Surface Science 0 North-Holland
64 (1977) 349-354 Publishing Company
COADSORPTION OF CARBON MONOXIDE AND HYDROGEN ON POLYC~Y~TALL~~~ PLATINLJb4 Received
in final form
Many studies have been made on the adsorption of pure gases on metal surfaces, but information is not fuIly available on the reaction and replacement of gases on metal surfaces. In order to make clear the nature of residual gases in ultra-high vacuum and catalytic process, it is very important to obtain information about the adsorption and reaction of different species in the presence of each other. Recently, the coadsoqtion of CO and Ha on polycrystalline or single crystal platinum surface has been studied by the combined techniques of thermal flash desorption and electron impact desorption [I I_ and low energy electron diffraction measurements [2j _ There is some uncertainty in our understanding of coadsorption, In the present note, the coadsortion of carbon monoxide and hydrogen on polycrystalline platinum filament has been studied by thermal desorption mass spectrometry in order to obtain further information on the kinetics of coadsorption and replacement of gases on metal surfaces. The experimental system made of glass was baked at about 350°C for 15 h and was evacuated to ultra-high vacuum by using a getter ion pump. The platinum filament used was 0.2 mm in diameter and 250 mm in length. Initially, the platinum fiiament was cleaned by heating at 1300 K for about 10 h in oxygen of 10-e Torr as tried by many jnvestigators f l,3], Before each experiment, the platinum filament was cleaned by thermal flashing to 1300 K. Thermal desorption has been used extensively by many investigators in studying adsorption and desorption of gas on metal surfaces and has been analysed in detail by Redhead  and Ehrlich . Thermal desorption runs were carried out by programming the approximately linear increase of the sample temperature. The filament temperature was increased fram the adsorption temperature (room temperature) to about 1300 K by resistively heating. The filament was heated at rates of abaut 1O-20 K/s. An omegatron type mass spectrometer was used to measure the sample gas compositions and the gas phase pressure during thermal desorption. For thermal desorption of gas, the sample filament was exposed to Rowing gas or gas mixture at a constant pressure at room temperature through bakeable metal 349
K. Kawasaki et al. / Coadsorption
of CO and H2 on polycrystalline
E t ;
Temperature, Fig. 1. Thermal posure time in overlap.
240 120 60
desorption spectra for CO on platinum exposed to CO at 5 X [email protected]
Torr. t,: cxminutes at room temperature. The zero levels of the curves are shifted to prevent
Fig. 2. Thermal desorption spectra for CO on platinum Torr with CO/H, ratio of about 4-5.
to the gas mixture
valves. In the continuous-flow system used here, no significant wall effects were observed during sample heating as already reported [6,7]. Our high pumping speed presumably minimizes these effects . The filament was exposed to flowing CO at pressure of 5 X lo-’ Torr at room temperature. A series of thermal desorption spectra of CO on platinum as a function of the exposure time to CO pressure of 5 X lop8 Torr are shown in fig. 1. It is seen that there are at least two distinct binding states of CO adsorbed on platinum, designated as PI and p2, with the maximum desorption rates at about 430 K (PI) and 540 K (&). This behavior for CO adsorption on polycrystalline platinum is similar to the result obtained by Winterbottom [S]. As the adsorption of CO proceeded, the adsorption states are filled in decreasing order of their temperatures. The desorption of CO from polycrystalline and single crystal platinum surfaces has been characterized in terms of first order processes from several binding states [ 1,8]. As shown in fig. 1, except for the initial adsorption process the peak temperatures observed are independent of the CO exposure time, indicative of a first order desorption process. It may be considered that CO is molecularly adsorbed [ 1,8]. From this assumption, the desorption energy of CO on platinum is estimated to be about 32 kcal/mole for fi2-C0 and 26 kcal/mole for flI-CO, respectively. The filament was exposed to flowing gas mixture of CO and Hz at pressures of 6 X lop9 - 2 X 10e8 Torr at room temperature with CO/H2 ratio of about 2-5. The residual gases except for CO and H2 in the mass spectrometer were Hz0 < -2 X lo-” Torr and CO2 < -3 X lo-” Torr.
K. Kuwasuki ef at. / Coadsorption of CO and Hz on polycrysta&‘ne pt
L.. 60 t,min
: I 400
Temperature Fig. 3. Thermal desorption as in fig. 1.
spectra for hydrogen on platinum surface at the same experiments
A series of CO deso~tion spectra as a function of exposure time to the gas mixture of 6 X 10e9 Torr with CO/Ha ratio of about 4-5 are shown in fig. 2. This figure shows similar behavior as the results mentioned above. With similar experiments as in fig. 2, a series of hydrogen desorption spectra as a function of the exposure time are obtained as in fig. 3. At first the peak height for the &_ state hydrogen,appeared at 380 K, increases with the exposure time to the gas mixture of CO and Ha and then passes through a maximum and decreases. This platinum filament was treated for 5 h at 1300 K in CO of 10M7 Torr in order to study the effects of surface cleaning on platinum. The platinum surface treated here was presumably contaminated with carbon. The desorption spectra of carbon monoxide or hydrogen on the contaminated surface as a function of exposure time to the gas mixture at pressures of 1.4 X lo-’ Torr were observed. The desorption spectra of CO showed a similar tendency to the results as shown in figs. 1 and 2. The desorption spectra of hydrogen are shown in fig. 4. The desorption behavior observed here is complex, indicating multiple binding states for hydrogen in the coadsorption process of CO and Ha. The hydrogen data show roughly two desorption peaks. At first the peak height for the /3~ state hydrogen appeared at the lower temperature side, increases with the exposure time to the gas mixture of CO and Ha, then passes through a maximum and decreases. When the & peak height decreases, the peak for /3~ state hydrogen appeared at the higher temperature side at about 560 K starts to increase. After this contaminated surface was treated by oxygen cleaning as mentioned above, desorption spectra of hydrogen similar to fig. 3 were obtained. The chemical
K, Kawasaki et al. / ~~adso~~ti#n of CO and Hz on ~o~~~c$ystaliineR
a.l b’O tc, min.
20 10 5 7
Fig. 4. Thermal desorption CO of 10e7 Torr.
spectra for hydrogen on platinum surface pretreated
nature of surface hydrogen platinum contaminated with results suggest the existence adsorption of CO and Hz on other observations.
at 1300 K in
giving rise to the desorption peak of PN hydrogen on carbon is not sure, but it is felt that the experimental of a complex state of hydrogen on the surface. The cosuch a contaminated surface will have to be studied by
tc , h Fig. 5. Adsorbed amount of carbon monoxide and hydrogen as a function of exposure time to the gas mixture of CO and H2 at 6 X low9 from figs. 2 and 3.
K. Kawasaki et al. / Coadsorption
of CO and H2 on po~yc~.ysta~l~nePt
Now let us consider the coadsorption of CO and Hz on the clean surface and the replacement of hydrogen adsorbed on the surface by carbon monoxide. Fig. 5 shows the dependence of the adsorbed amount for CO or Hz on clean surface upon the exposure time, using experimental results of figs. 2 and 3. In this case, the adsorbed amount of gas was estimated from the area of the thermal desorption curve. A figure similar to fig. 5 was seen in the result for (CO f Hz) on moiybdenum [9J. From fig. 5, it may be considered that carbon monoxide and hydrogen molecules are adsorbed on the platinunl surface with their partial pressures until the surface is almost saturated with the adsorbates and hydrogen adsorbed on the surface is replaced by carbon monoxide with the further exposure time, tc, more than to to the gas mixture of CO and Hz. It has been assumed by many investigators that hydrogen is adsorbed as atomic species on the platinum surface at room temperature. The rate of replacement of hydrogen adsorbed on the surface by carbon monoxide at t, 2 to may be expressed by -dn/dt,
= kn2 ,
where n is the concentration of hydrogen on the surface, k is the rate constant of desorption of hydrogen molecules, t, is the exposure time, and to is the exposure time at the nlaximum peak of adsorbed hydrogen. From eq. (1) 1,‘n .-- 1,‘rzo= k(t, ~ to)
where no is the concentration of hydrogen at to. A curve of (l/n - I/no) versus exposure time to the gas mixture of CO and H2 is shown in fig. 6, using the experimental results of fig. 5. The (1 /n -- l/no) versus t, plot is linear as a first approximation. From the results, the desorption of hydrogen replaced by CO shows a second order process. The nature of the rate constant k is not sure, although it may be considered from the experiments for replacement of hydrogen by CO on tungsten [lo] that the rate constant depends on the concentration of CO in the gas phase.
Further experiments are in progress to make clear the mechanism sorption of CO and H, on various metal surfaces.
of the coad-
Koji KAWASAKI, Tetsuo KODAMA *, Hirofumi MIKI, and Toshihide KIOKA
Deparment of Physics, Faculty of Science and Technology, Science Umiversityof Tokyo, Noda, Chiba, Japan
[ 11 V.H. Baidwin, Jr. and J.B. Hudson, J. Vacuum Sci. Technol. 8 (fY7I) 44.
[ 21 A.E. Morgan and G.A. Somorjai. Surface Sci. I2 (1968) 405. [3) J. Volter, M. Procop and H. Bern&, Surface Sci. 39 (1973) 453. 141 P.A. Redhead. Vacuum 12 (1962) 203. 151 G. Ehrlich, Advan. Catalysis 14 (1963) 255. [61 N. Hansen, Nuovo Cimento Suppl5 (1967) 389.  P.W. Tamm and L.D. Schmidt, J. Chem. Phys. 51 (1969) 5352.  W.L. Winterbottom, Surface Sci, 36 (1973) 195. f 91 T. Kioka and K. Kawasaki, unpublished data. [lOi J.L. Robins, Trans. AVS Vacuum Symp. 9 (1962) 510.
* Presentaddress: Japan.
industrial Research, Osaka University, Suita, Osaka,