1614
XOTES
-0.7081 x 10-6 c.g.s. unit, in order to determine the susceptihility of silicon compounds. The results are shown in Table I.
Table I : Diamagnetic Susceptibility of Silicon Compounds, - l o g c.g.s. units
XM >
Compounds
Dimethyldiethoxysilane Di-n-propyldiethoxysilane Trimethy lsilanol Dimethylsilanediol Diethylsilanediol a
Ref. 6.
obsd.
XM, X M , theor. prev. obtd. from recorded wavevalues‘ mech. Val.
XSiz
x8i
104.7 104.6* 104.4
17.0
149.8 66.4 58.3 81.1
16.4 18.6 16.9 16.1
150,8 66.3 58.4 81 0
mean value
Acknowledgment. The author wishes to express his sincere thanks to Dr. S. Dayal for the gift of organosilicon compounds, to Prof. A. Jlooltherji for allowing laboratory facilities and giving valuable suggestions, and to Prof. R. C. Mehrotra for encouragement. (6) A. Pacault, J. Hoarau, and A. Marchand, Aduan. Chem. Phys., 3 , 171 (1961). (7) W.Haberditzl, 2 . Chem., 1 , 225 (1961). (8) W. R. Angus, G . I. W. Llewelyn, and G. Stott, Trans. Faraday Soc., 5 5 , 887 (1959). (9) J. H. Van Vleck, “The Theory of Electric and Magnetic Susceptibilities,” Oxford University Press, 1932; J . Chem. Phys., 3 , 807 (1935).
17.0
* Ref. 3.
Oxidation of Carbon Monoxide on Thin Films of Nickel, Palladium, and an Alloy
by Earl G. Alexander and W. Walker Russell
Results and Discussion I n Table I we find that the mean values for xsl is 17.0 while xsi for individual compounds varies from 16.1 to 18.6 units. xsi has been calculated from the molecular susceptibilities of the compounds by subtracting from these values the atomic susceptibilities of hydrogen ( 2 , O ) 6 , 7 and oxygen (5.3)s and the group susceptibilities x c H a (13.45)8 and XCH$ (11.68).5 The value of xsi in triiiiethylsilanol is higher as compared to the values of xsl in other compounds. Trimethylsilanol contains only one Si-0 bond, while the remaining four compounds contain two Si-0 bonds. I n Si-0 bonds, the lone pairs of electrons on oxygen interact with vacant d orbitals of silicon. Such interactions bring about a decrease in Pauli’s diamagnetic term, thus lowering the value of xsi in the compounds having such bonds. We know that
I n the above Van Vleck quantum mechanical equationJ9 the symbols have their usual meaning. The first term in the expression is the usual Pauli diamagnetism and the second is Van Vleck’s high frequency paramagnetism. I n tetraalkyl silanes pT-dT bonding to the silicon atom is not specious so the value of xsi in tetraalkyl-substituted silanes will be conclusively higher than in the compounds in which the above mentioned type liaisons are present. There is very good agreement between the experimental and the theoretical values derived from the wave mechanical considerations. The Journal of Physical Chemistry
Department of Chemistry, Brown U n C e r s i t y , Providence, Rhode Island (Receired January 1 3 , 1964)
An effective method of studying the catalytic activity of metals has been to measure the reactivity of a gas preadsorbed on a metal film.’-3 Such a method should also be valuable in studying the catalytic activity of alloys. Because hydride formation may be a complication in studying hydrogeiiatioiis on palladium, the oxidation of CO to COZ was employed here. This reaction has already been studied by Stephens3 on palladium films. As preliminary experiments indicated that simultaneous or successive evaporation of two different metal filaments appeared incapable of producing an alloy film uniform throughout in chemical compositio~i,~ evaporation from a honiogeneous alloy pellet was used. Since nickel and palladium form a complete series of unordered-solid solutions, it was not unexpected that the behavior of their alloy was found to be intermediate to that of the pure component metals.
Experimental Unless otherwise stated, all temperatures are in degrees centigrade, all pressures are in mm., and all gas volumes are in ml. (STP). Materials and Apparatus. Wire form Ki (99.997,) (1) 0. Beeck, Discussions Faraday Soc., 8 , 118 (1950). (2) G. I. Jenkins and E . Rideal, J . Chem. SOC.,2490, 2496 (1955). (3) S. J. Stephens, J . Phys. Chem., 6 3 , 188 (1959). (4)M. K . Gharpurey and P. H. Emmett, J . Phys. Chem., 6 5 , 1182 (1961).
NOTES
and Pd (99.9%) were used for film preparation. Research grade (I2,CO, and helium, in sealed glass bulbs, were used as received. The all-glass apparatus comlprised (1) a two-stage mercury diffusion pump and cold traps; ( 2 ) a dosage system containing mercury manometers and traps; and (3) an adsorption system whiclh included the reaction vessel, two cold traps, and a Pfund gage6 with which pressures could be read down to 1 x 10-7. The adsorption system, which contained no stopcocks, could be isolated from the other parts of the apparatus by means of mercury manometers. Two types of reaction vesael were employed. Type I was constructed from a 200-ml. spherical, long-necked flask. Long molybdenum leads, introduced through a press seal a t the end of the flask neck, extended to the center of the flask bulb. A loop of coiled palladium wire (0.2 mm. diameter) or nickel wire (0.3 mm. diameter) was spot-welded between two molybdenum lead ends. The Type I1 vessel was an inverted 500ml., wide-necked, spherical flask. The spherical portion was joined to the neck through a, 45-mm. diameter shoulder around which a four-turn induction coil could be placed. Inside the shoulder portion a metal or alloy pellet was supported in a shallow quartz cup atop a vertical axial glass spindle. The induction coil was energized by a 4-kw. spark gap converter.
Preparation of Catalyst Films i. Filament Evaporation. Reaction vessel I and the two adjacent traps were evacuated and baked a t 360' for 10-14 hr. Other parts of the system were flamed. The filament was flashed several times during bakeout. Then the trap next to the reaction vessel was cooled with Dry Ice-2-propanol and the adjacent trap with liquid nitrogen. After an ice bath was positioned around the reaction vessel the film was evaporated with a filament current of 2.9 amp, for nickel and 1.6 amp. for palladium. The pressure during film evaporation was about 1 X 10-6 to 4 X 10-7. The evaporated film was stabilized by three times alternating between room temperature andl that of liquid nitrogen. The weight of a film was determined by weighing the filament before and after evaporation. Even though thc filament loop was placed close to the center of the spherical reaction bulb, the metal film was not uniform in thickness and showed a very thin streak in the plane of the filament loop. As simultaneous or successive evaporations from two different metal filaments could not be expected to produce a uniform alloy film under such conditions, evaporation from a preformed homogeneous alloy pellet was tried. 2 . Induction Evaporation. The preparation of suitable metal or alloy pellets is described elsewhere.6
1615
After sealing to the system a Type I1 reaction vessel containing a prepared pellet, the evacuation, bakeout, flaming and cooling of the traps was as in part 1. To remove dissolved gases, once during bakeout the pellet was induction heated almost to melting. The reaction vessel was packed in Dry Ice during film evaporation. The pellet was further outgassed three short periods of induction heating at low power, then a t maximum power the film was evaporated during about 5 niin. The film was evaporated in one continuous heating operation, since the presence of the metal film on the reaction vessel interior walls appeared to decrease the ability of the metal pellet to couple with the field of the induction coil. The pressure during film evaporation was about 1 X lo-+. The nickel films were stabilized as in part 1, but the Pd-Ni alloy film between -78' and 100'. Film weight was determined by weighing the pellet before and after evaporation. The films prepared by induction heating appeared brilliant and mirror-like as were the filament evaporated films. Although the thickness of the induction evaporated films appeared uniform with respect to cylindrical symmetry, film thickness decreased with distance from the pellet,. That the composition of the alloy film was little affected by film thickness, however, was demonstrated by polarographically analyzing, for both nickel and palladium, five different portions of the alloy film, each portion representing a different distance from the pellet. The film portion on the edge of the pellet cup was slightly richer in palladium, L e . , 7,4 atom %, but was so highly sintered that it probably contributed little to catalytic activity. The other film areas showed 6.0, 6.2, 6.1, and 6.2, respectively, yielding an average of 6.1 atom % palladium. Thus, it appeared that an alloy film having good uniformity in chemical composition throughout its extent could be prepared by the pellet evaporation technique. Procedure for Adsorption and Reacticn. With the reaction vessel cooled in ice, or Dry Ice-2-propanol, a known amount of oxygen was adsorbed on the metal film. Then after the gas phase was evacuated, the reaction vessel was therniostated at 0 or looo, and doses of CO were admitted. For one nickel film CO was adsorbed first and oxygen added subsequently. I n order to determine the amount of COz desorbed from a film,3 pressure measurements were made both with the second trap cooled in liquid nitrogen and a t room temperature. At the conclusion of the reactions on a film, helium expansions were made to determine necessary apparent volumes. (5) A. H. Pfund, Phys. Rev., 18, 78 (1921). (6) E. G. Alexander and W. W. Russell, to be published.
Volume 68, Number 6 June, 196.4
NOTES
1616
Table I : Summary of Adsorptions and Reactions on Films Reaction of CO with Adsorbed Oxygen ,--Temperature--Film
Wt., mg.
0% ads
(Pd)lc (Pd)2' (Pd)3' (Ni)4' (Ni)5' (Ni)7d
5 55 61 49 28 12
-78"
( Pd-Ni)lOd
21
-78"
COtotal adsb
Reaction
O0
0" 0" 0"
0"
...
0"
O0
0"
0"
0" 100"
looo
ads.
02"
COG ads.
1.49 0.31 0.16 7.08 9.34 4.10 ,
.
C O P desorb.
2.83 0.56 0.28
I
8.05
% 0 2 reacted
2.24 0.59 0.21
Coz desorb
75 95 65
2.3 2.0 2.3
...
...
...
...
0.39 0.47 0.33 1.12
0.07 0.02 0.68 0.82
0.4 0.2 8.3 5e
6.6 25 1.5 2.4
Reaction of On with Adsorbed CO Film
W t . , mg.
(Ni)Gd
30
~--Temperature--co&ds Reaction
0"
0"
M1. per g. of film. COt,tsl ads is sum of Co& plus COSdesorb. would have been significantly larger had run been continued longer.
Results and Discussion Typical results are summarized in Table I in which it is to be noted that the sixth column refers to the amount of CO or oxygen remaining adsorbed at the end of reaction on a film, while the eighth column lists the 7 0 of oxygen or CO reacted in forming the COZ desorbed. About "4 of the oxygen preadsorbed on a film was taken up rapidly to a low residual pressure (
