Furnace Cycle Testing....can you stand the heat?!

Furnace cyclic testing (FCT) or Thermo-cyclic furnace (TCF) testing is a performance evaluation commonly carried out on thermal barrier coatings (TBC) to validate their lifetime performance. The test may be carried out as part of coating research optimization, for coating qualification and validation as well as for regular quality/performance control.?

What is a FCT test?

Test samples are placed into a furnace at a desired test temperature for a set time period; for instance 50 minutes at 1100°C?. Once the exposure time is over then samples are removed from the furnace and cooled at a set rate.?The test samples may then be monitored for coating damage through inspection. Afterwards the samples are returned to the furnace for another exposure cycle. This process continues until the coatings show signs of spallation and failure.

The test could also be considered a cyclic oxidation or exposure test as the coating samples are not cooled during the furnace exposure so will reach an isothermal state the same as the furnace. While at the isothermal temperature, the sample will undergo oxidation of the bond coat (and substrate) as well as sintering of the TBC. The coating should be stress free at this point. At the end of the cycle, the samples are removed from the heating zone and are normally force cooled with compressed air. Cooling induces thermal stress due to thermal gradients and CTE miss-match between coating layers and the substrate. Failure for TBC samples is typically based on an accumulation of damage that occurs during each cool down (quench) cycle. Oxidation of the bond coat is the driver of failure and will determine the final lifetime. However, a more strain tolerant ceramic coating may extend the length of time the samples survive in the test.

The example below shows three TBC samples that have been tested for several hundred cycles. The comparison shows images of the samples hot (top image), just out of the furnace and after some minutes of cooling (bottom image).

You may note in the pictures that the hot samples display darker regions and lighter regions. Areas that are dark show the areas where the ceramic TBC has delaminated from the layer beneath. This material cools faster, leading to its darker appearance. This is important when monitoring the degree of spallation/delamination as the 'cold' samples show only the areas where coating has flaked off (grey areas) or cracked. The actual delamination area is far larger, as seen in the 'hot' sample image. The criteria for failure may vary, though 20% delamination by area is common. There could therefore be a big difference in the 'lifetime' depending on which state is used to check for delamination/spallation.

Why do we use FCT testing?

FCT testing is used to evaluate the relative performance of a thermal barrier coating system in a simulated environment. In a research and development process, prospective TBC coatings will be evaluated to see which will survive the longest. This may then be used as a factor for further coating development and optimization.?In a quality control or qualification situation, coatings may be required to achieve a certain average number of cycles without failure. Quality control testing of TBC's may occur at regular intervals or is required for new batches of powder or thermal spray equipment changes.?Essentially its another tool to help us gauge coating performance.

Test cycle time and temperature

The test cycle for an FCT test can vary greatly depending on the requirements of the TBC and whether oxidation or thermal cycling is of greater importance. For industrial gas turbines used in power generation, the test may require a long hold time of 24 hours due to a focus on long term oxidation performance. For aviation turbines, a short cycle time of 45-60 minutes may be more relevant for components with highly cyclic operation.

The test temperature used can be a critical factor. Oxidation is driven by kinetics so higher test temperatures will lead to more rapid oxidation.

• Higher test temperatures can shorten the test duration (shorter number of cycles).
• Low temperatures can be more representative of actual service temperatures, but may lead to extremely long times for test completion.

Compromise is usually selected whereby the test temperature is the maximum expected for the TBC in its service environment. Aviation turbines may experience higher peak temperatures (for short time frames) whereas IGT experience lower temperatures for long times. Exceeding the temperature capability of the bond coat can lead to non-representative oxidation phenomena and early failure, so care must be taken in the choice of test temperature.

As an example, an IGT component may be expected to have an operating bond coat surface temperature of 980°C (1796°F). At that temperature the coating should survive 45000 hours, or more than 5 years continuous exposure. This would be too long for the majority of evaluations so an accelerated test ?at higher temperatures is needed. A higher temperature of 1100°C is then chosen as test temperature.

Example cycles: -

• 60 minutes heating at 1100°C (2012°F) + 10 mins cooling
• 24 hours heating at 1100°C (2012°F) + 10 mins cooling
• 50 mins heating at 1121°C (2050°F) + 10 mins cooling
• 45 mins heating at 1135°C (2075°F) + 15 mins cooling

Cooling rate can also be considered an important criteria to the cycle program. More rapid cooling can introduce a greater degree of thermal shock to the test, whereas slower cooling will reduce this. Commonly there are requirements such as the sample should reach 100°C within 10 minutes of the cooling cycle.

Substrates and their geometry

The substrate used for testing can have a strong impact on the final lifetime of the coatings during testing. This comes from both the geometry of the sample and its material properties. The main concerns for substrate material properties are the chemistry of the material and its thermal expansion coefficient. Chemistry influences how the substrate itself oxidizes (it will on any uncoated surface!). It also determines how the substrate reacts with the bond coat through diffusion. Diffusion of elements between bond coat and substrate can also shorten the oxidation lifetime of the bond coat, so care should be taken here. The following is a comparison of two coating systems produced on two substrates, Hastelloy-X and Haynes 230. Samples were produced during the same deposition runs on TCF plates and testing was performed for 60 min cycles at 1100°C (2012°F).

Hastelloy-X is an older generation nickel based alloy that is often used in development work. Haynes 230 is a more modern alloy of the type used in industrial gas turbine parts.?As shown in the graph, use of a more modern alloy for the substrate gives a lifetime increase of almost 100% in this case. This can be attributed to the thermal expansion coefficient of the Haynes 230 being lower than Hastelloy-X; lowering the CTE mismatch with the TBC coating and therefore lowering stress buildup. It is important then when making comparisons between test results that there are no differences in substrate material that could influence the results.

The shape of the test samples is often dependent on the end user and their historical experience. Typical for thermal spray coatings is the use of ‘standard’ 1 inch diameter (25.4 mm) button in 3mm or 6mm thickness. Also common are ‘TCF plates’ of 50mm x 30mm x 6mm, originally developed by an industrial gas turbine OEM. In some cases rods, tubes, dummy vanes and even sectioned components are tested. An additional important factor is if the edge of the sample has a radius or chamfer as this will reduce stress at the edge of the sample/coating, preventing early failure.

Sample shape influences the stress state of the sample and coating during the test and does influence the cyclic lifetime. An example of two TBC coating produced on TCF plates and standard buttons is shown in the following example:-

In this case the same Hastelloy-X substrate material was utilized with two different geometries. It can be seen that the TCF plates give between 55% to 100% increase in measured lifetime versus standard buttons. It is thought that the lower lifetime for the 1 inch buttons is related to the greater amount of coating edge versus total coating area and the fact that the buttons do not flex due to their high stiffness. TCF plates by contrast, will bend during repeated cycling, potentially relieving some of the stress generated during oxidation and quenching from high temperature. More information on these specific coatings can be found in the open access paper:-

The Equipment

While FCT testing can be made using standard furnaces, the long duration of many tests, in the order of weeks to months, makes automation of the test process critical to successful evaluation. As such, a dedicated test machine is often needed. A highly capable example of such an equipment is from the company Entech AB in Sweden.

The latest generation of FCT furnace, the ECF 20/16-HV, has been refined over several iterations over more than a decade of providing test rigs to customers. Below is a recent example that was delivered to a major aero engine module supplier.

The test rig is designed to perform the test in a fully automated manner with minimal operator oversight. The PLC controlled furnace system allows for a fully programmable cycle including:-

• Time and temperature of exposure
• Heating ramp on furnace entry
• Cooling rate at the cooling stage position

The furnace temperature is monitored through primary and backup systems. An instrumented test sample or separate thermocouple can also be used to provide further logging of test cycle temperatures.

The large furnace plate test area has room to test up to 42 button samples with uniform heating. Samples are moved to and from the hot zone furnace automatically and are imaged with a integrated digital camera on arrival at the cooling stage after each cycle. This allows the user to track sample spallation with a record of sample condition. Further image processing can then be used to determine the failure lifetime by analyzing the images to determine the degree of spallation.

A number of these units are operating in research facilities as well as at coating providers and two OEM turbine equipment manufacturers. Notable research facilities using the equipment in their projects include University West (Sweden), Plasma Physics Institute (Czech Rep.) and Treibacher Industrie AG (Austria).

The equipment can be considered future proof in that it is capable of operating the test at temperatures in the range 1300°C?-1600°C. In this case well in excess of the limits of metallic components and in the range of operation of the next generation CMC/EBC components.?

Recommendations

While evaluating coating cyclic lifetime is a complex subject, some simple recommendations can be made to avoid basic issues:-

• Take care of substrate geometry and chemistry - always compare like with like
• Be aware of chemical reactions between the bond coat and the substrate
• Test many samples! – as with many evaluation tests for coatings, having sufficient number of samples to generate reliable data is important.
• Monitor the coatings while still hot, delaminations are more easily detected this way. when combined with image analysis of the data, this can provide highly accurate monitoring.
• An automated FCT test rig is highly recommended to manage the whole process.

I hope the information here is helpful to those starting to look at this type of test. Please get in touch if there is interest in the test equipment or test procedure in general.