Gas incorporation in sputtered and evaporated metal films

The argon and nitrogen gas concentrations in thin films have been measured for seven different metals. Sputtered and evaporated films are deposited on glass and tungsten substrates in a stainless steel ultra high vacuum system, which is designed to operate at background pressures in the 10-8 torr region, and at argon pressures of about 6 10-4 torr. The discharge is maintained by a triode type system and the plasma is confined with the aid of two magnet coils in a helmholtz configuration. The vacuum system is evacuated with ion and liquid nitrogen-cooled sublimation pumps. A second vacuum system is connected to the main deposition chamber by means of a gate valve so that the film samples can be withdrawn for study, using a rack and pinion mechanism, under continuously maintained vacuum conditions. The films are analysed by thorough outgassing in the second system, using an induction coil heater over a glass section of the system, so that desorption from other parts of the system is kept to a minimum. Gas analysis is achieved using a quadrupole mass spectrometer and the results are displayed on an XY recorder. Film thicknesses are monitored using a previously calibrated quartz crystal oscillator and the dimensions of samples and systems are known from careful measurement. The argon and nitrogen concentrations are measured as a function of gas pressure, film thickness. target voltage. target and substrate species and substrate bias voltage. The sputtered films are always found to contain more argon than the evaporated ones. This may be accounted for partly by the additional process of ion reflection, which takes place at the sputtered target surface but not at the evaporated one. Increasing the bias voltage causes an increase in the argon concentration in most of the films. This can be related to the noble gas clean-up effect, where energetic ions embed themselves in the film and become trapped by the continuous flux of metal atoms arriving at the surface. The results are in reasonable agreement with a theoretical model proposed by WINTERS et al. A comparison of the theoretical and experimental results makes it possible to derive sputtering yields for the gas. For argon, typical sputtering yields of ~ 10-16 atoms/ion are obtained at high bias ion energies, indicating that the gas release is being brought about by a gas sputtering mechanism. Yields of the same order as those of the metal suggest that gas release by a target sputtering process predominates at lower ion energies. However, the division between mechanisms is not well defined and depends to a great extent on the target species. The results for nitrogen incorporation support Winters proposal that metals can be divided into three classes. The results show that the simple hypothesis that bias sputtering always brings about a reduction in the impurity gas contamination does not hold for all gas/metal combinations. It is shown for nitrogen, that the gas concentration can increase or decrease depending on the class of the metal and on the partial pressure of nitrogen present during deposition. Some of the films have been examined using an electron microscope. A significant change in the surface appearance is observed in some the outgassed sputtered films which is not apparent in either the nonoutgassed films or in the evaporated samples. The reasons for this are not clear at present