Jörg Acker

Spectrometric analysis of process etching solutions of the photovoltaic industry—Determination of HNO3, HF, and H2SiF6 using high-resolution continuum source absorption spectrometry of diatomic molecules and atoms

The surface of raw multicrystalline silicon wafers is treated with HF-HNO(3) mixtures in order to remove the saw damage and to obtain a well-like structured surface of low reflectivity, the so-called texture. The industrial production of solar cells requires a consistent level of texturization for tens of thousands of wafers. Therefore, knowing the actual composition of the etch bath is a key element in process control in order to maintain a certain etch rate through replenishment of the consumed acids. The present paper describes a novel approach to quantify nitric acid (HNO(3)), hydrofluoric acid (HF), and hexafluosilicic acid (H(2)SiF(6)) using a high-resolution continuum source graphite furnace absorption spectrometer. The concentrations of Si (via Si atom absorption at the wavelength 251.611 nm, m(0),(Si)=130 pg), of nitrate (via molecular absorption of NO at the wavelength 214.803 nm, [Formula: see text] ), and of total fluoride (via molecular absorption of AlF at the wavelength 227.46 nm, m(0,F)=13 pg) were measured against aqueous standard solutions. The concentrations of H(2)SiF(6) and HNO(3) are directly obtained from the measurements. The HF concentration is calculated from the difference between the total fluoride content, and the amount of fluoride bound as H(2)SiF(6). H(2)SiF(6) and HNO(3) can be determined with a relative uncertainty of less than 5% and recoveries of 97-103% and 96-105%, respectively. With regards to HF, acceptable results in terms of recovery and uncertainty are obtained for HF concentrations that are typical for the photovoltaic industry. The presented procedure has the unique advantage that the concentration of both, acids and metal impurities in etch solutions, can be routinely determined by a single analytical instrument.

Analysis of gaseous reaction products of wet chemical silicon etching by conventional direct current glow discharge optical emission spectrometry (DC-GD-OES)

Major parts of the mechanism of the wet chemical etching of Si using mixtures of nitric and hydrofluoric acid are still in question. One of them is the issue about the formation of H and its relevance in the oxidation of Si. In the present work H evolution during acid etching of Si was estimated as a function of the HF/HNO3 mixing ratio and the temperature. A commercial GD-OES instrument with an unchanged Grimm type glow discharge chamber was used to determine the amount of formed H. A small fraction of the reaction gases, originated by etching of Si in a PTFE apparatus under Ar atmosphere, was fed by a constant Ar flow via a drying column into the discharge chamber. The gases were excited by the plasma of a continuously DC sputtered Fe sample. Numerous elements like Si, N, O, H, and even B from the B-doped Si were detected. However, only the H emissions at 121.5 nm and 656.3 nm were used for the estimation of the released H. The obtained results show that H is preferentially generated in HNO3-poor etch solutions (HNO3 < 40% (v/v)). In the range from −10 to 35 °C the temperature of the etch solution shows a negligible influence on the H generation, but affects the amount of nitrous gases. This indicates that particularly in HNO3-poor etch solutions the oxidation of Si proceeds to a considerable extent via the formation of H parallel to the already known pathway via the reduction of HNO3.

Development and validation of a new method for the precise and accurate determination of trace elements in silicon by ICP-OES in high silicon matrices

A new method for the accurate and precise determination of impurities in silicon was developed and statistically validated. Particular attention was paid to the correct determination of the non-metals boron and phosphorus. Instead a time-consuming open vessel digestion under mild conditions, the dissolution of silicon took place in a microwave-assisted high-pressure system. The essential innovation of the presented method is the direct use of the concentrated digestion solution for ICP-OES measurements. This approach avoids the commonly used, time-consuming method that requires the removal of silicon and acid matrix by volatilisation, which is the most critical step in the determination of boron; however, the ICP-OES measurement in such high silicon matrices requires an entirely new optimisation of the measuring conditions, including the careful selection of emission lines with respect to selectivity and, spectral and non-spectral inferences. For quantification of the impurities contents, the methods of matrix matching (MMC) and multiple standard addition (MSA) were used. After optimisation of the spike concentrations for MSA, the qualities of both methods were compared through a statistical analysis. For the metallic impurities Al, Mg, Ca, Ti, Cr, Mn, Fe, Ni, Cu, and Zr and P, the validation was performed against certified reference materials (IPT134, IPT135, NIST57b). To validate boron, 9 silicon samples with different contents of boron from three interlaboratory comparisons were used. The new procedure allows for the determination of the impurities of 4N-silicon (12 elements).

Impact of the chemical form of different fluorine sources on the formation of AlF molecules in a C2H2/N2O flame

The formation of diatomic AlF molecules was studied in a C2H2/N2O flame by means of a high-resolution continuum source flame absorption spectrometer using different fluorine containing compounds HF, H2SiF6, HBF4 and CF3COOH as fluorine sources. The fragmentation of these fluorine sources, as well the resulting impact on the AlF molecule formation, was derived from flame height distribution studies of the atomic and molecular species Al, AlO, Si, SiO, SiF, B and BF as a function of the fluorine concentration, the molar Al:F ratio and the burner gas composition. As a consequence, the used fluorine sources HF, H2SiF6, HBF4 and CF3COOH have been divided into two major groups. The first group of fluorine sources, covering HF, H2SiF6 and HBF4, decomposes during the drying of the aerosol under the formation of AlF3, which is the dominating species for the transport of aluminium into the flame. Its decomposition into AlF results in a high sensitivity of AlF molecular absorption at low flame observation heights. The second group of fluorine sources is exemplarily given by CF3COOH. In the upper parts of the flame the cleavage of the very stable C–F bond proceeds incompletely so that the sensitivity of the AlF molecular absorption is considerably lower than that for the other fluorine sources. In consequence, the AlF molecules are formed by the reaction between the fluorine atoms and the aluminium atoms, which are transported into the flame without the aid of fluorine, presumably via oxidic and/or carbidic species. The present investigations show that the sensitivity of the AlF molecular absorption and the pathway of AlF formation depend on the chemical form of the fluorine in the studied samples.