Article type: Full Paper

Modified Ceria for ‘Low-Temperature’ CO2 Utilization: A Chemical Looping

Route to Exploit Industrial Waste Heat

Vasudev Pralhad Haribal§1, Xijun Wang⧧1, Ryan Dudek1, Courtney Paulus1, Brian Turk2,
Raghubir Gupta2 and Fanxing Li1*
1
Department of Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, NC
2
Susteon Inc., Raleigh, NC
*Correspondence: [email protected]
§
first author ⧧co-first author

Abstract
Efficient CO2-utilization is key to limit global climate change. Carbon monoxide, which is a crucial
feedstock for chemical synthesis, can be produced by splitting CO2. However, existing thermochemical
routes are energy-intensive requiring high operating temperatures. We report a Hybrid Redox Process
(HRP) involving CO2-to-CO conversion using a lattice oxygen-deprived redox catalyst at relatively low
temperatures (<700 °C). The lattice oxygen of the redox catalyst, restored during CO2-splitting, is
subsequently used to convert methane to syngas. Operated at temperatures significantly lower than a
number of industrial waste heat sources, this cyclic redox process allows for efficient waste heat-utilization
to convert CO2. To enable the low temperature operation, we report lanthanum modified ceria (1:1 Ce: La)
promoted by rhodium (0.5 wt. %) as an effective redox catalyst. Near-complete CO2 conversion with a
syngas yield of up to 83% at low temperatures were achieved using Rh-promoted LaCeO4-x. While La
improves low-temperature bulk redox properties of ceria, Rh considerably enhances the surface catalytic
properties for methane-activation. Density Functional Theory calculations further illustrate the underlying
functions of La-substitution. The highly effective redox catalyst and HRP scheme provide a potentially
attractive route for chemical production using CO2, industrial waste heat, and methane, with appreciably
lowered CO2 emissions.

Keywords
CO2-utilization, syngas, chemical-looping, redox catalyst, reforming

1. Introduction
Fossil fuel utilization through human activities lead to over 35,000 teragrams (Tg) global emissions of
greenhouse gases each year, with CO2 being the primary component[1]. With the aim to limit global warming
below a 2oC increase, efficient usage of CO2 is a pressing issue. CO2 utilization is a broad term covering a
range of technologies that involves the consumption of CO2 to provide products or services with the main
objective of an economic benefit, ideally with additional environmental and social benefits[2,3]. CO2 is
viewed as a waste product in the context of flue gases, but it has the potential to be a valuable commodity.
Productive utilization of CO2 via conversion to valuable chemicals, is an attractive avenue[1,2,4]. With a
growing global population, cost-effective chemical synthesis via CO2-utilization is highly desired[5].
For the C1 molecules, Gibbs free energies of formation (∆Gf) are in the following order: CH4 > CH3OH >
HCHO > CO > HCOOH > CO2[6,7]. Strong double bonds with oxygen make carbon dioxide stable and
difficult to convert from a thermodynamic viewpoint. Geological CO2 sequestration represents one of the
few CO2 conversion processes that are energetically favored. However, it is technically challenging and
can be cost-intensive[1]. Some CO2 conversion reactions such as carboxylation do not require significant
energy input. However, the production of the carboxylation reactants such as alcohols and alkenes can be
quite energy intensive. Reactions producing reduced forms of CO2 like formaldehyde or carbon monoxide
also demand hefty energy input. The required energy, which is thermodynamically unavoidable, can be
supplied thermally, electrochemically or photo-chemically[7]. Electrochemical routes face challenges of low
energetic efficiency (or large over-potential), slow electron transfer kinetics, low selectivity and poor long-
term stability causing a major hurdle towards commercialization[8]. Similarly, high operating temperatures
and low photon efficiencies hinder the practical operation of the thermo-chemical and photo-chemical
processes, respectively[9–13]. From a sustainability standpoint, the required external energy should be
supplied by a renewable source. While such resources can supply the required energy, they still face
challenges related to available capacity, intermittency, and/or cost. Industrial waste heat, on the other hand,
is an abundant (5-13 quadrillion BTU/year, in the US) and low cost energy source. Discharged industrial
waste contains 20-50% of the consumed industrial energy[14]. For instance, the oxy-fuel glass furnace can
produce 1,427°C-exhaust containing 1.22 GJ of sensible waste heat per metric ton of glass product, which
is currently wasted[1,15]. Recovering this waste heat for CO2 utilization can result in significant net reduction
of greenhouse gas emissions. The challenge for industrial waste heat integration, however, resides in its
significantly lower quality (exergy-wise) compared to electric power, chemicals, or fuels. As such,
thermochemical CO2 utilization processes operated at low temperatures are necessary to enable efficient
utilization of industrial waste heat.

Commercially, carbon monoxide (CO) is an indispensable raw material for chemical production[16], being
the starting point for a plethora of carbonyl compounds. CO2-to-CO is a CO2-utilization scheme which can
help produce valuable chemicals sustainably. We propose a Hybrid Redox Process (HRP), shown in Figure
1a, which not only utilizes the CO2 to produce CO, but also has the potential to exploit the industrial waste
heat, due to its low-temperature operation. Additionally, it uses the abundant methane[17] to produce
methanol-ready syngas. HRP involves a two-step, open thermochemical loop facilitated by a redox catalyst
(Reactions 1 and 2). In the first step, a reduced redox catalyst (generically denoted as MeOy) reacts with
CO2 to produce high purity CO while being oxidized in the process. The oxidized redox catalyst (MeOx,
x>y) subsequently reacts with methane to yield synthesis gas (syngas, which is CO+H2) with a H2: CO ratio
(~2:1), ideal for methanol and Fischer-Tropsch synthesis[6,18,19]. Both syngas and CO are important
industrial feed-stocks for chemical production. Transformation of CO2, CH4, and industrial waste heat via
HRP thus offers various opportunities for chemical synthesis, as shown in Figure 1b. As an example, the
syngas products from HRP can readily be converted into methanol using Reaction 3. By combining this
methanol product with the CO product, generated from HRP, acetic acid can be produced by the commercial
carbonylation reaction (Reaction 4)[6,16,20]. Other potential products from HRP include methyl acetate, acetic
andrydride, etc.
CO2 + MeOy → CO + MeOx Reaction 1: CO2-splitting (Step 1)
CH4 + MeOx → 2H2 + CO + MeOy Reaction 2: Methane Partial Oxidation (Step 2)
CO + 2 H2 → CH3OH Reaction 3: Methanol from syngas
CH3OH + CO → CH3COOH Reaction 4: CativaTM process

The main challenge for HRP resides in efficient conversion of both methane and CO2 at relatively low
temperatures. The C-H bond in methane has a dissociation energy of 439.3 kJ/mol, making its activation
non-trivial, especially when lattice oxygen is used as the oxidant[21–24]. On the other hand, existing
thermochemical approaches for CO2-splitting require high temperatures and are often subjected to limited
CO2-to-CO conversion[10,25–29]. Sr and Fe-containing perovskite nanocomposites have exhibited exceptional
efficacy for both CO2-splitting and methane partial oxidation (POx), with >98% CO yield and 96% syngas
selectivity in respective steps[30]. However, these materials require high operating temperatures (>950oC),
which do not allow for efficient use of the waste heat. Ceria (CeO2) is another promising candidate for the
two reaction steps, owing to its ability to accommodate oxygen non-stoichiometry and fast oxygen diffusion
kinetics. But, it requires temperatures >900oC for high methane conversions[10,31–33]. The substitution in
ceria and surface promotion with platinum group metals have been attempted in previous studies to enhance
the activity of ceria-based redox catalysts[26,32]. However, effective redox catalysts for methane conversion
and CO2-splitting at low operating temperatures have yet to be reported.

For the HRP process, an ideal redox catalyst should possess (i) high oxygen release and storage capacity;
(ii) fast oxygen exchange rates; (iii) good stability over redox cycles;[21] and (iv) ability to operate at low
temperatures (< 700oC). In the current study, we report 0.5 wt. % Rh-promoted LaCeO4-x (Rh/LCO) as a
superior redox catalyst for HRP. High CO2 conversion (>95%) and CH4-to-syngas yield (83%) are
demonstrated at low temperatures (~650oC). It was determined that La-substitution decreases both the
surface and bulk oxygen vacancy formation energies in CeO2. Moreover, energy barriers for vacancy
migration are also significantly lowered. The low temperature methane activity is further enhanced by Rh-
promotion on the oxide surface. When operated at <700°C, a majority of the exergy (available work) in the
waste heat from a typical glass plant can be utilized. At this rate, waste heat from the U.S. glass industry
(2.9 million metric tons of glass/year) alone would be adequate to convert 2.3 million metric tons of CO2
each year, producing 2.71 million metric tons of acetic acid annually[1,14,15,34]. Integration of the waste heat
and low cost natural gas to convert CO2 can reduce the net CO2 emission for acetic acid production by 99%.
Through CO2 utilization and waste heat recovery, HRP and the low temperature redox catalyst reported in
this study also has the potential to produce value-added chemicals well beyond acetic acid.

2. Results and Discussion

2.1. HRP Redox Catalyst Design

Cerium-based oxides, with high oxygen anion (O2-) conductivity, are promising redox catalysts[26,35–37]. At
low temperatures, the concentration of oxygen vacancy and its energy barrier for migration are particularly
important since they determine the rate at which lattice oxygen can participate in the redox reactions. It is
noted that unmodified CeO2 has relatively low O2- conductivity and poor redox kinetics at the temperatures
of interest for HRP. However, the fluorite structure can accommodate cation dopants for improved O2-
conductivity and/or structural stability[35,38]. Among a variety of cations, trivalent lanthanum cation is a
good candidate due to its similarity in mass and size when compared to Ce4+. Meanwhile, addition of
lanthanum can also hinder coke formation by increasing the surface basicity. LaCeO4-x (or LCO) is chosen
in the current study since substitution of more than 50% Ce in CeO2 with La was reported to cause phase
segregation whereas LCO forms a stable structure[33,35].

2.1.1. Effect of La3+ substitution: DFT Calculation

To determine the effect of La-substitution, Density Functional Theory (DFT) calculations were performed
on (i) the bulk and surface oxygen vacancy formation energies (ΔEV) and (ii) energy barriers for oxygen
vacancy migration of both pure and La-substituted ceria (CeO2 and LCO, respectively). The former
quantifies the thermodynamic energy required for lattice oxygen removal whereas the latter depicts the ease
of vacancy migration from a kinetics-viewpoint. Both are important factors for the oxygen conductivity
that is required for macroscopic lattice oxygen removal and re-deposition in redox reactions. LCO unit cell
was constructed by replacing two Ce atoms of the pristine CeO2 unit cell by La atoms, and then removing
one O atom to form an oxygen vacancy (VO) as shown in Figure S1. As depicted in Figure 2a, the computed
ΔEV for LCO bulk is 2.26 eV, which is notably lower than that for pristine CeO2 (2.99 eV[39]). The ease of
O2- migration in LCO bulk can be correlated with a lower oxygen migration barrier of 0.30 eV (vs. 0.46 eV
for pure CeO2[35]) (Figure 2b), computed through transition-state calculations. In terms of vacancy creation
and migration on the surface, (111) surface was selected for both CeO2 and LCO since it is reported as the
most stable ceria surface[40]. Based on the X-ray photoelectron spectroscopy (XPS) results listed in Table
1, significant enrichment of La (~2:1 La:Ce atomic ratio) was observed near the oxide surface. As such, the
surface was constructed with a similar La: Ce ratio in DFT, as shown in Figure 2c. A negative value of
ΔEV was computed for LCO (111) surface, indicating a higher surface oxygen vacancy concentration than
the initially assumed LaCeO3.5 stoichiometry. The migration pathway of the sub-surface oxygen to the
vacancy site has a very low barrier of 0.16 eV, which is much lower than that of CeO2 (111) surface (0.47
eV). Overall, the DFT results indicate significantly easier oxygen migration and release, in La-substituted
ceria.

To further elucidate the effect of La-substitution, charge distributions of CeO2 and LaCeO3.5 were studied.
Lanthanum, a p-type dopant, can effectively reduce the electron donation by the cations compared to CeO2,
thereby decreasing the number of valence electrons (from 1.23 e- per O for pristine ceria to 1.19 e-) on
oxygen anions. The decrease in electron density can weaken the strength of the metal-oxygen (M-O) ionic
bond. This effect of La-substitution, combined with existing oxygen vacancies, reorganize the electronic
states for LCO when compared to CeO2. Such a reorganization of oxygen electronic states can be well
described by the oxygen p-band center (εp), which is related to the hybridization of metal and oxygen
orbitals and has been used as an effective descriptor for the activity of lattice oxygen[41,42]. Figure 2d shows
an upward-shift of εp after La-substitution, indicating stronger coupling between the electronic orbitals of
oxygen and metal atoms[43]. This leads to a higher activity of lattice oxygen in LCO. Hence, La-substitution
in ceria can be an effective approach to improve the oxide’s redox performance through the lowering of the
energy costs for both oxygen vacancy formation and migration.

2.1.2. Effect of Rh-promotion

Redox catalyst performance can be further enhanced by increasing the rate of surface oxygen removal,
which can be facilitated by catalytically active sites for methane activation. Our previous studies indicate
that the relative rates of lattice oxygen (O2-) conduction to the surface and the surface oxygen removal by
the gas–solid reactions determine the selectivity and activity of the redox catalysts for methane POx[44]. In
the present study, Rh is selected as the active metal to enhance the catalytic activity for methane activation
due to its high activity and resistance towards oxidation under the cyclic redox conditions. Figure 3a
illustrates the X-ray Diffraction (XRD) patterns of the as-prepared samples (0.5 wt. % Rh promotion on
both CeO2 and LCO), confirming the fluorite structure of ceria being maintained after both the La-addition
and Rh-promotion. Introduction of La in ceria leads to a shift in the position of the ceria peaks to smaller
2θ values. This is consistent with the anticipated lattice parameter increase, resulting from the increased
number of oxygen vacancies and larger cation diameter of La3+. The patterns also indicate that Rh is unlikely
to be incorporated into the bulk structure (Figure S13b), based on absence of lattice parameter change
between the reference and Rh-promoted ceria samples. This is consistent with the XPS results (Table 1),
which indicate more than 5 fold increase in near-surface Rh-content, when compared to the average atomic
composition of the bulk material. The significant surface enrichment of Rh is desirable for enhanced
methane activation.

Methane temperature programmed reduction (TPR) experiments were conducted to determine the effect of
promoting LCO with Rh. As shown in Figure 3b, Rh/LCO maintained the fluorite structure throughout the
reduction reaction in the presence of methane. Lattice oxygen removal resulted in increased oxygen non-
stoichiometry as opposed to the collapse of the fluorite structure, as indicated by the systematic shift in
peak positions towards smaller angles. Figure 3c shows that Rh/LCO activates methane at temperatures
considerably lower (300oC lower) than both un-promoted LCO and CeO2. Compared to Rh/CeO2, the onset
temperature for methane activation on Rh/LCO was also approximately 100 °C lower (Figure S14).
Although the cumulative oxygen release from Rh/CeO2 was higher than that for Rh/LCO at higher
temperatures due to the higher Ce content in Rh/CeO2, the superior activity at low temperatures for Rh/LCO
represents a distinct advantage for HRP. The ability to maintain phase stability while accommodating
significant oxygen vacancy formation under reducing conditions at relatively low temperatures (~ 350 °C
onset temperature) makes Rh/LCO a promising material for HRP.

2.2. Redox Catalyst Performance in HRP

For HRP, the desired redox catalyst should be able to produce high yields of syngas and CO at low
temperatures (<700oC). This would enable efficient waste heat integration. To that end, the activity and
selectivity of the redox catalysts during both the reduction and oxidation steps were tested in a quartz U-
tube reactor. Figure 4a compares the performance of Rh/LCO and Rh/CeO2 at various temperatures. As
can be seen, Rh/CeO2 exhibited comparable methane conversions at 700 °C and above (up to 99% at 900
°C). In comparison, Rh/LCO is significantly more active at lower temperatures (500 and 600 °C). For
instance, methane conversion on Rh/LCO more than tripled that on Rh/CeO2 at 500 °C, highlighting the
superior low-temperature performance of the La-substituted ceria sample. Based on the TPR results of
Rh/LCO (Figure 3b) and the requirement of HRP, a standard operating temperature of 650oC was selected
for redox performance evaluations. The reactor-feeds were switched between methane (reduction or POx
step) and CO2 (oxidation step), with an intermediate purge step. The calculated performance parameters
include methane conversion (%), CO selectivity (%) and syngas yield (during reduction) (mmol/g redox
catalyst) along with CO2 conversion (%, during oxidation). Figure 4b is a representative product
distribution during one redox cycle using Rh/LCO at 650oC, with Figure 4c compiling the results of the
two half-cycles over 25 complete redox cycles. Rh/LCO shows a stable methane conversion of ∼90% and
a high CO selectivity of 90%. CO2 conversion during the CO2-splitting step was also near complete. This
stable performance is also observed in Figure 4d, with the phases being maintained intact after repeated
cycles. At a low temperature of 650oC, a syngas yield of roughly 0.5 mmol/g material was observed for
each of the redox cycles. Figure 4d also confirms the stability of the LCO phase, as indicated by the
relatively small shifts in diffraction peaks for the reduced Rh/LCO compared to the oxidized sample.

2.3. Process Analysis

As shown in Figure 1a, the overall HRP has three main sections: (i) upstream HRP section, where syngas
and CO are produced; (ii) methanol synthesis section, that converts syngas into methanol; and (iii) acetic
acid synthesis section in which methanol and CO are converted to acetic acid by the CativaTM process. A
variety of downstream chemicals can be manufactured with this scheme (Figure 1b). We first focus on
acetic acid considering its growing market. Coal-based acetic acid synthesis, as an established commercial
process, is an ideal reference process for comparison. Production of syngas and CO via coal gasification
and chemical looping are broadly depicted in Figure 5a. We note that the downstream chemical production
section of the two processes, i.e. coal gasification (CG) and HRP, are identical. The main difference of the
two approaches resides in the upstream section, where syngas and CO are generated. In CG, water gas shift
is necessary, followed by CO2 removal to obtain methanol-ready syngas and pure CO. Energy required for
the CO2 separation is accounted for without further compression. Methane combustion is assumed to be the
energy source for the CG route, with H2 separation, CO2-separation and Air Separation Unit (ASU)
contributing to the overall process[21]. Based on the material performance shown of Figure 4c, HRP
provides a carbon efficiency of roughly 87% (Figure 5b) implying conversion of 87% of carbon in the
feeds (CH4 and CO2) to acetic acid, which is 51% higher than that of the CG. Using AspenPlus®[30], the
overall energy required for HRP and CG were calculated. After the integration of the industrial waste heat,
HRP demands 11.1 GJthermal for every metric ton of acetic acid produced. This is 68% lower than the CG
demand, which includes 8 GJthermal for CO2 separation. The high carbon efficiency of HRP manifests in high
productivity, where it produces 2.5 times as much acetic acid as CG, per GJ of feed fuel. After accounting
for HRP carbon credit for CO2 utilization, simulation results suggest that production of 1 metric ton of
acetic acid results in a negligible net CO2 production (0.022 metric tons) in the HRP case. In contrast, CG
leads to 2.1 metric tons of CO2 emission for each ton of acetic acid produced.

2.4. Potential Impacts

The U.S. industrial sector accounts for approximately one third of all energy used in the country; consuming
approximately 32 quadrillion Btu of energy every year[1,14]. Industrial waste heat refers to energy that is
generated in industrial processes without being put to practical use. This accounts for roughly 40% of total
annual industrial energy consumption[14]. Capturing and reusing this waste heat is an effective approach to
improve the overall energy efficiency and lowering CO2 emissions, as it is an emission-free substitute for
fossil fuels or electricity. The high temperature waste heat sources (1200-1700oC range)[1,14], which are
listed in Table 2, are exploitable considering their high quality from a thermodynamic second-law
standpoint. These major industrial processes consume a combined total of 120 TBtu of energy every year,
out of which approximately 30% ends up in waste heat. Assuming 50% of the sensible heat in the waste
heat stream is utilized, there is an annual supply of 18.6 TBtu of energy, efficiently exploitable at an
operating temperature of 650oC for HRP. The upstream reactor configuration of HRP demands 5.33 MBtu
of energy for every metric ton of produced acetic acid, which can be sufficed by supplying the available
waste heat. Considering the carbon efficiency of the process, 0.85 kg of CO2 is utilized for the production
of 1 kg of acetic acid. This makes HRP capable of utilizing roughly 3 million metric tons of CO2 to
manufacture 3.5 million metric tons of acetic acid each year via waste heat integration. The combined usage
of the abundant (i) high temperature industrial waste heat, (ii) methane and (iii) CO2 offers an excellent
opportunity for a sustainable process. The superior low temperature redox performance of Rh/LCO enables
efficient waste heat integration, making HRP an attractive route for sustainable CO2-utilization and
chemical production.

3. Conclusion
We report a hybrid redox process (HRP) that performs CO2-splitting and methane partial oxidation (POx)
in a synergistic two-step, thermochemical redox scheme. Ceria, being a promising material for such a redox
cycle, is modified by La-substitution and Rh-promotion. DFT calculations indicate that La substitution can
decrease the oxygen vacancy formation energy and migration barriers in both the bulk and surface of the
fluorite structured oxide. The reduced Rh/LaCeO3.5-x is capable of converting almost 100% of the CO2 to
CO at temperatures below 700oC. In the subsequent step, it converts 90% of methane selectively to syngas
with a 2:1 H2: CO ratio that is ideal for methanol synthesis, thus demonstrating superior methane partial
oxidation at low temperatures. The ability to operate redox cycles at low temperatures allows for efficient
integration of the industrial waste heat, which is an excellent emission-free energy source. Syngas and CO
are crucial feedstocks, for the production of methanol followed by a variety of carbonyl-containing products
like mono-ethylene glycol, acetic anhydride or acetic acid. When compared with the conventional coal-
based route, HRP offers a 68% reduction in the overall energy demand and 99% reduction in net CO2
emission for acetic acid production. Moreover, HRP has the potential to produce a variety of high-value
chemicals via carbonlyation. The ability to produce chemicals with high market volume and important
industrial applications makes HRP a promising avenue for efficient and economical CO2 utilization. At the
same time, conversion of low-cost natural gas, along with integration of the waste heat, can reduce the cost
of producing acetic acid and the associated CO2 production. HRP has the potential to utilize roughly 3
million metric tons of CO2 every year from the three major industrial waste heat sources, making it a
potentially suitable choice for sustainable CO2 fixation.

4. Experimental Procedures

4.1. Material synthesis

Two redox catalysts are synthesized for this study: CeO2 and LaCeO4-x. Samples are prepared using a
modified Pechini method. Stoichiometric amounts of nitrate salts of lanthanum [La(NO3)3, 6H2O] and
cerium [Ce(NO3)3, 6H2O] are dissolved in deionized water and stirred for 30 minutes at room temperature
to form a homogeneous solution. Citric acid, with 2.5 times the total moles of cations, is added and the
solution is kept under stirring for another half an hour at 40oC. Ethylene glycol with a molar ratio of 1.5: 1
with respect to citric acid is then added to the solution, and the temperature of the solution is increased to
80oC under stirring, until a homogeneous gel is formed. The gel is then dried overnight at 100oC and
annealed at 1000oC for 8 hours.
The surface of the synthesized redox catalysts is also promoted with rhodium using a wetness impregnation
method (0.5 wt. % Rh). Rhodium nitrate salt is dissolved in deionized water and appropriate stoichiometric
amounts of the solution are added to the samples in multiple steps. The impregnated samples are then dried
at 100oC for 4 hours and sintered at 1000oC for 6 hours. The prepared redox material is crushed and sieved
in the particle range of 90-225μm.

4.2. Material characterization

Powder X-ray diffraction (XRD) experiments are done to confirm the formation of the desired crystal
phases. XRD patterns are obtained using a Rigaku SmartLab X-ray diffractometer with Cu-Kα (λ=0.1542)
radiation operating at 40 kV and 44 mA. A stepwise approach with a step size of 0.1° and residence time
of 2.5 seconds at each step in 20-80 angle range (2θ) is used to generate the XRD patterns. The phase
identification is done using the HighScorePlus software from PANalytical. Fresh (as-prepared) and spent
(cycled) samples are analyzed.

To examine phase properties, in-situ XRD is conducted during methane temperature programmed reduction
(TPR) using an Empyrean PANalytical XRD using a similar Cu-Kα radiation operating at 45 kV and
40 mA. A 2θ range of 20–80o is used at a ramp rate of 0.1° holding each step for 0.1 s. A TPR experiment
is conducted in the in-situ XRD by heating the oxidized sample at 5 °C/min in 5% CH4 (balance N2). XRD
scans are conducted the entire length of the ramp from room temperature to 750°C.

X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific Inc.) with an Al-Ka X-ray source at an
operating voltage of 20 kV and a current of 10 mA is used to analyze near surface elemental compositions.
Survey spectra and the single element spectra were collected with 20 eV and 100 eV pass energy.

4.3. Redox experiments

Redox experiments are performed in a fixed bed configuration using a quartz U-tube reactor (1/8’’ ID) with
500 mg sample loading. Total flow rate is 25 ml min-1 with 10 vol% CH4 (reduction for 2 min), 2% CO2
(oxidation) and balance Ar. For the material comparison (Figure 4a), total flow is 50 ml min-1. The reactor
is purged with Ar after each half cycle to prevent mixing of the reactive gases. The samples are held in
place using quartz wool to prevent fluidization of the particles. The reactor effluent is analyzed using a
quadrupole mass spectrometer (Cirrus 2, MKS). To ensure the accuracy of the qualitative analysis, a single
point calibration is performed using a standard calibration gas (1% H2, 1% CH4, 1% CO, and 1% CO2,
balance Ar) before each experiment.

The reducibility of the samples in methane is characterized and tested through TPR experiments. The
experiments are performed in a quartz U-tube reactor in 10 vol% methane (balance Ar). The temperature is
increased from 350oC to 950oC with a 10oC min-1 ramping rate and kept isothermal at 950oC for 15 minutes.
The samples are oxidized at 950oC in 20% oxygen (ultra-dry, balance Ar) prior to reduction to keep them
completely oxidized.

4.4. DFT Computation Methods

All Density Functional Theory (DFT) calculations were implemented by the Vienna ab initio Simulation
package (VASP)[45] with the frozen-core all-electron projector augmented wave (PAW) model[46] and
Perdew-Burke-Ernzerhof (PBE) functions[47]. A kinetic energy-cutoff for the plane-wave expansion of the
electronic wave function is set to be 450 eV. The force and energy convergence criterion are set to be 0.01
eV/Å and 10-5 eV respectively. A Gaussian smearing of 0.1 eV is applied for optimization. A k-point grid
with 7 × 7 × 7 Gamma-centered mesh is chosen for the CeO2 and LaCeO3.5 bulk unit cell. For all the
supercells, a corresponding number of k-points is used to keep the k-mesh spacing constant across different
structures. The effect of Ce (4f) electron correlation is corrected by considering on-site Coulomb (U) and
exchange (J) interactions following the GGA + U approach[48] with Ueff = 5 eV[49]. Charge distribution
analysis is examined with the DDEC06 method, which is a refinement of the Density Derived Electrostatic
and Chemical (DDEC) approach[50]. The climbing image nudged elastic band (CI-NEB) method is applied
to search the transition state structures[51]. To avoid artificial interactions between adjacent slabs for the
surface models, a vacuum spacing of ~15Å is used.

Atomic model of pristine CeO2 unit cell (Figure S1a and b) is built with the optimized lattice constant of
5.418 Å, which is very close to the experimental reports (5.411 Å)[52]. LaCeO3.5 unit cell is built by replacing
two Ce atoms by La atoms, and then removing one O atom to form an oxygen vacancy (VO). Due to
symmetry, the selection of the La doping and VO sites in the CeO2 unit cell is unique. Subsequently, a 2 ×
2 × 2 supercell of LaCeO3.5 bulk structure is built to avoid artificial interactions induced by the periodic
boundary condition. To calculate ΔEV, all 7 non-equivalent sites depicted in Figure S2 are considered and
the most stable one (O1 site) is chosen for subsequent calculations. Upon the formation of the vacancy, all
adjacent 6 O atoms are allowed to migrate to the vacancy site, and the corresponding reaction energies are
calculated (Table S2). Among these possible oxygen mobility pathways, the migration of O4 and O6 are
no longer considered due to their very large reaction energies, which indicate even larger barriers.
Transition state calculations of the migration of the rest oxygen atoms (O2, O3, O4 and O6) are performed
(Figure S3). For the surfaces, to closely reproduce the experimental La:Ce ratio (64.1:35.9 ≈ 1.79)
obtained through XPS, one of the outermost Ce atoms is modified to La artificially, making the La:Ce ratio
of the first 6 layers to be 5:3 ≈ 1.67. Then a 2 × 1 × 1 supercell surface model is built, which shows only
two non-equivalent surface oxygen atom sites (A and B, Figure S4). For the migration of adjacent sub-
surface oxygen to the two vacancy sites (Figure S5), the corresponding reaction energies of all possible O
migration pathways are calculated (Table S3), in which the migrations of O2 or O4 to B site are no longer
considered due to the very large reaction energies. The transition state calculations of the rest oxygen
migration pathways are calculated as shown in Figure S6.

4.5. Process simulations

Overall, the process is split into three sections: syngas generation, methanol (MeOH) synthesis, and acetic
acid (AcOH) reactor. For the overall process, methane and CO2 are the feed and the product is AcOH.

Methanol is synthesized with syngas (with CO2) using a commercial Cu/Zn/Al2O3 catalyst. A packed
tubular reactor approximated using an RStoic model in AspenPlus®, is used. The MeOH reactor operates
at 265oC and 105 bar, with a 64% yield of CO⟶ MeOH and 17% yield of CO2⟶ MeOH. The synthesis
of AcOH using MeOH and CO is done using the CativaTM process. This commercial process produces
AcOH by the carbonylation of MeOH, via an iridium-containing complex and hydrogen iodide (HI). A two
reactor scheme has been used for modeling the process. The reactors are set at 190oC and 22 barg and
modeled as RStoic reactors. The first reactor involves reactions of MeOH with HI to give AcOH via AcI
(methyl iodide). HI is replenished and is not consumed in the process. The second reactor assumes the
formation of byproducts like propionic acid and CO2 via WGS reaction, from the remaining MeOH.
HRP is compared with the traditional coal-slurry gasifier (CG) process to obtain syngas, followed by MeOH
synthesis with subsequent AcOH synthesis. Coal is assumed to be composed of pure carbon (C) with feed-
water to be 40% more than the feed coal (by moles). An efficiency of 96% is assumed for the air separation
unit (ASU), with the final O2: C = 0.5. An adiabatic RGibbs reactor model is used for the entrained bed
gasifier, operating at 30 bar. All the following units are similar to the HRP process. The end-product is
AcOH. H2 and CO2 removal are required to produce syngas and CO streams. CO2 removal demands
additional energy, which is accounted for. The separated H2 is used as a fuel for the process energy
requirements. Key assumptions for these AspenPlus® models are based on a previous work and are
presented in the supplementary information[30]. Net CO2 production considers the contributions from natural
gas/coal upstream and energy conversion facility. The estimated upstream natural gas emission based on a
US-DOE report is 9.1 g CO2/MJ for CH4 and 10 g CO2/MJ for coal. The emission analysis from the two
processes is based on AspenPlus® model results. For both the processes, it is assumed that the thermal
energy requirements are met by burning the necessary amount of methane. While CO2 separated from coal
gasification syngas could be further purified and compressed for sequestration, this would incur significant
additional cost and energy consumptions. In the present study, we did not include these additional steps and
incorporated the CO2 in syngas into the net CO2 production calculations. For HRP, CO2 usage as a feed
leads to a CO2 credit.

5. Acknowledgements

This work was supported by the U.S. Department of Energy (Awards DE-FE0031703), National Science
Foundation (NSF CBET-1510900), and the North Carolina State University Kenan Institute for
Engineering, Technology and Science. The authors acknowledge the use of the Analytical Instrumentation
Facility (AIF) at North Carolina State University, which is supported by State of North Carolina and the
National Science Foundation (Award ECCS-1542015).

6. Author Contributions

F.L. conceived and supervised the study. V.P.H. and F.L. designed the study and wrote the manuscript.
X.W. carried out the DFT calculations. R.D carried out the XPS experiments and analyzed the results.
V.P.H. carried out the experiments with the help of C.P. and analyzed the data. V.P.H. carried out the
AspenPlus® simulation and collected the XRD data. B.T. and R.G. assisted in the process simulations.

7. Declaration of Interests

The authors declare no competing financial interest.

8. References
[1] Committee on Developing a Research Agenda for Utilization of Gaseous Carbon Waste Streams,
Board on Chemical Sciences and Technology, Division on Earth and Life Studies, National
Academies of Sciences, Engineering, and Medicine, Gaseous Carbon Waste Streams Utilization:
Status and Research Needs, National Academies Press, Washington, D.C., 2019.
[2] A. W. Zimmermann, R. Schomäcker, Energy Technology 2017, 5, 850.
[3] A. Zimmermann, n.d., 48.
[4] B. Smit, A.-H. A. Park, G. Gadikota, Frontiers in Energy Research 2014, 2, DOI
10.3389/fenrg.2014.00055.
[5] E. E. Stangland, Annual Review of Chemical and Biomolecular Engineering 2018, 9, 341.
[6] C. Mesters, Annual Review of Chemical and Biomolecular Engineering 2016, 7, 223.
[7] E. Alper, O. Yuksel Orhan, Petroleum 2017, 3, 109.
[8] Z. Sun, T. Ma, H. Tao, Q. Fan, B. Han, Chem 2017, 3, 560.
[9] O. K. Varghese, M. Paulose, T. J. LaTempa, C. A. Grimes, Nano Letters 2009, 9, 731.
[10] W. C. Chueh, C. Falter, M. Abbott, D. Scipio, P. Furler, S. M. Haile, A. Steinfeld, Science 2010, 330,
1797.
[11] D. Arifin, V. J. Aston, X. Liang, A. H. McDaniel, A. W. Weimer, Energy & Environmental Science
2012, 5, 9438.
[12] A. H. Bork, E. Povoden‐Karadeniz, J. L. M. Rupp, Advanced Energy Materials 2017, 7, 1601086.
[13] C. Muhich, A. Steinfeld, J. Mater. Chem. A 2017, 5, 15578.
[14] I. Johnson, W. T. Choate, A. Davidson, Waste Heat Recovery. Technology and Opportunities in U.S.
Industry, BCS, Inc., Laurel, MD (United States), 2008.
[15] S. U. Nimbalkar, A. Thekdi, B. M. Rogers, O. L. Kafka, T. J. Wenning, Technologies and Materials
for Recovering Waste Heat in Harsh Environments, 2014.
[16] J.-B. Peng, H.-Q. Geng, X.-F. Wu, Chem 2018, 0, DOI 10.1016/j.chempr.2018.11.006.
[17] R. A. Kerr, Science 2010, 328, 1624.
[18] J. A. Moulijn, A. van Diepen, M. Makkee, Chemical Process Technology, John Wiley & Sons Inc.,
Chichester, West Sussex, United Kingdom, 2013.
[19] K. C. Waugh, Catalysis Letters 2012, 142, 1153.
[20] J. H. Jones, n.d., 12.
[21] L.-S. Fan, Chemical Looping Systems for Fossil Energy Conversions, Wiley-AIChE, Hoboken, NJ,
2010.
[22] M. Tang, L. Xu, M. Fan, Applied Energy 2015, 151, 143.
[23] M. Rydén, A. Lyngfelt, T. Mattisson, Fuel 2006, 85, 1631.
[24] J. Adánez, A. Abad, T. Mendiara, P. Gayán, L. F. de Diego, F. García-Labiano, Progress in Energy
and Combustion Science 2018, 65, 6.
[25] P. Furler, J. Scheffe, M. Gorbar, L. Moes, U. Vogt, A. Steinfeld, Energy & Fuels 2012, 26, 7051.
[26] P. T. Krenzke, J. R. Fosheim, J. H. Davidson, Solar Energy 2017, 156, 48.
[27] P. G. Loutzenhiser, A. Meier, A. Steinfeld, Materials; Basel 2010, 3, 4922.
[28] S. Bhavsar, M. Najera, G. Veser, Chemical Engineering & Technology 2012, 35, 1281.
[29] M. Najera, R. Solunke, T. Gardner, G. Veser, Chemical Engineering Research and Design 2011, 89,
1533.
[30] J. Zhang, V. Haribal, F. Li, Science Advances 2017, 3, e1701184.
[31] K. Otsuka, T. Ushiyama, I. Yamanaka, Chem. Lett. 1993, 22, 1517.
[32] K. Otsuka, Y. Wang, E. Sunada, I. Yamanaka, Journal of Catalysis 1998, 175, 152.
[33] V. Bellière, G. Joorst, O. Stephan, n.d., 7.
[34] “Global Acetic Acid Market 2018-2022| Increasing Demand for Methanol to Drive Growth|
Technavio,” can be found under
https://www.businesswire.com/news/home/20181003005834/en/Global-Acetic-Acid-Market-2018-
2022-Increasing-Demand, 2018.
[35] L. Zhang, J. Meng, F. Yao, W. Zhang, X. Liu, J. Meng, H. Zhang, Inorganic Chemistry 2018, 57,
12690.
[36] X. Zhu, K. Li, Y. Wei, H. Wang, L. Sun, Energy Fuels 2014, 28, 754.
[37] X. Zhu, H. Wang, Y. Wei, K. Li, X. Cheng, Mendeleev Communications 2011, 21, 221.
[38] R. T. Dirstine, Journal of The Electrochemical Society 1979, 126, 264.
[39] Z. Yang, G. Luo, Z. Lu, K. Hermansson, J. Chem. Phys. 2007, 127, 074704.
[40] H.-Y. Li, H.-F. Wang, Y.-L. Guo, G.-Z. Lu, P. Hu, Chemical Communications 2011, 47, 6105.
[41] Y.-L. Lee, J. Kleis, J. Rossmeisl, Y. Shao-Horn, D. Morgan, Energy Environ. Sci. 2011, 4, 3966.
[42] W. T. Hong, M. Risch, K. A. Stoerzinger, A. Grimaud, J. Suntivich, Y. Shao-Horn, Energy Environ.
Sci. 2015, 8, 1404.
[43] Z. Liu, Y. Sun, X. Wu, C. Hou, Z. Geng, J. Wu, K. Huang, L. Gao, S. Feng, CrystEngComm 2019,
21, 1534.
[44] A. Shafiefarhood, J. Zhang, L. M. Neal, F. Li, Journal of Materials Chemistry A 2017, 5, 11930.
[45] G. Kresse, J. Furthmüller, Phys. Rev. B 1996, 54, 11169.
[46] P. E. Blöchl, Phys. Rev. B 1994, 50, 17953.
[47] null Perdew, null Burke, null Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[48] S. L. Dudarev, G. A. Botton, S. Y. Savrasov, C. J. Humphreys, A. P. Sutton, Phys. Rev. B 1998, 57,
1505.
[49] C. W. M. Castleton, J. Kullgren, K. Hermansson, The Journal of Chemical Physics 2007, 127,
244704.
[50] T. A. Manz, N. G. Limas, RSC Adv. 2016, 6, 47771.
[51] G. Henkelman, B. P. Uberuaga, H. Jónsson, The Journal of Chemical Physics 2000, 113, 9901.
[52] L. Gerward, J. Staun Olsen, L. Petit, G. Vaitheeswaran, V. Kanchana, A. Svane, Journal of Alloys
and Compounds 2005, 400, 56.
 Acetic Acid
a) CO2 MeOy CO
Synthesis

CH3OH
x>y

Waste Heat
Methanol
MeOx CO +H2 Synthesis
CH4

Upstream Downstream

b)

Ethylene Ube Process

CO2

Glycol

CO Acetic Acid

HRP
Upstream
CO + H2 Methyl Acetic
Methanol
Acetate Anhydride
Reactive Distillation Carbonylation
CH4

Gasoline
Fischer-Tropsch
Diesel
Synthesis
Light Olefins

Figure 1: a) Schematic of HRP (considering acetic acid as the final product); b) Paths to synthesize
chemicals from CO and syngas, obtained via HRP
 a) b)

c)
d)
CeO2 (111)
LaCeO3.5 (111)

Ce La O Selected O VO

Figure 2: a) Computed energies of oxygen vacancy formation (ΔEV); b) Energy barriers for oxygen
vacancy migration for bulk and (111) surface of CeO2 and LaCeO3.5; c) Energy potential profile along
the most favorable oxygen migration pathway on the (111) surfaces of CeO2 and LaCeO3.5; d) Electronic
projected density of states (PDOS) of CeO2 and LaCeO3.5
 a)
Rh-promoted samples
LaCeO3.5
CeO2 b)
% Relative intensity

pure CeO2

25 35 45 55 65
2θo

c)
MS Signal (Mass 28)

CeO2

Rh/LCO
LCO

350 450 550 650 750 850 950

Temperature (oC)

Figure 3: a) X-ray diffractogram for Rh-promoted ceria and LCO; b) CH4 TPR + in-situ XRD on
Rh/LCO; c) CH4 TPR of Rh/LCO, LCO and ceria
a) Rh/Ceria Rh/LCO b) 2.5
100
H2
H2 CO
CO CH4
CH4 CO2
CO2
90
2.0
80
70
X (CH4) (%)

Flow (ml/min)
1.5
60
50
40 1.0
30
20 0.5
10
0
0.0
500 600 700 900
0 5 10 15 20 25
Temperature (oC) Time elapsed (min)
c) d)
X (CH4) S (CO) X (CO2) Syngas yield
100 1.0
Rh/Ceria

Syngas yield (mmol/g redox catalyst)

90 0.9 (as prepared)
80 0.8 Rh/LCO
70 0.7 (as prepared)

60 0.6
%

50 0.5

40 0.4

30 0.3 Rh/LCO
Rh/LCO (cycled and oxidized)
20 0.2
(reduced)
10 0.1

0 0.0
12 13 14 15 16 17 18 19 20 21 22 23 24 25
Cycle # 25 30 35 40 45 50 55 60 2θ

Figure 4: a) Performance comparison of Rh/LCO with Rh/Ceria for methane conversion; b)

Representative product distribution for Rh/LCO under the HRP scheme at 650oC; c) Redox cycle results
of Rh/LCO at 650oC: CH4 and CO2 conversion (X), CO selectivity (S) and Syngas yield; d) Phases of as-
prepared and cycled Rh/LCO, showing stability over multiple redox cycles
 a) Coal + O2 + H2O CH4
CO2
via Coal via Chemical
Gasification Looping

CO production

CO production

Waste Heat
production

production
combustion
from CH4

Syngas

Syngas
CH4

Methanol
synthesis

Acetic Acid
Synthesis

Acetic Anhydride
Synthesis

Downstream Chemical Production

b)

y 23.1
/ 2.5x increase
l 58.6

Figure 5: a) Schematic of HRP and CG (coal gasification), with similar downstream units [CG route
requires H2 and CO2 removal to obtain syngas and pure CO]; b) Comparison of HRP with CG, using
various parameters
Table 1: Near surface atomic compositions of the redox catalysts (oxygen free basis, numbers shown in
parentheses correspond to expected compositions based on bulk stoichiometry)

Sample La 3d (%) Ce 3d (%) Rh 3d (%)

CeO2 100.0
LCO 64.1 (50) 35.9 (50)
Rh/LCO 61.0 (49.6) 34.7 (49.6) 4.3 (0.8)

Table 2: Exploitable High Temperature Waste Heat sources14 and CO2-utilization capacity

Temperature Waste Heata CO2 utilizable

Industry Furnace
(oC) (TBtu/year) (tonne/year)

Basic Oxygen Furnace

Steel 1700 13.6 2,162,207
(BOF)
Glass Oxy-fuel 1427 2.1 335,102
Electric Arc Furnace
Steel 1200 2.9 462,760
(EAF)
a
utilizable for an operation at 650oC
Graphical Abstract