
The reactions were quantitated in the same manner as before.
III. Results and Discussion A. Survey of the Reactions Between [MxOy]CaO and CCl4. The first goal of these experiments was to look at a small series of [MxOy]CaO samples to examine which, if any, would exhibit enhanced activity with the presence of transitionmetal oxides at the surface. The [MxOy]CaO systems studied were [Fe2O3], [V2O5], [NiO], [CuO], and [ZnO] all with CP764
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002
sample
SSA
breakthrough mol of CCl4/ no. g of CaO
reaction ratio (mol of CCl4/ mol of CaO)
CP-CaO 107 0.5% [Fe2O3]CP-CaO 94 1.0% [Fe2O3]CP-CaO 92 2% [Fe2O3]CP-CaO 96 5% [Fe2O3]CP-CaO 93
1 1 2 2 2
3.91 × 10-3 3.96 × 10-3 3.74 × 10-3 4.66 × 10-3 4.05 × 10-3
0.22 0.22 0.21 0.26 0.23
AP-CaO 0.5% [Fe2O3]AP-CaO 1.0% [Fe2O3]AP-CaO 2% [Fe2O3]AP-CaO 5% [Fe2O3]AP-CaO
1 1 9 5 4
5.34 × 10-3 2.34 × 10-3 6.67 × 10-3 7.72 × 10-3 6.53 × 10-3
0.30 0.13 0.37 0.43 0.37
130 103 125 117 112
CaO or AP-CaO as the core material. From this series, a single sample would be selected to perform further studies on other halocarbons. The data shown in Table 1 is for a series of the first row transition-metal oxides. The ideal reaction ratio for this reaction would be 0.50 because 1 mol of CCl4 will react with 2 mol of CaO. Although all the samples were of high surface area compared with normal CaO (CM-CaO), they can be listed in decreasing order
CP-CaO > [NiO] ≈ [ZnO] > [V2O5] ≈ [Fe2O3] > [CuO] and then compared to the list of decreasing reactivity
[Fe2O3] > CP-CaO ≈ [ZnO] ≈ [NiO] > [V2O5] ≈ [CuO] It is easily seen that there is a moderate correlation between the surface area and the reactivity of the particles. The same can be said for the AP-CaO samples
AP-CaO > [NiO] ≈ [V2O5] > [Fe2O3] > [ZnO] > [CuO] (by SSA) and
[Fe2O3] > [V2O5] > [NiO] > AP-CaO > [CuO] > [ZnO] (by reactivity) While the presence of [MxOy] on the CP-CaO samples seems to inhibit reaction with CCl4 (except for [Fe2O3]), three of the transition-metal oxides enhance the reaction with AP-CaO. These transition-metal oxides are Fe2O3, V2O5, and NiO with iron oxide increasing the reactivity by almost 1.5 times that of the CaO substrate alone. It is also seen that the ranking of the [MxOy]CaO samples do not correlate between the CPCaO and the AP-CaO series. In the next set of experiments, the effect of the percent loading of [MxOy]CaO was examined, and only the Fe2O3 system was examined (Table 2). The percent loading of the [Fe2O3] on CP-CaO does not seem to have much of an effect on activity. The different loadings gave an activity ranging from 0.21-0.26, which is just slightly greater than the value of 0.22 for the uncoated CP-CaO sample. The AP-CaO samples gave different results. With the exception of the 0.5% loading, the [Fe2O3]AP-CaO samples possessed a marked increase in activity. In fact, one of the samples, 2% [Fe2O3]AP-CaO, gave a result that indicates a nearly complete reaction of the calcium oxide with the CCl4. What is important to realize here is that the CaO sample is behaving as a stoichiometric reagent. The nanocrystals are not just reactive at the surface; they are allowing the diffusion of CCl4 deep within the particles and are essentially “coring” them out.
TABLE 3. Destructive Adsorption of CCl4 at 425 °C with Various 1% [MxOy]CaO sample
surface area
breakthrough no.
performance efficiency
CM-CaO CP-CaO [Fe2O3]CP-CaO
1-2 100 90
1 1 2
∼0 0.23 0.44
AP-CaO [Fe2O3]AP-CaO [NiO]AP-CaO [CuO]AP-CaO
140 130 126 78
1 9 3 10
0.31 0.51 0.39 0.30
A study of 1% [MxOy]CaO was also made with CCl4. These data are shown in Table 3. First, it is seen that even a 1% loading of [Fe2O3] gives a very noticeable increase in the activity of both types of nanocrystalline calcium oxide. The presence of [Fe2O3] caused the performance efficiency of the CP-CaO to increase from 0.23 to 0.44, even though the surface area of the sample has dropped by 10%. In the case of the [Fe2O3]AP-CaO, the efficiency increases from 0.31 to 0.51 (which is the maximum possible), even though the surface area of the sample drops by about 7%. A small amount of [NiO] also gives a notable increase but not quite to the extent of [Fe2O3]. [NiO] increased the activity from 0.31 to 0.39. The powder X-ray diffraction patterns of the solid products of the reaction of the nanocrystalline CaO also indicate a nearly complete reaction for the [Fe2O3]AP-CaO samples. The four most intense diffraction peaks of calcium chloride hydrate (mineral name is sinjarite) CaCl2‚2H2O were dominant. Residual CaO was observed in the patterns for the CPCaO samples with the three most intense peaks occurring at 32.2, 37.3, and 53.9° 2θ. The samples exhibit the hydrated form of calcium chloride as a product because of the very hygroscopic nature of the CaCl2 samples after reaction. These results show that nanocrystalline calcium oxide can be effective in the decomposition and mineralization of carbon tetrachloride. The presence of small amounts of transition-metal oxides deposited onto the surface of these CaO nanocrystals can have a marked affect on the activity with CCl4. It was determined that, with the addition of a thin layer of [MxOy], only iron gave a significant enhancement in activity for the CP-CaO sample, while [NiO], [ZnO], [Fe2O3], and [V2O5] gave enhancement for the AP-CaO samples. Next, the effect of different [Fe2O3] loadings were examined. It was determined that the 2% [Fe2O3]CaO (both AP and CP) samples were most effective. On the basis of the data obtained from these experiments, it is possible to
conclude that the presence of a small amount of [Fe2O3] is very beneficial to the activity of the nanocrystals. One tentative mechanism which may explain what is happening on the surface of the particles, is shown in Figure 1. The first step of the mechanism may involve reaction between carbon tetrachloride and [Fe2O3] to form iron chloride and carbon dioxide. In the next step, there is a ion exchange process by which the iron chloride exchanges chlorine for oxygen in the CaO substrate to regenerate Fe2O3. The renewed iron chloride can then react with the next pulse of carbon tetrachloride. The newly formed iron chloride can then migrate over the surface of the particle to “search” for a new, unreacted portion of CaO with which to perform another Cl-/O2- exchange. This cycle then repeats until the CaO substrate is saturated with chloride. B. Destructive Adsorption of Dimethyl Methyl Phosphonate (DMMP). The military has been interested in developing methods of disposal of chemical warfare agents, many of which are organophosphors. DMMP has been found to be a reasonable mimic for these toxic organophosphorus compounds (9, 10). While DMMP is relatively nontoxic, it is a good mimic because it possesses the same types of bonds that exist in the more hazardous war agents (11). It had been found earlier that DMMP can be destructively adsorbed by high surface area magnesium oxide (12, 13). This chemistry is limited to the surface of the particles and required two surface moieties of MgO to decompose one molecule of DMMP. DMMP decomposes on the surface to give the following products (12):
In the current work with CaO, the first set of experiments with DMMP was to determine the best temperature for destructive adsorption. To do this, the decomposition over AP-CaO and [Fe2O3]AP-CaO at temperatures of 200, 400, and 500 °C was examined. The results from these experiments are given in Table 4. It is clear from comparison of the data at 400 and 500 °C that temperature plays a very important role in the amount of DMMP decomposed, regardless of the type of sample. The decomposition of DMMP rises from essentially 0 at 400 °C to 70-78% at 500 °C, while there is a 6-fold increase in the breakthrough number. This result was also observed by Li and Klabunde in the experiments with MgO (9).
FIGURE 1. Illustration of Fe2O3 catalyst on nanocrystalline CaO for the destructive adsorption of carbon tetrachloride. VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
765
TABLE 4. Adsorption of DMMP with AP-CaO and [Fe2O3]AP-CaO at Various Temperatures temp % DMMP breakthrough mol of DMMP/ (°C) adsorbed no. mol of CaO
sample AP-CaO AP-CaO AP-CaO [Fe2O3]AP-CaO [Fe2O3]AP-CaO [Fe2O3]AP-CaO
200 400 500 200 400 500
4 27 84 0 5.1 79
2 6 26 1 6 26
0.03 0.09 0.37 0.02 0.03 0.39
TABLE 5. Destructive Adsorption of DMMP with [MxOy]CP/ AP-CaO at 500 °C sample
decomp ratio (mol of DMMP/ mol of CaO)
breakthrough no.
CP-CaO [Fe2O3]CP-CaO [NiO]CP-CaO [V2O5]CP-CaO [CuO]CP-CaO AP-CaO [Fe2O3]AP-CaO [NiO]AP-CaO [V2O5]AP-CaO [CuO]AP-CaO
0.18 0.16 0.13 0.14 0.17 0.40 0.42 0.37 0.39 0.29
10 6 4 12 8 24 26 24 26 14
TABLE 6. Destructive Adsorption of COS and CS2 with CaO at 500 °Ca sample
breakthrough no.
decomp ratio (mol of COS/ mol of CaO)
breakthrough no.
decomp ratio (mol of CS2/ mol of CaO)
CP-CaO [Fe2O3]CP-CaO AP-CaO [Fe2O3]AP-CaO
11 15 22 17
0.51 0.57 0.74 0.69
7 4 10 9
0.24 0.34 0.47 0.54
The next set of experiments were set up to examine the affect of the [MxOy] coating. All samples contained a 2% loading by total number of moles of CaO. The results from these experiments at 500 °C are shown in Table 5. Unfortunately, for a given temperature, the presence of transition-metal oxide does not seem to have a significant effect on the decomposition of DMMP. This is contrary to the results for [Fe2O3]MgO where there is a very definite activity enhancement due to the presence of [Fe2O3] (5). C. Destructive Adsorption of Carbonyl Sulfide and Carbon Disulfide. The next series of to be examined were simple sulfur-containing materials, such as carbonyl sulfide (COS), carbon disulfide (CS2), and sulfur dioxide (SO2). The result for sulfur dioxide were presented earlier (3). The data and results for the destructive adsorption of COS and CS2 with nanocrystalline CaO are given in Table 6. The amount of [Fe2O3] loading is 2%. On the basis of the balanced equations
CaO + COS w CaS + CO2
(3)
2CaO + CS2 w 2 CaS + CO2
(4)
and
the maximum reaction ratios for COS and CS2 would be 1.0 and 0.50, respectively. 9
During the COS and CS2 reactions, the nanocrystalline materials decreased in surface area considerably, for example, from about 100 m2/g to about 15. XRD analysis of the solid residue confirmed that the main solid product was calcium sulfide(oldhamite,JCPDS8,464),andforAP-CaOand[Fe2O3]AP-CaO, this was almost exclusively the product. For the CP-CaO and the [Fe2O3]CP-CaO materials, small amounts of residual CaO could be seen in the patterns (peaks at 32.2, 37.3, and 53.9° 2θ). Essentially the same results were observed in the powder XRD patterns of the samples allowed to react with COS or CS2. To summarize, the reactions of nanocrystalline CaO with COS and CS2 proceed quite well at 500 °C. The main gaseous products are CO2, and the solid products are almost exclusively CaS. D. Decomposition of Chloromethanes. The next set of experiments involved the chlorinated methane analogues methyl chloride (CH3Cl), methylene chloride (CH2Cl2), and chloroform (CHCl3). These experiments were intended to elucidate if CaO destructively adsorbed these materials and how this compared with strictly thermal processes. The ideal reaction ratios were based upon the following equations:
a The breakthrough numbers reported for CS are approximate 2 because 2.0 µL injections were used.
766
All of the nanocrystalline CaO samples worked well as destructive adsorbents for COS and CS2. Nearly all of the samples possessed activities above 50% of theoretical, and several essentially behaved as stoichiometric reagents. There was some difference in activity when [Fe2O3] was present on the surface of the particles. Figure 2 shows the destructive adsorption of COS as a function of injection number for the CaO samples. The diagonal black line is what would be expected for ideal 100% adsorption of each injection. The points where the data deviates from this line is the breakthrough number because at these points unreacted COS is present in the eluent of the reaction. It is seen that, even after 50 injections, the sample has not become completely saturated and is still capable of adsorbing small amounts of COS.
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002
3CaO + 2CHCl3 w 3CaCl2 + H2 + CO + CO2
(5)
CaO + CH2Cl2 w CaCl2 + H2 + CO
(6)
4CaO + 2CH3Cl w CaCl2 + 3CaH2 + 2CO2
(7)
There was essentially no reaction under these short contact time conditions between chloroform and AP-CaO and very little reaction with [Fe2O3]AP-CaO. The reaction with methylene chloride proceeded with better results. It was determined that about 30% of the AP-CaO interacted with the CH2Cl2, while just over 50% of the [Fe2O3]AP-CaO reacted. According to XRD, the solid product for this reaction was calcium chlorite hydrate, Ca(ClO)2‚4H2O (JCPDS 1,1165). In summary, the destructive adsorption of the chlorinated methanes gave mixed results. Neither chloroform nor chloromethane reacted to any great extent. The results for methylene chloride were a little better, with activities of about 30% for AP-CaO and 50% for [Fe2O3]AP-CaO, but they do not approach the results obtained with carbon tetrachloride. It should be noted, however, that the short contact times may be the reason for poor performance. Ongoing work with Freons has shown that sometimes long induction times before onset of reaction are encountered. Obviously, the formation of certain surface sites is necessary, and the complexities of this surface chemistry are currently being investigated and will be reported at a later time.
FIGURE 2. Destructive adsorption of COS at 500 °C as a function of injection number. E. Chlorinated Ethanes and Ethenes. Keeping with the investigation of chlorinated hydrocarbons, the next set of experiments involved chlorinated ethanes and ethenes. Specifically, tetrachloroethene (C2Cl4), tetrachloroethane (C2H2Cl4), and trichloroethene (C2HCl3) were examined. The results will be presented in the following order: C2Cl4, C2H2Cl4, and C2HCl3. The ideal reaction ratios were based upon the following reactions:
2CaO + C2Cl4 w 2CaCl2 + 2CO
(8)
2CaO + C2H2Cl4 w 2CaCl2 + 2CO + H2
(9)
3CaO + 2C2HCl3 w 3CaCl2 + 3CO + C + H2 (10) The reaction of C2Cl4 with AP-CaO and [Fe2O3]AP-CaO gave reaction ratios of 0.36 and 0.46, respectively. The ideal reaction ratio for this the reaction with C2Cl4 would be 0.5, so approximately 72% and 92% of the CaO was accessible for reaction in each of the samples. Thus, there was a small increase in the activity for the [Fe2O3]AP-CaO reagent. The diffraction patterns indicate a definite conversion of CaO to chlorinated product. The AP-CaO reaction displays a mixture of calcium chlorite hydroxide, Ca(OCl)2‚Ca(OH)2 (JCPDS 3,714); calcium chlorite tetrahydrate, Ca(OCl)2‚4H2O (JCPDS 1,1165); calcium chloride (hydrophilite), CaCl2 (JCPDS 24,223); and residual calcium oxide (lime) (JCPDS 37,1497). The [Fe2O3]AP-CaO sample exhibits primarily calcium chlorite hydroxide and small amount calcium chloride. The activities of the two calcium oxide samples toward 1,1,2,2-tetrachloroethane were nearly the same at 0.40 and 0.43 for AP-CaO and [Fe2O3]AP-CaO, respectively. These values are very near to the ideal reaction ratio of 0.50. The diffraction patterns of the two samples were nearly identical with the major solid product being calcium chloride dihydrate (sinjarite) (JCPDS 1,989). The results for trichloroethene were similar. Both calcium oxide samples gave impressive results, although there was not a dramatic difference when [Fe2O3] was present at the surface of the particles. The AP-CaO sample gave a destructive adsorption ratio of 0.58, while the [Fe2O3]AP-CaO sample gave a result of 0.60 (stoichiometric value would be 0.66). The XRD patterns of the reaction products were similar and, again, indicated a large extent of reaction between the test compounds and the CaO. The main solid product was
calcium chloride dihydrate (sinjarite) (JCPDS 1,989) with smaller amounts of calcium chloride (JCPDS 24,223) and calcium chloride tetrahydrate (JCPDS 25,1090). In summary, the destructive adsorption results of APCaO and [Fe2O3]AP-CaO were quite comparable between carbon tetrachloride and the chloroethane/ethenes. As in the case of CCl4, the CaO essentially behaves as a stoichiometric reagent toward the chlorocarbons. The solid products are almost always a form of calcium chloride. Often the calcium chloride is in the form of a hydrate which forms after the reaction by adsorbing atmospheric moisture during workup and acquisition of the powder XRD pattern. F. Nanocrystalline CaO as a Destructive Adsorbent/ Sorbent. It has been shown that nanocrystalline calcium oxides are effective destructive adsorbents for several classes of environmentally problematic compounds. The transitionmetal oxide coated second-generation nanocrystalline calcium oxides have also been utilized as destructive adsorbents for these compounds and often show a remarkable reactivity enhancement due to the presence of a small overlayer of iron oxide. In general, it can be concluded that the AP synthesis produces more active particles than the CP synthesis. It is almost always seen that the AP particles outperform the CP particles as destructive adsorbents. Also, it is often observed that the presence of a small amount of [Fe2O3], as little as 1% by mole, can have a dramatic influence on the reactivity of the nanoparticles. Often, samples with the presence of iron oxide react 1.5-2.0 times more than in the absence of the transition-metal oxide. Another interesting result is that the nanocrystals are capable of behavior as stoichiometric reagents and, thus, should be called “destructive sorbents.” This is a rare occurrence with solid-state reagents. Usually, only the surface of solid reagents will be reactive because of the formation of a protective layer of products. However, some reagents allow for sufficient solid-state diffusion that substantial yields can be obtained. One of the questions that can be asked at this point is why the nanocrystalline calcium oxides are so active. The first factor that has a large influence on the activity of the calcium oxide is the high surface area. High surface area material simply exposes more reagent for reaction. However, surface area is not the only factor. Surely, part of the great activity of the calcium oxide is due to the small size of the particles. This make sense because the small diameters of VOL. 36, NO. 4, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
767
the particles means that reagents need to diffuse only a small distance through the particles in order to access fresh substrate. A gaseous reagent may be able to penetrate the topmost 5-10 “layers” of substrate. For a “macro” particle, these topmost layers may only constitute 5% or less of the total molar mass of the particle. For a nanoparticle, while still only diffusing through the topmost 5-10 layers of the particle, this may be enough to reach the very center of the particle. Another factor that may be in play is the surface morphology. As stated earlier, it is believed that the nanocrystals possess high-index crystal faces. Because of the reduced coordination of the ions on these high-index sites, it is logical that they would be inherently more reactive than the 100 surface. Also, the particles would contain a larger fraction of low-coordination edge and corner sites than larger particles. These are believed to be more reactive than the normal CaO surface. The slight shift in the d spacing of the nanocrystals might give a clue to yet a third factor that may be enhancing the activity of these nanoparticles. This third factor is a result of a combination of the autocompensation of the surface and the small size of the crystallites. It is known that the surface of MgO “rumples” after autocompensation (14). The surface rumples when the oxygen ions contract inward and the magnesium ions expand outward from the surface of the particle. While this rumpling is typically small, between 2% and 8%, this is a dramatic change in a crystallite that is so small. It is conceivable that the rumpling of the surface can an affect on the lattice structure of the small crystals. If there is stress in the lattice of the crystal this would increase the lattice energy of the particles and destabilize the crystals. This stress is indicated by the shifts in the d spacings of the crystallites. The slight change in the crystal lattice and the added stress may act to enhance the diffusion of reagents through the particles. Another question to be answered is what role does the iron oxide play in enhancing the activity of these particles. The first thought is that it is a catalytic surface reagent. The mechanism by which it may catalyze the reaction with carbon tetrachloride (or other reagents) was described in the section discussing carbon tetrachloride. Another possibility is that the difference in the molar density of the iron oxide from the product creates fissures in the product layer during the reaction. The presence of the fissures would allow for additional gaseous reagent to diffuse to fresh reagent (15, 16).
768
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 4, 2002
Acknowledgments The support of the Army Research Office as well as the Hazardous Substance Research Center funded by EPA is acknowledged with gratitude.
Literature Cited (1) (a) Overall Evaluations of Carcinogenicity; Suppl. 7; IARC, International Agency for Research on Cancer: Lyon, France, 1987. (b) Guidelines for Drinking Water Quality 1, 2nd ed.; WHO, World Health Organization: Geneva, Switzerland, 1993. (c) Bouwer, E. J.; Rittmann, B. E.; McCarty, P. L. Environ. Sci. Technol. 1981, 15, 595. (d) Bouwer, E. J.; McCarty, P. L. Appl. Environ. Microbiol. 1983, 45, 1295. (e) An Exposure and Risk Assessment for Trichloroethylene; EPA-440/4-85-019; U.S. Environmetnal Protection Agency, U.S. Government Printing Office: Washington, DC, 1981. (f) Kirk-Othmer Encyclopedia of Chemical Technology, 4th ed.; Wiley: New York, 1994. (2) Koper, O. B.; Lagadic, I.; Volodin, A.; Klabunde, K. J. Chem. Mater. 1997, 9, 2468-2480. (3) Klabunde, K. J.; Stark, J. V.; Koper, O.; Mohs, C.; Park, D. G.; Decker, S.; Jiang, Y.; Lagadic, I.; Zhang, D.; J. Phys. Chem. 1996, 100, 12142-12153. (4) Decker, S.; Lagadic, I.; Klabunde, K. J.; Michalowicz, A.; Moscovici, J. Chem. Mater. 1998, 10, 674-678. (5) Jiang, Y.; Decker, S.; Mohs, C.; Klabunde, K. J. J. Catal. 1998, 180, 24-35. (6) Decker, S.; Klabunde, K. J. J. Am. Chem. Soc. 1996, 118, 1246512466. (7) “Purified” is defined as passing 99.97% pure helium through a series of filters to remove hydrocarbons, water, and oxygen. (8) The CCl4 was dried over a bed of 5 Å molecular sieves and distilled to remove sieve material from the liquid. (9) Li, Y. X.; Klabunde, K. J. Langmuir 1991, 7, 1388. (10) Yang, Y. C.; Baker, J. A.; Ward, J. R. Chem. Rev. 1992, 92, 1729. (11) (a) Beier, R. W.; Weller, S. W. Ind. Eng. Chem. Process Des. Dev. 1967, 6 (1), 380. (b) Graven, W. M.; Weller, S. W.; Peters, D. L. Ind. Eng. Chem. Process Des. Dev. 1966, 5, 183. (c) Graven, W. M.; Paton, J. D.; Weller, S. W. Ind. Eng. Chem. Process Des. Dev. 1966, 5, 34. (d) Tzou, T. Z.; Weller, S. W.; Annual Report to the Army CRDEC, 1988. (12) Lin, S. T.; Klabunde, K. J. Langmuir 1985, 1, 600-605. (13) Li, Y. X.; Schlup, J. R.; Klabunde, K. J. Langmuir 1991, 7, 1394. (14) Duke, C. B. Reconstructions of Solid Surfaces; Christman, K., Heinz, K., Eds.; Springer-Verlag: Berlin, Germany, 1990. (15) A core/shell morphology is most effective for low transitionmetal oxide loadings. However, high loadings of 10-20% for intimately mixed samples of Fe2O3 are also effective for catalyzed destructive adsorption of CCl4. (16) Transition-metal ions with variable oxidation states (V, Mn, Fe) are most effective (5).
Received for review March 13, 2001. Revised manuscript received October 5, 2001. Accepted October 22, 2001. ES010733Z
