Dust Particle Diagnostics in Rf Plasma Deposition of Silicon and Silicon Oxide Films (Invited)

Abstract

Particle contamination formed in reactive plasmas imposes an upper limit on the rate for particle-free deposition. Conversely, these plasmas could be exploited to produce nanometric clusters and particles for various applications. Infrared absorption spectroscopy has been applied to analyse the chemical composition of suspended particles. Mass spectrometry was also used to investigate cluster formation in these deposition plasmas. In pure silane plasmas, a random model reproduces the measured mass spectra, whereas the rich plasma chemistry in silane/oxygen mixtures shows a remarkable tendency to produce silasesquioxane anions.

Modelling of silicon hydride clustering in a low-pressure silane plasma

“…Also, the concentration of the neutral clusters is larger than that of the anions as observed earlier. We can qualitatively compare this with the results of Hollenstein et al [48], who have obtained mass spectra of silicon hydride clusters in their experiments with RF plasmas using pure silane or silane mixed with oxygen. The authors have used a reactor uniformly heated to 473 K. In their experiments, they observed anionic species which were more dehydrogenated than those predicted by our mechanism.…”

“…The presence of a bottleneck in the silyl anion clustering chain also makes this chain slower than the silylene anion clustering chain. Hollenstein et al [48] also observed an exponential decrease in the number density of anions for bigger clusters whereas our simulations indicate that all anionic clusters with more than five silicon atoms (except for anionic clusters with seven silicon atoms) show nearly equal density. However, in these discussions we should bear in mind that mass spectra predicted by our simulations and those observed by Hollenstein et al [48] correspond to different situations.…”

“…Hollenstein et al [48] also observed an exponential decrease in the number density of anions for bigger clusters whereas our simulations indicate that all anionic clusters with more than five silicon atoms (except for anionic clusters with seven silicon atoms) show nearly equal density. However, in these discussions we should bear in mind that mass spectra predicted by our simulations and those observed by Hollenstein et al [48] correspond to different situations. Our model predicts the spectra in an active steady state discharge whereas Hollenstein et al [48] had to observe their mass spectra in the afterglow of a pulsed discharge.…”

“…However, in these discussions we should bear in mind that mass spectra predicted by our simulations and those observed by Hollenstein et al [48] correspond to different situations. Our model predicts the spectra in an active steady state discharge whereas Hollenstein et al [48] had to observe their mass spectra in the afterglow of a pulsed discharge. The measured mass spectra are likely to be modified by electron attachment, which becomes very efficient in the cold afterglow of a plasma.…”

Polymorphous Silicon Films Deposited at 27.12 MHz

“…The formation of these powders increases as the deposition pressure increases, [3] and as the powders are negatively charged and can be retained close to the plasma frontiers, they disturb the plasma thus resulting in non-uniformities in the film. [15] A strong correlation of deposition pressure with film uniformity can, therefore, be expected, and indeed for the reactor configuration used, at pressures up to 240 Pa the uniformity of the films was better than 90 %, but decreased to » 70 % as the deposition pressure increased to values of about 290 Pa. For this deposition pressure, and with a power density below 110 mW cm ±2 , we achieved growth rates of about 0.3 nm s ±1 , about twice more than the values reported by Roca i Cabarrocas and coworkers, using pressures of the same magnitude and an excitation frequency of 13.56 MHz. …”