Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells

journal homepage: www.elsevier.com/locate/solmat

Spectral hemispherical reflectivity of nonstoichiometric cerium dioxide ^

Simon Ackermann, Aldo Steinfeld *

Department of Mechanical and Process Engineering, ETH Zürich, Sonneggstrasse 3, 8092 Zürich, Switzerland

CrossMark

ARTICLE INFO

Article history: Received 23 August 2016 Accepted 31 August 2016

Keywords: Ceria

Nonstoichiometry Reflectivity Spectroscopy Solar radiation

ABSTRACT

Nonstoichiometric ceria, CeO2-S, has emerged as a promising redox material for thermochemically splitting H2O and CO2 using concentrated solar energy. Knowledge of its radiative properties is crucial for the design of efficient solar reactors. Samples of various nonstoichiometries (0 r S r 0.0377) were prepared by thermal reduction in a thermogravimetric analyzer at high temperatures (T > 1473 K) and under low oxygen partial pressures (pO2 r 2.5 • 10atm). The spectral hemispherical reflectivity was measured using a spectroscopic goniometry system in the spectral range 300-2800 nm. A porous ceria sample with interconnected mm-sized pores showed comparable selectivity because of its high optical thickness. The total hemispherical reflectivity was computed for emission temperatures in the range 900-6000 K relevant to solar reactors.

© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license

(http://creativecommons.org/licenses/by/4.0/).

1. Introduction

Using concentrated solar radiation, thermochemical redox cycles based on nonstoichiometric metal oxides are capable of splitting H2O and CO2 to produce H2 and CO (syngas), the precursor for the synthesis of conventional transportation liquid fuels [1]. Ceria (CeO2) is currently considered the state-of-the-art redox material due to its favorable thermodynamics, rapid kinetics, and morphological stability over a wide range of temperatures [2-14]. The two-step thermochemical cycle based on nonstoichiometric ceria is represented by:

Reduction at Tred:

-CeO2_S ^

— s 2 Sox

îd aox

Oxidation at Tox: 1

2-sred

2-sred '

2-sred

where Sox and 8red are the oxygen nonstoichiometries before and after reduction, respectively. In the first step, Eq. (1), ceria is en-dothermally reduced in an atmosphere of low oxygen partial pressure, pO2, and at elevated temperatures, typically

* Corresponding author. E-mail address: aldo.steinfeld@ethz.ch (A. Steinfeld).

Tred > 1573 K, with process heat delivered by concentrated solar radiation. In the second step, Eq. (2), the reduced ceria is exo-thermally re-oxidized with H2O and/or CO2 to generate H2 and/or CO at lower temperatures, typically Tox < 1573 K. 8red and Sox strongly depend on Tred, Tox, and pO2. The difference 8red - Sox determines the maximum molar amount of fuel capable of being produced per cycle and per mole of ceria.

The design and optimization of solar chemical reactors for effecting the redox cycle demands the development of numerical heat transfer models using accurate radiative properties because radiation is the dominant heat transport phenomena at Tred [15,16]. Specifically for reticulated porous ceria structures, the accurate determination of the absorbed and scattered portion of radiation requires the knowledge of the hemispherical reflectivity of ceria rCeO2 as a function of wavelength A and nonstoichiometry 8:

a(X, S) = (1 - rceo,(, S))ß

S) = rceo2(i, S)ß

where a is the absorption coefficient, s the scattering coefficient, and ¡3 the effective extinction coefficient. On the other hand, ¡3 depends solely on the morphology and can be determined by applying pore-level Monte Carlo ray tracing on the exact 3D digital representation of the porous structures obtained by computed tomography [17]. To date however, there isn't any data reporting the spectral reflectivity of ceria as a function of its nonstoichiometry and measurement data on the optical properties of ceria are scarce. Hass et al. [18] investigated the refractive index and the absorption coefficient of evaporated CeO2 films in the wavelength range 0.22-

http://dx.doi.org/10.1016/j.solmat.2016.08.036

0927-0248/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Nomenclature Greek symbols

Cl speed of light [m s-1] a absorption coefficient [m- 1]

hp Planck constant [J s] ß extinction coefficient [m- 1]

kB Boltzmann constant [J K-1] 8 nonstoichiometry [ - ]

LCeO2 detector signal for the ceria sample [V] A wavelength [nm]

Lref detector signal for the reference sample [V] s scattering coefficient [m-1]

rCeO2J spectral hemispherical reflectivity of ceria [ - ]

RCeO2 total hemispherical reflectivity of ceria [ - ] Acronym

rrefJ spectral hemispherical reflectivity of the reference

sample [ - ] TGA thermogravimetric analyzer

T temperature [K]

1.0 mm. Touloukian et al. [19] reported normal spectral reflectivity of CeO2 of electron beam deposited films and of sintered powder. Fangxin et al. [20] and Marabelli et al. [21] measured the UV-visible reflectivity of nanocrystalline ceria and studied the absorption characteristics by Kramers-Kronig transformation. Ganesan et al. [22] determined the albedo and the scattering coefficient of stoi-chiometric CeO2 by measuring the directional-hemispherical reflectance in the range of 0.35-2.0 mm.

In this work, the spectral hemispherical reflectivity of sintered polycrystalline ceria was measured for a wide range of A and 8 relevant for the solar redox cycle. An empirical correlation is least-squares fitted to the total hemispherical reflectivity as a function of 8 and the blackbody radiation temperature. Additionally, the reflectivity of a porous ceria sample with ~ 0.26 porosity and a mean pore diameter of ~ 10 mm [23] is measured to investigate the effect of the mm-sized pores.

2. Methodology

2.1. Optical setup

Measurements were performed with a goniometry spectroscopic system [24,25], shown schematically in Fig. 1. It consists of: 1) Xe-arc lamp as light source that emits radiation in the range 170-3000 nm and approximates a blackbody at 6200 K in the visible range; 2) aspherical Czerny-Turner type double mono-chromator; 3) mechanical beam chopper to modulate at a frequency of 417 Hz for minimizing background noise; 4) imaging lens; 5) iris to adjust the ray cone angle; 6) mechanical slit; 7) integrating sphere; 8) sample holder; 9) photodetector; 10) lock-in amplifier, and 11) data acquisition system.

Fig. 1. Schematic of the spectroscopic system that consists of a: 1) xenon-arc lamp, 2) double monochromator, 3) chopper, 4) imaging lens, 5) iris, 6) mechanical slit, 7) integrating sphere, 8) sample holder, 9) photodetector, 10) lock-in amplifier, 11) data acquisition system.

Each measurement is performed with a ceria sample and a reference (calibrated) sample. The spectral hemispherical reflectivity of the ceria sample, rCeOli, is then calculated by:

LCeO, _ rCeO2.l

' refU

where LCeO2 and Lref are the detector signals for the ceria and reference samples, respectively. The spectral hemispherical reflectivity of the reference sample, rrefJ, is given by the manufacturer in the range of 250 < A < 2800 nm (Labsphere AS-01161-060, SphereOptics Zenith Polymer Diffuse Reflectance Standard - 50%).

2.2. Photodetectors

Three types of photodetectors were used to cover the wide range of wavelengths of interest: a SiC photodiode, a InGaAs photodiode and a PbS photoconductor covered the ranges 2501000 nm, 800-1700 nm and 1000-2800 nm, respectively, and 20, 50 and 100 consecutive measurements were performed at each wavelength (50 nm steps), respectively. An increasing number of consecutive measurements was performed at longer A to minimize the measurement error based on background noise because, for A > 1200 nm, the Xe-arc has a decreasing emission with wavelength leading to weak detector signals. The accuracy of the measurement was 95%, determined with the reference sample.

2.3. Ceria samples

Disk-shaped ceria pellets (23 mm-dia., 4 mm-height) were manufactured by pouring a slurry of Cerium (IV)-oxide (99.9%

Fig. 2. Nonstoichiometry (solid) and temperature (dashed) of ceria as a function of time during a TGA run.

Fig. 3. SEM of the ceria pellets: a) non-porous; and b) porous with 0.26 porosity and 10 mm mean pore diameter.

purity, grain size 5-40 mm, Sigma Aldrich) into a mold and sintering at 1873 K for several hours. The ceria slurry and heating rates were the same as the ones used for the production of reticulated porous ceramic structures used in the solar reactor [26] for the purpose of manufacturing pellets with equivalent surface conditions. The pellets were reduced in a thermogravimetric analyzer (TGA, NETZSCH SD 409) at Tred > 1473 K and pO2 < 2.5 • 10-4 atm. Fig. 2 shows the nonstoichiometry (left y-axis) and the temperature (right y-axis) as a function of the time during a TGA run. The ceria samples were heated to the desired temperature, isothermally reduced for 2 h, and then cooled and kept in a closed dry container. No re-oxidation was observed, even under ambient conditions over several weeks. The final non-stoichiometry was determined by weight measurements before and after the TGA runs (Mettler Toledo XS105 dual range balance) and verified with the weight measurements obtained with the TGA. Based on the O2 concentration measured downstream by gas chromatography (CP-4900 System, Agilent Technologies), the samples reached thermodynamic equilibrium during reduction.

Additionally, porous ceria pellets were manufactured by mixing carbon pore-forming agent particles (particle size 0.4-12 mm, HTW Hochtemperatur-Werkstoffe GmbH) in the ceria slurry to obtain an interconnected pore network with 0.26 porosity and 10 mm mean pore diameter after sintering. Fig. 3a and b show a scanning electron micrograph of the non-porous and porous stoichiometric (5=0) ceria pellets, respectively. The grain size for both samples was around 5-40 mm. The porous ceria pellets were used to investigate the effect of the mm-size pores on the hemispherical reflectivity.

Fig. 4. Mean value of the spectral hemispherical reflectivity of ceria with 8=0.0217. The error bars represent the standard deviation.

3. Results and discussion

Fig. 4 shows the mean value of rCeOia as a function of A for a ceria sample with 8= 0.0217. The error bars represent the standard deviation at 68.2% confidence error band for each wavelength. The accuracy of the mean value of rCeO2 i was + 0.02 for A r 1500 nm and + 0.05 for A > 1500 nm.

Fig. 5a shows the mean value of rc

CeO2,l

as a function of À for

various 5 in the range 0-0.0377. For 5=0, rCeOil increases sharply from 0.3 to 0.9 with increasing A throughout the visible spectrum and remained relatively constant around 0.9 for A > 1000 nm. This trend is not observed for reduced ceria. Instead, rCeOia increases monotonically over the IR spectrum, reaches a plateau around the 1700-2200 nm band, and decreases for A> 2200 nm. For the entire range of wavelengths considered, rCeOli decreases and the sample selectivity diminishes with increasing 5. Note that the colors of each curve shown in Fig. 5a were calculated with the red-green-blue combination of the measured reflectivities and represent the actual sample color. Fig. 5b shows rCeOl i vs. A for a non-porous and a porous stoichiometric ceria sample. For both samples, rCeOia showed similar values and selectivity. Thus, the effect of ^m-size porosity was negligible because of the high optical thickness of the porous sample. Its extinction coefficient was estimated to be 150,000 m-1 by applying pore-level Monte Carlo ray tracing [17] on a computer tomography scan of the mm-size morphology [23], resulting in an optical thickness of 4.6 and an attenuation of 99% of the incident radiation within the first 30 mm.

The total hemispherical reflectivity RCeO2 was computed by spectral integration of rCeOia weighted by Planck's blackbody radiation:

2nhpcl2rceO2,l(')s Jl=250nm ' u ~ 4

WT)L = ■

hpci eikBT - 1

Л = 2

hpci eikßT - 1

where hp is the Planck constant, kB the Boltzmann constant, c1 the speed of light, and T is the blackbody temperature in the range 900-6000 K. Note that blackbody radiation at T = 5780 K

Fig. 5. Spectral hemispherical reflectivity of sintered polycrystalline ceria pellets as a function of the wavelength for: a) non-porous for various nonstoichiometries; and b) non-porous and porous for 8=0. (For interpretation of the references to color in this figure, the reader is referred to the web version of this article.)

Table 1

Coefficients of the least-squared parameters a, b, and c of Eq. (7) given by a 3rd-order polynomial function f(T) = f + fjT1 + f2T2 + fjT3.

Parameter

Fig. 6. Total hemispherical reflectivity of ceria as a function of: a) blackbody radiation temperature for various nonstoichiometries; b) nonstoichiometry for various blackbody radiation temperatures. Markers are calculated by the exact equation, Eq. (6). Solid lines are calculated by the empirical correlation, Eq. (7).

square-fitted correlation as a function of 8 and T given by Eq. (7), with fitting parameters a, b and c given by 3rd-order polynomial functions of T listed in Table 1. e=10-10 to guarantee continuity of the correlation for 8=0. The normalized root-mean-square error is 1.36%. Fig. 6a shows RCeO2 vs. T for various 8 in the range 0-0.0377. Fig. 6b is a cross plot: it shows RCeO2 vs. 8 for various T in the range 1773-5780 K. Markers are calculated by the exact equation, Eq. (6). Solid lines are calculated by the empirical correlation, Eq. (7).

fo fl f2 f3

2.86880 • 10-02 -1.61441 • 10 -07 1.25380 • 10-9 -1.93330 • 10 -

4.25902 • 10 -01 3.55510 • 10-05 - 2.31996 • 10-08 2.37842 • 10 -

-1.83738 -1.45065 • 10 -03 4.85473 • 10 -07 - 4.01102 • 10 -

CeO#. T) =

(S + e)

+ C •

represents a good approximation of solar radiation. Spectral integration was performed with 50 nm steps up to a wavelength of 50 mm by extrapolating rCeO2,j with a mean value taken in the band 1500-2800 nm for each 8. As expected, RCeO2 decreases with T and 8 as a result of rCeO2,j decreasing with A and 8. Since RCeO2 is used in heat transfer modeling, it is convenient to have an empirical least-

4. Summary and conclusion

We have measured the spectral hemispherical reflectivity of sintered polycrystalline ceria for a wide range of non-stoichiometries and wavelengths. For stoichiometric ceria (8=0), rCeO2,j increases sharply from 0.3 to 0.9 with A throughout the visible spectrum and remained relatively constant around 0.9 for A > 1000 nm. A porous sample with mm-sized interconnected

pores exhibits similar values because of its high optical thickness. In contrast for nonstoichiometric ceria, rCeOia increases mono-tonically over the IR spectrum, reaches a plateau around the 17002200 nm band, and decreases for A> 2200 nm. As 5 increases, the value of rCeOia decreases and the selectivity diminishes. The total hemispherical reflectivity, computed by applying Planck's law, decreases with T and 5. An empirical correlation was derived for the purpose of heat transfer modeling of solar reactors.

Acknowledgments

We gratefully acknowledge the financial support by the Swiss Federal Office of Energy (Grant SI/501213-01) and the European Research Council under the European Union's ERC Advanced Grant (Project SUNFUELS - No. 320541).

References

[1] M. Romero, A. Steinfeld, Concentrating solar thermal power and thermo-chemical fuels, Energy Environ. Sci. 5 (2012) 9234-9245.

[2] S. Abanades, G. Flamant, Thermochemical hydrogen production from a two-step solar-driven water-splitting cycle based on cerium oxides, Sol. Energy 80 (2006) 1611-1623.

[3] S. Abanades, A. Legal, A. Cordier, G. Peraudeau, G. Flamant, A. Julbe, Investigation of reactive cerium-based oxides for H2 production by thermo-chemical two-step water-splitting, J. Mater. Sci. 45 (2010) 4163-4173.

[4] W.C. Chueh, C. Falter, M. Abbott, D. Scipio, P. Furler, S.M. Haile, A. Steinfeld, High-flux solar-driven thermochemical dissociation of CO2 and H2O using nonstoichiometric ceria, Science 330 (2010) 1797-1801.

[5] W.C. Chueh, S.M. Haile, A thermochemical study of ceria: exploiting an old material for new modes of energy conversion and CO2 mitigation, Philos. Trans. R. Soc. A 368 (2010) 3269-3294.

[6] P. Furler, J.R. Scheffe, A. Steinfeld, Syngas production by simultaneous splitting of H2O and CO2 via ceria redox reactions in a high-temperature solar reactor, Energy Environ. Sci. 5 (2012) 6098-6103.

[7] L.J. Venstrom, N. Petkovich, S. Rudisill, A. Stein, J.H. Davidson, The effects of morphology on the oxidation of ceria by water and carbon dioxide, J. Sol. Energy Eng. 134 (2012) 011005.

[8] R. Bader, L.J. Venstrom, J.H. Davidson, W. Lipinski, Thermodynamic analysis of isothermal redox cycling of ceria for solar fuel production, Energy Fuels 27 (2013) 5533-5544.

[9] V. Esposito, D.W. Ni, Z. He, W. Zhang, A.S. Prasad, J.A. Glasscock,

C. Chatzichristodoulou, S. Ramousse, A. Kaiser, Enhanced mass diffusion phenomena in highly defective doped ceria, Acta Mater. 61 (2013) 6290-6300.

10] S. Ackermann, J.R. Scheffe, A. Steinfeld, Diffusion of oxygen in ceria at elevated temperatures and its application to H2O/CO2 splitting thermochemical redox cycles, J. Phys. Chem. C. 118 (2014) 5216-5225.

11] J.E. Miller, A.H. McDaniel, M.D. Allendorf, Considerations in the design of materials for solar-driven fuel production using metal-oxide thermochemical cycles, Adv. Energy Mater. 4 (2014), 1300469-1300469-19.

12] J.R. Scheffe, A. Steinfeld, Oxygen exchange materials for solar thermochemical splitting of H2O and CO2: a review, Mater. Today 17 (2014) 341-348.

13] B. Bulfin, F. Call, M. Lange, O. Lubben, C. Sattler, R. Pitz-Paal, I.V. Shvets, Thermodynamics of CeO2 thermochemical fuel production, Energy Fuels 29 (2015) 1001-1009.

14] N. Knoblauch, L. Dorrer, P. Fielitz, M. Schmuecker, G. Borchardt, Surface controlled reduction kinetics of nominally undoped polycrystalline CeO2, Phys. Chem. Chem. Phys. 17 (2015) 5849-5860.

15] S. Haussener, A. Steinfeld, Effective heat and mass transport properties of anisotropic porous ceria for solar thermochemical fuel generation, Materials 5 (2012) 192-209.

16] P. Furler, A. Steinfeld, Heat transfer and fluid flow analysis of a 4 kw solar thermochemical reactor for ceria redox cycling, Chem. Eng. Sci. 137 (2015) 373-383.

17] J. Petrasch, P. Wyss, A. Steinfeld, Tomography-based monte carlo determination of radiative properties of reticulate porous ceramics, J. Quant. Spectrosc. Radiat. Transf. 105 (2007) 180-197.

18] G. Hass, J. Ramsey, R. Thun, Optical Properties and Structure of Cerium Dioxide Films, JOSA 48 (1958) 324-326.

19] Y.S. Touloukian, D.P. DeWitt, Thermal Radiative Properties - Nonmetallic Solids, IFI/Plenum, New York, 1972.

20] L. Fangxin, W. Chengyun, S. Qingde, Z. Tianpeng, Z. Guiwen, Optical properties of nanocrystalline ceria, Appl. Opt. 36 (1997) 2796-2798.

21] F. Marabelli, P. Wachter, Covalent insulator CeO2: optical reflectivity measurements, Phys. Rev. B 36 (1987) 1238-1243.

22] K. Ganesan, L.A. Dombrovsky, W. Lipinski, Visible and near-infrared optical properties of ceria ceramics, Infrared Phys. Technol. 57 (2013) 101-109.

23] S. Ackermann, J.R. Scheffe, J. Duss, A. Steinfeld, Morphological characterization and effective thermal conductivity of dual-scale reticulated porous structures, Materials 7 (2014) 7173-7195.

24] P. Coray, W. Lipinski, A. Steinfeld, Spectroscopic goniometry system for determining thermal radiative properties of participating media, Exp. Heat. Transf. 24 (2011) 300-312.

25] P. Good, T. Cooper, M. Querci, N. Wiik, G. Ambrosetti, A. Steinfeld, Spectral reflectance, transmittance, and angular scattering of materials for solar concentrators, Sol. Energy Mater. Sol. Cells 144 (2016) 509-522.

26] P. Furler, J. Scheffe, D. Marxer, M. Gorbar, A. Bonk, U. Vogt, A. Steinfeld, Thermochemical CO2 splitting via redox cycling of ceria reticulated foam structures with dual-scale porosities, Phys. Chem. Chem. Phys. 16 (2014) 10503-10511.