ZrB₂ powder from Nanografi was used. The powder has a purity of 99.5%, an average particle size of 5.5 µm, and was made of 80% zirconium and 18.9% boron in weight percent. Spark plasma sintering of zirconium diboride was performed using the FCT Systeme Gmbh SPS machine at Delft University of Technology.

Twelve cylindrical ZrB₂ samples were made in total with a diameter of 30 mm and thickness of 3 mm. Seven samples were sintered at a maximum dwelling temperature of 2000°C, and five samples were sintered at various maximum temperatures to change the densification level. The maximum sintering temperature was varied between 1800–2,000°C at increments of 50°C (1,800°C, 1,850°C, 1,900°C, 1,950°C, 2,000°C). As shown in Figure 1, the temperature was kept at the maximum for 10 min, after which it was cooled to 400°C during 64 min (25°C/min). On the other hand, the punch force was kept at 5 kN for 21 min, then held at 56 kN (∼80 MPa) for 15 min and 30 s. After that, it was reduced back to 5 kN for the remaining time. About 0.2 mm of material on each side was removed after the polishing was carried out. After the samples were polished to remove embedded graphite, they were re-polished using a standardized polishing procedure at the aerospace faculty. All the samples were cut into two-halves, one exposed to radiation and one kept unirradiated as a reference to analyze the influence of the radiation.

The electron radiation exposure was conducted using a Van de Graaf accelerator in the reaction institute at TU Delft. The electron accelerator uses a moving belt that collects charges on an insulated sphere, creating a high potential difference. The potential difference accelerates the electrons to high energies by creating an electron source. Halves of the SPS-ed ZrB₂ samples were exposed to electrons with an energy of 3 MeV at three different fluences while the other half of each pellet was kept as a reference. Electron fluences were varied by changing the irradiation time between 0 s (6 samples), 283.54 s (2 samples), 466 s (2 samples), and 1886.97 s (2 samples). This coincided with the Van de Graaf fluences and fluxes seen in Table 1. The Van de Graaf fluence was calculated assuming a 18.9, 23.0 and 28.4 μA beam current which could result in an average dose density of 59.7 kGy/cm² in 240 s on a Nylon material with a density of 1.15 g/cm³.

The ZrB₂ samples’ thermal and optical properties were measured after irradiation. Prior to the irradiation of the samples, XRD-analysis with a Rigaku Miniflex 600 and densification measurements using the Archimedes method were conducted. Thereafter, scanning electron microscopy (SEM) with a JEOL JSM7000F was conducted to investigate possible microstructural changes after irradiation at the highest fluence. Using a diamond cutter, the samples were cut into three 5 mm × 5 mm squares, then embedded into an epoxy resin with the thickness side facing up and polished. The embedded samples were thereafter sputtered in a 15 nm gold coating and had aluminum foil applied to the sides before being inserted into the SEM. Lastly, the thermo-optical properties were characterized. The coefficient of thermal expansion was measured using a PerkinElmer TMA 4000 thermo-mechanical analyzer with a temperature range of 100–600°C with a heating rate of 20 ^°C/min in a nitrogen flushed atmosphere. The thermal conductivity was measured using a Hot Disk TPS 2200. Room-temperature optical properties such as thermal emittance and solar absorption were measured at ESA-ESTEC in Noordwijk, the Netherlands. The thermal emittance was measured using an ET-100 handheld emissiometer measuring wavelengths between 2–3.5 µm, 3–4 μm, 4–5 μm, 5–10.5 µm and 10.5–21 µm. To measure the solar absorptance an Agilent UV-VIS-NIR spectrophotometer measured the spectra between 250–2,500 nm.

To ensure minimal impurities were present, XRD-analysis was carried out on the purchased powder from Nanografi as well as a sintered sample (crushed into powder) prior to irradiation. By comparing these two XRD-patterns as seen in Figure 2, it can be observed that there are minimal graphite impurities (Figure 2B). Using the XRD-data, the c and a lattice parameters of the crushed samples were calculated as listed in Table 2. It can be seen that these values are close to reference values (Villars et al., 2023). The density of the ZrB₂ samples were measured to investigate the densification effect of sintering temperature. As seen in Figure 3, there was no clear correlation between sintering temperature and density, which contradicts the relationship found in literature, where as sintering temperature increases from 1800 to 1950 ℃ , the sintered density changes from 87% to 97% (Guo et al., 2008). A possible explanation for difference was the much higher heating rate (>200 ℃ /min) and shorter dwelling time (3min) at maximum temperature in the literature, which makes the effect of maximum sintering temperature more prominent. In our case, due to relatively smaller heating rate (80°C/min), an intermediate dwelling of 2 min at 1,500°C and longer dwelling time (10min) at maximum temperature, all ZrB₂ samples achieved a relative density higher than 96% in the sintering temperature range we investigated.

Figure 4 shows SEM images of the cross sections of both prime and irradiated ZrB₂ samples. It can be seen that there is no apparent microstructural degradation in the bulk. While the grains are not obvious, it can be assumed from the arrangement of the small holes (<1 μ m in diameter, dark phase) that grain size remains largely unchanged after irradiation.

Figure 5 and Figure 6 display the thermal conductivity and coefficient of thermal expansion of the ZrB₂ samples plotted against increasing irradiation fluences. Despite the large uncertainties in the measurements, it can be seen that there is no obvious change. There has been quite large discrepancy in reported room temperature thermal conductivity values in literature, values as high as 88–100 Wm⁻¹ K⁻¹ were reported by Harrington et al. depending on the carbon content with a maximum 4.68 volume % C content (Harrington et al., 2015), while values as low as 56 W/m K was also reported by Zimmermann et al. (Zimmermann et al., 2008). A reason for this discrepancy could be contamination from balling media WC as a secondary phase in the reported literature (Zimmermann et al., 2008). Our measured thermal conductivity value (∼50 W/m/K) is on the lower end as well. An investigation into the carbon content through back scattered SEM reveals the volume fraction of graphite to be less than 1%, which is in agreement with our XRD analysis (that no obvious graphite peak was detected as shown in Figure 2). This low amount of carbon is unlikely to cause such drastic reduction of thermal conductivity compared to 88 W/m/K at 4.68 volume% C in Harrington’s study. While we did not use WC ball milling, we did use a different starting ZrB₂ powder which has 0.1% Si and 0.08%, 80% Zr and 18.9% B. Given the fact that additives and impurities can have a strong influence on the thermal conductivity of ZrB_2, it is not our intention to quantify what caused this low value but to seek possible change upon irradiation, which does not seem like the case.

On the other hand, our measured CTE value was ranging between 9E-6 to 1.2E-5 K^−1, which in fairly good agreement with reported literature values which has been measured as 6.8E-6 K⁻¹ between 300 K and 1300 K whereas our measurement was done between 373 K and 873 K (Zimmermann et al., 2008). It can be seen that there is a lack of observable damage upon increasing irradiation level in the bulk thermal properties (both the thermal conductivity and CTE) of ZrB₂, which could be due to the fact that the electron irradiation level was too low. NASA has claimed that electron irradiation fluence below 10¹⁸ e⁻cm⁻² would not cause detectable mechanical degradation for metals, whereas for ceramic materials, large changes in the thermal conductivity have been observed at neutron fluences of 10¹⁸ to 10¹⁹ n/cm² (Harold and William, 1970). However, no experimental data was provided to prove the electron radiation resistance of ceramic materials in terms of thermal properties. Albeit the fluence in the present study amounts to 10¹⁵ e⁻cm⁻², it is still three orders of magnitude lower compared to the value claimed for neutrons from the NASA report, while neutrons in general cause much greater degradation compared to electrons because of heavier mass. It is speculated that if we increase the radiation time to 1,000 longer or change the radiation source to heavier particles such as proton, neutron or alpha particles, observable changes in bulk properties such as thermal conductivity and CTE may occur.

There was no significant influence on the optical properties of the increasing electron radiation fluence for the three fluences studied as seen in Figure 7 and Figure 8. It is worth noting that the fluctuation in optical characteristics did not exceed the measurement devices’ accuracy values. All the data points fell between the upper and lower uncertainty levels. It had been reported that proton irradiation on olivine, another inorganic material, generated a shift in the reflectance spectrum (Kanuchová et al., 2010) which is not observed in our case. Since there are numerous differences between the present study and that in (Kanuchová et al., 2010) (material used, irradiation particle, particle energy, particle fluence), it is difficult to pinpoint the reason why such an effect was not observed here. Moreover, a previous study also reported a color change in ZrB₂ under Helium irradiation, which would coincide with a change in absorption properties in the visual spectrum range (Garrison et al., 2018b). In this case, the irradiation energy was 30 keV, but the radiation fluence was significantly higher with values such as 8.4 × 10²¹ He/m² and 5 × 10²² He/m². The main difference between study (Garrison et al., 2018b) and the present study was the use of heavier ion irradiation at higher fluences instead of electron irradiation, which causes significantly more damage to the material. Therefore, the absence of ion implantation and ion sputtering could be contributing factors as to why there was no damage to the optical properties by electron irradiation. Additionally, the electrons have a larger penetration depth so could have caused damage deeper in the ZrB₂ samples and consequently limited damage on the surface.

Figure 9 displays the R_a of the irradiated samples against the irradiation fluence. This is only plotted for samples produced at the same sintering temperature of 2000°C. No apparent influence on R_a can be observed from the three irradiation fluences tested. Similarly, to the thermal property measurements, within the same irradiation fluence (see 3.15E14 e⁻/cm²), there was a more significant change in Ra than with increasing fluence. There was limited research which investigated the influence of irradiation on the surface roughness of materials, and none with ZrB₂. In the study by Carter and Vishnyakov, it was found that silicon had a significantly lower surface roughness under normal incidence Xe⁺-ion irradiation compared to irradiation with incidence angles between 0–40^◦ (Carter and Vishnyakov, 1996). Although surface roughness values were lower for irradiated samples in the present study, it is yet to be further investigated if this is indeed caused by irradiation such as electron sputtering, or rather an artifact of polishing. More intense irradiation would have caused larger-scale damage to the surface of ZrB₂, which should have been visible in SEM imaging after irradiation and thus help clarify this question.

Figure 10 displays the absorptance and R_a for all the measured samples sintered from different temperatures. Based on the linear fit, there was a weak negative correlation between the two variables, albeit with a meager r-squared value. A similar case was seen with the hemispherical thermal emittance measurements in Figure 11, where a weak positive correlation was observed.

According to Wen et al. (Wen and Mudawar, 2006), there exists a relationship between the surface roughness, resistivity and infrared emissivity of thin films, which can be explained by the following mathematical models, R r = R p ⁡ exp − 4 π σ λ 2

Where Rr and Rp represent the reflectance of a rough and smooth surface, respectively, σ corresponds to the root-mean-square roughness and λ refers to the wavelength. According to this theoretical model, higher surface roughness will lead to lower reflectance and thus higher emissivity at a specific wavelength. Zhang et al. (Zhang et al., 2020) discovered that when the surface roughness of ZrB₂ films was in nanometer-scale, it did not show a remarkable effect on the infrared emissivity of ZrB₂ thin films. In our study, the surface roughness is in the micrometer range (0.6–2.6 µm), and we observed a weak dependence, as shown in Figure 10 and Figure 11. It is reasonable for us to conjure if the roughness range were enlarged, greater optical property change would potentially be observed.

What we can conclude also, is that the irradiated samples generally had a lower mean arithmetic roughness than the unirradiated samples. This could indicate a smoothing surface effect from the irradiation originating from effects such as electron sputtering.