There has been increasing demand for polymer film formation by physical vapor deposition for optical and optoelectronic devices. Most of the polymers do not evaporate, and have to be synthesized on the substrate surface. We are describing two types of polymeric film formation by using the ionization-assisted deposition (IAD) polymerization method.

The first method co-deposits two bifunctional monomers to cause stepwise reaction on the substrate surface. As an example, polyurea films were obtained by polyaddition of diisocyanate and dipiperidine. The IAD method has a special feature that the polymerization and film growth proceed under a strong electric field that is generated by the substrate bias voltage and the ionic charges accumulated on the film surface. This results in the dipole orientation of the film molecules and induces optical nonlinearity. As a consequence, the IAD-polymerized polyurea film showed second harmonic generation and electrooptic effect in as-deposited status without the conventional poling process that has been required to orient the dipoles in the ordinary polymer films.

Another method is to deposit single material that has a polymerizable group and to activate stepwise reaction by generating initiators through the ionization process. It was found that the IAD can form polymer films of some vinyl compounds that cannot be polymerized by the conventional vacuum deposition method. This method was utilized to prepare a triphenylamine-containing polymer for the carrier transport layer of organic light emitting diode. The device stability was improved by using the polymerized film instead of its monomer counterpart.

The plasmapolymerised thin films are of increasing importance in the field of transparent packaging. In particular, hydrocarbon coatings prepared by the method of reactive magnetron sputtering combined with plasma-stimulated gas-phase polymerisation have proved to have excellent diffusion barrier properties for gases and water vapour. Due to their intrinsic flexibility, this barrier type is a potential candidate to widen the existing application fields.

However, the non-linear relationship of stretch failure and permeation properties is a challenge for modelling the functional coatings so as to meet the required specifications of the product. An improvement of the functionality of these thin films has been achieved using the general regression neural network GRNN. The modelling of the plasma process allows to tailor the properties of the functional coatings. Without knowledge of the mathematical relationship concerning the process parameters and the coating properties, GRNN generates an optimised set of input parameters and supports the scientist in maximizing the return on experiments.

A deeper understanding of the nature of organic thin films is necessary in order to explain the unique properties of this diffusion barrier type. The degree of the polymer-like phase in the structure and its effect on the functional performance will be presented. The investigation of the composition and structure was performed using elastic recoil detection, Rutherford backscattering, nuclear magnetic resonance, and Raman spectroscopy. As a conclusion, the combination of the analysis of the coatings and the modelling of the plasma processes enables the production of thin films in a more controlled and systematic manner.

The nucleation and growth of hydrocarbon thin films through molecular cluster-surface collisions is examined using atomistic simulations with a well-known reactive empirical-bond order potential for hydrocarbons. Experimental evidence exists that strongly adhering diamond-like carbon or polymer thin films can be generated through the impact of small molecular organic clusters with mica, diamond, or glass surfaces. Our goal is to better understand the dependence of the film structure the reaction conditions. The simulations show the atomic-scale mechanisms by which the films nucleate and reveal the conditions needed to tailor the structure of the film.

Supported by the National Science Foundation (CHE-9708049), the Petroleum Research Fund and the University of Kentucky Center for Computational Science