Characterization of Thin Films Using High Definition Confocal Microscopy

Thin films are employed in many different industrial fields including tool and die production for machining, corrosion and wear prevention, as well as functional and decorative coatings. As an example, the lenses in all modern optical systems, like microscopes, binoculars, or eye glasses, are coated with multi-layered films which have various functions (optical, mechanical, or corrosion resistance).

A wide variety of technologies are available for the production of thin films. They employ very different physical-chemical mechanisms and result in very different outputs. Physical Vapor Deposition (PVD), Chemical Vapor Deposition (CVD), Atomic Layer Deposition (ALD), Molecular Beam Epitaxy (MBE), Sol Gel methods, and Electroless Deposition are some of the processes used to make coatings from oxides, nitrides, carbides, metals, and complex compounds with thicknesses varying from a few nanometers up to several micrometers.

The microscale structure (microstructure) of those coatings can be amorphous, crystalline, nanocrystalline, or columnar, and it depends strongly on the deposition process used, even for the same deposited material. For example, one of the attractive features of PVD technology is the possibility to produce coatings with different microstructures by varying the appropriate process parameters. The coating’s surface morphology reflects the microstructure of the underlying bulk material and can act as a fingerprint for each type.

The use of advanced methodologies for surface characterization of such systems is among the most important issues for the development  and optimization of innovative products.

Therefore, the analysis of coatings consists of chemical and physical characterization, including analysis of the mechanical properties and surface morphology.

Non-contact optical profilometers are multipurpose and flexible instruments for surface characterization of materials. Their principle of operation, normally based on confocal imaging and/or interferometry, enables the user to measure features of interest on a sample over many different scales of size and surface roughness, as well as types of optical characteristics.

In particular, in the field of thin films, a single instrument combining confocal microscopy and interferometric optical profilometry can be useful for a wide range of applications, not only for surface morphology, but also for film thickness measurement, analysis of failure modes, and better understanding of mechanical properties.

Examples are reported here which show the use of a confocal microscope and optical profilometer in synergy with other, completely different methods, such as indentation and scratches at the micro-/nano-scale, which are often exploited to assess the mechanical properties of thin films.

All the confocal microscope and optical profilometry results presented here were obtained with the Leica DCM 3D, a dual technology (confocal and interferometric) optical microscope/profilometer.

The deposition technologies used to manufacture the samples investigated in this report were: magnetron sputtering PVD (MS-PVD) for the production of chromium (Cr), titanium nitride (TiN) and aluminum nitride (AlN) films on polymers, steel and aluminum (Al) substrates or silicon (Si) wafers, respectively, RF (radio frequency) plasma assisted CVD (RF-PACVD) for the production of Diamond Like Carbon (DLC) coatings on silicon wafers, Cathodic Arc deposition for the production of titanium nitride coatings on M2 high speed steel substrates.

For the mechanical test involving indentations or scratches, the instruments employed were: a Mitutoyo HM-124 micro hardness tester with Vickers tip and a Agilent G200 Nanoindenter (XP head) with Berkovitch tip for indentations, as well as a CSM Revetest Xpress for scratch testing with an 800 µm radius Rockwell tip.

All the measurements were made according to standards ASTM E 384 [1] and ISO 6507-1 [2] for microhardness tests, ISO 14577 [3] for nanoindentations, CEN/TS 1071-3 [4] for scratch tests, and CEN/TS 1071-11 [5] for residual stress evaluation. 

Some measurements made with the Leica DCM 3D were also checked with an FEI Helios Nanolab 600 FIB/SEM microscope which was used to image a coating’s cross section and nanomill the surface.

For all roughness measurements reported here, the Leica DCM 3D was used, operated in the confocal microscope mode using a 20x–50x objective lens. For surface roughness measurement, the parameter, R_a, essentially the average height of the peaks and average depth of the valleys on the surface, which is calculated with the equation:

was used, where the function Z refers to the height of the surface at a particular point along the contour denoted by x and l_r is the total length of the contour along the direction of x.

A method of deriving R_a is as follows:

• a mean line denoted X–X is fitted to the measurement data;
• the portions of the profile within the sampling length, l, and below the mean line are then inverted and placed above the line; and
• R_a is the mean height of the profile above the original mean line.

This derivation method for R_a is illustrated in Figure 1 below.

Controlling the substrate roughness

The surface roughness of the substrate has a strong influence on the properties of the deposited film, in particular its adhesion, microstructure and final topography. As demonstrated in the following images, a chromium (Cr) coating deposited onto a polymer substrate by magnetron sputtering (MS-PVD) can appear shiny and reflective like a mirror (Figure 2a) or opaque gray-blue (Figures 2b). In Figures 2c–f, it is possible to see how the microscope/profilometer measured accurately the surface roughness of the different polymer substrates employed in the preparation of the samples. If a specular reflective surface is desired, the roughness parameter, R_a, must be less than 100 nm. Careful analysis with the microscope/profiler (Leica DCM 3D) made it possible to check surface roughness in every step of the sample production, ensuring predictability of results [6].

Another important variable to take into account when producing thin films is the deposition rate. This rate is most easily estimated by dividing the final film thickness by the total deposition time needed to produce it (the average deposition rate). As can be seen in Figure 3a–b, if the film thickness is large enough for optical detection, it is easy to measure the height difference across the border between the masked and unmasked areas after film deposition. For this case, a 1.25 µm thick DLC (diamond-like carbon) film was deposited onto a silicon substrate with RF-PA-CVD [7].

Fig. 3: A 1.25 µm thick DLC (diamond-like carbon) film deposited onto a silicon substrate with RF-PA-CVD: a) topography of the film near the edge at the boundary of the masked and unmasked areas and b) height profile across the boundary. The film height near the border is about 1.2 µm.

Microindentation is often used to assess the elasto-plastic behavior of thin films. The elastic modulus of the sample (thin film) can be obtained from mechanical models using the data of the indentation force and the size and shape of the resulting mark. For this application, the microscope/profiler is capable of making 3D topographies of the entire indentation mark, giving information about elastic recovery and pile-up of the film.

In Figure 4a–c, the 3D reconstruction of Vickers indentation marks are reported. The sample is a TiN film on stainless steel deposited by MS-PVD [8]. The topographic image gives clear information about plastic deformation occurring during indentation.

Fig. 4: 3D reconstruction of Vickers indentation marks on a TiN film deposited on stainless steel by MS-PVD: a) bright field image; b) 3D indentation topography (image size = 51 × 33 × 1.8 µm); and c) profile of the indentation along a diagonal showing a depth of 1.7 µm.

The scratch test method is often used to assess the adhesion strength between deposited films and the substrate (EN 1071-3) [4]. When a scratch occurs on a coated plane surface, depending on the material properties (brittle or ductile), there are standard failure modes where the performance of the film is mainly defined by the ultimate load at which failure happens. Even if a microscope is already integrated into the scratch testing device, it is still possible to extract much more information using a confocal microscope/profilometer which allows the user to obtain complete 3D profiles of the produced scratch. In addition, a 3D image can be obtained with all the scratch features in focus, irrespective of their position (the Color Infinite Focus feature of the Leica DCM 3D). The images in Figure 5a–b of a TiN film on M2 high speed steel show clearly how the fine details of the scratch (trench) at all depths can be seen. The image in Figure 5a was taken without Infinite Focus for comparison. It has a limited depth of field due to the high magnification. The second set of images, Figures 5c–d, show the scratch track topography. The third set of two images, Figures 5e–f, demonstrate the powerful correlative capabilities of the microscope/profiler where it can be seen with high resolution what happened at the track points, i.e. the places where the scratch instrument noticed a change in the friction coefficient.

Fig. 5: Scratch test on a TiN film deposited onto M2 high speed steel: a) bright field image of scratch end; b) image of scratch end using color infinite focus; c) image showing scratch topography; d) 3D view of scratch (637 µm x 3.1 mm x 37.8 µm); and higher resolution image showing e) point (blue) where cracking of the coating is seen and f) point (yellow) where coating delamination is seen.

As already discussed above for microindentations, the same principles can be applied successfully to nanoindentations as well. In this case, the profilometer can be used to map the cracks and spallations induced by an indentation or the blistering induced by the residual stress. It is also possible to infer the fracture toughness of the film and its adhesion onto the substrate. Some examples are shown in Figure 6a–b, images taken of an AlN film on silicon made by MS-PVD [9].

The wear test is another classical method to analyze films/coatings. A wear test is performed by putting the coated sample into contact with a counterpart substrate, applying a known load normal to their surfaces, and then putting the sample in motion relative to the counterpart substrate. After a certain time, the motion is stopped and the volume of material worn away is measured. One of the most used ones is the crater grinding test, made by putting a revolving ball in contact with a surface. A crater grind test was performed on a TiN film deposited onto a stainless steel substrate. Figures 7a–b show how the profilometer can measure with high precision the geometry of the resulting craters and the amount of material worn away. The traditional way to estimate this volume with conventional microscopy, measuring only the radius of the crater, can lead to inaccurate data, especially in the case of craters with non-circular shapes [10].

Fig. 7: Crater grinding wear test on a TiN film on stainless steel: a) topographic image of crater produced after wear test and b) profile of the crater where the crater depth is 9 µm.

A quantity of interest for thin film producers is the residual stress that remains after film deposition. This stress affects significantly both the "apparent" fracture toughness and the adhesion of the film to the substrate. One of the standard ways (CEN/TS 1071-11) [5] to measure residual stress is to coat a thin (<1 mm) silicon wafer and let it bend under the force applied by the coating. The substrate curvature which is measured after deposition can be directly related to the residual stress by knowing the elastic properties of the material and using Stoney’s equation:

where the subscripts s and f refer to substrate and film variables, E_s is the elastic modulus for the substrate, h is the thickness, R is the radius of curvature, and ν is Poisson’s ratio.

The equation above is accurate only for isotropic materials. A more complex equation has to be used if anisotropic crystals, whose mechanical properties vary between the different crystallographic axes, are considered, for example silicon wafers which have the crystal orientation Si(100).

In Figures 8a–b, it is possible to see how a 1 µm thick DLC coating can cause the underlying silicon wafer to bend. The degree of bending is clearly measurable with the confocal microscope/profiler. After measuring the radius of curvature due to the bending, the value can be used in Stoney’s equation to calculate the stress with a high degree of precision [11].

Fig. 8: 1 µm thick DLC coating which induces bending of an underlying thin Si wafer: a) raw data measurement showing the profile of the bent wafer (maximum height difference = 100 µm) and b) leveled data (subtraction of the non-planarity of the specimen holder) showing the profile of the bent wafer (maximum height difference = 16 µm).

The work reported here demonstrates how advantageous a 3D confocal microscope/optical profiler can be in studying mechanical behavior when exploited for the characterization of thin films, especially when coupled with other 2D techniques (SEM, FIB) or mechanical tests. Applications include characterization of film surface morphology, analysis of film defects, 3D reconstruction of microscale and nanoscale indentation marks to determine film hardness and elasticity, investigation of the modes of film failure and quantification of film adhesion to the substrate from the data of scratch tests and residual stress analysis, and the measure of material loss resulting from wear tests. Optimal results are normally achieved when using both high definition confocal microscopy and interferometric optical profilometry. Both of these modes are provided by the Leica DCM 3D.

The authors would like to express their gratitude to James DeRose of Leica Microsystems for review and proof-reading of the manuscript.

1. ASTM E384-11e1: Standard Test Method for Knoop and Vickers Hardness of Materials. Publisher: ASTM (formerly known as the American Society for Testing and Materials).
2. ISO 6507-1:2005: Metallic materials – Vickers hardness test – Part 1: Test method. Publisher: ISO (International Organization for Standardization).
3. ISO 14577 Part 1 to 4 (ISO 14577-1:2002, ISO 14577-2:2002, ISO 14577-3:2002, ISO 14577-4:2007): Metallic materials – Instrumented indentation test for hardness and materials parameters – Part 1 to 4. Publisher: ISO (International Organization for Standardization).
4. EN 1071-3:2005: Advanced technical ceramics – Methods of test for ceramic coatings – Part 3: Determination of adhesion and other mechanical failure modes by a scratch test. Publisher: CEN (European Committee for Standardization).
5. CEN/TS 1071-11:2005: Advanced technical ceramics – Methods of test for ceramic coatings – Part 11: Determination of internal stress by the Stoney formula. Publisher: CEN (European Committee for Standardization).
6. Activities funded by the project FIT-REACH (Italy).
7. Activities funded by the project ITALNANONET (Italy).
8. Activities funded by industry contracts with STM, University Rome 3.
9. Activities funded by the joint project NAMESUR (Italy – Argentina).
10. Activities funded by the project FIT-REACH (Italy).
11. Activities funded by the project ITALNANONET (Italy).