Langmuir 1991, 7, 2140-2145
2140
Catalytic Behavior of Noble Metal/Reducible Oxide Materials for Low-Temperature CO Oxidation. 2. Surface Characterization of Au/MnO, Steven D. Gardner,+Gar B. Hoflund,' and Mark R. Davidson Department of Chemical Engineering, University of Florida, Gainesville, Florida 3261 1
Herbert A. Laitinen Department of Chemistry, University of Florida, Gainesville, Florida 3261 1
David R. Schryer and Billy T. Upchurch NASA Langley Research Center, Hampton, Virginia 23665 Received January 9,1991. In Final Form: April 18, 1991 Au/MnO, and MnO, surfaces used for low-temperatureCO oxidation have been characterized by using ion scatteringspectroscopy(ISS),Auger electron spectroscopy(AES),and X-ray photoelectronspectroscopy (XPSor ESCA) before and after pretreatment in He at 55 O C . The pretreatment enriches the surface of Au/MnO, with 0 and Au and decreases the surface 0 concentration on MnO,. It also causes complex changes in the chemical state of the Mn.
Introduction Recent studies of low-temperature CO oxidation over noble metal/reducible oxide (NMRO) catalysts have identified several promising materials for use in C02 lasers, chemical sensors, and air purification A promising catalyst with regard to activity is gold supported on manganese oxide (Au/MnO,). The reaction data's3 indicate that Au and MnO, interact synergistically exhibiting high long-term CO oxidation activity near ambient temperature with negligible activity decay. However, little is understood about the nature of the Au-MnO, interaction and the corresponding CO oxidation mechanism on these seemingly complex surfaces. Reviews of the utilization of gold in c a t a l y s i ~ l *cite ~*~ numerous investigations of Au particles dispersed on supports such as MgO, A1203, and SiOz. These studies show that the catalytic properties of Au can be significantly altered when supported on various oxides. Relative to unsupported Au? supported Au particles often exhibit a marked increase in surface reactivity particularly toward oxygen-bearing molecules. Numerous experimental resultshave been obtained which indicate that this enhanced reactivity results from an interaction between the Au and the support material. Studies utilizing extended X-ray absorption fine structure spectroscopy (EXAFS)indicate that Au-0 species are present on Au/MgO and Au/A1203 ~ a t a l y s t s .Infrared ~ ~ ~ spectroscopy data taken from Au/ + Present address: Department of Chemical Engineering, Mississippi State University, Mississippi State, MS 39762. (1) Gardner, S. D.; Hoflund, G. B.; Schryer, D. R.; Schryer, J.; Upchurch, B. T.; Brown, D. R. Langmuir, preceding paper in this isaue. (2) Upchurch, B. T.; Kielin, E. J.; Miller, I. M. In Low-Temperature CO Oxidation Catalysta for Long-Life COz Lasers. Proceedings of a workshopheld at NAsALangleyResearchCenter,Hampton, VA, October 17-19, 1989; Schryer, D. R., Hoflund, G. B., Eds.; NASA Conference Publication 3076; NASA Washington, DC, 1989; p 69. (3) Gardner, S. D.; Hoflund, G. B.; Upchurch, B. T.; Schryer, D. R.; Schryer, J.; Kielin, E. J. J. Catal. 1991, 129, 114. (4) Schwank, J. Gold Bull. 1983, 16, 103. (6) Wache, I. E. Gold Bull. 1983, 16, 98. (6) Outka, D. A.; Madix, R. J. Surf. Sci. 1987, 179, 351. (7) Cocco,G.;Enzo,S.;Fagherazzi,G.;Schiffiii, L.;Bassi, I. W.;Vlaic, G.; Galvagno, S.; Parravano, G . J . Phys. Chem. 1979,83, 2527.
MgO yield CO absorption bands which appear at significantly lower wavenumbers than those previously observed on Au/SiOz and Au films.9 Reaction experiments have revealed that Au/SiOz and Au/MgO undergo extensive isotopic oxygen exchange whereas under comparable conditions metallic Au, Si02, and MgO do not.1° Data from Mossbauer spectroscopy and X-ray photoelectron spectroscopy (XPS or ESCA) are consistent with Ausupport interactions as well.11-13 While the nature of the Au-support interaction is not well understood, it appears to be a unique property dependent upon the support material. In fact, a recent study has indicated that strong metal-support interaction (SMSI) behavior characteristic of group VI11 metals supported on TiO2, VZOS,and NbzOs does not occur a t Au/TiOz surfaces.14 Nevertheless, Ausubstrate interactions may still occur for Au supported on other reducible metal oxides, and the degree of interaction may depend upon the presence of impurities which alter surface free energies.15 Although a support effect appears to be well substantiated for a considerable number of Au catalysts, there is evidence that the Au particle size may also affect the reactivity of these surfaces. Experiments with numerous Au catalysts including Au/FezO3, Au/C-Os, Au/NiO, Au/ A&, and Au/Si021s have revealed an overall trend wherein CO oxidation activity increases with decreasing Au particle size. These results are consistent with the observation that small Au particles exhibit an increased affinity for 02.17 However, some exceptions to the former (8) Bassi, I. W.; Lytle, F. W.; Parravano, G. J. Catal. 1976, 42, 139. (9) Schwank, J.; Parravano, G.; Gruber, H. L. J. Catal. 1980,61,19. (10) Schwank, J.; Galvangno, S.;Parravano, G. J. Phys. Chem. 1980, 63, 415. (11) Delgasa, W. N.; Boudart, M.; Parravano, G . J. Phys. Chem; 1968, 72, 3563. (12) Liang, K. S.; Salaneck,W. R.; Aksay, I. A. Solid State Commun. 1976, 19, 329. (13) Batista-Leal,M.; Lester, J. E.; Luccheei, C. A. J. Electron Spectrosc. Relat. Phenom. 1977,11, 333. (14) Shastri, A. G.; Datye, A. K.; Schwank, J. J. Catal. 1984,87,266. (15) Ajayan, P. M.; Marks, L. D. Nature 1989,338, 139. (16) Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. J. Catal. 1989, 115, 301. (17) Cha, D. Y.; Parravano, G. J. Catal. 1970,18, 200.
0743-7463/91/2407-214~~02.50/0 0 1991 American Chemical Society
Langmuir, Vol. 7, No. 10, 1991 2141
Surface Characterization of AulMnO,
trend are observed which may be due to support effects. 10% Au/MnO, ISS It is possible that ultrafine Au particles are a prerequisite for observation of a Au-aupport interaction.18 Under this I n II I assumption the reactivity of supported Au may ultimately bedetermined by the methodof preparation. Apparently, precipitation techniques yield smaller Au particles and more active catalytic surfaces16J7than impregnation. The increased reactivity of small, supported Au particles has renewed interest in Au as a major catalyst constituent. While there has been considerable research on the interaction of Au with relatively inactive supports, Au supported on reducible oxides has received little attention. E/ E. This is probably due to the fact that the catalytic activity Figure 1. ISS spectra obtained from 10 atom 76 Au/MnO, (a) of these materials has only recently been d i s c ~ v e r e d . ~ J ~ J ~ before and (b) after He pretreatment at 55 "C for 1 h. This present paper describes a detailed surface characterization study which has been undertaken to provide information which will be useful in understanding the mechanism of low-temperature CO Oxidation on Au/ MnO,. Auger electron spectroscopy(AES),ion scattering spectroscopy(ISS),and X-ray photoelectronspectroscopy (XPS or ESCA) have been utilized to investigate Au/ B MnO, surfaces in order to determine the composition and chemical states of the surface species and the nature of their interaction during reductive pretreatment.
Experimental Section The catalytic surfaces investigated in this study were 10 atom ?% Au/MnO, (on aAu:Mn basis) anda bare MnO, support. Details regarding their preparation have been described previously.' Briefly, the samples were prepared by coprecipitation from HAuCb and/or Mn(NO3) solution. The appropriate precursor solutions were added dropwise to a stirred solution of Na2C03 at room temperature. After being washed and dried the crushed precipitates were calcined in air at 400 "C for 4 h. The as-prepared samples were pressed into tin specimen cups and inserted into an ultrahigh vacuum (UHV) system (base pressure below 10-10 Torr) for initial surface characterization. In order to correlate the surface characterization data with the CO oxidation activity curves,9 the samples were subsequently transferred to a preparation chamber connected to the UHV system and pretreated in 760 Torr of He for 1h a t 55 "C using a customdesigned platform heating element.20 After pretreatment the samples were returned to the UHV analytical chamber without air exposure for further characterization. Energy analysis for the ISS,AES, and XPS experiments was accomplishedby using a Perkin-Elmer PHI Model 25-270 doublepass cylindrical mirror analyzer (CMA)with an internal, movable aperture which allowed for variation of the polar acceptance angle for incoming particles in ISS. ISS spectra were collected in the nonretarding mode using a 147" scattering angle (fixed by the experimental geometry) and pulse counting detection.21 A 100nA, 1-keV 'He+ primary ion beam was defocused over a 1-cm2 area and spectra collection periods were kept as short as possible to minimize sputter damage. AES was performed in the nonretarding mode using a 3-keV, 10-pAprimary electron beam with a 0.2-mm spot diameter. Survey and high-resolutionXPS spectra were recorded with Mg Ka excitation in the retarding mode using 50- and 25-eV pass energies, respectively.
Results and Discussion The ISS spectra taken from the Au/MnO, and MnO, samples appear in Figures 1 and 2, respectively. The spectra yield information regarding the composition of the outermost atomic layers of (a) the air-exposed surfaces and (b) the surfaces after pretreatment in He at 55 "C. Prominent spectral features due to Au, Mn, and 0 appear in Figure 1 and features due to Mn and 0 are evident in (18)Bond, G.C.;Sermon, P. A. Gold Bull. 1973,6, 102. (19)Kobayashi, T.; Haruta, M.; Sano, H. Chem. Express 1989,4,217. (20)Hoflund, G.B.; Davidson, M. R.; Cordo, G.R. J. Vac. Sci. Technol., A 1991, 9, 2412. (21)Gilbert, R. E.; Cox, D. F.; Hoflund, G. B. Reu. Sci. Znatrum. 1982, 53,1281.
01
0.5
0.0
EIE,
Figure 2. ISS spectra obtained from MnO, (a) before and (b) after He pretreatment at 55 "C for 1 h.
Figure 2. Also, a significant surface concentration of Na is observed. The Na most likely originates from aqueous Na2C03 which was used in the preparation of both samples. The lack of significant spectral features near E/Eo = 0.66 indicates the absence of extensive C1 surface contamination. The potential sources of C1 include the HAuC4 precursor, glassware, and atmospheric species. The small amount of C1 on these surfaces indicates that washing with hot water is effective in removing most of the C1from these surfaces. Suggestionshave been made that Clshould be avoided in the preparation of Pt/SnO, low-temperature CO oxidation catalysts since the use of C1-containing precursors decreases the activity.22 Considering the ISS spectra in Figure 1and the Au/MnO, reaction data,'*$ the preparation method used in this study appears to avoid significant complications due to C1 contamination from HAuCb. The relationship between surface Na concentration and the low-temperature CO oxidation activity of Au/MnO, is currently being investigated. Figures 1 and 2 also indicate that the surface compositions of Au/MnO, and MnO, change as a result of the He pretreatment at 55 "C. This behavior is interesting since these conditions are very mild. The relative amounts of Na on both catalyst surfaces are substantially reduced. On the basis of peak areas (see Table I), the O/Mn concentration ratio of Au/MnO, only slightly increases during pretreatment whereas the O/Mn concentration ratio decreases for MnO,. Furthermore, as shown in Figure 1and Table I, pretreatment results in a significant increase in surface Au concentration on Au/MnO,. Although these ISS data are consistent with some type of interaction (22)Catalpta for LongRange COz Later Sptem. A progrm report from a catalyst development program carried out by GEC Avionics, U.K. Atomic Energy Authority Harwell and UOP, Ltd.,under the auspices of the Royal Sign& and Radar Establishment, November, 1988. Contract NO. SDIO 84-87-C-0046. (23)Davis, L. E.,MacDonald, N. C., Pnlmberg, P. W.,Riach, G. E., Weber, R. E., Eds. Handbook of Auger Electron Spectroscopy; PerkinElmer Corp.: Eden Prairie, MN, 1972.
Gardner et al.
2142 Langmuir, Vol. 7, No. 10,1991 Table I. ISS Peak Area Ratios Au/MnO,
MnO,
air-exposed
pretreated
air-exposed
pretreated
0.26 0.12 0.12
0.30 0.09 0.25
0.30 0.69 nla
0.26 0.40 n/a
O/Mn Na/Mn Au/Mn MS
100
10% AWMnO,
900
500 7w KINETIC ENERGY (ev)
Figure 3. AES spectra obtained from 10 atom % Au/MnO, (a) before and (b) after He pretreatment at 55 O C for 1 h. AES
MnOx
I
I
Mn
O1
Na
Mn
100
300
500 700 KlNEllC ENERGY (ev)
BM)
Figure 4. AES spectra obtained from MnO, (a) before and (b) O C for 1 h.
after He pretreatment at 55
between the Au and MnO,, AES and XPS experiments provide additional information. The AES spectra obtained from Au/MnO, and MnO, both before (a) and after (b) the pretreatment appear in Figures 3 and 4, respectively. Distinct Mn and 0 features appear whereas features due to Au, C1, and Na are present but somewhat less discernible. The AES spectra are in many ways consistent with the ISS data. Firstly, although the Na AES signals are small, the spectra indicate that the Na concentration is greater on the MnO, surface region than the Au/MnO, surface region and that the amounts on both samples decrease during the pretreatment. Secondly, the trend found in ISS and discussed above concerning the surface 0 concentration is observed. After the pretreatment, Figure 3 indicates that the concentration of 0 on the Au/MnO, surface increases relative to Mn whereas Figure 4 indicates that the relative surface concentration of 0 decreases on MnO,. By use of standard sensitivity factors, an estimate of the surface compositions of the two samples after pretreatment (spectra shown in Figures 3b and 4b) indicates an O/Mn atomic ratio which corresponds approximately to Mn304 assuming all of the 0 is associated with the Mn. This is reasonable since only a very small amount would be associated with Na. Thirdly,
Figure 3 depicts an increase in Au concentration after the Au/MnO, has been pretreated. The spectrum shown in Figure 3b indicates a Au concentration near 11 atom 5% based upon Au and Mn only which compares favorably with a Au concentration of 10 atom ?& based upon Au and Mn contained in the preparation solutions. However, the AES spectra in Figure 3 reveal significant features due to C1. Since ISS probes the outermost atomic layer and AES probes at least 20 A beneath the surface, the C1 mostly lies beneath the surface. This is consistent with the assertion that washing the catalyst removes C1 from the surface. Also, Figures 3 and 4 indicate that Au/MnO, contains more C1than MnO,, which is consistent with the suggestion that HAuC14 is the primary source of C1. Similar to that observed for Na, the amount of C1 present decreases after the pretreatment. Finally, the Mn Auger peak shapes change during the pretreatment. The changes are apparent on the high kinetic energy shoulders near 560,610, and 650 eV. This indicates a change in the chemical state of the Mn, and the peak shape changes appear to be similar for both samples. In order t,ounderstand the CO oxidation mechanism on Au/MnO,, it is necessary to determine not only the surface composition but also the chemical states of the surface species. Furthermore, knowledge of the Au and Mn chemical states with respect to both Au/MnO, and MnO, should provide insight regarding the synergistic interaction between Au and MnO,. XPS is a particularly useful technique for obtaining chemical state information, and it complements ISS and AES. On the basis of the data discussed above, Au/MnO, and MnO, respond differently to a pretreatment in He at 55 "C. The surface of MnO, loses 0 whereas the surface of Au/MnO, is enriched in both 0 and Au. The change in the surface 0 concentration may also result in a change in the Mn or Au oxidation state(s) which can be detected by XPS. Manganese-oxygen systems have previously been studied by using XPS.24-28Unfortunately, XPS peak binding energies alone usually are not sufficient to accurately determine the chemical state of Mn in manganese oxides because the Mn 2p features are relatively broad for MnO, compounds so binding energy shifts due to chemical state changes often cannot be observed. This problem is compounded by the fact that manganese oxides are usually complex and may contain Mn which is present in several chemical states along with considerable amountsof water.= As a result, inconsistencies appear in the reported Mn 2p3 2 binding energiesfor manganese oxides.%-* However, additional information is available from XPS which has proven useful in determining the Mn oxidation state. This information consists of the location of Mn 2p shake-up lines, the extent of Mn 3s multiplet splitting, XPS valenceband structure, and XPS 0 1s s p e ~ t r a . ~ ~ - ~ ~ The Mn 2p XPS spectra obtained from Au/MnO, and MnO, taken (a) before and (b) after pretreatment in He at 55 "C appear in Figures 5 and 6, respectively. Although the spectra for both samples appear similar in overall character, there are visible differences between them. The spectra shown in Figure 5 indicate that the Mn 2p3p peaks of Au/MnO, have a binding energy of about 641.1 eV and (24) Oku, M.; Hirokawa, K.; Ikeda, S. J. Electron Spectrosc. Relat. Phenom. 1976, 7,465. (25) Hu, H. K.; Rabalais,J. W. Surf. Sci. 1981, 207, 376. (26) Foord, J. S.; Jackman, R. B.; Allen, G. C. Philos. Mag. 1984,49, 6.51. .
(27) Young, L. 2.;Young, V. J. Electron Spectrosc. Relut. Phenom. 1984,34,45. (28) Castro, V. D.; Polzonetti,G . J. Electron Spectrosc. Relat. Phenom. 1989,48, 117. (29) Levason, W.; McAuliffe, C. A. Coord. Chem. Reu. 1972, 7, 353.
Surface Characterization of AulMnO,
Langmuir, Vol. 7, No. 10, 1991 2143
peaks since strong shake-up lines are observed on these types of surfaces only when stoichiometric oxides are presente26Accordingly,the spectra shown in Figure 5 (and Figure 6) suggest the presence of multiple Mn oxidation states. Spectra similar to those in Figures 5 and 6 exhibiting numerous shoulders on the Mn 2p peaks have been observed for discontinuous films consisting of mixtures of Mn metal and Mn0.27 B The Mn 2p peaks obtained from MnO, appear at slightly higher binding energy (about 641.4 eV) than those obtained from Au/MnO,, and they do not shift during the pretreatment. Furthermore, the shoulders on the high binding '1 2 energy sides of the 2p peaks (arrows 4 and 5 , for example) appear to be more distinct and numerous, again perhaps an indication of the presence of several Mn chemical states. m em a60 040 830 Small shake-up structures (arrows 1 and 2) are visible SNDINQENEMY (ev) approximately 6 eV from the 2p peaks which within Figure 5. XPS Mn 2p spectra obtained from 10 atom % Au/ experimental error are consistent with the presence of MnO, (a) before and (b) after He pretreatment at 55 "C for 1h. MnO. Similar to the Mn 2p features obtained from Au/ MnO,, and Mn 2p spectra shown in Figure 6 exhibit a XPS W X high-intensity background above 660 eV which masks M2P potential shake-up structures indicative of Mn203,Mn304, and MnO2. The XPS Mn 2p binding energies measured for Au/ MnO, and MnO, are near those which have been reported for both MnO and M11304.~~This also supports the assertion that multiple Mn oxidation states are present on these surfaces. However, a shake-up structure approximately 10-11 eV above the Mn 2p1p peak has been shown to be indicative of the presence of Mn304.24128Such a feature is not observed in Figures 5 and 6 which is probably due to the large intensity of the Mn LMM Auger peak. XPS Mn 2p spectra obtained from M n ~ 0 reveal 4~~ that the magnitudes of the shake-up structures are small m em a60 040 690 enough so that they are insignificant against the high BINDINQENERQY (ow intensity of the Mn LMM Auger feature. The shake-up Figure 6. XPS Mn 2p spectra obtained from MnO, (a) before structures should be further attenuated if additional forms and (b) after He pretreatment at 55 O C for 1 h. of Mn are present on a complex surface.26*a Therefore, it is possible that the shake-up features are present but that they do not shift during pretreatment. In addition, too small to be resolved in this study. several features appear on the high binding energy side of Previous studies have shown that Mn oxidation state the 2p peaks. In Figure 5a a small shoulder is present at information may be realized from the extent of the Mn 3s 645.5 eV (arrow 11, and two small shake-up peaks (arrows multiplet ~ p l i t t i n g . ~ ~Unfortunately, *~~~8 XPS Mn 3s 2 and 3) appear which are shifted about 5.5 eV from their spectra can only be obtained from MnO, surfaces because respective 2p peaks. Subsequent to the pretreatment, the Mn 3s peaks of Au/MnO, are obscured by the Au 4f Figure 5b shows that the two shake-up peaks and the peaks. In order to compensate for surface charging, the shoulders remain and a shoulder on the 2~112peak emerges Mn 3s spectra were referenced to the valence band of MnO (arrow 4). Shake-up structures which might correspond which exhibits a peak at approximately 4 eV below the to MnzO3, Mn304, or MnO2 would appear a t a binding Fermi level due to localized 3d s t a t e ~ . ~ ~ This ! ~ 8is a suitable energy near 664 eV.2s Unfortunately, as exhibited in reference since the Mn 3d peak is essentially stationary Figures 5 and 6, the use of Mg K a radiation introduces a for compositions ranging from MnO to Mn2O3.% Charging large Mn LMM Auger peak which would effectively mask is not a severe limitation in this case since the magnitude any shake-up features present in this region of the XPS of the Mn 3s splitting is measured relative to the Mn 3s spectra. Many of the previous e x p e r i m e n t ~ have ~ ~ l ~ ~ ~signal ~ itself. The Mn 3s spectra obtained from MnO, (a) incorporated A1 Kru radiation which removes the Mn LMM before and (b) after pretreatment appear in Figure 7. The Auger peak from the location where these shake-up broad appearance of the spectra particularly after prestructures are observed. Nevertheless, after pretreatment treatment is consistent with the presence of multiple Mn significant changes are observed in the shape of the Mn chemical states. It also appears that the multiplet splitting LMM feature obtained from both samples which indicates features contain contributions from different oxidation a change in the Mn chemical state. This is consistent states of Mn. As a result, it is difficult to determine a with the changes observed in the AES Mn spectra in representative value of the extent of Mn 3s multiplet Figures 3 and 4 discussed above. The small shake-up splitting. The spectrum in Figure 7a, which is the most structures indicated by arrows 2 and 3 in Figure 5 are easily resolved, contains two significant features which probably due to the presence of Mn0.a However, the are separated by approximately 5 eV. This value is close positions of the shoulders designated by arrows 1 and 4 to that which has been measured for Mn304 and Mn203.24 do not correspond to previously measured Mn 2p binding Valence-band XPS spectra obtained from Au/MnO, energies or shake-up peak separations.2c28 Although these and MnO, surfaces are shown in Figures 8 and 9, features are not fully understood, they may be an indication respectively. The spectra were taken from the surfaces of the complexity of these MnO, surfaces. This assumption (a) before and (b) after pretreatment at 55 "C. The shapes is consistent with the small magnitude of the shake-up and relative intensities of features in these spectra yield r n , l , l , , , , , , . " " 1 I , I
O
I
I
I
1
,
1
,
#
1
1
1
1 I I I
I 1 1 I I ? I I I I I I I I I I , ,
B
l . . 1 . / 1 I , . 1 ~ I ~ I I I I I 1 1 I . I / I , I / I I I . I , , . I I 1 I , . I I . . I I I
I
,
Gardner et al.
2144 Langmuir, Vol. 7, No. 10,1991
1oWkRlnO.
XPS 0 la
B
I
w
.
,
.
,
I
.
,
.
,
l
,
,
,
65
I
.
,
.
I
1
l
.
I
l
I
,
l
.
.
,
.
.
j
w
n
.
/
.
/
.
,
,
,
j
,
,
,
,
590 BlNDlNQ ENERGY (ev)
BHDINQENERGY (ev)
Figure 7. XPS Mn 3s spectra obtained from MnO, (a) before and (b) after He pretreatment at 55 "C for 1h.
,
,
,
,
520
Figure 10. XPS 0 1s spectra obtained from 10 atom 5% Au/ MnO, (a) before and (b) after He pretreatment at 55 "C for 1h. I
"
'
.
'
'
>
l
~
l
.
,
l
l
'
-
1
°
~
C
'
"
"
1
"
'
'
xP8 0 18
B
540
Figure 8. Valence-band XPS spectra obtained from 10 atom ?6 Au/MnO, (a) before and (b) after He pretreatment at 55 O C for 1 h.
Figure 9. Valence-band XPS spectra obtained from MnO, (a) before and (b) after He pretreatment at 55 O C for 1 h. information about the chemical nature of these surfaces and changes which occur during pretreatment. The MnO, spectra in Figure 9 exhibit a low density of states (DOS) near the Fermi level (0-10 eV) which is characteristic of an oxide. Two predominant peaks appear in these spectra at about 4 and 7 eV. The former has been attributed to the localized 3d states of Mn and the latter to 0 2p electron^.*^*^'*^^ Small peaks are also present at 14.5,17.5, and 19 eV. After pretreatment the 0 2p feature is significantly reduced, and large changes are observed in
,
l
.
l
,
.
,
,
.
550 Bl"Q ENERaY (eW
l
,
l
l
l
l
,
l
l
620
Figure 11. XPS 0 1s spectra obtained from MnO, (a) before and (b) after He pretreatment at 55 "C for 1 h. the 14-20-eV region which is indicative of a change in Mn chemical state. Both of these observations are consistent with the ISS,AES, and XPS data. The Au/MnO, and MnO, spectra are different in the 0-10-eV range but similar in the 14-20-eV range. As shown in Figure 8, the very large DOS in the 0-10 eV range is characteristic of a metal suggesting that the Au has a metallic nature but the as-prepared sample DOS cuts off sharply near 1eV below the Fermi level. This indicates that the Au in the as-prepared sample is present as an oxide, probably Au2O3. This is further supported by the fact that the as-prepared catalyst is soluble in concentrated HC1 but metallic Au is not. The MnO, peaks contribute in this region, but they are small compared to the Au peaks. The 14-20 eV features are quite similar to the corresponding MnO, features. This point allows for an estimation of the size of the Au peaks compared to the MnO, peaks. Further support for this assertion lies in the fact that the 14-20 eV features on Au/MnO, behave quite similarly to those on MnO, during pretreatment. However, the 7-eV feature obtained from the Au/MnO, surface increases in size during pretreatment. This is consistent with an increase in the surface 0 concentration as observed by using the other techniques. The spectrum in Figure 8b is similar to that of Mn& while the spectrum in Figure 9b resembles that of MnO.nls The 0 1s XPS spectra obtained from Au/MnO, and MnO, (a) before and (b) after pretreatment are shown in Figures 10 and 11,respectively. The binding energies of the predominant features are near 529.1 eV for Au/MnO, and approximately 529.3 eV for MnO,, and they remain
Surface Characterization of AulMnO,
A
Langmuir, Vol. 7, No. 10, 1991 2145 1oY AWnO,
Figure 12. XPS Au 4f spectra obtained from 10 atom % Au/ MnO, (a) before and (b) after He pretreatment at 55 O C for 1h.
unchanged after pretreatment. The binding energies and peak shapes are similar to 0 1s spectra recorded for MnsO,." Shoulders which originate from hydroxyl groups and adsorbed water appear on the high binding energy sides of the predominant oxidic peaks. Similar peak shoulders have been observed in XPS 0 1sspectra obtained from Pt/SnO, surfaces.3(+32 Although the predominant peaks apparently do not change during pretreatment, changes occur with regard to the shoulders on both samples. For MnO, the high binding energy shoulder is decreased in size after pretreatment whereas it is increased for Au/ MnO,. These spectra indicate that adsorbed water and hydroxyl groups desorb from MnO, surface during the 55 "C pretreatment whereas they accumulate during the pretreatment on Au/MnO,. This finding is consistent with the hypothesis that hydroxyl groups play an important role in the low-temperature CO oxidation mechanism on Pt/SnO, and perhaps other NMRO surface^.^^@-^^ In fact, characterization studies of Pt/SnO, catalysts indicate that optimum CO oxidation performance coincides with a pretreatment condition which maximizes the surface Pt(0H)Z concentration.30~34 Given the data above it is possible that the increased 0 concentration on the surface of pretreated Au/MnO, may be due to Au, which is present in an oxidized state, perhaps as Au hydroxide. Therefore, it was anticipated that information from the Au 4f spectra of Au/MnO, might prove beneficial. The XPS Au 4f spectra of Au/MnO, (a) before and (b) after He pretreatment appear in Figure 12. The Au 4f7p peak of the air-exposed Au/MnO, surface is located near 85.3 eV binding energy whereas upon He pretreatment the peak shifts to approximately 85.7 eV. However, when the spectra are superimposed, they appear essentially identical in overall peak shape and peak width. Such behavior is indicative of differential surface charging, and as a result the binding energies of the spectra in Figure 12 may not be representative of the true chemical state of Au. Nevertheless, the binding energies are between (30) Drawdy, J. E.; Hoflund, G. B.; Gardner, S. D.; Yngvadottir, E.; Schryer, D. R. Surf. Interface Anal. 1990, 16, 369. (31) Cox, D. F.; Hoflund, G. B.; Laitinen, H. A. Appl. Surf. Sci. 1984, 20.30. ,~
(32) Hoflund, G. B.; Grogan, A. L., Jr.; Asbury, D. A.; Schryer, D. R. Thin Solid Film 1989,169,69. (33) Schryer, D. R.; Upchurch, B. T.; Van Norman, J. D.; Brown, K. G.; Schryer, J. J. Catal. 1990,122, 193. (34) Gardner, S. D.; Hoflund, G. B.; Schryer, D. R.; Upchurch, B. T. J. Phys. Chem. 1991,95, 835. (35) Schryer, D. R.; Upchurch, B. T.; Sidney, B. D.; Brown, K. G.; Hoflund, G. B.; Herz, R. K. J . Catal. 1991, 130, 314.
those reported for metallic Au at 83.8 e V 6 and Au203 at 86.3 eV.16 A literature search was unable to locate an established binding energy for Au hydroxide. There is additional information to consider with respect to Figure 12 and its interpretation. Investigations of Au deposited on A1203 and Si0212*37 have reported XPS Au 4f peaks which are shifted to higher binding energy (as high as 85 eV) when small Au particles (or islands) with dimensions on the order of 30A or less are present. When the Au particle dimensions became larger, the Au 4f peaks were progressively shifted toward the binding energy of bulk metallic gold. Similar binding energy shifts have been reported recently for other metal/support systems.% These observations have been attributed to matrix effects of the support material and perhaps real differences in the electronic structure of small Au particles relative to bulk Au. The latter is consistent with the increased reactivity of small Au particles. Since the Au particle size distribution has not been measured for the Au/MnO, sample, relationships between the Au particle size and the Au 4f binding energy cannot be assessed in the present study. On the contrary, Haruta and co-workers16have studied a Au/FezOa catalyst which was prepared by a coprecipitation technique similar to the one used in this study. The surface was characterized by ultrafine Au particles (mean diameter of approximately 36 A) and the measured Au 4f7~2binding energy was 83.9 eV. Furthermore, no detectable differences were observed between the Fe 2p and the 0 1s spectra of the Au/FezOa surface and the bare Fez03 support. The Mn 2p and 0 1s spectra of Au/MnO, and MnO, in the present study, however, do exhibit significant differences, and this may correlate with the increased binding energies observed in the Au 4f spectra of Figure 12. These differences are consistent with reaction datal which indicate that Au/MnO, is superior to Au/ Fez03 with respect to measurements of low-temperature CO oxidation activity using stoichiometric CO and 02. Summary Au/MnO, and MnO, surfaces used for low-temperature CO oxidation have been characterized before and after pretreatment in He a t 55 "C by using ion scattering spectroscopy (ISS), Auger electron spectroscopy (AES), and X-ray photoelectron spectroscopy (XPS). The surfaces are complex and appear to contain multiple phases of Mn with an average composition estimated near Mn304. Significant concentrations of Na with much smaller amounts of C1 remain on the surfaces as a result of the preparation. In addition, hydroxylated species and adsorbed water appear to be present. Subsequent to the He pretreatment, the surface of Au/MnO, is enriched in 0 and Au whereas the surface of MnO, becomes oxygendepleted. Coincidentally,the XPS spectra before and after pretreatment exhibit significant differences which may be indicative of a Au-MnO, interaction. However, the extreme heterogeneous nature of the Au/MnO, and MnO, surfaces precludes an exact determination of the Mn and Au chemical states present. The relative surface concentrations observed for Au, Mn, 0,Na, and C1must correlate with the exceptional low-temperature CO oxidation activity exhibited by Au/MnO,, but such a correlation has not been investigated in this study. Registry No. Au, 7440-57-5; MnO,, 11129-60-5;CO, 630-080; He, 7440-59-7. (36) Wagner, C. D., Riggs, W. M., Davis, L. E., Moulder,J. F., Muilenberg, G. E., Eds.Handbook of X-rayPhotoelectronSpectroscopy;PerkinElmer Corp.: Eden Prairie, MN, 1972. (37) Kim, K. S.; Winograd, N. Chem. Phys. Lett. 1975,30,91. (38) Gonzalez-Elipe,A. R.; Munuera, G.;Eepinos, J. P. Surf.Interface Anal. 1990,16, 375.
