The primary aim of this thesis is to contribute to the knowledge of the plasma assisted growth mechanism of carbon based thin films. This is achieved by means of studying the plasma chemistry in an Ar/C2H2 Expanding Thermal Plasma (ETP) used for deposition of hydrogenated amorphous carbon (a-C:H) and relating the measured hydrocarbon radical density quantitatively to the observed growth rate and film properties. The plasma composition has been studied by means of two ultra-sensitive techniques, which have been specifically designed and constructed for this purpose: Cavity Ringdown absorption Spectroscopy (CRDS) and Threshold Ionization Mass Spectrometry (TIMS). The ETP technique has been selected for its unusual high deposition rate (up to 70 nm/s), low electron temperature allowing to neglect all electron impact induced dissociation processes, the absence of energetic ion bombardment, the near ideal plug-down geometry suitable for straightforward quasi one-dimensional modeling of the plasma chemistry and the remote plasma character enabling an easy independent parameter study. CRDS has been applied to measure with high sensitivity C, CH and C2 radicals and to study the broadband absorption feature observed in the ETP. However, the range of the species, which can be measured by means of CRDS, is limited by existence of optical transitions in the spectral region accessible by the laser system used. A triple stage TIMS diagnostics was designed and constructed to complete the understanding of the plasma chemistry. In order to make quantitative measurements a novel background signal correction and an absolute density calibration procedure was developed and applied. In this study previous results were corroborated, i.e. the primary C2H2 decomposition is argon ion induced in subsequent charge transfer (CT) and dissociative recombination (DR) steps, and new insights in the plasma chemistry have evolved. The measurements presented in this thesis lead further to the hypothesis that C and C2 radicals are formed in similar amount as C2H in the primary C2H2 decomposition, a consequence of additional internal energy available in the CT and DR reactions. It was further confirmed that the plasma composition and film properties are determined by the ratio of the C2H2 flow and the argon ion and electron fluence emanating from the plasma source. Two limiting cases related to the ratio between the injected C2H2 flow and argon and electron fluence are observed. On the one hand, if the C2H2 flow into the reactor is smaller then emanating argon ion and electron fluence, the C2H2 is fully depleted and decomposed into C, C2, C2H and CH radicals. These radicals are then responsible for the growth of soft polymer-like a-C:H films, as confirmed by film property measurements. On the other hand, if the C2H2 flow is larger than the argon ion and electron fluence, C2H2 is abundantly present in the gas phase and the reactions of the primarily formed radicals (C, CH, C2 and C2H) with C2H2 determine the plasma composition close to the substrate. Two plasma chemistry branches have been identified in the latter case. In the first branch reactions of C2 and C2H radicals with C2H2 lead to the formation of stable diacetylene (C4H2) and further reactive C4 and C4H radicals. Stable hydrocarbon molecules with an even number of carbon atoms (C2nH2), which are commonly measured in C2H2 plasmas, are formed in this branch. The observed high gas phase reactivity of C2H, C4 and C4H hydrocarbon radicals excludes them from being the dominant growth precursors in an Ar/C2H2 ETP, falsifying the previous film growth hypothesis based on the C2H radical. The other branch involves reactions of C and CH radicals with C2H2 to form C3 and C3H radicals as products. These radicals are found to be abundantly present close to the substrate, since they are resonantly stabilized and hence unreactive in the gas phase with C2H2 and other stable hydrocarbons. The C5 and C5H radicals, formed most probably in reactions of C and CH with C4H2, show a similar behavior. The proposed plasma chemistry scheme has been successfully tested by means of a quasi one-dimensional plasma chemistry simulation model. Quantitatively the C3 radical density close to the substrate is found to be the highest among all radicals measured. Since the C3 density behavior as function of the acetylene flow mimics the measured film growth rate, the C3 radical is proposed to contribute significantly to the growth of hard a-C:H film. The film analysis results imply that hydrogen has to be incorporated into the growing film in order to explain the hydrogen content of about 30% for the best quality films. However, further research is required to understand whether hydrogen atoms or additional hydrogen containing species are incorporated during the film growth process.
|Qualification||Doctor of Philosophy|
|Award date||6 Oct 2004|
|Place of Publication||Eindhoven|
|Publication status||Published - 2004|