No abstract

This paper constitutes an analysis of the forces of adhesion of small particles to surfaces, most specifically as applied to semiconductor surfaces. The primary forces of adhesion of small, less than 50 ~m diam particles on a dry surface are van der Waals forces. These van der Waals forces of adhesion can increase as a function of time due to particle and/or surface deformation which increases the contact area; micron-size particles can be held to surfaces by forces exceeding 100 dyn, which corresponds to pressures of 109 dyn/cm'-' or more. Total forces of adhesion for micron-size particles exceed the gravitational force on that particle by factors greater than 106. Electrostatic forces only become important and predominate for particles larger than 50 ~m diam. Immersion of the adhered particle system can, in some cases, greatly reduce the total adhesion force, first by shielding of the electrostatic and van der Waals attractions, and also by adding double layer repulsion because of dipolar alignment of liquid molecules or dissolved ions at the surfaces. Double layer interactions may, however, also add to the attractive forces if dipoles align properly for attraction. An important consideration is the possibility that if the particles are not removed by the liquid immersion, then a liquid bridge can be formed by capillary action between the particle and surface upon removal from the liquid. This would add a very large capillary force to the total force of adhesion. This capillary force has been shown to remain, in some cases, even when the system is baked at above the liquid boiling point for more than 24h. Removal of these small particles from surfaces is in theory possible but is in practice extremely difficult. It is clear that emphasis should be placed on prevention of particle deposition rather than on counting on achieving subsequent removal.

The chemical vapor deposition (CVD) and reflow of phosphosilicate glass (PSG) have been reviewed. The CVD reaction from silane and phosphine has been simulated by SOLGAS thermochemical calculations. This equilibrium calculation method predicts stoichiometric conversion of silane and phosphine. In comparison to experimental determinations, the calculations underestimate the glass phosphorus content by about 1–3 weight percent. In addition, recent observations that the reflow of PSG is greatly enhanced by the presence of a steam ambient during heating have been examined. A possible mechanism of this reflow enhancement has been discussed. Experimental evidence, which support this mechanism, by electron microprobe, FTIR, and SIMS studies of PSG films, are presented.

This paper will describe theoretical calculations of the gravitational fall of particles in vacuum as a function of pressure and particle size, and will give confirming experimental observations of particle deposition on silicon wafers in a test vacuum system. Also, using silicon wafers as test vehicles, observations will be described about the effects of wafer orientation, pumpdown and venting rates, and moving parts on the generation and behavior of particles in a vacuum system. Further confirmation of these results will be presented through the use of a modified laser particle counter in vacuum to directly measure particles in the vacuum chamber.

When HF/H2O vapor derived from an azeotropic solution of HF/H2O was used to etch various oxides of silicon, the etch rates were found to have qualitatively similar dependence on the wafer temperature. However, the sensitivites of the dependence were found to vary greatly among the different oxides. Consequently, the etch selectivities among different oxides could be controlled by changing the wafer temperature. In particular, the phosphosilicate glass to thermal oxide etch selectivity increased from about 18:1 to at least 2900:1, when the temperature was raised from around 25°C to around 50°C.