Monitoring of High Selectivity Chemistry for Si3N4 Etch

Abstract

Concentrated phosphoric acid at nearly boiling temperature (165°C) has been used for many years in Selective Silicon Nitride etch process. Traditionally, inorganic silicate is added to improve the selectivity of Si3N4 over SiO2. Recent semiconductor processes such as 3D Memory and FinFET logic require an enhancement in Si3N4:SiO2 selectivity in excess of 100. The low solubility of Si in H3PO4 limits this Si-addition approach for process that requires high Si3N4:SiO2 selectivity. In recent years, H3PO4 etchant containing organosilicate has been developed to improve the solubility of Si, enabling both high selectivity and desirable etch rate. Another important function of organosilicate is to prevent etched SiO2 from re-deposition to the wafer surface. This high selectivity process using organosilicate presents new metrology challenges. A reagent electrochemical method was developed to monitor Si in hot phosphoric process tanks. The same method with improved reagent was also successfully applied for monitoring organosilicate. The accuracy of this is evaluated by off-line ICP-MS method, which is shown in Figure 1. Samples were obtained from an actual process tank. The results from this improved fluoride method match those from ICP-MS. Good stability of organosilicate results by the same method is shown in Figure 2. This test was performed on samples with a known prepared concentration for a period of one week. All measured results have an accuracy of less than 2% (solid lines on Figure 2). A recent study with the combination of the two methods above demonstrates the capability of measuring two different Si forms. In this study, two separate measurements using two different reagents were done first. Two equations were derived from signals/results obtained from these two tests, to calculate both organo- and inorganic Si. Figure 3 shows these calculated organo- and inorganic Si results in comparison with expected values. In addition, ICP provides an alternative metrology option to measure both silicon and carbon elements. The organo-Si portion can be derived from carbon concentration assuming no other source of carbon in the matrix. However, organic solvents are commonly used in preparation of H3PO4 etchant containing organosilicate, and evaporation at high process temperature would break the correlation between organosilicate and total carbon. Figure 4 compares the ICP results before and after the heat treatment, and a dramatic decrease in the total carbon concentrations is observed, while the same trend was not seen in total Si concentrations. Slight increase in Si concentrations is due to the loss of solvent during heating. Therefore, ICP measurement of organosilicate targeting carbon element needs to be scrutinized case by case to avoid possible solvent interference. Reference: [1] S.J. Baffat, M.S. Lucey, M.R. Yalamanchilli Hot Phosphoric Acid APC for Silicon Nitride Etch. Semiconductor International, 8/1/2002 [2] Hong et al. Compositions for etching and methods of forming a semiconductor device using the same US Patent Application 20130092872 [3] Cho et al. Etching composition and method for fabricating semiconductor device using the same US Patent Application 20130157427 [4] Shalyt et al. Analysis of silicon concentration in phosphoric acid etchant solutions US Patent 8,008,087 [5] Shalyt et al. Analysis of silicon concentration in phosphoric acid etchant solutions United States Patent Application 20160018358 [6] Bai et al. Metrology for High Selective Silicon Nitride Etch Solid State Phenomena 255:81-85 Figure 1