Chapter 2: A retrofittable heating module for use in in-situ EBSD tensile experiments
A. Detector Screen Cooling
In previous work, it has been found that the quality of the obtained images from the EBSD detector decreased over time for elevated temperatures. This decrease originated in the heating of the detector screen due to the incident radiation from the hot sample and its surroundings [18]. When the deposited phosphor heats up, it starts to randomly emit photons, which disturbs the Kikuchi pattern imaging since less contrast is found and the patterns are less pronounced. A previously proposed solution is to perform an EBSD scan and to monitor the screen temperature parallel to the experiment. When the phosphor screen reaches a pre-specified temperature, the detector screen is retracted and allowed to cool down before continuing the experiment [23]. While this is effective, it does also hinder the continuous data acquisition and therefore there is a limitation in choice of heating sequences.
A.1 Design
To solve this issue, a cooling system is proposed. To cool the phosphor screen, active water cooling is chosen. The implementation is quite straightforward as there is already circulating water present for the cooling of the heating device. Due to the poor thermal conductivity of glass, upon which the phosphor screen is deposited in the conventional detector set-up, and difficulties regarding poor surface contact between cooler and screen, no cooling can be added to the existing system.
Therefore a system is designed where the water is passed in between two parallel glass plates. Once assembled, a phosphor layer can be deposited on the front of the cooling assembly, yielding perfect contact between phosphor layer and the cooling assembly. On the backside, a ring is mounted to the glass. This ring is placed over the end of the EBSD detector tube, fixing the phosphor screen to the detector in the same manner as for a regular phosphor screen. The assembly is shown in Figure 28.
Figure 28: Glass substrate for the cooling of the phosphor screen with a ring on the backside to connect the cooling to the detector tube, enabling similar mounting to a regular phosphor screen.
A.1.1 Diffraction
An assembly of this type brings along several challenges that must be tackled in order to ensure safety and usefulness in the experimental EBSD set-up. Where, for the conventional phosphor screen, the light created by the incident electrons only has to travel through 1mm of glass, in the new situation, it has to travel through two layers of glass and a thin layer of water. As the assembly is
Page 43 used in vacuum, the glass has to be at least 4mm thick in order to safely withstand atmospheric pressure from the water. In addition, a 2mm layer of water is added to convey heat away from the screen, which increases the amount of distance travelled through a refracting medium tenfold.
Depicted in Figure 29 below is the difference in the diameter of the exiting cone of light due to the addition of diffracting media. The difference in exiting diameter can also be thought of as the light originating from a virtual point. As the virtual origin lies further away from the back of the phosphor screen, the camera and lens must be placed closer to the detector screen.
Figure 29: Depiction of emitted light through diffracting media of different thickness. The normal phosphor screen and substrate and cooled substrate are superimposed and aligned where they enter the detector tube, here on the left. The
white points depict the virtual origins.
The distance that the lens and camera must positioned closer to the back of the phosphor screen is equal to D2 minus D1 divided by the tangent of the exit angle of the light. All three variables can be expressed as a function of the original exit angle of the light from the phosphor screen, the distance is given by:
π· = (2β2β β1) β tan(π) + βπ€β tan (sinβ1(1.5
1.33 β sin(π))) tan (sinβ1(1.5 β sin(π)))
Here, h1 is the thickness of the original substrate, h2 is the thickness of one sheet of cooled glass, hw is the thickness of the flowing water, and Ξ is the exit angle at the phosphor screen. If it is not possible to move the lens towards the detector screen, according to the lens maker's equation, the camera must be positioned closer to the lens. A combination of both abovementioned solutions is also possible.
Another factor to keep in mind is the fact that the light will be passing through moving water in the new situation as opposed to static glass for the original situation. This moving water must remain a laminar flow to let the light pass unscattered; as soon as turbulence starts occurring, the Kikuchi patterns would not be recognizable and therefore useless. In addition, the entire cooling conduit must be free of any gas that could potentially forms bubbles as this would have a similar effect on pattern quality. To verify whether the flow is turbulent or laminar, the dimensionless Reynolds number is used:
π π = π’ β π·
βπ ; π·
β= 2βπ€ (β + π€)
Here, u is the mean velocity of the water, D
his the hydraulic diameter defined by the height
h and width w of the cross-section, and Ξ½ is the kinematic viscosity of water. Given thedimensions of the previously mentioned flow channel, this results in maximum velocity of
Page 44
0.6 m/s for water of 20
Β°C. This value is determined for a Reynolds number of 1800 as this is the beginning of the transition regime from laminar to turbulent flow. While it is generally accepted that turbulent flow starts at a Reynolds number of 2300 [37], it is better to be on the safe side as any disturbance could render the Kikuchi patterns unusable.A.1.2 Water circuit
In order to keep the amount of appliances and flanges to a minimum, the choice has been made to use only one pump and only one flange to feed water into the vacuum chamber. Due to this, the detector cooling and heating device cooling are connected in series. As the glass from the detector screen is a bad thermal conductor, the water first passes through the EBSD detector screen cooling for maximum cooling to occur. The slightly warmer water, which will have had a temperature increase of several degrees, passes through the copper cooling conduits of the heating device.
Because of the magnitude of the radiation and good thermal conductivity of the copper, several degrees do not matter as much as it would for glass in the reverse order of cooling.
A.2 Validation
To validate the usefulness of the EBSD detector cooling system, several requirements must be met. If the Kikuchi patterns observed by the EBSD detector camera cannot be indexed due to the screen cooling, it can be concluded that the cooled screen isnβt of use. In addition, if the continuous water flow does not cool the screen sufficiently to keep a constant and low temperature, it can be concluded that the cooling is inadequate and with that the cooling doesnβt adequately serve its purpose. Therefore, the pattern quality should be verified first, and subsequently, the cooling should be analysed.
A.2.1 Pattern quality
To verify pattern quality, several EBSD scans should be made using the conventional detector screen and determining the average value of some key parameters such as image quality and confidence index. These values are generally accepted as measures for the quality of EBSD scans [38]. In addition, the inverse pole figures (IPFs) may be observed to look for any anomalies. Subsequently, the conventional detector screen is taken out, and the cooled screen is placed. Again, several scans should be made and the same values as before should be determined to be compared with the case of the conventional detector screen. With the cooled detector screen in place, this should be done both with still water, i.e. without flow, and with flowing water in order to verify which role water itself and water flow play, independently of each other.
Since still water shouldnβt affect the imaging, the main potential influencer is the active flow of the water as disruptions can originate from changes in flow properties such as local velocity gradients or turbulence. To determine whether the potential decrease in quality due to this flow is worth it, this decrease should be compared to the decrease of quality due to the thermal noise that is avoided by the use of flowing water. If quality decreases less due to flow with respect to the quality decrease due to thermal noise, the use of the detector screen cooling is validated. This is under the assumption that the cooling extracts enough heat, which should also be validated.
A.2.2 Cooling
To verify whether the cooling is effective, experiments are conducted while heating a sample. The sample should be heated up slowly while monitoring the temperature of the conventional phosphor screen. In addition to the temperature of the phosphor screen, image quality and confidence index
Page 45 should also be monitored. When the thermal radiation from the sample results in too much thermal noise, the experiment should be aborted. This procedure should be repeated several times for a good average of temperature, confidence index and image quality relation to be formed.
Subsequently, the conventional detector screen is swapped out for the cooled detector screen and this procedure is repeated. While the main focus lies on the temperature of the detector screen, the confidence index and image quality should also be monitored and can be compared to the values of the previous validation experiment, i.e. without heating. If the temperature of the screen remains low, it can be concluded that the cooling works. In addition, if the confidence index and image quality resemble the values resulting from the non-heated experiments but deteriorate over time, it can be concluded that the screen cooling can be used to extend heated experiment duration. However, there will be a limitation in time when confidence index and image quality become too low. Finally, if these parameter resemble the non-heated values and do not deteriorate, the detector screen cooling can be used in heating experiments of undetermined duration.
A.3 Conclusion
While keeping in mind possible effects of a flowing medium on light exiting a phosphor EBSD detector screen and applying safety measures with respect to glass and vacuum, a cooling device has been designed. Due to the deposition of the phosphor screen directly onto the glass substrate, maximal thermal conductivity has been achieved. While flowing water may have an effect on the scan quality, the cooling effect of this water should cause the temperature of the detector screen to remain low. With that, the decrease in imaging quality should be negligible compared to the effect that an increase in temperature has on conventional phosphor screens. Therefore the use of a actively water-cooled EBSD detector screen is deemed potentially useful and necessary for EBSD research at elevated temperatures and should be validated in the future.
Page 46