An alternative to elevated temperatures and powerful molecular reducing agents is a technique referred to as electron chemical vapor deposition (e-CVD), which uses plasma electrons instead of a molecular reducing agent for the redox chemistry. However, only a phenomenological understanding of the process is available. This thesis aims to study the details of the e-CVD process to build a better understanding for the chemical and physical processes governing it. To start comprehending the surface chemistry, an electrically modified quartz crystal microbalance (QCM) probe was developed for the e-CVD process, a sensor capable of sensing very small mass differences. This gave us information concerning precursor adsorption, net mass gain, and the dynamics of the process. We noted that pulsed e-CVD yielded denser films than continuous e-CVD. To untangle the plasma characteristics of the e-CVD plasma discharge, an RF Sobolewski probe was employed in order to gain information of the electron temperature and electron density in the plasma along with the plasma- and plasma sheath potential. These results provide a more solid understanding of the boundaries for the plasma chemical reactions, which were later used to correlate various decomposition reactions of the precursor ferrocene. To compare metal-carbon (M–C) coordination and metal-nitrogen (M–N) coordination in the e-CVD process, a deposition study using Fe–C (ferrocene) and Fe–N (iron amidinate) were performed. The results show that it does appear to be certain differences when these types of precursors are used for iron deposition. Importantly, this thesis also shows that the carbon contamination can be mitigating when pulsing the process. Finally, deposition of copper containing films was studied, revealing self-limiting characteristics using the electrically modified QCM.
The findings in this thesis gives knowledge of the plasma dynamics in the e-CVD process and for plasma CVD processes in general. In addition, this thesis contributes with instrumental efforts that can be employed in any e-CVD scheme to understand the deposition puzzle.