Lanthanum hexaboride for solar energy applications

We investigate the optical properties of LaB6 – based materials, as possible candidates for solid absorbers in Concentrating Solar Power (CSP) systems. Bulk LaB6 materials were thermally consolidated by hot pressing starting from commercial powders. To assess the solar absorbance and spectral selectivity properties, room-temperature hemispherical reflectance spectra were measured from the ultraviolet to the mid-infrared, considering different compositions, porosities and surface roughnesses. Thermal emittance at around 1100 K has been measured. Experimental results showed that LaB6 can have a solar absorbance comparable to that of the most advanced solar absorber material in actual plants such as Silicon Carbide, with a higher spectral selectivity. Moreover, LaB6 has also the appealing characteristics to be a thermionic material, so that it could act at the same time both as direct high-temperature solar absorber and as electron source, significantly reducing system complexity in future concentrating solar thermionic systems and bringing a real innovation in this field.

characterization of LaB 6 ceramics. For thermodynamic solar applications, the knowledge of optical properties of LaB 6 is of interest because, as recalled above, it is used as additive for UHTC borides.
However, we point out that the innovation potential of LaB6 is much higher. In fact, compared to previously investigated UHTCs, LaB 6 has the appealing characteristics to be a thermionic material, so that it could act at the same time both as direct high-temperature solar absorber and as electron source, significantly reducing system complexity in future concentrating solar thermionic systems and bringing a real innovation in this field. Moreover, if surface impurities will be carefully controlled, e.g. with a proper control of production processes and with operation in vacuum, so that the work function is maintained below 4.13 eV (corresponding to light wavelengths longer than 300 nm), direct electron emission (photoemission) from LaB 6 induced by absorption of the UV-blue spectral portion of concentrated sunlight could be additionally and simultaneously exploited (Fig. 1). Thus, to assess the potential of LaB 6 for this radically new concept of solar absorber we carried out our study as a function of the sample porosity and surface finishing.

Results and Discussion
Microstructural features. Table 1 lists investigated samples, while Fig. 2 shows microstructures and EDS spectra. As expected, the monolithic material (LaB 6 _p), Fig. 2a,b contains a significant fraction of open porosity, in agreement with density measurements. The grain size ranges from 3 to 10 µm. Traces of La-B-O spurious phases were recognized along the grain boundaries by SEM-EDS analysis. (see inset of Fig. 2b and corresponding EDS spectra). These impurities are typical contaminants of the boride powders and hindered the material densification during hot pressing 40 .
From both Fig. 2a,b and Table 1 the difference of surface roughness between as sintered and polished surface can be appreciated (Ra decreased from 0.58 to 0.09 µm for unpolished and polished surfaces, respectively).
As for the LaB 6 _d sample, addition of ZrB 2 and B 4 C notably improved the densification. The final material was completely dense and the microstructure can be observed in Fig. 2c-e. According to XRD (not shown) no extra phases formed. Dark spots in Fig. 2c,d belong to B 4 C, while the in secondary electron imaging LaB 6 and ZrB 2 display similar contrast. By in-lens signal, see Fig. 2e, dark contrasting grains belong to LaB 6 , whilst light contrasting grains belong to ZrB 2 . The improvement in the densification was mainly attributed to the addition of B 4 C, via cleaning of surface oxides such as La x O y , B x O y from the boride particles, as observed for similar ceramics 40 . Increase in the final density also affected the quality of surface polishing and consequently the final roughness decreased to Ra 0.03 µm, see Table 1.  Optical characterization. Figure 3 shows the reflectance spectra of LaB 6 samples, compared to previously analyzed dense hafnium and zirconium borides 41 . The inset of Fig. 3 shows the reflectance in the sunlight spectral region, the sunlight spectrum is also superimposed for reference. LaB 6 reflectance directly increased with decreasing of the surface roughness. This is particularly evident comparing LaB 6 _p rough and LaB 6 _p polished. For the dense composite material, the reflectance was further improved due to the almost complete elimination of porosity. The spectrum shape presents very similar features in pure LaB 6 and the LaB 6 -based composite containing ZrB 2 and B 4 C, except for a local minimum at round 3 μm wavelength and a couple of secondary small minima between 7 and 10 µm, present in both spectra of porous samples. These features could be ascribed to oxide phase impurities in porous LaB 6 ceramics, like La-O phases, B-O, or mixed La-B-O phases (as those observed in Fig. 2b). In contrast, these spurious phases were not observed in the optical spectra of LaB 6 _d owing to their removal by B 4 C, as previously mentioned. No specific features in the spectrum that could be attributed to B 4 C inclusions were recognized in LaB 6 _d.  Compared to hafnium and zirconium borides, LaB 6 has a considerable lower reflectance (i.e. a higher absorbance) near the peak of sunlight spectral distribution. However, thanks to peculiar spectral features of LaB 6 , the rough porous sample, LaB 6 _p rough remains more absorptive up to about 1 μm wavelength. Towards longer wavelengths, the reflectance of porous pellets is about 20% lower than that of other borides, while the dense sample shows a similar reflectance at the infrared plateau region.
From the experimental room-temperature hemispherical reflectance ρ ∩ (λ) we calculated the total solar absorbance, α: where B(λ, 1100 K) is the blackbody spectral radiance at 1100 K temperature and λ1 = 0.3 µm and λ2 = 16.0 µm. The α/ε ratio (sometimes called spectral selectivity) is a parameter assessing the material potential for solar receiver applications, and ideally should be taken as high as possible. Figure 4 shows absorbance and spectral selectivity for the investigated LaB 6 samples, together with the two reference borides. From Fig. 4 we can immediately appreciate the high potential of LaB 6 as solar absorber material: both solar absorbance and spectral selectivity are comparable or higher than those of comparison materials ZrB 2 and HfB 2 41 . Even more, the rough porous sample (LaB 6 _p rough) shows a remarkable solar absorbance of 0.7, which is only slightly lower than that of the most advanced solar absorber material to date in actual plants, namely silicon carbide (SiC, α ≈ 0.8 41 ) (see ref. 43 and references therein). As for spectral selectivity, for all samples it remains higher than that of SiC (α/ε ≈ 1 41 ) and for most of them it is comparable to that of the two reference borides, with the LaB 6 dense polished sample showing a remarkable value of α/ε = 6.4. If we compare LaB 6 pellets each other, we can say that a high solar absorbance is obtained at the expense of spectral selectivity. However, even in its less Scientific RepoRts | 7: 718 | DOI:10.1038/s41598-017-00749-w absorptive form (dense polished), LaB 6 results promising, as said before, with respect to previously investigated borides, as it shows a comparable absorbance and a remarkably higher spectral selectivity.
For a more reliable assessment of LaB 6 potential for solar absorber applications, we measured the thermal emittance at high temperature. In fact it is known that emittance calculated from room-temperature reflectance spectra are a very useful tool for a preliminary material evaluation, but underestimate the value at high temperature 44 . Figure 5 compares the experimental spectral normal emittances (solid lines) with the values calculated from room-temperature reflectance data ε calc (λ) = 1 − ρ ∩ (λ) (dashed values). As previously mentioned, the measured emittance is higher than the calculated one. Samples maintain the hierarchy among them, as expected, with the dense pellet showing the lowest emittance (spectrally integrated value of 0.2), the polished porous LaB 6 the intermediate value (0.4) and the rough porous the highest value (0.6). As a term of comparison, it should be noticed that LaB 6 favourably compares to SiC also if experimental emittances are considered, because the obtained values are significantly lower than those experimentally obtained for SiC pellets (0.9 at 1100 K).
If the spectral shape of emittance curves is concerned, we can observe that for the dense sample, calculated and experimental curves are very similar, while porous samples show some differences. In particular, for porous pellets, we cannot find in experimental emittances the wide bands of local maxima peaked at around 3, 7 and 9 µm wavelengths, that can be seen, on the contrary, in the curves calculated from room temperature spectra. Even if a quantitative evaluation requires additional microstructural measurements after high temperature experiments, which are beyond the scope of this work and will be the subject of a further study, we can explain the obtained results in terms of degassing of impurities still present in porous samples. For instance, according to studies carried out on the boron oxide vapor pressure 45 , 1100 K and ~10 −6 mbar cause boron oxide species vaporization in boride samples, which could be the case of the LaB6_p sample subjected to high vacuum/high temperature during the emissivity measurement.
In summary, the study we cave carried out both at room and high temperature show that LaB 6 not only is comparable or better than other boride UHTCs, but in its porous rough form is even better than SiC showing a similar solar absorbance and twice the spectral selectivity, as inferred from room temperature measurements, and a lower measured emittance at 1100 K temperature (0.6 versus 0.9). As already mentioned, SiC is the most advanced solar absorber actually used in plants to date.

Conclusions
This work is devoted to the study of the spectral reflectance and thermal emittance of pure LaB 6 and LaB 6 -ZrB 2 -SiC composites in order to evaluate their potential as novel solar absorbers. Pure LaB 6 materials contained a porosity of about 20%, which resulted in a lower reflectance compared to composite materials. In dense composites containing ZrB 2 as secondary phase, LaB 6 has an intrinsic solar radiation selectivity with high reflectance plateau at wavelengths longer than 1 µm and a remarkable α/ε ratio approaching 6.4. Moreover, the porous sample with rough surface favorably compares even to SiC from all points of view, as it shows a similar solar absorbance and a doubled spectral selectivity. Finally, if we consider the thermal emittance at 1000-1100 K temperature, even the most emissive LaB 6 porous rough sample has a significantly lower emittance than SiC.
However we emphasize that, in our opinion, the innovation potential of LaB 6 is much higher because, in addition to the promising characteristics listed above, it is also a widely documented thermionic material, and with one of the highest known electron emissivities. Thanks to these properties it could successfully act, at the same time, both as direct high-temperature solar absorber and as electron source, significantly reducing system complexity in future concentrating solar thermionic systems. This could make it a really revolutionary material in solar technology. The powder mixture was then densified by hot pressing at 1900 °C, 40 MPa and holding time 15 min. After sintering, the density of the ceramics was experimentally determined by the Archimede method. The sintered pellets were polished using diamond paste up to 1 µm. The samples' surface were analysed by SEM-EDS (FE-SEM, Carl Zeiss Sigma NTS Gmbh, Oberkochen, DE) and energy dispersive X-ray spectroscopy (EDS, INCA Energy 300, Oxford instruments, UK). The mean surface roughness (R a ) and the distance between the highest asperity, peak or summit, and the lowest valley (R t ) was measured according to the European standard CEN 624-4 using a commercial contact stylus instrument (Taylor Hobson mod. Talysurf Plus) fitted with a 2 µm-radius conical diamond tip over a track length of 8 mm and with a cut-off length of 0.8 mm.
The hemispherical reflectance spectra were acquired using two instruments: a double-beam spectrophotometer (Lambda900 by Perkin Elmer) equipped with a Spectralon ® -coated integration sphere for the 0.25-2.5 µm wavelength region and a Fourier Transform spectrophotometer (FT-IR "Excalibur" by Bio-Rad) equipped with a gold-coated integrating sphere and a liquid nitrogen-cooled detector for the range 2.5-16.5 μm. Thermal emittance has been measured using the setup described in detail in ref. 44 and here briefly recalled for convenience. The apparatus consists of a high-vacuum furnace interfaced either to a Fourier-transform infrared spectrophotometer (Bio-Rad Excalibur) and to a reference blackbody (C.I. Systems SR-2) by means of a splitting optical system. The ultimate pressure limit is few 10 −6 mbar. The temperature is read on the sample upper surface by means of three thermocouples. The uncertainties are ±20 K on the temperature and ±5% on the spectral emittance. As for the temperature, it should be observed that the furnace heater is able to reach up to 1200 K, but the temperature obtained on the measurement surface of each sample is lower because of losses at thermal contact.