Characterization of a-CSi:H films prepared by PECVD in terms of adhesion

Highlights

• Comparison and discussion of critical normal load and work of adhesion results.

• Critical normal load increased with an increase of power by up to 2.7 times.

• Work of adhesion decreased mainly due to a significant increase of Young's modulus.

• Work of adhesion was found to be more appropriate as a measure of adhesion.

• High stability of films with respect to adhesion was observed over 5 years.

Abstract

Mechanical properties of thin films and, in particular, adhesion are crucial engineering parameters which must be determined to allow their successful application both in the industry and science. Deposited hydrogenated amorphous carbon-silicon (a-CSi:H) films prepared by plasma-enhanced chemical vapour deposition (PECVD) were tested using a nanoscratch test. The load at which adhesion failure occurs is known as the critical normal load and is used as a semi-quantitative measure of adhesion. However, results in this article suggest that the critical normal load is not an appropriate measure of adhesion for these materials, and instead, the work of adhesion should rather be considered as an alternative method. This is due to the considerable variation of Young's modulus, which affects the critical load figures, while the friction coefficient and film thickness also do not have negligible influence. Moreover, the critical load is considerably influenced by intrinsic and extrinsic parameters. Furthermore, it has been found that the critical load increased with increasing a-CSi:H film thickness, but this did not apply to the work of adhesion. On the contrary, it remained constant unless a change of film failure mode occurred. Adhesion was monitored for 5 years to determine the effect of aging and the results were found to be practically identical over this duration which indicates high film stability.

Introduction

Film adhesion together with its mechanical properties is one of the most important material properties for industrial and scientific applications. Thin films prepared for example by PVD (physical vapour deposition) or CVD (chemical vapour deposition) with controlled adhesion are essential in many applications such as anti-scratch, barrier, transparent, wear-resistant, bioactive and anti-reflective coatings or in unconventional optics for surface modified materials to achieve additional functional properties [[1], [2], [3], [4], [5], [6], [7]]. The greatest potential of this research for industrial applications is in functionalized surfaces, protective coatings and mainly in the glass-fibre reinforced polymer composites without sharp interfaces, mimicking natural and biological systems [8].

Advanced materials, such as composites, are based on the principle of mutual reinforcement of their components. This means that transfer of stress takes place from the matrix to the reinforcement (particulates or fibres) when the material is under stress. To ensure efficient transfer of stress, a strong adhesion among all components is necessary. In the case of glass fibres, it is therefore beneficial to modify them with a thin layer (or a gradient multilayer), resulting in compatibility and bonding between the fibres and the polyester matrix which allows to increase adhesion, hence improving the properties of the composite [[8], [9], [10]]. This can be achieved uniquely by means of films prepared by plasma-enhanced chemical vapour deposition (PECVD) using a tetravinylsilane precursor [[11], [12], [13]].

The value of the load at which film adhesion failure is detected from the nanoscratch test is known as the critical normal load (L_c) and is used as a semi-quantitative measure of film adhesion to the substrate [[14], [15], [16]]. However, this value usually provides sufficient information for comparative purposes [16,17]. The L_c also corresponds to the first significant fluctuation of the lateral force component in the loading curves. This marked fluctuation in the lateral force corresponds to the point of adhesion failure, which is usually also verified by either Optical Microscopy (OM), Scanning Electron Microscopy (SEM) [18] or Atomic Force Microscopy (AFM) [19]. The choice of these microscopic methods depends on the dimensions of the scratches, as well as on the thicknesses and the character of films. It is also possible to employ the measurement of acoustic emissions during the scratch test to detect L_c [15].

Since the L_c value is influenced by many extrinsic and intrinsic parameters [14,20], further proposed determination of the work of adhesion seems to be more appropriate criterion used to characterize film adhesion to a substrate. One of the first models used in determination of the work of adhesion was the energy balance approach; this assumes that the film under compressive stress in the front of the indenter releases its accumulated elastic energy by cracking, spallation of the film and delamination from the substrate at L_c. The stress responsible for adhesion failure can then be related to the work of adhesion W_adh = σ_appl(x)²·t/2E, where σ_appl(x) is the applied stress at the leading edge of the indenter, t is the film thickness and E is the Young's modulus of the film [[21], [22], [23]]. However, this model does not adequately describe the stresses in materials that exhibit plastic deformation [24]. Burnett and Rickerby identified three contributing factors responsible for the film detachment from the substrate during the nanoscratch test, namely ploughing, internal stress and adhesive friction. The friction force F is then a linear sum of these three contributions and is related to the W_adh [25]. Based on later studies by Bull et al.; F is equal to L_c·μ_c, where μ_c is friction coefficient at L_c. Finally, work of adhesion can be expressed from equation

$$L_{\mathrm{c}}=\frac{\pi {d_{\mathrm{c}}}^2}{2\nu {\mu }_{\mathrm{c}}}{\left(\frac{2E{W}_{\mathrm{adh}}}{t}\right)}^{\frac{1}{2}}$$

where d_c is the critical scratch width and ν is Poisson's ratio [24].