The reliability of novel flexible opto-electronic devices depends heavily on the resilience of a thin ceramic oxide layer deposited on a polymer substrate. Currently, one of the most popular combinations consists of a thin layer of Indium Tin Oxide (ITO) on a polyester substrate, such as polyethylene terephthalate (PET). The ITO layer, usually around a few hundred nanometers in thickness, is very susceptible to cracking. As this layer experiences cracking and delamination from the substrate, the resistance of this layer sharply increases and it is rendered useless.
Characterization of the mechanical properties of this oxide layer after deposition is very important. The properties of ITO deposited on glass have been previously investigated, but because the ITO layer has an amorphous structure, the properties of the ITO can be quite different than when deposited on glass. A large mismatch in modulus between the ITO and the polymer substrate can also affect adhesion to the substrate and the measured hardness values. For this reason, indentation and scratch testing of the ITO-coated PET system is very valuable, but straightforward testing might not always be an option.
There are a number of challenges when performing both indentation and scratch testing on a system consisting of a thin hard coating on a soft polymeric substrate. Care must be taken to ensure substrate effects do not influence the coating data.
Figure 1: Load depth curves for 3 coating thicknesses.
Figure 2: Optical micrographs of residual indents for 4 applied normal loads and 3 coating thicknesses (1000x magnification).
The techniques described here include nanoindentation with a spherical indenter to promote circumferential cracking of the brittle layer and nanoscratch testing to promote adhesive failure.
For nanoindentation testing, a 20 μm spherical indenter was oaded to normal loads up to 200 mN with a pause of 10 seconds. The goal of this style of indentation testing was to promote cracking of the ITO layer. In all cases, the first visible crack appeared at approximately 40 mN. At 100 mN a second circumferential crack was observed, while at 150 mN a third crack was present. Radial cracking was also observed at a load of 200 mN for the coating thicknesses of 50 nm and 100 nm. Severe damage of the 50nm thick coating was observed at 200mN. Optical micrographs of each indent can be seen in Figure 2.
Penetration depths of several microns were observed for the films. The load-depth curves presented in Fig. 1 show a small variation between samples due to coating thickness. The diameter of each crack was measured optically. The primary circumferential crack diameter for all samples and loads was equal to the diameter of the indenter itself (20μm). This shows cracking was promoted by the compliance of the polymeric substrate.
As the indenter first makes contact and load is increased, a primary crack is formed. Further loading elastically deforms the substrate while causing cracking and delamination of the ceramic coating. Future work will model this contact with the goal of understanding this failure mechanism in more detail.
Figure 3: Panoramic comparison of scratches on each sample, (500x magnification). Applied load range was 0.08 – 5 mN.
Nanoscratch testing was performed using a 5 μm radius spherical diamond indenter. Samples were adhered to glass slides for testing. Low-load scratching was performed using the High Resolution cantilever of the Nano Scratch Tester (NST). Critical loads were determined using optical methods and were compared for several coating thicknesses.
Two primary failure mechanisms were observed for all samples. The first mode of failure during testing was rupture of the ITO layer. Further failure occurred in the form of spallation of the coating and scarring of the PET substrate. A panoramic comparison of a scratch performed on each sample is presented in Figure 3.
Scanning Force Microscopy (SFM) was performed at the critical failure points of the sample with a coating thickness of 250 nm and is presented in Figure 4.
The load at failure was also plotted against coating thickness and is shown in Figure 4. This graph shows that the failure mechanism of spallation of the coating has a greater dependence on coating thickness than a failure characterized as rupturing. Scratch width at the critical loads was also measured using optical methods for each scratch and was plotted against film thickness. This plot can be seen in Figure 5. Scratch widths at the critical loads appear to be less dependent on film thickness than the critical load values themselves.
When attempting to determine the mechanical properties of a transparent oxide deposited on a thin polyester film, it is necessary to adapt indentation and scratch testing methods. Indentation testing utilizing a spherical indenter to promote circumferential cracking and low-load scratch testing with a high- resolution friction table were used to characterize and compare the mechanical properties of the composite films. Results show that these methods can accurately characterize differences in film thickness. Further developments in these testing methods will allow for a more flexible range of tests that can be conducted on thin composite films. This will allow correlations to be made between laboratory sample testing and the actual in-service performance of devices which utilize ITO technology (e.g., touchscreens, flexible solar cells, flexible LED lighting, etc.)
1) K. Zeng, et al. Thin Solid Films 443 (2003) 60-65
2) B. R. Lawn, Journal of Materials Research, Vol 17, No. 12, Dec 2002
3) H.Chai et al. Journal of Materials Research, Vol 19, No. 6, Jun 2000