Coarse Design

2.4.1 Design Components

Following the previously described design elements, the heating stage is designed rotationally symmetric around the mutual centreline of the sample and the supporting 3C-SiC disk, see Figure 3a.

The sample assembly is placed inside a nest of horizontal springs (k1 in Figure 3) that form a thermally centred support by axisymmetrically exerting a radial, compressive force on the sample and SiC disk.

Simultaneously, three springs exert a vertical force on the SiC from behind (k2 in Figure 3). The three legs of the nest of springs have a V-groove at the bottom that rests on top of three recessed barrels that are part of the base plate., see Figure 3. This is done in order to accommodate for the radial thermal expansion of the nest of springs, while keeping the thermal centre in position, as shown in Figure 3b andFigure 4a. Both the nest of springs and the base plate are manufactured out of tantalum, see Figure 4b, due to its low thermal conductivity and high melting temperature [9].

The vertical force exerted by the springs, k2, is countered by a hood with three pawls that is placed over the nest of springs and is attached to the base plate by fixing each pawl in a groove with alumina bolts. With respect to of all the heating stage components, the hood has the largest contact area with the sample, and is thus made out of aluminium-oxide (Al2O3) for its low thermal conductivity [9]. The working of the entire assembly is depicted in Figure 3c in the form of a free body diagram, where the depicted situation occurs three times rotated around the central axis of symmetry. Since this assembly makes use of the thickness of the sample to fixate the sample, this

Page 14 means that the sample should always be 1mm thick. However, as the possibility of thinner samples is also interesting, a tantalum ring may be placed below the SiC-block. With such a ring, the laser light can still reach the SiC through the orifice of the tantalum ring and the positioning of the sample continues to work as intended. In order for the laser to be positioned close enough to the SiC-disk for it to shine all of its light on it, the base plate has a concave shape, as can be seen in Figure 3c and Figure 4b.

Figure 3: (a)The nest of springs holder, shown in grey, around the sample and 3C-SiC block, shown in turquoise and black respectively (b) cross-section of the 3D-model of the assembly with the holder on the base plate and covered with the alumina hood, where the sample is depicted in turquoise and the SiC disk in black. The labels correspond to those used in

the free body diagram of (c). (c) Free body diagram of the cross section of the heating assembly along one of three rotational symmetric plane (c),

Figure 4: (a) Nest of springs, with the specimen shown in turquoise, on the base plate, (b) bottom view of the copper cooling heatshield, shown in brown, with the baseplate assembly covered by the hood, shown in white, (c)heatshield with

two tapered conical openings for the electrons to pass through.

Figure 5: Front (a) and rear (b) view of the heatshield together with the heating assembly mounted on the steel frame. On the right, the ferrule clamp(black) and ferrule can be seen.

In Figure 4a, the nest of springs and base plate are shown without the hood. Since the centre assembly of the heating stage, consisting of sample holder, hood and base plate, becomes very hot during operation of the stage, this assembly must be isolated from the rest of the stage. To this end,

Page 15 the base plate is connected to the rest of the stage via three thin legs that are part of the base plate.

Due to their limited cross-sectional area, they forms thermal resistors. In addition, they are connected to the actively cooled heatshield, made of copper and shown as the brown cover in Figure 4. Therefore, the heat dissipating from the sample through the legs cannot reach the rest of the stage or the SEM as the cooling circuit acts as a barrier between them and the stage. The heatshield is actively cooled by a coil that is attached to the back of the heatshield and through which water flows. Besides shielding the sides of the heated assembly, this coil also extracts the dissipated energy from the heating stage. For it to cool the heating stage and shield the heating assembly as well as possible, the heatshield and coil are created from one piece.

Shown in Figure 4c, are the conical openings in the heatshield for the electrons to pass through. The two openings in the heatshield are tapered so that the outgoing electron cone isn’t hindered, and the stage sample tilt can be varied from 69 to 75 degrees. In previous high-temperature EBSD methods, heatshields were typically designed to be open over the entire surface of a sample [16-26], therefore, the benefit of this heatshield has been validated while taking into account a limited manufacturing precision, see in Appendix D. It was found that this heatshield with tapered conical openings yields no significant decrease in data quality while it drastically decreases the amount of exiting radiation.

With respect to a heatshield that is open over the entire surface of the sample, this novel heatshield reduces the emitted thermal radiation by 35%, for which the validation is found in Appendix D. Heat Shielding Design

In order to connect the heatshield and base plate to the moveable stage of the SEM, the heatshield is attached to a steel frame as can be seen in Figure 5. In order to guide the laser to the sample, a ferrule clamp is attached to the frame so the termination of the optical fibre is not in direct contact with the hot components. This is important because the optical fibre termination of commercially available lasers are typically made of rubber.

2.4.2 Thermal Simulations

In order to verify whether the design would yield a heating stage capable of meeting the set requirements, thermal simulations have been conducted. To accurately model the heating stage, various assumptions and simplifications have been made. First of all, the laser is modelled to be a simple heat flux incident on the backside of the SiC-disk for which the chosen value equals the energy put in to the system. As the experiments are conducted in vacuum, convection to the environment is neglected and black body radiation is assumed. For this radiation, all surfaces that can ‘see’ each other are coupled via radiation. The only location convection occurs is in the cooling circuit of the heatshield for which the convective heat transfer coefficient is chosen to be 500W/m²K [32]. All components that are in contact with each other are modelled to be in perfect contact. Also, the small bolts that hold various parts in place are neglected. Finally, the assumption is made that the heating stage does not thermally interact with the SEM as the model is refined to the point that the base of the heating stage frame remains at approximately the initial temperature. These choices are further motivated in Appendix E. Thermal simulations

From the simulations, of which the results are shown in Figure 6, it is found that the sample temperature reaches a maximum of 987°C in steady state at 100W laser power. This is more than 500°C less than the goal temperature of 1500°C. Moreover, the inside of the cooling circuit of the heatshield becomes too hot, namely 167°C. This is not allowed to happen as the water in the cooling

Page 16 circuit will start to boil. In addition, the rise in temperature reduces the effect of the actively cooled heatshield as a thermal barrier and leads to a rise in temperature up to 126°C at the bottom of the frame.

Figure 6: The base plate assembly with hood (left) and without hood (right) for an input power of 100W to the back-side of the SiC disk. The heatshield, cooling circuit and frame are not shown as they obstruct the view of the assembly.

This creates a significant temperature difference between SEM and heating stage and therefore, the assumption with regard to the absence of thermal interaction between SEM and heating stage becomes invalid. Since the amount of radiation coming from the sample is rather constant as it is a function of component temperature, the conductive losses should be investigated as those are simpler to optimize. Finally, it is found that due to the open backside of the heating stage, the radiative heat transfer to the optical fibre delivering the light leads to a significant temperature increase in the fibre ferrule as well. Where the maximum operational temperature of the fibre is 80°C, it becomes 189°C in the simulation.

To create a useful heating stage, both the issues regarding energy dissipation from the sample and the fibre temperature must solved.