THE JOURNAL OF
PHYSICAL CHEMISTRY (Registered in U. 8. Patent Office)
VOLUME 62
(0Copyright,
1959, by the American Chemical Society)
JANUARY 5, 1959
NUMBER12
CHANGES OF ELECTRIC RESISTANCE OF NICKEL FILMS WITH ADSORPTION OF HYDROGEN BY ZENJIRO ODAAND HIKARUARATA Electrical Communication Laboratory, Nippon Telegraph and Telephone Public Corporation, Musashino-shi, Tokyo, Japan Received December 13, 1067
The changes of electric resistance of annealed nickel films due to the adsorption of hydrogen have been investigated by changing the condition of preparation. The resistance increases on very clean films, while it decreases on contaminated films. Experiments show that this decrease is due not to the oxygen adsorbed on the surface, but to that which has diffused into nickel lattice. There are two sorts of reversal in resistance change during adsorption, one occurs on a very clean film a t - 183’ and the other on a film containing a small amount of oxygen a t 0’. The former is considered to be an inherent property of nickel film, while the latter is a spurious one due to oxygen. The degree of vacuum which can guarantee the cleanliness of the film is discussed.
Introduction The changes of electric resistance of evaporated metal films due to gas adsorption have been measured by several authors to elucidate the electronic interaction in adsorption1+ and the mechanism of catalytic on the film. Especially the effect of hydrogen on nickel film was investigated persistently by Suhrmann and his coll a b o r a t o r ~ . ~ -They ~ have found that there are two trends in the change of resistance depending on the “Ordnungszustand” of the film; the resistance decreases on the well-annealed “geordnete” films, while it increases on the “ungeordnete” films evaporated a t 90°K. and not annealed. However, more recently the same authorslo have reported that even on a “geordnete” film the resistance increases a t very low pressure of hydrogen (2 X (1) Z. Oda, Bull. Chem. SOC.Japan, 87, 465 (1954). (2) Y. Mizuahima, 2. Oda and 0. Ochi, J . Phys. SOC.Japan, 12, 355 (1957). (3) R. Suhrmann and K. Schulz, Z . physik. Chem., 1, 69 (1954); J . Colloid Sci. Suppl., 1, 50 (1954). (4) R. Suhrmann, Advances in Catalysis, 7 , 303 (1955). (5) R. Suhrmann and K. Schulz, Nalurwiss., 48, 340 (1955): R. Suhrmann, Z. Eleklrochem., 60, 804 (1956). (6) W. M. H. Sachtler, J . Chem. Phys., 86, 751 (1956); Z. Elsktrochsm., 60, 838 (1956). (7) G. Reinacker and N. Hansen, Z . anorg. allgem. Chem., 884, 162 (1956): Z. Elektrochem., 60, 887 (1956). (8) J. H. Singleton, THISJOURNAL, 60, 1606 (1956). (9) R. Suhrmann and G. Wedler, Z. Elektrochem., 60, 892 (1956); Advancer in Catalysie, 9, 223 (1957). (10) R. Suhrmann. G. Wedler and D. Sohliephake, 2. physik. Chem., 18, 128 (1957).
10-6 mm.) and decreases subsequently a t higher pressures. This phenomenon was also interpreted in terms of surface heterogeneity which would be caused by some disordered regions remaining on the ordered film. However, Singletons has obtained just the opposite result in a conventional high vacuum system. A t very low equilibrium pressures (2 X mm.) the adsorption of hydrogen gives rise to a decrease in resistance except a t the beginning of adsorption. This was followed by reversible adsorption over a pressure range from loR5 to 600 mm., causing an increase in resistance. On the other hand, Sachtlera has indicated that adsorption of hydrogen always increases the resistance of the film when it is prepared in a carefully evacuated high vacuum system, while on the film deposited in a poor vacuum system (10-8 mm.), it gives the opposite effect. From these results, he has concluded that the different degree of “contamination” is primarily responsible for the different behavior of the two sorts of films and has questioned the interpretation of Suhrmann. These results suggest that there may be at least two factors which affect the sign of resistance change. If this is the case, they would appear simultaneously in the experiments of previous authors referred above, making the phenomenon complex. Hence, in order to clarify the situation, an attempt should be made t o separate the effect of each factor under a suitable condition. Firstly,
1471
1472
ZENJIRO ODA AND HIKARU ARATA
Vol. 62
of vacuum is required to produce and keep the clean film without severe contamination. Experimental Experimental Techniyes._The vacuum apparatus used is shown in Fig. 1 whic is similar to that of Beckerll and could be evacuated to less than 1 X 1O-O mm. by repeating the bake-out of glass assembly and outgassing of ionization gages. The peak pressure evolved during evaporation was reduced to less than 2 X 10-8 mm. by extensive outgassing of nickel wire. Nickel film was evaporated from an electrically heated pure nickel wire,‘* 0.5 mm. in diameter and 4 cm. in length, Fig. 1.-Experimental apparatus: V, evaporation bulb; onto the inner wall of a 4 cm. diameter tube, and had a uniGLC, greaseless cock; F, orifice; TI li uid O2trap, (IG), form thickness except a t the edges. The apparent area of and the average thickness was varied (IG)a, Bayard-Alpert ionization gage. E?2 is passed into the the film was 70 system through a capillary and a trap not shown in this from 50 to 200 atomic layers. Two platinum foil electrodes were sealed in the wall of the bulb center, and the figure. resistance of the film between these electrodes was measured on a Wheatstone bridge. The d.c. current diY I&AoLECULES rection was reversed to check for errors due to contact resistance and thermoelectric effects. Two methods were employed for preparation of lo-’ the ordered films: (1) Along our previous method,” the film was evaporated onto the bulb wall a t 0’ and then annealed a t 85-90’ for 10-30 minutes. (2) FolCi lowing Suhrmanns the film was evaporated onto the t bulb wall a t liquid oxygen temperature and then anA nealed at 110”for 1 hour. Contaminated films were produced intentionally by flowing oxygen or hydrogen into the system during or - 10-4 after evaporation. Pure hydrogen was obtained by heating a Pdlo-’ thimble electrically in coal-gas and reserved for each 0 20 40 60 80 100 run. The purity of the gas thus obtained was checked Time, min. by a mass-spectrometer. Oxygen was produced by the thermal decomposition of potassium permanganate Fig. 2.-Change of electric resistance of a nickel film evappowder and was not purified further* orated under extremely good vacuum condition by method 1. The adsorption of hydrogen in a pressure range = 00; A e hydrogen adfission; E = evacuation sticking probabilit of hydrogen, ns = number of hydrogen l ~ ~ w ~ ~c ~ ~ ~ l ; ~ ~ i molecules adsorbe$. pressures were taken in a constant volume by closing a main stopcock and capillary leak, but the adsorbed amount could not be obtained in this method. 7 1 Procedure of Constant Flow Method.-When a film is not evaporated, the gas flowed into the system through a leak is taken up by the ionization gage (IG)1, pumped out through the orifice F, and the remainder is accumulated in the volume V. This is lo-* yi expressed quantitatively as follows. Let E = the number of hydrogen molecules entering the volume V per second, through thec apillary leak. The numlo-’ ber of molecules leaving the system per second by 66.00 pumping = (F/IcT)(PI - Pz), where F = “conductance” of the orifice F ( = 5/cc./sec. for HL), PI and P1 are pressures in V and outside the orifice, re65.60 10-6 spectively. Then, if we neglect the clean-up effect of the ionization gage (IG)‘ the change of pressure in the 0 20 40 60 80 100 volume V is given by Time, min. F Fig. 3.-Change of electric resistance of a nickel film evap(PI - P*) (1) E T - E -IcT orated under extremely good vacuum condition by method 2. TA = -183”. A t stationary state, dPl/dt = 0, so that P the effect of contamination should be examined E = - (PI - Pz)o (2) with “geordnete” films, and secondly that of kT “Ordnungszustand” with very clean films. The When a film1 is evaporated on the inner wall of the tube, work of Sachtler is regarded as an attempt made the number of molecules adsorbed per second by this = (Sp/lcT)Pi, where S, = pumping speed of the film. Since along the first principle. However, it is not clear the clean-up effect of ionization gage is remarkable, the whether his films were annealed enough to reach leaving molecules due to this effect must be taken into the “geordnete” state of Suhrmann or not. I n account. Let 8, be the pumping speed of (IG)I, this correcaddition, the film thickness, the pressure of hy- tion term = (Sg/kT)P1. Thus
Pump
~
e
~
~
I
.
-
drogen and the adsorption temperature were different for two sorts Of films, SO that his COnClUSiOn seems to be premature. It is the purpose of this work to investigate the effect Of Oxygen contamin’tion on “geordnete” films in order t o check the conclusion of Sachtler and to know what degree
(11) J. A. Beoker and C. D. Hartmen, THIS JOURNAL, 57, 157 (1953): Advances zn Catalysis, 7 , 135 (1955). (12) Spectroscopic analysis showed that the purity is estimated to be 99.9%. The impuritiea were Co, Mg, Mu, Al, Fe and Sn of 10-8 % order, and cu, ca, and Ag of 10-4 % order. (13) E. Oda, BUZZ.Cham. SOC.Japan, as, 281 (1955).
s.
IN ELECTRIC RESISTANCE OF NICKELFILMS WITH HYDROGEN ADSORPTION 1473 Dec., 1958 CHANGES
From eq. 2 and 3, Spand the number of admoles n, are given by
SP
=
1
'is; jP(Pl - Pt)o - F(P1 - P,) sap1
-
-
VSl
(4)
and (5) Usually the term V(dPl/dt) is small compared with other terms and can be neglected. Hence 1 sg (4)' -P 1 (F(P1- P 2 ) O - F(P1 P2))
s
-
-
'.
, I 1 10-6 40 60 100 110 Time, min. Fig. 4.--Change of electric resistance of a nickel film evaporated in hydrogen atmosphere of 10+-10-* mm. by method 1: TA = -183". 21.90
I
1
Values of 8, have been obtained by subsidiary experiments and found to be more thap 500 cc./sec. at first for the electron 'emission of 0.1 mA., but decrease rapidly to 10 cc./sec. with increase in clean-up amount. This variation of 8,is also taken into account. The pumping speed 8, is dependent on the sticking probability s of the gas. If we use the latter, the adsorption rate is given by
Y
1
I
I
I
20
0
50.60 F-
I
an, = S.v.PIA dt where v = l/dZnrnkT and A = apparent surface area of the film. This should be equal to (Sp/kT')P1,so that 0
Results and Discussion Very Clean Film at O".-If hydrogen was adsorbed on a very clean nickel film a t 0", the resistance of film increased with adsorption and decreased with desorption independent of film thickness and preparing method. Figure 2 shows an example of thick films prepared by method 1. On a fresh surface with high sticking probability, the resistance increases parallel
X%"
20
40
80
60
100
MOLECULES
-I7
59.00
-lo
-
4
E
I -6
G58.50
A
+
)?
- - - F - - - u - = - ~ 7x10
'mh
,o:l
10-2
7.5, -
10-8
carkrbced
lo-' 58.00 -
.
10-
yi
1474
ZENJIROODA AND HIKARUARATA
macroscopic geometrical area as its surface area.I4 From these facts it is anticipated that the low-temperature adsorption of hydrogen on clean nickel films comprises two stages, rapid and slow, corresponding to the magnitude of sticking probability. Furthermore, the rapid stage is separated into two types, the one gives rise to an increase in resistance and the other to an opposite effect. The second slow adsorption with a trend t o increase the resistance always takes place on the surface covered by rapid adsorption. I n fact, when a film was evaporated and annealed in hydrogen atmosphere of to lo--' mm. the initial increase of resistance disappeared and only the decrease and second increase were observed as shown in Fig. 4. It seems to be probable that the hydrogen taken up during preparation attacked the clean sites on which the initial rapid adsorption would increase the resistance. Film Evaporated in Oxygen Atmosphere.-When a film was evaporated in a flow of oxygen a t a residual pressure of about lo-' mm. order, hydrogen adsorption caused a monotonous decrease in resistance. Figure 5 shows a result on such an oxygenated film prepared by method 1. This film contained so much oxygen that the sticking probability was lower than After ad- and desorption of hydrogen were repeated 3 times, this film was again exposed to oxygen flow, yielding a marked increase in resistance. The effect of hydrogen on resistance was not altered by this treatment. With a lower rate of oxygen flow during evaporation (residual pressure of about lops mm.), the change of resistance showed a reversal in course of hydrogen adsorption. The result is shown in Fig. 6 , which suggests that a fraction of the surface remains in a clean state. Hydrogen is adsorbed preferentially on clean sites due to their high sticking probability, inducing an initial rise of resistance. Oxygen Adsorbed Film.-Figure 7 shows a result obtained when a clean film was partially covered with oxygen at O", and then exposed to hydrogen at the same temperature. In this situation the effect of oxygen contamination was not so clearly observed as in Fig. 5 and 6. However, when such a film was annealed a t 85" for 10 minutes in vucuo we could find the decrease in resistance with adsorption of hydrogen. This heat treatment may cause a diffusion of adsorbed oxygen atoms into metal lattice to produce an oxygenated film such as evaporated in oxygen flow. This fact indicates that the decrease of resistance by hydrogen adsorption is due not to the oxygen adsorbed on the surface but to the oxygen which has diffused into the nickel layer. The Effect of Contamination.-The results ob(14) For the average density of surface atoms 1.7 X 10'6 Ni atoms/ cm.3 the number of adsorbed Hn molecules in monolayer is evaluated to be -6 X 1018 molecules when A = 7 0 cm.2 and adsorption ratio Ni/Hz = 2.
Vol. 62
tained above indicate that when hydrogen is adsorbed on the very clean "geordnete" films, their electric resistance is increased irrespective of film thickness, temperature of adsorption (0 and - 183') and pressure of hydrogen (
