- Email: [email protected]
The rate of formation of nickel carbonyl from carbon monoxide and nickel single crystals
Holland Publishing Company
S U R F A C E S C I E N C E LETTERS T H E RATE OF F O R M A T I O N OF N I C K E L CARBONYL F R O M CARBC M O N O X I D E AND NICKEL S I N G L E CRYSTALS K. LASCELLES and L.V. R E N N Y INCO Europe Ltd., Clvdach, Swansea, SA6 5QR, UK Received 10 November 1982
We have measured the rate of Ni(CO) 4 formation at a total pressure of 1 atm on Ni(100), ( 1 and (111) single crystals cleaned by argon ion bombardment and UHV annealing. The (1 surface reacts significantly faster than the other two surfaces. The partial pressure dependence the reaction and the effect of impurities are also discussed.
The reaction between nickel and carbon monoxide is very importa forming the basis for a commercial process for the refining of nickel, t example, I N C O Ltd. operates refineries at Clydach, Wales, UK, and at Copl Cliff, Sudbury, Ontario, producing many thousands of tonnes of high put nickel annually by processes based upon this reaction. There have been ma studies of the reaction Ni + 4CO --* N i ( C O ) 4 , but only recently has nickel b~ used in the form of single crystals. Some of this work was stimulated by clai that a magnetic field affects the rate of reaction [1-5], although subsequ work by several groups of workers have failed to reproduce any such eff [6-8]. Observations of faceting on the surface of low index nickel sin crystals during the reaction are also relevant to interpretations of the react mechanism on the atomic scale [9]. In this paper we report for the first time measurements of the rate carbonylation of three low index nickel single crystals cleaned by argon i b o m b a r d m e n t and U H V annealing. Unlike previously reported work, crystals gave significantly different reaction rates in a short time. Nickel single crystal discs, 25 m m diameter and 0.7 m m thick, of (1¢ (110) and (111) orientation, from Metals Research Limited were used. crystal was mechanically polished to 0.25 /zm diamond and then elect polished before mounting in a small glass reaction vessel which formed parl an U H V system. It was cleaned on both sides by many cycles of argon : b o m b a r d m e n t (10 3 Torr argon, 600 eV, 3 /~A cm -2, 10 min) and anneal (10 -9 Torr, 700°C, 20 min). The crystal was heated with an external RF c 0039-6028/83/0000-0000/$03.00 © 1983 North-Holland
16~
K. Laseelh,s. L. V. Renny / Rate of formation of nickel earho,(vl
Crystals p r e p a r e d in this way a p p e a r e d perfectly smooth in a scanning e microscope at magnifications of up to 30,000. It could not be d e m o n that they were atomically clean, but it is k n o w n that this cleaning tec gives atomically clean surfaces [10], while deliberately a d d i n g impuriti c o m m o n l y c o n t a m i n a t e " c l e a n " nickel surfaces (e.g. carbon, sulphur or ¢ gave quite different c a r b o n y l a t i o n characteristics. A t h e r m o c o u p l e was spot welded to the crystal s u p p o r t i m m e d i a t e l y ent to the crystal. The true crystal temperature, d e t e r m i n e d in a sc calibration experiment with a d d i t i o n a l thermocouples spot welded crystal, was only a few degrees higher. The true crystal t e m p e r a t u r e is t quoted. C a r b o n m o n o x i d e was purified by passing over heated copper, s u p p o r t e d p l a t i n u m catalyst at 200°C, over p a l l a d i u m and finally thr, cold trap at - 165°C. The c a r b o n m o n o x i d e contained < 1 p p m oxyg~ no i m p u r i t y measurable by a q u a d r u p o l e mass s p e c t r o m e t e r a t t a c h e d reaction vessel was detected. A r g o n was purified to < 1 p p m total im t using a BOC rare gas purifier. The nickel crystals were c a r b o n y l a t e d in a flow system. The concen of nickel c a r b o n y l was measured by UV s p e c t r o p h o t o m e t r y , using a silica cell glass-blown to the U H V system. 0.2 p p m of nickel carbonvl c( measured at a wavelength of 225 nm. Reaction rates were flow d e p e m low carbon m o n o x i d e flows but were i n d e p e n d e n t of the c a r b o n mo flow rate at 400 ml min l, c o r r e s p o n d i n g to a gas velocity of 0.34 ( across the crystal. All measurements were m a d e using this flow rate. A f t e r only a few minutes of reaction the crystals gave c a r b o n y l a t i o that were steady with time for up to 6 h at constant t e m p e r a t u r e and pr Fig. 1, curve (a), shows the reaction rate on a function of time for i", This type of b e h a v i o u r is different from any previously r e p o r t e d : t ferences are p r o b a b l y due to different methods of crystal p r e p a r a t i o n . discussed later. Table 1 shows the specific reaction rates at the t e m p e r a t u r e of the ma c a r b o n y l a t i o n rate, T,.... , which was 130_+ 2.5°C for all crystals. The i m a g n i t u d e s of the reaction rates for the single crystals do not bear a relationship to the atomic packing densities of the surfaces. By me electron m i c r o s c o p y we established that no faceting occurred on an}' crystals, although the total a m o u n t of material removed was always sm~ a p p a r e n t activation e n e r g y for the reaction was 8.5 +_ 0.5 kcal mole l three crystals in the t e m p e r a t u r e range 5 0 - 1 0 0 ° C . Diluting the c a r b o n m o n o x i d e with argon, at a total pressure a t m o s p h e r e , gave the pressure d e p e n d e n c e of the reaction rate. The r~ rate was p r o p o r t i o n a l to P~go; a d e p e n d e d on temperature~ and was I 65°C, 1.9 at 130°C and 2.9 at 160°C for all three crystal surfaces. It is well k n o w n that small a m o u n t s of impurities can affect the carl:
~e .~ c m
-2
m l n "¸
4xlO -9 (b)
2 x l O -9
(a)
2o
4O
60 Time,
8O
lC~
min
Fig. 1. Effect of oxygen on the c a r b o n y l a t i o n of Ni(110) at 130°C. (a) < 1 p p m oxygen in carl monoxide. (b) 3 p p m in c a r b o n monoxide.
tion reaction. W h e n using carbon monoxide containing 3 p p m of oxygen fact, before the platinum catalyst was added to the carbon monoxide purifi tion system), the results were as shown in fig. 1, curve (b). The reaction J was clearly greater than with the oxygen-free carbon monoxide but the valu, Tin, ~ was unchanged. Sulphur was pre-adsorbed by exposing the crystal to hydrogen sulphid 65°C; the exposure was in excess of that required to saturate the surface [ The initial reaction rate increased by a factor of twenty five, but the decayed with time. Also T,,,ax decreased to 80°C and the activation em decreased to 5.4 kcal mole-~ (30-60°C). Pressure dependence measurem on the sulphided crystal gave values of ~x = 1.06 at 59°C and 2.1 at 80°C. Table 2 shows a summary of data obtained in different laboratories on
Table 1 R a t e s of nickel c a r b o n y l f o r m a t i o n at 130°C Crystal
Specific reaction rate (mole cm -2 r a i n - 1)
N i ( l 11) Ni(100) N i ( I 10)
4.4×10 9 1.4× 10 9 1.7 X 1 0 - 9
_ ~
X X
,~
× X X
~
~°
,~"
,
X
X X ×
~
o
t
o
r o
i
.-u
E
o
c-. r,.-~
o r-,i
z~ Z
,--.
oo
o,
~" ~.,
:.E I.-.
carbonylation of nickel single crystals. All measurements were made at atm, spheric pressure apart from those of Greiner and Menzel [12] who worked 316 mbar pressure. Kirnchik et al. [3] used a discontinuous chemical method determine nickel carbonyl, Mehta et al. [7] used a photoionisation detector b all other workers have used UV spectrophotometry. Each group of workers h used a different method of surface preparation. It is unlikely that reducti~ removes all impurities from the nickel surface, and probably leaves the out few atomic layers damaged, with reactive sites [13]. This would account for high initial carbonylation rate decreasing to a more or less steady rate observ by most workers. Our method of argon ion b o m b a r d m e n t and annealing is standard method of preparing a nickel surface for L E E D examination, with damaged outer layers [10]. We observe what one might expect for the carbo~ lation of perfect crystal surface, a rapid increase to a constant carbonylati rate. C o m p a r i n g data in table 2, there seems to be general agreement that T,..... about 130°C; Greiner and Menzel's value of 120°C, obtained at a worki pressure of 316 mbar would be about 10°C higher at atmospheric pressure. lower value of Tm~~ could well indicate sulphur contamination of the surfa There is considerable scatter in the apparent activation energy determir by different groups of workers. It is impossible to assign a "correct" value, [ the majority of results on all forms of nickel [9] lie within the range 5 - 1 2 k mole ~. One must also note the problem of diffusional resistances in kinetic measurements. In this present study these were overcome by workin~ a sufficiently high gas velocity across the crystal surface. It is impossible to whether this was true for the other studies cited. Our measurements of the partial pressure dependence of the reaction r do not agree with those of De G r o o t et al. [9]. We agree with the results L o s s m a n n [14] on porous nickel sheet and of Trivin [15] on nickel pow~ L o s s m a n n obtained a value of a = 1 at < 70°C, and calculated that c~ incre~ with temperature and is equal to 4 at equilibrium. Trivin obtained a valu( ~x = 0.85 for P c o = 0,1-1 atm at 4 0 - 9 0 ° C . Trivin proposed a reaction sch~ where Ni(CO) 2 is the largest surface species formed, and the rate c o n t r o l process is either the k a n g m u i r - H i n s h e l w o o d type interaction between adja~ N i - C O species to form Ni(CO)2 or a combination of this step and desorption of Ni(CO)2 from the surface. This is the only scheme consis~ with his measured value of a, with which we agree. In the presence of oxygen, as shown in fig. 1, the reaction rate goes thro a maximum, and then decays to a steady value. This shows that tJ impurities in the gas phase can significantly affect the reaction. Presum~ some of the oxygen is chemisorbed and incorporated into the nickel lattice analogy with sulphur, which is of course in the same group as oxygen in Periodic Table, the presence of oxygen activates nickel atoms for carbon tion.
t~ /~
ix,
~,~.t-tt(:,~,
l~, v .
Ixc'f~rtk
/
lx.ut~
¢JJ J(Jlgflggll(.'rl
*.JJ , ' l l ~ . i ~ l
(uf{)(lfl}l
Of the five sets of data on the absolute rate of the carbonyl fo reaction, four are very similar after allowing for the differences in the t, ture and pressure of the measurements. The values obtained by Krinch are about two orders of magnitude greater, which together with tl~ activation energy and low value of T..... would suggest sulphur contan of the crystal surface. The values given by De Groot et al. [9] d ° not re true values for each crystal plane, since they could only achieve stable t rates after several days, when extensive faceting to (111) planes ha~ place. Their most likely explanation for this was that the (111) plane faster than the (100) or (110) planes, but they were unable to confirm t believe that we have measured a true reaction rate for each of the three surfaces, verifying that the (111) surface does, in fact, react faster tt (100) or (110) surface. The authors wish to express their gratitude to the management and d of Clydach Refinery, Inco Europe Limited and Inco Metals Comp~ permission to publish this paper.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
A.Ya. Kipnis, N.F. Mikhailova and R.A. Shvartsman, Russ. J. Phys. Chem. 46 (1971 G.S. Krinchik, R.A. Shvartsman and A.Ya. Kipnis, JETP Letters 19 (1974) 231. G.S. Krinchik and R.A. Shvartsman, Soviet Phys.-JETP 40 (1975) 1153. A.Ya. Kipnis, N.F. Mikhailova and R.A. Shvartsman, Kinetika Kataliz 15 (1974) 13 G.S. Krinchik, Umschau Wiss. Tech. 78 (1978) 54. P. de Groot and K. Dransfield, Z. Anorg. Allgem. Chem. 446 (1978) 39. R.S. Mehta, M.S. Dresselhaus, G. Dresselhaus and H.J. Zeiger, Surface Sci. 78 (197 G. Greiner and D. Menzel, Surface Sci, 109 (1981) L510. P. de Groot, M. Coulon and K. Dransfield, Surface Sci. 94 (1980) 204. H . H Madden, J. Kiippers and G. Ertl, J. Chem. Phys. 58 (1973) 3401. M. Perdereau and J. Oudar, Surface Sci. 20 (1970) 80. G. Greiner and D. Menzel, to be published. H. Mazurek, R.S. Mehta, M.S. Dresselhaus, G. Dresselhaus and H.J. Zeiger, Surface (1982) 530. [14] G. Lossmann, PhD Thesis, Berlin (1965); O. Knacke and G. Lossmann, Z. Physik. Chem. 53 (1967) 272. [ 15] H. Trivin, PhD Thesis, Grenoble ( 1973); H. Trivin and L. Bonnetain, Compt. Rend. (Paris) C270 (1970) 13.
S U R F A C E S C I E N C E LETTERS T H E RATE OF F O R M A T I O N OF N I C K E L CARBONYL F R O M CARBC M O N O X I D E AND NICKEL S I N G L E CRYSTALS K. LASCELLES and L.V. R E N N Y INCO Europe Ltd., Clvdach, Swansea, SA6 5QR, UK Received 10 November 1982
We have measured the rate of Ni(CO) 4 formation at a total pressure of 1 atm on Ni(100), ( 1 and (111) single crystals cleaned by argon ion bombardment and UHV annealing. The (1 surface reacts significantly faster than the other two surfaces. The partial pressure dependence the reaction and the effect of impurities are also discussed.
The reaction between nickel and carbon monoxide is very importa forming the basis for a commercial process for the refining of nickel, t example, I N C O Ltd. operates refineries at Clydach, Wales, UK, and at Copl Cliff, Sudbury, Ontario, producing many thousands of tonnes of high put nickel annually by processes based upon this reaction. There have been ma studies of the reaction Ni + 4CO --* N i ( C O ) 4 , but only recently has nickel b~ used in the form of single crystals. Some of this work was stimulated by clai that a magnetic field affects the rate of reaction [1-5], although subsequ work by several groups of workers have failed to reproduce any such eff [6-8]. Observations of faceting on the surface of low index nickel sin crystals during the reaction are also relevant to interpretations of the react mechanism on the atomic scale [9]. In this paper we report for the first time measurements of the rate carbonylation of three low index nickel single crystals cleaned by argon i b o m b a r d m e n t and U H V annealing. Unlike previously reported work, crystals gave significantly different reaction rates in a short time. Nickel single crystal discs, 25 m m diameter and 0.7 m m thick, of (1¢ (110) and (111) orientation, from Metals Research Limited were used. crystal was mechanically polished to 0.25 /zm diamond and then elect polished before mounting in a small glass reaction vessel which formed parl an U H V system. It was cleaned on both sides by many cycles of argon : b o m b a r d m e n t (10 3 Torr argon, 600 eV, 3 /~A cm -2, 10 min) and anneal (10 -9 Torr, 700°C, 20 min). The crystal was heated with an external RF c 0039-6028/83/0000-0000/$03.00 © 1983 North-Holland
16~
K. Laseelh,s. L. V. Renny / Rate of formation of nickel earho,(vl
Crystals p r e p a r e d in this way a p p e a r e d perfectly smooth in a scanning e microscope at magnifications of up to 30,000. It could not be d e m o n that they were atomically clean, but it is k n o w n that this cleaning tec gives atomically clean surfaces [10], while deliberately a d d i n g impuriti c o m m o n l y c o n t a m i n a t e " c l e a n " nickel surfaces (e.g. carbon, sulphur or ¢ gave quite different c a r b o n y l a t i o n characteristics. A t h e r m o c o u p l e was spot welded to the crystal s u p p o r t i m m e d i a t e l y ent to the crystal. The true crystal temperature, d e t e r m i n e d in a sc calibration experiment with a d d i t i o n a l thermocouples spot welded crystal, was only a few degrees higher. The true crystal t e m p e r a t u r e is t quoted. C a r b o n m o n o x i d e was purified by passing over heated copper, s u p p o r t e d p l a t i n u m catalyst at 200°C, over p a l l a d i u m and finally thr, cold trap at - 165°C. The c a r b o n m o n o x i d e contained < 1 p p m oxyg~ no i m p u r i t y measurable by a q u a d r u p o l e mass s p e c t r o m e t e r a t t a c h e d reaction vessel was detected. A r g o n was purified to < 1 p p m total im t using a BOC rare gas purifier. The nickel crystals were c a r b o n y l a t e d in a flow system. The concen of nickel c a r b o n y l was measured by UV s p e c t r o p h o t o m e t r y , using a silica cell glass-blown to the U H V system. 0.2 p p m of nickel carbonvl c( measured at a wavelength of 225 nm. Reaction rates were flow d e p e m low carbon m o n o x i d e flows but were i n d e p e n d e n t of the c a r b o n mo flow rate at 400 ml min l, c o r r e s p o n d i n g to a gas velocity of 0.34 ( across the crystal. All measurements were m a d e using this flow rate. A f t e r only a few minutes of reaction the crystals gave c a r b o n y l a t i o that were steady with time for up to 6 h at constant t e m p e r a t u r e and pr Fig. 1, curve (a), shows the reaction rate on a function of time for i", This type of b e h a v i o u r is different from any previously r e p o r t e d : t ferences are p r o b a b l y due to different methods of crystal p r e p a r a t i o n . discussed later. Table 1 shows the specific reaction rates at the t e m p e r a t u r e of the ma c a r b o n y l a t i o n rate, T,.... , which was 130_+ 2.5°C for all crystals. The i m a g n i t u d e s of the reaction rates for the single crystals do not bear a relationship to the atomic packing densities of the surfaces. By me electron m i c r o s c o p y we established that no faceting occurred on an}' crystals, although the total a m o u n t of material removed was always sm~ a p p a r e n t activation e n e r g y for the reaction was 8.5 +_ 0.5 kcal mole l three crystals in the t e m p e r a t u r e range 5 0 - 1 0 0 ° C . Diluting the c a r b o n m o n o x i d e with argon, at a total pressure a t m o s p h e r e , gave the pressure d e p e n d e n c e of the reaction rate. The r~ rate was p r o p o r t i o n a l to P~go; a d e p e n d e d on temperature~ and was I 65°C, 1.9 at 130°C and 2.9 at 160°C for all three crystal surfaces. It is well k n o w n that small a m o u n t s of impurities can affect the carl:
~e .~ c m
-2
m l n "¸
4xlO -9 (b)
2 x l O -9
(a)
2o
4O
60 Time,
8O
lC~
min
Fig. 1. Effect of oxygen on the c a r b o n y l a t i o n of Ni(110) at 130°C. (a) < 1 p p m oxygen in carl monoxide. (b) 3 p p m in c a r b o n monoxide.
tion reaction. W h e n using carbon monoxide containing 3 p p m of oxygen fact, before the platinum catalyst was added to the carbon monoxide purifi tion system), the results were as shown in fig. 1, curve (b). The reaction J was clearly greater than with the oxygen-free carbon monoxide but the valu, Tin, ~ was unchanged. Sulphur was pre-adsorbed by exposing the crystal to hydrogen sulphid 65°C; the exposure was in excess of that required to saturate the surface [ The initial reaction rate increased by a factor of twenty five, but the decayed with time. Also T,,,ax decreased to 80°C and the activation em decreased to 5.4 kcal mole-~ (30-60°C). Pressure dependence measurem on the sulphided crystal gave values of ~x = 1.06 at 59°C and 2.1 at 80°C. Table 2 shows a summary of data obtained in different laboratories on
Table 1 R a t e s of nickel c a r b o n y l f o r m a t i o n at 130°C Crystal
Specific reaction rate (mole cm -2 r a i n - 1)
N i ( l 11) Ni(100) N i ( I 10)
4.4×10 9 1.4× 10 9 1.7 X 1 0 - 9
_ ~
X X
,~
× X X
~
~°
,~"
,
X
X X ×
~
o
t
o
r o
i
.-u
E
o
c-. r,.-~
o r-,i
z~ Z
,--.
oo
o,
~" ~.,
:.E I.-.
carbonylation of nickel single crystals. All measurements were made at atm, spheric pressure apart from those of Greiner and Menzel [12] who worked 316 mbar pressure. Kirnchik et al. [3] used a discontinuous chemical method determine nickel carbonyl, Mehta et al. [7] used a photoionisation detector b all other workers have used UV spectrophotometry. Each group of workers h used a different method of surface preparation. It is unlikely that reducti~ removes all impurities from the nickel surface, and probably leaves the out few atomic layers damaged, with reactive sites [13]. This would account for high initial carbonylation rate decreasing to a more or less steady rate observ by most workers. Our method of argon ion b o m b a r d m e n t and annealing is standard method of preparing a nickel surface for L E E D examination, with damaged outer layers [10]. We observe what one might expect for the carbo~ lation of perfect crystal surface, a rapid increase to a constant carbonylati rate. C o m p a r i n g data in table 2, there seems to be general agreement that T,..... about 130°C; Greiner and Menzel's value of 120°C, obtained at a worki pressure of 316 mbar would be about 10°C higher at atmospheric pressure. lower value of Tm~~ could well indicate sulphur contamination of the surfa There is considerable scatter in the apparent activation energy determir by different groups of workers. It is impossible to assign a "correct" value, [ the majority of results on all forms of nickel [9] lie within the range 5 - 1 2 k mole ~. One must also note the problem of diffusional resistances in kinetic measurements. In this present study these were overcome by workin~ a sufficiently high gas velocity across the crystal surface. It is impossible to whether this was true for the other studies cited. Our measurements of the partial pressure dependence of the reaction r do not agree with those of De G r o o t et al. [9]. We agree with the results L o s s m a n n [14] on porous nickel sheet and of Trivin [15] on nickel pow~ L o s s m a n n obtained a value of a = 1 at < 70°C, and calculated that c~ incre~ with temperature and is equal to 4 at equilibrium. Trivin obtained a valu( ~x = 0.85 for P c o = 0,1-1 atm at 4 0 - 9 0 ° C . Trivin proposed a reaction sch~ where Ni(CO) 2 is the largest surface species formed, and the rate c o n t r o l process is either the k a n g m u i r - H i n s h e l w o o d type interaction between adja~ N i - C O species to form Ni(CO)2 or a combination of this step and desorption of Ni(CO)2 from the surface. This is the only scheme consis~ with his measured value of a, with which we agree. In the presence of oxygen, as shown in fig. 1, the reaction rate goes thro a maximum, and then decays to a steady value. This shows that tJ impurities in the gas phase can significantly affect the reaction. Presum~ some of the oxygen is chemisorbed and incorporated into the nickel lattice analogy with sulphur, which is of course in the same group as oxygen in Periodic Table, the presence of oxygen activates nickel atoms for carbon tion.
t~ /~
ix,
~,~.t-tt(:,~,
l~, v .
Ixc'f~rtk
/
lx.ut~
¢JJ J(Jlgflggll(.'rl
*.JJ , ' l l ~ . i ~ l
(uf{)(lfl}l
Of the five sets of data on the absolute rate of the carbonyl fo reaction, four are very similar after allowing for the differences in the t, ture and pressure of the measurements. The values obtained by Krinch are about two orders of magnitude greater, which together with tl~ activation energy and low value of T..... would suggest sulphur contan of the crystal surface. The values given by De Groot et al. [9] d ° not re true values for each crystal plane, since they could only achieve stable t rates after several days, when extensive faceting to (111) planes ha~ place. Their most likely explanation for this was that the (111) plane faster than the (100) or (110) planes, but they were unable to confirm t believe that we have measured a true reaction rate for each of the three surfaces, verifying that the (111) surface does, in fact, react faster tt (100) or (110) surface. The authors wish to express their gratitude to the management and d of Clydach Refinery, Inco Europe Limited and Inco Metals Comp~ permission to publish this paper.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
A.Ya. Kipnis, N.F. Mikhailova and R.A. Shvartsman, Russ. J. Phys. Chem. 46 (1971 G.S. Krinchik, R.A. Shvartsman and A.Ya. Kipnis, JETP Letters 19 (1974) 231. G.S. Krinchik and R.A. Shvartsman, Soviet Phys.-JETP 40 (1975) 1153. A.Ya. Kipnis, N.F. Mikhailova and R.A. Shvartsman, Kinetika Kataliz 15 (1974) 13 G.S. Krinchik, Umschau Wiss. Tech. 78 (1978) 54. P. de Groot and K. Dransfield, Z. Anorg. Allgem. Chem. 446 (1978) 39. R.S. Mehta, M.S. Dresselhaus, G. Dresselhaus and H.J. Zeiger, Surface Sci. 78 (197 G. Greiner and D. Menzel, Surface Sci, 109 (1981) L510. P. de Groot, M. Coulon and K. Dransfield, Surface Sci. 94 (1980) 204. H . H Madden, J. Kiippers and G. Ertl, J. Chem. Phys. 58 (1973) 3401. M. Perdereau and J. Oudar, Surface Sci. 20 (1970) 80. G. Greiner and D. Menzel, to be published. H. Mazurek, R.S. Mehta, M.S. Dresselhaus, G. Dresselhaus and H.J. Zeiger, Surface (1982) 530. [14] G. Lossmann, PhD Thesis, Berlin (1965); O. Knacke and G. Lossmann, Z. Physik. Chem. 53 (1967) 272. [ 15] H. Trivin, PhD Thesis, Grenoble ( 1973); H. Trivin and L. Bonnetain, Compt. Rend. (Paris) C270 (1970) 13.