pubs.acs.org/Langmuir © 2009 American Chemical Society
Internal Surface Modification of MFI-Type Zeolite Membranes for High Selectivity and High Flux for Hydrogen Zhong Tang and Junhang Dong* Department of Chemical and Materials Engineering, University of Cincinnati, Cincinnati, Ohio 45221
Tina M. Nenoff Surface and Interface Sciences, Sandia National Laboratories, Albuquerque, New Mexico 87185 Received February 7, 2009. Revised Manuscript Received March 18, 2009 MFI-type zeolite membranes were modified by depositing molecular silica at a small number of active sites in the internal surface by in situ catalytic cracking of silane precursor. The limited silica deposition reduced the effective size of the zeolitic channels that dramatically enhanced the H2 selectivity without causing a large increase in H2 transport resistance. The modified zeolite membrane achieved an extraordinary H2/CO2 permselectivity of 141 with a high H2 permeance of 3.96 10-7 mol/m2 3 s 3 Pa at 723 K. The effect of pore modification on the gas transport behavior was studied on the basis of single gas permeation data.
The recent search for high-temperature hydrogen (H2)-permselective membranes has been largely driven by the idea to produce H2 with simultaneous CO2 capture through a single-step water gas shift (WGS) of fossil fuel- and biomass-derived syngas. The highly siliceous zeolite membranes are attracting growing interest because of their necessary sulfur tolerance and hydrothermal stability not possessed by other candidates such as the palladium alloy and amorphous silica membranes.1,2 Currently, the main challenge for the zeolite membranes is the incompatibility between selectivity and flux for H2 separation from the complex gas mixtures involved in the catalytic reaction systems.2 The main components in gas streams from hydrogen production by catalytic conversion of biomass and fossil fuels include H2, CO2, CO, CH4, H2O, and common impurity H2S. These small gases are essentially nonadsorbing in the siliceous zeolites at high temperature (HT) (>573 K). Therefore, HT H2 separation through the zeolite membranes must rely on the differentiation of molecular diffusivity and/or the size exclusion effect depending on the ratio (λ) of the molecular kinetic diameter (dk) to the membrane pore size (dp) (i.e. λ = dk/dp).3-6 In recent years, the MFI-type zeolite membranes with a large Si/Al ratio and the all-silica DDR-type zeolite membranes have been particularly investigated for HT H2 separation.7-11 The primary mass-transport channels in MFI-type zeolites have an *Corresponding author. Phone: (513) 556-3992. Fax: (513) 556-3474. E-mail:
[email protected]. (1) Ockwig, N. W.; Nenoff, T. M. Chem. Rev. 2007, 107, 4078–4110. (2) Dong, J.; Lin, Y. S.; Kanezashi, M.; Tang, Z. J. Appl. Phys. 2008, 104, 121301–121317. (3) Y. Gu, Y.; Oyama, S. T. Adv. Mater. 2007,, 19, 1636–1640. (4) Xiao, J.; Wei, J. Chem. Eng. Sci. 1992, 47, 1123–1141. (5) Krishna, R.; van Baten, J. M. Chem. Eng. Sci. 2008, 63, 3120–3140. (6) de Vos, R. M.; Verweij, H. Science 1998,, 279, 1710–1711. (7) Tomita, T.; Nakayama, K.; Sakai, H. Microporous Mesoporous Mater. 2004, 68, 71–75. (8) Zheng, Z.; Hall, A. S.; Guliants, V. V. J. Mater. Sci. 2008, 43, 2499–2502. (9) Gu, X.; Tang, Z.; Dong, J. Microporous Mesoporous Mater. 2008, 111, 441– 448. (10) Kanezashi, M.; O’Brien, J.; Lin, Y. S. AIChE J. 2008,, 54, 1478–1486. (11) Hong, M.; Falconer, J. L.; Noble, R. D. Ind. Eng. Chem. Res. 2005, 44, 4035–4041.
4848
DOI: 10.1021/la900474y
effective diameter of 0.56 nm, which offers high selectivity by size discrimination for critically sized molecules such as xylene isomers.12 However, the transport of H2, CO2, and other small gases in the MFI zeolitic pores is dominated by activated gaseous diffusion4,10 resulting in high permeance but low selectivity for H2. The DDR-type zeolite has a pore structure of cages connected by small windows with an effective size of 0.4 nm. However, the DDR-type zeolite membrane also exhibited low H2 selectivity with high permeance because of the mesoscale intercrystalline spaces inevitably existing in the polycrystalline film.8 Modifications of the DDR and MFI types of zeolite membranes have been reported in an attempt to enhance the H2 selectivity. The DDR-type membranes were modified by counterdiffusion chemical vapor deposition (CVD) of silica using tetraethyl orthosilicate (TEOS) as a precursor to reduce the size of the intercrystalline pores.8,10 The MFI-type zeolite membranes were modified by catalytic thermal cracking of preadsorbed methyldiethoxysilane (MDES) to deposit molecular silica in the intracrystalline pores and intercrystalline spaces.11,13 The MDES molecule is nearly linear with a kinetic size of 0.4 nm 0.91 nm, which is small enough to enter the zeolitic pores (dp = 0.56 nm). TEOS and tetramethoxyl silane (TMOS) are common precursors for the modification of MFI zeolite external surfaces and intercrystalline pores in membranes by the CVD method because these molecules are too large to enter the intracrystalline MFI zeolite pores.14,15 The TMOS and TEOS molecules are nearly spherical with large sizes of 0.89 and 0.96 nm, respectively. These membrane modifications resulted in significant increases in HT H2/CO2 selectivity but caused unacceptable losses of H2 permeance of about an order of magnitude. Figure 1 presents the (12) Lai, Z.; Bonilla, G.; Diaz, I.; Nery, J. G.; Sujaoti, K.; Amat, M. A.; Kokkoli, E.; Terasaki, O.; Thompson, R. W.; Tsapatsis, M.; Vlachos, D. G. Science 2003, 300, 456–460. (13) Masuda, T.; Fukumoto, N.; Kitamura, M. Microporous Mesoporous Mater. 2001, 48, 239–245. (14) Lu, D.; Kondo, J. N.; Domen, K.; Begum, H. A.; Niwa, M. J. Phys. Chem. B 2004, 108, 2295–2299. (15) Nomura, M.; Yamaguchi, T.; Nakao, S. Ind. Eng. Chem. Res. 1997, 36, 4217–4223.
Published on Web 4/6/2009
Langmuir 2009, 25(9), 4848–4852
Letter
Figure 1. H2 permeance and selectivity over CO2 in zeolite membranes at >473 K (O, MFI unmodified;9,11,16-19 b, MFI modified; 9,11,13 4, DDR unmodified;8,10 and 2, DDR modified8,10).
reported selectivities of H2 over CO2 in the unmodified and modified zeolite membranes, which shows serious incompatibility between H2 selectivity and permeance. The high H2/CO2 selectivity of >100 in the modified MFI-type membranes indicates that certain size exclusion effects between H2 (dk = 0.286 nm) and CO2 (dk = 0.33 nm) can be realized through pore size reduction by the deposition of molecular silica. However, the reduced pore size causes a drastic increase in the H2 transport resistance in the channels that are fully modified over the entire membrane thickness. In this letter, we report the controlled partial modification of the subnanometer channels in MFI-type zeolite membranes to realize high H2 selectivity and meanwhile preserve high permeance as shown in Figure 1.
Experimental Section The MFI zeolite membranes were synthesized on the inner surface of R-alumina tubes (Pall Corp.) as we previously reported.20 The tube is 80 mm long with an i.d. and o.d. of 7 and 10 mm, respectively. The two ends of the tube are sealed by glass covering 10 mm at each end. The inner side of the tube has a 10-μm-thick, 0.2-μm-dp top layer with a porosity of 35-40%. The zeolite membrane was synthesized from an aluminum-free precursor solution containing SiO2, NaOH, H2O, and template tetrapropylammonium hydroxide (TPAOH) by in situ hydrothermal crystallization at 453 K for 20 h. The resultant zeolite membrane had a thickness of 2 to 3 μm. The MFI-type zeolite membranes synthesized under these particular conditions are known to contain a low level of aluminum in the framework as a result of the slight dissolution of the R-alumina surface in the basic synthesis solution.21,22 The small amount of isomorphous substitution of Si4+ by Al3+ in the MFI framework creates a limited number of surface defects in the zeolitic channels (i.e., [(tSi-O-)H+]). The membrane was modified by in situ catalytic cracking deposition (CCD) of MDES molecules at the sites of [(tSiO-)H+] during the operation of H2/CO2 separation. The experimental apparatus and the basic operation procedure for (16) Burggraaf, A. J.; Vroon, Z. A. E. P.; Keizer, K.; Verweij, H. J. Membr. Sci. 1998, 144, 77–86. (17) Algieri, C.; Bernardo, P.; Golemme, G.; Barbieri, G.; Drioli, E. J. Membr. Sci. 2003, 222, 181–190. (18) Min, J. S.; Kiyozumi, Y.; Itoh, N. Ind. Eng. Chem. Res. 2003, 42, 80–84. (19) van de Graaf, J. M.; Kapteijn, F.; Moulijn, J. A. J. Membr. Sci. 1998, 144, 87–104. (20) Gu, X.; Dong, J.; Nenoff, T. M.; Ozokwelu, D. E. J. Membr. Sci. 2006, 280, 624–633. (21) Wegner, K.; Dong, J.; Lin, Y. S. J. Membr. Sci. 1999, 158, 17–27. (22) Kanezashi, M.; O’Brien, J.; Lin, Y. S. Microporous Mesoporous Mater. 2007, 103, 302–308.
Langmuir 2009, 25(9), 4848–4852
membrane modification have been described elsewhere.9 The MDES vapor (1.026 kPa, saturated at 296 K) was carried by an equimolar H2/CO2 mixture (100 cm3/min) flowing over the membrane surface at a pressure of 1.5 bar and a temperature of 723 K. The permeate stream was continuously analyzed by online GC-MS (GC, Agilent GC 6890N and MS Agilent 5975B) to monitor the catalytic cracking products and the pore modification effect on the H2/CO2 separation results. To investigate the mechanism of MDES CCD in the MFI zeolite and confirm the molecular silica deposition at the active sites of the internal surface, MFI zeolite crystals with a Si/Al ratio of 111 were synthesized and modified under different conditions. FTIR examination was performed for the zeolite samples after CCD modification at different temperatures to identify the form of deposited species. The FTIR spectra were recorded on a BioRad Digilab Division FTS 40 spectrometer between 400 and 4000 cm-1. The weight gain of the zeolite during CCD of MDES was measured by a thermogravimetric analyzer (TGA, SDT Q500) at different temperatures to help understand the role of the carrier gas. The zeolite samples modified by MDES and TEOS were examined by the temperature-programmed desorption of ammonia (NH3-TPD) to confirm the silica deposition at the internal active sites. The NH3-TPD measurements were performed by a Micromeritics Autosorb 2910 unit.
Results and Discussion In this work, the in situ CCD modification was done twice with a 10 h annealing period in between. The results of onlinemonitored H2/CO2 separation during the modification process are presented in Figure 2. After the first modification, the H2 permeance decreased from 3.75 10-7 to 2.7 10-7 mol/m2 3 s 3 Pa whereas the H2/CO2 separation factor (SF = (yH2/yCO2)permeate/ (yH2/yCO2)feed) increased from 3.4 to 68. The H2 permeance increased to 3.6 10-7 mol/m2 3 s 3 Pa, and the separation factor stabilized at a value of 57 after 10 h of annealing in the H2/CO2 stream without MDES vapor. The changes in H2 selectivity and permeance during annealing may be attributed to the continued decomposition of the remaining chemisorbed organosilyl species in the membrane porosity. Further enhancement of the membrane separation capabilities can be achieved with a repeated CCD of MDES. The second modification of the membrane resulted in a further increase in the H2 separation factor from 57 to a stable value of 123 under the feed containing MDES vapor, with only a small decrease in H2 permeance to 2.2 10-7 mol/m2 3 s 3 Pa. After the termination of feeding MDES, the H2 separation factor stabilized at ∼108 in about 24 h with a virtually unchanged H2 permeance and a slightly increased CO2 permeance. The stabilization of the H2 selectivity and permeance in the presence of MDES suggest the absence of noncatalytic thermal cracking-deposition of MDES on the membrane external and internal surfaces, which allows for good controllability for the modification. The ability of MDES to modify the intracrystalline pores was also demonstrated through the modification of a disk membrane first with TEOS and then with MDES in the H2/CO2 carrier gas at 723 K. The online-monitored H2/CO2 separation resulting during the membrane modification are shown in Figure 3. After CCD modification with TEOS, the H2 permeance decreased by ∼33%, but the H2/CO2 selectivity increased only from 3.4 to 5.2 because the large TEOS molecules (dk = 0.96 nm) are unable to enter the intracrystalline pores but can access and modify the nanometer-scale intercrystalline pores. After switching the precursor from TEOS to MDES, the H2/CO2 selectivity dramatically increased to 22.6 with only an additional DOI: 10.1021/la900474y
4849
Letter
Figure 2. Separation results for an equimolar H2/CO2 mixture during the CCD modification. (I) Heating from ∼298 to 723 K; (II) dwelling at 723 K; (III) first MDES CCD; (IV) annealing in H2/ CO2 feed without MDES; (V) second MDES CCD; and (VI) H2/ CO2 feed without MDES.
Figure 4. FTIR spectra of MFI zeolite samples (a) before modification, (b) for MDES adsorbed at 423 K, (c) for MDES adsorbed at 573 K, and (d) modified by MDES at 723 K.
∼5% decrease in H2 permeance. The results indicate that there was no silica film deposited on the TEOS-modified membrane surface to block the pore entrances for MDES molecules. The NH3-TPD measurements for the TEOS- and MDESmodified MFI zeolite particles (Si/Al=111) revealed that the number of acidic sites in the MDES-modified sample was substantially smaller than that in the TEOS-modified sample. This suggests that the internal surface acidic sites were eliminated by MDES whereas only the external surface acidic sites were eliminated by TEOS CCD. The zeolite particles (Si/Al = 111) were also examined by BET measurement before and after the CCD modification with MDES. The BET surface area, pore volume, and mean pore size were 454 ( 4.6 m2/g, 0.187 cm3/g, and 0.561 nm, respectively, before modification and were 428 ( 3.8 m2/g, 0.175 cm3/g, and 0.521 nm, respectively, after modification. The small decreases in pore volume and mean pore size indicate limited silica deposition in the zeolite channels, which is the key to maintaining low-transport-resistance high H2 permeance. The mechanisms of CCD of silane molecules on the aluminumcontaining MFI zeolite surface have been studied in the literature.23 In this study, the online GC-MS analysis found H2, CO2, CO, CH4, C2H4, H2O, and C2H5OH in the permeate stream during CCD of MDES. These molecular species are expected
in accordance with the reactions of silane chemisorption and decomposition on the zeolite surface except for CO.13,23 The appearance of CO is likely a result of the reverse WGS reaction of the H2/CO2 carrier gas at elevated temperatures (>573 K) (i.e., H2 + CO2 S CO + H2O). The FTIR spectra of the unsupported zeolite samples (Si/Al = 111) are shown in Figure 4. Silanols (∼3740 cm-1) were found in the fresh zeolite. After adsorbing MDES vapor at 423 K, the silanol peaks disappeared, and peaks for Si-OC2H5 (∼2990 cm-1), Si-H (∼2160 cm-1), and Si-CH3 (∼1405 cm-1) appeared as a result of the chemical and physical adsorption of MDES, which is consistent with the three sorbate species proposed by Masuda et al.13 In the sample treated at 573 K, the peaks for Si-H and SiCH3 disappeared whereas the silanol peak reemerged and the SiOC2H5 peak remained strong. When the modification temperature increased to 723 K, the Si-OC2H5 peak also disappeared with a strong silanol peak remaining at ∼3740 cm-1. This suggests that the final deposits are likely to be (OH)3Si[O-Sit]framework in the H2/CO2 carrier gas.23 The peaks of bridging hydroxyl groups (Si-OH-Al) at 3610 cm-1 and hydrogen-bonded species at ∼3475 cm-1 23,24 existed in all samples with very small variations in intensity because of water adsorption when in contact with the atmosphere during ex situ sample preparation and measurements. Masuda et al.13 suggested that the effective size of the MFI zeolitic channels was 0.36-0.47 nm at locations deposited with molecular SiO2 (after firing at 823 K in air). Therefore, when the deposits are in the form of silanols, the effective pore size is likely to be 573 K may be facilitated by the reverse-WGSproduced H2O. This is evidenced by the results of thermogravimetric analysis for the CCD modification of the MFI zeolite particles with a Si/Al ratio of 111 using dry He and H2/CO2 carrier gases, respectively. The MDES sorption rates were virtually
:: (23) O’Connor, C. T.; Moller, K. P.; Manstein, H. CatTech 2001, 5, 172–182.
(24) Li, L.; Guan, N. Microporous Mesoporous Mater. 2009, 117, 450–457.
Figure 3. H2/CO2 separation results during the CCD modification of a disk membrane using TEOS and MDES.
4850
DOI: 10.1021/la900474y
Langmuir 2009, 25(9), 4848–4852
Letter
Figure 5. Single gas permeance as function of molecular size in the MFI zeolite membrane (a) before and (b) after modification.
identical in the two carrier gases at 423 K where the reverse WGS reaction and decomposition of chemisorbed species do not occur. As the temperature increased to 573 and 723 K, the rate of weight gain was substantially faster in the H2/CO2 carrier gas than in He in the initial stage of treatment. At elevated temperatures, the H2O produced by the reverse WGS reaction can directly participate in the hydrolysis-decomposition23 of the three chemisorbed species in the confined space of zeolite channels. The modified tubular membrane presented in Figure 2 was tested with single gas permeation for H2, He, CO2, Ar, N2, CO, and CH4 in a temperature range of 298-723 K. The permeation experiments were performed at a feed pressure of 2 bar and a permeate pressure of 1 bar without using sweep gas. The results are presented in Figure 5 for four selected temperatures in comparison with the results obtained before modification. At all temperatures, the unmodified membrane exhibited permselectivity (Roi/j = ji/jj, where ji and jj represent √ the pure gas permeance) below the Knudsen factors [RK = (Mw,i/Mw, H2)] for H2 over the other gases. The modification caused a drastic decreases in permeance for the gas molecules with dk > 0.3 nm but caused only a small decreases in permeance for H2 and He, which have dk < 0.3 nm. The H2/CO2 permselectivity was 141 with a high H2 permeance of 3.96 10-7 mol/m2 3 s 3 Pa at 723 K. The cutoff for permeance between H2 and CO2 suggests that the effective pore size of the modified MFI zeolite membrane was close to the CO2 kinetic diameter of 0.33 nm, which agrees with the earlier estimate. It was also observed that, although the decreases in permeance were small for both H2 and He in the modified membrane, the permselectivity between H2 and He reversed after modification. The unmodified membrane was H2-permselective over He because of the dominant Knudsen diffusion mechanism by which the lighter H2 has greater diffusivity. In the modified membrane, gas transport became dominated by the molecular size-dependent activated diffusion in the modified locations of the channel. Thus, the membrane became He-permselective over H2 because He has a smaller kinetic size than H2. In the modified membrane, the permeance of CO2 was higher than that of CH4 but lower than those of Ar, N2, and CO (Figure 5b), which does not follow the order of their nominal molecular dynamic sizes. Recent theoretical studies have revealed that in small-pore zeolites such as LTA and DDR types, molecular diffusivities are in the order of Ar > CO2 > CH4 at 300 K,5 which coincides with the order of permeance in the current modified membrane at 298 K. The modified MFI-type zeolite channels may be thought of as a chamber-and-window structure, where the windows are places deposited with molecular silica. However, because of the few Al3+-associated active sites in the MFI framework, the number of narrow windows formed by silica Langmuir 2009, 25(9), 4848–4852
deposition is rather small, resulting in much lower transport resistance for H2 and He in the modified MFI zeolite as compared to that for the LTA and DDR zeolites. Utilizing the model of molecular transport in zeolites developed in the literature, we undertook the analysis of mass transport for the current membrane system. The pure gas transport diffusivity (Dc) is determined by the molecular load (q) and the jump diffusivity (Do), Dc = qDo.25 For small gases at HT, q may be given by q = βP, where β is a constant of the gas molecule load in the zeolite when ideal gas behavior can be assumed.10,26 The gas permeance is then given by φ j ¼ ðD0 βÞ δ
Do ¼
1 R 8RT 2 -Ed exp z πMw RT
ð1Þ
ð2Þ
where φ is a constant depending on the porosity and tortuosity, δ is the membrane thickness, and R and z are the single jump distance and diffusion coordination number, respectively. Because pore size reduction occurs only in a very small portion of the zeolitic channel, φ, R, and z can be considered to be unchanged after the modification. Thus, the change in the actual molecule load under the nonequilibrium permeation condition (i.e., βafter/βbefore) and the diffusion activation energy (Ed) can be evaluated from the ln j ∼ (1/T) relations before and after modification. The actual β under permeation conditions depends on the rate of molecules entering the pores and the rate of molecules diffusing through the pores. β has the maximum or equilibrium value of (1/ RT)10 when diffusion is the rate-determining step. β decreases when the rate of molecules entering the pores becomes lower than the rate of diffusion in the pores. Table 1 shows the Ed and βafter/ βbefore values calculated from the HT (>573 K) single gas permeance data of this study. The significant increase in the diffusion activation energy for all gases is caused by the reduction of the effective pore size in the modified zeolite membrane. The modification resulted in changes in molecular loads that are different between the group of small molecules (dk < 0.3 nm, i.e., He and H2) and the group of slightly larger molecules (dk > 0.3 nm, i.e., CO2, Ar, N2, CO, and CH4). The β values of H2 and He increased (βafter/βbefore > 1) after membrane modification suggesting that the effect of diffusivity reduction for these molecules overpowered the effect of the (25) Babarao, R.; Jiang, J. Langmuir 2008, 24, 5474–5484. (26) Gu, Y.; Hacarlioglu, P.; Oyama, S. T. J. Membr. Sci. 2008, 310, 28–37.
DOI: 10.1021/la900474y
4851
Letter Table 1. Single Gas Permeation Properties in Membrane before and after Modification molecule dk, nm Mw, g/mol Ed, kJ/mol
before after
βafter/βbefore
He
H2
CO2
Ar
N2
CO
CH4
0.260 4
2.89 2
0.330 44
0.340 40
0.364 28
0.376 28
0.381 16
7.60 10.97
6.56 13.39
3.38 12.55
6.98 14.97
8.23 19.04
4.76 13.14
5.87 9.15
1.46
2.32
0.103
0.176
0.264
0.099
0.027
ji, 10-7 mol/m2 3 s 3 Pa RH20/j
before, at 723 K
4.88
5.34 1.09
1.25 4.27
1.31 4.08
1.43 3.73
1.59 3.36
1.43 3.73
ji, 10-7 mol/m2 3 s 3 Pa RH20/j
after, at 723 K
4.06
3.96 0.97
0.0283 141
0.0609 65
0.0629 63
0.0395 100
0.0220 180
reduction in their rate of entering the pore space. For molecules with dk > 0.3 nm, the drastic reduction in molecular load (βafter/ βbefore , 1) indicates a size exclusion effect that makes the process of molecules entering the zeolite pores the rate-determining step.
Conclusions We demonstrated an in situ surface modification method for zeolite membranes that results in exceptional H2 selectivity over small molecules involved in biomass- and fossil-derived fuel gases. The developed CCD method is effective for the controlled modification of the subnanometer pores in low-Al-content MFI-type zeolite membranes where only a few active sites exist for the deposition of molecular silica species. The thus-modified MFI membrane overcomes the H2 selectivity/flux incompatibility commonly observed in zeolite membranes. The high-temperature H2 selectivity and permeance in the CCD-modified membrane have exceeded the current target values set for microporous porous membranes by the U.S. Department of Energy (RH2/CO2 > 50 and jH2 > 2.7 10-7 mol/m2 3 s 3 Pa, T > 673 K). The membrane reported in this letter is currently being tested
4852
DOI: 10.1021/la900474y
for the high-temperature WGS membrane reaction for simulated coal-derived syngas. Acknowledgment. This research was supported by the Ohio Air Quality Development Authority (AY08-09-C21-N) and the U.S. DOE/NETL (grant DE-FG36-GO15043). Partial support also came from the LDRD of Sandia National Laboratories. Sandia is a multiprogram laboratory operated by Sandia Corp., a Lockheed Martin Company, for the U.S. DOE’s NNSA under contract DE-AC04-94AL85000. Supporting Information Available: Results of NH3-TPD measurements for the TEOS- and MDES-modified zeolites. Results of the thermogravimetric analysis of the CCD modification in different carrier gases. Single gas permeance as a function of temperature. Membrane permeance for H2 and He before and after modification. ln j ∼ (1/T) for pure gases at >573 K. Proposed CCD reaction mechanisms. This material is available free of charge via the Internet at http:// pubs.acs.org.
Langmuir 2009, 25(9), 4848–4852
