Thin film materials are widely used in fields such as semiconductors, coatings, energy storage, medical equipment, and photovoltaics, and their performance and reliability are highly dependent on the interfacial bonding strength. Traditional measurement methods such as peel tests and micro scratch tests are often limited by the testing conditions and material properties, making it difficult to achieve quantitative characterization of interfacial bonding energy. To this end, Dr. Shatha et al. from Oxford University systematically studied the interfacial delamination behavior of tungsten coatings with different thicknesses on silicon substrates using two methods: nanoindentation and nanoscratching. The energy calculation was corrected through finite element simulation, providing new ideas and methodological support for the quantitative evaluation of interfacial energy.

This article describes tungsten coatings with different thicknesses (1.5 μ m and 1.9 μ m) prepared by magnetron sputtering on silicon substrates, and conducts nanoindentation and scratch testing. Nanoindentation tests were conducted using Berkovich indenters in the low load range of 50-400 mN and the high load range of 100-1000 mN, respectively. Three loading rates of 52, 104, and 260 mN/s were set to investigate the effect of strain rate. The nano scratch test also uses a Berkovich indenter, conducting scratch experiments in two directions: "edge forward" and "facing forward", under a linear normal load of 0-70 mN. The scratch length is set to 250 μ m and the time is 30 seconds.

Figure 1 SEM images of three indentation failure stages of thin films with different thicknesses (left: 1.5 µ m, right: 1.9 µ m): (a-b) plastic deformation; (c-d) Interface separation; (e-f) Coating fragmentation

Figure 2. Indentation force displacement curves of films with different thicknesses (left: 1.5 µ m, right: 1.9 µ m): (a-b) low load; (c-d) High load

Figure 3: Changes in Irreversible Work (Wirr) with Normal Load for Thin Films of Different Thicknesses (Top: 1.5 µ m, Bottom: 1.9 µ m)

By observing the indentation morphology through scanning electron microscopy, it was found that the tungsten silicon film exhibited a typical three-stage failure behavior during the indentation process: the initial stage was plastic deformation and the generation of surface radial cracks (Figure 1 (a-b)); As the load increases, it enters the interface delamination stage, forming a circular bulge (Figure 1 (c-d)); Thin film fragmentation occurred in the final stage (Figure 1 (e-f)). The corresponding load displacement curve shows that the curve is relatively smooth under low load conditions, while there is a significant abrupt change under high load conditions, corresponding to the fragmentation process of the thin film (Figure 2). To quantify the interfacial delamination energy, the Malzbender energy analysis method was used to linearly extrapolate the irreversible work (Wirr) at each stage, decomposing the total Wirr into plastic work (Wp), delamination work (Wd), and fracture work (Wf). However, the traditional linear extrapolation method mistakenly assumes that the plastic work always increases linearly with the load, ignoring the contribution of plastic deformation of the soft substrate as the indentation depth increases. As the depth of indentation increases, the energy dissipated by the plastic deformation of the substrate is erroneously included in the interfacial delamination energy, resulting in a significant overestimation of the delamination energy. To avoid this error, the article conducted finite element simulations without interface delamination, corrected the plastic work of the thin film system, and effectively eliminated the interference of substrate deformation (Figure 4). In addition, by comparing the nanoindentation test results under different loading rates, it can be found that the indentation force displacement curve is basically consistent with the indentation morphology, and the difference in peel energy measurement results is small (Table 1).

Figure 4 Wd of the first and second stages after FEM (interface not separated) correction

Figure 5: Nano scratch test results of thin films with different thicknesses (top: 1.5 µ m, bottom: 1.9 µ m): (a) SEM images of surface morphology; (b) The variation of normal load and friction coefficient over time

In the nano scratch test, the coating suddenly fractures upon reaching the critical load, accompanied by a sudden change in friction coefficient (Figure 5). The study uses the Laugier model to convert critical loads into interfacial bonding strength and is based on elastic mechanicsStrain energy release rateBasic theory, calculate stripping energy. The results showed that the peel energy obtained from scratch testing was significantly affected by the test parameters: an increase in the ratio of normal load to scratch length (dL/dx) would lead to an increase in critical load and a decrease in peel energy (Table 2); The "edge forward" direction of the indenter will result in higher critical loads and peel energies compared to the "face forward" direction (Table 3). Although the scratch method can be used for comparative studies between different samples, its results are generally higher than those of nanoindentation due to the inclusion of film fracture energy, and it is extremely sensitive to testing conditions, so there are certain limitations in quantitative evaluation.

Table 1 Changes in peel energy measured by nanoindentation at different loading rates

Table 2 Changes in Peel Energy of Nano Scratch Measurement under Different Scratch Parameters

Table 3: Changes in Peel Energy of Nano Scratches Measured at Different Scratch Angle

Comparing the two methods comprehensively, nanoindentation combined with finite element correction can provide stable and quantitative interface energy data, and is insensitive to loading rate, demonstrating good reliability; Although nano scratches can effectively simulate actual friction conditions, they are more suitable as qualitative or semi quantitative research tools due to their complex energy composition and strong parameter dependence. On this basis, further research can be conducted on tough transparent coatings (such as PC), combining in-situ observations and load curves to distinguish between interface delamination and coating body fracture processes, in order to improve the accuracy of critical load identification and obtain more accurate interface delamination energy.