A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol Autocatalytically

Article abstract of DOI:10.1016/j.chempr.2020.04.008

Nature utilizes multicomponent catalyst systems to convert simple, abundant starting materials into complex molecules that are essential for life. In contrast, synthetic chemical transformations rarely adopt this strategy because it is difficult to replicate the sophisticated supramolecular assemblies used by biology for active-site separation and substrate trafficking. Here, we describe a method for multicomponent catalyst separation that involves encapsulating transition-metal complexes in nanoporous materials called metal-organic frameworks. The multicomponent catalyst system was highly active for converting hydrogen and carbon dioxide to methanol, and it could be formulated to be readily recyclable. Moreover, we uncovered an autocatalytic feature that was possible only when we utilized the multicomponent catalyst strategy. These results open avenues for obtaining fuel from abundant and renewable resources. Methanol is a promising renewable fuel that can be adapted to the current liquid fuel infrastructure. It can be produced from hydrogen and carbon dioxide, mitigating greenhouse gas emissions and storing hydrogen in the process. However, the industrial production of methanol through this hydrogenation reaction currently requires elevated temperatures and pressures and can produce significant amounts of unwanted byproducts. Here, we employ a bioinspired tandem catalytic system to efficiently hydrogenate carbon dioxide to methanol selectively at low temperatures. We achieved superior performance by eliminating catalyst incompatibility through encapsulating at least one of the catalysts involved in the tandem process in nanoporous materials called metal-organic frameworks. In the long term, this method could be applied to other tandem catalytic processes, allowing more efficient access to alternative fuels, commodity chemicals, and valuable pharmaceutical products. Tsung and co-workers describe a three-component tandem catalytic process for the hydrogenation of carbon dioxide to methanol. The bioinspired process is enabled by encapsulation of at least one of the two ruthenium-based catalysts required in the metal-organic framework (MOF) UiO-66. The reaction was found to have an autocatalytic feature that enables the reaction to be carried out without superstoichiometric additives. Encapsulating both ruthenium-based catalysts in the MOF allowed the catalyst to be recycled.

Full text of DOI:10.1016/j.chempr.2020.04.008

ll
Article
A Bioinspired Multicomponent Catalytic
System for Converting Carbon Dioxide into
Methanol Autocatalytically
Thomas M. Rayder, Enric H.
Adillon, Jeffery A. Byers,
Chia-Kuang Tsung
byersja@bc.edu (J.A.B.)
frank.tsung@bc.edu (C.-K.T.)
HIGHLIGHTS
Tandem catalysis results in
efficient hydrogenation of CO2 to
methanol
Autocatalysis is enabled through
catalyst sequestration in metal-
organic frameworks
Encapsulating multiple catalyst
species makes catalysts recyclable
and productive
Tsung and co-workers describe a three-component tandem catalytic process for
the hydrogenation of carbon dioxide to methanol. The bioinspired process is
enabled by encapsulation of at least one of the two ruthenium-based catalysts
required in the metal-organic framework (MOF) UiO-66. The reaction was found to
have an autocatalytic feature that enables the reaction to be carried out without
superstoichiometric additives. Encapsulating both ruthenium-based catalysts in
the MOF allowed the catalyst to be recycled.
Rayder et al., Chem 6, 1–13
July 9, 2020 ª 2020 Elsevier Inc.
https://doi.org/10.1016/j.chempr.2020.04.008

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
A Bioinspired Multicomponent Catalytic
System for Converting Carbon Dioxide
into Methanol Autocatalytically
1,2,
Thomas M. Rayder,1 Enric H. Adillon,1 Jeffery A. Byers,1, and Chia-Kuang Tsung
*
*
SUMMARY
The Bigger Picture
Methanol is a promising
Nature utilizes multicomponent catalyst systems to convert simple,
abundant starting materials into complex molecules that are essen-
tial for life. In contrast, synthetic chemical transformations rarely
adopt this strategy because it is difficult to replicate the sophisti-
cated supramolecular assemblies used by biology for active-site
separation and substrate trafficking. Here, we describe a method
for multicomponent catalyst separation that involves encapsulating
transition-metal complexes in nanoporous materials called metal-
organic frameworks. The multicomponent catalyst system was high-
ly active for converting hydrogen and carbon dioxide to methanol,
and it could be formulated to be readily recyclable. Moreover, we
uncovered an autocatalytic feature that was possible only when
we utilized the multicomponent catalyst strategy. These results
open avenues for obtaining fuel from abundant and renewable
resources.
renewable fuel that can be
adapted to the current liquid fuel
infrastructure. It can be produced
from hydrogen and carbon
dioxide, mitigating greenhouse
gas emissions and storing
hydrogen in the process.
However, the industrial
production of methanol through
this hydrogenation reaction
currently requires elevated
temperatures and pressures and
can produce significant amounts
of unwanted byproducts.
Here, we employ a bioinspired
tandem catalytic system to
efficiently hydrogenate carbon
dioxide to methanol selectively at
low temperatures. We achieved
superior performance by
INTRODUCTION
Biological organisms have evolved complex networks of chemical reactions that are
necessary for their survival. For example, the Calvin cycle employs a series of redox
and condensation reactions to convert carbon dioxide into glyceraldehyde-3-phos-
phate (G3P), a precursor to the sugars that organisms use as fuel (Figure 1A).1 The
cycle involves multiple catalytic transformations controlled precisely by supramolec-
ular assemblies that traffic substrates between isolated active sites. This strategy al-
lows biological systems to overcome inherent limitations to selectivity, reactivity,
and compatibility that would otherwise restrict the sequences of chemical reactions
needed to convert carbon dioxide to fuel. As humans face their own rising energy
demands, it is logical to take inspiration from biology in order to produce fuel
from readily available, renewable starting materials. This endeavor has resulted in
the development of many biomimetic synthetic catalytic systems, some of which rival
their natural analogs2–7 but most of which primarily mimic the identity of the catalyt-
ically active sites rather than the supramolecular protein assemblies that host multi-
ple active sites.8 As a result, synthetic systems rarely capitalize on pathways that
would otherwise benefit from multistep processes.9–11
eliminating catalyst
incompatibility through
encapsulating at least one of the
catalysts involved in the tandem
process in nanoporous materials
called metal-organic frameworks.
In the long term, this method
could be applied to other tandem
catalytic processes, allowing more
efficient access to alternative
fuels, commodity chemicals, and
valuable pharmaceutical
products.
Inspired by the supramolecular protein assemblies that nature uses for fuel produc-
tion, we now report a synthetic catalyst system for the hydrogenation of carbon di-
oxide to methanol. The system utilizes multiple catalytically active sites assembled
and isolated from one another in a three-dimensional (3D) crystalline nanoporous
material called a metal-organic framework (MOF) (Figure 1B). MOFs are composed
of inorganic nodes and organic bridging linkers, which make them attractive host
Chem 6, 1–13, July 9, 2020 ª 2020 Elsevier Inc.
1

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Figure 1. Biological Inspiration for the Reported Cascade Hydrogenation of Carbon Dioxide to
Methanol
(A) Top: the light-dependent reactions in photosynthesis use multistep reactions to produce the
energy carriers ATP and NADPH for use in the Calvin cycle (Pi = –OPO3ꢀ). Bottom: the Calvin cycle
utilizes multiple isolated catalytically active sites separated by supramolecular assemblies to store
energy from light as fuel by converting carbon dioxide and water to glucose precursor G3P.
(B) The MOF UiO-66 (pictured as a single cage of an extended 3D network) is used as a catalyst and
as a host material for isolating two transition-metal-based catalysts required for producing
methanol from hydrogen and carbon dioxide. Because the multistep reaction occurs in a single
reaction vessel, it benefits from autocatalysis that results when the methanol product is
incorporated in the second step of the overall transformation.
materials for catalysis because their porous structure can be engineered by a com-
bination of organic chemistry and self-assembly principles.12 The site isolation im-
parted by the strategy for encapsulating catalytically active sites in MOFs allows
multistep, cascade catalysis to be carried out in a single vessel, and it is amenable
to reaction engineering that targets an individual step in a multistep process.
Here, this capability was exploited for catalyst optimization, which revealed an
autocatalytic feature of carbon dioxide hydrogenation that has thus far not been re-
ported. Moreover, the activity of the catalyst rivals the best synthetic systems re-
1Department of Chemistry, Merkert Chemistry
Center, Boston College, Chestnut Hill, MA 02467,
USA
ported to date13,14 at the lowest reported temperature15 and with a catalyst that
is readily recyclable.
2Lead Contact
Coupled with catalytic water splitting,16 the hydrogenation of carbon dioxide to
*Correspondence: byersja@bc.edu (J.A.B.),
frank.tsung@bc.edu (C.-K.T.)
methanol is a promising method for selectively obtaining easily transportable, en-
ergy-dense fuel from readily available, renewable resources.17–20 Although
https://doi.org/10.1016/j.chempr.2020.04.008
2
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
heterogeneous catalysts are industrially employed for this transformation,17 reac-
tions are carried out at high temperatures and pressures, and the ill-defined active
sites inherent to most heterogeneous catalysts adversely affect reaction selec-
tivity.21 Conversely, homogeneous molecular catalysts have been developed for
carbon dioxide hydrogenation.22–26 These catalysts feature well-defined active sites
amenable to logical optimization and beneficial for reaction selectivity. However,
few examples of homogeneous catalysts are known for the hydrogenation of carbon
dioxide to methanol, and none of the reported systems are recyclable.13–15,27–31 All
catalysts for this transformation face a formidable challenge in traversing four
different oxidation states of carbon, each thermodynamically uphill until the ultimate
reduction of formaldehyde to methanol.32
The most efficient molecular catalysts for carbon dioxide hydrogenation reported
to date involve combining a ruthenium-based catalyst with superstoichiometric
amounts of amine additives.27–29 This combination aids methanol formation by
first converting formic acid to a formamide intermediate. Although this approach
significantly improved catalyst turnover compared with that of previous homoge-
neous molecular catalysts, the stability of the formamide intermediate often led
to mixtures of methanol and formamide.28 Subsequent work improved product
separation by using a polyamine29 or sterically encumbered amine additives
that force the reaction to proceed through a thermodynamically less stable
and therefore more reactive formamide intermediate.13 Despite these develop-
ments, all iterations of this strategy require superstoichiometric amine additives.
There is one example of a molecular homogeneous catalyst that does not
require any additives for catalytic turnover,30 but this cationic ruthenium-based
catalyst is less active and requires higher temperatures than the systems that uti-
lize amines as additives.
An alternative that has been less extensively explored is using multiple catalysts to
affect the conversion of carbon dioxide to methanol via a cascade of chemical re-
actions. Sanford and co-workers demonstrated that such an approach is feasible.31
In this system, a ruthenium-based hydrogenation catalyst was first used to convert
carbon dioxide to formic acid. Next, a Lewis acid catalyst incorporated the meth-
anol solvent to convert formic acid into methyl formate. After the first two steps,
methyl formate was distilled to a second reactor containing another ruthenium-
based catalyst that reduced methyl formate to methanol. This reaction engineering
was necessary to address catalyst incompatibility that led to low catalyst activity.
More recently, Goldberg and co-workers have built upon this concept using
more stable homogeneous complexes by employing catalysts with improved sta-
bility, resulting in a 20-fold further increase in activity.32 Although the turnover
of this three-component system was lower than when amines were used as addi-
tives, carbon dioxide hydrogenation reactions that proceed through a formate
ester rather than a formamide intermediate are potentially beneficial because
esters are less thermodynamically stable than amides.33 Moreover, because an
alcohol is the product in the reaction, autocatalysis is possible if the reactions
could be carried out in a single reaction vessel because methanol can be used
as a reactant to drive the second step of the transformation. We hypothesized
that high turnover of carbon dioxide to methanol is achievable if reactor engineer-
ing is circumvented so that the three steps of the cascade catalysis occur within a
single reaction vessel. Inspired by the efficacy of protein superstructures used in
biological systems, we pursued a catalyst system for the cascade hydrogenation
of carbon dioxide to methanol, which involved isolating multiple active sites in a
nanostructured assembly (Figure 1B).
Chem 6, 1–13, July 9, 2020
3

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
RESULTS AND DISCUSSION
Cascade Catalysis for Converting Carbon Dioxide to Methanol Requires
Component Separation
Over the past decade, MOFs have emerged as attractive materials for their use in catal-
ysis.34 The highly tunable nature of MOFs makes them ideal materials for hosting tran-
sition-metal-based catalysts for efficient catalytic transformations.8,35–39 As a further
demonstration of the utility of using MOFs in catalysis, we envisioned that it would be
possible to effectively separate multiple catalysts by exploiting the aperture-opening
approach developed by our group to encapsulate catalytically active transition-metal
complexes into MOFs.36 This method relies on diffusion of a transition-metal complex
into pores of a MOF as a result of linker dissociation, which occurs readily in protic sol-
vents.36,40–42 Given that linker dissociation is very slow in polar aprotic or nonpolar sol-
vents, the host-guest constructs synthesized in protic solvents can be used as catalysts in
polar aprotic or nonpolar solvents without complications from catalyst leaching. Previ-
ously, we used this strategy to encapsulate the pincer complex (tBuPNP)RuH(CO)Cl
(1)26 in the MOF UiO-6643 (i.e., 1@UiO-66). The host-guest construct was an excellent
catalyst for the hydrogenation of carbon dioxide to formate, and it demonstrated
some advantages over the homogeneous analog: it prevented bimolecular decomposi-
tion, it was readily recycled, and it was less prone to poisoning.36
Although this method was useful for the conversion of carbon dioxide to formate, it did
not require the simultaneous operation of multiple catalysts that we hypothesized were
necessary for the conversion of carbon dioxide to methanol. To extend this method to
the three-step cascade hydrogenation of carbon dioxide to methanol, we first made
use of the transition-metal complex and the MOF in 1@UiO-66. 1 served as an active
catalyst to hydrogenate carbon dioxide to formic acid,26 and the Lewis acidic UiO-66
zirconium oxide nodes served as a catalyst to convert formic acid to a formate ester
when an alcohol additive was present (Table S1).44 To produce methanol as the ulti-
mate product, we added a homogeneous ruthenium-based complex (tBuPNN)
RuH(CO)Cl (2) known to catalyze ester hydrogenation31 to the reaction mixture with
1@UiO-66. This combination of catalysts and ethanol as a superstoichiometric additive
(10 mmol) produced methanol as the only detectable product (by 1H NMR spectros-
copy) from carbon dioxide (3 bar) and hydrogen (37 bar) with a turnover number
(TON) (expressed as mmol methanol per mmol 1) of 4,710 G 150 in N,N0-dimethylfor-
mamide (DMF) at 70ꢁC after 16 h (Table 1, entry 1). No other liquid products were
detectable by 1H NMR spectroscopy. Analysis of the reactor headspace by gas chroma-
tography revealed no gaseous products other than trace amounts of carbon monoxide
(CO) (1.9 G 1.1 ppm), which was not statistically distinct from the background produc-
tion of CO observed in the absence of catalyst (1.2 G 0.6 ppm). The small amount of CO
observed could be a consequence of the reverse water-gas shift reaction.
We evaluated the necessity for each component in this cascade reaction by carrying
out reactions under otherwise identical conditions by using each component individ-
ually (Table 1, entries 2–5). In these control experiments, 1, UiO-66, and 2 were
found to be inactive when used independently for the hydrogenation of carbon di-
oxide. Moreover, and importantly, all combinations of 1, 2, and UiO-66 resulted
in similar inactivity, including a combination of all three components in solution (Ta-
ble 1, entries 6–9). This last control experiment highlighted the benefits of the site
isolation achieved by encapsulation of at least one catalyst precursor in the MOF.
The site isolation achieved by the heterogeneous catalyst avoids possible bimolec-
ular decomposition pathways between catalysts 1 and 2 that would hinder the pro-
ductivity of the homogeneous system.45
4
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Table 1. Activity of Carbon Dioxide Hydrogenation Reactions under Various Cascade Catalysis
Conditions
(Catalysts)
CO2
+
H2
CH3OH
DMF, EtOH (10 mmol)
70ºC, 16h
(3 Bar)
(37 Bar)
Catalyst
Entry
1
2
3
4
5
6
7
8
9
–
U
–
–
–
U
U
–
U
1
–
–
U
–
–
U
–
U
U
UiO-66
U
–
–
–
–
U
–
–
–
U
U
U
2
U
–
U
–
–
–
1@UiO-66
TON
4,710
0
0
0
0
0
0
0
0
Reactions demonstrate the necessity of component separation for catalyst activity: 1 (2.23 3 10ꢀ7 mmol)
or 1@UiO-66 (10 mg, [Ru] = 2.23 3 10ꢀ7 mmol) for the hydrogenation of CO2 to formate, UiO-66 (10 mg)
for the esterification of formic acid to formate ester, and 2 (2.23 3 10ꢀ7 mmol) for the reduction of formate
ester to methanol. Check marks indicate which species are present in solution for each reaction under the
conditions specified in the scheme at the top of the table. Reaction mixtures were analyzed by 1H NMR
spectroscopy, which revealed no detectable liquid products other than methanol. Error in entry 1 (TON G
150) was determined as the average error of three runs (Figure S1).
Identity of Alcohol Additive Affects Catalytic Activity
We next evaluated the identity of the alcohol additive to further improve catalyst ac-
tivity. Results from these reactions revealed three factors important to catalyst per-
formance: the length of the alcohol, the degree of branching in the alcohol, and the
acidity of the alcohol (Figure 2A). When linear alcohols were used, a modest
decrease in TON was observed with increasing chain length (Figure 2A, blue).
Compared with linear alcohols, branched alcohols led to lower turnover for analogs
of the same molecular weight (Figure 2A, yellow). Finally, a significant increase in ac-
tivity was observed when the acidity of the alcohol was increased (Figure 2A, red). In
particular, greatly increased TON was observed with 2,2,2-trifluoroethanol (TFE),
leading to the highest turnover of any alcohol employed (TON = 6,600).
Chem 6, 1–13, July 9, 2020
5

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Figure 2. Optimization in the Cascade Hydrogenation of Carbon Dioxide to Methanol
(A) The length of linear alcohols (blue), the degree of branching (yellow), and the pKa of the alcohol
(red) affect activity. Solid circles are scaled and shaded proportionally to turnover to methanol,
which is also represented numerically, and the dotted circle denotes turnover to and through
formate ester. Error values displayed were determined as the average error of three runs.
(B) A Hammett plot for carbon dioxide hydrogenation reactions carried out with various para-
substituted aromatic alcohols suggests negative charge buildup (r = 0.166) during the rate-
determining step (conditions for all reactions: 2.23 3 10ꢀ7 mmol Ru per catalyst, 10 mmol alcohol
additive, 37 bar H2, 3 bar CO2, 70ꢁC, and 16 h).
Slower reaction rates with increasing alcohol size and branching suggested that
mass transport was important for the steps involving the alcohol additive. The faster
reaction rates observed for more acidic alcohols are consistent with a mechanism
that proceeds through an ester intermediate because the electrophilic formate ester
formed is expected to be more reactive. Interestingly, when 2-nitroethanol was used
as an additive, a significant amount of the formate ester intermediate was observed,
which was not observed when any other additive was used (MeOH/ester = 1.2:1).
This intermediate was most likely due to atypical formate ester stability.46 In this
case, turnover to or through the formate ester was similar to that of the two other
substituted ethanols tested (TONester = 6,100).
6
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Although a direct correlation between pKa and TON was not observed with the
aliphatic alcohols, the size differences for these alcohol additives are most likely
also important (see above). To better understand the electronic effects of the alcohol
on the reaction and to minimize the influence from steric bulk, we carried out reac-
tions by using para-substituted phenols as additives. Consistent with the results
obtained with aliphatic alcohols, electron-deficient phenols resulted in higher cata-
lytic turnover. A Hammett plot was generated from these results, revealing a good
correlation between TON and the Hammett s-value (Figure 2B). A r-value
of +0.166 was obtained from the slope of the line relating TON to s, which sug-
gested buildup of negative charge in the rate-determining step of the process. In
contrast, Hammett plots previously obtained for Fischer esterification reactions
have a negative r-value,47 suggesting that the rate-limiting step in the three-step
cascade hydrogenation reaction is most likely not esterification.
Catalyst Can Be Recycled through Encapsulation of Both Transition-Metal
Complexes
We next evaluated the importance of substrate trafficking by altering the identity of
the encapsulated catalyst (Figure 3). To this end, 2@UiO-66 was synthesized through
an aperture-opening encapsulation similar to that used for 1@UiO-66 (see Supple-
mental Information). When 2@UiO-66 and 1 were utilized with TFE as the additive,
a turnover of 5,700 was observed, which was lower than when 1@UiO-66 and 2
were used in the reaction (Figure 3A, rows 1 and 2). When the two heterogenized
constructs were used together, a TON of 3,500 was observed, which was just over
half the turnover achieved by 1@UiO-66 and 2 or 2@UiO-66 and 1 (Figure 3A, row
3). The combination of mass-transport limitations and the reversibility of the esteri-
fication step most likely led to lower catalyst turnover in this scenario. Nevertheless,
the fully heterogeneous system is advantageous because it can be readily recycled
(Figure 3B, left). No appreciable loss in activity was observed over five cycles, lead-
ing to an effective TON of approximately 17,500. Analysis of the catalyst construct
by scanning electron microscopy (Figure S2) and powder X-ray diffraction (Figure S3)
showed no degradation or significant change in the MOF host, and analysis of the
supernatant by inductively coupled plasma optical emission spectroscopy (ICP-
OES) revealed no detectable ruthenium. Moreover, the ruthenium loadings within
the MOF were similar before (4.55 ppm) and after (4.52 ppm) recycling.48
Finally, we evaluated the case where both pincer catalysts were encapsulated within
the same framework (Figure 3A, bottom). Because complex 1 is incompatible with
acetonitrile and 2 is incompatible with neat methanol, we synthesized the co-encap-
sulated construct [1,2]@UiO-66 by using sequential encapsulation involving encap-
sulation of 1 in methanol followed by encapsulation of 2 in acetonitrile (see Supple-
mental Information). Although the synthesis of [1,2]@UiO-66 is straightforward, a
unique analytical challenge exists in ensuring that 1 does not leach from 1@UiO-
66 during encapsulation of 2. We addressed this concern by using a similarly sized
iridium complex in place of 1 as a model complex for loading (Statement S2 and Ta-
ble S2). Because we could easily detect iridium or ruthenium concentrations simul-
taneously by ICP-OES, this model complex allowed us to assess whether complex
leaching was occurring during the second step of the sequential encapsulation strat-
egy. Satisfyingly, the co-encapsulated construct was confirmed to have been synthe-
sized without leaching. The activity of [1,2]@UiO-66 was notably higher than that
of mixtures of 1@UiO-66 and 2@UiO-66 (Figure 3A, row 4) but lower than the activity
of both partially homogeneous combinations. As was the case with the mixtures of
1@UiO-66 and 2@UiO-66, [1,2]@UiO-66 could be recycled through five cycles
Chem 6, 1–13, July 9, 2020
7

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Figure 3. Effect that the Identity of the Heterogenized Catalyst(s) Has on the Cascade
Hydrogenation of Carbon Dioxide to Methanol
(A) TON observed for reactions using 1 and/or 2 encapsulated in UiO-66. TON is expressed as
mmol methanol per mmol 1 and represents an average of three reactions; error is expressed as
average error.
(B) Recyclability studies for five cycles in the two fully heterogenized systems, 1@UiO-66 + 2@UiO-
66 and [1,2]@UiO-66. Each cycle is denoted by a black outlined box and numbered in white. TON
for each reaction is represented to the right of each column. Also see Figure S4.
with no appreciable decrease in activity or loss in ruthenium loading, leading to a cu-
mulative TON of 21,000 (Figure 3B, right).
Evidence for Autocatalysis Is Observed at Low Additive Loadings
Our working mechanistic hypothesis for the conversion of carbon dioxide to meth-
anol involves three steps: (1) hydrogenation of carbon dioxide to formic acid, (2)
conversion of formic acid to formate ester aided by the alcohol additive, and (3)
hydrogenation of the formate ester to give methanol and reform the alcohol
additive. As mentioned previously, because both products from the third step
of the reaction can serve as reactants in the second step of the reaction, the
overall transformation can be autocatalytic. This autocatalytic feature was previ-
ously unobservable because the reactions were performed in two separate ves-
sels31 or proceeded through a reaction mechanism where autocatalysis is not
possible.13,14,28–30 Although Goldberg and co-workers made use of a system in
which autocatalysis could be operative,32 the reaction conditions that were tested
prohibited observation of this effect.
To interrogate the possibility for autocatalysis, we carried out carbon dioxide hydro-
genation reactions at progressively decreasing loadings of TFE by using 1@UiO-66
and 2 as the multicomponent system (Figure 4). Turnover with respect to 1@UiO-66
decreased initially from 6,600 to 4,500 when TFE loading was decreased from 10 to
1 mmol (Figure 4, section i). However, as TFE loading was decreased further from 1
to 2.23 3 10ꢀ7 mmol, no appreciable change in TON was observed (Figure 4,
section ii). In this range, productivity of methanol with respect to TFE increased by
6 orders of magnitude from 1.08 3 10ꢀ3 to 4,500 mmol of product methanol per
mmol of TFE added, making the reaction catalytic in TFE. Below TFE loadings of
2.23 3 10ꢀ7 mmol, TON with respect to 1@UiO-66 began to decrease (Figure 4, sec-
tion iii), whereas productivity of methanol with respect to TFE remained constant.
8
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Figure 4. Conversion of Carbon Dioxide to Methanol at Various Concentrations of TFE Additive
The plot illustrates the effect of varying the concentration of TFE additive on catalyst productivity
with respect to ruthenium (orange) and TFE (red). Data points collected when the reaction is
catalytic in TFE are denoted with an open marker. Data are represented as an average of three runs,
and error bars represent average error.
These three regimes reflect complex behavior in the multicomponent system, which
can be explained if one considers the reaction mechanism. At high TFE loadings
(Figure 4, section i), the additive made up a major percentage of the solvent (19%
to 2% v/v), which significantly changed the solvent polarity. Solvent dielectric is ex-
pected to be an important contributor to the rate of the reaction considering that
negative charge buildup occurs during the rate-limiting step of the transformation
(Figure 2B). At intermediate TFE concentrations (Figure 4, section ii), the amount
of methanol produced remained constant regardless of the amount of TFE added.
However, the amount of methanol produced with respect to the TFE added steadily
increased as TFE decreased until [TFE] < 7.3 3 10ꢀ5 mM. Autocatalysis is indicated
by this observation because above a critical TFE concentration, enough methanol
can be produced so that the reaction proceeds predominately through a methyl
formate intermediate. In such a scenario, the reaction would not depend on the
amount of TFE, thereby allowing the additive to be used in catalytic quantities.
Finally, at very low TFE loadings (Figure 4, section iii), the decrease in TON as TFE
decreases and the constant amount of methanol produced with respect to TFE are
evidence that at these low TFE concentrations, the rate-determining step changes
so that the esterification reaction becomes rate limiting.
The observed autocatalytic behavior highlights a major advantage of our system
over reactions that proceed through a formamide intermediate. The major differ-
ence between these systems stems from the thermodynamic stability and (presum-
ably) kinetic reactivity of the ester and amide intermediates (Figure 5). Our results
mostly suggest that the conversion of the ester or amide to formaldehyde is rate
limiting (when [TFE] > 2.23 3 10ꢀ7 mmol). Thermodynamically, the pathway through
N,N0-diisopropylformamide (blue path) is 1 kcal/mol more stable than the pathway
through 2,2,2,-trifluoroethyl formate (orange path). The difference favors faster
Chem 6, 1–13, July 9, 2020
9

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
Figure 5. Hypothesized Mechanistic Root for Observed Activity in the Cascade Catalytic
Conversion of Carbon Dioxide to Methanol
Reaction coordinate diagram for converting carbon dioxide to methanol through a N,N0-
diisopropylformamide (blue path), 2,2,2-trifluoroethyl formate ester (orange path), or methyl
formate ester (red path) intermediate. Energies shown are enthalpic heats of reaction (DHrxn, kcal/
mol), which were obtained directly from thermochemical data or calculated with Benson group
increments.46
reactions at lower temperatures by proceeding through the formate ester interme-
diate rather than the formamide intermediate. This thermodynamic preference is
most likely further magnified in the transition state (not shown) for ester hydrogena-
tion compared with the amide reduction because the latter is more sterically encum-
bered and the former is expected to have greater electrophilicity. Compounding the
thermochemical and kinetic advantages, the transformation reported here benefits
from access to a second pathway through a methyl ester intermediate inaccessible
to reactions requiring the formation of an amide (red path, Figure 5). Although
methyl formate is thermodynamically uphill in relation to carbon dioxide and
hydrogen, the barrier can be overcome with sufficient buildup of methanol, which
ultimately results in autocatalysis.
Conclusion
The nanostructured assembly of three catalysts reported here for carbon dioxide
cascade hydrogenation to methanol is among the most active systems reported to
date. The reaction proceeds with high turnover at 70ꢁC and compares favorably
with the state-of-the-art homogeneous catalyst system (Table S3). The multicompo-
nent reaction is particularly active at low temperatures and with substoichiometric
amounts of TFE additive where it is autocatalytic. The high catalyst turnover
observed is consistent with lower thermodynamic barriers incurred for reactions
that proceed through formate ester intermediates than for formamide intermedi-
ates. Moreover, because multiple catalysts are used in this system, a step in the
transformation could be independently optimized without significantly affecting
the activity of the catalysts used for other steps of the transformation. This capability
highlights the modularity of the multicomponent approach and is reminiscent of how
biological systems evolve to improve reaction efficiency and selectivity. The ability
to modulate activity by changing the identity and positioning of the catalysts in
the multicomponent system is also similar to the supramolecular protein assemblies
used by enzymes. Finally, the benefits of the approach are highlighted by its ability
to access unique reaction manifolds, such as the autocatalysis observed here. This
last point is particularly exciting for the discovery of reactivity that has previously
10
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
remained dormant as a result of the one-catalyst-per-reaction paradigm that has
predominated synthetic catalyst design.
EXPERIMENTAL PROCEDURES
Resource Availability
Lead Contact
Further information and requests for resources and reagents should be directed to
and will be fulfilled by the Lead Contact, Chia-Kuang Tsung (frank.tsung@bc.edu).
Materials Availability
The hybrid catalysts developed in this study can be synthesized according to pro-
cedures available in the accompanying Supplemental Information. Characterization
data for these catalysts are available upon request from the authors.
Data and Code Availability
All data (except NMR data) used for evaluating catalyst TON are available in the main
text or Supplemental Information. NMR data are freely available to anyone through
contact with the corresponding authors.
SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.
2020.04.008.
ACKNOWLEDGMENTS
This research was supported by the Catalysis Program of the Basic Energy Sciences
program of the US Department of Energy Office of Science under award DE-
SC0019055. We acknowledge Charuni Gunathunge and Matthias Waegele for their
assistance with acquiring gas chromatography data. E.H.A. acknowledges a Beck-
man Scholars undergraduate fellowship through the Beckman Foundation.
AUTHOR CONTRIBUTIONS
Data Curation, T.M.R.; Formal Analysis, T.M.R., J.A.B., and C.-K.T.; Investigation,
T.M.R. and E.H.A.; Visualization, T.M.R., E.H.A., J.A.B., and C.-K.T.; Writing – Orig-
inal Draft, T.M.R., J.A.B., and C.-K.T.; Writing – Review and Editing, T.M.R., E.H.A.,
J.A.B., and C.-K.T.; Conceptualization, J.A.B. and C.-K.T.; Funding Acquisition,
J.A.B. and C.-K.T.; Project Administration, J.A.B. and C.-K.T.; Supervision, J.A.B.
and C.-K.T.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: November 19, 2019
Revised: January 14, 2020
Accepted: April 13, 2020
Published: May 5, 2020
REFERENCES
1. Calvin, M. (1956). The photosynthetic cycle.
Bull. Soc. Chim. Biol. 38, 1233–1244.
3. Morrow, S.M., Colomer, I., and Fletcher, S.P.
(2019). A chemically fuelled self-replicator. Nat.
Commun. 10, 1011.
hybrid biological | inorganic system.
J. Photochem. Photobiol. A 358, 411–415.
2. Chen, J., Gong, X., Li, J., Li, Y., Ma, J., Hou, C.,
Zhao, G., Yuan, W., and Zhao, B. (2018). Carbonyl
catalysis enables a biomimetic asymmetric
Mannich reaction. Science 360, 1438–1442.
5. Helm, M.L., Stewart, M.P., Bullock, R.M.,
DuBois, M.R., and DuBois, D.L. (2011). A
synthetic nickel electrocatalyst with a turnover
4. Liu, C., Colo´ n, B.E., Silver, P.A., and Nocera,
D.G. (2018). Solar-powered CO2 reduction by a
Chem 6, 1–13, July 9, 2020
11

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
frequency above 100,000 s⁻1 for H₂ production.
Science 333, 863–866.
transformation reactions in standard and
hybrid processes. Nat. Catal. 2, 381–386.
of CO2 to methanol. J. Am. Chem. Soc. 133,
18122–18125.
6. Chabolla, S.A., Machan, C.W., Yin, J.,
Dellamary, E.A., Sahu, S., Gianneschi,
N.C., Gilson, M.K., Tezcan, F.A., and
Kubiak, C.P. (2017). Bio-inspired CO2
reduction by a rhenium tricarbonyl
bipyridine-based catalyst appended to
amino acids and peptidic platforms:
incorporating proton relays and hydrogen-
bonding functional groups. Faraday
Discuss. 198, 279–300.
32. Chu, W.-Y., Culakova, Z., Wang, B.T., and
Goldberg, K.I. (2019). Acid-assisted
hydrogenation of CO2 to methanol in a
homogeneous catalytic cascade system. ACS
Catal. 9, 9317–9326.
19. Wang, W.H., Himeda, Y., Muckerman, J.T.,
Manbeck, G.F., and Fujita, E. (2015). CO2
hydrogenation to formate and methanol as
an alternative to photo- and
electrochemical CO2 reduction. Chem. Rev.
115, 12936–12973.
33. Ahmad, K., and Upadhyayula, S. (2019).
Greenhouse gas CO2 hydrogenation to fuels: a
thermodynamic analysis. Environ. Prog.
Sustain. Energy 38, 98–111.
20. Roy, S., Cherevotan, A., and Peter, S.C.
(2018). Thermochemical CO2 Hydrogenation
to single carbon products: scientific and
technological challenges. ACS Energy Lett.
3, 1938–1966.
34. Huang, Y.B., Liang, J., Wang, X.S., and Cao, R.
(2017). Multifunctional metal-organic
framework catalysts: synergistic catalysis and
tandem reactions. Chem. Soc. Rev. 46,
126–157.
7. Foster, S.L., Bakovic, S.I.P., Duda, R.D.,
Maheshwari, S., Milton, R.D., Minteer, S.D.,
Janik, M.J., Renner, J.N., and Greenlee, L.F.
(2018). Catalysts for nitrogen reduction to
ammonia. Nat. Catal. 1, 490–500.
21. Gogate, M.R. (2019). Methanol synthesis
revisited: reaction mechanisms in CO/CO2
hydrogenation over Cu/ZnO and DFT
analysis. Petrol. Sci. Technol. 37, 603–610.
35. Chen, L., and Xu, Q. (2019). Metal-organic
framework composites for catalysis. Matter 1,
57–89.
8. An, B., Li, Z., Song, Y., Zhang, J., Zeng, L.,
Wang, C., and Lin, W. (2019). Cooperative
copper centres in a metal-organic framework
for selective conversion of CO2 to ethanol. Nat.
Catal. 2, 709–717.
22. Heimann, J.E., Bernskoetter, W.H., and Hazari,
N. (2019). Understanding the individual and
combined effects of solvent and Lewis acid on
CO2 insertion into a metal hydride. J. Am.
Chem. Soc. 141, 10520–10529.
36. Li, Z., Rayder, T.M., Luo, L., Byers, J.A., andTsung,
C.K. (2018). Aperture-opening encapsulation of a
transition metal catalyst in a metal–organic
framework for CO2 hydrogenation.J. Am. Chem.
Soc. 140, 8082–8085.
9. Lima, C.G.S., Moreira, N.M., Paixa˜o, M.W., and
Correˆ a, A.G. (2019). Heterogenous green
catalysis: application of zeolites on
multicomponent reactions. Curr. Opin. Green
Sustain. Chem. 15, 7–12.
23. Tanaka, R., Yamashita, M., and Nozaki, K.
(2009). Catalytic hydrogenation of carbon
dioxide using Ir(III)-pincer complexes. J. Am.
Chem. Soc. 131, 14168–14169.
37. Mon, M., Adam, R., FerrandoꢀSoria, J., Corma,
A., Armentano, D., Pardo, E., and LeyvaꢀPe´ rez,
A. (2018). Stabilized Ru[(H2O)6]3+ in confined
spaces (MOFs and zeolites) catalyzes the
Imination of primary alcohols under
10. Jose´ Climent, M.J., Corma, A., and Iborra, S.
(2012). Homogeneous and heterogeneous
catalysts for multicomponent reactions. RSC
Adv. 2, 16–58.
24. Valde´ s, H., Garcıa-Eleno, M.A., Canseco-
´
Gonzalez, D., and Morales-Morales, D. (2018).
Recent advances in catalysis with transition-
metal pincer compounds. ChemCatChem 10,
3136–3172.
atmospheric conditions with wide scope. ACS
Catal. 8, 10401–10406.
11. Bruggink, A., Schoevaart, R., and Kieboom, T.
(2003). Concepts of nature in organic synthesis:
cascade catalysis and multistep conversions in
concert. Org. Process Res. Dev. 7, 622–640.
38. Qin, J.S., Yuan, S., Lollar, C., Pang, J.,
Alsalme, A., and Zhou, H.C. (2018). Stable
metal-organic frameworks as a host
platform for catalysis and biomimetics.
Chem. Commun. 54, 4231–4249.
25. Wiedner, E.S., and Linehan, J.C. (2018).
Making a splash in Homogeneous CO2
hydrogenation: elucidating the impact of
solvent on catalytic mechanisms. Chemistry
24, 16964–16971.
12. Furukawa, H., Cordova, K.E., O’Keefe, M., and
Yaghi, O.M. (2013). The chemistry and
applications of metal-organic frameworks.
Science 341, 1230444.
39. Buru, C.T., Wasson, M.C., and Farha, O.K. (2020).
H5PV2Mo10O40 polyoxometalate encapsulated
in NU-1000 metal-organic framework for aerobic
oxidation of a mustard gas simulant. ACS Appl.
Nano Mater. 3, 658–664.
26. Filonenko, G.A., van Putten, R., Schulpen,
E.N., Hensen, E.J.M., and Pidko, E.A. (2014).
Highly efficient reversible hydrogenation of
carbon dioxide to formates Using a
ruthenium PNP-pincer catalyst.
13. Everett, M., and Wass, D.F. (2017). Highly
productive CO2 hydrogenation to methanol –
a tandem catalytic approach via amide
intermediates. Chem. Commun. 53, 9502–
9504.
ChemCatChem 6, 1526–1530.
40. Kim, M., Cahill, J.F., Fei, H., Prather, K.A., and
Cohen, S.M. (2012). Postsynthetic ligand and
cation exchange in robust metal–organic
27. Kar, S., Kothandaraman, J., Goeppert, A.,
and Prakash, G.K.S. (2018). Advances in
catalytic homogeneous hydrogenation of
carbon dioxide to methanol. J. CO2 Util. 23,
212–218.
14. Kar, S., Sen, R., Kothandaraman, J., Goeppert,
A., Chowdhury, R., Munoz, S.B., Haiges, R., and
Prakash, G.K.S. (2019). Mechanistic insights
into ruthenium-pincer-catalyzed amine-
assisted homogeneous hydrogenation of CO2
to methanol. J. Am. Chem. Soc. 141, 3160–
3170.
frameworks. J. Am. Chem. Soc. 134, 18082–18088.
41. Morabito, J.V., Chou, L.Y., Li, Z., Manna, C.M.,
Petroff, C.A., Kyada, R.J., Palomba, J.M., Byers,
J.A., and Tsung, C.K. (2014). Molecular
encapsulation beyond the aperture size limit
through dissociative linker exchange in metal-
organic framework crystals. J. Am. Chem. Soc.
136, 12540–12543.
28. Rezayee, N.M., Huff, C.A., and Sanford, M.S.
(2015). Tandem amine and ruthenium-
catalyzed hydrogenation of CO2 to methanol.
J. Am. Chem. Soc. 137, 1028–1031.
15. Ribeiro, A.P.C., Martins, L.M.D.R.S., and
Pombeiro, A.J.L. (2017). Carbon dioxide-to-
methanol single-pot conversion using a C-
scorpionate iron(II) catalyst. Green Chem 19,
4811–4815.
29. Kothandaraman, J., Goeppert, A., Czaun,
M., Olah, G.A., and Prakash, G.K.S. (2016).
Conversion of CO2 from air into methanol
using a polyamine and a homogeneous
ruthenium catalyst. J. Am. Chem. Soc. 138,
778–781.
42. Karagiaridi, O., Bury, W., Mondloch, J.E.,
Hupp, J.T., and Farha, O.K. (2014). Solvent-
assisted linker exchange: an alternative to the
de novo synthesis of unattainable metal-
organic frameworks. Angew. Chem. Int. Ed. 53,
4530–4540.
16. Chu, S., Li, W., Yan, Y., Hamann, T., Shih, I.,
Wang, D., and Mi, Z. (2017). Roadmap on solar
water splitting: current status and future
prospects. Nano Futures 1, 022001.
30. Wesselbaum, S., Moha, V., Meuresch, M.,
Brosinski, S., Thenert, K.M., Kothe, J., vom
43. Cavka, J.H., Jakobsen, S., Olsbye, U., Guillou,
N., Lamberti, C., Bordiga, S., and Lillerud, K.P.
(2008). A new zirconium inorganic building
brick forming metal organic frameworks with
exceptional stability. J. Am. Chem. Soc. 130,
13850–13851.
Stein, T., Englert, U., Holscher, M.,
¨
17. Olah, G.A., Goeppert, A., and Prakash, G.K.S.
(2009). Chemical recycling of carbon dioxide to
methanol and dimethyl ether: from
Klankermeyer, J., and Leitner, W. (2015).
Hydrogenation of carbon dioxide to
methanol using a homogeneous ruthenium–
triphos catalyst: from mechanistic
investigations to multiphase catalysis.
Chem. Sci. 6, 693–704.
greenhouse gas to renewable, environmentally
carbon neutral fuels and synthetic
hydrocarbons. J. Org. Chem. 74, 487–498.
44. Cirujano, F.G., Corma, A., and Llabre´ s i
Xamena, F.X. (2015). Conversion of levulinic
acid into chemicals: synthesis of biomass
derived levulinate esters over Zr-containing
MOFs. Chem. Eng. Sci. 124, 52–60.
18. Tackett, B.M., Gomez, E., and Chen, J.G.
(2019). Net reduction of CO2 via its
thermocatalytic and electrocatalytic
31. Huff, C.A., and Sanford, M.S. (2011). Cascade
catalysis for the homogeneous hydrogenation
12
Chem 6, 1–13, July 9, 2020

Please cite this article in press as: Rayder et al., A Bioinspired Multicomponent Catalytic System for Converting Carbon Dioxide into Methanol
Autocatalytically, Chem (2020), https://doi.org/10.1016/j.chempr.2020.04.008
ll
Article
45. When alcohol additive was replaced with
formic acid (10 mmol), no turnover to
methanol was observed (Figure S5),
suggesting the presence of ester as
an important intermediate in turnover
to methanol and highlighting the
importance of the Lewis acid catalyst
in the system.
46. Domalski, E.S., and Hearing, E.D. (1993).
Estimation of the thermodynamic properties of
C-H-N-O-S-halogen compounds at 298.15 K.
J. Phys. Chem. Ref. Data 22, 805–1159.
effects on esterification kinetics. J. Chem.
Educ. 87, 845–847.
48. Catalyst loadings were manipulated to
best achieve a 1:1 ratio of 1 to 2 for
1@UiO-66 + 2@UiO-66 and for [1,2]@UiO-
66 and do not reflect maximum possible
loadings.
47. Smith, J.B., Byrd, H., O’Donnell, S.E., and
Davis, W. (2010). Hammett parameter and
molecular-modeling correlations of substituent
Chem 6, 1–13, July 9, 2020
13