Spectrally selective ultra-high temperature ceramic absorbers for high-temperature solar plantsElisa Sani, Luca Mercatelli, Paola Sansoni, Laura Silvestroni, and Diletta Sciti Citation: J. Renewable Sustainable Energy 4, 033104 (2012); doi: 10.1063/1.4717515 View online: http://dx.doi.org/10.1063/1.4717515 View Table of Contents: http://jrse.aip.org/resource/1/JRSEBH/v4/i3 Published by the American Institute of Physics. Related ArticlesHeat transfer and friction factor of solar air heater having duct roughened artificially with discrete multiple v-ribs J. Renewable Sustainable Energy 4, 033103 (2012) Heat transfer and friction characteristics of dimple-shaped roughness element arranged in angular fashion (arc)on the absorber plate of solar air heater J. Renewable Sustainable Energy 4, 023112 (2012) Experimental investigation of energy and exergy efficiencies of domestic size parabolic dish solar cooker J. Renewable Sustainable Energy 4, 023111 (2012) Photon reabsorption in fluorescent solar collectors J. Appl. Phys. 111, 076104 (2012) Mathematical modeling of a prototype of parabolic trough solar collector J. Renewable Sustainable Energy 4, 023110 (2012) Additional information on J. Renewable Sustainable EnergyJournal Homepage: http://jrse.aip.org/ Journal Information: http://jrse.aip.org/about/about_the_journal Top downloads: http://jrse.aip.org/features/most_downloaded Information for Authors: http://jrse.aip.org/authors

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

Spectrally selective ultra-high temperature ceramicabsorbers for high-temperature solar plants

Elisa Sani,1,a) Luca Mercatelli,1 Paola Sansoni,1 Laura Silvestroni,2

and Diletta Sciti21CNR-INO, National Institute of Optics, Largo E. Fermi, 6, I-50125 Firenze, Italy2CNR-ISTEC, Institute of Science and Technology for Ceramics, Via Granarolo 64,I-48018 Faenza, Italy

(Received 31 August 2011; accepted 26 April 2012; published online 11 May 2012)

Ultra-high temperature ceramics are the ideal materials for extreme conditions

owing to their very high melting points and good thermo-mechanical properties at

high temperatures. For these reasons, they are widely known as materials for

aerospace applications. This paper presents a comparative spectral characterization

of zirconium, hafnium, and tantalum carbides ultra-high temperature ceramics for

concentrating solar power applications. Room-temperature reflectance spectra have

been measured from the ultraviolet wavelength region to the mid-infrared band.

Using these spectral properties, the ceramics were evaluated as sunlight absorbers

in receivers for high-temperature thermodynamic solar plants. VC 2012 AmericanInstitute of Physics. [http://dx.doi.org/10.1063/1.4717515]

I. INTRODUCTION

Different concepts for solar receiver systems have been developed. In addition, also within

a given approach, different solutions are possible. For instance, if we consider linear parabolic

collectors operating at low and medium temperatures, they can convert incident solar radiation

into heat in different ways. The most diffused systems use sunlight absorption by blackened or

specially developed absorbing surfaces, from which heat is collected and conducted to a heat

transfer fluid. Another possibility is the direct absorption of sunlight by a heat transfer fluid

flowing through transparent tubes (a black liquid1 and, more recently, a nanofluid2,3). Different

collector architectures with associated working temperature ranges require tailored systems sol-

utions.4 It is a general rule that the efficiency of a collector increases with the concentration

factor and, using a turbine, the efficiency to produce electricity improves with the working tem-

perature. In solar tower plants, the heliostat field concentrates the collected sunlight onto the re-

ceiver, and its high temperature capability strongly determines the system performance. Several

receiver schemes and materials have been studied. They can be basically distinguished in three

groups: receivers with sunlight absorption by particles suspended in a gas stream,5 cavity

receivers with absorption by a bulk material and either employing water/steam or molten salt as

heat transfer fluid,4 and volumetric receivers with absorption by a porous material and heat

transfer to a gas flowing through it.6 Each solution has advantages and drawbacks and the best

tradeoff generally depends on the specific application. Advantages of the proposed bulk

absorber cavity scheme are low heat loss, easy control, high heat capacity, and the possibility

to exploit mature technologies of conventional fossil fuel power plants.7

Ultra-high temperature ceramics (UHTCs) are a family of materials including borides, car-

bides, and nitrides of hafnium, zirconium, and tantalum. UHTCs are characterized by some of

the highest melting points of known materials: in particular, monocarbides of Ta and Hf exhibit

the highest melting temperatures (>3900 �C) among every known compound. In addition, all

a)Author to whom correspondence should be addressed. Electronic mail: elisa.sani@ino.it. Tel.: þ39(0)55.23081.

Fax: þ39(0)55.233775.

1941-7012/2012/4(3)/033104/13/$30.00 VC 2012 American Institute of Physics4, 033104-1

JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

UHTCs are very hard, have good wear resistance, mechanical strength, good chemical stability,

strength at high temperatures, and high thermal conductivity. Due to these properties, UHTCs

are ideal elements for thermal protection systems, especially those requiring chemical and struc-

tural stability at extremely high operating temperatures. Nowadays, UHTCs are being investi-

gated in the aerospace industry for hypersonic vehicles, rocket motor nozzles, or atmospheric

entry probes, compatible with the most extreme entry conditions.8–11 The ultra-high melting

points of UHTCs, together with the unique combination of good thermal conductivity and

chemical stability, suggest consideration of these materials for application in solar furnaces.12,13

For evaluation of the potential of UHTCs as selective sunlight absorbers in solar cavities

operating at very high temperatures, it is essential to evaluate the spectral absorber properties

of these materials. The ideal absorber material is spectrally selective with respect to thermal

radiation: it has a low reflectance (ideally approaching zero) at solar spectrum wavelengths

(k< 3 lm) and a high reflectance (ideally approaching 100%) for longer wavelengths. The ther-

mal emission spectrum requirements, at the transition wavelength range from low to high re-

flectance, change according to the plant operating temperature.12,14 Studies of spectral emissiv-

ity characteristics of either bulk15–18 or film samples19 of UHTCs dated back to the 1960 s and

were conducted in the framework of space or military applications. Due to the intrinsic diffi-

culty to perform high-temperature emissivity measurements and to the fact that emissivity

strongly depends on the peculiar material composition, fabrication method,18 and surface

finishing,15–17 available literature data are difficult to interpret and to compare. For example,

they show different emissivity values for materials produced by different suppliers, while they

do not give information about several important parameters, such as actual material composi-

tion, densification level, surface finishing, and porosity. In addition, available data often refer to

different physical quantities (total hemispherical, total normal, or spectral normal emissivities)

or may show significant disagreement. Moreover, many of the UHTCs are significantly non-

stoichiometric20 and the precise compositions and phases are often unclear from the literature.

While optical properties of films and coatings of some transition metal carbides and nitrides

were investigated,21–25 to the best of the author’s knowledge, reports on optical properties of

UHTC carbides are available in the literature only for single-crystalline samples of tantalum

carbide26 and zirconium carbide.27,28 However, polycrystalline UHTC carbides are easier to

produce than single crystals for solar absorber applications. Compared to single crystals, poly-

crystalline ceramics are expected to exhibit different optical properties due to light scattering

effects from ground boundaries and pores.29 The present paper provides a preliminary compara-

tive analysis of several ceramic ultra-high temperature carbide materials aimed to identify the

most promising for concentrating solar power (CSP) applications. We describe the optical char-

acterization of various zirconium, hafnium, and tantalum carbide-based compositions. Their op-

tical properties are compared to those of more conventional ceramic materials such as SiC-

based ceramics, which were used in volumetric solar absorbers.30 Room-temperature reflectance

optical spectra have been measured from ultraviolet (UV) to mid-infrared (MIR) wavelength

regions. In particular, UV to near infrared (NIR) spectra have enabled determination of the sun-

light absorbing properties of samples, whereas investigations in the MIR region have permitted

a preliminary evaluation of the materials properties as high-temperature thermal emitters. It

should be pointed out that the absorber emissivity properties in the medium infrared band are

related to the MIR reflectivities. These properties are required to evaluate new approaches to

increase the operating temperature of thermal solar plants.

II. EXPERIMENTAL

A. Materials preparation

Commercial powders were used to prepare the ceramic materials: ZrC Grade B (H.C.

Starck, Germany), mean particle size 3.8 lm; HfC (Cerac, Inc., Milwaukee, WI), particle size

range 0.2–1.5 lm; TaC (Cerac Inc, Milwaukee, WI), particle size range 0.2–1.5 lm; TaSi2

(ABCR, GmbH & Co, Karlsruhe, Germany), �45 lm; MoSi2 (<2 lm, Aldrich, Steinbeim, Ger-

many). Compositional details are reported in Table I. Some samples were prepared by sintering

033104-2 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

the as-received powders at temperatures lower than 2000 �C, in order to produce porous materi-

als. Other materials were prepared mixing the carbide powder with sintering aids such as

MoSi2 or TaSi2 in amounts between 5 and 20% vol. in order to get fully dense pellets. 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 lm screen. The pellets

were sintered at temperatures around 1900 �C. Further details on materials’ preparation are

available in Refs. 31–35.

After sintering, the bulk densities were measured by the Archimedes method. Crystalline

phases were identified by x-ray diffraction (XRD, Siemens D500, Karlsruhe, Germany). Bulk

compositions and surface morphology were analysed by scanning electron microscopy (SEM,

Cambridge S360, Cambridge, UK) and energy-dispersive spectroscopy (EDS, INCA Energy

300, Oxford Instruments, High Wycombe, UK). Image analysis (IMAGE Pro-Plus 7.0, Media

Cybernetics, Silver Spring, MD) was also performed on SEM micrographs to quantify the

amount of secondary phases. Flat surfaces were prepared by grinding and polishing the sintered

materials with diamond pastes with particle size down to 1 lm.

The mean surface roughness (Ra) and the distance (Rt) between the highest asperity, peak

or summit, and the lowest valley were measured according to the European standard CEN 624-

4 using a commercial contact stylus instrument (Taylor Hobson mod. Talysurf Plus) fitted with

a 2 lm-radius conical diamond tip over a track length of 8 mm and with a cut-off length of

0.8 mm. Surface roughness values are reported in Table I.

B. Optical measurements

Optical reflectance spectra in the 0.25–2.5 lm wavelength region were acquired using a

double-beam spectrophotometer (Lambda900 by Perkin Elmer) equipped with a 150-mm diame-

ter integration sphere for the measurement of both the total and purely diffuse reflectance.

Specular reflectance spectra in the 1.5–25 lm wavelength region were acquired using a Fourier

Transform FT-IR "Excalibur" Bio-Rad spectrophotometer.

As for the evaluation of relative weight of diffuse versus specular reflectance, at first, it

should be observed that ideal perfectly polished surfaces show only specular reflection, whereas

real surfaces always have a diffuse reflectance component that becomes increasingly negligible

as surface roughness decreases and light wavelength increases. Since our goal was to identify

the most promising UHTC families for solar thermal applications, the characterization of sur-

face scattering was beyond the scope of the present work. For this reason, this paper mentions

TABLE I. Compositions, density, and roughness of the investigated samples, Ra: roughness, Rt: peak to valley distance.

The MoSi2 sample was included for comparison.

Sample

label

Starting

composition (vol. %) Density (%)

Secondary phases

detected after sintering

Ra

(10�3 �lm)

Rt

(lm)

ZC99 100 ZrC 99 C (1%) 23 6 2 0.68 6 0.08

ZCM100 90 ZrCþ 10MoSi2 100 SiC (5%) MoSi2 (�6%–7%) 16 6 3 0.46 6 0.16

ZCM96 80 ZrCþ 20MoSi2 96 MoSi2 (>15%) SiC (2%–3%) 48 6 3 1.8 6 0.4

HC67 100 HfC 67 HfO2 (3%) 60 6 10 1.26 6 0.43

HCT85 95 HfCþ 5 TaSi2 85 TaSi2 (�5%) HfO2 (3%) 20 6 4 0.35 6 0.14

HCM95 90 HfCþ 10 MoSi2 95 MoSi2 (10%) SiC (<1%), HfO2 (3%) 13 6 2 0.20 6 0.06

HCM100 85 HfCþ 15 MoSi2 100 MoSi2 (15%) HfO2 (3%) 12 6 1 0.10 6 0.01

TC76 100 TaC 76 — 65 6 10 1.4 6 0.5

TC92 100 TaC 92 Ta2O5 (�3%) 14 6 1 0.27 6 0.21

TCT97 85 TaCþ 15 TaSi2 97 TaSi2 (15%) SiO2 (<3%) 15 6 1 0.16 6 0.02

TCM100 85 TaCþ 15 MoSi2 100 MoSi2 (15%) Amorphous SiO2 (2%–3%) 10 6 1 0.10 6 0.01

SC100 98 SiCþ 1 B4Cþ 1 C 100 B4C (1%), C (1%) 10 6 1 0.10 6 0.01

MoSi MoSi2 100 Amorphous SiO2 (5%) � �

033104-3 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

only some basic considerations on surface quality, just to be able to estimate the confidence

level of acquired specular reflectance spectra. To do that, we followed the convention, applied

in the market of optical components, that the surface is considered to be an acceptable mirror

for the light wavelength k if it has peak-to-valley features with maximum k/4 depth.36 On the

base of the measured roughness values, summarized in Table I, we concluded that all the car-

bide samples satisfy the k/4 criterion for wavelengths longer than 0.26 lm.

To assess the response of our ceramics to the sunlight spectrum, we measured both the

total (hemispherical) and the purely diffuse reflectance values in the UV to NIR spectral band.

On the other hand, for all samples, the MIR specular reflectance spectrum can be practically

considered as total reflectance within the scope of the present analysis, due to the combined

effect of surface roughness and light wavelength, as already discussed. Therefore, the MIR

spectra provide useful information about the spectral curve, which allows discriminating the

spectral behavior of different UHTC families. The reflectance of MoSi2 was measured consider-

ing that this intermetallic was used as additive for the densification of other ceramics.

III. RESULTS

A. Microstructural features of the sintered samples

The UHTC carbide ZrC, HfC, and TaC are crystalline compounds of a host metal and car-

bon; they have an NaCl-type structure and are also referred to as interstitial carbides. These car-

bides have the physical properties of ceramics and the electronic properties of metals, i.e., high

hardness and strength with high thermal and electrical conductivities as well as metal-like

reflectivity spectra with a high expected reflectivity value in the infrared band. Furthermore,

they have the highest melting points of any group of materials, combined with high thermal

and chemical stability. Some basic properties are summarized in Table II (data from Ref. 37).

A drawback of this class of materials is the poor resistance to oxidation. All these carbides start

to readily oxidize at T> 700 �C in air, forming porous scales together with large volume expan-

sions, cracking and detachment from the substrate.31,33

On the other hand, SiC is a hard semiconducting material and has been used in ceramics

for many years as abrasive, wear resistant, oxidation resistant, and corrosion resistant parts. SiC

is very refractory, tending to dissociate at temperatures above 2500 �C. For this reason, it is cur-

rently employed in high temperature structural components. In comparison with other carbides,

SiC is quite difficult to densify without additives because of the covalent nature of Si-C bond-

ing and the low self-diffusion coefficient.38 In contrast to the UHTC carbides, SiC has a very

good oxidation resistance up to 1800 �C due to the development of a protective silica layer.39

The typical microstructural features of tested materials are briefly summarized in the next

paragraphs. Starting compositions, final densities, compositions, and surface roughness values

are reported in Table I.

ZrC-based materials: Pure ZrC labelled as ZC99 was densified by hot pressing at

1930 �C and is characterized by a regular microstructure with equiaxed ZrC grains with dimen-

sions around 10–20 lm (Fig. 1(a)). Residual porosity was 1 vol. %. Traces of graphite are visi-

ble in the microstructure and are due to C contamination in the starting powder. The sample

labelled as ZCM100 is a composite initially containing 10 vol. % MoSi2 and was sintered by

hot pressing at 1900 �C, reaching the full density (Fig. 1(b)). During densification complex

TABLE II. Summary of the characteristics of investigated materials (data available in the literature).

Density

(g/cm3)

Crystal

structure

Melting

point ( �C)

Electrical resistivity

(X�cm)

Thermal

conductivity (W/mK)

Electronic

properties

ZrC 6.234 Cubic 3420 43� 10�6 20.5 Metal-like

HfC 12.67 Cubic 3930 37� 10�6 20.0 Metal-like

TaC 14.5 Cubic 3950 25� 10�6 22.1 Metal-like

SiC 3.2 Cubic or hexagonal 2500 102–104 110 Semiconductor

033104-4 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

oxidation and carburization reactions involving spurious oxide phases, MoSi2 and C27 resulted

in formation of some SiC. According to x-ray diffraction and SEM-EDS analyses, black con-

trasting features visible in the polished microstructure were SiC particles. Their amount was

estimated through image analysis to be around 5 vol. %, while residual MoSi2 was around 6–7

vol. %. The sample labelled as ZCM96 (Fig. 1(c)) is a composite initially containing 20 vol. %

MoSi2. It was sintered in a conventional graphite furnace under an inert gas flux at 1950 �Cwithout application of external pressure. The obtained sample l density was around 96%. MoSi2amount was higher than 15 vol. %, whilst the amount of formed SiC was around 2–3 vol. %.33

HfC-based materials: A monolithic sample of HfC was consolidated by hot pressing at

1850 �C reaching a final density of 67%. This porous sample, labelled as HC67, contains a

very fine open porosity. Rounded pores have dimensions usually lower than 1 lm (Fig. 2(a)).

HCT85 is a composite material containing 5 vol. % of TaSi2 and a final porosity of about 15

vol. % (Fig. 2(b)). HCM95 and HCM100 are composite materials consolidated by hot pressing

at 1900 �C, with a starting MoSi2 content of 10 and 15 vol. % and with densities of 95% and

100%, respectively (Figs. 2(c) and 2(d)). In the sintered microstructures, the MoSi2 content is

very close to the initial composition, as confirmed by x-ray diffraction spectra.31,34 A low

FIG. 1. Polished surfaces of ZrC-based materials: (a) ZC99, (b) ZC-10 M-100, and (c) ZC-20 M-96.

FIG. 2. HfC-based materials. (a) HC67, (b) HC-5 T-85, (c) HC-10 M-95, and (d) HC-15 M-100.

033104-5 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

amount of SiC was also detected by EDS-analysis (1 vol. %). All the HfC-carbide samples con-

tained 3–4 vol. % of hafnium oxide due to oxide impurities in the starting powder.

TaC-based materials: Microstructural features of the TaC-based materials are shown in

Figs. 3(a)–3(d). TC76 is a monolithic material consolidated by hot pressing at 1800 �C and is

characterized by a fine microstructure and fine porosity and a final relative density of 76%

(Fig. 3(a)). TC92 is a monolithic material consolidated at 1900 �C and reached a much higher

density, around 92% (Fig. 3(b)). TaC grains with size around 10–20 lm are recognizable in the

microstructure together with an amount of very fine closed porosity with pore dimensions in

the range 0.3–1 lm.

TCT97 is a composite material initially containing 15 vol. % TaSi2. It was consolidated by

hot pressing at 1750 �C reaching the full densification. In the matrix, TaSi2 phase is recogniz-

able as the one with irregular features and slightly darker contrast compared to the matrix (Fig.

3(c)). According to x-ray diffraction and image analysis, the content was estimated to be close

to the starting content. Black features in Fig. 3(c) are amorphous phases based on SiO2, arising

from reaction of TaSi2 with oxygen species during the sintering cycle. Their amount is around

2–3 vol. %.31

TCM100 is a composite material initially containing 15 vol. % MoSi2. The material was

consolidated by hot pressing at 1900 �C reaching the full densification. In the microstructure,

beside the constituent phases, a significant fraction of amorphous silica (2%–3%) was

detected31 (Fig. 3(d)). According to x-ray diffraction and image analysis, the MoSi2 content

was very close to the initial content.

SiC-based materials: The SiC-based sample labelled as SC100 is a fully dense monolithic

material sintered at 2150 �C with a low percentage of B4C and carbon as sintering agents. In

the sintered microstructure, SiC grains (around 10 lm) are recognizable, along with a small

fraction of C and B4C impurities (Fig. 4).

Generally speaking, except for pure materials, ZC99, HC67, TC76, and TC92, all compo-

sites contain major amounts of secondary phases as indicated in Table I, deriving from the

addition of sintering agents, MoSi2 or TaSi2. These silicides remained mostly confined as

FIG. 3. Polished surfaces of TaC-based materials: (a) TC67, (b) TC92, (c) TC-15 T-97, and (d) TC-15 M-100.

033104-6 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

distinct grains in the carbide matrices. To a lesser degree, they reacted with oxide species or C

giving rise to SiO2 or SiC species, respectively, depending on the sintering environment/condi-

tions as previously reported.31–34 Concerning the surface preparation, it can be observed from

Table I that the Ra values are very similar for most of investigated materials. Finally, only for

comparative purposes, a pure MoSi2 material was also analyzed, as it is the most widely used

secondary phase for the preparation of these carbides.

B. Optical spectra

Room-temperature optical reflectance spectra of samples under investigation are shown in

Figs. 5–7. The spectra of SiC and MoSi2 are also reported for comparison, as they affect the

reflectivity profile in the majority of samples, as discussed below. The reference SiC-based

sample is characterized by the well recognizable reststrahlen reflectance peak at around 12.4 lm

FIG. 4. Polished surface of the SiC sample.

FIG. 5. Ultrarefractory zirconium carbides reflectance spectra and comparison with a SiC-based material and MoSi2 sinter-

ing aid. The UV to NIR spectrum of MoSi2 could not be measured due to the small dimensions of the sample.

033104-7 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

(hollow square symbols in Fig. 5). On the other hand, MoSi2 has a main reflectivity minimum

at around 2.7 lm and secondary minima at around 6.1, 7.3, and 10.1 lm.

ZrC- based materials: As for zirconium carbides (Fig. 5), pure ZrC (ZC99) shows a step-

like reflectivity curve, with two minima around 0.38 and 1.35 lm and with no other peaks or

additional features in the near and medium infrared.

In both the ZrC-based composites spectra (namely those of ZCM100 and ZC-M96), it is

possible to identify a secondary peak at around 12.4 lm wavelength due to the presence of re-

sidual SiC in the microstructure, as it can be easily recognized by comparing their curves to the

spectrum of the SiC-based material. This peak is less pronounced in ZCM96, due to the lower

amount of SiC (see Table I). In this sample, the spectrum is also affected by the presence of

MoSi2, as inferred from the additional minimum around 2.7 lm.

HfC-based materials: Fig. 6 reports the reflectivity spectra of HfC-based materials. Pure

hafnium carbide (HC67) shows two relative reflectance minima in the UV-visible region, at

around 0.38 and 1.25–1.30 lm, and a monotonic reflectivity increase towards the infrared zone,

FIG. 6. Reflectance spectra of hafnium ultra-refractory carbides. The inset shows the spectra in the wavelength range

0.25–2.5 lm.

FIG. 7. Reflectance spectra of tantalum ultra-refractory carbides. The inset shows the spectra in the wavelength range

0.25–2.5 lm.

033104-8 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

with a small feature around 8.6 lm. A similar UV-VIS spectrum is shown by the sample HC-

5 T-85, produced with the TaSi2 sintering aid. As for the samples produced with the MoSi2 sin-

tering aid, the one with the highest amount (HCM100) shows the typical MoSi2 reflectance

shoulder at around 2.7 lm, thus confirming that the presence of silicide residuals, in the amount

of 10 vol. % or higher, affects the optical spectrum.

As for the dependence of reflectivity characteristics on porosity, Fig. 6 compares materials

with similar surface finishing levels but with different densities (67%, 85%, 95%, 100%, see

Table I). Besides the discussed differences in the spectral distribution due to the use of MoSi2or TaSi2 as sintering aids, the absolute values of reflectance for all samples with densities

higher than 85% appear similar, while the reflectivity of highest porosity samples is lower. This

means that samples with porosities up to 15% actually behave in a similar way, if the optical

reflectivity is concerned. However for larger differences in porosity, like those shown by HC67,

it is reasonable to deduce a reflectivity decrease as the porosity increases (or, equivalently, as

the density decreases). In fact, it seems reasonable to expect some light-trapping-like effects

due to surface pores.

TaC-based materials: the reflectivity curves of pure TaC (TC76 and TC92 samples in Fig. 7)

show a reflectivity minimum at around 0.4 lm (which was detected for all TaC-based materials), a

monotonic reflectivity increase up to 2.3 lm, then a lower slope increase up to around 8 lm, and

finally a nearly constant reflectivity plateau for longer wavelengths. The different reflectivity val-

ues of TC92 and TC76 could be likely ascribed to a porosity effect (see Sec. IV).

As for the doped samples, the spectrum confirms the presence of MoSi2 residuals in the

TCM100 sample (the pronounced shoulder in the reflectivity curve of this sample at around

2.7 lm if compared to those of pure TaC).

The TCM100 and TCT97 samples have similar density (100% and 97%, respectively).

They have been produced with different sintering aids: MoSi2 for TCM100 and TaSi2 for

TCT97. The spectrum of TCT97 shows a pronounced rise edge below 5 lm, with no concavity

change for longer wavelengths. It can be observed that the presence of significant fractions of

TaSi2 in the final sample does not seem to affect the spectral curve, which is different from

that of MoSi2-containing sample. Regarding the effect on reflectance of sample porosity, the

values are very similar for densities higher than 90%, while the values for lower density sam-

ples were significantly lower, as it was also found with the HfC compositions. This means that

about 10% of porosity difference (i.e., for samples with 92% and 100% densities, respectively)

does not significantly alter the optical reflectivity. On the other hand, the significantly higher

porosity of TC76 results in a reflectivity decrease.

IV. DISCUSSION

As general conclusions of this optical characterization, it is apparent from Figs. 5–7 that

ultra-refractory carbides such as ZrC, HfC, and TaC materials behave very differently from

SiC. SiC only has a well-recognizable reflectance peak at around 12 lm of wavelength, with a

slight reflectivity decrease from 15 to 25 lm, in agreement with some of the samples analyzed

in Ref. 18. On the contrary, all UHTCs, even with some differences among them-selves, display

a step-like increase of reflectance from visible to infrared, with presence of a wavelength reflec-

tance cutoff.

The difference in the shape of reflectivity curves of UHTCs with respect to that of SiC

arises from their different energy level structure. UHTCs have metal-like electronic structure

that, from the point of view of optical characteristics, generates a smooth spectral reflectivity

curve, which increases almost monotonically from VIS to infrared, and shows a high value of a

reflectivity plateau for a wide spectral range in the infrared region. On the other hand, the semi-

conducting character of SiC is connected to the existence of discrete vibrational states arising

in big steps in the reflectivity curve with sudden changes from very low to very high reflectivity

values. For UHTCs, the wavelengths corresponding to a reflectance of 50% of the respective

peak value are listed in Table III. Zirconium carbide-based samples have cutoff wavelengths

roughly around 1.1 and 1.6 lm, while tantalum carbide-based materials show a cutoff in the

033104-9 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

visible range, roughly around 0.6 lm for all the investigated density levels. Hafnium-carbide

based sample show a larger variability in the cutoff wavelengths, as they range from 1.08 lm

for the dense HCM95 to 3.47 lm for the porous pure sample HC67. For a single sample of this

series (HCM100), the measurement of the spectrum in the UV-visible range was not possible

because of the very small specimen size.

A good absorber for a solar furnace is a material providing a low ratio between thermal

emittance and solar absorbance at the operating temperatures. For opaque materials, like in our

case (in all cases, we measured transmittance values lower than 0.1% in the wavelength range

from 1.4 to 25 lm, with sample thicknesses of few millimeters), the Kirchoff’s law applies.

Therefore, the directional spectral emissivity e0(k) is equal to the directional spectral absorptiv-

ity a0(k) and for the energy conservation it is given by

e0ðkÞ ¼ 1� q0 \ðkÞ; (1)

where q0\ (k) is the spectral directional hemispherical reflectivity (i.e., the spectral reflectivity

measured for light coming from a single incidence direction and for light reflected into the

whole half space in front of the surface, like measurements performed with an integrating

sphere). The directional emissivity e0k1,k2 in the wavelength range (k1, k2) is related to the

spectral directional emissivity e0 (k) by the equation

e0k1;k2 ¼

ðk2

k1

eðkÞ � BðkÞdk

ðk2

k1

BðkÞdk

; (2)

where B(k) is the blackbody spectral radiance.

The total solar directional absorptivity a0S of the sample is calculated integrating the spec-

tral directional absorptivity a0(k) on the Sun emission spectrum S(k), according to the equation

a0S ¼

ðkmax

kmin

ð1� q0 \ðkÞÞ � SðkÞdk

ðkmax

kmin

SðkÞdk

; (3)

TABLE III. Room-temperature spectral cutoff wavelengths, corresponding to a reflectivity of

half of the maximum value.

Sample Cutoff (lm)

ZC99 1.61

ZCM100 1.11

ZCM96 1.42

HC67 3.47

HCT85 1.56

HCM95 1.08

HCM100 N.A.

TC76 0.62

TC92 0.56

TCT97 0.54

TCM100 0.54

SC100 Not applicable

033104-10 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

Hence, we can conclude that the measurement of spectral reflectivity is the starting point

to estimate both sunlight absorption and thermal emission characteristics of samples.

It should be noticed that, for a precise assessment of the material potential in solar furna-

ces, the characterization of the spectral behavior must be done at the furnace operating tempera-

ture (that is expected to be lower than 1400 K: notice that the materials requirements for volu-

metric solar absorber are a resistance to temperatures of about 1300 K).40 Since high-

temperature reflectivity or emissivity measurements require the use of complex apparatuses and

long measurement times,41–48 a preliminary choice among the various materials to identify the

most promising ones is absolutely mandatory, and this was the purpose of the present work. In

fact, literature data on many different materials (metals,49 semiconductors,50 dielectrics51) show

that if the material does not undergo phase or surface changes, the spectral reflectivity or emis-

sivity curve shapes do not dramatically change. In this case, the curves showed changes in in-

tensity while retaining the qualitative shape of the room temperature curve. Therefore, S-shaped

curves do not translate into band structures and conversely, even if, for curves of the former

type, a change in the cutoff spectral position can be expected.15–17

From the acquired hemispherical reflectance spectra, we calculated the room-temperature

solar directional absorptivity a0S according to Eq. (3), using as S(k) the CIE Sun spectrum with

air mass¼ 1.5 (Ref. 52) and kmin¼ 0.3 lm, kmax¼ 2.3 lm. The calculated values of a0S are

listed in Table IV. TaC-based materials show the lowest solar absorptivity, with similar values

lying around 0.4 for all samples, independently from sample porosity and sintering aid used

during synthesis. ZrC-based samples have a slightly higher solar absorptivity (somewhat higher

than 0.5), with 0.59 the maximum value obtained for the pure dense sample ZC99 and with no

differences between the two other investigated samples (both being nearly fully dense and with

MoSi2 addition). HfC-based materials show a higher spread among the investigated samples.

Among UHTCs, the best solar absorber is HC67, with a solar hemispherical absorptivity near

to 0.9, even higher than that of silicon carbide (SC100, about 0.8). From these calculations,

even if they were performed on room-temperature spectra as explained, we can conclude that

for UTCHs, the spectral features in the wavelength region of solar spectrum strongly depend on

the base material, with practically no dependence on the sintering aid (except for the HfC case

that needs additional investigations). As for the effect of sample porosity, it seems that solar

absorptivity is higher if the sample density is lower than about 70%, with no absorptivity de-

pendence on porosity for density values in the 90%–100% range (10%–0% porosities). Even if

it seems reasonable to expect some light trapping effects within the surface pores, the study of

optical properties on a higher range of porosity levels is needed to obtain conclusive results

about this effect, so it will be the subject of further investigations.

TABLE IV. Calculated directional solar absorptivity and thermal emissivity.

Sample a0S e0(0.3–25lm)

ZC99 0.59 0.23

ZCM100 0.56 0.22

ZCM96 0.56 0.31

HC67 0.88 0.55

HCT85 0.63 0.30

HCM95 0.51 0.19

HCM100 N.A. N.A.

TC76 0.48 0.30

TC92 0.39 0.16

TCT97 0.38 0.14

TCM100 0.40 0.17

SC100 0.78 0.80

033104-11 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

When the comparison between SiC and carbide UHTCs is concerned, SiC appears a better solar

absorber owing to its lower VIS-NIR reflectivity. UHTCs are expected to show a lower thermal emis-

sivity than silicon carbide, thanks to their higher reflectivity in the medium-infrared region. Thus to

have also a qualitative estimation of the expected thermal emissivity of samples, we calculated from

the acquired spectra the directional emissivity in the wavelength range 0.3–25 lm using Eq. (2). In

the calculation, we considered as B(k) the blackbody spectrum at 1200 K. Results are summarized in

Table IV. The lowest emissivity (lower than 0.2) is shown by TaC-based dense materials, while the

majority of ZrC-based samples show values around 0.2. Like in the solar absorptivity calculation,

HfC-based samples show a spread of emissivity values, ranging from about 0.2 to 0.3 for dense pel-

lets and about 0.5 for the porous one. In general, samples with a significant porosity have a consider-

ably larger thermal emissivity than similar dense materials (see TC76 and HC67 cases), as expected.

SiC has by far the largest thermal emissivity among the materials under study, corresponding to a

a0/e0 ratio of about 1. These qualitative considerations are confirmed by recent high-temperature

measurements reporting for ZrC, TaC, and HfC a considerable lower emissivity than SiC.53,54 It

should be noticed that for solar receiver applications, a low thermal emissivity is a very important pa-

rameter, as it governs radiative thermal losses in the heated absorber and thus the maximum allowable

operating temperature. Moreover UHTCs have satisfactory thermo-mechanical properties at high

temperature (equivalently to SiC). On the other hand, possible concerns for the use of carbides for

this application are poor resistance to oxidation, which implies their use under vacuum or inert atmos-

phere, and current high costs of raw materials, especially for HfC powders.

From these very preliminary considerations, it appears clear that the quantitative assess-

ment of UHTC performances for this new solar application is a very complex task that includes

the evaluation of several parameters. In this work, we give a basic evaluation of the spectrally

selective characteristics of UHTCs, together with the identification of the most relevant parame-

ters to take into account for the proposed application. The final identification of the best per-

forming material(s) necessarily needs complex high temperature measurements and a choice

between different tradeoffs. When the comparison between UHTCs and SiC is concerned, it

should be noticed that although thermo-mechanical requirements are equivalently satisfied by

both classes of ceramics, the more favorable optical characteristics and above all the low emis-

sivity at high temperatures of ultra-refractory carbides seem to be the most important advantage

with respect to SiC that has been evidenced by this study.

V. CONCLUSIONS

This work examined the intrinsic spectral selectivity of different carbide samples of

UHTCs for new applications as sunlight absorbers in tower solar plants. The investigation has

been performed in the wavelength range from 0.25 to 25 lm, which allows showing a high

intrinsic spectral selectivity for the majority of tested samples, as they reach a low reflectivity

at the wavelengths of solar emission and a high reflectivity at the wavelengths of thermal infra-

red. Directional solar absorptivity has been calculated for most of them, showing that, for a

given carbide type, the use of sintering aids does not affect the solar absorber features. Never-

theless, the various investigated carbide families showed distinctive features in the UV-VIS

spectra that result in different performances in the field of sunlight absorption. Among the

investigated materials, the highest value of solar absorptivity has been measured for hafnium

carbide-based porous samples, while the lowest corresponds to dense tantalum carbide. Direc-

tional emissivity in the wavelength range 0.3–25 lm has been calculated as well, allowing to

expect from UHTCs a considerably lower emissivity with respect to the reference SiC ceramics.

These results will be cross-compared with high temperature emissivity investigations for a

quantitative assessment of UHTC potentialities and performances as solar absorbers and solar

energy storage materials.

ACKNOWLEDGMENTS

The authors thank Dr. Franco Francini for useful discussions about optical properties of

UHTCs, R. Renis for samples preparation, and C. Melandri for surface roughness measurements.

033104-12 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions

1H. Yuncu, E. Paykoc, and Y. Yener, Solar Energy Utilization (Kluwer Academic, Boston, 1987).2Y. Wenhua, D. M. France, J. L. Routbort, and S. Choi, Heat Transfer Eng. 29, 432 (2008).3E. Sani, S. Barison, C. Pagura et al., Opt. Express 18, 5179 (2010).4S. A. Kalogirou, Prog. Energy Combust. Sci. 30, 231 (2004).5T. Yan and Y. Chen, Renewable Sustainable Energy Rev. 14, 265 (2010).6T. Fend, R. Pitz-Paal, O. Reutter, J. Bauer, and B. Hoffschmidt, Sol. Energy Mater. Sol. Cells 84, 291 (2004).7M. Romero, R. Buck, and J. E. Pacheco, J. Sol. Energy Eng. 124, 98 (2002).8S. R. Levine, E. J. Opila, M. C. Halbig, J. D. Kiser, M. Singh, and J. A. Salem, J. Eur. Ceram. Soc. 22, 2757 (2002).9M. B. Dickerson, P. J. Wurm, J. R. Schorr, W. P. Hoffman, P. G. Wapner, and K. H. Sandhage, J. Mater. Sci. 39, 6005(2004).

10F. Monteverde and R. Savino, J. Eur. Ceram. Soc. 27, 4797 (2007).11R. Borrelli, A. Riccio, D. Tescione, R. Gardi, and G. Marino, Acta Astronaut. 65, 442 (2009).12W. F. Bogaerts and C. M. Lampert, J. Mater. Sci. 18, 2847 (1983).13C. E. Kennedy, “Review of mid- to high-temperature solar selective absorber materials,” National Renewable Energy

Laboratory Report No. NREL/TP-520-31267, 2002.14C. M. Lampert, Sol. Wind Technol. 4, 347 (1987).15P. T. B. Shaffer, “Development of ultra refractory materials,” Progress Report No. 25, Carborundum Co., Niagara Falls,

New York, 1 May-31 July 1961, 16 August 1961 (Contract NOrd-17175) ASTIA AD-261 471.16G. M. Kibler, T. F. Lyon, M. J. Linevski, and V. J. Desantis, “Carbonization of plastics and refractory material research,”

Vol. II, Technical Report No. WADD-TR-60-646 Part III Vol. II, 1964.17G. M. Kibler, T. F. Lyon, M. J. Linevski, and V. J. Desantis, “Refractory material research,” Technical Report No.

WADD-TR-60-646 Part IV, 1964.18D. P. DeWitt, T. K. Riddle, and R. E. Taylor, “Spectral and total emissivity and reflectivity at high temperatures,” Report

No. PRL 152, School of Mechanical Engineering, Purdue University, West Lafayette, Indiana, 1978.19W. H. Askwyth, R. J. Hayes, R. D. House, and G. Mikk, “Determination of emissivity of materials,” Volume I Interim

final report, Jul. 1, 1959-Dec. 31, 1962, Report No. PWA-2206, Pratt & Whitney Aircraft, East Hartford, Connecticut,1962.

20K. E. Storms, The Refractory Carbides (Academic, New York, 1967).21B. Karlsson, J. E. Sundgren, and B. O. Johansson, Thin Solid Films 87, 181 (1982).22B. Karlsson, R. P. Shimshock, B. O. Seraphin, and J. C. Haygarth, Sol. Energy Mater. 7, 401 (1983).23L. Roux, J. Hanus, J. C. Francois, and M. Sigrist, Sol. Energy Mater. 7, 299 (1982).24J. C. Francois, G. Chassaing, P. Gravier, R. Pierrisnard, and A. M. Bonnot, Thin Solid Films 127, 205 (1985).25J. Spitz, Thin Solid Films 45, 31 (1977).26F. A. Modine, R. W. Major, T. W. Haywood, and G. R. Gruzalski, Phys. Rev. B 29, 836 (1984).27J. F. Alward, C. Y. Fong, M. El-Batanouny, and F. Wooten, Phys. Rev. B 12, 1105 (1975).28F. A. Modine, T. W. Haywood, and C. Y. Allison, Phys. Rev. B 32, 7743 (1985).29J. G. J. Peelen and R. Metselaar, J. Appl. Phys. 45, 216 (1974).30C. C. Agrafiotis, I. Mavroidis, A. G. Kostandopoulos, B. Hoffschmidt, P. Stobbe, and W. Fernandez-Quero, Sol. Energy

Mater. Sol. Cells 91, 474 (2007).31D. Sciti, L. Silvestroni, S. Guicciardi, D. Delle Fabbriche, and A. Bellosi, J. Mater. Res. 24, 2056 (2009).32D. Sciti, S. Guicciardi, and M. Nygren, Scr. Mater. 59, 638 (2008).33L. Silvestroni and D. Sciti, J. Mater. Res. 23, 1882 (2008).34D. Sciti, L. Silvestroni, and A. Bellosi, J. Am. Ceram. Soc. 89, 2668 (2006).35S. Dutta, “Densification and properties of a-silicon carbide,” J. Am. Ceram. Soc. 68, C-269 (1985).36See http://www.cvimellesgriot.com for CVI-Melles-Griot catalogue of supplied optical components.37L. E. Toth, “Transition metal carbides and nitrides,” in Refractory Materials, A Series of Monographs, edited by J. L.

Margrave (Academic, New York, 1971), p. 6–10.38D. Sciti and A. Bellosi, J. Mater. Sci. 35, 3849 (2000).39K. L. Luthra, J. Am. Ceram. Soc. 74, 1095 (1991).40T. Fend, B. Hoffschmidt, O. Reutter, J. Sauerhering, and R. Pitz-Paal, in Proceedings of ASME ICNMM2006, 4th Interna-

tional Conference on Nanochannels, Microchannels and Minichannels (ASME, 2006), p. 1–8.41M. Balat-Pichelin, D. Hernandez, G. Olalde, B. Rivoire, and J. F. Robert, J. Sol. Energy Eng. 124, 215 (2002).42J. Randolph, J. Ayon, R. Dirling et al., Carbon 37, 1731 (1999).43T. Paulmier, M. Balat-Pichelin, D. Le Queau, R. Berjoan, and J. F. Robert, Appl. Surf. Sci. 180, 227 (2001).44T. Paulmier, M. Balat-Pichelin, and D. Le Queau, Appl. Surf. Sci. 243, 376 (2005).45M. Balat-Pichelin, J. F. Robert, and J. L. Sans, Appl. Surf. Sci. 253, 778 (2006).46L. Scatteia, R. Borrelli, G. Cosentino, E. Beche, J. L. Sans, and M. Balat-Pichelin, J. Spacecr. Rockets 43, 1004 (2006).47L. Scatteia, D. Alfano, F. Monteverde, J. L. Sans, and M. Balat-Pichelin, J. Am. Ceram. Soc. 91, 1461 (2008).48D. Alfano, L. Scatteia, S. Cantoni, and M. Balat-Pichelin, J. Eur. Ceram. Soc. 29, 2045 (2009).49J. E. Janssen and R. H. Torborg, Measurement of Thermal Radiation Properties of Solids, edited by J. C. Richmond

(National Aerospace and Space Administration, Washington DC, 1963).50D. L. Stierwalt, Appl. Opt. 5, 1911 (1966).51G. E. Guazzoni, Appl. Spectrosc. 26, 60 (1972).52CIE (Commission Internationale de l’Eclairage), “Solar spectral irradiance,” Technical Report No. 85, 1989.53E. Sani, L. Mercatelli, F. Francini, J. L. Sans, and D. Sciti, Scr. Mater. 65, 775 (2011).54E. Sani, L. Mercatelli, D. Fontani, J.-L. Sans, and D. Sciti, J. Renewable Sustainable Energy 3, 063107 (2011).

033104-13 Sani et al. J. Renewable Sustainable Energy 4, 033104 (2012)

Downloaded 16 May 2012 to 193.205.148.190. Redistribution subject to AIP license or copyright; see http://jrse.aip.org/about/rights_and_permissions