Cyclic impact tests are used as model tests for assessing coating durability under dynamic loading
[4][5][6]. Bulk materials and coatings systems often undergo fatigue deformation mechanisms in the multi-cycle tests that are not observed in single-cycle tests
[7][8]. In an impact test on a coated system, the test severity and positions of peak impact-induced stresses relative to the coating–substrate interface can be controlled by varying the impact energy and the geometry of the test probe. Fatigue mechanisms can vary with the ratio of coating thickness
t to the indenter radius
R [9][10] (
t/
R), so it can be very useful to perform impact tests with different contact sizes to obtain data over a range of
t/
R. Therefore, macro-scale, micro-scale and nano-scale impact tests have been developed. The differences between them and how these influence the observed behaviour is discussed in more detail in later sections. Deformation and failure mechanisms depend on applied load and indenter sharpness.
t/
R values are very low (≈0.001) in macro-scale tests of thin physical vapour deposition (PVD) coatings with cemented carbide or hardened steel spherical indenters with 1–3 mm end radius. The peak von Mises stresses that result in plastic deformation are located deep into the substrate, and hence, the fatigue behaviour can be strongly influenced by the substrate properties
[1][4][5][6][7][11]. Although they can be useful, macro-scale impact tests have some limitations. An alternative approach to determining coating fatigue resistance is to perform nano-scale impact tests at higher
t/
R with much sharper probes. These accelerated tests are of much shorter duration than macro-scale tests and subject coatings to more severe conditions that replicate the high stresses generated in actual operating conditions. The position of peak von Mises stresses relative to the coating-substrate interface is completely different in the nano- and macro-scale tests. Nano-impact tests are very sensitive to small differences in coating properties and have shown excellent correlation to coating performance in applications. In particular, there have been many studies on Al-rich (Ti,Al)N-based PVD coatings on cemented carbide that have shown strong correlation between the wear of coated tools in high-speed machining applications and the fracture resistance found in the nano-impact test
[12][13][14][15][16][17][18][19][20][21][22][23].
2. Nano-Impact—Experimental Setup, Test Basics and Test Metrics
In the nano-impact test, a diamond indenter is withdrawn to a set distance from the sample surface and then rapidly accelerated to produce a high strain rate impact event. The depth-sensing capability of a commercial nanoindentation system (NanoTest, Micro Materials Ltd., Wrexham, UK) is used to monitor the degradation of surface from repeated localised stresses at high rates of strain, which are orders of magnitude higher than in normal (quasi-static) nanoindentation. The configuration is shown schematically in
Figure 1a.
Figure 1. (
a) Schematic representation of experimental configuration for repetitive nano-impact test. (
b) Example of instantaneous probe position in a repetitive micro-impact test.
An initial surface contact by the impact probe under a minimum contact load determines the depth zero at the beginning of the nano-impact test experiment. The actuated (static) coil force is then applied, producing elastoplastic deformation by indentation. The corresponding initial indentation depth under load,
h0, which includes elastic and plastic deformation, is used to confirm that the depth zero is measured correctly and the test did not impact in an anomalous region of the surface. Repetitive contacts are produced by electromagnetic actuation where the impact probe is rapidly withdrawn from the surface (e.g., to 10 μm above the surface, as shown schematically on the left-hand side of
Figure 1b) and then accelerated over this distance to impact the surface (right-hand side of
Figure 1b), producing true high strain rate impact events (see also
Figure 2) where the probe leaves the surface between each subsequent impact. The under-load impact depth,
h, is always larger than
h0, as the dynamic impact force is significantly larger than the static impact load, due to inertia. Once the probe comes to rest, it is retracted, and with periodic actuation, the surface re-impacted at the same position at a set frequency, typically at 4 s intervals, to produce a cyclic impact test. The position of the impacting diamond probe under load is recorded throughout the test, allowing the progression of damage to be monitored cycle by cycle. An example is shown in
Figure 1b.
Figure 2. Strain rate in a nano-impact test on alumina with a cube corner indenter at 5 mN.
Strain rate in a nano-impact test on alumina with a cube corner indenter at 5 mN.
Several
resea
uthorchers have shown that the strain rate at contact in the nano-impact test can be extremely high, typically in the region of 10
4–10
5 s
−1 [26][27][28][39][40]. To illustrate this,
Figure 2 shows how the strain rate varies with time during a single impact event on a bulk alumina sample when impacted by a cube corner diamond probe (three repeats are shown). Although the strain rate reduces after contact, it remains at a high level throughout the majority of the impact event.
Experimental parameters such as the test probe geometry, applied load, acceleration distance and the total number of impact cycles and their frequency are user-controlled in the nano-impact test to alter the severity of the test and its duration. A cube corner diamond indenter (with a small end radius of ≈50 nm) has been the most popular choice of impact probe, as its geometry produces high contact strain, which is beneficial in driving impact-induced fracture within a short test time. The applied load and accelerating distance control the impact energy delivered to the sample. Typical nano-impact test parameters that have been used for testing hard coatings are: (i) cube corner diamond impact probe, (ii) 90° impact angle, (iii) 25–150 mN applied load, (iv) 15 μm accelerating distance, (v) 0.25 Hz impact frequency, (vi) 300 s test duration (i.e., 75 impacts in total), (vii) 5–10 repeat tests at each load, and (viii) normal laboratory temperature (e.g., 22 °C). The general procedure for micro-impact tests is the same.
Qin and co-workers
[29] have split the impact process into three stages: (i) acceleration, (ii) indentation, and (iii) rebound. High-resolution analysis of probe depth vs. time data is used to determine a range of metrics from single impacts including (i) coefficient of restitution (
Vin/
Vout) and (ii) from knowledge of the effective mass of the pendulum, fractional potential energy (
KEin/
KEout)
[41]. Analysis of single nano-impacts
[29] provided an estimate of the fraction of the impact energy transferred to the sample as ≈0.7 and the fraction lost through losses to the system, i.e., transmission into the pendulum, vibration, and air damping as ≈0.3.
The response of a material to repetitive contact in the nano-impact test depends on its ductility or brittleness. On a ductile material, there is a gradual increase in probe penetration depth. The rate of depth increase slows with continued impacts, particularly for strongly work-hardening materials. In contrast, for a brittle material, there are often several abrupt increases in probe depth during the test due to cohesive and/or interfacial failures. A typical example on a coating system is shown in
Figure 3. At 50 mN, there was no clear failure, but at higher loads, the abrupt increases due to fracture are clear.
Figure 3. Variation in probe depth with number of impacts in nano-impact tests at 50–150 mN on Ti
0.25Al
0.65Cr
0.1N PVD coating on cemented carbide with a sharp cube corner indenter.
Since it is trivial to set up multiple tests in an automated schedule, typically, multiple replicate tests are performed at different positions (e.g., in a grid array) on the sample surface to improve the statistical significance of the results. The impact resistance of different coatings can be assessed by the number of impacts required for failure to occur in 50% of the tests. Rebound impacts are essentially elastic
[42] so that only the initial impact in each cycle is counted. Failure probability can be estimated by ranking the number of impacts-to-failure events in order of increasing fatigue resistance and then assigning a probability of failure to the
nth ranked failure event in a total sample size of
N, according to Equation (1), in an analogous approach to the treatment of distributions of failure stresses in Weibull statistics.
By combining failure probability data at different loads, a plot of the number of impacts required for failure to occur in 50% of tests vs. the impact force can be obtained. Failure in the nano-impact test can be strongly load-dependent. As an example,
Figure 4 illustrates how the failure probability changes with load and number of impacts for an 80 nm ta-C coating on Si when impacted by a spherical indenter with a 4.6 μm end radius
[43].
Figure 4. Failure probability vs. load for 80 nm ta-C on Si.
Failure probability vs. load for 80 nm ta-C on Si.
Several methods have been used to confirm that the abrupt changes in depth (as shown in
Figure 3) are due to the onset of fracture. Jennett and Nunn
[41] have used high-resolution analysis to monitor the change in fractional potential energy absorbed with continued impacting, showing a marked increase in energy absorbed for the impacts that resulted in abrupt increases in probe depth. In nano-impact tests on the bulk ceramic materials, alumina and partially stabilised zirconia, simultaneous acoustic emission (AE) detection has been used to reveal which impacts cause cracking
[44][45]. Although there was a correlation between impacts that caused a large increase in depth being accompanied by bursts of AE, the in situ monitoring of AE revealed a more complex behaviour with crack systems developing over several impacts before a larger burst of AE for the impact resulting in material removal under the impact probe. Shi and co-workers
[46] used high-resolution data acquisition of single nano-impacts on CrN to reveal changes in depth–time data when fracture occurred. In cyclic impact, more stochastic behaviour is observed from the onset of cracking. An indirect but practically useful indication of fracture is the onset of variability in depth vs. number of impacts in repeat tests
[43]. When there was no fracture in any of the tests, the reproducibility in probe depth was typically very good. Higher variability begins once fracture occurs after a certain number of impacts in some tests but not others.
3. Comparison between Nano-, Micro- and Macro-Scale Impact Tests
The general procedure for nano- and micro-impact tests is the same. In micro-impact tests on hard coatings, the experimental parameters are typically the same except for accelerating distance, applied load and probe geometry. The accelerating distance is typically set at 40 µm so that differences in impact energy are obtained by altering the applied load (0.5–5 N). Sphero-conical diamond test probes with end radii of 8–100 μm have been most commonly used
[36][47][48]. The impact energy is given by the product of the impulse force and accelerating distance. Since the accelerating distance is typically kept constant, it is common to report data in terms of the actuated impact force. Due to higher forces and accelerating distances, the energy supplied in micro-impact is typically from ×100 to ×1000 greater than in the nano-impact test, which enables spherical probes to be used effectively, causing fracture rapidly.
Typical experimental parameters in each type of test are summarised in
Table 1. Although the principles behind nano-/micro- and macro-scale impact tests are common, there is a fundamental difference of approach in the nano-/micro- tests, which are depth-sensing, as the change in depth under load is monitored throughout the test with a capacitive sensor, and the macro-impact tests, which are not depth-sensing. Instead, coating durability in macro-scale impact tests has been assessed by post-test evaluations of damage such as crater volume
[49] or failed-area ratio (defined as area of substrate exposure divided by the total contact area)
[4][50]. Nano- and micro-impact tests are accelerated tests that are typically much shorter and probe coating system behaviour under more severe conditions where there is greater coating strain. Detailed information on the fatigue failure mechanisms in nano/micro-impact tests is obtained through setting up automated arrays of rapid repeat tests at different loads (e.g., 5–10 repeats per load) with cycle-by-cycle monitoring of the damage providing a precise measure of the number of impact cycles to coating failure in each test.
Table 1.
Comparison between nano-, micro- and macro-scale impact test techniques (1).