Measuring Mechanical Properties Up To 800 °C with High Temperature Ultra Nanoindentation Tester

In recent years, there has been much debate about the best way to perform elevated temperature nanoindentation tests, with thermal drift, tip erosion and noisefloor taking pride of place in the list of issues which can obstruct such experiments. Anton Paar TriTec has been at the forefront of this development work, which has resulted in the launch of its High Temperature Ultra Nanoindentation Tester (UNHT3 HTV).

Earlier work [1-2] has demonstrated that, besides oxidation, thermal drift is one of the key issues that will cause error in elevated temperature tests, with drift rates tending to increase with an increase in temperature. Solving this issue in the UNHT3 HTV has taken an important development and has needed many modifications to account for all possible variables.

Instrumentation

Based on Anton Paar’s longstanding experience in nanoindentation, the heart of the UNHT3 HTV is based on the greatly successful and patented Ultra Nanoindentation Tester (UNHT3) which has already established itself as an instrument with unparalleled measurement stability [3-4].

Schematic representation of the UNHT3 HTV system.

Figure 1. Schematic representation of the UNHT3 HTV system.

Its measurement head has been optimized in order to obtain high temperature operation and incorporated with a patent-pending sample stage which enables measurements to be carried out at any temperature within the operating range with the highest possible thermal stability.

As shown schematically in Figure 1, the measuring head, optical video microscope and sample stage are mounted in a high vacuum chamber capable of being pumped down to 10-7 mbar using a turbomolecular secondary pump and a primary pump. The two key advantages of operating within a vacuum are:

(i) Removal of the effects of oxidation, meaning that it is possible to test sample materials without their surface mechanical properties becoming modified by the development of oxides. In addition, it is also possible to use indenter materials which would otherwise be inadapted to an oxidizing environment: for instance, diamond is the indenter material of choice at room temperature but it oxidizes above ~400 °C, and then softens and becomes easily blunted thus making it virtually useless for nanoindentation.

(ii) Limiting heat losses by convection within the chamber, thus significantly aiding thermal stabilization.

The main disadvantage of vacuum operation is that the working of the valves and pumps will introduce extra mechanical noise into the measurements and hence, specific actions have been taken in order to minimize this as best possible, including the following:

(a) Materials choice: the frame’s internal construction has been optimized by using a blend of aluminum, cast iron and stainless steel which allows optimum mechanical damping.

(b) Incorporation of a vacuum buffer linked between the backing valve and the secondary pump to allow operation over many hours without the requirement for the primary pump. This can maintain a 10-6 mbar vacuum for over 10 hours.

(c) Use of a 5-axis magnetically-levitated turbomolecular pump with low friction bearings to minimize mechanical vibration.

(d) Anti vibration: the whole vacuum chamber is mounted on a 4-point anti-vibration table which uses active compressed air to “float” the chamber, thus eliminating the majority of vibrational noise.

(e) Stiffening the springs of the UNHT3 HTV measuring head providing a spring constant of 6 Nmm-1 (as compared to 3 Nmm-1 with a standard UNHT3) thus maintaining an acceptable noisefloor and compensating for the extra mass of the indenter and reference.

The UNHT3 HTV measuring head.

Figure 2. The UNHT3 HTV measuring head.

The UNHT3 HTV measuring head, shown in Figure 2, is maintained at low temperature with the help of a turbulence-free water cooling circuit and heat shields. Based on the UNHT3, the vital parts of the measurement system (e.g., Zerodur elements) are thus maintained within the temperature range where they are considered to be the most stable (typically < 50 °C).

The heating system is made up of three Infrared (IR) heating elements which are used to heat the indenter, the sample and the reference. An IR bath is thus established around the test area of the sample, as shown in Figure 3. The temperature of the three elements is separately monitored and controlled using three thermocouples, two embedded in the reference tips and indenter respectively, and the third placed against the back-face of the sample support disk. A fourth thermocouple can be landed onto the sample surface prior to the indenter and reference approach in order to monitor the top-face of the sample. All thermocouples (type N) are optimized for a extremely fast response time.

Principle of the Infra-Red (IR) bath.

Figure 3. Principle of the Infra-Red (IR) bath.

The temperature gradient between the sample surface and the indenter tip is therefore minimized, lowering its effect on the thermal drift. Due to the low mass of the indenter and reference tips, the target temperature can be achieved in tens of seconds, although the standard tip heating rate is typically 50 °C/min and can be defined by the user. The UNHT3 HTV head is covered by two pending patents [6-7] which relate to the manner in which heat is managed and the novel design of the indenter shafts. Figure 4(a) shows the design of the reference/ indenter. The indenter tip is held in a molybdenum holder which also comprises of an embedded thermocouple. Molybdenum was selected as it is a good conductor and also a high absorber of IR radiation. The heat produced around the tip is then blocked by a ceramic (macor) heat sink and a heat dissipater, thus reducing transmission of heat into the measuring head.

Detail of the indenter design (a) to minimize heat dissipation into the UNHT3 HTV head and (b) an FEM simulation proving that heat dissipation is extremely low.

Figure 4. Detail of the indenter design (a) to minimize heat dissipation into the UNHT3 HTV head and (b) an FEM simulation proving that heat dissipation is extremely low.

The sample heating stage (patent pending [8]) is shown in Figure 5 and includes several innovative design features, including:

i. Easy sample changing

ii. Turbulence-free water cooling circuit

iii. Large sample diameter (up to 50 mm)

iv. Optimized infrared heater for rapid sample heating (typically 50 °C/min)

v. Novel clamping system to rigidly hold the sample over the entire temperature range

High temperature heating stage with cooling block, graphite sample support disk and retaining ring.

Figure 5. High temperature heating stage with cooling block, graphite sample support disk and retaining ring.

It is essential to note that the objective of most high temperature nanoindentation tests is to test the sample surface at a stable, measurable temperature. Trials have highlighted that mounting a thermocouple on the sample surface (usually placed between the indenter and reference) can become complicated as the thermocouple is within the IR bath and may not provide accurate readings owing to the irradiation. Practice has established the fact that the surface thermocouple is not in fact necessary because the IR bath is maintained by the reference and heated indenter, as well as the sample heater underneath. The sample surface will be at the same temperature as that of the indenter and reference, when the system reaches stability at a given temperature, as they are both in direct contact and their respective thermocouples are extremely close to the surface (thermocouples are mounted at a distance of ~500 mm from the apex of a typical Berkovich indenter).

The principle of the IR bath can be understood in a better manner by considering a concrete example, as shown in Figure 6 for a test at 800 °C on a molybdenum sample, where temperature stability has been reached in a vacuum of 4.6 x 10-6 mbar. In this particular case, the indenter is at 815.6 °C, the backside of the sample is at 678.9 °C and the reference is at 810.4 °C. The indenter and reference do not show the same temperature because their areas of contact are extremely different: In this case, the indenter is a Berkovich diamond (pyramidal) whereas the reference is a tungsten carbide sphere of diameter 1.5 mm, so their heat dissipation will differ. Finite Element Modelling (FEM) simulations (as shown in Figure 4) have established that the IR bath encapsulates the whole lower portion of the molybdenum tip holder (Figure 4(a)) meaning that a stable thermocouple reading 810 °C means that the actual sample surface is also at 810 °C.

Heat management software window showing the value of each thermocouple when stability has been reached at 810 °C on the surface of a molybdenum sample.

Figure 6. Heat management software window showing the value of each thermocouple when stability has been reached at 810 °C on the surface of a molybdenum sample.

System Validation

Validation of the UNHT3 HTV system has been executed on a wide range of sample materials over the complete temperature range of the instrument, in order to understand the typical power requirements for stable heating and the limits of stability. The example in Figure 7 shows typical stability tests carried out on a High Conductivity Oxygen Free (HCOF) copper sample which was selected for its high conductivity in order to analyze the “worst case scenario” of high heat dissipation. Experience has proved that the higher the conductivity of the material, the higher the potential for thermal drift and it becomes more difficult to stabilize the temperature. This example demonstrates that indents made with 100 mN maximum load using a diamond Berkovich indenter with a 60s pause at 10% of the maximum load in order to monitor the average drift rate from the penetration depth data. This same test was repeated at temperatures from Room Temperature (RT) up to 600 °C in 100 °C increments, in every single case the average drift rate being measured as < 1 nm/minute (~17 pms-1) over a hold period of 60 seconds.

Thermal drift measurements (raw data) on High Conductivity Oxygen Free (HCOF) copper showing < 1 nm/min. average drift rate at every temperature.

Figure 7. Thermal drift measurements (raw data) on High Conductivity Oxygen Free (HCOF) copper showing < 1 nm/min. average drift rate at every temperature.

Owing to its tendency to soften dramatically as its melting temperature (~1085 °C) is approached, the 600 °C maximum test temperature was not exceeded on copper. Such softening leads to instabilities in the material which can be wrongly interpreted as thermal drift events. For this reason, various other higher temperature conducting materials are usually chosen for validation of drift rate at temperatures > 600 °C, e.g., Molybdenum, Chromium, Vanadium, etc.

A similar test on Molybdenum (melting temperature ~2623 °C) at 810 °C is shown in Figure 8, where the pause at 10% maximum load has been extended to 300 seconds (5 minutes) and over this period the average drift rate was measured as 1.25 nm/minute. This establishes the fact that once the power values of the indenter, sample heaters and reference have been optimized and a stable IR bath maintained, the stability of the system is exceptional and thus it is possible to envisage far longer term creep tests without reservation, even on highly conductive materials. It was also observed during successive tests at an elevated temperature that can contact drift (drift occurring due to heat flow between sample at contact and the indenter) was practically nil because the IR bath encapsulates the sample and indenter even when they are separated for a few minutes between tests. This is a vital feature of the system as it permits high thermal stability to be maintained even during matrices of hundreds of indentations, useful in cases where mapping of surface properties may be of interest.

Thermal drift measurements (raw data) on pure Molybdenum at 810 °C showing < 2 nm/min. average drift rate over a 300 s pause at 10% of maximum applied load.

Figure 8. Thermal drift measurements (raw data) on pure Molybdenum at 810 °C showing < 2 nm/min. average drift rate over a 300 s pause at 10% of maximum applied load.

Another valuable factor refers to the impact of indentation on the vacuum level in the chamber. In a typical experiment, the chamber vacuum pressure might be stabilized to 10-4 – 10-6 mbar based on the level of outgassing from the sample material and how much time is available for pumping down. After establishing the IR bath and stabilizing the system, the pressure of the indenter as it indents the sample surface may cause major outgassing which manifests itself as a momentary reduction in vacuum pressure. This is frequently due to the material being compressed, causing the removal of previously trapped sub-surface species. In extreme cases, this may momentarily drop the vacuum level by an order of magnitude (observed in porous industrial ceramic coatings, for example) and hence might influence the capability of the vacuum to protect the surface. It is thus important to pump adequate vacuum to minimize such effects, mainly when testing highly reactive materials, e.g. tungsten.

Results

Some typical load vs. depth curves at 23 °C and at 810 °C for the Molybdenum sample are shown in Figure 9 together with optical images of the indents at the two temperatures. It should be noted that the room temperature indentation was carried out after the 810 °C test, explaining the fact that the sample had cooled down and had thus been “heat cycled”.

Optical images of indents on pure Molybdenum at (a) 23 °C and (b) 810 °C using a diamond indenter and vacuum 4.6 x 10-6 mBar. Corresponding load-depth data is shown in (c).

Figure 9. Optical images of indents on pure Molybdenum at (a) 23 °C and (b) 810 °C using a diamond indenter and vacuum 4.6 x 10-6 mBar. Corresponding load-depth data is shown in (c).

It is interesting to observe the relative differences in creep at maximum load during the 300 second hold period, where creep at 810 °C totaled 777 nm and creep at 23 °C totaled 122 nm due to the obvious softening effect at that elevated temperature.

Load-depth data on High Conductivity Oxygen Free (HCOF) copper corresponding to the measurements shown in Figure 7. Hold period was 5 seconds at 100 mN maximum load and 60 seconds at 5 mN prior to unloading.

Figure 10. Load-depth data on High Conductivity Oxygen Free (HCOF) copper corresponding to the measurements shown in Figure 7. Hold period was 5 seconds at 100 mN maximum load and 60 seconds at 5 mN prior to unloading.

A similar effect was observed in HCOF Copper as shown in Figure 10 for measurements from room temperature (23 °C) up to 600 °C. The softening effect is visible while loading at the highest temperatures, manifesting itself as irregularities along the loading curve.

Acknowledgements

The Authors would like to thank Johannes Michler and Guarav Mohanty at EMPA, Switzerland and Jeff Wheeler at ETHZ, Switzerland for ongoing collaboration in the development of the UNHT3 HTV and many useful discussions.

References

1. C. A. Schuh, C. E. Packard and A. C. Lund, J. Mater. Res., 21 (2006) 725

2. J. C. Trenkle, C. E. Packard and C. A. Schuh, Hot nanoindentation in inert environments, Review of Scientific Instruments 81.7 (2010) 073901

3. J. Nohava, N. X. Randall and N. Conte, Novel ultra nanoindentation method with extremely low thermal drift: principle and experimental results, J. Mater. Res., Vol. 24, No. 3 (March 2009) 873-882

4. US Patent 7,685,868 B2, EP 1828744

5. J. M. Wheeler and J. Michler, Indenter materials for high temperature nanoindentation, Review of Scientific Instruments 84 (2013) 101301

6. European Patent EP14191443.2 Heating arrangement for a material testing device

7. European Patent EP14191442.4 Surface measurement probe

8. European Patent EP16151845.1 Sample holder arrangement

9. J. M. Wheeler, D. E. J. Armstrong, W. Heinz and R. Schwaiger, High temperature nanoindentation: the state of the art and future challenges, Current Opinion in Solid State and Materials Science, 19, 6 (2015) 354-366

This information has been sourced, reviewed and adapted from materials provided by Anton Paar TriTec SA.

For more information on this source, please visit Anton Paar TriTec SA.

Ask A Question

Do you have a question you'd like to ask regarding this article?

Leave your feedback
Submit