No abstract

The recombination coefficient y for oxygen. atoms on silica has been determined for the temperature range 20" to 600°C. The employment of Smith's method using two Wrede-Harteck gauges together with their associated Pirani gauges in a system in which there are temperature differences between the various parts is discussed and a satisfactory procedure devised. The activity of the surface increased from y = 1.6 x 10-4 at 20" to 1.4 x 10-2 at 600°C. The graph of log y against l/T was not a straight line showing that it is not possible to explain the change with temperature in terms of a single energy of activation. The significance of the present results in relation to the changes that are known from other data to occur in the surface of silica as the temperature is raised was discussed. It was concluded that the activity and its change with temperature could be accounted for by supposing that oxygen atoms from the gas phasecombined with loosely bound oxygen atoms in the surface, the defects then being replaced at once by other oxygen atoms. In these experiments oxygen atoms were produced by a micro-wave-maintained discharge. The apparatus for doing this is described. In an appendix details of some experiments that were carried out to study the usefulness of this method are given.

No abstract

The surface recombination coefficient y of oxygen atoms has been determined for a number of surfaces (20 oxides and As, Sb and Bi). The atoms were produced by an electrodeless discharge maintained by an r.f. transmitter (8 Mc/sec). The decay in atom concentration down a side-tube into which tubes coated with the materials could be introduced was determined by using two Wrede-Harteck gauges. The activities ranged from y = 3-1 x 10-5 for Pyrex to 4300 x 10-5 for CuO. In the first long period, the activities of the oxides of Cay V, Zn, Gay Ge and As were small while those of the five metals from Mn to Cu were large. This suggested that the presence of metal ions with incomplete d-shells conferred activity on the surface. However, examination of the results for the other oxides suggested that another factor affecting the activity was the " acidity " or " alkalinity " of the oxide. The former favoured low and the latter high activity. The results obtained here were compared with conclusions that have been drawn regarding the relative abilities of different surfaces to break chains by the removal of atoms and radicals from gaseous mixtures reacting chemically. It is found that the correspondence between the two types of data is surprisingly close.In a previous paper of this series (part 3) 1 the recombination coefficient y, which is defined as the fraction of the total number of collisions which lead to recombination, was determined for oxygen atoms on several oxides. It was shown moreover that the reaction was first order. In this paper values of y were obtained for Pyrex glass, 20 oxides, and arsenic, antimony and bismuth, which were probably converted to oxides by the atomic oxygen. The Wrede-Marteck gauges described in part 4 2 were used and it appears that this does provide a real improvement in the determination of y for the inactive surfaces where it is important to reduce the rate of removal of atoms by the detector. In addition the modified theoretical treatment presented in part 4 was applied to the determination of the order of the removal process on Pyrex. This eliminated an unsatisfactory feature which had existed when Smith's formula was used.3 EXPERIMENTAL METHODThe general lay-out of the apparatus was the same as in parts 1,2 4 and 3.1 The oxygen was produced by electrolysis and its purification was the same as described in part 1. It was saturated with water vapour at 26-5"C before passing through the needle valve into the discharge tube. The discharge was maintained with an RCA high-frequency communications transmitter type ET-4336-B as described in part 3 but operating at a frequency of 8 Mc/sec. Its power rating was 350 W but at maximum power the discharge spread up the side-arms. This difficulty has been encountered by other workers.5 In practice, the maximum power used was about 150 W. The pumping system was a conventional one, consisting of a high-speed single-stage mercury-diffusion pump in series with a standard two-stage diffusion pump backed by an Edwards Speedivac IS/150 rotary oil pump. The...

Linnett and Marsden's apparatus, with some modifications, was used for determining the accommodation coefficients for the removal of oxygen atoms at surfaces. The results at room temperature for 28 surfaces (elements, oxides and halides) are presented and discussed. The conclusions may be summarized as follows. (i) For most of the surfaces the decay of atom concentration down the side tube is, within our experimental error, exponential, showing that the wall removal-process is usually first order. In the few cases where the decay appears not to be strictly exponential the surface may not be uniform.(ii) The range of values for the recombination coefficient is from 10-4 to 2 X 10-1. (iii) The activity cannot be related to any single parameter associated with the surface material though some of the comparisons can be understood on the basis of a particular mechanism.Linnett and Marsden192 measured the efficiencies of the recombination of oxygen atoms on various surfaces using the method Smith 3 had employed for hydrogen atoms. Pyrex (Hysil) glass, potassium chloride, lithium chloride, molybdenum trioxide and lead monoxide were examined; the effect of varying the temperature was also studied. Here the same general method was used. Metal and non-metal, oxide and salt surfaces were examined at room temperature and their efficiencies calculated. The efficiencies y were expressed as a fraction of the collisions of oxygen atoms at the surface, assumed smooth, which led to their recombination. The object of examining many materials was to see whether the relative efficiencies would indicate the mechanism of removal.The apparatus was similar to Linnett and Marsden's; 1 certain important differences are described in the next section. EXPERIMENTAL GAS SUPPLYDry electrolytic oxygen was used, the gas being passed through a trap cooled in liquid air and then over phosphorus pentoxide. The oxygen flow was controlled by a capillary tube into which a tungsten wire, almost filling the tube, could be moved by a screw. At this valve, the pressure dropped from about 760 to 0.05 mm Hg. Pressures were measured on a McLeod gauge. ELECTRIC DISCHARGEAn electrodeless discharge was used to produce atoms.1 The discharge tube of Pyrex, 4.2 cm diam. and 70 cm long, had two side-arms attached 55 cm from the end at which gas entered. The discharge coil 8 an diam. consisted of six turns of 6 mm &am. copper tubing. The r.f. generator was the oscillator and output stages of a converted RCA high frequency communications transmitter, type ET-4336-B. The frequency was controlled at 14mc/sec, and the output at 50-100 W. The discharge was coupled to the output coil by a single coupling loop and parallel feeder. The output circuit was series tuned by a condenser. To reduce stray radiation, the discharge coil, tuning condenser and parallel feeder were enclosed in an earthed screen of perforated zinc sheet. SIDE ARMS AND SURFACESOne side arm was 14.5 mm diam. and 55 cm long, and the other, 12 mm diam. and 35 cm long. Both were fitted with water jackets main...