Environ. Sci. Technol. 1996, 30, 3593-3596
X-ray Photoelectron Spectra of the Palladium-Iron Bimetallic Surface Used for the Rapid Dechlorination of Chlorinated Organic Environmental Contaminants ROSY MUFTIKIAN,† KENNETH NEBESNY,† Q U I N T U S F E R N A N D O , * ,† A N D AND NIC KORTE‡ Department of Chemistry, University of Arizona, Tucson, Arizona 85721, and Environmental Sciences Division, Oak Ridge National Laboratory, Grand Junction Office, Grand Junction, Colorado 85103
Palladized iron (Pd/Fe) has been successfully used for the rapid dechlorination of organic environmental contaminants in aqueous solutions. We have investigated the nature of the Pd/Fe surface by X-ray photoelectron spectroscopy. Our results indicate that the reactive Pd/Fe surface is formed by the stepwise reduction of Pd(IV) in solution to Pd(II), which replaces protons on the hydroxylated iron oxide surface and forms Pd(II)-O-Fe bonds. These bonds are unstable and collapse spontaneously to yield the reactive palladized iron in which the palladium is in the elemental state. Prolonged exposure of this Pd/Fe surface to a saturated solution of aqueous TCE results in the growth of the hydroxylated iron oxide film that deactivates the Pd/Fe surface. The thick hydroxylated iron oxide film can be removed, and the original activity of the Pd/Fe surface can be restored by washing the surface with a dilute acid solution.
Introduction The deposition of small amounts of palladium (0.05% w/w) on the surface of iron particles results in a bimetallic surface (Pd/Fe) that has a dramatic effect on the reductive dehalogenation of organic environmental contaminants (1). We have shown that aqueous solutions containing 1,1,2trichloroethylene (TCE), 1,1-dichloroethylene, cis- and trans-1,2-dichloroethylene, and tetrachloroethylene are very rapidly hydrodechlorinated to ethane when contacted with the Pd/Fe bimetallic surface. The chloromethanes (CCl4, CHCl3, and CH2Cl2) are also rapidly converted to * Author to whom all correspondence should be addressed. Telephone: (520) 621-2105; fax: (520) 621-1807; e-mail: fernandq@ ccit.arizona.edu. † University of Arizona. ‡ Oak Ridge National Laboratory.
S0013-936X(96)00289-1 CCC: $12.00
1996 American Chemical Society
methane, and as expected, the rate of reaction decreases as the number of chlorine substituents decreases (2). We have shown also that water-methanol solutions of the polychlorinated biphenyls are rapidly dechlorinated on the Pd/Fe bimetallic surface (3). A broad range of chlorinated compounds, including many herbicides and pesticides, have been readily dechlorinated in our laboratory. These reactions have obvious commercial implications, and it is important therefore to investigate the factors that affect the rate of dechlorination. It has become evident from our work that these reductive dehalogenation reactions occur on the bimetallic surface and involve free radicals that are adsorbed on the Pd/Fe surface and, therefore, cannot be detected in solution by electron spin resonance spectroscopy. Our attempts to study these surface reactions by Raman scattering spectroscopy have also been unsuccessful probably because this method has insufficient sensitivity to detect the low concentrations of organic species that are adsorbed on the Pd/Fe surface. The technique that has been successful in probing the Pd/Fe surface is X-ray photoelectron spectroscopy (XPS). We have been able to obtain information about the nature of the iron surface before and after palladium atoms have been deposited on it. We have been able to characterize the surface changes that occur when a chlorinated organic compound such as TCE is completely and reductively dechlorinated on the Pd/Fe surface and converted to ethane. We have observed the progressive growth of oxide layers on the Pd/Fe surface in the course of the reductive dechlorination in an aqueous solution. We have been able to show that these oxide layers can be successfully removed and the reactivity of the surface restored by an acid wash of the Pd/Fe surface with a dilute solution of an inorganic or organic acid. The results of our XPS studies that are reported below should prove useful for the design of pilot and full-scale systems for the removal of VOCs from groundwater.
Experimental Section X-ray Photoelectron Spectra. The XPS measurements were performed by using a Vacuum Generators ESCALAB MKII photoelectron spectrometer (East Grinsted, U.K.) with an AlKR1,2 (1486.6 eV) X-ray source and a hemispherical 150mm mean radius electron analyzer with a take-off angle of 90°. The energy resolution E1/2 (full width at half maximum) of the XPS peaks was less than 1.0 eV for the Ag 3d5/2 peak with the AlKR1,2 source that was operated with a power of 300 W. The binding energies of the photoelectrons were determined by assuming that the carbon 1s electrons had a binding energy of 284.6 eV. The data were recorded digitally, and all peak scans were signal averaged until an acceptable signal-to noise ratio was obtained. During the data acquisition, the pressure in the sample chamber did not exceed 5 × 10-10 Torr. The relative atomic ratios of palladium, oxygen, and carbon on the Pd/Fe surface were calculated from the Pd 3d, O 2s, and C 1s peak areas after a standard background subtraction procedure was applied, and integration of the baseline-corrected peak area was performed. The peak areas were corrected for cross section, excitation, and inelastic mean free path by employing well-established algorithms (4). All reported ratios are relative to the O 2s peak area because this peak was free of any overlap with
VOL. 30, NO. 12, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3593
FIGURE 1. X-ray photoelectron spectra of the iron surface (a) before and (b) after etching with argon ions.
other XPS peaks. The more commonly used and more intense reference peak is the O 1s, but it overlaps the Pd 3p5/2 peak and, therefore, cannot be used to calculate elemental ratios. The calculated atomic ratios are only useful in identifying trends in the experimental results. For example, the Pd:Fe ratio increased as expected, from 0.07 to 0.4, when the iron surface was exposed to potassium hexachloropalladate solution from 1 to 30 s, whereas the O:Fe ratio remained approximately constant. Sample Preparation. A 1 cm × 1 cm pure iron foil, 0.254 mm thick, obtained from Aesar (Ward Hill, MA) was spot-welded to a stainless steel stub, and the surface of the foil was etched with 8 keV of argon ions from an argon ion gun. The argon ion beam was incident on the iron surface at an angle of ∼15° to the surface normal. The iron foil was exposed to a solution of potassium hexachloropalladate in water and to the wash solutions by moving the sample into the load lock chamber. The chamber was brought up to a pressure of 1 atm with ultrahigh-purity argon. The potassium hexachloropalladate solution was added to the sample surface with the aid of a micropipet and was allowed to react for a predetermined length of time. The iron surface was then rinsed with ultrapure water that was added from a second micropipet. The sample was pumped down in the load lock chamber and then repositioned in the sample chamber, and the surface analysis was performed as described above.
Results and Discussion The photoelectron spectrum of the surface of the iron foil before and after etching with argon ions is shown in Figure 1, panels a and b, respectively. Curve fitting to extract the chemically shifted component peaks from the Fe 2p and O 1s X-ray photoelectron spectra obtained from various types of iron samples has been performed by previous investigators, who have deduced that an iron surface exposed to humid air is covered by a multilayer hydroxylated oxide film (5). The outermost layer of this film is composed of FeO(OH); the intermediate oxide layers formed on the elemental iron surface have iron:oxygen ratios varying from 1.5 to 1.3, and the thicknesses of each of these layers were found to depend on the length of exposure of the iron surface to oxygen and water vapor. The Fe 2p spectrum shown in Figure 1a strongly resembles Fe 2p spectra that have been reported previously (5). It is reasonable to conclude, therefore, that the surface of the iron foil used in our work is also covered by a hydroxylated oxide film.
3594
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 12, 1996
FIGURE 2. (a) Binding energies of the Pd 3d3/2 and the Pd 3d5/2 peaks in the X-ray photoelectron spectrum of elemental Pd deposited on the iron surface. (b) Pd 3d peaks of elemental Pd and Pd2+ species after deposition of a very low surface concentration of palladium. The presence of Pd-O-Fe bonds is evident from the binding energies of the Pd2+ species. (c) Pd 3d peaks after the iron surface with the Pd-O-Fe bonds was allowed to stand overnight in an ultra high vacuum.
After the film is removed by etching under vacuum with argon ions, the X-ray photoelectron spectrum of the “clean” iron surface was recorded. The Fe 2p spectrum shown in Figure 1b is that of the clean iron surface or elemental iron in which the binding energies of the Fe 2p electrons are as follows: Fe 2p1/2 ) 720.3 eV and Fe 2p3/2 ) 707.1 eV. Exposure of the clean iron surface to a solution of K2PdCl6 (0.010 g of Pd/100 mL of ultrapure water) for varying lengths of time resulted in the deposition of Pd on the Fe surface as a result of the redox reaction:
PdCl62- + 2Fe0 f 2Fe2+ + Pd0 + 6Cl-
(1)
The longer the time of contact between the iron surface and the palladizing solution, the more palladium is deposited on the iron surface. This has been observed in previous experiments in which 10 µm particles of iron with a very large surface area have been palladized with a solution of K2PdCl6, which is deep orange in color (2). Deposition of Pd on the iron surface via the redox reaction (eq 1) can be followed by observing a rapid decrease in the intensity of the orange color of the K2PdCl6 solution. After the redox reaction has ceased, the deposition of Pd0 on the iron surface is complete, and the solution becomes colorless. The presence of Pd0 on the iron surface, after exposure of the iron surface to the palladizing solution for about 30 s, is confirmed by the binding energies of the Pd 3d peaks in the photoelectron spectrum shown in Figure 2a (Pd 3d3/2 ) 340.3 and Pd 3d5/2 ) 335.0 eV). If, however, the iron surface is exposed to a very low concentration (0.005 g of Pd/100 mL) of the palladizing solution for about 1 s or less, the presence of a higher oxidation state of palladium, in addition to the Pd0, is clearly seen in the X-ray photoelectron spectrum of the iron surface shown in Figure 2b. This indicates that a hydroxylated film is formed on the iron surface as soon as it is contacted with an aqueous solution and that the protons on the -OH groups readily exchange with palladium ions that are present in solution. A similar cation exchange phenomenon has been observed previously where the protons on the
FIGURE 3. Fe 2p photoelectron spectrum of an iron surface covered with a thin hydroxylated iron oxide film. The presence of elemental iron is confirmed by the presence of Fe 2p peaks at 707.1 and 720.3 eV.
-OH groups on an iron surface exchange with K+ ions in solution (5). The Pd-O-Fe bonds that are formed on the iron surface are however unstable, and the palladium ions undergo further reduction to Pd0 by electrons provided by the oxidation of iron that is present in the relatively porous hydroxylated iron oxide film. The spectra shown in Figure 2b,c, which confirm the eventual formation of Pd0 on the iron surface, were obtained when the iron surface with the Pd-O-Fe bonds was allowed to stand overnight in the sample chamber maintained at ultra-high vacuum. It is postulated from these results that Pd(IV) in the palladizing solution is first reduced to Pd2+ ions by the redox reaction:
PdCl62- + Fe0 f Fe2+ + Pd2+ + 6Cl-
(2)
The Pd2+ then exchanges with protons from the hydroxylated iron oxide surface and is reduced further as shown in eqs 3 and 4:
2FeOH + Pd2+ f 2FeOPd + 2H+
(3)
2H+ + 2FeOPd + Fe0 f Fe2+ + 2FeOH + Pd0 (4) The hydroxylated layer on the iron oxide surface has a highly complex structure that probably consists of many types of structures with -OH groups; although in eq 4 Fe0 is shown as the reducing agent, it is entirely possible that iron in some low oxidation state on the iron surface provides the electrons for this reduction reaction. Equations 2 and 4 are merely representations of overall reactions that occur when palladium is deposited by a redox reaction on the surface of the hydroxylated iron oxide surface. There are no significant changes in the Fe 2p spectrum that was recorded in the course of the palladium deposition. The line shapes and peak maxima remain essentially unchanged from those shown in Figure 1a. It may be concluded that the structure of the hydroxylated iron oxide surface remains the same and that the depth of the hydroxylated iron oxide film is also unchanged because the Fe 2p1/2 and the Fe 2p3/2 peaks of the elemental iron surface that lie beneath the hydroxylated oxide film are clearly seen in all the Fe 2p spectra (Figure 3). Two important observations can be made from these data. First, the Fe 2p spectra remain unchanged during and after the
FIGURE 4. Pd 3d photoelectron spectrum of a heavy deposit of elemental palladium on a hydroxylated iron oxide surface. The peak shapes and peak widths show that only elemental palladium is present.
deposition of Pd. Second, after the Pd(IV) is reduced to Pd0, the Pd 3d spectra are unaffected by the presence of iron or oxygen on the surface. It may be concluded, therefore, that the palladium atoms are located in the matrix of randomly distributed iron and oxygen atoms on the iron oxide surface and that there is no electronic interaction between the palladium and the iron and oxygen atoms. In the course of our experiments with palladized iron, it was observed that the rate of hydrodechlorination decreased with continued use of the palladized iron particles. We have attributed this to an increase in thickness of the hydroxylated oxide film, which makes the palladium atoms less accessible to the reactant molecules in the aqueous solution. We have also found that rinsing the surface of the iron particles with a dilute solution of an inorganic or organic acid restores the original activity of the palladized iron surface. We have probed the iron surface with the XPS technique to seek an explanation for our experimental observations. The Pd 3d and the Fe 2p spectra of the iron surface after it is exposed to the palladizing solution (0.05 g of Pd/100 mL of H2O) for 5 min is shown in Figures 4 and 5a, respectively. The binding energies and the shape of the Pd 3d peaks indicate that the palladium is present on the iron surface only in the elemental state, Pd0, and not as Pd0 atoms and Pd2+ ions. It is also evident that these Pd atoms are present in the surface film within a depth of about 40 Å, which is within the limit that can be probed by the XPS technique. The Fe 2p spectrum (Figure 5a) indicates that the surface of the elemental iron is almost completely covered by the hydroxylated oxide film and the presence of elemental iron below this film is barely visible as a shoulder at 706.8 eV. Exposure of this palladized iron film to air for about 3 weeks resulted in an additional growth of the surface film, which almost completely obscured the elemental iron (Figure 5b). The Pd 3d spectrum, however, remained unchanged, thereby indicating that the Pd atoms were accessible to molecules in solution. A freshly prepared palladized iron foil was exposed to an aqueous solution saturated with trichloroethylene until the product from the hydrodechlorination reaction (i.e., ethane) could no longer be detected. The full-scan photoelectron spectrum of the Pd/Fe surface exposed to the aqueous solution saturated with TCE is shown in Figure 6a. A significant decrease in the calculated Pd/Fe ratio, from
VOL. 30, NO. 12, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3595
FIGURE 5. (a) Fe 2p photoelectron spectrum of the hydroxylated iron oxide film. The presence of a shoulder at 707.1 eV indicates the presence of elemental iron in addition to the multiple oxidation states of iron. (b) The Fe 2p photoelectron spectrum of the hydroxylated iron oxide film after its thickness was allowed to increase in moist air for 3 weeks. The Fe 3p peaks of elemental iron are obscured by the thick hydroxylated iron oxide film.
as a result of the prolonged exposure to the aqueous solution of TCE (Figure 6b). It may also be concluded that the elemental palladium is buried in the thick hydroxylated oxide film. A full scan of the photoelectron spectrum of the same sample washed twice with 3 M HCl is shown in Figure 6c. It is evident that the intensities of the Pd 3d peaks have increased by a factor of 3 from 0.02 to 0.06 and that the underlying iron surface has also been exposed by a reduction in the thickness of the hydroxylated oxide film. This is confirmed by a reduction in the O:Fe ratio by a factor of 2 from 0.06 to 0.03. The reactivity of the Pd/Fe surface is almost completely restored by the two-step acid wash; this can probably be attributed to the formation of surface pores to the underlying elemental iron and palladium. In none of these experiments have we observed a loss of palladium from the Pd/Fe surface. The attenuation of the Pd 3d electrons by the hydroxylated oxide layer that has formed on the Pd/Fe surface is responsible for the decrease in the Pd 3d peak area in Figure 6a in comparison with the Pd 3d peak area in Figure 6b.
Acknowledgments This work was supported financially by the Environmental Restoration Programs at the Paducah Gaseous Diffusion Plant and the Portsmouth Gaseous Diffusion Plant, both operated by Lockheed-Martin under contract to the Department of Energy. The authors are grateful to Dames and Moore, Phoenix, AZ; the Materials Characterization Program at the University of Arizona; and the National Institute of Environmental Health Sciences, NIH Grant ES 04940, for additional financial support. The contents of this publication are solely the responsibility of the authors and do not necessarily represent the official views of NIH.
Literature Cited
FIGURE 6. (a) Full-scan photoelectron spectrum of the Pd/Fe surface. (b) Full scan photoelectron spectrum of the Pd/Fe surface after exposure to an aqueous solution of saturated TCE. (c) Full-scan photoelectron spectrum of the Pd/Fe surface in panel b washed twice with 3 M HCl.
0.6 to 0.02, clearly indicates that there is an overall decrease in elemental palladium that is accessible to the aqueous solution of TCE. The Fe 2p peak is obscured by the presence of the hydroxylated oxide film that has increased in thickness
3596
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 30, NO. 12, 1996
(1) Fernando, Q.; Muftikian, R.; Korte, N. Patent pending (patent assigned to Research Corporation Technologies Inc., Tucson Arizona 85711). (2) Muftikian, R.; Fernando, Q.; Korte, N. Water Res. 1995, 29, 2434. (3) Grittini, C.; Malcomson, M.; Fernando, Q.; Korte, N. Environ. Sci. Technol. 1995, 29, 2899. (4) Briggs, D., Seah, M., Eds. Practical Surface Analysis; John Wiley and Sons: New York, 1983. (5) Kurbatov, Cr.; Darque-Cerett, E.; Acouturier, M. Surf. Interface Anal. 1992, 18, 811.
Received for review March 29, 1996. Revised manuscript received July 14, 1996. Accepted July 30, 1996.X ES960289D X
Abstract published in Advance ACS Abstracts, October 1, 1996.
