189
ferred to the titration flask while open to the air. This rule is iimited, of course, by the hydrolyzabitity of specific halosilanes with atmospheric moisture. For example, although hexachlorodisilane and dichlorodiisopropoxysilanc boil only xx0 apart, 146" and 157" respectively, the hexachlorodisilane cannot be sampled except by sealed ampoule, whereas the dichlorodiisopropoxysilane can be sampled quantitatively in the open by acting quickly. Bromo-and iodosilanes, whatever the boiling points, hydro'tyzc rapidly with atmospheric moisture and may only be sampled by means of ampoules.
Iodosilan(es have been determined, following reaction wit?1 excess ethanolic base, by hack-titration with standard acid 7. This method precludes the formation of ethyl iodide. An analysis of an iodosilane was also included in this paper since it was probably of a higher purity than was necessary for the above reference. Ioclosilane assays by the direct nanaqueous method of this paper were not quantitative, When aqueous tcrt,-hutanol was substituted for both the ethanol (tetrabutyl-phosphonium hydroxide as base) and the isopropanol solvent, quantitative results were approached. Too much time, however, was required to reach a stable end-point.
Hyclrogcn gas was not detected by mass spectrometry during titrations of chloromethylpllenylsilane and hcxachlorodisilane.
Undetected quantities of hydrogen may possibly have evolved from the reactions between silicon-hydrogen or disilane linkages and alcol~ol or moisture during the addition of base. The role of t.his base would have been catalytic, however, and no net loss of titer would have resulted. Potassium halides were precipitated during the chloro-and bromosilane titrations but in no way obscured the end-points or lessened the accuracy of the method. Addition of a few ml of water following the titrations cleared the solutions, even in the cases of silicon tctrachloride and hcxnchlorodisilanc.
Halide precipitates were eliminated by titration of halosilanes with tetrabutylan~monium hydroxide. Under comparable conditions, however, the yellow to blue encl-point of this base was not as sharp as when using the potassium hydroxide titrant.
The authors thank IL N. M~HEEIXR for technical assistance.