Paper published on: Solar Energy Materials and Solar Cells, vol. 109, pp. 8-16 (2013), DOI: 10.1016/j.solmat.2012.10.004 http://www.sciencedirect.com/science/article/pii/S0927024812004710

Suitability of ultra-refractory diboride ceramics as absorbers for solar energy applications

Diletta Scitia, Laura Silvestronia, Luca Mercatellib, Jean-Louis Sansc and Elisa Sani*b

a CNR-ISTEC,Institute of Science and Technology for Ceramics, Via Granarolo 64, I-48018 Faenza (Italy)

b CNR-INO National Institute of Optics, Largo E. Fermi, 6, I-50125 Firenze, Italy b PROMES-CNRS Processes, Materials and Solar Energy Laboratory,7 rue du Four Solaire, 66120

Font Romeu, France * corresponding author: Tel: +39(0)55.23081 Fax:+39(0)55.233775 email elisa.sani@ino.it

Abstract: Ultra-refractory diborides are currently studied mainly as thermal protection materials for aerospace and military applications. However, their favourable properties (very high melting points and good thermo-mechanical properties at high temperatures) can be advantageously exploited to increase the operating temperature of thermodynamic solar plants in concentrating solar power systems. This paper reports on the spectral reflectance characterization of hafnium and zirconium diborides containing MoSi2 as secondary phase to evaluate their potential as novel solar absorbers. To assess the spectral selectivity properties, room-temperature hemispherical reflectance spectra were measured from the UV wavelength region to the mid-infrared, considering different levels of porosity for each system. Moreover, for zirconium diboride and hafnium diboride composites containing 10 vol% of MoSi2 sintering aid, the thermal emittance was measured in the 1100-1400 K temperature range. Room temperature spectral characteristics and high temperature emittance were compared to that of a monolithic silicon carbide.

Keywords: borides; optical properties; emittance; solar absorbers; solar plants; concentrating solar power.

1. Introduction

Concentrating Solar Power (CSP) is considered to be one of the most promising and sustainable technologies for electricity production in the future [1]. For thermodynamic sunlight exploitation, the sun radiation must be collected and concentrated upon a receiver [2-7]. The heat is used to produce steam that drives an electricity turbo-generation unit. The ideal sunlight absorber material is spectrally selective against the thermal radiation: it has a low reflectance

1

(ideally approaching zero) at the solar spectrum wavelengths (λ<3 μm) and a high reflectance (ideally approaching 100%) for longer wavelengths. The requirements about the transition wavelength from the low to the high reflectance spectral regime change according to the required plant operating temperature, which determines the thermal emission spectrum of the heated absorber [8-10].

As it is a general rule that the efficiency of solar thermal systems rapidly increases with increasing working temperature, there is a large effort to increase operating temperatures by developing novel solutions [3-6]. For solar tower approach, a critical parameter for temperature increasing is the material that constitutes the receiver, namely the absorber, where the whole heliostat field concentrates the collected sunlight. Among the solid materials devoted to absorbers, graphite is one of the main ones [6]. While having the highest melting temperature of any known material, it burns at high temperature in the presence of oxygen, and this limits achievable temperatures and thus efficiencies in practical solar plants. Other ceramic materials currently under study mainly for volumetric absorbers, are alumina [5], which is characterized by very high thermal stability, oxidation resistance and high refractoriness, but, being white, it shows non-optimal sunlight absorption, and silicon carbide (SiC) [4], a grey semiconductor with good sunlight absorption and high oxidation resistance.

Diborides of transition metals such as ZrB2, HfB2 are commonly referred to as ultra-high temperature ceramics (UHTCs) for their extremely high melting temperatures exceeding 3000°C, solid-state stability, good thermochemical and thermo-mechanical properties, high hardness, high electrical and thermal conductivities [11-13]. UHTCs are ideal for thermal protection systems, especially those requiring chemical and structural stability at extremely high operating temperatures. Up to now, UHTCs are mainly employed in the aerospace industry for hypersonic vehicles, rocket motor nozzles or atmospheric entry probes capable of the most extreme entry conditions [11-13].

In the framework of CSP thermodynamic sunlight exploitation, the ultra-high melting point of UHTCs, together with the unique combination of good thermal conductivity and chemical stability appear intriguing for employing them in high temperature novel solar furnaces. To assess the potential of UHTCs as selective sunlight absorber in solar furnaces devoted to work in the high temperature regime, the spectral absorber properties of these materials need to be evaluated. Studies of spectral emittance characteristics of either bulk [14-17] or film samples [18] of UHTCs dated back to the 60s-70s and were conducted in the framework of space or military applications. Due to the intrinsic difficulty to perform high-temperature emittance measurements and to the fact that emittance strongly depends on the particular material composition, fabrication method [17] and surface finishing [14 -17], available literature data, if any, are also difficult to interpret and to compare. For example, they show different emittances for materials produced by different suppliers while not giving information about several important parameters such as actual material composition, densification level, surface finishing, porosity etc. In addition, available data often refer to different physical quantities (total hemispherical, total normal or spectral normal emittances) or disagree among each other. Very recently we characterized the room temperature spectral reflectance and high temperature total emittance properties of some UHTC carbides [19-21], while, to the best of our knowledge, reports on the optical properties of ceramic diboride materials and focused on their application for solar energy exploitation are lacking.

In the present work we describe the comparative optical characterization of various zirconium and hafnium diborides, with different density levels. We measured the room-temperature

hemispherical reflectance spectra from the ultraviolet (UV) to the mid-infrared (MIR) wavelength regions. In particular, UV-visible-near infrared (NIR) spectra allowed us to characterize the sunlight absorbing properties of the samples, whereas investigations in the MIR region allowed a preliminary evaluation of the material properties for use as high-temperature thermal emitter. Moreover, for selected ZrB2 and HfB2 samples, we measured the total hemispherical emittance in the wavelength range from 0.6 to 40 μm for temperatures ranging from 1100 to 1400 K. For a more significant assessment of actual UHTC potential as solar absorber, the optical properties of UHTC samples are compared with those of a sample of monolithic SiC.

2. Experimental

2.1 Materials preparation and characterization Commercial powders were used to prepare the ceramic materials. Compositional details are

reported in Table I. ZrB2 and HfB2-based materials were either pure samples or prepared with MoSi2 as sintering aid, with a typical content of 5-10 vol%. In the latter case, the powder mixtures were milled in absolute ethanol using ceramic milling media, subsequently dried in a rotary evaporator and sieved through a 250 µm screen. The final density and porosity were tailored through addition of the sintering aid and/or adjusting the thermal cycle. Further details on materials’ preparation are available in refs [22-26]. Samples constituted by monolithic SiC were also produced. After sintering, flat surfaces were prepared by grinding and polishing the sintered materials with diamond pastes up to 1 µm. The mean surface roughness (Ra) 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. Surface roughness values for the polished samples are reported in Table I. Microstructural features were analysed by scanning electron microscopy (SEM, Cambridge S360, Cambridge, UK) and energy-dispersive spectroscopy (EDS, INCA Energy 300, Oxford Instruments, High Wycombe, UK) and image analysis (Image pro plus 7.0).

Two nearly dense ZrB2- and HfB2-based samples containing 10 vol% MoSi2 were selected for high temperature emittance measurements. To this purpose, discs of 40 mm diameter, 2 mm thickness and 300-400 nm surface roughness were prepared. For the sake of comparison, SiC discs were also prepared with 300 nm surface finishing. Before and after the test the surface morphology was analysed by SEM-EDS and X-ray diffraction spectra were also collected to detect any variation in crystallographic phases.

2.2 Room temperature reflectance measurements Optical reflectance spectra in the 0.25-2.5 µm wavelength region were acquired using a

double-beam spectrophotometer (Lambda900 by Perkin Elmer) equipped with a 150-mm diameter integration sphere for the measurement of the hemispherical reflectance. The spectra in the wavelength region 2.5-14.3 μm have been acquired using a Fourier Transform spectrophotometer (FT-IR "Excalibur" by Bio-Rad) equipped with a gold-coated integrating sphere and a liquid nitrogen-cooled detector. In all cases the reflectance spectra are acquired for

quasi-normal incidence.

2.3 High temperature emittance measurements High-temperature hemispherical emittance has been measured using the MEDIASE setup

developed at PROMES-CNRS laboratory [27-29]. MEDIASE (Moyen d’Essai et de DIagnostic en Ambiance Spatiale Extreme) is composed by a vacuum chamber with ultimate pressure limit of 10-6 mbar, equipped with a hemispherical silica glass window of 35 cm diameter. Figure 1 shows the scheme of the experimental chamber, that can be equipped with various additional instruments if required by the experiment. The water-cooled sample holder is set to keep the specimen at the focus of the 1-MW solar furnace. Care is taken in mounting the sample in such a way as to minimize the thermal losses towards the holder. Directional radiance was measured on the rear face of the sample at different angles thanks to a movable, computer-controlled three-mirror system, using a radiometer as light detector. The acquired wavelength range is 0.6–40 μm. The spectral response of the whole optical system was calibrated against a reference blackbody. The definition of the total (i.e. integrated on the whole available wavelength range) emittance ε(θ, T) is the following:

( , )( , )( )B

I TTI Tθε θ = (1)

where I(θ,T) is the directional total radiance emitted by the sample at the angle θ and temperature T and IB(T) is the blackbody directional total radiance at the same temperature T, both acquired for the whole sensitivity range of the detector. (Total) hemispherical emittance values ε∩(T) were obtained by integration of the (total) angular values over a half space, assuming hemispherical symmetry of the emission lobe. The samples were heated by concentrated solar power and the radiance was collected from a small region of about 7 mm diameter (for 0° angle of incidence) around the sample center using a Cassegrain optical system in front of the radiometer. The surface temperature was approximated by the color temperature [30] measured using a pyro-reflectometer specifically developed at PROMES-CNRS [31,32] and read near the center of the sample. The relative standard uncertainties of the temperature measurement are typically from 0.5 to 1%, while the relative standard uncertainty of the emittance measurement is 2.5%.

Figure 1: scheme of the MEDIASE setup. 1: hemispherical window; 2: water-cooled thermal shield and sample holder; 3: specimen; 4: optical fiber; 5: 3-mirror goniometer; 7: measurement window; 8: bichromatic pyro-reflectometer; 9: radiometer; 6, 10, 11, 12: holders for additional instruments (not used for the present experiment).

2.4 Calculation of the spectral directional emittance In the following we recall some useful definitions. More details can be found in textbooks of

radiative heat transfer, for instance in ref. [33]. For opaque materials, as in our case, the Kirchoff's law applies. Therefore, the spectral

directional emittance ε’(λ,T) referred at a temperature T is equal to the spectral directional absorbance α’(λ, T), and for the energy conservation it is given by

ε’(λ,T)= 1-ρ’∩(λ,T) (2)

where ρ’∩(λ,T) is the spectral directional hemispherical reflectance (i.e. the spectral reflectance measured for light coming from a single incidence direction and reflected into the whole half space in front of the surface, like measurements performed with an integrating sphere), measured at the same temperature T.

The directional emittance ε’λ1,λ2(T) in the wavelength range (λ1,λ2) is thus related to the spectral directional emittance ε’(λ,T) by the equation:

2

11, 2 2

1

'( , ) ( , )' ( )

( , )

T B T dT

B T d

λ

λλ λ λ

λ

ε λ λ λε

λ λ

⋅=∫

∫ (3)

where B(λ,T) is the blackbody spectral radiance at the same temperature T. The total solar directional absorbance α’S(T) of the sample is obtained integrating the spectral

directional absorbance α’(λ,T) on the Sun emission spectrum S(λ), according to the equation:

max

min

max

min

(1 ' ( , )) ( )' ( )

( )S

T S dT

S d

λ

λλ

λ

ρ λ λα

λ λ

∩− ⋅=∫

∫

λ

(4)

To give an estimation of the absorption and emission properties of the investigated samples from room temperature measurements, we considered the measured quasi-normal hemispherical reflectance spectra (i.e. the hemispherical reflectance measured for quasi-normal light incidence) at room temperature ρ+∩(λ, Trt) and we calculated:

a) the room-temperature solar quasi-normal absorbance α+S (Trt) according to Eq. (4), using as

S(λ) the CIE Sun spectrum with air mass=1.5 [34] and λmin=0.3 µm, λmax=2.3 µm; b) an approximated quasi-normal emittance ε+

λ1,λ2 (Trt,T) integrated in the wavelength range λ1=0.3-λ2=14.3 µm, considering in the calculation of Eq. (3) the blackbody emission at T=1400 K. It should be noticed that the obtained values are only an approximate indication of the quasi-normal hemispherical emittance at the temperature T. Anyway, the obtained approximate values can be useful for a first comparison between the different investigated materials when high temperature measurements are unavailable.

3. Results

3.1 Microstructural features of borides

Microstructural details of the samples under investigation are illustrated in Fig. 2. To obtain fully dense ZrB2 and HfB2 samples, at least an amount of 5-10 vol% of MoSi2 and a temperature of 1850°C are needed. Lower amounts of MoSi2 and/or lower sintering temperatures result in materials with residual porosity. ZrB2 without sintering aids resulted in samples with 25% (ZB75) porosity as confirmed by both density and porosity evaluation. The density was raised to 85% (1800°C) by adding just a 5 vol% of MoSi2. Increasing both the sintering temperature to 1850°C and the MoSi2 content to 10%, the final relative density reached 95-97%. Pore shape and morphology is apparent in the micrograph of Fig. 2. Pore diameter frequency distribution is illustrated in Fig. 3. Interconnected pores with average diameter up to 10 µm and irregular morphology can be seen in ZB75. Open porosity converts into closed porosity as the density increases. At the same time, the maximum and average pore dimensions decrease with increasing density as reported in Table I. If the constituent phases are considered, after sintering the dominant crystalline phases are ZrB2 and MoSi2 [22]. The final volumetric amount of MoSi2 is slightly lower than starting composition due to the reaction during sintering arising in the formation of MoB and Mo-Si-B mixed phases, with a total amount of 1-2 vol%. In the final microstructure ZrB2 and MoSi2 are not distinguishable, because they display similar contrast both in optical and electron microscopy. However, ZrB2 grains have a rounded shape, whilst MoSi2 phase is characterized by irregular shape with concave angles (see inset in Fig. 2c).

As for HfB2-based materials, without any sintering aid a sample with a final porosity of about 25% (HB75) was obtained after sintering at 1750°C (Fig. 2d-f). With addition of 3-5 vol% of MoSi2 and progressive increase of the sintering temperature, densities of 85, 95 and 99% (sintering at 1900°C) were obtained. Pore dimensions in HfB2 samples are usually finer than

ZrB2-based materials due to the finer grain size of the starting boride powders with respect to ZrB2 powders (Fig.3b). According to X-ray diffraction, only HfB2 was detected along with small traces of HfO2 and SiC, with no MoSi2 residuals due to the conversion of the silicide into SiC [25]. A SiC sample was obtained by sintering at 1800°C with small percentage of Alumina and Yttria as sintering aids. Matrix Sample

label Density (%)

Sintering additive amount (vol%)

Mean pore diameter (µm)

Min-max pore diameter (µm)

Surface roughness Ra (µm)

ZrB2 ZB75 75 0 3.5 0.5 - 11 ~0.05 ZrB2 ZB85 85 5 2.3 0.5 - 9 ~0.05 ZrB2 ZB95 95 10 1.5 0.4 - 4 ~0.03 ZrB2 ZB97 97 10 0.4 0.1 - 4 ~0.04 HfB2 HB75 75 - 1.8 0.5 – 8 ~0.03 HfB2 HB85 85 3 0.9 0.5 - 4.0 ~0.05 HfB2 HB95 95 5 1.9 0.7- 8 ~0.03 HfB2 HB99 99 5 <0.1 - ~0.02 Table 1: Main features of the UHTC samples for the reflectance spectra.

Figure 2. Microstructural features of borides. Pore amount and morphology in a) ZB75, b) ZB85, c) ZB95. Inset in fig 2c represents the typical microstructure of ZrB2-MoSi2 composites. Pore amount and morphology in d) HB75, e) HB85. f) Typical microstructure of HB99.

Figure 3. Pore diameter distributions: a) ZrB2- based materials; b) HfB2-based materials.

3.2 Room temperature reflectance spectra The room-temperature hemispherical reflectance spectra of samples under investigation are

shown in Figure 4,a-c. The inset in each picture shows the detail of the 0.3-2.5 µm wavelength range. Figure 4a compares the reflectance curves of dense ZrB2 and HfB2 materials with that of SiC. Borides exhibit step-like reflectance curves similar each other, characterized by low reflectance values in UV-VIS and high reflectance values in the MIR, typical of metallic materials described by the Drude-Lorentz model [35]. On the contrary, SiC shows the typical spectrum of a semiconductor, with a well defined reststrahlen band at around 12 μm and a considerably lower value of the reflectance for longer infrared wavelengths [35]. Figure 4b shows the reflectance spectra of zirconium diborides for compositions with different densities. It can be appreciated a gradual increase of reflectance with increasing density, due to the decrease of light trapping by the material pores with increasing sample density. When the density reaches 90-95% there is no appreciable difference amongst the samples. All the samples show a secondary reflectance minimum at wavelengths slightly shorter than 2 µm, with a small wavelength shift of ZB85 towards shorter wavelengths. Samples with the highest MoSi2 sintering aid content, namely

ZB97 and ZB95, evidence a secondary reflectance minimum at around 2.7 µm wavelength, which has been measured for pure MoSi2 [21], thus confirming the presence of some residuals of this compound in the final samples, with a larger residual amount of MoSi2 in ZB97 than in ZB95. This however does not led to a consistent difference for the calculated solar absorbance (about a 4% difference) while larger differences can be found for the calculated emittance (cfr. Table III).

Concerning hafnium-based diborides, from the spectra in Fig. 4c they can be divided in two groups, according to their reflectance values for wavelengths shorter than about 10 µm. Samples with densities in the range 75-85% have a lower reflectance, with almost no spectral differences among them, while samples with higher densities show higher reflectance curves very similar each other in the wavelength range 3-10 µm, with some differences in the range 1-3 µm that can be reasonably ascribed to light trapping within surface pores. For wavelengths longer than 10 µm the reflectance curves of all samples converge to a value higher than 95%, which increases with the wavelength. Similarly to ZrB2 samples, a secondary reflectance minimum at wavelengths slightly shorter than 2 µm was detected in all the HfB2 spectra. Finally, it should be noticed that the reflectance spectra of zirconium diboride-based samples show quite scattered curves with respect to the hafnium-based ones, that displays a more defined step-like behavior. This could be probably ascribed to the higher pore size dispersion described above.

Figure 4. Reflectance curves: a) dense ZrB2, HfB2 and SiC samples; b) ZrB2 samples with different level of porosity; c) HfB2 samples with different level of porosity

3.3

is

ly lower of that of silicon carbide (about a half of it in the temperature range 1300-400 K).

Sample label Family

ty oughness

µm)

High temperature emittance Table II summarizes some microstructural details for the composites selected for high

temperature emittance. Fig. 5 shows the measured temperature-dependent hemispherical emittance of ZB97* and HB99*. The obtained data are compared with a sample of a 90% dense silicon carbide with 0.3 μm surface roughness, labeled as SiC’. Both borides show similar high-temperature emittance values, in agreement with the similarity of their room-temperature spectralreflectance curves. On the other hand, from Fig. 5 we can appreciate that emittance of boridesignificants

1

Relasitive

den(%)

Sinteringadditive (% vol)

Surface rRa (

ZB97* ZrB2 97 MoSi2 (10) 0.3

HB99* HfB2 99 MoSi2 (10) 0.4

Table II: Samples for high temperature emittance characterization. The asterisk indicates a different surface finishing with respect to the samples used for room-temperature reflectance measurements.

Figure 5. Comparison of the measured total hemispherical emittance of zirconium and hafnium diborides and silicon carbide as a function of the temperature.

To assess the stability of samples under high-temperature exposure in vacuum, we compared

the surface features before and after the tests. ZrB2 and HfB2 showed similar results, apart from removal of residues of the final machining. Fig. 6 shows SEM and EDS analyses performed of ZB97* and HB99* before and after the high-temperature test. No significant variation of surfamorphology were detected. X-ray diffraction (not shown) performed before and after the test confirmed the presence of the starting phases ZrB2 and MoSi2 with no chemical reaction between them, as expected. EDS analyses performed on the surface before and after tests revealed a slighvariation of boron and oxygen peaks, which could be related to form

ce

t ation of amorphous boron

xide resulting from oxidation of ZrB2 or HfB2. o

Figure 6. (a)-(b): Surface morphology of ZB97* a) before and b) after the emittance tests and corresponding EDS spectra. (c)-(d) Surface morphology of HB99* c) before and d) after the emittance tests and corresponding EDS spectra.

3.4-

t

aterials, as discussed below. Calculations for a dense SiC sample are lso included for reference.

Results of spectral calculations The values of α+

S(Trt) (Eq. 4) and ε+λ1,λ2 (Trt,T) (Eq. 3), calculated from the acquired room

temperature reflectance spectra are listed in Table III, together with their ratio and with the available measured values of quasi-normal emittance at 1400 K temperature for comparison. Ishould be noticed that the calculated values should be considered approximate because, for a quantitative assessment, the spectra should be acquired at the absorber operating temperature. Anyway they can be helpful when high temperature data are not available, or for a preliminary screening among different ma

Sample label ative sity

rt) (Trt ,T)

3 μm; .3 μm 00 K

+λ3,λ4(T)

3=0.6 μm; 4=40.0 μm

Trt)/ε+λ1,λ2(Trt ,T)Rel

den(%)

α+S(T

ε+

λ1,λ2

λ1=0.λ2=14T= 14

ελλT=1400K

α+S(

ZB75 75 0.57 0.34 1.7 ZB85 85 0.53 0.26 2.0 ZB95 95 0.47 0.20 2.3 ZB97 97 0.49 0.25 ZB97* sample

2.0 0.44 HB75 75 0.57 0.30 1.9 HB85 2.0 85 0.59 0.29 HB95 95 0.48 0.19 2.5 HB99 99 0.48 0.19 HB99* sample

0.46 2.5 SiC 100 0.78 0.80 SiC’ sample

0.82 1.0 Table III: Calculated values of total solar quasi-normal absorbance and quasi-normal thermal emittance in the range 0.3-14.3 μm. The quasi-normal emittance measured in the range 0.6-40.0 μm for ZB97*, HB99* and SiC’ is also included for reference.

Zirconium and hafnium diborides appear to show similar characteristics both in sola

absorbance and thermal emittance, as confirmed by the high temperature measurementssolar absorbance and thermal emittance are higher for porous samples, as expected. If the α+

S/ε+λ1,λ2 ratio is concerned, porous samples appear less performing (i.e. the value of the

α+S/ε+

λ1,λ2 ratio is lower) than dense ones and hafnium-based samples appear similar to zirconium-containing pellets. The reference SiC has a higher solar absorbance, but the considerably higher emittance unfavourably compares it to borides for solar applications, as itsα+

S/ε+λ1,λ2 ratio is about unity. High temperature measurements confirm the similarity b

ZrB2 and HfB2-based samples in emittance behaviour although measured values are much hthan those ca

r . Both

etween

igher lculated from room temperature spectra. Moreover, both calculations and high-

mperature measurements confirm the very different behavior of borides with respect to SiC. his suggests that the room-temperature spectral characterization can be helpful for a

luation of different samples and when high temperature measurements are

ero) er

teTcomparative evaunavailable.

4. Discussion

4.1 Suitability of borides as solar absorbers The ideal solar absorber should agree with several requirements, as already said: it has to be

spectrally selective against the thermal radiation with a low reflectance (ideally approaching zat the solar spectrum wavelengths and a high reflectance (ideally approaching 100%) for long

wavelengths. A low radiative emission at the operating temperature is required to minimize energy losses and to avoid heating of elements surrounding the absorber. In fact, energy losses via thermal emission significantly affect the overall system efficiency. Moreover, as thermal emittance increases with the temperature, a large emittance value negatively affects also thdesired increase of the operating temperature for the thermodynamic solar plant. Furthermore, for the practical use, the absorber melting point should be far higher than operative temperaturematerial should possess mechanical resistance up to the o

e

s. The perative temperatures and a critical

the n to face possible s,

etctransfer the collected heat to a thermo-convective fluid.

1)

2) tigated temperature range is very low. This result ere

re. In o

3) erature range from 500 to 800 MPa, depending on the sintering

procedure [22-25]. At 1773 K, specifically for composites containing MoSi2 in suitable amounts (10-20 vol%), flexural strength can reach 650 MPa and 550 MPa in air for zirconium and hafnium diborides respectively [22]. Tthe general form

rmal shock resistance compatible with the concentrated solar radiatiooperating temperature variations (caused, for example, by variations of weather condition, cloud

). Finally, a high thermal conductivity is desired because the absorber should efficiently

In this respect, borides seem to accomplish all these requirements. The metal-like optical behavior of the investigated borides can be clearly recognized by step-like reflectance curves with low reflectance in the UV-VIS and high reflectance values in themedium infrared. The emittance of ZrB2 and HfB2 in the invesis quite surprising because borides are indicated as material for aerospace applications whhigh emittance values are desired to favor heat dissipation during re-entry in atmosphethe latter case however, the investigated samples are SiC-reinforced materials and undergmuch higher temperatures, i.e. T>1800 K. Thermo-mechanical requirements are satisfied: for borides, typical values of mechanical strength at room temp

hermal shock resistance can be evaluated through ula:

(1 )cE

dT ν στ

−= ii (5)

where ν is the Poisson coefficient, σ the flexural strength, E the Young’s modulus, τ the thermal expansion coefficient. Using typical values found in the literature we get a critical thermal shock temperature ΔTc of the order of 150-200 K, which is suitable for this kind of application. Chemical stability: present conditions adopted for high temperature characterization are vermild, both for the chosen temperatures and for the high level of vacuum (10-6 mbar). As a consequence, very little surface modification was observed as expected. In the case that Zr-HfB2 composites are exposed to air at higher temperatures, oxidation phenomena are expeto occur. For the specific case of ZrB2-MoSi2, oxidation involves formation of an external amorphous borosilicate layer resulting from oxidation of MoSi2, producing silica, and ZrB2producing boron oxide, when the temperature is

4) y

or cted

, higher than 1450 K [36]. As emittance of

ld be operated in vacuum or under inert atmosphere.

silica is as high as 0.8, it is expected that oxidation phenomena can significantly raise the emittance of the absorber. The oxidation behavior of borides will be the subject of further investigations. However, this possible drawback could be relaxed if the absorber cou

5) Thermal conductivity: according to recent results in the literature, thermal conductivity of 6 W(m K)- 1 at room temperature and increases gradually to 67 W (m K)-1 . For HfB values as high as 104 W(m K)- 1 were reported [11].

s,

high e the ability of the material to absorb the solar

diation is tailoring the material porosity to efficiently capture the radiation. Data obtained so far incomplete material sintering are partially

as

p to

monolithic ZrB2 is 5up to at 1675 K [37] 2

4.2 Effect of porosity Experimental data indicate that either the total amount of porosity and the pore size affects

the reflectance of borides. ZB75 and ZB85 are characterized by pores and cavities as large as 10-15 µm. Due to this, they are able to partially trap radiation and their reflectance values are always lower than those of ZB95 and ZB97 in the overall range of investigated wavelengths.

On the other hand, porous HfB2 samples (namely HB75 and HB85), show reflectance valueswhich are similar to those of dense samples for λ>8-10 µm, due to the fact that pores are much smaller than those of ZrB2 samples with similar levels of porosity. Thus, in this class of materiala difference in reflectance are only seen in the range 0-8 µm. It is worthy to note that in HB95 pore dimension range is similar to HB75. Nevertheless, its reflectance is closer to that of a fully dense material such as HB99. In this case, reflectance values suggest that total amount of pores rather than their dimensions dictates the reflectance properties. Although borides show the mentioned step-like reflectance behavior, room temperature reflectance values are still quitein the sunlight spectral range. One way to improvrashowed that the porosity obtained by leaving an effective in improving the material absorbance.

4.3 Effect of secondary phases: MoSi2 vs SiC Because of strong covalent bonding and low self-diffusion, high temperatures and external

pressures are required to densify diborides. MoSi2 and SiC particulates are the most common additives used to improve densification and properties of ZrB2. The most important advantage of using MoSi2, is that through the addition of at least 10 vol% of this silicide, dense ZrB2 and HfB2ceramics can be produced even in the absence of applied pressure [22]. This represents an enormous improvement from the technological point of view, because it significantly reduces the costs of materials production. During heating, a transient liquid phase forms due to the reaction between surface oxide impurities on the diboride particles. Specifically, B2O3 from the diboride particle surface reacts with MoSi2 to form a transient liquid phase, which helps the matter transferphenomena. In the final microstructure, although a part of MoSi2 can react with the matrix, forming small amounts of intermediate phases, most of this secondary phase remains confined particles amongst the boride grains, as can be observed in the microstructure [26]. Another advantage of the MoSi2 addition is that ceramics can retain their room temperature strength uat least 1800 K in oxygen or inert atmosphere.

From the optical properties point of view, the presence of MoSi2 is not detrimental for the overall optical properties of the composite, as we assessed that the room temperature optical spectra are nearly not affected by the presence of an amount of MoSi2 up to 10% vol.

On the other hand, historically, SiC particulate additions have been the most common approach to improve the strength and oxidation resistance of ZrB2-based ceramics. Although we

have not measured the room temperature spectrum of a ZrB2-SiC composite, emittance data inthe present work can be compared with those found in ref. [38], obtained for a ZrB2-SiC system with 15vol% SiC. It can be seen that emittance of the ZrB2-SiC composite is significantly higher than the ZrB2-MoSi2 composites investigated in the present work at the same operating temperature. The increase in emittance is probably due to the presence of the SiC particulate, which, h

aving an intrinsic high emittance as confirmed by our measurements, raises the overall mittance of the composite in comparison with that of ZrB2-MoSi2. This suggests that MoSi2 is ore suitable than SiC as sintering aid and secondary phase in borides to be used for solar energy

the increase in emittance is useful for aerospace but detrimental for solar ene

at,

. The room temperature optical spectrum of SiC is reported in Fig. 3 along with that of b C is

ove. pared

to b

e

usly des

/ε ratio for borides over SiC.

ls 0

cts of oxidation in air are expected for both the classes of materials, thanks to the

sly

of solid-state sintered SiC is now

emapplications. In fact,

rgy.

4.4 Borides vs SiC

SiC has been recently started to be employed as absorber in solar applications. In spite of thvery little information is available in the literature on its optical properties, especially at high temperature

orides. Differently from borides, which show a metal like behavior, the spectrum of Sitypical of semiconductor materials with a well defined reststrahlen band, as underlined abFurthermore, SiC has a much higher absorbance at solar spectrum wavelengths when com

orides. The most striking difference displayed by the two classes of materials lies in the high

temperature regime, because in all the investigated temperature range (1100-1450 K) themittance of SiC is significantly higher than that of the tested boride-MoSi2 composites.

As for the α/ε ratio, in the simplified assumptions underlying the calculations previocribed, we get that almost all the borides show α/ε ratios of about 2, with the minimum value

(1.7) obtained for a porous ZrB2. On the other hand, the same ratio for SiC is about 1. Considering that α can be increased by surface texturing without increasing the thermal emittance[8], potentially we should be able to further increase the α

As for the thermo-mechanical properties, borides perform better as far as mechanical strengthis concerned, especially at high temperature. Strength at 1800 K is hardly higher than 400 MPa for solid-state sintered SiC, whilst strength for ZrB2- and HfB2 –based composites can reach 500 and 600 MPa, respectively, as summarized in Table IV.

Concerning the material degradation at high temperature, it has to be recalled that emittance tests were conducted in high vacuum. However, to understand the suitability of these materiafor high temperature absorbers, their oxidation behavior in air should be considered. Up to 140K, very limited effe

development of a silica surface layer [40]. Thermal conductivity of SiC is much higher than ZrB2 at room temperature but rapidly decreases at high temperature. For instance at 1400 K, thermal conductivity is 45 W (m K)- 1 for SiC, and still 60 W (m K)- 1 for ZrB2, as previouindicated [11,37].

Finally, as for the cost of the ZrB2 products, processing temperatures of borides can be even lower than those required for SiC despite the common belief that UHTCs are very difficult to densify. For instance, conventional sintering for production

adays 2400 K, whilst it has been demonstrated that ZrB2-MoSi2 composites can be sintered at 2100 K. Although the cost of final components is the result of many production costs, including

the raw materials, industrial process, final machining, etc, the possibility of sintering at lowetemperatures is considered a key factor for cost reduction.

Which materials is preferable between borides and SiC for solar absorber applications is san open issue and depends on many parameters, including the desired operating temperature and desired system operating conditions (e.g. the concentration ratio [41]). However, a series of favourable characteristics, including the higher α/ε ratio and higher thermal stability of UHTCover SiC, good processability and higher mechanical strength and spectral selectivity suggest they have a good potential for this kind of application. Pitz-Paal et al. have demonstrated thatuse of spectrally selective absorber

r

till

s

the s improves the solar receiver performances. The efficiency

provement is particularly significant for low concentration systems and high temperatures. Mor given concentration rati ly i he re r temperature [41], thus motivating the search for nov terials able t jugate ultra-hi mperature resis y.

imeover, for a o, it nonlinear ncreases with t ceive

el ma o con gh tetance and spectral selectivit

Property ZrB2-MoSi2 HfB2-MoSi2 SiC Emittance (mean value 1100-1400K)

0.4 0.4 0.7

Flexural strength at RT (MPa) 500-700 [22,23] ] 600-800 [24,25 400-700 [39] Flexural strength at 1800 K (MPa) 500-600 [24,25] 500 [22,23] 400 Processing temperature (K) 1, 1, 2100-2400 [1

22, 23] 2200-2400 [122, 23]

2250-2400

Thermal conductivity (W/mK) 110 (RT) – 40 )

57 (RT) – 67 (1800K) [11] (1800 K

104 (RT) [11]

Oxidation products Amorphous borosilicate layer up to 1800 K

Amorphous borosilicate layer up to 1800 K

Amorphous silica

Table IV: Property comparison between borides and SiC. RT = room temperature.

o

ion

ver SiC,

d to

avelength range 0.6-40 μm, obtaining an

5. Conclusions In this work we evaluated, for the first time to the best of our knowledge, the intrinsic spectral selectivity of different Ultra High Temperature Ceramics, hafnium and zirconium diborides, tevaluate their potential for novel sunlight absorbers in tower solar plants operating at higher temperature. We measured the room-temperature hemispherical reflectance spectra from 0.25 to14 μm, evidencing typical curves with a low reflectance in the UV-VIS-NIR wavelength regand a high reflectance plateau at the wavelengths of the thermal infrared. From the acquired reflectance spectra, we calculated the quasi-normal solar absorbance and an approximate quasi-normal thermal emittance. We evidenced favorable spectral characteristics of borides othat reasonably arise in a higher absorbance over emittance ratio and thus in superior solar absorber performances. Within the boride family, hafnium and zirconium boride are expecteshow similar performances. In general, dense samples appear more promising than porous ones.Moreover, for samples of dense zirconium and hafnium boride, we measured the high-temperature hemispherical emittance integrated in the w

emittance value around 0.4 for temperatures from 1100 to 1450 K. Both room-temperature reflectance spectra and high-temperature emittance favourably compare UHTC borides against

HTC borides as good candidates for novel absorbers able to raise the

FERA Grant Agreement n. 228296) is gratefully acknowledged, as well s the PROMES Director and PROMES Researchers for the use of facilities. Authors thank Dr.

Franco Francini for useful discussions about optical properties of UHTCs, C. Melandri for easurements, R. Renis for materials preparation, Mauro Pucci and Massimo D’Uva

for

Ref

Citro, Decarbonization: Achieving near-total energy independence and y in

30

nofluids as direct sunlight absorbers. Optics Express 18 (2010) 5179-87.

hic honeycombs as

y

ic conversion system, Solar Energy 58 (1996) 191-5

aterials.

SiC, thus proposing Uoperating temperature of thermodynamic solar plants.

Acknowledgments Financial support by the Access to Research Infrastructures activity in the 7th Framework Programme of the EU (Sa

roughness mtechnical assistance.

erences

[1] R. Shinnar, F.near-total elimination of greenhouse emissions with available technologies, TechnologSociety 30 (2008) 1–16 [2] S.A. Kalogirou, Solar thermal collectors and applications. Prog Energy Combust Sci(2004) : 231-95. [3] E. Sani, S. Barison, C. Pagura, L. Mercatelli, P. Sansoni, D. Fontani, et al. Carbon nanohorns-based na[4] C.C. Agrafiotis, I. Mavroidis, A.G. Kostandopoulos, B. Hoffschmidt, P. Stobbe, M, Fernandez- V. Quero, Evaluation of porous silicon carbide monolitvolumetric receivers/collectors of concentrated solar radiation. Sol Energy Mater. Sol. Cells 91 (2007) 474-88. [5] J. Karni, A. Kribus, R. Rubin, P. Doron, The “porcupine”: a novel high-flux absorber for volumetric solar receivers,J. Solar Energy Eng. 120 (1998), 85-95. [6] T. Yan, Y. Chen, Review of study on solid particle solar receivers, Renew. Sust. EnergRev. 14 (2010) 265-76. [7] H. Naito, Y. Koshaka, D. Cookie, H. Arashi, Development of a solar receiver for a high-efficiency thermoionic/Thermoelectr[8] W.F. Bogaerts, C.M. Lampert, Materials for photothermal solar energy conversion, J. Mater. Sci. 18 (1983) 2847-75. [9] C.M. Lampert, Advanced optical materials for energy efficiency and solar conversion, Sol. Wind Techn. 4 (1987) 347-79. [10] C.E. Kennedy, Review of mid- to high-temperature solar selective absorber mNational Renewable Energy Laboratory report no. NREL/TP-520-31267 (2002) [11] W.G. Fahrenholtz, G.E. Hilmas, I.G. Talmy, J.A. Zaykoski, Refractory Diborides of Zirconium and Hafnium, J. Am. Ceram. Soc. 90 (2007) 1347–1364 [12] C. Mroz, Zirconium diboride, Am. Ceram. Soc. Bull. 73 (6) (1994) 141–142.

[13] E. Wuchina, M. Opeka, S. Causey, K. Buesking, J. Spain, A. Cull, J. Routbort, E. Gutierrez-Mora, Design for Ultra high temperature applications: The mechanical and thermal properties of HfB2, HfCx, HfNx and Hf(N), J. Mater. Sci., 39(2004) 5939-49.

ra 1961. 16 August 1961. (Contract

nd al report no. WADD-TR-60-646 Part III Vol. II

al report no. WADD-TR-60-646 Part IV(1964) t

eering, Purdue

- ut

li, F. Francini, J.-L. Sans, D. Sciti, Ultra-refractory ceramics for

solar plants, J. Renew. Sust.

2 d by pressureless sintering, Scripta Mater. 57 (2007) 165-168

es, ng A, 434 (2006) 303-306

s

253-3258

ico-chemical 180

[14] P.T.B. Shaffer, Development of Ultra Refractory Materials, Carborundum Co., NiagaFalls, New York. Progress report no. 25, 1 May - 31 JulyNo. rd-17175) ASTIA AD-261 471.15 [15] G.M. Kibler, T.F. Lyon, M.J. Linevski, V.J.Desantis, Carbonization of plastics arefractory material research. Vol. II. Technic(1964) [16] G.M. Kibler, T.F. Lyon, M.J. Linevski, V.J.Desantis, Refractory material research. Technic[17] D.P. DeWitt, T.K. Riddle, R.E. Taylor, Spectral and total emissivity and reflectivity ahigh temperatures. Report no. PRL 152, School of Mechanical EnginUniversity, West Lafayette, Indiana (1978). [18] Determination of emissivity of materials, Volume I Interim final report, Jul. 1, 1959Dec. 31, 1962. Report no. PWA-2206, Pratt & Whitney Aircraft, East Hartford, Connectic(1962) [19] E. Sani, L. Mercatelhigh-temperature solar absorbers, Scripta Mater. 65 (2011) 775–778 [20] E. Sani, L. Mercatelli, D. Fontani, J.-L. Sans, D. Sciti, Hafnium and tantalum carbides for high temperature solar receivers, J. Renewable Sustainable Energy 3(2011) 063107. [21] E. Sani, L. Mercatelli, P. Sansoni, L. Silvestroni, D. Sciti, Spectrally-selective Ultra-High Temperature Ceramic absorbers for high-temperature Energy, in press (2012) [22] L. Silvestroni, D. Sciti, Effects of MoSi2 additions on the properties of Hf- and Zr-Bcomposites produce[23] D. Sciti, F. Monteverde, S. Guicciardi, G. Pezzotti, A. Bellosi, Microstructure and mechanical properties of ZrB2-MoSi2 composites produced by different sintering techniquMaterials Science and Engineeri[24] D. Sciti. L. Silvestroni, A Bellosi, Fabrication and properties of HfB2-MoSi2 compositeproduced by hot pressing and spark plasma sintering, Journal of Materials Research, 21 (2006) 1460-1466[25] D. Sciti, G. Bonnefont, G. Fantozzi, L. Silvestroni, Spark plasma sintering of HfB2 with low additions of silicides of molybdenum and tantalum, Journal of the European Ceramic Society, 30, (2010) 3[26] L. Silvestroni, H.-J. Kleebe, S. Lauterback, M. Müller, D. Sciti, Transmission Electron Microscopy on Zr- and Hf-borides with MoSi2 Additions, Journal of Materials Research 25 (2010) 828-834. [27] M. Balat-Pichelin, D. Hernandez, G. Olalde, B. Rivoire, J. F. Robert, Concentrated solar energy as a diagnostic tool to study materials under extreme conditions, J. Solar Energy Eng. 124 (2002) 215–22. [28] T. Paulmier, M. Balat-Pichelin, D. Le Quéau, R. Berjoan, J. F. Robert, Physbehavior of carbon materials under high temperature and ion irradiation, Appl. Surf. Sci.(2001) 227–45.

rature of opaque materials using a

v. Sci. Instrum. 66 (1995) 5548-5551.

ine the true temperature on opaque surface without

lar Spectral Irradiance. CIE (Commission Internationale de l’Eclairage) Technical

1983).

7-

D. Porter, H.

ect of Machining .

ves on densification, microstructure nd properties of liquid-phase sintered silicon carbide, J. Mater. Sci. 35 (2000) 3849-3855.

[40] J. W. Fergust, W. L. Worrell, Effect of Carbon and Boron on the High-Temperature Oxidation of Silicon Carbide, J. Am. Ceram. Soc. 78 (1995) 1961–1964. [41] R. Pitz-Paal, High temperature solar concentrators, in Solar Energy conversion and photoenergy systems Vol. 1, J. Blanco and S. Malato eds. (Encyclopedia Of Life Support Systems EOLSS, http://www.eolss.net)

[29] T. Paulmier, M. Balat-Pichelin, D. Le Quéau, Structural modifications of carbon–carbon composites under high temperature and ion irradiation, Appl. Surf. Sci., 243 (2005) 376–93. [30] D. Hernandez, A concept to determine the true tempetricolor pyroreflectometer, Review of Scientific Instruments 76 (2005) 024904. [31] D. Hernandez, G. Olalde, A. Beck, E. Milcent, Bicolor pyroreflectometer using an optical fiber probe, Re[32] D. Hernandez, J.L. Sans, A. Netchaieff, P. Ridoux, V. Le Sant, Experimental validation of a pyroreflectometric method to determhampering reflections, Measurement 42 (2009) 836–843. [33] M. F. Modest, Radiative heat transfer, 2nd edition, Academic Press, 2003 [34] SoReport no. 85 (1989). [35] C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, New York, [36] D. Sciti, M. Brach, A. Bellosi, Long-term oxidation behaviour and mechanical strength degradation of a pressurelessly sintered ZrB2-MoSi2 ceramic, Scripta Mater. 53 (2005) 1291302. [37] J.W. Zimmermann, G.E. Hilmas, W. G. Fahrenholtz, R. B. Dinwiddie, W.Wang, Thermophysical Properties of ZrB2 and ZrB2–SiC Ceramics, J. Am. Ceram. Soc., 91(2008) 1405–1411 [38] L. Scatteia, D. Alfano, F. Monteverde, J.-L. Sans, M. B. Pichelin, Effmethod on the catalicity and emissivity of ZrB2 and ZrB2-HfB2-based composites, J. AmCeram. Soc., 91 (2008) 1461-1469 [39] D. Sciti, A. Bellosi, Effects of additia