Article

pubs.acs.org/EF

Thermodynamic Analysis of Syngas Production via the Solar

Thermochemical Cerium Oxide Redox Cycle with Methane-Driven
Reduction
Peter T. Krenzke and Jane H. Davidson*
Department of Mechanical Engineering, University of Minnesota, Minneapolis, Minnesota 55455, United States

ABSTRACT: The present work considers the convergence of two approaches for syngas production: solar fuels via the cerium
oxide (ceria) redox cycle and the partial oxidation of methane. The chemical thermodynamics of the ceriamethane system
reveal that coupling the reduction of ceria to the partial oxidation of methane enables isothermal cycling at temperatures as low as
1223 K with the additional production of high-quality syngas during the reduction step. The equilibrium non-stoichiometry of
the oxidation step has a substantial impact on the conversion of the oxidizer to fuel, with important implications for cycle
eciency. A model of the process thermodynamics is used to evaluate the eciency of the cycle and its sensitivity to oxidation
non-stoichiometry, temperature, and concentration ratio. Reduction with methane enables signicant gains in eciency over
other proposed approaches, with plausible solar-to-fuel eciencies reaching 40% without any heat recovery.

1. INTRODUCTION
Concentrated solar power is a sustainable source of high-
temperature process heat, enabling the solar-driven operation
of many useful endothermic processes. Of particular interest is
the storage of solar energy in chemical bonds via the splitting of
water and carbon dioxide to produce hydrogen and carbon
monoxide, referred to collectively as syngas. Syngas and its
constituents are valuable feedstocks used for production of
ammonia, methanol, dimethyl ether, or FischerTropsch
products.1,2 The versatility of syngas has motivated the study
of various pathways for converting methane to syngas,
especially variants of natural gas oxyforming,3 including steam
reforming,4 dry reforming with carbon dioxide,1 and partial
oxidation.5 The present work considers the convergence of two
approaches for syngas production: solar fuels via the cerium
oxide (ceria) redox cycle and the partial oxidation of methane.
In comparison to the solar-only ceria cycle, this two-prong
approach lowers the required reduction temperature, which
allows for operation at a lower solar concentration ratio and
increases the eciency of conversion of sunlight to fuel.
Of the metal oxide cycles, the ceria cycle is particularly
attractive because ceria retains a stable cubic uorite crystal
structure for large non-stoichiometries, < 0.25.6 Figure 1 is a
conceptual sketch of the solar-only ceria redox cycle.
Concentrated sunlight provides process heat to reduce ceria
(R1). Ceria is then reoxidized when H2O and CO2 are split to
produce H2 and CO (R2).
CeO2 ox CeO2 red + /2O2 (R1)
CeO2 red + CO2
CeO2 ox + CO + ( 1)CO2 (R2a)
Figure 1. In the ceria redox cycle, concentrated sunlight drives the
endothermic reduction of ceria, releasing O2 and increasing the non-
stoichiometry from ox to red. The reduced ceria is used to split H2O
and CO2 to produce H2 and CO in the exothermic oxidation reaction.
The enthalpy change of reactions R1 and R2 varies with the
change in the non-stoichiometry of ceria between the reduced
and oxidized states, red ox. The gravimetric
production of CO and H2 is directly proportional to . The
oxidation step requires the use of excess oxidizer, which is
quantied by the stoichiometric oxidizer ratio
nox,in
n fuel (3)
dened as the molar ratio of oxidizer supplied to fuel produced
in reaction R2. For > 1, the oxidation products contain a
mixture of CO and H2 and unreacted oxidizer.
Solar-only redox cycling of ceria has been studied over a
range of conditions, generally categorized as either a two-
temperature approach, where reduction is carried out at a
higher temperature and lower oxygen partial pressure than
oxidation,713 or an isothermal approach, where reduction and
oxidation temperatures are equal or near isothermal when
oxidation is only a few hundred degrees (K) lower than

CeO2 red + H 2O Received: March 18, 2014

Revised: May 15, 2014
CeO2 ox + H 2 + ( 1) H 2O (R2b) Published: May 16, 2014

2014 American Chemical Society 4088 dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095
Energy & Fuels
reduction. In the latter approach, the dierence in oxygen
partial pressure between reduction and oxidation provides the
driving force for the change in non-stoichiometry and, thus,
must be greater than for the two-temperature approach.1416
Isothermal operation of the solar-only cycle obviates the need
for solid-phase heat recuperation at the expense of very large
stoichiometric oxidizer ratios, > 100 at 1773 K, and more
demanding heat recovery requirements.17 Both approaches
require reduction temperatures above 1623 K to reach adequate
red to enable meaningful fuel production. The primary
methods proposed for attaining low oxygen partial pressure
during reduction are the ow of an inert sweep gas,9,16 which
introduces signicant energy requirements for purifying and
heating the sweep gas, or operation at reduced total pressure,13
which also imposes an energy penalty and may prove dicult to
implement at high temperature. Predicted solar-to-fuel
eciencies for reactors carrying out the solar-only redox cycle
vary depending upon the assumptions but are optimistically on
the order of 1020%,9,13,16 where eciency is
n fuel
HHVfuel
Q sol (4)
Prior experimental studies have demonstrated that excess
oxidizer is necessary,6,14,18,19 and thus, energy not accounted for
in eq 4 will be required to separate hydrogen or carbon
monoxide from the oxidant in the product stream. The solar-
only ceria cycle has been demonstrated in prototype solar
reactors using a variety of ceria morphologies, with reported
cycle eciencies reaching 1.7%.1012
In the present work, we propose coupling the partial
oxidation of methane to the solar-driven reduction step of the
ceria redox cycle to attain high solar-to-fuel eciency. This
coupled option lowers the required temperature for reduction,
thus permitting ecient operation of a solar reactor at a lower
solar concentration ratio and under isothermal (or near
isothermal) conditions (i.e., reduction and oxidation are carried
out at the same temperature). The approach was acknowledged
by Abanades and Flamant,7 when they rst proposed the solar-
driven ceria cycle for hydrogen production, but no prior work
has considered the thermodynamics of the cycle. Carbothermal
reduction with methane is a sustainable pathway for syngas
production when biogenic methane is used as the reductant. An
analysis of the chemical and process thermodynamics of the
ceriamethane system are used to evaluate the proposed
approach, and a comparison is given to solar-only approaches
to the ceria cycle.
The proposed coupled cycle is shown in Figure 2, in which
reduction occurs according to reaction R5.
Figure 2. Ceria redox cycle with ceria reduction coupled to the partial
oxidation of methane. In addition to the fuel produced by H2O and
CO2 splitting during the oxidation reaction, high-quality syngas is
produced from CH4 partial oxidation in the ceria reduction reaction.
4089
Article
CeO2 ox + CH4 CeO2 red + 2 H 2 + CO
(R5)
The evolved oxygen is used in the partial oxidation of methane,
reaction R6, rather than simply being removed.
CH4 + 1/2O2 2H 2 + CO (R6)
Although the partial oxidation of methane (R6) is exothermic,
reduction (R5) is endothermic, allowing the solar energy stored
in the reduced ceria to produce fuel in the same oxidation step,
(R2) as the cycle in Figure 1. Unlike conventional processes for
converting methane to syngas, which combust a portion of the
feedstock, the proposed cycle produces syngas with a larger
heating value than the methane feedstock. Carbothermal
reduction eliminates the need for an inert sweep gas or
vacuum operation. Furthermore, in the reduction step,
additional H2 and CO are produced in a ratio that facilitates
conversion to liquid fuels via FischerTropsch reactions or
methanol synthesis. The use of a metal oxide as the oxygen
source for reaction R6 avoids the challenge of nitrogen
handling in downstream processes, which prohibits commercial
viability of partial oxidation of methane in air.5
The feasibility of using metal oxides as the oxygen source for
methane partial oxidation has been demonstrated experimen-
tally for cerium-based oxides2024 as well as tungsten
oxide,2528 iron oxide,29 and zinc oxide.30,31 Partial oxidation
with undoped ceria has a high selectivity toward the desired
syngas products when operated at 1123 K.23 In a broader
context, the solar carbothermal reduction of other metal oxides
has been studied. Thermodynamic analyses indicate the
promise of the carbothermal reduction of SnO32 and ZnO33
as ecient means to store solar energy in chemical bonds. The
solar carbothermal reduction of iron, zinc, and magnesium
oxides has been demonstrated,3337 including the reduction of
ZnO on a 300 kW pilot scale.35
2. CHEMICAL THERMODYNAMICS
For the proposed coupled cycle, chemical equilibrium calculations are
used to determine the ceria non-stoichiometry that is achieved by the
reduction and oxidation steps and to dene the required amounts of
methane and oxidizer to accomplish the respective reactions. The non-
stoichiometric states accessible to thermochemical cycling are related
to the operating temperature and the oxygen partial pressure of the
gaseous environment according to eq 7.
pO
2
hO 2() T sO 2() = RT ln
pref (7)
The partial molar enthalpy and entropy of oxygen in ceria, hO2 and
sO2, depend only upon non-stoichiometry for 0.001 0.27.38
Curve ts of the experimental data for hO2 and sO2 are used to
calculate equilibrium non-stoichiometries for ceria as a function of the
temperature and oxygen partial pressure.
For reduction with methane, the oxygen partial pressure is
determined from the equilibrium of reaction R6 at 1 bar using the
software code HSC39 (Figure 3). Above 1223 K, syngas comprises
more than 99 mol % of the products and the oxygen partial pressure is
quite low, 1.8 1020 pO2 9.8 1017 bar for 1223 T 1773 K.
Below 1223 K, the equilibrium mixture contains complete combustion
products, solid carbon, and unreacted methane. To promote high-
quality syngas production, reaction R5 is evaluated for temperatures
above 1223 K. Deviations from the 2:1 ratio of CH4/O2 lead to
thermodynamic equilibrium compositions that also contain complete
combustion products for lower ratios and unreacted methane and solid
carbon for higher ratios at temperatures above 1223 K. In practice, the
dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095
Energy & Fuels
Figure 3. Equilibrium product distribution of reactants supplied in
reaction R6 as a function of the temperature at 1 bar total pressure.
Above 1223 K, syngas constitutes at least 99 mol % of the products. At
1223 K, the oxygen partial pressure is 1.8 1020 bar.
rate of oxygen release from ceria aects the ratio of methane to
oxygen. Experimental studies of methane-driven reduction of ceria in
the temperature range of 6731123 K indicate that the ratio of CH4/
O2 tends to exceed 2:1;2024 i.e., excess methane is used, but
accumulation of solid carbon can be avoided.20,21 The experimental
studies also indicate that increasing the reaction temperature increases
the rate of oxygen release and improves the conversion of methane to
syngas. The reaction of methane with undoped ceria has high
selectivity toward syngas products and methane conversion near 25%
at 1123 K.23 Improved methane conversion to syngas is expected along
with increased oxygen release rates at temperatures above 1223 K. In
addition to ideal operation of reaction R5, a case with experimentally
demonstrated methane conversion of 25% is also evaluated in the
present study to show eciencies that can be attained on the basis of
the state of the art.
The equilibrium oxygen partial pressure of the oxidation step is
established by the oxidizer dissociation reactions
CO2 CO + 1/2O2 (R8a)
H 2O H 2 + 1/2O2 (R8b)
with the corresponding equilibrium conditions
yCO yO 1/2 psys
1/2
K CO2(T ) = 2
yCO pref
2 (9a)
yH yO 1/2
psys 1/2
K H2O(T ) = 2 2
yH O
2
pref (9b)
for which temperature-dependent equilibrium constants are tabulated
using HSC.39 In the limit as , the equilibrium mole fractions,
yi , are established directly by the oxidizer dissociation reactions,
resulting in yO2 = 6.1 106 for CO2 and yO2 = 4.7 106 for H2O at
1223 K. However, the aim of the oxidation step is to convert an
appreciable fraction of the oxidizer to fuel, requiring nite .
Converting half of the oxidizer to fuel, for example, corresponds to
= 2. For this case, the equilibrium mole fractions of oxidizer and fuel
are equal, yCO2 = yCO = 0.5, and the oxygen mole fraction is
approximately 10 orders of magnitude lower than for the innite
oxidizer case: yO2 = 9.2 1016 for CO2 and yO2 = 3.9 1016 for H2O
at 1223 K.
The eect of conversion on the equilibrium oxygen mole fraction
translates to important dierences in equilibrium non-stoichiometry,
which is calculated with eq 7 for reduction according to eq R5 (dash-
dot line), oxidation with CO2 (solid lines), and oxidation with H2O
(dashed lines) and shown in Figure 4. The reduction non-
stoichiometry is red = 0.2 at 1223 K and reaches red = 0.25 above
4090
Article
Figure 4. Accessible non-stoichiometries for reduction with CH4
(dash-dot line), oxidation with CO2 (solid lines), and oxidation with
H2O (dashed lines).
1273 K. In this study, the reduction temperature is limited to 1273 K
to ensure that ceria remains in the uorite phase.6 The black curves are
lower limits for the accessible oxidation non-stoichiometry that
correspond to innite . Along these curves, ceria is in equilibrium
with the supplied oxidizing gas and no further reaction occurs. For
oxidation with innite quantities of oxidizer below 1373 K, the lower
limit for ox is less than 0.001 and the limit on the non-stoichiometry
swing is virtually equivalent to red. As the oxidation temperature is
increased, the lower limit for ox increases. At 1773 K, ox 0.021 is
accessible with CO2 and ox 0.026 with H2O. To convert the
oxidizer to fuel, a nite quantity of oxidizer is used and oxidation non-
stoichiometries must be larger than the limiting values. The red curves
correspond to oxidation with = 2, and the blue curves correspond to
= 10. As the stoichiometric oxidizer ratio is decreased, the
equilibrium non-stoichiometry increases because of the change in the
gas composition. It follows that choosing to limit oxidation to
increasingly large ox will improve the equilibrium conversion of
oxidizer to fuel, and the stoichiometric oxidizer will approach unity.
This hypothesis is consistent with an experimental demonstration of
improved eciency for the isothermal ceria cycle with sweep gas when
the reduction and oxidation reactions are terminated before ceria
equilibrates with the supplied gas.40 In the present analysis, ox is
shown to be a deciding factor for the process eciency and limiting
oxidation to large ox is viewed as a strategy for improving the oxidizer
conversion.
Thus, chemical equilibrium thermodynamics point to isothermal
operation of the cycle for large ox at temperatures from 1223 to
1273K, where the lower temperature limit allows for high-quality
syngas production during the ceria reduction step and the upper limit
avoids phase transition in ceria. Moreover, this temperature range is
well-matched to operating temperatures possible with commercial
central receiver solar facilities.41
3. THERMODYNAMIC PROCESS MODEL
The schematic in Figure 5 represents the model thermodynamic
system for the isothermal non-stoichiometric ceria redox cycle and the
partial oxidation of methane. The system consists of two reactors
(both could be embodied in a single device) and two heat exchangers.
Thermodynamic states are numbered 18 and labeled with the species
present. Concentrated sunlight provides the heat to drive the cycle,
Q sol. Ceria is cycled continuously between the reduction and oxidation
reactors. Methane is preheated by heat exchange with reduction
product gases in HXred prior to entering the solar reactor, where it is
further heated to the reduction temperature by solar energy and reacts
with ceria. The reduction products enter HXred at T3 and exit at T4.
The oxidizer is preheated in HXox by the oxidation products before
entering the oxidation reactor. The products of the oxidation reaction
exit the system via HXox. Fully mixed ow is assumed for both reactors,
such that the gas composition in each reactor is uniform. This
assumption represents a conservative case for oxidizer conversion and
is consistent with the thermodynamic analysis of the two-temperature
cycle, in which a nitrogen sweep gas is used to maintain the low partial
dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095
Energy & Fuels Article

ni , j
yi , j =
nj (16)
where the subscript i represents the species and the subscript j
represents the state in Figure 5. With a single subscript, nj refers to the
total molar ow rate at state j.
Applying the fully mixed model and introducing the reaction extent
of reaction R8a, , the species molar ow rates are
nCO
2,7 = (1 )nceria
(17)
nCO,7
= nceria
(18)
1
nO 2,7 = nceria

2 2 (19)
n7 = nCO
2,7 + nCO,7
+ nO 2,7 (20)
Figure 5. Schematic of a thermodynamic system for implementing the Treating ox and T as parameters enables calculation of and with
proposed cycle in Figure 2. Solid arrows represent mass ows, and eqs 7 and 15.
block arrows represent energy ows. With known species gas ow rates, energy balances on the heat
exchangers allow for calculation of the temperature of gases leaving the
system boundary. The methane is preheated to T2 by the reduction
euent in HXred, characterized by heat recovery eectiveness, HXred.
pressure of oxygen over the ceria.9 Reactants are supplied to the The reduction euent is cooled to T4 in the same heat exchanger. The
system at ambient temperature, T1 = T5 = T. Both reactions occur at amount of heat exchanged is
temperature T and the products exit the reactors at the reaction
temperature, i.e., T3 = T7 = T. The total pressure of the system and its n1 (h1(T ) h1(T)),
surroundings is assumed to be 1 bar. Q HX = HX red min

n3 (h3(T ) h3(T))
red
The system level energy balance is (21)
Q sol = Q s,red + Q s,ox + Q c,red + Q c,ox + Q reject + Q rerad where the molar enthalpy of the mixture at state j is

+ Q loss (10)
hj(Tj) = yi ,j hi ,j(Tj)
i (22)
The supplied solar energy, Q sol, must provide the enthalpy change of
T2 and T4 are calculated with eqs 23 and 24.
the reduction and oxidation gas ows, which include sensible heating,
Q s,red and Q s,ox, and chemical energy storage, Q c,red and Q c,ox. A Q HX = n1 (h2(T2) h1(T1))
red (23)
portion of the solar input is rejected to maintain isothermal operation
during oxidation, Q reject In addition, heat is lost by reradiation from the Q HX = n3 (h3(T3) h4(T4))
receiver, Q rerad, and convection to the ambient, Q loss. red (24)
Calculating the energy requirements for the reduction and oxidation The heating rate required to produce the total enthalpy change of the
gas ows requires the ow rates and gas compositions, which are reduction gases entering and leaving the system boundary is
linked to the ceria non-stoichiometry. In the reduction step, ceria is
reduced to red, as shown in Figure 4, and methane is supplied at the Q g,red = n4 h4(T4) n1 h1(T1) (25)
molar ow rate
To evaluate the impact of operating with excess methane, a special
nCH
4,1 = nceria
(11) case is considered with an increased reduction gas heating require-
ment. Li et al.23 demonstrated the partial oxidation of methane over
to achieve the 2:1 CH4/O2 ratio in reaction R6. The reduction
ceria with high syngas selectivity and methane conversion near 25%.
products are formed at the rates
On the basis of the methane conversion demonstrated by Li et al., the
nCO,3
= nceria
(12) additional sensible heating requirement is

nH
2,3 = 2nceria
Q CH ,excess = 3nCH
4,1hCH4(T4) 3nCH
4,1hCH4(T1) (26)
(13) 4

In the model system, gas compositions are modied only in the The oxidation side heat exchanger is evaluated in the same manner to
reactors, e.g., nCO,3 = nCO,4. determine Q HXox, T6, and T8. The heating rate required by the
The following equations for the oxidation step are presented only oxidation gases is
for CO2 splitting. The equations are identical for H2O splitting if one
simply replaces CO2 with H2O and CO with H2. In terms of the Q g,ox = n8 h8(T8) n5 h5(T5) (27)
quantities in Figure 5, the ratio of CO2 supplied to the stoichiometric
amount is Note that Q g,red and Q g,ox include both the sensible and chemical
energy storage terms. The net energy stored in chemical bonds for the
nCO
2,5 reduction and oxidation steps is
=
nceria
(14) Q c,red = (ni ,4 ni ,1)HHVi
The equilibrium condition for CO2 thermolysis becomes i (28)

yCO,7 yO ,71/2 psys 1/2 Q c,ox = (ni ,8 ni ,5)HHVi

K CO2(T ) = 2
i (29)
yCO ,7 po (15)
2
and the sensible heating requirement of the gases, Q s,red and Q s,ox, is
and the mole fractions are given by

4091 dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095

Energy & Fuels Article

Q g,red = Q c,red + Q s,red (30)

Q g,ox = Q c,ox + Q s,ox (31)

There is no net energy released by the solid ceria because of its cyclic
oxidation and reduction within the system. However, if the enthalpy
released by the solid during oxidation
ox
nceria
hceria,ox = nceria

red
hO 2 d
(32)
exceeds the enthalpy change of the oxidation gases between states 6
and 7, heat must be rejected to maintain operation at T.
Q reject = nceria
hceria,ox + n6 h6(T6) n7 h 7(T7) (33a) Figure 6. Eciency as a function of the concentration ratio for T =
1223 K and ox = 0.1 with no heat recovery for CO2 splitting (solid
If the enthalpy change in the gas phase is larger than that of the solid, line) and H2O splitting (dashed line).
heat rejection is not required.
Q reject = 0 (33b) yields eciencies for CO2 and H2O splitting of about 40%.
There is very little benet to raising the concentration ratio
The solar receiver is assumed to be eectively black, so that there are beyond 1000. For example, eciency increases to only 43% at
no losses associated with absorption of the concentrated sunlight. C = 3000. This result demonstrates one of the key benets of
Losses because of emission are a fraction of the solar input methane reduction. Without carbothermal reduction, ecient
Tred 4 operation of the ceria-based redox cycle requires a concen-
Q rerad = Q sol tration ratio of 3000,9,13,16 only achievable with costlier dish
CG (34)
systems.42 For the results that follow, we x the concentration
The thermal loss from the system components by convection to the ratio to 1000 to match commercial central receiver systems.
ambient is treated as a fraction, F, of the absorbed solar energy. The inuence of ox on the stoichiometric oxidizer ratio, ,
Q loss = F(Q sol Q rerad) (35) dened by eq 14, is shown in Figure 7 for CO2 splitting (solid
The solar-to-fuel eciency in terms of the above quantities is
Q c,red + Q c,ox
=
Q sol (36)
where the numerator is the net dierence in the higher heating value
of the reaction products and the supplied methane and the
denominator is the solar energy supplied to drive the cycle.

4. RESULTS
To illustrate the benets of the proposed cycle, we present the
results of a parametric analysis of the eects of operating
parameters on eciency. We show that the oxidation non-
stoichiometry (ox) has a decisive impact on eciency, because
it prescribes the amount of oxidizer that must be supplied and
heated to the oxidation temperature. We also report the eects
of the concentration ratio (C) and heat recovery eectiveness
(HXox and HXred) on the cycle eciency. Acknowledging that
complete methane conversion has not yet been demonstrated,
the impact of unreacted methane on eciency is also shown.
Results are reported for the cycle operated isothermally at
temperatures from 1223 to 1273 K. Although higher temper-
ature operation may be possible, raising the temperature oers
no thermodynamic benet for the proposed cycle. The solar
energy input for the reported results is 1 kW, and results may
be scaled to arbitrary input power. The fraction of convective
losses from the system, F = 0.2, is selected to match previous
analyses of ceria redox cycles.9,16
The impact of the solar concentration ratio on eciency is
shown in Figure 6 for CO2 splitting (solid lines) and H2O
splitting (dashed lines) at 1223 K with ox = 0.1. These data are
for the simplest system without heat exchangers, i.e., HXred =
HXox = 0. The trend is the same for all temperatures and ox
considered. A concentration ratio of 1000, which is available in
commercial central receiver concentrating solar power plants,41
4092
Figure 7. Stoichiometric oxidizer ratio required for oxidation from red
to ox with CO2 (solid lines) and H2O (dashed lines).
lines) and H2O splitting (dashed lines). Two important trends
are evident. First, with increasing ox, decreases and
approaches the ideal value, = 1, corresponding to complete
conversion of the oxidizer to fuel. Reducing is desirable for
ecient operation, because it reduces the amount of unreacted
oxidizer that is contained in the product gas mixture leaving the
system at state 8. Second, although the stoichiometric oxidizer
ratio is larger for water than for carbon dioxide, the dierence is
signicantly reduced at larger oxidation non-stoichiometries.
Both trends show thermodynamic benets of operating with
large oxidation non-stoichiometries. In the range considered,
the importance of the operating temperature also depends
upon the oxidation non-stoichiometry. For CO2 splitting with
ox = 0.005, = 27 at 1223 K and = 56 at 1273 K. For ox =
0.1, oxidation is nearly stoichiometric; = 1.1 at 1223 K, and
= 1.2 at 1273 K. For H2O splitting with ox = 0.1, = 1.2 at
1223 K and = 1.4 at 1273 K.
Selection of ox has a substantial impact on the process
eciency, because of the energy requirement for heating the
dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095
Energy & Fuels
oxidizer, which increases in proportion to the stoichiometric
oxidizer ratio. The eciency for the proposed cycle without
heat recovery is shown in Figure 8 for a range of ox. For CO2
Figure 8. Solar-to-fuel eciency for CO2 splitting (solid lines) and
H2O splitting (dashed lines) versus Tox and ox with no heat recovery.
splitting at 1223 K and ox = 0.1, the eciency is 39%; for ox =
0.005, the eciency is 11%. The trends are the same for water
splitting. For ox 0.02, the eciency is higher for CO2
splitting than for H2O splitting, because H2O splitting requires
a larger stoichiometric oxidizer ratio. However, at ox = 0.1, the
eciency is slightly higher for H2O splitting, because the higher
heating value of H2 is larger than that of CO and the dierence
in the stoichiometric oxidizer ratio is negligible.
Although eciencies of 39% are possible without recovering
sensible heat from the process gases, heat recovery on the
reduction side provides a modest improvement. For CO2
splitting with ox of 0.1, the sensible heating requirement of
the reduction gases, Q s,red, is on the order of 10% of the solar
input. The eciency increases from 39 to 44% at 1223 K with
perfect heat recovery from the reduction gases. Recovering heat
from the oxidation gases is only benecial when Q reject = 0, but
for CO2 splitting with ox = 0.1, the stoichiometric oxidizer ratio
remains near unity for the considered temperature range and
Q reject > 0.
To identify the eciency possible with the experimentally
demonstrated conversion of methane to syngas over ceria, a
special case is considered, where 75% of the supplied methane
does not react and is contained in the reduction products. The
eciencies for this case are shown for ox = 0.1 with no heat
recovery (Figure 9) and 90% heat recovery from the reduction
gases (Figure 10). The additional sensible heating requirement
Figure 9. Eciency for the proposed cycle for CO2 splitting (solid
lines) and H2O splitting (dashed lines) without heat recovery and 25%
methane conversion to syngas.
4093
Article
Figure 10. Eciency for the proposed cycle for CO2 splitting (solid
lines) and H2O splitting (dashed lines) with HXred = 0.9 and 25%
methane conversion to syngas.
increases the sensitivity of eciency to heat recovery from the
reduction gases. With no heat recovery, the eciency is near
30% at 1223 K. Implementing heat recovery from the reduction
gases with HXred = 0.9 enables 39% eciency for the cycle with
excess methane.
In comparison to solar-only approaches to the ceria cycle
(Figure 1), the proposed coupled cycle with methane-driven
reduction (Figure 2) provides signicant eciency and
operational benets. The eciency of the proposed cycle at
the recommended operating conditions, ox = 0.1 at 1223 K,
and the eciency reported for an isothermal solar-only
approach to the ceria cycle that uses sweep gas for reduction
and operates at 1773 K9 are given in Table 1. The eciency
Table 1. Comparison of Reported Solar-to-Fuel Eciencies
CO2 H2O HXred HXox
process (%) (%) T (K) C (%) (%)
isothermal with 11 5 1773 3000 90 90
sweep gas
proposed CH4 39 40 1223 1000 0 0
reduction
predicted for the proposed cycle without heat recovery is 39%,
nearly quadruple the eciency predicted for the solar-only
cycle with ambitious levels of heat recovery and a much higher
concentration ratio. In contrast to the solar-only approach,
water splitting may be carried out just as eciently as carbon
dioxide splitting with the proposed cycle, because the
thermodynamic disadvantage for water splitting is overcome
by a substantial increase in the accessible reduction non-
stoichiometry when methane is used. Because high eciencies
are predicted for the proposed coupled cycle operated
isothermally and without gas-phase heat recovery, the need
for high-temperature solid-phase heat recuperators and gas heat
exchangers is avoided. The lower operating temperature of the
proposed approach also allows for higher eciencies to be
attained at lower concentration ratios than for the solar-only
approaches, which has positive economic implications for
commercial implementation.
5. CONCLUSION
The present thermodynamic study shows that coupling the
partial oxidation of methane to the reduction step of the solar
ceria redox cycle enables ecient conversion of solar energy to
fuel at temperatures as low as 1223 K. The coupled cycle
produces high-quality syngas by the partial oxidation of
dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095
Energy & Fuels Article

methane in the ceria reduction step in addition to the carbon red = reduction
monoxide and hydrogen produced by splitting carbon dioxide reject = heat rejected from oxidation reactor
and water in the oxidation step. The predicted solar-to-fuel rerad = reradiation
eciency is 39% for carbon dioxide splitting and 40% for water s = sensible energy storage
splitting for the cycle without heat recovery. The need for heat sol = solar
recovery from the oxidation euent is eliminated completely Abbreviations
by operating with large oxidation non-stoichiometry, ox, to HX = heat exchanger

allow for nearly complete conversion of the oxidizer to fuel.
Using biogenic methane sources, the proposed approach can be REFERENCES
operated sustainably. The high eciencies possible without
(1) Havran, V.; Dudukovic, M.; Lo, C. Ind. Eng. Chem. Res. 2011, 50,
heat recovery and the relatively low operating temperature 70897100.
make the proposed cycle an attractive alternative to solar-only (2) Agrafiotis, C.; von Storch, H.; Roeb, M.; Sattler, C. Renewable
approaches to storing solar energy with the ceria cycle.

Sustainable Energy Rev. 2014, 29, 656682.

(3) York, A. P. E.; Xiao, T.; Green, M. L. H.; Claridge, J. B. Catal. Rev.
AUTHOR INFORMATION 2007, 49, 511560.
Corresponding Author (4) Rostrup-Nielsen, J. R. Catal. Today 1993, 18, 305324.
*E-mail: jhd@me.umn.edu. (5) Christian Enger, B.; Ldeng, R.; Holmen, A. Appl. Catal., A 2008,
346, 127.
Notes (6) Chueh, W. C.; Haile, S. M. Philos. Trans. R. Soc. London 2010,
The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
This research was supported by the National Science
Foundation (EFRI-1038307) and the University of Minnesota
Initiative for Renewable Energy and the Environment. Peter T.
Krenzke is supported by the National Science Foundation
Graduate Research Fellowship Program (NSF-GRFP) under
Grant 00039202.
368, 32693294.
(7) Abanades, S.; Flamant, G. Sol. Energy 2006, 80, 16111623.
(8) Chueh, W. C.; Haile, S. M. ChemSusChem 2009, 2, 735739.
(9) Lapp, J.; Davidson, J. H.; Lipinski, W. Energy 2012, 37, 591600.
(10) Chueh, W. C.; Falter, C.; Abbott, M.; Scipio, D.; Furler, P.;
Haile, S. M.; Steinfeld, A. Science 2010, 330, 17971801.
(11) Furler, P.; Scheffe, J. R.; Steinfeld, A. Energy Environ. Sci. 2012,
5, 60986103.
(12) Furler, P.; Scheffe, J. J. R.; Gorbar, M.; Moes, L.; Vogt, U.;

Steinfeld, A. Energy Fuels 2012, 26, 70517059.

(13) Ermanoski, I.; Siegel, N. P.; Stechel, E. B. J. Sol. Energy Eng.
NOMENCLATURE 2013, 135, 031002.
C = concentration ratio (14) Hao, Y.; Yang, C.-K.; Haile, S. M. Phys. Chem. Chem. Phys. 2013,
F = heat loss factor 15, 1708417092.
(15) Muhich, C. L.; Evanko, B. W.; Weston, K. C.; Lichty, P.; Liang,
G = direct normal irradiance (W m2)
X.; Martinek, J.; Musgrave, C. B.; Weimer, A. W. Science 2013, 341,
h = molar enthalpy (J mol1) 540542.
HHV = higher heating value (J mol1) (16) Bader, R.; Venstrom, L. J.; Davidson, J. H.; Lipinski, W. Energy
K = equilibrium constant Fuels 2013, 27, 55335544.
ni = molar ow rate (mol s1) (17) Ermanoski, I.; Miller, J. E.; Allendorf, M. D. Phys. Chem. Chem.
pi = species i partial pressure (bar) Phys. 2014, 16, 84188427.
pref = reference pressure (bar) (18) Venstrom, L. J.; Davidson, J. H. J. Sol. Energy Eng. 2011, 133,
psys = total pressure (bar) 011017.
Q = heat transfer rate (W) (19) Venstrom, L. J.; Petkovich, N.; Rudisill, S.; Stein, A.; Davidson,
R = ideal gas constant (8.314 J mol1 K1) J. H. J. Sol. Energy Eng. 2012, 134, 011005.
(20) Otsuka, K.; Sunada, E.; Ushiyama, T.; Yamanaka, I. Stud. Surf.
T = temperature (K)
Sci. Catal. 1997, 107, 531536.
y = mole fraction (21) Otsuka, K.; Wang, Y.; Sunada, E.; Yamanaka, I. J. Catal. 1998,
Greek Letters 175, 152160.
= non-stoichiometry (22) Otsuka, K.; Wang, Y.; Nakamura, M. Appl. Catal., A 1999, 183,
hO2 = partial molar enthalpy of oxygen in ceria (J mol1) 317324.
= change in non-stoichiometry (23) Li, K.; Wang, H.; Wei, Y.; Yan, D. Appl. Catal., B 2010, 97, 361
sO 2 = partial molar entropy of oxygen in ceria (J mol1) 372.
= heat exchanger eectiveness (24) Fathi, M.; Bjorgum, E.; Viig, T.; Rokstad, O. Catal. Today 2000,
63, 489497.
= solar-to-fuel eciency (25) Sim, A.; Cant, N. W.; Trimm, D. L. Int. J. Hydrogen Energy 2010,
= stoichiometric oxidizer ratio 35, 89538961.
= reaction extent (26) Kwak, J. H.; Han, G. Y.; Yoon, K. J. Int. J. Hydrogen Energy 2013,
Subscripts 38, 113.
1, 2, ... = state (27) Kodama, T.; Ohtake, H.; Matsumoto, S.; Aoki, A. Energy 2000,
c = chemical energy storage 25, 411425.
excess = exceeding reaction requirement (28) Kodama, T.; Shimizu, T.; Satoh, T.; Shimizu, K.-I. Energy 2003,
28, 10551068.
fuel = species having a heating value (29) Kodama, T.; Shimizu, T.; Satoh, T.; Nakata, M.; Shimizu, K. Sol.
g = gas Energy 2002, 73, 363374.
i = species (30) Steinfeld, A.; Brack, M.; Meier, A.; Weidenkaff, A.; Wuillemin,
j = state D. Energy 1998, 23, 803814.
loss = convection losses to ambient (31) Kraupl, S.; Steinfeld, A. J. Sol. Energy Eng. 2001, 123, 237243..
ox = oxidation (32) Forster, M. Energy 2004, 29, 789799.

4094 dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095

Energy & Fuels Article

(33) Steinfeld, A.; Larson, C.; Palumbo, R.; Foley, M. I. I. I. Energy

1996, 21, 205222.
(34) Steinfeld, A.; Kuhn, P.; Karni, J. Energy 1993, 18, 239249.
(35) Wieckert, C.; Frommherz, U.; Kraupl, S.; Guillot, E.; Olalde, G.;
Epstein, M.; Santen, S.; Osinga, T.; Steinfeld, A. J. Sol. Energy Eng.
2007, 129, 190196.
(36) Epstein, M.; Olalde, G.; Santen, S.; Steinfeld, A.; Wieckert, C. J.
Sol. Energy Eng. 2008, 130, 1450414505.
(37) Murray, J.; Steinfeld, A.; Fletcher, E. Energy 1995, 20, 695704.
(38) Panlener, R. J.; Blumenthal, R. N.; Garnier, J. E. J. Phys. Chem.
Solids 1975, 36, 12131222.
(39) Roine, A. HSC Thermochemistry, 7.0; Outotec: Espoo, Finland,
2009.
(40) Venstrom, L. J.; De Smith, R. M.; Hao, Y.; Haile, S. M.;
Davidson, J. H. Energy Fuels 2014, 28, 27322742.
(41) Alexopoulos, S.; Hoffschmidt, B. Renewable Energy 2010, 35,
13521356.
(42) Romero, M.; Steinfeld, A. Energy Environ. Sci. 2012, 5, 9234
9245.

4095 dx.doi.org/10.1021/ef500610n | Energy Fuels 2014, 28, 40884095