Oscillographic Method for Studying Gas Adsorption Phenomena*

An essential feature of many studies on adsorption, evaporation, surface diffusion, etc., is the observation of the process by which an equilibrium state is established after a change in the gas pressure in the system or after a change in the temperature of the solid phase.

Obtaining correct data on the time course of this process is possible, however, only if the rate of action of the energetic processes on the indicator and the rate at which the latter reproduces them are sufficiently small in comparison with the rate of the reaction itself.

This condition becomes difficult to fulfill in the study of processes proceeding in thin (monatomic) films, and especially when the solid phase is represented in the form of a filament, as is often necessary in order to obtain, by heating, a genuinely clean surface. Here neither optical, nor micromanometric, nor thermal methods are applicable.

Precisely for the study of processes occurring on the surface of metallic filaments, the method described below is intended,

* Johnson and Vick, Proc. Roy. Soc. (A), 151, 296, 1935.

developed by M. Johnson and F. Wick of the University of Birmingham.

The indicator here is the thermoelectron-emission current of a metal filament. The method consists in the fact that the emission current of the filament (tungsten), on whose surface the process under study takes place, creates a voltage drop across a noninductive resistance. The potential difference that arises is fed to one pair of deflecting plates of a cathode oscillograph; to another pair of plates a sawtooth voltage is applied. The potential difference between the sweep plates is a linear function of time. Recording is carried out by photographing the path of the electron beam on the oscillograph screen. On the same photographic plate, by simply switching a commutator, one can obtain a photograph of the motion of the beam under the action of only one sweeping field (giving the abscissa axis) and a sinusoid from a standard generator, serving as the time scale.

Since the effect of the reaction on the surface of the filament is reflected in the magnitude of the emission current immediately, the delay of the pulse at the output resistance does not exceed the time of passage of the electrons from the filament to the collector, which is negligible. A more significant lag is introduced by the oscillograph and the circuit. However, even here it is possible to achieve reproduction with a delay of no more than \(10^{-7}\) sec. Some further limitation on the speed of reproduction is imposed by the photographic part of the apparatus, since at excessively high speeds the spots the plates are sensitive to do not permit a sharp image to be obtained. In the authors’ experiments, the maximum linear speed of the spot available for recording was \(2 \cdot 10^{6}\) cm/sec.

Fig. 1

Thus the inertia of the instrument proves to be very small in comparison with the time intervals required for changing the state of the gaseous (pressure) or solid (temperature) phase that is the cause of the reaction. For the case of a change in the temperature of the filament this is illustrated by Fig. 1. Here the upper curve depicts the change in the thermionic current of a pure tungsten filament when its temperature is changed from 1065 to 2445° K (this change occurs owing to an instantaneous change in the current strength). The lower curve of the same figure depicts the change in the thermionic current of the same filament for the same change in temperature, but for the case when the surface of the filament is covered with an adsorbed layer of oxygen. The gradual rise of this curve indicates gradual evaporation of the adsorbed oxygen. The slope of this line in the section beyond the rising part of the upper curve (after the final temperature is reached) is a measure of the rate of evaporation of the adsorbed oxygen at a temperature of 2445° K.

Another example of an oscillogram is provided by Fig. 2. In contrast to the method of forming a layer of adsorbed oxygen described above, the measurement reproduced by the curve of Fig. 1, here oxygen was introduced in excess. This led, as a result of oxidation of the surface, to a fall in emission from the level shown by the ...

...of the upper horizontal straight line being displayed (taken beforehand) down to the level corresponding to the lower one. Recording of the oscillogram begins with the onset of evaporation of tungsten oxide, marked by an increase in emission, which rises to a value exceeding the initial value owing to the fact that, when the oxide evaporates, a clean tungsten surface is revealed. However, the presence of a small amount of oxygen leads to a new reaction, expressed in a decrease of emission as the more volatile oxygen layer (as compared with the adsorbed oxygen layer) is removed. Since, at the filament temperature in the experiment to which Fig. 2 refers (\(\sim 2000^\circ\) K), the process of evaporation of adsorbed oxygen (represented in Fig. 1) would take several minutes, the rapid (within 0.5 sec.) rise of the curve in Fig. 2 with an inaccessibly steep slope in the former methods clearly demonstrates the difference between the two types of action of oxygen on the surface (formation of an adsorbed layer and formation of an oxide).

Seconds

Fig. 2

The authors successfully applied the described method to determine the duration of oxygen remaining on the filament in the adsorbed state and to determine the heat of evaporation of adsorbed oxygen.*

N. Khlebnikov, Moscow

* Johnson and Vick, Proc. Ray. Soc. (A), 151, 308, 1935.