Document [original]

Effect of Conversion, Temperature and Feed Ratio on In2O3/

In(OH)3Phase Transitions in Methanol Synthesis Catalysts:

A Combined Experimental and Computational Study

Philipp Kampe,[a] Anne Wesner,[a] Patrick Schühle,[b] Franziska Hess,*[c] and Jakob Albert*[a]

Catalytic hydrogenation of CO2to methanol has attracted lots

of attention as it makes CO2useable as a sustainable carbon

source. This study combines theoretical calculations based on

the dummy catalytic cycle model with experimental studies on

the performance and degradation of indium-based model

catalysts for methanol synthesis. In detail, the reversibility of

phase transitions in the In2O3/In(OH)3system under industrial

methanol synthesis conditions are investigated depending on

conversion, temperature and feed ratio. The dummy catalytic

cycle model predicts a peculiar degradation behavior of In(OH)3

at 275°C depending on the water formed either by methanol

synthesis or the competing reverse water-gas-shift reaction.

These results were validated by dedicated experimental studies

confirming the predicted trends. Moreover, X-ray diffraction

and thermogravimetric analysis proved the ensuing phase

transition between the indium species. Finally, the validated

model is used to predict how hydrogen drop out will affect the

stability of the catalyst and derive practical strategies to prevent

irreversible catalyst degradation.

Introduction

The economic and ecologic supply of the growing energy

demand is a global challenge. In view of the climate change the

future energy supply has to be based on renewable energies as

well as alternative feedstocks.[1] Electrolytically produced H2is of

particular importance as a molecule to store energy and drive

chemical catalytic processes. However, fluctuations in the H2

flux due to the intermittent nature of renewable energy sources

(hydrogen drop out) can impose severe strain on catalysts due

to the change of reactant ratio in the feed and subsequent

drop of conversion and catalyst bed temperature. Such a drop

out may cause irreversible degradation, from which the catalyst

may not be able to recover without additional reactivation

steps after hydrogen supply is restored. In order to stabilize

catalysts operating under intermittent conditions against such

degradation phenomena, a deep understanding of how tem-

perature, conversion and reactant feed ratio affect catalyst

stability is required. This enables the integrated design of new

catalysts and of the linked processes to prevent irreversible

degradation due to hydrogen drop out.

Besides hydrogen, a carbon source is required for methanol

synthesis. Due to its significant contribution to the greenhouse

effect, a special focus has to be drawn on the utilization of CO2

from emissions. One possible strategy hereby is carbon capture

and utilization.[2] CO2is an abundant, non-toxic and renewable

chemical. It is thermally stable and chemically inert with a

standard formation enthalpy of 394 kJ/mol. Consequently, a

large energy input for chemical conversion is necessary.[3]

Therefore, a promising way is the catalytic valorization of CO2

with electrolytically produced H2from renewable energies to

methanol, as shown in Eq. 1:[4]

CO2þ3 H2)

*CH3OH þH2O

DHR¼  50 kJ=mol (1)

In the commercialized process, methanol is produced from

syngas (CO/CO2/H2) via a Cu/ZnO/Al2O3catalyst. Under typical

reaction conditions (T=200–300°C, p=50–100 bar and CO/

CO2/H2of 28/2/70) a selectivity of SMeOH =30-70% can be

achieved.[5] Besides the exothermic hydrogenation of CO/CO2

(Eq. 1–2) into methanol, also the endothermic reverse water-

gas-shift (RWGS) reaction (Eq. 3) plays a role in the context of

methanol synthesis and CO2hydrogenation:[6]

[a] P. Kampe, A. Wesner, J. Albert

Institute of Technical and Macromolecular Chemistry

Universität Hamburg

Bundesstraße 45

20146 Hamburg (Germany)

E-mail: [email protected]

[b] P. Schühle

Institute of Chemical Reaction Engineering

Friedrich-Alexander-Universität Erlangen-Nürnberg

Egerlandstraße 3

91058 Erlangen (Germany)

[c] F. Hess

Institut für Chemie

Technische Universität Berlin

Straße des 17. Juni 124

10623 Berlin (Germany)

E-mail: [email protected]

Supporting information for this article is available on the WWW under

https://doi.org/10.1002/cplu.202300425

an open access article under the terms of the Creative Commons Attribution

Non-Commercial NoDerivs License, which permits use and distribution in

any medium, provided the original work is properly cited, the use is non-

commercial and no modifications or adaptations are made.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 124/136] 1

ChemPlusChem

www.chempluschem.org

Research Article

doi.org/10.1002/cplu.202300425

CO þ2 H2)

*CH3OH

DHR¼  91 kJ=mol (2)

CO2þH2)

*CO þH2O

DHR¼ þ41 kJ=mol (3)

The RWGS reaction as undesired pathway is the thermody-

namically preferred one under the applied reaction conditions

and favored by high temperatures. For shifting the equilibrium

to higher methanol selectivities, higher pressures and lower

temperatures are necessary.[7] However, not only thermodynam-

ics but also kinetics and stability of the catalysts define the

efficiency of a renewable methanol synthesis process. Kinetic

modeling is widely used for the understanding of CO2hydro-

genation and therefore, helps to develop efficient catalysts and

reaction concepts.[8]

Recently, various In2O3based catalysts for CO2hydrogena-

tion have been developed.[9,10] Pure In2O3shows only a low

selectivity to methanol of SMeOH =63% (T=300°C, p=50 bar)

due to the predominant competing RWGS reaction.[11] The

superior selectivity towards RWGS can be explained by the

higher activation energy for the synthesis of methanol

compared to RWGS.[12] In further DFT and experimental studies,

oxygen vacancies turned out to be the active sites for methanol

formation, whereby one oxygen vacancy surrounded by three

indium atoms leads to activation of CO2and the heterolytic

splitting of H2.[12,13,14] Despite that, In2O3has only a poor ability

for H2activation.[14,15] Previous experimental studies have further

shown that the surface of bulk In2O3is reduced at temperatures

above 220°C, leading to metallic indium species in a H2-rich

atmosphere.[12,16]

In2O3based catalysts possess different chemical stability in

CO2hydrogenation reactions depending on the reaction

conditions. Tsoukalou et al. reported three distinct catalytic

regimes (activation, stable performance, deactivation) during

CO2hydrogenation by combining X-ray absorption spectro-

scopy (XAS), X-ray powder diffraction (XRD) and in situ trans-

mission electron microscopy (TEM). Hereby, a reductive

amorphization of the In2O3-x nanocrystallites progresses with

time on stream, leading to an over-reduction to molten In0

being responsible for the deactivation.[17]

Aside from over-reduction to In0, In2O3-based catalysts also

suffer from partial conversion to less active In(OH)3, or from

poisoning with feed gas impurities, such as sulfur and nitrogen-

containing compounds.[10] For the phase transition between

In2O3and In(OH)3species by hydration, the composition of the

surrounding process gas plays a decisive role. Due to the

thermodynamics of hydration reactions, we propose that the

propensity of In2O3to hydrate is intimately coupled to temper-

ature and conversion in the catalyst bed. In this contribution,

we investigate the influence of important reaction parameters

on the stabilities of In2O3and In(OH)3under typical methanol

synthesis conditions by applying a computational model based

on Gibbs free energies of reaction. We subsequently test the

predictions by dedicated model experiments employing pure

In2O3and In(OH)3catalysts. We further discuss possible reactor

operation strategies to prevent catalyst degradation by phase

transformation when hydrogen drop out occurs.

Results and Discussion

Theoretical modeling

Modeling the phase transitions of catalysts in fixed-bed reactors

is challenging because the gas phase surrounding the catalyst

is never at equilibrium. Likewise, the catalyst can never reach an

equilibrium state and this is one of the core principles of

catalytic processes. However, despite the lack of a clearly

defined thermodynamically stable state of the catalyst, a steady

state exists, which can be either static or dynamic, e.g.,

oscillating. In catalysts that undergo phase transitions, the

steady state is determined by the kinetics of these phase

transitions and the principle of minimum entropy production. If

no information is available about how these phase transitions

occur, something can still be learned about the stability of the

phases based on the reaction conditions (like temperature, feed

ratio, conversion, and selectivity): this can be achieved by

studying the driving forces, specifically the changes in Gibbs

free energy, and how they relate to the reaction conditions.

However, when the gas feed is not at equilibrium, the

degradation and reactivation proceed via different reactions

due to the Gibbs free energy difference between the reaction

products and reactants of the overall catalytic reaction. The

reactions associated with degradation and reactivation then

form a closed catalytic cycle, where the degraded catalyst

represents an intermediate. In the case of In2O3/In(OH)3, the

hydration of In2O3, can occur via one of two pathways:

methanol formation (Eq. 4) or CO formation (Eq. 5) through

RWGS:

0:5 In2O3þ1:5 CO2þ4:5 H2!InðOHÞ3þ1:5 CH3OH (4)

0:5 In2O3þ1:5 CO2þ1:5 H2!InðOHÞ3þ1:5 CO (5)

Reactivation of In(OH)3occurs via a simple dehydration

reaction by releasing water:

InðOHÞ3!0:5 In2O3þ1:5 H2O(6)

Adding Eq. 4 and Eq. 6, or Eq. 5 and Eq. 6 results in two so-

called dummy catalytic cycles as illustrated in Figure 1a). Note

that for the application of this model, it is not a prerequisite

that the dummy catalytic cycle describes the dominant reaction

path, i.e., that the catalytic reaction proceeds via a solid-state

reaction. However, the utility of this concept becomes immedi-

ately obvious if we recall the physical meaning of the Gibbs free

energy as the maximal amount of work that a chemical reaction

can exert. This insight can be used to evaluate the thermody-

namic viability of proposed reaction mechanisms based on their

intermediates with the highest and lowest Gibbs free energy. In

the case of catalyst degradation, the Gibbs free energy of the

intermediate in relation to the reactants indicates how rever-

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 125/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

sible a phase transition under nonequilibrium conditions is, as

illustrated in Figure 1b).

Here, the level on the left represents the active catalyst,

In2O3, along with the two reactants. The two levels depicted to

the far right represent again the active catalyst with the two

possible products of the reaction, CH3OH/H2O (orange), and CO/

H2O (red). Methanol formation and RWGS have different ΔrG as

indicated by black arrows connecting the reactant and product

Gibbs free energies, which places the final states at different

levels. The ΔrG for the two reactions are a function of

conversion, selectivity, temperature, and reactant feed ratio.

The intermediates for the two reactions are In(OH)3and either

CH3OH or CO. Similarly, we can compute the hypothetical Gibbs

free energy of this intermediate. If the intermediate is above

the top dashed line, In(OH)3will not be formed to a large

extent, i.e., In2O3is stable. If the intermediate is below the

bottom dashed line, In(OH)3is stable, i.e., the formation of

In(OH)3will be irreversible. If the CH3OH or CO intermediate is

in between the first and second line or between the first and

third line, respectively, the formation of In(OH)3is reversible.

This means in practice that the catalyst enters a steady state,

which can be either one or the other, and stabilization of either

phase can be accomplished by tuning the kinetics of these

phase transitions, for instance, by the appropriate selection of

promotors or catalyst support.

We note that the application of this model, while not

limited to any particular material or catalytic reaction,[18] can

also be applied to mixed oxides,[19] and possibly also to

supported catalysts. This requires, however, that the thermody-

namic data are known, i.e., that the influence of a catalyst

support on the free enthalpy of reaction, can be estimated with

sufficient reliability. Similarly, the kinetics of phase transitions

can be taken into consideration quantitatively, for instance, to

understand the influence of particle morphology on the life

time of a catalyst.[20] Kinetics can play a crucial role in

determining the long-term stability of thermodynamically

unstable materials in catalysis and they also determine the

steady-state phase fractions if degradation is reversible.[21] In

the case of In(OH)3/In2O3, we observe no remarkable differences

between our expectations from the purely thermodynamic

treatment and our experimental observations. As a conse-

quence, considering the kinetics in addition would add little

additional insight at the present stage.

In the following, stability calculations regarding In2O3under

typical reaction conditions of p=75 bar and a CO2/H2ratio=1/

3 are undertaken. Therein, the influence of temperature in

methanol synthesis without RWGS and with RWGS reaction on

the hydration/dehydration behavior of In2O3are investigated.

These results were validated by experimental investigations in a

fixed bed reactor due to different positions of the bed and a

two-segment configuration. The model was applied to predict

the influence of hydrogen drop out on the stability of the

catalyst.

Stability of In2O3under methanol synthesis without RWGS

Figure 2 shows the Gibbs free energy curves for three different

reaction temperatures (200°C, 250°C, 300°C) for a stoichiomet-

ric p(H2)/p(CO2) mixture of 3:1 with a total pressure of 75 bar

(i.e., p0(CO2)=18.75 bar, p0(H2)=56.25 bar) as a function of the

CO2conversion, assuming that only CH3OH and H2O are formed

as reaction products, i.e., the RWGS does not take place.

In each graph, the blue curve represents the free enthalpy

of the dehydration reaction as given by Eq. 6. Negative values

mean that dehydration of In(OH)3is thermodynamically favor-

able and will occur to a large extent, while for positive values,

dehydration is not thermodynamically favorable, but may still

occur to a smaller extent. Rehydration of In2O3due to the

formation of CH3OH via Eq. 4 will usually assume negative

values, as long as the conversion remains below a certain

threshold that depends on the reaction temperature. Naturally,

the overall Gibbs free energy of the CO2reduction is given by

the sum of dehydration and rehydration, as these reactions

together form a full catalytic cycle (displayed by the green lines

in Figure 2). The equilibrium conversion is indicated by a thick,

black line, which is only visible in the diagram for 300°C

because equilibrium conversions at lower temperatures are

over 30% and, therefore, outside the plot range.

Figure 2a shows the free enthalpy curves for 200°C. Here,

dehydration (blue curve) is exergonic at conversions below

2.3%. Dehydration becomes less favorable with increasing CO2

conversion due to the presence of H2O as a reaction product in

the gas stream, which suppresses the release of H2O from

In(OH)3. The rehydration due to the formation of CH3OH (orange

curve) shows a similar trend, because the CO2content in the

gas stream decreases and the CH3OH content increases with

Figure 1. Illustration of the dummy catalytic cycle. a) Reactions converting

CO2and H2via a two-step reaction into the products (H2O, CH3OH, and CO).

b) Gibbs free energy profiles for dehydration and product formation in the

CH3OH and RWGS pathways.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 126/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

increasing CO2conversion. Therefore, we conclude that dehy-

dration is not expected to occur at conversions over 2.3% at

200°C (red shaded region), while at lower conversion, dehy-

dration to form In2O3is possible, but reversible, because both

dehydration and rehydration are exergonic. This result indicates

that at low temperatures, the phase transition will inevitably

occur at conversions larger than 2.3% due to the formation of

In(OH)3. Because In(OH)3is favored at lower temperatures, this

kind of phase transition may be self-accelerating due to lack of

local heat production and resulting cold spot formation in

partially degraded areas. Furthermore, the local heat production

due to the exothermic CO2hydrogenation reaction depends on

the local reaction rate, which tends to decrease with increasing

reactant conversion. Note that, even at low conversion, the

driving force for dehydration given by the blue curve is rather

small, while the driving force for rehydration (orange curve) is

much larger. This result alone, however, is insufficient to

conclude whether In(OH)3or In2O3will dominate in the steady

state at low conversion. However, we can conclude that the

fraction of In(OH)3in the steady state will increase along the

catalyst bed, i.e., the higher the local conversion, the more

In(OH)3will be present because the driving force for dehydra-

tion quickly approaches zero, while the driving force for

rehydration remains at a large negative value within the whole

range of X�2.3%.

Higher temperatures present a vastly different picture, as

illustrated in Figure 2b, which shows the free enthalpy curves

for 250°C. At higher temperature, dehydration (blue curve) is

possible over a larger range of conversions (X<12.2%), because

higher temperature favors In2O3over In(OH)3. Rehydration

(orange curve) is still exergonic over the whole range of

conversion. It is less favored with increasing conversion due to

the accumulation of the reaction product CH3OH, in the gas

stream, and depletion of the reactants, CO2and H2. This means

that dehydration and rehydration are thermodynamically rever-

sible up to a conversion of 12.2%. At higher conversion (red

shaded area), In(OH)3is thermodynamically stable and will not

be converted to In2O3.

At 300°C (Figure 2c), the situation is quite different because

the dehydration curve now lies below zero over the whole

conversion range, while the curve representing rehydration by

methanol formation is endergonic at X>17%, indicating that

In(OH)3will dehydrate irreversibly, resulting in the formation of

In2O3. At conversions below 17%, rehydration is still possible,

indicating reversible phase transition or coexistence of In2O3

and In(OH)3, with lower conversions favoring In(OH)3. Such

reversible phase transitions may also contribute to faster

sintering or particle reshaping.

The modeling results clearly show that the In2O3catalyst will

display enhanced stability against hydration at higher reaction

temperatures. The influence of conversion on the stability of

In2O3is less straightforward; at low temperatures, the section of

the catalyst bed that comes in contact with the feed at low

conversion (i.e., close to the reactor inlet) will be less

susceptible to phase transformation via In(OH)3formation, while

at high temperatures, the catalyst section in the high

conversion zone, i.e., close to the reactor outlet, will be more

stable.

Stability of In2O3under methanol synthesis with RWGS

A realistic indium-based catalyst applied in CO2reduction does

not have perfect selectivity towards methanol. In practice, the

RWGS will dominate in the hot zone of the catalyst bed as it is

thermodynamically favored over methanol formation at high

temperatures. Therefore, the gas stream can be expected to

contain additional H2O formed by the RWGS; since the H2O

content in the gas stream is one of the main factors in the

In2O3/In(OH)3equilibrium, we assume that the occurrence of the

RWGS can change the phase transition behavior. To determine

the influence of RWGS on the phase transition, we assume that

the catalyst operates at a certain selectivity towards RWGS that

is extracted from the measured selectivities of our catalytic

experiments presented below. For the following computations,

Figure 2. ΔGcurves for catalyst dehydration and rehydration without

considering RWGS. a) 200°C, b) 250°C, c) 300°C for a CO2/H2mixture of 1/3

at a total pressure of 75 bar. Black line: equilibrium conversion, blue curve:

ΔrGfor dehydration reaction, orange curve: ΔrGfor rehydration through

methanol formation, green curve: overall ΔrGfor methanol formation.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 127/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

we thus employ averaged RWGS selectivities at 104% (CO

conversion was below the detection limit), 10%, and 25% for

200°C, 250°C, and 300°C, respectively. Figure 3 shows the

resulting ΔG curves for these temperatures. Now there are two

additional curves in each diagram, indicating the driving force

for rehydration of the catalyst through CO formation (Eq. 5, red

curves), and the total ΔGRWGS (purple curve), which is typically

more negative than ΔGMeOH due to the lower selectivity of the

indium-based catalyst towards RWGS (i.e., the actually pro-

duced amount of CO is less than expected from pure

thermodynamics). The low CO content in the gas feed generally

results in a high thermodynamic driving force toward CO

formation.

With the RWGS, there is an additional reaction path for the

catalyst to rehydrate (Eq. 5), while the dehydration reaction is

the same as for methanol formation. Similar to the discussion

on Figure 2, we employ the driving forces of these reactions,

particularly the intersections of the blue, red, and orange

curves, with the abscissa, to assess whether the phase transition

from In(OH)3to In2O3is reversible or irreversible, and now

extend this approach by considering additional contributing

reactions.

Figure 3a shows the ΔG curves for 200°C, for which the

dehydration and methanol curves look exactly the same as in

Figure 2 a because the RWGS selectivity is close to zero. We

observe two additional curves drawn in red (rehydration via

RWGS) and purple (overall ΔGRWGS), which appear in the

negative range, indicating a high thermodynamic driving force

for RWGS. In terms of phase transition, including the RWGS in

our considerations does not alter the conclusions drawn from

Figure 2 a, because the phase transition is still reversible at

conversions lower than 2.3% due to dehydration, and both

rehydration reactions having negative Gibbs free energies.

Similarly, above a conversion of 2.3%, In(OH)3is stable because

the dehydration is now endergonic, indicating that In(OH)3is

unable to release water at high conversions.

At 250°C, (Figure 3b) we observe that dehydration is

exergonic at X <12.2%, and this curve is also unchanged

compared to the case without considering the RWGS, as both

reactions produce one molecule of H2O per consumed molecule

of CO2, i.e., the partial pressure of H2O as a function of CO2

conversion is independent of selectivity. Rehydration through

methanol formation (orange curve) and RWGS (red curve) are

both exergonic over the whole conversion range. Just like the

case without RWGS, In(OH)3is thermodynamically stable at

conversions over 12.2%, while reversible conversion to In2O3is

possible at lower conversions.

Figure 3c displays the ΔG curves for 300°C, where we

observe that dehydration is exergonic over the whole range of

conversion due to the high temperature. The formation of In2O3

is reversible through the methanol formation and RWGS

reaction up to conversions of 21% and 31%, respectively. Due

to the increased RWGS selectivity, we now observe that the

orange curve indicating rehydration through methanol forma-

tion has shifted a little to higher CO2conversion, enlarging the

reversible (white) region. Furthermore, without considering

RWGS in the modeling, it was concluded that In2O3is fully

stable at higher conversions (Figure 2c). If RWGS is considered

as well, it is revealed that rehydration via the RWGS is still

possible at conversions over 21% (blue hatched region), i.e.,

In2O3is favored at higher conversion, but not fully thermody-

namically stable.

From the computational results, we derive two central

hypotheses about the phase stability: (I) The In(OH)3content

post reaction will decrease with temperature, and (II) the

In(OH)3content in the catalyst bed at steady state will vary with

conversion; more specifically: at low temperatures, more In(OH)3

will be present at low conversion, while at high temperatures,

more In(OH)3will be present at high conversion. To verify these

hypotheses and demonstrate this unexpected behavior, we

conducted a series of dedicated stability experiments.

Figure 3. ΔG curves with RWGS based on experimental selectivities. a)

200°C, b) 250°C, c) 300°C for a CO2/H2mixture of 1/3 at a total pressure of

75 bar. Black line: equilibrium conversion, blue curve: ΔrG for dehydration

reaction, orange curve: ΔrG for rehydration through methanol formation,

green curve: overall ΔrG for methanol formation, red curve: ΔrG for

rehydration through RWGS. purple curve: overall ΔrG for RWGS.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 128/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

Experimental investigations on catalyst stability in a fixed-bed

reactor

For experimental evaluation of the thermodynamic predictions

from the dummy catalytic cycle, different model catalysts as

well as reactor configurations have been tested. All tests were

performed under industrially relevant conditions, using a total

pressure of p=75 bar and a stoichiometric feed ratio of CO2/

H2=1/3 in a fixed-bed reactor (for details see Figure S1 in the

Supporting Information). The reaction studies were conducted

at 200°C, 250°C, 275°C and 300°C, respectively.

Catalyst testing with experimental setup 1 (top, middle, and

bottom configuration)

In the first experimental setup, commercial In2O3or In(OH)3was

used as a catalyst material, representing the two sides of the

dummy catalytic cycle, i.e. the two possible starting points of

the rehydration/dehydration mechanism. The three different

reaction temperatures, 200°C, 250°C, and 300°C were adjusted

one after another according to a defined heating ramp, without

cooling down the reactor or changing the catalyst material in

between. In2O3and In(OH)3were positioned either in top,

middle or bottom position in relation to the thermocouple

inside the reactor (Figure 4). The different positions modify the

theoretical temperature profiles (Figure S2) in the reactor to

evaluate the effects of temperature on the stability of In2O3and

to compare the results with the predicted thermodynamic

models. The ensuing phase transition between the indium

species and crystal structure were determined via XRD and

thermogravimetric analysis (TGA).

Moreover, real fixed-bed reactors have a temperature profile

in the catalyst bed. At the entrance of the catalyst bed, the

reaction rate is the highest, decreasing with increasing bulk

height, leading to an integral temperature profile, due to

reaction heat.

The integral temperature profile will increase over the

height of the catalyst bed. The position of the thermocouple is

either at the beginning and coldest spot of the catalyst bed

(top position), in the middle (middle position) or at the end

(bottom position) and therefore the hottest zone of the reactor

(Figure S2).

For the first experimental setup, experimental data for In2O3

as a model catalyst are summarized in Table 1. It was observed

that CO2conversion and methanol selectivity depend on the

segment position for all investigated temperatures (middle

¼bottom >top). The middle and bottom position showed the

highest CO2conversion and methanol selectivity for each

temperature (Table 1) and a more homogenous temperature

distribution (Figure S2). Table 2 shows the temperature depend-

ence of CO2conversion and MeOH selectivity for In2O3and

In(OH)3in the middle configuration of the experimental setup 1

(Figure 4).

Generally, conversion increased with a higher temperature

while the selectivity of methanol decreased due to the

competing endothermic RWGS reaction. Moreover, also the final

partial pressure of water increased with increasing CO2con-

version as both possible reactions contribute equally to water

formation (Table S1).

For 200°C, no CO formation by RWGS was observed and

CO2conversions were similar for In2O3(0.13%) and for In(OH)3

(0.15%). Here, only methanol is detectable, resulting in a

selectivity S(MeOH) of 100%.

Figure 4. Schematic reactor configuration for the first experimental setup

(top, middle or bottom position).

Table 1. Experimental data for In2O3and standard deviation (see Eq. S4) determined in the first experimental setup, with T=200, 250 and 300°C hold for

3 h each, for a CO2/H2mixture of 1/3 (1200 Nmlmin1) and a total pressure of 75 bar.

X (CO2)/% S (MeOH)/%

T/°C top middle bottom top middle bottom

In2O3200 0.08�0.00 0.13�0.00 0.14�0.02 100�0 100�0 100�0

250 0.66�0.00 0.98�0.02 0.94�0.01 88�0 87�0 87�0

300 2.40�0.03 3.56�0.04 3.26�0.05 76�0 73�1 68�1

Table 2. Comparison of the experimental data of In2O3and In(OH)3with standard deviation (see Eq. S4) determined in the middle configuration, T=200,

250 and 300°C, for a CO2/H2mixture of 1/3 at a total pressure of 75 bar.

In2O3In(OH)3

T/°C XCO2 (%) SMeOH (%) XCO2 (%) SMeOH (%)

200 0.13�0.00 100�0 0.15�0.02 100�0

250 0.98�0.02 87�0 1.16�0.00 86�0

300 3.56�0.04 73�1 4.25�0.13 71�1

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 129/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

When temperature was increased to 250°C, the RWGS

selectivity rises to 13% for In2O3and to 14% for In(OH)3. An

average selectivity of 10% was determined for all positions

/starting materials (see Table S2) and used for the simulations in

Figure 3b). For T=250°C, In2O3is mostly stable because the

dehydration step is exergonic over the whole conversion range.

The dehydration behavior is very sensitive to the temperature,

both in terms of kinetics and thermodynamics. Therefore, the

temperature distribution in the catalyst bed may be crucial,

with regard to transformationto In(OH)3at cold spots. For In2O3,

CO2conversion increases to 0.98% and for In(OH)3to 1.16%,

when going from 200 to 250°C. Through the overall higher CO2

conversion due to the higher temperature, an increased heat

formation and water content in the system was registered.

For 300°C, the RWGS selectivity reached 27% (In2O3) and

29% (In(OH)3), respectively. The average selectivity of 25% for

all positions and starting materials was used for the modeling

studies in Figure 3c). The CO2conversions for both starting

materials are somewhat different with 3.56% (In2O3) and 4.25%

(In(OH)3). This might be due to the different accessibility of

oxygen vacancies in both materials. The water partial pressure

significantly increased due to the overall higher activity of the

reaction system (Table S1).

The catalytic results suggest that In(OH)3has been con-

verted completely to In2O3in the middle and bottom position

at 300°C. In comparison, In2O3is not rehydrated to In(OH)3. In

top position the catalytic performance was lower due to the

overall lower catalyst bed temperature and the resulting higher

In(OH)3content. This can be explained with the observations

carried out by XRD (Figure 5) and TGA-measurements (Figure 6)

taken before and after reaction. XRD was used to identify the

crystal structures of In2O3and In(OH)3. Fresh and pure In2O3in

Figure 5a) (black line) was indexed to the cubic structure (blue

square) whereas pure In(OH)3in b) was indexed to the cubic

structure (red oval) as well. After reaction, the XRD patterns

with In2O3as catalyst material displayed no reflections for

In(OH)3. No reflection shift and no phase transition in any

configuration were noticed. This was also verified by TGA,

where no mass loss was measured with In2O3as starting

material (Figure 6a). For In(OH)3as starting material (black line)

Figure 5. XRD for first experimental setup. In2O3a) and In(OH)3b) (before reaction, starting materials) for reactor configuration top, middle, bottom (after

reaction) at 300°C (after temperature ramp). Diffraction patterns for In2O3( ), In(OH)3( ) and InOOH ( ).

Figure 6. TGA data for the first experimental setup at 300°C. In2O3a) and In(OH)3b) (before reaction, black) and after reaction in top (green), middle (blue)

and bottom (red) configuration.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 130/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

the XRD showed characteristic reflections in Figure 5b) indicat-

ing the existence of a cubic In2O3phase (blue square) except for

In(OH)3(red oval), which is not fully converted into In2O3in top

position. No shift in the XRD reflections and no structure

transition were observed, therefore only water was emitted.

This was confirmed by TGA (Figure 6b), where a mass loss of

2.42% of In(OH)3in top position was observed (see Table S3),

suggesting a near-total conversion to In2O3at 300°C.

Regarding the integral temperature profile of the catalyst

bed and the position of the thermocouple, there are new

phenomena to consider. Therefore, we revealed the validity of

the thermodynamic model about the stability of In2O3.

Catalyst testing with experimental setup 2 (two-segment-

configuration)

Computational results in the section Theoretical modeling

predict strong dependence of the In(OH)3dehydration behavior

on the conversion due to the formed H2O. At lower temper-

atures, we expect more In(OH)3to be present at high

conversions, while at higher temperatures, more In(OH)3is

expected at lower conversion. These hypotheses from the

dummy catalytic cycle were studied in a dedicated experimen-

tal setup (Figure 7). In this second experimental setup (two-

segment configuration), the catalyst bed was split into two

identical segments, physically separated by a layer of glass wool

to enable independent characterization of model catalysts from

the low-conversion and high-conversion zone post reaction. As

catalyst material, just In(OH)3was used and the catalytic tests

were performed under the same conditions as previously,

adding an additional temperature step at 275°C. For this

experiment, the sample was extracted from the reactor after

each temperature step and fresh In(OH)3was used for the next

experiment with a different temperature. In this reactor

configuration, In(OH)3was applied to examine the phase

transition behaviour for different conversion ranges (and there-

by, water contents) and temperatures. Depending on the CO2

conversion in the first (top) segment, the second (bottom)

segment operates with effluent gas in comparison to the first

one, i.e., it is exposed to the reaction products, such as H2O,

methanol, and CO. Thereby, the influence of conversion on the

rehydration/dehydration behaviour of the In2O3/In(OH)3system

could be examined for different temperatures. The separation

of the catalyst bed prevents an overheating due to CO2

hydrogenation and a lower integral temperature profile (see

Figure S3). The phase transition that occurred and the crystal

structure of the catalyst were determined via XRD and TGA for

both layers separately.

For the second experimental setup, all experimental data of

CO2hydrogenation and the calculated partial pressure of water

are summarized in Table 3.

It was observed that the CO2conversion and methanol

selectivity both depend on the temperature, as expected: The

conversion of CO2increased with increasing temperature from

0.04% (200°C) up to 3% (300°C), while the selectivity of

methanol decreased due to competing RWGS with higher

temperature from 100% at 200°C down to 76% at 300°C.

For 200°C in the second experimental setup, the CO2

conversion (X (CO2)=0.04%) as well as methanol-selectivity (S

(MeOH)=100%), are consistent with the results obtained for

the top position of the first experimental setup. XRD measure-

ments (Figure 8) show only a slight phase transition to In2O3(at

about 30°and 35°), with the low-conversion section (top)

containing more In2O3than the high-conversion section

(bottom). TGA shows an In(OH)3amount of 82.11% of the low-

conversion sample and 89.91% In(OH)3in the high-conversion

one (Table 4) after reaction, consistent with the XRD results. To

explain this observation, one needs to consider that the upper

section produces water through methanol synthesis, which

influences the phase transition equilibrium during reaction in

the lower segment. This reduces the thermodynamic driving

force for dehydration as shown in Figure 3a) (blue) curve with

increasing conversion. Dehydration is still exergonic at the

experimentally determined 0.04% conversion, but rehydration

is also possible and highly exergonic. We note that at 0.04%

conversion, the influence of a temperature gradient due to the

heat of reaction is negligible or the phase transition is too

sluggish.

For 250°C the CO2conversion reached 0.52%, which is in

between the values reached in top and middle configuration of

Figure 7. Schematic reactor configuration for the second experimental setup

(two-segment-configuration).

Table 3. Experimental data and standard deviation (see Eq. S4) of setup 2 for temperatures of 200, 250, 275 and 300°C for a CO2/H2mixture of 1/3 at a total

pressure of 75 bar. *simulated with ASPEN PLUS (for details see Table S1 in the Supporting Information).

T/°CXCO2/% SMeOH/% p(H2O)/bar*

200 0.04�0.00 100�0 0.007

250 0.52�0.00 100�0 0.093

275 0.84�0.06 82�1 0.150

300 3.0�0.02 76�1 0.535

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 131/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

the first experimental setup (Table 3). The measured RWGS

conversion was negligible, i.e., water was only formed by

methanol synthesis (p(H2O)=0.093 bar). The XRD patterns in

Figure 8 show more distinct peaks assigned to In2O3in

comparison to 200°C. The peak intensity ratio between In2O3

and In(OH)3is smaller in the high-conversion segment, consis-

tent with the TGA results (Table 4, Figure 9). Herein, the bottom

segment still shows a higher In(OH)3amount (69.41%)

compared to the top segment (55.77%), resulting from water

release by phase transition of In(OH)3to In2O3. This is due to the

formation of water in the segment above, which inhibits the

release of water by phase transformation of In(OH)3. This is in

good agreement with the prediction for this temperature and

conversion from theoretical modeling.

For 275°C the CO2conversion reached 0.84% and the

MeOH selectivity decreased to 82% due to RWGS. Water

formation increased (p(H2O)=0.15 bar) because of methanol

synthesis and RWGS. In both segments the XRD patterns

indicate a mixture of In(OH)3and In2O3. The peak intensity ratio

between In2O3and In(OH)3is higher in the bottom segment

(Figure 8), i.e., at high conversion. This is consistent with the

TGA results, where the top and bottom segments show an

In(OH)3amount of 62.06% and 22.65%, respectively. For 275°C,

we detect substantially more In(OH)3in the low-conversion

segment, indicating that the trend has reversed compared to

200°C and 250°C (see Figure 9, green curve). This surprising

finding can be explained on the basis of our computational

model. The stability of In2O3generally increases with increasing

temperature due to the dehydration reaction, which is now

exergonic over the whole accessible conversion range. Dehy-

dration still shows the same dependence on the conversion as

it did at lower temperature, i.e. the driving force decreases with

increasing conversion. However, the phase transition from

In(OH)3to In2O3is controlled not only by the dehydration, but

also by the reverse reaction, i.e., hydration due to the formation

of methanol and RWGS. The curves for the MeOH formation

and RWGS indicate exergonic reaction at low conversion and

endergonic reaction at high conversion, showing that the

dehydration of In(OH)3becomes irreversible at high conversion

beyond a certain temperature threshold. We note that the

experimentally measured conversion levels are below the

intersection of the MeOH/RWGS curve at 275°C. This is due to

the significant uncertainty regarding the thermodynamic data

of In(OH3) and In2O3, which gives rise to inaccurate temperature

dependence of Gibbs free energies of reaction. However, the

principal trends, i.e., that In(OH)3is stable at low temperature

and high conversion, while In2O3is stable at high temperature

and high conversion, and reversible phase transition is possible

at both high and low temperature at low conversion, is

unaffected by these inaccuracies.

For 300°C, the CO2conversion further increased to 3%

(Table 3), which is in between the values of the top and middle

configuration of the first experimental setup for In(OH)3

(Table S1). The separation in two segments inhibited the RWGS

reaction due to less heating of the bed and shorter residence

time in each bed. The observed RWGS selectivity reached 24%,

which matched with the studies in Figure 3 c). From the XRD

data, we infer that In(OH)3undergoes a complete phase

transition to In2O3in the high-conversion section (bottom),

while some In(OH)3remains in the low-conversion section (top).

The TGA data also showed no mass loss, indicating no

In(OH)3present in the bottom section, and only a small amount

of In(OH)3(1.91%) in the top section, and therefore confirmed

the trend for the stability of In2O3(Table 4, Figure 9). From the

Figure 8. XRD data for the second experimental setup. In(OH)3(before

reaction) and for 200°C (red), 250°C (blue). 275°C (green), 300°C (purple)

(after reaction) and In2O3(reference), with diffraction patterns for In2O3( ),

In(OH)3( ) and InOOH ( ).

Table 4. In(OH)3amount before and after reaction for second experimental

setup calculated with the TGA results (Table S3, Supporting Information).

low-conversion zone,

top/%

high-conversion zone,

bottom/%

Pure 95.68 95.68

200°C 82.11 89.91

250°C 55.77 69.41

275°C 62.06 22.65

300°C 1.91 0.00

Figure 9. TGA data for second experimental setup. In(OH)3(before reaction,

black) and for 200°C (red), 250°C (blue), 275°C (green), 300°C (purple) (after

reaction, top).

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 132/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

computational modeling, we expect that the conversion to

In2O3is irreversible at high conversions, while reversible

transition is expected at low conversion, indicating a good

agreement between theory and experiment.

Phase transition during hydrogen drop out

Finally, the model was employed to investigate how the catalyst

will behave when the hydrogen content in the feed drops due

to the instability of intermittent hydrogen sources. For a H2-

deficient p(H2)/p(CO2) ratio of 1/1, the free energy curves for

different temperatures are shown in Figure 10.

The first noticeable consequence of the reduced hydrogen

content in the feed is a decrease of the equilibrium conversion

to 11.7%, 8.5%, and 6.7% at 200°C, 250°C, and 300°C,

respectively. This will result in a reduced reaction rate and may

lead to a temperature drop in the reactor. At 200°C (Fig-

ure 10a), the reversible zone (white) still goes up to 2.3%

conversion, indicating irreversible In(OH)3formation at higher

conversion. However, due to the reduced equilibrium conver-

sion and reduced reaction rate under H2deficit, the white range

now spans over a larger fraction of the catalyst bed, which

would enhance the stability of In2O3compared to a stoichio-

metric reactant ratio. At 250°C (Figure 10b), irreversible In(OH)3

formation does not occur because dehydration is exergonic

below equilibrium conversion. Rehydration via methanol for-

mation and RWGS is exergonic up to 7.3% conversion, leading

to reversible In(OH)3formation at low conversion. Above 7.3%,

only the RWGS enables rehydration, indicating that In2O3will be

favored at higher conversion for H2deficit. The reversible

formation of In(OH)3through RWGS could be suppressed by co-

feeding CO during hydrogen dropout. This would move the red

curve representing In(OH)3formation via RWGS up in Gibbs free

energy, moving the region where In2O3is fully stable to lower

conversion. At 300°C (Figure 10c), the result looks quite similar,

with the reversible zone stretching from 1.4% to beyond

equilibrium conversion. In2O3is fully stable only for conversions

beyond the equilibrium line, what is irrelevant in practice. In

summary, our model calculations for hydrogen-deficit operation

suggest that stability is generally enhanced compared to a

stoichiometric feed composition, both at high and low temper-

ature.

The results give us practical guidance about reactor

operation during hydrogen dropout. First of all, a drop of

temperature below 250°C should be prevented, as lower

temperature generally favors In(OH)3formation. Due to the

reduced heat of reaction at decreased reaction rates, stronger

heating of the reactor may be required to retain the desired

temperature and suppress In(OH)3formation. At elevated

temperature, In(OH)3can be formed reversibly through RWGS,

and this mode dominates over the largest fraction of the

catalyst bed, possibly contributing to lower activity in the

middle/high conversion zone of the reactor. If such degradation

is observed in practice, co-feeding small amounts of CO (1%)

could be considered to protect the catalyst from In(OH)3

formation. Naturally, reduction to In0must also be taken into

account in these considerations, as this is another possible

mode of degradation. Increased In0formation is, however, not

expected, because the reactant feed at p(CO2)/p(H2)=1 with

e.g. 1% CO is still less reducing than the regular stoichiometric

feed at p(CO2)/p(H2)=3 (see Figure S4).

Conclusions

Employing a combination of thermodynamic considerations

and dedicated stability experiments, we have shown that the

phase transformation of In2O3-based catalysts by conversion

into In(OH)3can be predicted by considering the reversibility of

Figure 10. ΔG curves for variable p(H2)-deficient operation at p(H2)/p(CO2)

=1/1. a) 200°C, b) 250°C, c) 300°C at a total pressure of 37.5 bar. Black line:

equilibrium conversion, blue curve: ΔrG for dehydration reaction, orange

curve: ΔrG for rehydration through methanol formation, green curve: overall

ΔrG for methanol formation, red curve: ΔrG for rehydration through RWGS.

purple curve: overall ΔrG for RWGS.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 133/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

dehydration and rehydration reactions. The thermodynamics of

these reactions depend sensitively on the reaction conditions,

most notably, on the temperature and conversion, due to the

formation of H2O, which suppresses the dehydration reaction at

lower temperatures. The selectivities toward RWGS and meth-

anol formation have only a minor influence on the phase

transition because both reactions form equal amounts of H2O

per turnover. However, the relationship between In(OH)3

formation and conversion is not straightforward; at lower

temperature, In(OH)3formation is favored at higher conversion,

while at higher temperature, lower conversion favors In(OH)3

formation. This peculiar phase transition behavior was con-

firmed by dedicated model experiments, where the catalyst bed

was split into a high-conversion and low-conversion section,

and the two sections were characterized separately post-

reaction. This approach allows to connect structural changes

with the local gas atmosphere the catalyst is exposed to. These

experiments confirm the aforementioned trend. Hereby, the

inversion occurs between 250°C and 275°C, while the computa-

tional model predicts it at 285°C. This small deviation can be

explained by the uncertainty associated with the experiment-

based thermodynamic data underlying our calculation. The

basic trends, however, are not sensitive to these uncertainties.

In addition to a change of the In(OH)3fraction along the catalyst

bed, the catalyst is subject to reversible phase transitions at low

conversion, which may contribute to fast catalyst particle

sintering or reshaping.

We finally employed our model to examine how a change

of p(CO2)/p(H2) ratio possibly influences catalyst degradation.

Our results indicate that In(OH)3formation is not generally

promoted by hydrogen dropout. Quite in contrast, at low

temperature In2O3is expected to be more stable due to lower

water content in the gas feed. At higher temperatures, only

RWGS is thermodynamically able to contribute to catalyst

hydration, and even this could be suppressed by co-feeding CO

during hydrogen drop out. The main risk for catalyst stability

during hydrogen drop out stems from a temperature reduction

due to reduced heat of reaction. This is important for the

reactor design of methanol synthesis, because large scale fixed

beds may have cold spots leading to locally increased In(OH)3

formation.

Deactivation of heterogeneous catalysts in technical proc-

esses depends sensitively on the reaction conditions and local

conversion levels. The methodology described here provides an

intuitive access to understand the influence of reaction

conditions on catalyst deactivation involving phase transitions

or stoichiometry changes. This approach offers practical

guidance on optimizing reactor operation to prevent catalyst

degradation.

Experimental Section

Theoretical modeling

The Gibbs free energy change DrGfor each reaction is computed

explicitly as a function of the reaction conditions:

DrG T;p0;CO2;H2;:::ð Þ;X;SMeOH

 �

¼X

ni�DfGiT;p0;CO2;H2;:::ð Þ;X;SMeOH

 � (7)

Here, X and SMeOH represent the CO2conversion and selectivity

toward CH3OH, respectively. Trepresents the temperature and p0;CO2

and p0;H2represent the initial partial pressures of CO2and H2. We

assume that no product is present initially, with the exception of

H2O, which is present as an impurity with a concentration of 5 ppm,

which is a typical impurity level for laboratory gas. Finally, niand

DfGistand for the stoichiometric coefficient and Gibbs free energy

of formation of species i.

The Gibbs free energy of formation is computed via:

DfGiT;p0;ðCO2;H2;:::Þ;X;SMeOH

 �

¼DfH0

iT�DfS0

iþR�T�log ai

ð Þ:(8)

The standard enthalpies and entropies of formation, DfH0

iand DfS0

are taken from experimental data as given in Table 5. airepresents

the activity of species i, which is approximated as ai¼pi=p0for the

gas phase species. The activities of pure solid phases are defined as

1. Rand p0are the general gas constant and standard pressure

(105Pa). Inserting Eq. 7 into Eq. 8 results in Eq. 9:

DrGiT;p0;ðCO2;H2;:::Þ;X;SMeOH

 �

¼DrG�Tð ÞþR�TX

niln ai

ð Þ:(9)

The thermodynamic data for the reactant molecules in the gas

phase, In2O3and In(OH)3are taken from the NIST database,[22] the

Springer Materials data collection,[23] and primary literature,[24]

respectively (Table 5). We note that there is some uncertainty

regarding the thermodynamic data of In(OH)3because it is not a

well-studied material, and its thermodynamic data has to our

knowledge not been assessed in the form of curated data

collection. Because experimental measurements of the standard

entropy of In(OH)3are not available in the literature to the best of

our knowledge, we estimated it from the onset of the decom-

position temperature of In(OH)3obtained from our TGA experi-

ments on fresh In(OH)3(220°C). Errors in the enthalpies of

formation and standard entropies will mostly affect the temper-

ature where transition from In(OH)3to In2O3takes place, but not

the trends, i.e., how a change of temperature, conversion or CO/H2

ratio will affect the stability.

Table 5. Thermodynamic data employed in the model.

DfH0/(kJmol-1)S0/(JK1mol1) Reference

CO2393.5 213.8 [22]

H20 130.7 [22]

CH3OH(g) 201.3 239.9 [22]

H2O(g) 241.8 188.8 [22]

CO 110.5 197.7 [22]

In(OH)3927 127.1a[23]

In2O3923 101.8 [24]

aEstimate from the onset of In(OH)3decomposition in our TGA experiment

(220°C).

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 134/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

Catalytic experiments

Catalytic CO2hydrogenation was conducted in a high pressure

continuous flow fixed-bed reactor (Figure S1). The reactor with an

inner diameter of 20 mm was made of stainless steel (1.4571). A gas

mixture of 25% CO2(4.5 grade) and 75% H2(5.0 grade) from

Westfalen, as well as N2(5.0 grade) from Air Liquide were used. The

reactant gas mixture with a CO2/H2stoichiometric ratio of 1/3 was

introduced by mass-flow controllers (Bronkhorst Prestige FG-201 CV)

and flows downwards through the reactor tube. The reaction

pressure was adjusted using a back-pressure regulator (Dutch

Regulators). The temperature was controlled with a surrounding

heating mantle and a thermocouple. Gas lines at the inlet and

outlet of the reactor were heated to 180°C in order to preheat the

reaction gas and to prevent condensation of methanol and water in

the outlet. Gas-flows and temperatures were automatically con-

trolled by a LabVIEW (National Instruments TM) interface. The outlet

gas composition was sampled every 30 min and analyzed using an

online gas chromatograph (Bruker 450-GC), set up with four

columns (Restek Q-Bond, Restek U-Bond, Bruker Swax, Bruker

Molsieve 5 Å), a methanizer (for CO2and CO quantification), two

flame ionization detectors (FIDs) and one thermal conductivity

detector (TCD).

For the catalyst test experiments, Indium(III) oxide (99.9% metals

basis) from Alfa Aesar and Indium(III) hydroxide (99.8% metals

basis) from Thermo Scientific were used. For the first experimental

setup, 5.0 g of In2O3or In(OH)3was mixed homogeneously with

quartz spheres and loaded into the reactor. The segment was

positioned in top, middle or bottom position in relation to the

thermocouple and held in place by a bed of quartz wool (see

scheme in Figure 4). Prior to reaction, the dummy catalyst was

pretreated at 200°C under flowing N2(300 Nml min1) for 1 hour.

Afterwards, the reactor was heated to the desired reaction temper-

ature (200°C, 250°C, 300°C) one after another in a ramp and each

temperature was hold for 3 hours. For the second experimental

setup, 5.0 g of In(OH)3was mixed homogeneously with quartz

spheres and split in two identical segments (see scheme in

Figure 7). The segments were fixed and held in place by a bed of

quartz wool. For each experimental setup, prior to reaction, the

dummy catalyst was pretreated to 200°C under flowing N2(300

Nml min1) for 1 hour. Then the reactor was heated to the reaction

temperature (200°C, 250°C, 300°C).

For every experimental setup, a CO2/H2flow of 1200 Nml min1was

passed through the reactor to start CO2-hydrogenation. The

product gas analysis was performed by GC every 30 min under

steady state reaction conditions. After the reaction, the reactor was

cooled down (3 K min1) under a continuous flow of nitrogen (1000

Nml min1) and the catalyst was removed and stored under Argon

(4.6 grade, Heide Gas). The detailed calculation of _

nand the partial

pressure of water (pH2O) are shown in the Supporting Information.

CO2conversion X(CO2), yield Y(MeOH) and selectivity S(MeOH) of

methanol were calculated applying Eqs. 10–12.

XCO2¼_

nCO2;in _

nCO2;out

nCO2;in �100 % (10)

YMeOH ¼_

nMeOH

nCO2;in �100 % (11)

SMeOH ¼YMeOH

XCO2

(12)

Catalyst characterization

Thermogravimetric analysis was carried out using a SETSYS

Evolution TGA-DTA from Setaram Instrumentation. The samples

(~30 mg) were heated up from ambient temperature to 110°C

under nitrogen and the temperature was hold for one hour in order

to remove all water residues. Afterwards, the samples were heated

up to 300°C with 2 K/min and hold for three hours to determine

the mass loss, which was attributed to the formation of water by

conversion of In(OH)3to In2O3. X-Ray diffraction was carried out

using a Panalytical MPD X’Pert Pro, with a Cu-Kα-source. The

measuring range was 10–80°, with a step size of 0.013°and a

counting time of 73 seconds.

Acknowledgements

We thank the Central Analytics Department of the University of

Hamburg (UHH) for carrying out the XRD measurements. Open

Access funding enabled and organized by Projekt DEAL.

Conflict of Interests

The authors declare no conflict of interest.

Data Availability Statement

Research data are not shared.

Keywords: catalyst degradation ·CO2hydrogenation ·dummy

catalytic cycle ·indium ·methanol synthesis

[1] a) N. Armaroli, V. Balzani, Angew. Chem. Int. Ed. 2007,46, 52–66; b) D.

Gielen, F. Boshell, D. Saygin, Nat. Mater. 2016,15, 117–120.

[2] a) Z. Zhang, Renewable and Sustainable Energy Reviews 125 2020,

109799; b) K. M. K. Yu, I. Curcic, J. Gabriel, S. C. E. Tsang, ChemSusChem

2008,1, 893–899; c) M. Aresta, A. Dibenedetto, A. Angelini, Chem. Rev.

2014,114, 1709–1742.

[3] a) P. Gao, Nat. Chem. 2017,9, 1019–1024; b) A. Gulzar, A. Gulzar, M. B.

Ansari, F. He, S. Gai, P. Yang, Chemical Engineering Journal Advances

2020,3, 100013.

[4] a) D. S. Marlin, E. Sarron, Ó. Sigurbjörnsson, Front. Chem. 2018,6, 446;

b) V. Dieterich, A. Buttler, A. Hanel, H. Spliethoff, S. Fendt, Energy

Environ. Sci. 2020,13, 3207–3252; c) P. Gao, L. Zhang, S. Li, Z. Zhou, Y.

Sun, ACS Cent. Sci. 2020,6, 1657–1670.

[5] a) V. D. Dasireddy, B. Likozar, Renewable Energy 2019,140, 452–460;

b) R.-P. Ye, Nat. Commun. 2019,10, 5698.

[6] a) M. Bowker, ChemCatChem 2019,11, 4238–4246; b) J. Zhong, X. Yang,

Z. Wu, B. Liang, Y. Huang, T. Zhang, Chem. Soc. Rev. 2020,49, 1385–

1413.

[7] a) W.-J. Shen, K.-W. Jun, H.-S. Choi, K.-W. Lee, Korean J. Chem. Eng. 2000,

17, 210–216; b) S. Kanuri, S. Roy, C. Chakraborty, S. P. Datta, S. A. Singh,

S. Dinda, Intl J of Energy Research 2022,46, 5503–5522; c) K. Stangeland,

H. Li, Z. Yu, Ind. Eng. Chem. Res. 2018,57, 4081–4094.

[8] S. Ghosh, J. Sebastian, L. Olsson, D. Creaser, Chem. Eng. J. 2021,416,

129120.

[9] a) O. Martin, A. J. Martín, C. Mondelli, S. Mitchell, T. F. Segawa, R. Hauert,

C. Drouilly, D. Curulla-Ferré, J. Pérez-Ramírez, Angew. Chem. Int. Ed.

2016,55, 6261–6265; b) P. Schühle, S. Reichenberger, G. Marzun, J.

Albert, Chem. Ing. Tech. 2021,93, 585–593; c) K. Lau, P. Schühle, S.-X.

Liang, F. de Kock, J. Albert, S. Reichenberger, ACS Appl. Energ. Mater.

2021,4, 9206–9215.

[10] P. Schühle, M. Schmidt, L. Schill, A. Riisager, P. Wasserscheid, J. Albert,

Catal. Sci. Technol. 2020,10, 7309–7322.

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 135/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425

[11] Z. Lu, J. Wang, K. Sun, S. Xiong, Z. Zhang, C. Liu, Green Chemical

Engineering 2022,3, 165–170.

[12] M. S. Frei, M. Capdevila-Cortada, R. García-Muelas, C. Mondelli, N. López,

J. A. Stewart, D. Curulla Ferré, J. Pérez-Ramírez, J. Catal. 2018,361, 313–

321.

[13] a) M. Dou, M. Zhang, Y. Chen, Y. Yu, Catal. Lett. 2018,148, 3723–3731;

b) A. Cao, Z. Wang, H. Li, J. K. Nørskov, Catalysis Letters2021,11, 1780–

1786.

[14] J. Ye, C. Liu, D. Mei, Q. Ge, ACS Catal. 2013,3, 1296–1306.

[15] S. Dang, B. Qin, Y. Yang, H. Wang, J. Cai, Y. Han, S. Li, P. Gao, Y. Sun, Sci.

Adv. 2020,6, eaaz2060.

[16] T. Bielz, H. Lorenz, W. Jochum, R. Kaindl, F. Klauser, B. Klötzer, S. Penner,

J. Phys. Chem. C 2010,114, 9022–9029.

[17] A. Tsoukalou, P. M. Abdala, D. Stoian, X. Huang, M.-G. Willinger, A.

Fedorov, C. R. Müller, J. Am. Chem. Soc. 2019,141, 13497–13505.

[18] F. Hess, ACS Catal. 2022,12, 1, 497–511.

[19] C. Li, F. Hess, I. Djerdj, G. Chai, Y. Sun, Y. Guo, B. M. Smarsly, H. Over, J.

Catal. 2018,357, 257–262.

[20] C. Li, Y. Sun, F. Hess, I. Djerdj, J. Sann, P. Voepel, P. Cop, Y. Guo, B. M.

Smarsly, H. Over, Appl. Catal. B 2018,239, 628–635.

[21] Y. Sun, F. Hess, I. Djerdj, Z. Wang, T. Weber, Y. Guo, B. M. Smarsly, H.

Over, ChemCatChem 2020,12 (21), 5511–5522.

[22] P. J. Linstrom, W. G. Mallard, Eds., NIST Chemistry WebBook, NIST

Standard Reference Database Number 69, National Institute of Stand-

ards and Technology, Gaithersburg MD, 20899, https://doi.org/10.

18434/T4D303.

[23] E. H. P. Cordfunke, R. J. M. Konings, W. Ouweltjes, J. Chem. Thermody-

namics 27 1995 431–433.

[24] Landolt-Börnstein, Springer Materials, New Series IV/19 A4, 2001, DOI:

10.1007/b71435.

Manuscript received: August 3, 2023

Revised manuscript received: August 25, 2023

Accepted manuscript online: August 25, 2023

Version of record online: ■■■,■■■■

Wiley VCH Mittwoch, 13.09.2023

2309 / 319485 [S. 136/136] 1

ChemPlusChem

Research Article

doi.org/10.1002/cplu.202300425