Abstract：A portable X-ray fluorescence spectrometer was used for on-site analysis of titanium in geological samples. Due to software design deficiency, the determination results of titanium in barium-containing samples were to the high side. The causes for these results were discussed. It was believed that the characteristic lines of barium overlapping with titanium led to the higher results. Through regression trend analysis and power regression, an empirical equation was proposed: y=0.487 8x0.911 6 (x stood for the barium content determined by X-ray fluorescence spectrometry| y stood for the content of false titanium content affected by barium). The linear correlation coefficient of scatter diagram of false titanium content and barium in the sample was: R2=0.999 3. Based on this empirical equation, the content of false titanium affected by barium in the sample was worked out. The actual content of titanium was thus acquired through deducting the false titanium content by the total titanium content obtained by determination. Thirty-four mineral samples and four national certified reference materials were verified, whose results after correction were consistent with those of inductively coupled plasma atomic emission spectrometry or certified values. The linear correlation coefficient of the two methods results was R2=0.936 2. The detection limit was 0.021%. The relative standard deviation (RSD) was 1.7%. The proposed method could realize semi-quantitative or even quantitative on-site analysis of titanium in field. It can trace abnormity rapidly, providing efficient method for fast delineation of mineral body.
