Philips J. Res. 36,195-209,1981

R 1040

THE REACTION OF BROMINE AND OXYGEN WITH A TUNGSTEN SURFACE PART Ill: PIT FORMATION ON TUNGSTEN IN BROMINE GAS by G. C. J. ROUWELER and B. J. DE MAAGT

Abstract A description is given of a remarkable phenomenon occurring during exposure of tungsten to bromine gas at relatively high temperatures. Generally, the phenomenon consists of the formation of hemispherical pits in the tungsten. The influence ..of reaction time, temperature and bromine partial pressure on pit growth was investigated. Also studied was the influence of a thermal pretreatment of the measuring probes on pit formation. From the results and supporting experiments using Scanning Auger Microprobe (SAM) techniques, it is proposed that a local impurity in the tungsten, the nature of which remains unknown, acts as a catalyst for the tungsten-bromine reaction.

1. Introduetion Recently, we published experimental results on the kinetics of the tungstenbromine 1), the tungsten-bromine-oxygen 1), and the tungsten-bromine-oxygen-hydrogen 2) reactions. During these investigations heavy local erosion of the measuring probes was also observed, especially in the tungsten-bromine system 1) at relatively high bromine partial pressures. Generally, this strong erosion consisted of the formation of hemispherical craters. Analogous to the nomenclature of similar phenomena occurring during corrosion of metals and. alloys in solutions 3-12) and gases 13,14) we classify this phenomenon as pitting. The present study deals with the conditions that lead to pit formation on tungsten exposed to bromine. The influence of the experimental parameters reaction time, reaction temperature, and bromine partial pressure on pit formation was investigated. In addition the effect of a thermal pretreatment was studied. Furthermore, an attempt was made to determine the influence of several surface impurities using Scanning Auger Microprobe (SAM) techniques. From the results we conclude. that pitting has to be ascribed to a fast tungsten-bromine reaction catalyzed by some impurity, which can be neutralized substantially by an oxidizing thermal pretreatment. Philips Journalof Research Vol. 36 No.3

1981

195

G. C. J. Rouwe/er and B. J. de Maagt 2. Measuring system The experiments were carried out in a gas-flow system similar to that used in earlier experiments 1) (see also fig. I). The argon (at atmospheric pressure) used as carrier gas for bromine was purified by means of a Messer-Griesheim adsorption cell. Typically, the argon fiux had a value of 0.5 cm3s-1 (STP). The mole fractions of oxygen and water in the gas were less than 0.2 x 10-6 and less than 0.5 x 10-6, respectively. Reducing components were present up to a maximum of 0.5 X 10-6 in H2 equivalents. Bromine was obtained by electrolysis of a melt of silver bromide. Oxygen, which was used during the thermal pretreatment of the tungsten probes, was obtained from a zirconia solidelectrolyte cell1). Ga

Fig. 1. Schematic description of the gas-flow system with argon cylinder (A), adsorption cell for gas purification (B), valves (C), bromine (D) and oxygen (E) generation cells, mixing vessel (F), reaction vessel (G) with tungsten rods (Ga), measuring probe (Gb) and Pyrex window (Ge), bromine absorption vessel (H), and soap film flowmeter (1). Samples were observed with a photomicroscope (K).

The reaction vessel contained two tungsten rods, between which the measuring probe was spot-welded. The measuring probe consisted of a tungsten ribbon with a thickness of 30 urn (manufacturer Cross, material A) or 25 urn (manufacturer Philips, materials B and C). The width of material A was 0.10 cm, that of materials B and C, 0.10 and 0.20 cm, respectively. The ribbon was coiled on a fiat spine in order to enlarge the surface accessible to microscopie observation. The dimensions of the spine were 0.245 cm by 0.05 cm, the distance between the turns of the coil being about 0.02 cm. In a few experiments Philips tungsten wire (diameter 192.5 urn) was used (material D). The impurity concentrations of the different materials were determined mainly by spectroscopie analysis (table I). 196

Philips Journalof

Research

Vol.36

No.3

1981

I

The reaction of bromine and oxygen with a tungsten surface. Part JII TABLE I Impurity concentrations in the different materials (0/0 by weight)

y

~

A Cross ribbon

B Philips ribbon

C Philips ribbon

D Philips wire

0.0015 0.021 0.0082 0.0003 0.0015