Spurious Peaks in Total Reflection X-Ray Fluorescence Analysis

Abstract

This paper presents details concerning such spurious peaks as Fe Ka and Ni Ka radiation, which are often observed in the trace determination of metallic impurities on silicon wafers by using a monochromatic total reflection X-ray fluorescence (TXRF) analyzer. The intensity of a spurious peak varies along with changes in the incident azimuth angle and/or detected intensity of the primary X-ray beam. The origin of this phenomenon is impurities existing along the path of Xrays. This phenomenon influences the accuracy of trace TXRF analysis, and a practical detection limit of a TXRF analyzer becomes worse by approximately one order, becoming 10" atoms/cm2, than the calculated detection limit (2 -8X1010 atoms/cm2) in our case.

A Review of Standardization Issues for Total Reflection X-Ray Fluorescence and Vapor Phase Decomposition/Total Reflection X-Ray Fluorescence

“…25 It had been believed that the preparation of appropriate reference materials for VPD/TXRF would be easier than for TXRF, because of the earlier experience of residue analysis in Europe using fixed-angle TXRF instruments and spiked residues.26 However, some recent reports question this assumption. Yakushiji et al 11 have presented calculations and data showing the effect of the finite gap between the sample and the detector on the TXRF detection of small residues as a function of the residue diameter. Kondo et a1.12 have revealed that a diluted AAS droplet of several millimeters can dry along with the segregation of Ni metal to less than 100 µm in diameter, and that this can lead to unexpected mass-absorption effects6 in the quantification of TXRF.…”

Accuracy of Total Reflection X-Ray Fluorescence Spectrometry near the Detection Limit

“…Since the deviation seemed to be parallel to the ideal 45˚ line, we considered that the deviation originated in spurious (or instrumental) peaks. 7 The spurious peak originated from the impurities in the detection system (especially in the Be window 8 ). When the scattered or diffracted W-L b struck the detector, the impurities in the detector were excited and their fluorescent X-rays were detected as spurious peaks.…”

“…7 The proportionality coefficient was obtained as follows. The center of a blank wafer was measured several times at different azimuth angles of incident W-L b X-rays.…”

“…Yakushiji et al pointed out that the existence of a spurious peak itself followed the deterioration of the detection limit of TXRF. 11 In order to minimize the spurious peaks, they proposed the use of an x-y movement sample stage instead of r-q stage. 11 It became possible to fix the incident X-ray azimuth for any measuring point by using x-y stage.…”

“…11 In order to minimize the spurious peaks, they proposed the use of an x-y movement sample stage instead of r-q stage. 11 It became possible to fix the incident X-ray azimuth for any measuring point by using x-y stage. The TXRF analysis could be performed at the azimuth angle at which the intensity of W-L b was minimized, and consequently the spurious peak was also minimized.…”