10938
J. Phys. Chem. C 2008, 112, 10938–10942
Probing the Surface Heterogeneity of Polycrystalline Zinc Oxide by Static Adsorption Microcalorimetry. 1. The Influence of the Thermal Pretreatment on the Adsorption of Carbon Dioxide Xinyu Xia, Jennifer Strunk, Wilma Busser, Raoul Naumann d’Alnoncourt, and Martin Muhler* Laboratory of Industrial Chemistry, Ruhr-UniVersity Bochum, D-44780 Bochum, Germany ReceiVed: January 25, 2008; ReVised Manuscript ReceiVed: May 2, 2008
The adsorption of carbon dioxide on differently pretreated polycrystalline ZnO was studied by thermodynamic and kinetic methods. The uptake of CO2 observed in a static Tian-Calvet microcalorimeter reached saturation at about 5 µmol/m2, corresponding to about half of the exposed Zn2+ sites after a thorough thermal pretreatment at 450 °C for 4 h. The saturation uptake was found to be correlated inversely with the amount of residual hydroxyl groups on the ZnO surfaces. At room temperature, the adsorption of CO2 was found to occur in two steps. Initially, the adsorption was nonactivated, and the exposed surfaces were saturated at a very low equilibrium pressure (p , 1 Pa) with an initial differential heat of adsorption (qdiff) of 100-120 kJ/mol, a standard entropy of -190 J mol-1 K-1, and an adsorption rate constant of 10-5 Pa-1 s-1. During the second stage, an inhibiting effect was observed; the equilibrium coverage increased slowly with increasing pressure, qdiff decreased rapidly with increasing coverage, and the rate of adsorption was low. Temperature-programmed desorption measurements indicated the formation of strongly adsorbed polydentate carbonates at higher temperatures with an adsorption energy between 120 and 160 kJ/mol. 1. Introduction Zinc oxide is an important functional material, which is applied in different fields for various purposes. Due to its semiconducting and optical properties, it is used as a front electrode in thin-film solar cells and as a sensor for hydrogen.1–6 ZnO nanoparticles are applied to improve the properties of rubber7 and also frequently in heterogeneous catalysis. ZnO is an essential component of the ternary copper-based catalysts for the modern low-pressure methanol synthesis process, in which carbon monoxide, carbon dioxide, and hydrogen are converted into methanol, the third-most large-scale chemical product.8,9 ZnO is also a catalyst for the water gas shift reaction, converting carbon monoxide with water into carbon dioxide and hydrogen.10 Carbon dioxide is a useful probe molecule to study the basicity of metal oxide surfaces. Heats of adsorption of CO2 on ZnO were reported in several studies, but the values varied over a broad range. Measurements on polycrystalline samples yielded higher values of 80-120 kJ/mol based on temperatureprogrammed desorption (TPD)11 or the Clausius-Clapeyron equation.12,13 Other than the very old data, the only calorimetrically measured heat of adsorption of CO2 on polycrystalline ZnO amounts to 130 kJ/mol as reported by Auroux and Gervasini.14 Detailed information on structure sensitivity is accessible by studying single-crystal surfaces.15 On the nonpolar ZnO(101j0) plane, which is dominantly exposed on polycrystalline ZnO particles, the adsorption energies in the literature are quite divergent. Both large values of 70-140 kJ/mol16 and small values of 31-45 kJ/mol17,18 were found. Wang et al.19 reported recently for the adsorption of CO2 on ZnO(101j0) that all three atoms in a CO2 molecule interact with Zn and O surface * To whom correspondence should be addressed. Phone +49 234 32 28754. Fax: +49 234 32 14115. E-mail:
[email protected]. URL: http://www.techem.rub.de.
atoms, forming a tridentate carbonate species. The adsorbate structures include a weakly adsorbed (1 × 1) structure with a TPD peak maximum at 200 K using a heating rate of 1 K/s and a more strongly adsorbed (2 × 1) structure with a TPD peak maximum at 320 K. The corresponding desorption energy barriers are 53 and 87 kJ/mol supposing that Ad ) 1013 s-1.19 Thermal pretreatment is a necessary step prior to the adsorption of probe molecules on polycrystalline ZnO samples due to the high reactivity of the exposed surfaces and especially of the hydroxyl groups, which are inevitably present. Polar faces of ZnO are unstable because of electrostatic reasons requiring additional stabilization.20 Recently, it was found that on Znterminated ZnO(0001) faces, either nanoscaled triangular islands with O-terminated steps exist21,22 or a layer of hydroxyl groups is present, 23 depending on the preparation conditions. The oxygen-terminated ZnO(000-1) faces, on the other hand, are found to be essentially always covered by hydrogen atoms.24,25 The desorption energy of water from ZnO(101j0) is as high as 99 kJ/mol, and the elimination of hydroxyl groups from ZnO(101j0) is even more difficult.26,27 The differential heat of adsorption (qdiff) of water on polycrystalline ZnO is much larger than that of CO2; Nagao et al.28,29 reported that the calorimetrically measured qdiff of water on polycrystalline ZnO pretreated at 450 °C changes from over 150 to 100-120 kJ/ mol at different coverages. Yasumoto’s work30 based on the Clausius-Clapeyron equation yielded a much higher isosteric heat of adsorption of water compared to that of CO2 on ZnO; the former amounted to 90-110 kJ/mol at high coverages of 4.8-5.5 mol/m, while the latter decreased to 60 kJ/mol at coverages of 5.0 µmol/m2. Despite the necessity of thermal pretreatment, there is no commonly established procedure for polycrystalline ZnO powder; heating temperatures of 400,14,31–34 450,35 or 500 °C36 were chosen, and the duration was also variable. As a consequence,
10.1021/jp8007464 CCC: $40.75 2008 American Chemical Society Published on Web 07/01/2008
Thermal Pretreatment on the Adsorption of CO2
J. Phys. Chem. C, Vol. 112, No. 29, 2008 10939
the amount of residual hydroxyl groups and strongly bound carbonate species on the investigated ZnO samples may vary significantly, rendering it difficult to compare adsorption results obtained in different studies. For this reason, calorimetric adsorption measurements of CO2 on ZnO samples were performed after a systematic series of different thermal pretreatments. The resulting differences in thermodynamics and kinetics provide clear evidence for the influence of residual surface hydroxyl groups on the adsorption of CO2 on polycrystalline ZnO. The interaction of polycrystalline ZnO powder samples with carbon monoxide is reported in the following second part.37 2. Experimental Section The sample used in this study was nanocrystalline ZnO (NanoTek ZnO, provided by Nanophase Technologies). It was prepared by physical vapor synthesis based on the oxidation of vaporized metallic Zn followed by condensation of ZnO. After heating at 450 °C for 4 h, its specific surface area determined by N2 physisorption applying the BET equation amounted to 10 m2/g. The sample was pressed and sieved, and the sieve fraction of 250-355 µm was used. A metal-tightened U-tube reactor connected to a flow setup was applied for the sample pretreatment as described in our previous work,38 with helium (flow rate 25 NmL/min) or diluted oxygen (10% in argon, flow rate 25 NmL/min) at ambient pressure. All flow rates are reported as normal mL/min (NmL/ min, normal conditions 0 °C, 1013 mbar). The effluent gas mixture was analyzed by an online quadrupole mass spectrometer (GAM 400, Balzers). Three different thermal processes were chosen, (1) 250 °C for 4 h, (2) 450 °C for 30 min, and (3) 450 °C for 4 h using a heating ramp of 5 K/min to reach these temperatures. BET measurements proved that no further loss in specific surface area occurred due to these processes. A thermogravimetric analysis showed that the weight loss for the third procedure is 0.8% and that there is still an additional 0.02% weight loss when the sample is heated from 450 to 1000 °C. The water loss during the pretreatment was about 35 µmol/m2, which is significantly larger than the density of ions on ZnO surfaces. On the basis of the dimensions of the unit cell of ZnO, the density of Zn2+ sites on ZnO(101j0) faces was 9.8 µmol/ m2. Therefore, a water multilayer must have been present on the ZnO surfaces prior to the pretreatment. In addition, the release of CO2 was observed during the thermal treatment with a molar ratio of about H2O/CO2 ) 4/1. After the pretreatment, about 0.5 g of the ZnO sample were directly sealed in a capsule filled with helium and introduced into the sample cell in the microcalorimeter setup, strictly excluding any contact with air. The microcalorimetry system and the measurement procedures were already described in our previous work.39,40 The adsorption temperature was T ) 30 °C. The volume of the adsorption cell was V ) 129 ( 1 cm3, and it was calibrated for each measurement. The purity of the adsorptive gas CO2 was 99.9995%, and N2 (purity 99.9999%) was used as the inert gas for pressure compensation. Additionally, in the microcalorimeter cell, a rehydrated sample was obtained by exposing the ZnO sample pretreated at 450 °C for 4 h to saturated water vapor at 40 °C, followed by evacuating at 120 °C for 72 h. Temperature-programmed desorption (TPD) of CO2 from the ZnO sample was performed in order to obtain more insight into the desorption and adsorption kinetics of CO2. The applied flow TPD setup was described elsewhere.41 An amount of 0.2 g of the ZnO sample was exposed to 4% CO2 (purity: 99.9995%) in
Figure 1. CO2 adsorption isotherms on polycrystalline ZnO as a function of the pretreatment. Thermal conditions: (A) 450 °C, 4 h; (B, C) 450 °C, 30 min; (D) 400 °C, overnight (data from ref 14); (E) 250 °C, 4 h; (F) hydrated. Pretreatment gas: (A, C-E) 10% O2; (B) He. Solid line: best-fitting result for trace A using a uniform energy distribution with εmax ) 108 kJ/mol, εmin ) 70 kJ/mol, and ∆S0 ) -190 J mol-1 K-1.
helium (purity: 99.9999%) at ambient pressure at 200 K followed by flushing in helium and heating with a ramp of 5 K/min. Two ZnO samples were investigated; for the first sample, the flow rate was 10 NmL/min. Two TPD experiments were performed, and during the second measurement, the heating ramp was stopped at 440 K, and the sample was cooled down to 300 K and heated up again to study the TPD from a partially covered surface. For the second sample, the flow rate was 50 NmL/min. 3. Results and Discussion 3.1. Microcalorimetry. Figure 1 shows the equilibrium isotherms of CO2 adsorption on ZnO as a function of the pretreatment. The shapes of the isotherms are quite similar; the surface was saturated at a very low pressure of CO2, which is so small that it appears to be virtually zero for the volumetric analysis in the experiment. The uptake at saturation increased when the pretreatment was performed at higher temperatures and for longer periods of time, reaching a maximum coverage of 5 µmol/m2. Correspondingly, the hydrated sample had the lowest CO2 uptake. The thermogravimetric analysis showed that there is still a further weight loss of 0.02% after the pretreatment 450 °C for 4 h, which may roughly correspond to 1 µmol/m2 hydroxyl groups on the ZnO surfaces. The same value can be derived from the work by Nagao and Morimoto42 on water release from ZnO Kadox formerly supplied by New Jersey Zinc Co. It indicates that the residual strongly bound hydroxyl groups mainly occupy the sites on polar faces and defects, which may account for about 20% of the total ZnO surfaces,36 whereas the nonpolar faces are basically free of OH groups. Therefore, according to the Zn2+ density on ZnO(101j0) faces of 9.8 mol/m and the CO2 saturation amount of 5 mol/m, around half of the exposed Zn2+ sites should be occupied. This estimate agrees quite well with the results by Wang et al.,19 who observed a stable (2 × 1) structure of adsorbed CO2 on ZnO(101j0) at room temperature. When a thermal pretreatment is not strong enough, as demonstrated by the rehydration, also the nonpolar faces are largely occupied by OH groups, leading to a decrease in the saturation uptake of CO2. Figure 1 also shows that there is no obvious difference between the sample pretreated in helium (Figure 1, trace B) and that in dilute oxygen (Figure 1, trace C) for the same thermal process. When using oxygen or reducing gases, only the defects
10940 J. Phys. Chem. C, Vol. 112, No. 29, 2008
Figure 2. Differential heats of CO2 adsorption on polycrystalline ZnO as a function of the pretreatment. Thermal conditions: (A) 450 °C, 4 h; (B, C) 450 °C, 30 min; (D) 400 °C, overnight (data from ref .); (E) 250 °C, 4 h; (F) hydrated. Pretreatment gas: (A, C-E) 10% O2; (B) He. Solid line: best-fitting result for trace A using a uniform energy distribution with εmax ) 108 kJ/mol, εmin ) 70 kJ/mol, and ∆S0 ) -190 J mol-1 K-1.
Xia et al.
Figure 4. TPD spectra of CO2 adsorbed on polycrystalline ZnO. Flow rates: (A) 50 NmL/min; (B, C1, C2) 10 NmL/min. Heating ramps: 5 K/min; at the end of trace C1, the ZnO sample was cooled quickly down to room temperature to desorb CO2 from a partially covered sample (trace C2).
Figure 3. Dependence of qdiff on the dosing pressure of CO2 (circles) during successive doses of CO2.
and the polar faces should be affected, whereas the stable ZnO(101j0) faces should not change significantly. Therefore, the different gas atmosphere used in the pretreatment did hardly influence the adsorption of CO2, which mainly involves the ZnO(101j0) faces. Figure 2 reports the change of qdiff with coverage including the ZnO Kadox data obtained by Auroux and Gervasini.14 The initial value is about 100-120 kJ/mol and independent of the sample type and pretreatment method. When the coverage approaches the saturation value, qdiff decreases rapidly. During the calorimetric measurements, an overlapping endothermic process was observed. When a smaller dosing pressure of CO2 was applied, as shown in Figure 3, an endothermic peak could be more clearly distinguished in the heat flow profile of the calorimeter, and as a consequence, the heats of adsorption derived from Figure 3 became smaller, although the coverage during these doses was not close to saturation. However, this small endothermic contribution can be neglected when the exothermic adsorption is dominating. The endothermic process is presumably due to slow structural changes of the ZnO surfaces induced by the adsorption of CO2. By using a uniform distribution model to describe the energetically heterogeneous adsorption sites on the exposed ZnO surfaces,43 the following thermodynamic parameters originated from a least-squares fit for the ZnO sample after pretreatment at 450 °C for 4 h; the adsorption energy of the most strongly adsorbing sites was εmax ) 108 kJ/mol, that of the weakest adsorption sites was εmin ) 70 kJ/mol, with nm ) 5 µmol/m2 (uptake for a monolayer), and the standard entropy of adsorption
Figure 5. Analyzing the TPD profile (upper plot I) with a distribution of energetically heterogeneous adsorption sites (lower plot II). (a-e): peak labels; solid line in upper plot I: experimental data; dashed line: calculated from the energy distribution derived from the calorimetric measurement; dashed-dotted line: theoretical values using the site distribution shown in lower plot II; dotted lines: contributions from each kind of adsorption site.
was ∆S0 ) -190 J mol-1 K-1. The value of ∆S0 is close to that of CO2 adsorption on MgO (-183 ( 10 J mol-1 K-1),46 which is also a strong adsorption (qdiff > 200 kJ/mol at low coverage). Calculated results based on these fitting parameters are shown in Figure 1 and 2 as solid lines, and the consistent agreement with the experimental data provides further evidence that the assumption of a uniform distribution is valid. 3.2. Temperature-Programmed Desorption. The broad TPD profiles shown in Figure 4 essentially cover the whole temperature range, thus indicating clearly that the exposed ZnO surfaces are energetically heterogeneous. The TPD profiles obtained in different measurements demonstrate that the adsorption and desorption of CO2 are fairly reproducible. The shift of the peak position of the measurements with different flow rates points to the influence of readsorption on the TPD profiles. By integrating the desorbed amount of CO2, the coverage achieved
Thermal Pretreatment on the Adsorption of CO2
J. Phys. Chem. C, Vol. 112, No. 29, 2008 10941
amounts to 3.7 µmol/m2, which is close to the saturation amount in the calorimetric measurements. In the TPD profile (Figure 4, trace B), five desorption peaks can be distinguished, which are marked as peak a to peak e with increasing desorption temperature in Figure 5. Figure 5 also shows that they do not resemble the profile derived from the energy distribution from the microcalorimetric measurement (dashed line). Thus, the energy distribution of the adsorption sites reflected by each of the peaks has to be considered. An equation derived previously is applied for the desorption and readsorption on each kind of site,40 and the amount of energetically different adsorption sites within its own kind (type i) is considered to be uniform regarding the adsorption energy. A CSTR model is assumed to relate the pressure of the desorbed gas (p) to the desorption rate
{
dΘi ) (kap - Ad,iKi) (1 - Θi) dT dΘi mARTa p ) -β nm,i dT V i
β
∑
(
)
(1)
where β is the heating rate (5 K/min), Θ is the fractional coverage, T is the temperature, ka is the adsorption rate constant in the Wigner-Polanyi equation, which is equal to the preexponential factor of adsorption (Aa) for nonactivated adsorption, Ad is the pre-exponential factor of desorption, nm, i is the density of each kind of adsorption site in mol/m2, R is the ideal gas constant, Ta ) 298 K is ambient temperature, m ) 0.20 g is the sample mass, A ) 10 m2/g is the specific surface area, and V˙ ) 10 NmL/min is the volumetric flow rate. Ki is a function of coverage and surface heterogeneity44
Ki )
exp[-(εmax,i - Θεδ,i)/RT] - exp(-εmax,i/RT) 1 - exp[-(1 - Θi)εδ,i/RT]
(2)
where εmax,i is the adsorption energy of the strongest adsorption sites within kind i and εδ,i is the scale parameter of the energy distribution of adsorption site kind i. In the calculation, Aa ) 1 × 10-5 Pa-1 s-1 is applied (see the following subsection), and Ad is determined by43
Ad ) Aap0 exp(-∆S0/R)
(3) mol-1
K-1
where is the standard pressure, ) -190 J is applied for desorption peaks a, b, and c, and ∆S0 ) -200 J mol-1 K-1 is applied for desorption peaks d and e. A best fitting gives the distribution of adsorption sites shown in the lower part of Figure 5. For a better agreement with the experimental TPD profile, a more complex adsorption site distribution model would be required, which is beyond the scope of this investigation on polycrystalline ZnO samples. Although the energy scale and the amount of adsorption sites is flexible to a certain degree, the TPD profile analysis provides more information on the polycrystalline ZnO surfaces. The energies corresponding to peaks a, b, and c match closely with the energy range obtained with the microcalorimeter. The pronounced peak b has an energy range of 72-90 kJ/mol, and on the basis of its narrow shape and the similarity to the desorption energy from ZnO(101j0) (87 kJ/mol19), it is reasonable to correlate this kind of adsorption site with the exposed ZnO(101j0) surfaces of the polycrystalline sample. Peaks d and e imply much higher adsorption energies than that obtained by microcalorimetry. Similar high desorption temperatures were also observed in the literature.11 They may originate from a structural rearrangement of carbonates presumably adsorbed at defects during the TPD measurements. Since p0
∆S0
Figure 6. Kinetics of CO2 adsorption on ZnO pretreated at 450 for 30 min in diluted oxygen. Crosses: experimental data; solid lines: calculated curves based on eq 4 with ka/(10-5 Pa-1 s-1) ) 1.2, 2.5, 5.9, and 8.0 for the curves after dose 1, 3, 5, and 6.
such a restructuring involves both the adsorbed carbonate and the surface atoms, it should be a strongly activated process, which is unlikely to occur during a calorimetric measurement at room temperature but can easily take place at higher temperatures during a TPD experiment. Although a tridentate carbonate can already be formed at room temperature, the orientation of this carbonate species is vertical to the surface,19 having led to an assignment as bidentate species in previous studies, 45 which may rearrange to a more horizontal geometry. This hypothesis is supported by the increase of peak e shown in Figure 4 during the repeated TPD measurement with the same sample. In the infrared study by Saussey et al.,45 an increase of the so-called polydentate carbonate bands (1520 cm-1) was observed when the CO2-covered ZnO was heated up to 450 °C. Unfortunately, it was not possible to perform TPD experiments with varying heating ramps to derive thermodynamic parameters because this overlapping rearrangement of carbonates would make such an analysis unreliable. 3.3. Kinetics of CO2 Adsorption on ZnO. The value of ∆S0 ) -190 J mol-1 K-1 indicates that the adsorbed CO2 species are highly localized. At the experimental temperature, the standard entropy of CO2 is 214 J mol-1 K-1, and consequently, the entropy of the adsorbates amounts to only 24 J mol-1 K-1, meaning that the majority of the adsorbed CO2 species should be bound in the strongly localized polydentate mode. If they were in the monodentate state, then the vibration entropy would be significantly larger due to the contributions of rocking and bending vibrations of the adsorbates. This result agrees with recent DFT calculations identifying the most stable mode for CO2 adsorption on ZnO(101j0) as the tridentate carbonate.19 Surface hydroxyl groups and other residual strongly bound species also affect the adsorption kinetics of CO2 on ZnO. Figure 6 shows the increase in coverage versus time after several doses in one series of measurements on the ZnO sample pretreated at 450 °C in diluted oxygen. When the coverage is far from saturation, the uptake reaches equilibrium within several hundred seconds, and the time dependence can be well described by the following kinetic equation47
(
n(t) ) neq - (neq - n0)exp -kanm
mART t V
)
(4)
where n is the uptake in mol/m2, n0 and neq are the initial and equilibrium uptakes after this dose, respectively, nm is supposed to be 5 µmol/m2, T ) 303 K, V is the volume (129 cm3), and m ) 0.49 g. The adsorption rate constant ka is obtained on the order of magnitude of 10 Pa-1 s-1. When the coverage is close
10942 J. Phys. Chem. C, Vol. 112, No. 29, 2008
Xia et al.
TABLE 1: Kinetic and Thermodynamic Parameters of the Two Stages of CO2 Adsorption on ZnO at Room Temperature
References and Notes
ka / qdiff / ∆S0/ measurements (Pa-1 s-1) (kJ mol-1) (J mol-1 K-1) characteristics
(2) Mu¨ller, J.; Schope, G.; Kluth, O. Thin Solid Films 2003, 442, 158. (3) Go¨pel, W. Prog. Surf. Sci. 1985, 20, 9. (4) Hishimuma, N. ReV. Sci. Instrum. 1981, 52, 313. (5) Krummel, C.; Freiling, A.; Schmidt, R.; Kelleter, J.; Wollnik, H.; Kohl, C. D. Fresenius’ J. Anal. Chem. 1995, 353, 521. (6) Henrich, V. E.; Cox, P. A. The Surface Science of Metal Oxides; Cambridge University Press: Cambridge, U.K., 1994. (7) Wang, J.; Chen, Y. J. Appl. Polym. Sci. 2006, 101, 922. (8) Hansen J. B. In Handbook of Heterogeneous Catalysis; Ertl G., Knoezinger, H., Weitkamp, J., Eds.; VCH: Weinheim, Germany, 1997. (9) Campbell, C. T. Surf. Sci. Rep. 1997, 27, 1. (10) Hinrichsen, O.; Kochloefl, K.; Muhler, M. In Handbook of Heterogeneous Catalysis 2nd ed.; Ertl G., Knoezinger, H., Weitkamp, J., Eds.; VCH: Weinheim, Germany, 2008. (11) Bowker, M; Houghton, H; Waugh, K. C.; Giddings, T.; Green, M. J. Catal. 1983, 84, 252. (12) Hart, P. M. G.; Serba, F. Trans. Faraday Soc. 1960, 56, 551. (13) Kokes, R. J.; Glemza, R. J. Phys. Chem. 1965, 69, 17. (14) Auroux, A.; Gervasini, A. J. Phys. Chem. 1990, 94, 6371. (15) Wo¨ll, Ch. Prog. Surf. Sci. 2007, 82, 55. (16) Hotan, W.; Go¨pel, W.; Haul, R. Surf. Sci. 1979, 83, 162. (17) Esser, P.; Go¨pel, W. Surf. Sci. 1980, 97, 309. (18) Go¨pel, W.; Bauer, R. S.; Hansson, G. Surf. Sci. 1980, 99, 138. (19) Wang, Y.; Meyer, B.; Qiu, H. S.; Kotsis, K.; Staemmler, V.; Traeger, F.; Langenberg, D.; Muhler, M.; Wo¨ll, Ch. Angew. Chem., Int. Ed. 2007, 46, 7315. (20) Tasker, P. W. J. Phys. C 1979, 12, 4977. (21) Dulub, O.; Diebold, U.; Kresse, G. Phys. ReV. Lett. 2003, 90, 016102. (22) Dulub, O.; Boatner, L. A.; Diebold, U. Surf. Sci. 2002, 519, 201. (23) Valtiner, M.; Borodin, S.; Grundmeier, G. Phys. Chem. Chem. Phys. 2007, 9, 2406. (24) Kunat, M.; Gil Girol, St.; Becker, Th.; Burghaus, T.; Wo¨ll, Ch. Phys. ReV. B 2002, 66, 081402. (25) Meyer, B. Phys. ReV. 2004, 69, 045416. (26) Meyer, B.; Marx, D.; Dulub, O.; Kunat, M.; Langenberg, D.; Wo¨ll, Ch. Angew. Chem., Int. Ed. 2004, 43, 6641. (27) Wang, Y.; Muhler, M.; Wo¨ll, Ch. Phys. Chem. Chem. Phys. 2006, 8, 1521. (28) Nagao, M.; Morimoto, T. J. Phys. Chem. 1969, 73, 3809. (29) Nagao, M.; Yunoki, K.; Muraishi, H.; Morimoto, T. J. Phys. Chem. 1978, 82, 1032. (30) Yasumoto, I. J. Phys. Chem. 1984, 88, 4041. (31) Giamello, E.; Fubini, B. J. Chem., Soc. Faraday Trans. 1 1983, 79, 1995. (32) Dent, A. L.; Kokes, R. J. J. Phys. Chem. 1969, 73, 3772. (33) Baran´ski, A.; Galuszka, J. J. Catal. 1976, 44, 259. (34) Chiorino, A.; Ghiotti, G.; Boccuzzi, F. Vacuum 1990, 41, 16. (35) Chang, C. C.; Dixon, L. T.; Kokes, R. J. J. Phys. Chem. 1973, 77, 2634. (36) Scarano, D.; Spoto, G.; Bordiga, S.; Zecchina, A.; Lamberti, C. Surf. Sci. 1992, 276, 281. (37) Xia, X.; Strunk, J.; Busser, W.; Naumann d’Alnoncourt, R.; Muhler, M. J. Phys. Chem. C 2008, 112, 10931. (38) Naumann d’Alnoncourt, R.; Bergmann, M.; Strunk, J.; Lo¨ffler, E.; Hinrichsen, O.; Muhler, M. Thermochim. Acta 2005, 434, 132. (39) Naumann d’Alnoncourt, R.; Kurtz, M.; Wilmer, H.; Lo¨ffler, E.; Hagen, V.; Shen, J.; Muhler, M. J. Catal. 2003, 220, 249. (40) Xia, X.; Naumann d’Alnoncourt, R.; Strunk, J.; Litvinov, S.; Muhler, M. J. Phys. Chem. B 2006, 110, 8409. (41) Strunk, J.; Naumann d’Alnoncourt, J.; Bergmann, M.; Litvinov, S.; Xia, X.; Hinrichsen, O.; Muhler, M. Phys. Chem. Chem. Phys. 2006, 8, 1225. (42) Nagao, M.; Morimoto, T. J. Phys. Chem. 1980, 84, 2054. (43) Xia, X.; Litvinov, S.; Muhler, M. Langmuir 2006, 22, 8063. (44) Xia, X.; Strunk, J.; Litvinov, J.; Muhler, M. J. Phys. Chem. C 2007, 111, 6000. (45) Saussey, J.; Lavalley, J. C.; Bovet, C. J. Chem. Soc., Faraday Trans. 1 1982, 78, 1457. (46) Beruto, D.; Botter, R.; Searcy, A. W. J. Phys. Chem. 1987, 91, 3578. (47) Xia, X.; Naumann d’Alnoncourt, R.; Strunk, J.; Litvinov, S.; Muhler, M. Appl. Surf. Sci. 2007, 253, 5851. (48) Pajares, J. A.; Garcia Fierro, J. L.; Weller, S. W. J. Catal. 1978, 52, 521.
stage 1 stage 2
10-5 ,10-5
80-120 ,80
-190
nonactivated activated
to saturation, it takes several hours to reach equilibrium, and the temporal course cannot be described by the simple eq 4. The strong increase from 10 to 104 s indicates that there is an overlapping process which is much slower than adsorption. Presumably, this process is related to the diffusion of residual strongly bound species or to structural changes of the adsorption sites. On the basis of the changes of the adsorption rate and the heat of adsorption of CO2 on ZnO, two stages of adsorption of CO2 on ZnO at room temperature are identified, as summarized in Table 1. During the first stage, CO2 adsorbs on vacant Zn-O pairs on the empty surfaces, which is therefore an essentially nonactivated adsorption. During the second stage, the number of empty surface Zn-O pairs is limited, and the adsorption of CO2 becomes an activated process. Residual hydroxyl groups may contribute to this process in the following way; first, bidentate adsorption of CO2 on O2- occurs, then OH groups diffuse from the neighboring Zn2+ ions onto vacant neighboring sites, and finally, the bidentate carbonate species are converted into a strongly adsorbed tridentate one. The effect of surface OH groups on CO2 adsorption was also found in the study on other metal oxides, such as scandia, which can adsorb more CO2 when pretreated at a higher temperature.48 Structural changes at high degrees of coverage and the activated population of defect sites may also contribute to the lowered rate and heat of adsorption. Further studies employing FTIR spectroscopy are in progress to identify differently bound carbonate species. 4. Conclusions The adsorption of CO2 on ZnO at room temperature leads to polydentate carbonates as the majority surface species, and at saturation, about half of the Zn2+ sites are occupied. Surface hydroxyl groups hinder the adsorption of CO2 on polycrystalline ZnO and decrease the saturation amount of adsorbed CO2 on the surfaces. When the carbonate coverage is not close to saturation, the inhibiting effect of surface OH groups is weak. The adsorption is nonactivated in this case, qdiff is 80-120 kJ/ mol, the adsorption rate constant is about 10 Pa-1 s-1, and the standard adsorption entropy is about -190 J mol-1 K-1. When the coverage is close to saturation, the activation energy of adsorption increases, and the rate of adsorption decreases. Therefore, CO2 adsorption experiments on ZnO are strongly affected by the thermal pretreatment. Only in TPD measurements was a higher adsorption energy of over 150 kJ/mol observed, representing polydentate carbonates presumably bound at steps which were formed at a higher temperatures during the TPD experiments. Acknowledgment. The authors thank Bernd Meyer, Yuemin Wang, and Christof Wo¨ll for fruitful discussions and the Deutsche Forschungsgemeinschaft (DFG) for financial support within the Collaborative Research Center (SFB 558) “MetalSubstrate Interactions in Heterogeneous Catalysis”.
(1) Hosono, W.; Fujihara, S.; Kimura, T. Key Eng. Mater. 2002, 216, 69.
JP8007464
