Multi-batch Preparation of Standard Samples from a Single Doped Solution for Cross-checking in Surface Metal Analyses of Silicon Wafers

Abstract

Metallic impurities on semiconductor surfaces degrade the electrical properties of large-scale integrated circuits (LSIs). To determine such impurities, total reflection X-ray fluorescence spectrometry (TXRF) is widely used. 1 TXRF requires reference standard samples for quantification. We have proposed a method for preparing standard samples for this purpose using immersion in an alkaline hydrogen peroxide solution (IAP). 2,3 In the IAP method, base silicon wafers are immersed in an alkaline hydrogen peroxide solution in which known amounts of metal ions are doped. The metal ions adsorb on the surface silicon dioxide layer. Since the metal adsorption is controlled by a natural equilibrium, 4 the IAP wafers that are prepared in a single solution have high wafer-to-wafer uniformity.2 Because of this characteristic, the IAP wafers can also be used for crosschecking. 5,6 However, the concentrations of the surface metal are not precisely the same between different production batches, even if the concentration of added metal is the same. This situation limits the maximum number of available identical wafers to 25, which is the typical capacity of a wafer cassette. While the reason for such inconsistency is not clear, more than 25 identical samples are often required for large-scale crosscheck experiments. The aim of this paper is to propose an improved method of production that provides more than 25 identical IAP wafers. ExperimentalThe silicon wafers used in the present work were CZ n-type single crystals with polished (1 0 0) faces. The wafers were cleaned with a 0.5% hydrofluoric acid solution (r.t., 2 min), followed by a mixture of hydrochloric acid and hydrogen peroxide (standard cleaning solution-2: SC-2 7 ) before the IAP process.The baths for the IAP and rinse were made of high-purity quartz glass. That for the IAP was equipped with a circulation system. All apparatus that came into contact with either the wafers and/or the solutions were made of high-purity quartz glass or fluorocarbon polymer.All chemicals, except for the metal solutions, were of semiconductor industry grade. The metal solutions were diluted standard solutions for chemical analysis.The solution used for IAP was a 1:1:5 (volume ratio) mixture of 29% NH3, 31% H2O2, and H2O (standard cleaning solution-1: SC-1 7 ). A certain amount of Fe and Ni solution was added to the solution. After all the chemicals were poured into the bath, the solution was circulated for 5 min to become homogeneous. The pre-cleaned wafers were then immersed for 120 s at 60˚C, followed by a quick dump rinse bath (QDR, 180 s), and then an overflow bath (OFR, 270 s). The wafers were finally dried with a spindrier (SPD, 600 rpm, 270 s). The timetable is shown in Fig. 1.The NH3 and H2O2 concentrations were monitored with a continuous-flow spectrophotometric analyzer (CS-220, COS Ltd. ) every 5 min.The adsorbed metal concentration was measured with TXRF (TXRF300, Rigaku corp.). The excitation X-ray was Au-Lβ

Trends in total reflection X-ray fluorescence spectrometry for metallic contamination control in semiconductor nanotechnology

“…In the immersion method a variety of procedures is in practice with a mechanism depending on the nature of the chemical solution: chemisorption of cations at neutral and slightly acidic pH [77,78], chemical deposition and chemisorption of hydroxide species at alkaline pH, deposition and incorporation in the SiO 2 using NH 3 /H 2 O 2 solutions (also known as the immersion in alkaline hydroxide peroxide solution (IAP) method [69,79]), electrochemical outplating in HF solutions [80], etc. In the spin-coating method, a diluted metal standard solution is dispensed on the wafer, positioned on a wafer spinner.…”

“…This process is known to cause quantification errors up to a factor of 3 after a few months aging. The IAP prepared wafers are shown to be less sensitive to variations over time as the contamination is buried into the SiO 2 layer, making them more suitable as calibration standard [69,79].…”

Whole-Surface Analysis of Semiconductor Wafers by Accumulating Short-Time Mapping Data of Total-Reflection X-ray Fluorescence Spectrometry

“…The accuracy of TXRF measurement strongly depends on the accuracy of the glancing angle . Figure is a typical angle scan profile of a wafer intentionally contaminated by immersion in alkaline hydrogen peroxide solution (IAP) method. − In this case, only a 0.01° shift in the glancing angle around 0.08° causes a ±14% error in fluorescence, so a high-precision angle adjustment is required for accurate analysis. Because it is difficult to make a perfect angle adjustment mechanically, a feedback control to make the Si Kα intensity from the substrate consistent is commonly employed in commercially available TXRF instruments.…”

Laser ablation ICP-MS to determine Cu on a Si wafer prepared by ion sputtering

“…Several reference wafers have been used in the semiconductor industry, prepared by such methods as dipping, spin-coating, and particle deposition. 8,9 In the dipping method, a wafer was soaked in a bath containing known concentrations of contaminants for a limited time and then dried. In spin-coating, a wafer was coated with standard solutions by a spin coater.…”