Analyst, July, 1968, Vol. 93, $9. 441-442 441 The Rapid Determination of Tungsten in Ores by X-ray Fluorescence Analysis BY K. G. CARR-BRION AND K. W. PAYNE ( Warren Spring Laboratory, Stevenage, Herts.) An X-ray fluorescence powder method involving the use of tungsten Ka radiation enables tungsten to be rapidly determined in ores of different mineralogical composition. Good agreement with chemical assays is obtained. WHEN tungsten is determined in powdered ores by X-ray fluorescence analysis, marked heterogeneity effects1 can occur with the particle sizes produced by normal grinding methods. These effects are caused by tungsten being found as two minerals with markedly different X-ray absorption coefficients in the tungsten L wavelength region. The minerals are scheelite (CaWO,) and wolframite (Fe,Mn)WO,.The difference in tungsten La intensity per unit concentration from scheelite to wolframite containing samples was found to be about 20 per cent. in the middle of the concentration range examined. If the ratio of the two minerals is known and remains effectively constant, analyses can still be carried out with tungsten L radiation by using mineral powder standards. However, if the ratio of the two minerals varies, considerable errors can result. These effects can be overcome by fusion or minimised by ultra-fine grinding. Fusion is time consuming and may prove difficult with some ores, for instance, those containing large amounts of arsenic. The fineness of grinding required to eliminate these effects is beyond the range of rapid conventional grinding equip- ment : the samples examined had already been ground to less than 300 mesh in a “swing mill.” A third method of overcoming the effect is to use tungsten Ka radiation for the determination.The much lower X-ray absorption coefficients encountered enable the determination to be made directly on the powdered ores without measurable heterogeneity effects. The use of hafnium as an internal standard guards against any matrix effects with both the tungsten L and K radiations. EXPERIMENTAL INSTRUMENTAL CONDITIONS- A Phillips P.W. 1212 X-ray spectrometer, equipped with a gold tube operating at 100 kV, 20 mA and a lithium fluoride 220 analysing crystal was used: 28 angles used were tungsten Kq,, 8-39’, hafnium‘ Kcc,,, 8.92’ and background 8.09’: counting times for each position, 30 seconds (in two increments of 15 seconds).SAMPLE PREPARATION- Samples were ground to less than 300 mesh in a “swing mill.” Those expected to contain more than 5 per cent. of tungsten were diluted 1 + 7 w/w with a potassium sulphate buffer powder containing 5 per cent. of hafnium dioxide; those expected to contain less than 5 per cent. of tungsten were diluted 1 + 1 with the buffer. Two-gram buffered samples were examined directly by hand tamping in the sample holders. RESULTS Some typical results obtained are shown in Table I. TABLE I EXAMINATION OF ANALYSED SAMPLES Tungsten found by X-ray method, Sample per cent. 6516 53.1 5543 43.5 5544 16.3 6371 2-07 8344 0.84 4618 0.52 0 SAC; Controller, H.M. Stationery Office. Accepted tungsten concentration, per cent.52.5 42.8 15.8 1.92 0.73 0.5 1442 CARR-BRION AND PAYNE REPRODUCIBILITY AND ACCURACY- The reproducibility obtained on successive sampling was better than 1 per cent. relative. The error relative to the composition determined by standard chemical methods was 1.8 per cent. for ten samples in the concentration range 16 to 60 per cent. of tungsten, and 6 per cent. for six samples in the concentration range 0.5 to 2 per cent. The exact mineralogical com- position of these samples was unknown, but wolframite and scheelite standards gave equal intensities per unit concentration. Few samples were examined in the intermediate range. CONCLUSIONS The proposed method enables tungsten in ores of widely varying mineralogical com- position to be determined rapidly. The concentration range covered is between 0.5 and 60 per cent. of tungsten. The sensitivity is limited by the relatively poor signal-to-background ratio found in the tungsten K region of the spectrum. The use of tungsten L radiation would give a much lower limit of determination, but would require the use of a fusion technique to eliminate heterogeneity effects. REFERENCE 1. Claisse, F., and Samson, C., Adv. X-ray Analysis, 1962, 5, 335. Received December 20tk, 1967