AVS2012 Session PS1-TuA: Plasma Deposition and Plasma Enhanced ALD

Although very promising results in terms of moisture and oxygen permeation barrier properties have been achieved by organic/ inorganic multi-layers, the impact of the organic interlayer on the global barrier performance is still under discussion. It is generally considered that the organic interlayer acts as smoothening layer for the macro-defects present at the surface or in the barrier layer. It is also hypothesized that the organic interlayer infiltrates into the nano-pores present in the barrier layer, therefore affecting the barrier itself at microstructure level.

In the present work the moisture permeation barrier performance of multi-layers deposited by means of initiated- and plasma enhanced- CVD methods [1] is addressed, with the aim of unraveling the role of the organic interlayer. Calcium test measurements allowed discriminating between the water permeation through the macro-defects/pinholes and the permeation through the matrix porosity. It has been found that the improvement in terms of barrier properties, due to the filling/infiltration of the SiO₂ layer nano-pores by an organosilicon interlayer [2], universally correlates with the residual open micro- (i.e. < 2 nm) porosity in the inorganic (SiO₂ or Al₂O₃) layer, independently of the deposition method (PECVD or plasma-assisted ALD). The residual nano-porosity has been evaluated by means of ellipsometric porosimetry which allows carrying out adsorption/desorption isothermal of a probe molecule in the open porosity. In detail, only for WVTR values of the SiO₂-like layer larger than 10^-3 g m^-2 day^-1, a barrier improvement factor is observed upon deposition of the organic interlayer, although limited to values between 5 and 10. This outcome, in combination with the evaluation of the macro-defect density upon organosilicon layer deposition on the SiO₂ layer, provides the evidence that smoothening plays a major role in the improvement of the barrier performance.

Next to the growth and characterization studies of the organosilicon/SiO₂ multi-layer, the control on the residual open nano-porosity of high growth rate roll-to-roll atmospheric pressure plasma deposited SiO₂ layers is addressed. Upon 2 nm thick Al₂O₃ layer deposition, the water vapor transmission rate of the SiO₂ layer decreases of a factor 10³, due to the filling of the residual SiO₂ nano-porosity.

[1] G. Aresta, J. Palmans, M.C.M. van de Sanden, M. Creatore, (2012), accepted for publication in J.Vac. Sci. Technol.

Plasma polymers have been used since the 1960s and have found a wide range of applications. However the basic processes and mechanisms leading to plasma polymer film growth are poorly understood. Due to this lack of understanding, external parameters such as RF power and monomer flowrate are typically quoted in the literature, which provide no information on processes at the molecular level. Therefore reproducing films with different reactors is usually done by trial-and-error as the critical parameters determining growth processes are largely unknown.

Here we will discuss the roles of ions and neutrals in contributing mass to the film with a view to understanding the molecular processes by which plasma polymer films grow. The pivotal role of monomer chemical structure, specifically sites of unsaturation, in determining which growth mechanisms dominate will also be demonstrated.

Plasma deposition enables synthesis of optoelectronic thin films on flexible substrates at low temperature. The ultimate performance of these materials is controlled by defects such as pinholes or cracks. This work focuses on understanding the formation of such defects during both film synthesis as well as during the application of stress. A simple solvent-etch based technique is presented which can be used to visualize and quantify defects in both single and multi-layered films deposited on flexible substrates. This approach allows for quantitative evaluation of crack/defect density evolution. This technique is applied to both sputtered indium tin oxide (ITO) as well as alumina-silicone nanolaminates deposited on flexible polycarbonate films by plasma-enhanced chemical vapor deposition. The use of the nanolaminate architecture is shown to significantly attenuate pinhole density relative to single layers. In the case of sputtered ITO, an optimum thickness is identified that balances electronic performance with mechanical integrity. For both materials their performance under applied stress was examined using an automated bending test setup that explored mechanical failure as a function of applied strain and number bending cycles. The threshold strain for failure under the application of tensile/compressive stress was quantified by bending samples to a given radius of curvature. Below a critical strain these materials demonstrate robust performance during cycling, displaying their strong potential for device applications.

Biodegradable and derived from renewable resources, polylactide (PLA) has gained enormous attention as an alternative to conventional synthetic packaging materials. Nevertheless, PLA presents high water and gas permeability, which leads to a high degradation rate by hydrolysis of the ester bonds and therefore substantially limits its application for food packaging.

In order to improve the barrier properties of PLA substrate, we propose to cover it with a highly cross-linked PLA-based plasma polymer film (PPF) synthesized by Plasma Enhanced Chemical Vapor Deposition (PECVD). The control of both chemical composition and cross-linking degree would allow to tune the gas permittivity and, as a consequence, the degradation rate of the PLA substrate.

The present work reports on the study of the influences of the experimental parameters (RF power, working pressure and precursor flow rate) on the PPF films properties using a design of experiments tool. The PPF are studied in terms of chemistry and cross-linkage by XPS, FTIR spectroscopy and Tof-SIMS measurements. In addition, in-situ IR spectroscopy is used to probe the plasma in order to get a better understanding on the plasma-surface interactions during the growth process. It comes out that, in our conditions, the applied power is the key parameter controlling the PPF properties. By increasing it, the C/O ratio in the coatings increases from 2.5 to 9.3. The decrease of the oxygen content when applied power is increased is correlated with the loss of O-C=O and C-O functions from 15% to 0 % and from 21% to 5 %, respectively. The in situ FTIR data support this observation and suggest the formation of alkynes in the plasma.

Carbon materials possess various functional properties by forming various structures of carbon atoms. Especially diamond-like carbon (DLC) film is one of the primary materials as alternative of metal electrodes in various devices, in contract a carbon doped silicon oxide (SiCOH) is known as a low-k dielectric in ultralarge-scale integration (ULSI) devices . However, it is difficult to precisely control their properties using conventional plasma-enhanced chemical vapor deposition (PECVD). To solve this problem, we have developed a neutral-beam-enhanced chemical vapor deposition (NBECVD) process as an alternative to the conventional PECVD process. NBECVD can almost completely eliminate the irradiation of UV photons and electrons to the substrate surface by a carbon aperture, resulting in a damage-free deposition process. Moreover, the NBECVD can form a film through surface polymerization caused by bombardment of an energy-controlled Ar neutral beam on the surface with absorbed precursors. We previously proposed controlling the molecular-level structures in SiOCH film on 2 inch wafer by using NBECVD process, which can control the film properties (k-value and modulus). Since the bombardment energy of the neutral beam can be precisely controlled by Ar plasma above a carbon aperture, selective dissociation of weak chemical bonds in the precursors is possible with applicable plasma source, which enables us to design the film structure by controlling the precursor structure.

In this study, we developed NBECVD equipment for 8 inch wafer selecting microwave plasma source and deposited DLC and SiCOH films. For DLC film, we used toluene as precursors to control sp²/sp³ bonds ratio and hydrogen content in the film. DMOTMDS was selected for SiCOH film as reported previously. As the results, we can form high-quality DLC films and high density and low-k SiCOH films on 8 inch wafer by controlling the film structure precisely.

Plasma Enhanced Atomic Layer Deposition (PEALD) relies on plasma generated species as co-reactants for one half of the cyclic, two-step deposition process. Plasma generated species include radicals, ions, electrons, and photons. PEALD research is still in its early stages and the community is just beginning to elucidate the roles of these species in the deposition process and ultimately in the resulting deposited film characteristics.

Plasmas can exhibit quite complex behavior. Without excellent Faraday screening, Inductively Coupled Plasmas (ICPs) will have some component of capacitive coupling between the induction coil and the plasma. This leads to various amounts of extension of the "remote" plasma into the ALD reactor as currents originating at the high capacitive potentials of the plasma source seek out a grounded counter electrode. The sink location for the capacitive currents would be expected to vary substantially depending on if the ALD reactor is coated with a conducting or insulating film.

Ideally the plasma source will deliver a high flux of radicals to the substrate surface. Recombination of radicals on reactor wall surfaces is one of mechanisms by which the radical flux is reduced between the remote source and the substrate. The wall/radical recombination rate is a function of the radical species, the wall material, and the wall temperature.

In a PEALD system which sees multiple film chemistries, at any given time the exposed coating on the reactor walls may be of various conducting or insulating materials. The plasma properties and the radical recombination properties, and thus the flux and characteristics of plasma generated species seen by the substrate, could be substantially different depending on the details of the wall films. Additionally, the wall characteristics could change during the deposition leading to drift in the process over the course of the deposition. This, is turn, could lead to within wafer and wafer-to-wafer non-uniformities.

We have investigated the role of various insulating and conducting wall films on the spatial radical flux and spatial ion flux and energy distribution. Measurements were made on multiple ICP sources designed to deliver radicals to a PEALD reactor in a remote plasma configuration. Measurements were made over a range of rf powers (50-300W), gas compositions (Ar, O2, N2, and mixtures), gas flow rates (10-400sccm), wall films (304L SS, Al2O3, Pt, …), wall temperatures (RT - 200C), and probe/plasma source distances (150 - 450mm). Relative radical flux measurements are derived from the mass loss of photoresist coated QCM crystals. Ion data are derived from Langmuir probe measurements.