Role of ligand-bound CO2in the hydrogenation of CO2to formate with a (PNP)Mn catalyst

Article abstract of DOI:10.1021/acscatal.1c01709

Herein, we describe the catalytic hydrogenation of CO2 to formate with (PNP)Mn-H (PNP = 2,6-bis(di-tert-butylphosphinomethyl)- pyridine; Mn = Mn(CO)2). Contrary to the established mechanism for CO2 hydrogenation, mechanistic studies indicate that CO2 does not insert into the Mn-H bond of (PNP)Mn-H to give the formate species, (PNP)Mn- OCHO. The lack of reactivity is confirmed by thermochemical studies that show that (PNP)Mn-H is not sufficiently hydridic to reduce CO2. Deprotonation of the hydride to give [(?PNP)Mn-H]- ((? indicates the deprotonated ligand) enhances the hydricity by ~17 kcal·mol-1 and hence should be sufficiently hydridic to hydrogenate CO2. This reactivity is not observed, and CO2 instead binds to the backbone to generate another anionic hydride species [(CO2-PNP)Mn-H]. The formate is lost only from this species, through hydride transfer to an external CO2. These findings are unexpected because substrate binding to the backbone of catalysts that can undergo metal-ligand cooperativity (MLC) is thought to be detrimental to catalysis; this work suggests that alternative mechanisms should be considered. The enhanced hydricity observed upon deprotonation may be broadly applicable to systems capable of undergoing MLC. Moreover, this work shows an example of how thermochemical analysis can be used to advance mechanistic understanding in (de)hydrogenation catalysis.

Full text of DOI:10.1021/acscatal.1c01709

pubs.acs.org/acscatalysis
Research Article
Role of Ligand-Bound CO2 in the Hydrogenation of CO2 to Formate
with a (PNP)Mn Catalyst
Kevin Schlenker, Elizabeth G. Christensen, Asylbek A. Zhanserkeev, Gabriel R. McDonald,
Emily L. Yang, Kevin T. Lutz, Ryan P. Steele, Ryan T. VanderLinden, and Caroline T. Saouma*
Cite This: ACS Catal. 2021, 11, 8358−8369
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Herein, we describe the catalytic hydrogenation of CO2 to
formate with (PNP)Mn−H (PNP = 2,6-bis(di-tert-butylphosphinomethyl)-
pyridine; Mn = Mn(CO)2). Contrary to the established mechanism for CO2
hydrogenation, mechanistic studies indicate that CO2 does not insert into the
Mn−H bond of (PNP)Mn−H to give the formate species, (PNP)Mn−
OCHO. The lack of reactivity is confirmed by thermochemical studies that
show that (PNP)Mn−H is not sufficiently hydridic to reduce CO2.
Deprotonation of the hydride to give [(*PNP)Mn−H]− (* indicates the
deprotonated ligand) enhances the hydricity by ∼17 kcal·mol−1 and hence should be sufficiently hydridic to hydrogenate CO2. This
reactivity is not observed, and CO2 instead binds to the backbone to generate another anionic hydride species [(CO2-PNP)Mn−H].
The formate is lost only from this species, through hydride transfer to an external CO2. These findings are unexpected because
substrate binding to the backbone of catalysts that can undergo metal−ligand cooperativity (MLC) is thought to be detrimental to
catalysis; this work suggests that alternative mechanisms should be considered. The enhanced hydricity observed upon
deprotonation may be broadly applicable to systems capable of undergoing MLC. Moreover, this work shows an example of how
thermochemical analysis can be used to advance mechanistic understanding in (de)hydrogenation catalysis.
KEYWORDS: CO2 hydrogenation, catalysis, metal−ligand cooperativity, mechanisms, thermodynamic parameters
makes use of MLC.18 The dearomatized (or deprotonated)
catalyst heterolytically adds H2, with ligand protonation
INTRODUCTION
■
Advancing the hydrogenation of CO2 to formate or formic acid
(FA) is motivated by increased global energy demands.1−3
concomitant to formation of the metal hydride. CO2 then
inserts into the metal hydride to give a formate-bound species.
Formic acid is primarily produced from CO on a 720 000 t
annual scale,4 and the development of formic acid fuel cells,5
The base-mediated formate loss ensues via ligand deprotona-
using FA as a H2 carrier,6 and further hydrogenation to
tion to regenerate the dearomatized (or deprotonated) species.
MeOH1 will only increase demand. Despite numerous
Alternative mechanisms that do not rely on MLC have also
been proposed,30,31 as indicated by the dashed arrows in
examples of homogeneous catalysts that hydrogenate CO2 to
FA, better catalysts should be developed that allow for high
turnover numbers (TONs) and frequencies (TOFs) under
mild conditions.1
Scheme 1.
An underexplored pathway for CO2 hydrogenation involves
activation of CO2 through MLC to generate a (CO2−L)M
adduct. In this reaction, the Lewis-basic ligand attacks the
Lewis-acidic carbon of CO2 to give a ligand-bound species and
one oxygen of the bound CO2 coordinates the metal, giving
the metal an electron count of 18 (Scheme 1, right).
Ru,14,18,32,33 Mn,27 and other metals25,34−36 have been shown
to activate CO2 in this manner. Moreover, other hydrogenation
substrates can coordinate to catalysts capable of MLC
The CO2 insertion into a M−H to give M−OCHO is
thought to be a critical step in the mechanism of converting
CO2 to formate or FA.7−10 Insight into this process has been
gleaned from kinetic studies that explore how solvent, Lewis
acids, and ligands impact the rate of CO2 insertion,11,12 as well
as thermodynamic analysis of intermediate metal hy-
drides.13−17
The insertion of CO2 into the M−H is also pertinent for
catalysts that are capable of undergoing metal−ligand
cooperativity (MLC). Pincer-ligated Ru,14,18−20 Co,21,22
Fe,23−25 and Mn26,27 complexes have emerged as promising
catalysts for the hydrogenation of CO2 to formate, with some
Ru catalysts capable of converting CO2 to MeOH14,19,28,29 in
the presence of amines. A proposed mechanism is shown in
Scheme 1 (left) for the hydrogenation of CO2 to formate that
Received: April 14, 2021
Revised: June 5, 2021
Published: June 23, 2021
https://doi.org/10.1021/acscatal.1c01709
© 2021 American Chemical Society
ACS Catal. 2021, 11, 8358−8369
8358

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Scheme 1. (Left): Proposed Mechanism for CO2 Hydrogenation to Formate by (PNE)-Ligated Ru (E = PtBu2 or NEt2) That
Undergoes MLC (Black Arrows); the Alternative Direct (dashed) Pathway is also Shown. (Right): Ligand-Bound CO2 is
Thought to be Detrimental to Catalysis or to Engage in an Alternative, Minor Pathway
analogously,37−39 suggesting that this reactivity may occur
more broadly.
Catalytic Hydrogenation of CO2. The hydrogenation of
CO2 in the presence of 100 equiv of KOtBu was examined
using two Mn complexes, the putative active hydride,
(PNP)Mn−H,38 and the ligand-bound CO2 species, (CO2-
PNP)Mn. This species is readily prepared by treatment of
(*PNP)Mn with 0.85 atm.45 of CO2 (the dearomatized ligand
is designated by *).
For CO2 hydrogenation at Ru, mechanistic studies by
Sanford suggest that a CO2-bound species may hydrogenate
CO2 under an alternative mechanistic scheme (Scheme 1,
right);18 this pathway is thought to be minor relative to that
shown on the left of Scheme 1. However, density functional
theory (DFT) studies by Pidko on a related catalyst suggest
that a pathway with a ligand-bound CO2 is not favored,40 and
Milstein suggests that the analogous species on Fe is off-
cycle.25
In the presence of 100 equiv of KOtBu, (PNP)Mn−H
hydrogenates CO2 to formate (Table 1, entries 1−4). With
Table 1. Hydrogenation of CO2 by Mn Catalysts
Given that binding of CO2 is competitive to H2 binding at
Ru,14 we sought to unequivocally establish if such complexes
are pertinent to the hydrogenation of CO2. Herein, we describe
the catalytic hydrogenation of CO2 to formate with a Mn
catalyst. Mechanistic and thermochemical studies are not
consistent with the established step of CO2 inserting into a
neutral metal hydride, akin to that shown on the left of Scheme
1. While thermodynamically favorable, hydrogenation of CO2
to formate by an anionic Mn−H also does not occur. Rather,
our studies suggest that formate is produced from an anionic
Mn−H that has a ligand-bound CO2, analogous to that shown
in Scheme 1 (right). Compared to the prior work on Ru,18 our
studies suggest that formate loss occurs from the ligand-bound
intermediate. Given the ability to deprotonate catalysts that
undergo MLC, and the range of catalysts that react with the
substrate via MLC at the backbone,14,18,25,27,32−39 the results
are broadly relevant to advancing (de)hydrogenation mecha-
nisms at catalysts that can undergo MLC.
b
PCO
(bar)
PH
2
(bar)
T
TON
2
a
entry
(°C)
catalyst
formate
1
2
3
4
8
8
8
8
50
100
50
100
100
100
100
100
100
(PNP)Mn−H
(PNP)Mn−H
(PNP)Mn−H
(PNP)Mn−H
(PNP)Mn−H
(CO2-PNP)Mn
--
9 (8)
17 (6)
5 (3)
58 (15)
2 (1)
49 (15)
0
15
15
15
15
15
15
0
15
15
15
15
15
0
c
5
6
7
8
9
(PNP)Mn−H
(PNP)Mn−H
0
0
15
a
b
Reactions were run in 10 mL of tetrahydrofuran (THF). Quantified
via ion chromatography, see the Supporting Information (SI).
Average of 3 runs, with the standard deviation given in parentheses.
c
No base added.
RESULTS AND DISCUSSION
■
Choice of the System. Given that Mn has been shown to
hydrogenate26,27,41 and electrocatalytically reduce CO2 to
formate,16,42 as well as dehydrogenate formic acid to CO2
and H2,43,44 we sought to explore the hydrogenation of CO2
equivalent pressures of CO2 and H2, the turnover number
(TON) increases as the temperature is increased from 50 to
100 °C and with increasing the overall pressure of the gases. As
shown in entry 5, addition of a base is required for the
formation of formate (or FA).
Next, the catalytic activity of (CO2-PNP)Mn was consid-
ered. Under identical conditions to (PNP)Mn−H (entries 4 vs
6), (CO2-PNP)Mn gives the same TON within error.
Omission of the catalyst, H2, or CO2 gives no formate (entries
7−9). The unoptimized catalytic trials clearly indicate that
using (PN P)Mn −H (PNP
= 2, 6-bis(di-tert-
butylphosphinomethyl)pyridine; Mn = Mn(CO)2).38 This
ligand was chosen so that the thermochemical parameters
can be compared to the Ru analogue, (PNP)Ru(CO)H2,14
which differs in metal identity and replacement of one CO
with a hydride.
8359
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Scheme 2. Plausible Mechanisms Consistent With the Catalytic Hydrogenations
both (CO2-PNP)Mn and (PNP)Mn−H can access a catalytic
cycle.
of the free ligand (Scheme 3). Attempts to stabilize
(PNP)Mn−OCHO by placement under a CO2 atmosphere
Though ligand-bound CO2 species have been invoked to be
detrimental to catalysis,25,40 these results are consistent with
the following 3 scenarios. First, catalysis ensues through
(PNP)Mn−H (Scheme 2, bottom), and (CO2-PNP)Mn is off-
cycle but can enter the cycle through loss of CO2. This
scenario is consistent with established mechanisms for CO2
hydrogenation. Second, catalysis ensues through (CO2-
PNP)Mn (Scheme 2, top) and (PNP)Mn−H is off-cycle but
can enter through the loss of H2. Such a mechanism has no
literature precedence and is included for completeness. Third,
both (CO2-PNP)Mn and (PNP)Mn−H can be transformed
to a catalytically viable intermediate (Scheme 2, right). Given
that catalysis proceeds under basic conditions and prior work
on Ru,18 [(CO2-PNP)Mn−H]− is considered a plausible
intermediate that is accessible from both (CO2-PNP)Mn and
(PNP)Mn−H (Scheme 2, right). All three scenarios invoke
MLC at the ligand.
Scheme 3. (Top): Stability of (PNP)Mn−OCHO in THF.
(Bottom): Reactivity of (PNP)Mn−OCHO toward the Base
Stoichiometric Reactions Pertinent to Formate
Production via (PNP)Mn−H. To determine if the cycle
that proceeds through (PNP)Mn−H is correct (Scheme 2,
bottom), stoichiometric studies were conducted to determine
feasibility. Independent synthesis of (PNP)Mn−OCHO was
undertaken to identify its formation and reactivity toward the
base. Treatment of equimolar (*PNP)Mn and formic acid in
pentanes results in the precipitation of a yellow powder. The
were not successful (see Figure S19). This suggests that the
observed reactivity is not simply a deinsertion of CO2; the
product (PNP)Mn−H is stable under N2 and CO2.
To determine if CO2 inserts into the Mn−H bond to give a
formate species, a THF solution of (PNP)Mn−H was placed
under 0.85 atm of CO2.45 As ascertained by 31P NMR
spectroscopy, no reaction occurred. When the reaction was
repeated using 30 bar of CO2 at 50 °C, the only species present
upon cooling and replacing under a N2 atmosphere was (CO2-
PNP)Mn (Scheme 4).
It is conceivable that small amounts of (PNP)Mn−OCHO
form in the equilibrium of (PNP)Mn−H with CO2, and that
deprotonation of (PNP)Mn−OCHO produces formate and
(*PNP)Mn (Scheme 2, bottom cycle). Indeed, treatment of
(PNP)Mn−OCHO with 1 equiv of KOtBu rapidly produces
(*PNP)Mn and KOCHO (Scheme 3). If this mechanism is
1
solid has an IR stretch at 1625 cm−1, and the H NMR
spectrum has a resonance at 8.47 ppm, both consistent with a
formate species. A single resonance at 105.9 ppm is observed
in the 31P NMR spectrum. Taken together, this is consistent
with the product being (PNP)Mn−OCHO.
Unlike the Ru analogue, (PNP)Ru(CO)(H)(OCHO),14
THF solutions of (PNP)Mn−OCHO are not stable; over the
course of hours, conversion to (PNP)Mn−H occurs with loss
8360
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
mechanism shown in the top of Scheme 2, provided that
once (*PNP)Mn is formed it reacts with H2 to generate
(PNP)Mn−H and enters the bottom cycle (reversibly).
However, (CO2-PNP)Mn is readily deprotonated with 1
equiv of KOtBu to give [(*CO2-PNP)Mn]− (Scheme 5). The
anionic species further reacts with H2 to give [(CO2-
PNP)Mn−H]− (shown in Scheme 2, right, and vide infra),
whereas neutral (CO2-PNP)Mn requires elevated pressures of
H2 to give (PNP)Mn−H and (PNP)Mn−OCHO (vide
supra). These results are not consistent with the mechanism
shown at the top of Scheme 2 under ambient conditions.
Taken together, the mechanistic studies suggest that a
mechanism that proceeds via neutral (CO2-PNP)Mn (with
no deprotonation, top of Scheme 2) is not viable under mild
conditions.
Scheme 4. Stoichiometric Reactions Pertinent to the
(PNP)Mn−H Cycle
correct, then treatment of (PNP)Mn−H with 1 equiv of
KOtBu at 0.85 atm of CO2 should result in formate loss and
the formation of (*PNP)Mn (Scheme 4); no reaction is
observed (Figure S48). Taken together, this suggests that the
mechanism that proceeds via neutral (PNP)Mn−H is not
operational under the mild conditions investigated. These
conditions differ from those of Table 1, where the elevated
temperature may allow for kinetic barriers to be overcome
and/or for different thermodynamics parameters to be
present,17 while the elevated pressures may shift equilibria
between the catalyst and gaseous substrates.
Stoichiometric Reactions Pertinent to Formate
Production via (CO2-PNP)Mn. Though there is no
precedence for H2 “insertion” into species akin to (CO2-
PNP)Mn, it is considered as a mechanistic possibility (Scheme
2, top). Under ambient conditions (0.85 atm H2, 20 °C), no
reaction occurs between (CO2-PNP)Mn and H2. After
exposure to 30 bar of H2 and 50 °C, the major species in
solution upon returning to ambient conditions is (PNP)Mn−
H, with ∼13% of the desired (PNP)Mn−OCHO (Scheme 5).
Treatment of (CO2-PNP)Mn with 1 equiv of KOtBu and
0.85 atm H2 results in formate loss and the formation of
(PNP)Mn−H (Scheme 5). This is consistent with the
Thermochemical Analysis Pertinent to Catalysis via
Neutral Cycles. To better understand the observed reactivity
that is pertinent to the mechanisms of Scheme 2, the equilibria
shown in Scheme 6 were measured or extrapolated. These
measurements were all done in THF, to best describe their
values under catalytic conditions. The equilibrium of
(*PNP)Mn with H2 or CO2 to give (PNP)Mn−H or (CO2-
PNP)Mn, respectively (K1 or K2), was monitored by UV−vis
spectroscopy. To observe equilibrium mixtures, very low
pressures of either gas had to be introduced via a calibrated
bulb (see the SI), and the solution allowed to equilibrate.
As shown in Table 2, the binding of H2 (−3.3 kcal·mol−1)
and CO2 (−3.7 kcal·mol−1) is favorable under standard
conditions. These values are comparable to those obtained
for (*PNP)Ru(CO)(H), whereby similar energies are
associated with both H2 (−4.1 kcal·mol−1) and CO2 (−3.8
kcal·mol−1).14 Under 30 bar of gas, the free energies for both
equilibria decrease by 2 kcal·mol−1, further stabilizing (PNP)-
Mn−H and (CO2-PNP)Mn relative to (*PNP)Mn.
The equilibrium of (*PNP)Mn with FA to give (PNP)Mn−
OCHO (K3, ΔGFA) was monitored by 31P NMR spectroscopy
Scheme 5. Stoichiometric Reactions Pertinent to the (CO2-PNP)Mn Cycle
8361
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Scheme 6. Equilibrium Constants and Thermodynamic Parameters Measured (K1, K2, K3, K4, K11) and Extrapolated (K5, K6,
K12) in THF
Table 2. Equilibrium Constants and Free-Energy Changes
Associated With the Equilibria of Scheme 6
Scheme 7. Equilibria Employed to Determine the
Hydricity
a
a
d
e
equilibrium
K1 (ΔGH
K
ΔG° (kcal·mol−1
−3.3 0.2
)
ΔG (30 bar)
−5.3
b
b
)
280 10
480 10
2
K2 (ΔGCO
)
−3.7 0.2
−5.7
2
b
c
K3 (ΔGFA
)
∼1100
∼−4.1
K4 (ΔG4)
1.6 × 10−19
9.8 × 10−35
∼9.1 × 10−6
6.3 × 10−32
2.5 × 10−22
25.6 0.3
46.4 0.4
∼6.9
42.5 0.3
29.5 0.3
c
c
c
c
−
K5 (ΔGH
K6 (ΔGOCHO
K11 (ΔG11)
K12 (ΔGH (anion)
)
a
Free energies are in units of kcal·mol−1. The value of ΔG7 is from ref
−
)
46.
−
)
a
b
c
All measurements were done at 25 °C. Units in M−1. Units in M.
From the sum of equations shown in Scheme 7, the hydricity
of (PNP)Mn−H in THF is found to be 46.4 kcal·mol−1.
Compared to (PNP)Ru(CO)(H)2, the Mn hydride is 1.8 kcal·
d
e
Under standard conditions. Conversion to nonstandard conditions:
ΔGx = ΔGx°− RT ln(Y) where Y is the partial pressure of gas.14 Units
in kcal·mol−1.
mol−1 less hydridic (higher ΔGH value). This difference in
−
hydricity has also been corroborated by direct equilibrium
measurements made between the Mn and Ru systems (see
Figure S43).
The lack of reactivity between (PNP)Mn−H and CO2 to
give formate and [(PNP)Mn]+ indicates that (PNP)Mn−H is
less hydridic than formate. While the hydricity of formate is
not known in THF, we previously reported a lower limit of
44.7 kcal·mol−1.14 Hence, direct hydride transfer from
(PNP)Mn−H to CO2 is <1.7 kcal·mol−1 uphill.
The above analysis assumes that the resulting formate does
not bind to the metal. As shown by Miller48 and us,14 the
effective hydricity can be increased by binding of an anion to
the resulting cation, in this case, formate binding yields
(PNP)Mn−OCHO. Hence, the effective hydricity is the sum
of the hydricity of the hydride, in this case (PNP)Mn−H, and
the energy associated with binding the anion to the cation, in
this case formate binding to [(PNP)Mn]+ (sum of vertical
equilibria in Scheme 8).
(Scheme 6, top-middle to bottom-left). This approach allows
for quantification of (PNP)Mn−H and other degradation
products that arise from the instability of (PNP)Mn−OCHO
(Scheme 3). Because the reaction of (*PNP)Mn with FA is
fast to equilibrate and the subsequent reactivity is slow, the
value obtained is assumed to be ∼ representative of K3 and
found to be ∼1100 M−1 under standard conditions (see the SI
for details). Hence, (PNP)Mn−OCHO is favored over
(*PNP)Mn by ∼4.1 kcal·mol−1 with 1 M FA. The analogous
equilibrium on Ru14 is favored by 3.5 kcal·mol−1 under
standard conditions. Despite the equilibrium to yield (PNP)-
Mn−OCHO being more favorable than on Ru, degradation is
only observed from (PNP)Mn−OCHO.
To determine if hydride transfer from (PNP)Mn−H to CO2
is thermodynamically favorable, the hydricity of (PNP)Mn−H
must be determined. This is done following the thermochem-
ical cycle in Scheme 7, whereby the first two equations
correspond to equilibria that are pictorially shown in Scheme
6.
To determine the effective hydricity upon formate binding,
the energy associated with formate release from (PNP)Mn−
OCHO to give [(PNP)Mn]+ must be known (ΔGOCHO , K6).
−
Titration of [(PNP)Mn]+ with various equivalents of the
phosphazene base, (OMeC5H4N)P(pyrr)3, (pKip = 16.8),47
gives a pKip value of 18.8 0.2 for [(PNP)Mn]+. The cation is
more acidic than [(PNP)Ru(CO)H]+, which has a pKip of
20.7 0.2.14 The increased acidity on Mn suggests that the
anionic charge on the ligand is better stabilized, likely through
enhanced π-backbonding interactions relative to the Ru
analogue.
From the equations of Scheme 9, the binding of formate to the
−1
−
cation (−ΔGOCHO ) is favorable by ∼6.9 kcal·mol ; this value
compares well to that at other Mn(I) complexes.16 Hence, the
effective hydricity (Scheme 8) is ∼39.5 kcal·mol−1. If the
effective hydricity is greater than the hydricity of formate, then
CO2 insertion into (PNP)Mn−H to give (PNP)Mn−OCHO
((K10)−1) is thermodynamically favorable. This analysis shows
that by accounting for formate coordination, the reaction with
8362
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
ligand. These findings suggest that (PNP)Mn−(OCHO) may
not be relevant to the catalytic cycle, particularly under mild
conditions.
Scheme 8. Equilibria Pertinent to CO2 Deinsertion and the
Effective Hydricity
a
The combined reactivity and thermochemical studies
suggest that the two catalytic cycles proposed in Scheme 2
may not be operational. As catalysis occurs in the presence of
excess KOtBu, we hypothesized that under catalytic conditions,
both (PNP)Mn−H and (CO2-PNP)Mn may give rise to a
common anionic hydride intermediate (Scheme 2, right).
Deprotonation of (PNP)Mn−H. Isoelectronic anionic
metal hydrides are more hydridic than neutral counterparts,
and reduction of a metal hydride to give an anionic congener
also increases the hydricity.49 While there are no examples of
the effect that a ligand deprotonation has on the hydricity of
complexes that undergo MLC, mechanistic studies by Kempe
indicate that a dianionic Mn−H is the active catalyst for
hydrogenation of imines.50 Hence, it is expected that an
anionic congener of (PNP)Mn−H will be more hydridic.
n
Treatment of (PNP)Mn−H with BuLi or Schlosser’s base
results in clean conversion to [(*PNP)Mn−H]− (only partial
conversion is observed with KOtBu, see Figure S46). The
formally 20-electron species is stable as a solid, provided there
is no trace moisture in the glovebox atmosphere. Though we
have not been able to obtain an X-ray diffraction (XRD)
structure, this species has been thoroughly characterized by
NMR and IR spectroscopy. Consistent with ligand deproto-
nation, the 31P NMR spectrum shows an AB pattern with two
a
An alternative thermocycle to obtain K10 is indicated by the product
of equilibrium constants.
Scheme 9. Equilibria Employed to Determine Formate
Release
a
1
doublets at 124.6 and 114.9 ppm (J = 56 Hz). The H NMR
spectrum is diagnostic of dearomatization, with the formally
aromatic peaks shifted to 5.23−6.04 ppm. Notably, a hydride
resonance that is split into a doublet of doublets is present at
−4.57 ppm (JHP = 56, 64 Hz). The anion can be protonated by
the addition of a weak acid (HBF4·BrC6H4P1(pyrr)) to
quantitatively regenerate (PNP)Mn−H (see Figure S60).
The hydricity of [(*PNP)Mn−H]− is related to the
hydricity of neutral (PNP)Mn−H, as shown in Scheme 6
and eq 1.
a
Free energies are in units of kcal·mol−1. The value of ΔG8 is from ref
14.
CO2 should be favored by ∼5.2 kcal·mol−1. As this reactivity is
not observed, it suggests that the barrier for CO2 insertion is
too high. Moreover, that hydride transfer ensues from the Ru
analogue that has an effective hydricity of 40.9 kcal·mol−1,
reinforces the conclusion that in terms of thermodynamics,
CO2 should insert into (PNP)Mn−H to give (PNP)Mn−
OCHO (Scheme 8, top).
From the equilibria shown in Scheme 8, the free energy
associated with CO2 deinsertion from (PNP)Mn−(OCHO)
to give (PNP)Mn−H under standard conditions is 4.8 kcal·
mol−1. This value can be modulated by changing the partial
pressure of CO2. The instability of (PNP)Mn−(OCHO)
described in Scheme 3 is thus a consequence of the other
products that are formed.
The thermochemical analysis provides two key results. First,
(PNP)Mn−H is not sufficiently hydridic to hydrogenate CO2
to give [(PNP)Mn]+ and OCHO−. Subsequent formate
binding modulates the effective hydricity to now make it
thermodynamically favorable, indicating a kinetic barrier (at
room temperature), which is proposed to be overcome under
catalytic conditions run at elevated temperatures. This suggests
that another hydride may form under catalytic conditions.
Second, if (PNP)Mn−(OCHO) forms under elevated
pressures & temperatures of catalysis, then excess base must
drive the loss of formate (see Figure S44)14,18 at a faster rate
than that of (PNP)Mn−(OCHO) degradation that gives free
−
−
ΔGH + 1.364·pKa = ΔGH (12) + 1.364·pKa(11)
(1)
From Hess’s law, knowing the pKa of [(*PNP)Mn−H]−
−
allows for determination of it’s hydricity, ΔGH (anion)
.
Titration of (PNP)Mn−H with the Wittig base, Ph3P=CH-
(Me)2 (pKip = 28.7),51 gave a pKip value of 31.2 0.2. From
eq 1, the hydricity of [(*PNP)Mn−H]− is found to be 29.5
( 0.4) kcal·mol−1. Hence, deprotonation of the ligand
backbone enhances the hydricity by 16.9 kcal·mol−1. The
magnitude of this shift is impressive given that the hydricity
scale for 1st row transition-metal hydrides in MeCN spans ∼40
kcal·mol−1.49
Reactivity of [(*PNP)Mn−H]− toward CO2. It is
anticipated that treatment of [(*PNP)Mn−H]− with CO2
will result in formate loss via direct hydride transfer, as the
hydride transfer is favorable by 15.2 kcal·mol−1. Indeed, the
reaction of [(*PNP)Mn−H]− with 0.85 atm of CO2 proceeds
to produce formate (84% yield), but the product of direct
hydride transfer, (*PNP)Mn, is not observed (Figure 1, top).
Rather, (CO2-PNP)Mn is the major species in solution
(∼92%), and (PNP)Mn−H is also present (∼8%). Given
the moisture sensitivity of [(*PNP)Mn−H]−, the neutral
hydride is postulated to form from trace moisture.
Monitoring the reaction by 31P NMR spectroscopy indicates
that within 3 min of CO2 addition, [(*PNP)Mn−H]− is
consumed and replaced by two new species that have
8363
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 1. (Top): Observed reactivity of [(*PNP)Mn−H]− toward CO2. (Bottom): 31P NMR spectra taken at various times.
inequivalent phosphines (minor: δ = 151.5 & 134.6 ppm, J =
61 Hz; major: δ = 145.2 & 128.5 ppm, J = 59 Hz). Peaks
associated with (CO2-PNP)Mn are present at 30 min and
continue to grow (as the intermediate peaks decrease) until
∼24 h, when the reaction is complete. When the reaction of
[(*PNP)Mn−H]− is repeated with higher CO2 pressure (∼1.5
atm), the reaction is complete within 4 h. Inclusion of H2 to
the reaction does not impact the reaction rate.
minimized structures of these two isomers in implicit solvent
indicate that the up conformer is lower in Gibbs free energy by
0.82 kcal·mol−1. Although this energy gap is likely near the
error bars of these DFT computations, the existence of two
isomers of nearly equal energy is nonetheless consistent with
the experimental observations.
Hence, direct hydride transfer from [(*PNP)Mn−H]− to
CO2 does not occur. The lack of reactivity is speculated to be
due to [(*PNP)Mn−H]− being a formally 20-electron species.
CO2 insertion that proceeds via an inner-sphere mechanism
necessitates transient coordination to the metal;11 in the
absence of hemilabile ligands, this would increase the formal
electron count to 22. While an outer-sphere mechanism,
whereby the hydride directly attacks CO2 may be feasible, the
observed reactivity indicates that the HOMO of [(*PNP)-
Mn−H]− resides primarily on the ligand, consistent with the
DFT calculations (Figure S79).
Addition of 1 equiv of CO2 to [(*PNP)Mn−H]− gives rise
to the same two intermediates observed above (Figure S66).
Over the course of hours, resonances associated with the
kinetic product (major species at 3 min) decrease as it is
converted to the thermodynamic product (minor species at 3
min). Of note, the resonances associated with the kinetic
product are broadened with fewer CO2 equiv, consistent with a
fluxional process. Two hydride resonances are observed in the
1H NMR spectrum, with integration that matches that
observed in the 31P NMR spectrum. When the reaction is
repeated with 13CO2, the 13C NMR spectrum shows two
resonances at 176.6 and 175.1 ppm, which correspond to
ligand-bound CO2. Taken together, the NMR data suggest that
the two intermediates are isomers of [(CO2-PNP)Mn−H]−,
whereby one has the CO2 pointing “up” toward the hydride,
and the other pointing “down” away from the hydride. DFT-
Deprotonation of (CO2-PNP)Mn. Given that both (CO2-
PNP)Mn and (PNP)Mn−H are precatalysts, if the mechanism
proceeds via [(CO2-PNP)Mn−H]−, then this species should
be accessible starting from (CO2-PNP)Mn. Deprotonation of
(CO2-PNE)Ru(CO)(H) occurs at the ligand to yield [(*CO2-
PNE)Ru(CO)(H)]−.14,18 Sanford previously observed that
introduction of 1 bar of H2 and CO2 to [(*CO2-PNN)Ru-
8364
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(CO)(H)]− generates (CO2-PNN)Ru(CO)(H) along with a
15% yield of KOCHO.18
value is more acidic compared to that of (PNP)Mn−H (pKip =
31.2).
Treatment of (CO2-PNP)Mn with 1 equiv of KOtBu
generates [(*CO2-PNP)Mn][K]. The 31P NMR spectrum has
doublets at 118.7 and 113.7 ppm (J = 104 Hz). Furthermore,
Reactivity of [(*CO2-PNP)Mn]− With H2. Addition of
0.85 atm of H2 to [(*CO2-PNP)Mn]− results in a color
change and precipitation of a colorless solid over 24 h. After
heating to 65 °C, NMR spectroscopy indicates that the major
Mn species in solution is (PNP)Mn−H. IR and NMR
spectroscopy was used to identify the solid as KOCHO.
Quantification of formate by ion chromatography gives 93%
yield.
1
the H NMR spectrum of (CO2-PNP)Mn shows that the
aromatic pyridine peaks of the (CO2-PNP)Mn shift to 6.21,
5.81, and 5.39 ppm, indicating dearomatization. Single-crystal
X-ray diffraction confirmed the dearomatization, with the bond
length of the pyridine carbon and the attached methylene
spacer decreasing from 1.499 (3) to 1.39 (1)Å (Figure 2). Of
To gain further insight, the reaction was monitored by 31P
and 1H NMR spectroscopy (Figure 3). After addition of H2 to
[(*CO2-PNP)Mn]−, the 31P NMR spectrum shows a pair of
doublets that correspond to the thermodynamic isomer of
[(CO2-PNP)Mn−H]− described above. Of note, this species
remains a minor component of the reaction mixture, with
[(*CO2-PNP)Mn]− present at all time points; higher H2
pressures are required to shift the equilibrium (see the SI).
Over time, signals associated with this species decay as those
for (PNP)Mn−H grow. The reaction requires heating and
days for completion. No other identifiable intermediates are
observed.
The conversion of [(*CO2-PNP)Mn]− to formate and
(PNP)Mn−H formally requires 2 equiv of H2. When only 1
equiv of H2 is added to [(*CO2-PNP)Mn]−, ∼30% converts
to (PNP)Mn−H.
Proposed Mechanism. We postulate that loss of formate
from [(CO2-PNP)Mn−H]− and conversion to (PNP)Mn−H
could occur by 3 different mechanisms. Bimolecular mecha-
nisms, whereby 2 equiv of anionic [(CO2-PNP)Mn−H]− must
react with one another, are not considered due to steric and
electrostatic arguments. Likewise, a mechanism that proceeds
from the neutral (PNP)Mn−OCHO (gray arrows) is not
thought to be operative, at least for low-temperature catalysis
(vide infra), given the thermodynamic and mechanistic studies
presented above. As catalysis ensues from [(CO2-PNP)Mn−
H]−, whereby the ligand is deprotonated, revised Noyori-type
mechanisms52,53 (whereby the protonated ligand does not
transfer a proton but rather serves as a hydrogen-bond donor)
are also not considered.
Figure 2. Thermal ellipsoid plots (50%) of (top): (CO2-PNP)Mn
and (bottom): [(*CO2-PNP)Mn][K]. Hydrogen atoms and solvent
molecules are removed for clarity, and only one Mn from the unit cell
is shown for the anionic species. Select bond distances (Å) for (CO2-
PNP)Mn: Mn1−O3: 2.074(2); O3−C3: 1.287(2); C3−O4:
1.227(3); C3−C4: 1.554(3); C4−C5: 1.500(3); and C9−C10:
1.499(3). Select bond distances (Å) for [(*CO2-PNP)Mn][K]:
Mn1−O3: 2.099(7); O3−C10: 1.273(2); C10−O4: 1.23(1); C9−
C10: 1.53(1); C4−C7: 1.51(1); C3−C4: 1.39(1). Similar bond
distances are obtained about Mn2 of the unit cell.
The first mechanism considered is unimolecular formate loss
from [(CO2-PNP)Mn−H]− to give (*PNP)Mn, which then
reacts with CO2 or H2 to generate (CO2-PNP)Mn or
(PNP)Mn−H, respectively (Scheme 10, tan arrows). Given
that the reaction of (*PNP)Mn with either gas is complete
within minutes and similar thermodynamics associated with
the two reactions, if this mechanism is correct then the rate of
formate loss should have no dependence on CO2 or H2. The
data presented are not consistent with this mechanism.
Alternatively, H2 addition across the C−C bond of [(CO2-
PNP)Mn−H]− generates (PNP)Mn−H and formate in a
single step (Scheme 10, blue arrows). If this mechanism is
correct, then addition of H2 to the reaction of [(CO2-
PNP)Mn−H]− and CO2 should be faster (compared to that
with no H2 added). That there is no rate enhancement, and
that the reaction of [(*CO2-PNP)Mn]− with excess H2 is
significantly slower (and requires heating), suggests that this
mechanism is also incorrect.
note in the crystal structure of [(*CO2-PNP)Mn][K] is the
∼0.02 Å increase in the Mn−O bond distance relative to
(CO2-PNP)Mn. Furthermore, the IR stretch associated with
the bound CO2 shifts from 1651 to 1605 cm−1 upon
deprotonation.
The anion, [(*CO2-PNP)Mn]−, exists as a potassium-
bridged dimer in the solid state. Treatment with 18-crown-6
yields the monomeric potassium adduct, confirmed through
single-crystal X-ray diffraction. Attempts to prepare analogous
sodium or lithium species were not successful and showed
limited stability (see the SI). This suggests that the potassium
is essential in stabilizing the anion.
Finally, the hydride of [(CO2-PNP)Mn−H]− can react with
free CO2 to generate formate and (CO2-PNP)Mn (Scheme
10, pink arrow). This is consistent with the observed rate
enhancement with more CO2 pressure. Moreover, the
Titration of (CO2-PNP)Mn with the Wittig base, Ph3P
CH(CH3)2 (pKip = 28.7),51 yielded a pKip = 28.7 0.2. This
8365
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 3. (Top): Observed reactivity of [(*CO2-PNP)Mn]− with H2. (Bottom): 31P NMR taken at different time points. The spectrum collected
after 7 days was heated at 65 °C for 6 days.
requirement of heat in the reaction of [(*CO2-PNP)Mn−H]−
and H2 suggests that CO2 may first have to be lost.
mechanisms shown in the top and bottom of Scheme 2 (and
top of Scheme 10) are likely not operative; this may not hold
for catalysis under harsher conditions that employ elevated
temperatures and higher pressures.
Application to Room-Temperature Hydrogenation of
CO2. The use of a first-row transition-metal catalyst for room-
temperature hydrogenation of CO2 is relatively rare.1,54,55 The
stoichiometric reactions described above suggest that room-
temperature hydrogenation of CO2 may be feasible. Addition
of CO2 to [(*PNP)Mn−H]− readily occurs under ambient
conditions (0.85 atm), while addition of H2 to [(*CO2-
PNP)Mn]− is thermodynamically less favorable (both
reactions give [(CO2-PNP)Mn−H]−). This difference is due
to the difference in pKa values (see the SI). Given that [(CO2-
PNP)Mn−H]− reacts with ambient CO2 to lose formate to
generate (CO2-PNP)Mn, it is anticipated that higher H2
pressures may be required relative to CO2 to achieve turnover.
Indeed, with 100 equiv of KHMDS, (PNP)Mn−H catalyti-
cally hydrogenates CO2 at room temperature using 2 atm of
CO2 and 15 atm of H2 (TON = 5) (see SI). While the
unoptimized TON is modest, it suggests that knowing the
thermodynamic parameters of the system can lead to catalysis
under milder conditions. These hydrogenation conditions are
similar to those of the stoichiometric studies described (vide
supra), indicating that under these mild conditions, the
CONCLUSIONS
■
Herein, we presented the catalytic hydrogenation of CO2 to
formate at a PNP-ligated Mn catalyst. Both reactivity studies
and thermochemical analysis indicate that the traditional
mechanism, whereby CO2 inserts into (PNP)Mn−H, is not
likely to occur (Scheme 10, gray arrows). When formate
binding is considered, the effective hydricity of the catalyst is
sufficient to reduce CO2, yet this reactivity is not observed.
This contrasts with studies on the Ru analogue.14,18 These
results underscore how subtle changes in the catalyst can alter
the reactivity.
Additionally, it was found that deprotonation of the ligand
backbone of (PNP)Mn−H generates [(*PNP)Mn−H]−,
which is more hydridic by 16.9 kcal·mol−1. This appears to
be an underexplored design strategy and may be pertinent to
many published hydrogenation reactions. Notably, mechanistic
work by Kempe indicates that the reactive hydride is dianionic
in a related system that can undergo MLC.50 Bergens showed
8366
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
[(CO2-PNP)Mn−H]− can be accessed and is sufficiently
hydridic to reduce CO2. This intermediate is 18-electron and
lacks an electrophilic site on the ligand, allowing for hydride
transfer to occur. While the hydricity of this species cannot be
measured due to its inherent instability, addition of the
negative charge increases the hydricity. Given that other
carbonyl substrates bind MLC catalysts analogously, mecha-
nisms that proceed with the “bound” substrate should be
considered, particularly when the equilibrium toward H2 rivals
that toward the substrate (CO2, in this case). We emphasize
that this mechanism may not be dominant under harsher
conditions (increased pressure and temperature), and that it
may not hold for other MLC catalysts, particularly those that
undergo deprotonation at an amine.52,53 Nonetheless,
analogous mechanisms should be considered in mechanistic
schemes, particularly those where deprotonation occurs at a
nonpolar C−H bond.
Scheme 10. Possible Mechanisms for CO2 Hydrogenation to
Formate
a
Overall, this work is pertinent to advancing mechanistic
schemes and catalyst design strategies for (de)hydrogenation
catalysts. The results emphasize that caution must be taken in
considering mechanisms, particularly in systems that are
capable of undergoing MLC; substrate binding to the
backbone and hydride transfer from a deprotonated species
were identified in the system investigated. Consideration of
such deviations from established mechanistic hypotheses
should be considered more broadly. This is particularly true
for CO2 hydrogenations at catalysts capable of undergoing
MLC, which are run under basic conditions.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.1c01709.
■
sı
a
All reactions/equilibria with solid arrows have been observed, and
All experimental details, including syntheses, spectro-
scopic data, equilibrium plots (PDF)
Crystallographic data (CIF)
those with dashed lines are included for completeness. The
mechanism shown at the bottom is consistent with all experimental
data.
AUTHOR INFORMATION
Corresponding Author
that hydride transfer to amides and imines at a Noyori-type Ru
catalyst occurs after deprotonation of the ligand (to give an
anionic catalyst);56 it was found that catalytic amounts of the
base were required as the conjugate acid, which then
protonated the substrate. Given that most catalytic trans-
formations mediated by catalysts capable of undergoing MLC
occur in the presence of catalytic amounts of the base,57−59 it is
conceivable that hydride transfer occurs from an anionic
deprotonated species.
Despite now having sufficient thermodynamic driving force
to hydrogenate CO2 (in the absence of formate binding to the
Mn), no formate is directly produced from [(*PNP)Mn−H]−.
Rather, CO2 coordinates the backbone, and formate is lost
from [(CO2-PNP)Mn−H]−. It is speculated that the lack of
CO2 insertion is due to the high-electron count about the Mn,
with the reactivity occurring at the ligand instead of at the Mn.
These findings reinforce the importance of considering relative
kinetic barriers and just not only the thermodynamic
parameters in predicting reactivity.
■
Caroline T. Saouma − Department of Chemistry, University of
Utah, Salt Lake City, Utah 84112, United States;
orcid.org/0000-0003-1170-6175;
Email: caroline.saouma@utah.edu
Authors
Kevin Schlenker − Department of Chemistry, University of
Utah, Salt Lake City, Utah 84112, United States
Elizabeth G. Christensen − Department of Chemistry,
University of Utah, Salt Lake City, Utah 84112, United
States
Asylbek A. Zhanserkeev − Department of Chemistry,
University of Utah, Salt Lake City, Utah 84112, United
States
Gabriel R. McDonald − Department of Chemistry, University
of Utah, Salt Lake City, Utah 84112, United States
Emily L. Yang − Department of Chemistry, University of Utah,
Salt Lake City, Utah 84112, United States
Kevin T. Lutz − Department of Chemistry, University of Utah,
Salt Lake City, Utah 84112, United States
Ryan P. Steele − Department of Chemistry, University of
Utah, Salt Lake City, Utah 84112, United States;
orcid.org/0000-0002-3292-9805
The mechanism that is most consistent with the
experimental data proceeds via a catalytic cycle that
necessitates a ligand-bound CO2 (Scheme 10, bottom).
While such species have been invoked as detrimental,25,40
both thermochemical and reactivity studies suggest that it is
required for the hydrogenation of CO2. Now, anionic hydride
8367
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(10) Waldie, K. M.; Ostericher, A. L.; Reineke, M. H.; Sasayama, A.
F.; Kubiak, C. P. Hydricity of Transition-Metal Hydrides:
Thermodynamic Considerations for CO2 Reduction. ACS Catal.
2018, 8, 1313−1324.
(11) Heimann, J. E.; Bernskoetter, W. H.; Hazari, N.; Mayer, J. M.
Acceleration of CO2 insertion into metal hydrides: ligand, Lewis acid,
and solvent effects on reaction kinetics. Chem. Sci. 2018, 9, 6629−
6638.
Ryan T. VanderLinden − Department of Chemistry,
University of Utah, Salt Lake City, Utah 84112, United
States; orcid.org/0000-0002-0128-1284
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.1c01709
Author Contributions
(12) Heimann, J. E.; Bernskoetter, W. H.; Hazari, N. Understanding
the Individual and Combined Effects of Solvent and Lewis Acid on
CO2 Insertion into a Metal Hydride. J. Am. Chem. Soc. 2019, 141,
10520−10529.
The manuscript was written through contributions of all
authors. All authors have given approval to the final version of
the manuscript.
Funding
(13) Ceballos, B. M.; Yang, J. Y. Directing the reactivity of metal
hydrides for selective CO2 reduction. Proc. Natl. Acad. Sci. U.S.A.
2018, 115, 12686−12691.
The authors gratefully acknowledge start-up funding from the
University of Utah & USTAR and NSF CAREER (1945646).
The NMR results included in this report were recorded at the
David M. Grant NMR Center, a University of Utah Core
Facility. Funds for construction of the Center and the helium
recovery system were obtained from the University of Utah
and the National Institutes of Health awards 1C06RR017539-
01A1 and 3R01GM063540-17W1, respectively. NMR instru-
ments were purchased with the support of the University of
Utah and the National Institutes of Health award
1S10OD25241-01.
(14) Mathis, C. L.; Geary, J.; Ardon, Y.; Reese, M. S.; Philliber, M.
A.; VanderLinden, R. T.; Saouma, C. T. Thermodynamic Analysis of
Metal-Ligand Cooperativity of PNP Ru Complexes: Implications for
CO2 Hydrogenation to Methanol and Catalyst Inhibition. J. Am.
Chem. Soc. 2019, 141, 14317−14328.
(15) Taheri, A.; Thompson, E. J.; Fettinger, J. C.; Berben, L. A. An
Iron Electrocatalyst for Selective Reduction of CO2 to Formate in
Water: Including Thermochemical Insights. ACS Catal. 2015, 5,
7140−7151.
(16) Bhattacharya, M.; Sebghati, S.; VanderLinden, R. T.; Saouma,
C. T. Toward Combined Carbon Capture and Recycling: Addition of
an Amine Alters the Product Selectivity from CO to Formic Acid in
Manganese Catalyzed Reduction of CO2. J. Am. Chem. Soc. 2020, 142,
17589−17597.
(17) Hu, J.; Bruch, Q. J.; Miller, A. J. M. Temperature and Solvent
Effects on H2 Splitting and Hydricity: Ramifications on CO2
Hydrogenation by a Rhenium Pincer Catalyst. J. Am. Chem. Soc.
2021, 143, 945−954.
(18) Huff, C. A.; Sanford, M. S. Catalytic CO2 Hydrogenation to
Formate by a Ruthenium Pincer Complex. ACS Catal. 2013, 3, 2412−
2416.
(19) Kothandaraman, J.; Goeppert, A.; Czaun, M.; Olah, G. A.;
Prakash, G. K. S. Conversion of CO2 from Air into Methanol Using a
Polyamine and a Homogeneous Ruthenium Catalyst. J. Am. Chem.
Soc. 2016, 138, 778−781.
(20) Filonenko, G. A.; van Putten, R.; Schulpen, E. N.; Hensen, E. J.
M.; Pidko, E. A. Highly Efficient Reversible Hydrogenation of Carbon
Dioxide to Formates Using a Ruthenium PNP-Pincer Catalyst.
ChemCatChem 2014, 6, 1485.
(21) Daw, P.; Chakraborty, S.; Leitus, G.; Diskin-Posner, Y.; Ben-
David, Y.; Milstein, D. Selective N-Formylation of Amines with H2
and CO2 Catalyzed by Cobalt Pincer Complexes. ACS Catal. 2017, 7,
2500−2504.
(22) Scheuermann, M. L.; Semproni, S. P.; Pappas, I.; Chirik, P. J.
Carbon Dioxide Hydrosilylation Promoted by Cobalt Pincer
Complexes. Inorg. Chem. 2014, 53, 9463−9465.
(23) Langer, R.; Diskin-Posner, Y.; Leitus, G.; Shimon, L. J. W.; Ben-
David, Y.; Milstein, D. Low-Pressure Hydrogenation of Carbon
Dioxide Catalyzed by an Iron Pincer Complex Exhibiting Noble Metal
Activity. Angew. Chem., Int. Ed. 2011, 50, 9948−9952.
(24) Zhang, Y.; MacIntosh, A. D.; Wong, J. L.; Bielinski, E. A.;
Williard, P. G.; Mercado, B. Q.; Hazari, N.; Bernskoetter, W. H. Iron
catalyzed CO2 hydrogenation to formate enhanced by Lewis acid co-
catalysts. Chem. Sci. 2015, 6, 4291−4299.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Dr. Peter Flynn for the help with NMR experiments.
Royce Song, Maxwell Reese, and Mallory Philliber helped
obtain preliminary hydrogenation data.
REFERENCES
■
(1) Wang, W. H.; Himeda, Y.; Muckerman, J. T.; Manbeck, G. F.;
Fujita, E. CO2 Hydrogenation to Formate and Methanol as an
Alternative to Photo- and Electrochemical CO2 Reduction. Chem. Rev.
2015, 115, 12936−12973.
(2) Burkart, M. D.; Hazari, N.; Tway, C. L.; Zeitler, E. L.
Opportunities and Challenges for Catalysis in Carbon Dioxide
Utilization. ACS Catal. 2019, 9, 7937−7956.
(3) Hepburn, C.; Adlen, E.; Beddington, J.; Carter, E. A.; Fuss, S.;
Mac Dowell, N.; Minx, J. C.; Smith, P.; Williams, C. K. The
technological and economic prospects for CO2 utilization and
removal. Nature 2019, 575, 87−97.
(4) Artz, J.; Muller, T. E.; Thenert, K.; Kleinekorte, J.; Meys, R.;
̈
Sternberg, A.; Bardow, A.; Leitner, W. Sustainable Conversion of
Carbon Dioxide: An Integrated Review of Catalysis and Life Cycle
Assessment. Chem. Rev. 2018, 118, 434−504.
(5) Yu, X.; Pickup, P. G. Recent advances in direct formic acid fuel
cells (DFAFC). J. Power Sources 2008, 182, 124−132.
(6) Grasemann, M.; Laurenczy, G. Formic acid as a hydrogen source
− recent developments and future trends. Energy Environ. Sci. 2012, 5,
8171−8181.
(7) Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.; Hazari, N.
Secondary Coordination Sphere Interactions Facilitate the Insertion
Step in an Iridium(III) CO2 Reduction Catalyst. J. Am. Chem. Soc.
2011, 133, 9274−9277.
(25) Rivada-Wheelaghan, O.; Dauth, A.; Leitus, G.; Diskin-Posner,
Y.; Milstein, D. Synthesis and Reactivity of Iron Complexes with a
New Pyrazine-Based Pincer Ligand, and Application in Catalytic Low-
Pressure Hydrogenation of Carbon Dioxide. Inorg. Chem. 2015, 54,
4526−4538.
(8) Ziebart, C.; Federsel, C.; Anbarasan, P.; Jackstell, R.; Baumann,
W.; Spannenberg, A.; Beller, M. Well-Defined Iron Catalyst for
Improved Hydrogenation of Carbon Dioxide and Bicarbonate. J. Am.
Chem. Soc. 2012, 134, 20701−20704.
(9) Wesselbaum, S.; Moha, V.; Meuresch, M.; Brosinski, S.; Thenert,
K. M.; Kothe, J.; Stein, T. V.; Englert, U.; Hölscher, M.;
Klankermayer, J.; Leitner, W. Hydrogenation of carbon dioxide to
methanol using a homogeneous ruthenium-Triphos catalyst: from
mechanistic investigations to multiphase catalysis. Chem. Sci. 2015, 6,
693−704.
(26) Kar, S.; Goeppert, A.; Kothandaraman, J.; Prakash, G. K. S.
Manganese-Catalyzed Sequential Hydrogenation of CO2 to Methanol
via Formamide. ACS Catal. 2017, 7, 6347−6351.
(27) Kumar, A.; Daw, P.; Espinosa-Jalapa, N. A.; Leitus, G.; Shimon,
L. J. W.; Ben-David, Y.; Milstein, D. CO2 activation by manganese
8368
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
pincer complexes through different modes of metal-ligand coopera-
tion. Dalton Trans. 2019, 48, 14580−14584.
Complexes and the Manganese-Mediated Decomposition of Formic
Acid. Organometallics 2016, 35, 2049−2052.
(45) Atmospheric pressure in Salt Lake City, Utah, is ∼ 0.85 atm.
(46) Brereton, K. R.; Jadrich, C. N.; Stratakes, B. M.; Miller, A. J. M.
Thermodynamic Hydricity across Solvents: Subtle Electronic Effects
and Striking Ligation Effects in Iridium Hydrides. Organometallics
2019, 38, 3104−3110.
(47) Kaljurand, I.; Saame, J.; Rodima, T.; Koppel, I.; Koppel, I. A.;
Kogel, J. F.; Sundermeyer, J.; Kohn, U.; Coles, M. P.; Leito, I.
Experimental Basicities of Phosphazene, Guanidinophosphazene, and
Proton Sponge Superbases in the Gas Phase and Solution. J. Phys.
Chem. A 2016, 120, 2591−2604.
(48) Pitman, C. L.; Brereton, K. R.; Miller, A. J. M. Aqueous
Hydricity of Late Metal Catalysts as a Continuum Tuned by Ligands
and the Medium. J. Am. Chem. Soc. 2016, 138, 2252−2260.
(49) Wiedner, E. S.; Chambers, M. B.; Pitman, C. L.; Bullock, R. M.;
Miller, A. J. M.; Appel, A. M. Thermodynamic Hydricity of Transition
Metal Hydrides. Chem. Rev. 2016, 116, 8655−8692.
(50) Freitag, F.; Irrgang, T.; Kempe, R. Mechanistic Studies of
Hydride Transfer to Imines from a Highly Active and Chemoselective
Manganate Catalyst. J. Am. Chem. Soc. 2019, 141, 11677−11685.
(51) Saame, J.; Rodima, T.; Tshepelevitsh, S.; Kutt, A.; Kaljurand, I.;
Haljasorg, T.; Koppel, I. A.; Leito, I. Experimental Basicities of
Superbasic Phosphonium Ylides and Phosphazenes. J. Org. Chem.
2016, 81, 7349−7361.
(52) Mandal, S. C.; Rawat, K. S.; Pathak, B. A computational study
on ligand assisted vs. ligand participation mechanisms for CO2
hydrogenation: importance of bifunctional ligand based catalysts.
Phys. Chem. Chem. Phys. 2019, 21, 3932−3941.
(53) Dub, P. A.; Gordon, J. C. The role of the metal-bound N−H
functionality in Noyori-type molecular catalysts. Nat. Rev. Chem.
2018, 2, 396−408.
(54) Vollmer, M. V.; Ye, J.; Linehan, J. C.; Graziano, B. J.; Preston,
A.; Wiedner, E. S.; Lu, C. C. Cobalt-Group 13 Complexes Catalyze
CO2 Hydrogenation via a Co(−I)/Co(I) Redox Cycle. ACS Catal.
2020, 10, 2459−2470.
(55) Jeletic, M. S.; Mock, M. T.; Appel, A. M.; Linehan, J. C. A.
Cobalt-Based Catalyst for the Hydrogenation of CO2 under Ambient
Conditions. J. Am. Chem. Soc. 2013, 135, 11533−11536.
(56) John, J. M.; Takebayashi, S.; Dabral, N.; Miskolzie, M.;
Bergens, S. H. Base-Catalyzed Bifunctional Addition to Amides and
Imides at Low Temperature. A New Pathway for Carbonyl
Hydrogenation. J. Am. Chem. Soc. 2013, 135, 8578−8584.
(57) Alig, L.; Fritz, M.; Schneider, S. First-Row Transition Metal
(De)Hydrogenation Catalysis Based On Functional Pincer Ligands.
Chem. Rev. 2019, 119, 2681−2751.
(28) Rezayee, N. M.; Huff, C. A.; Sanford, M. S. Tandem Amine and
Ruthenium-Catalyzed Hydrogenation of CO2 to Methanol. J. Am.
Chem. Soc. 2015, 137, 1028−1031.
(29) Khusnutdinova, J. R.; Garg, J. A.; Milstein, D. Combining Low-
Pressure CO2 Capture and Hydrogenation To Form Methanol. ACS
Catal. 2015, 5, 2416−2422.
(30) Filonenko, G. A.; Conley, M. P.; Copéret, C.; Lutz, M.;
Hensen, E. J. M.; Pidko, E. A. The impact of Metal−Ligand
Cooperation in Hydrogenation of Carbon Dioxide Catalyzed by
Ruthenium PNP Pincer. ACS Catal. 2013, 3, 2522−2526.
(31) Filonenko, G. A.; Smykowski, D.; Szyja, B. M.; Li, G.; Szczygieł,
J.; Hensen, E. J. M.; Pidko, E. A. Catalytic Hydrogenation of CO2 to
Formates by a Lutidine-Derived Ru−CNC Pincer Complex:
Theoretical Insight into the Unrealized Potential. ACS Catal. 2015,
5, 1145−1154.
(32) Vogt, M.; Gargir, M.; Iron, M. A.; Diskin-Posner, Y.; Ben-
David, Y.; Milstein, D. A new mode of activation of CO2 by metal-
ligand cooperation with reversible C-C and M-O bond formation at
ambient temperature. Chem. - Eur. J. 2012, 18, 9194−9197.
(33) Johnson, M. C.; Rogers, D.; Kaminsky, W.; Cossairt, B. M. CO2
Hydrogenation Catalyzed by a Ruthenium Protic N-Heterocyclic
Carbene Complex. Inorg. Chem. 2021, 60, 5996−6003.
(34) Feller, M.; Gellrich, U.; Anaby, A.; Diskin-Posner, Y.; Milstein,
D. Reductive Cleavage of CO2 by Metal−Ligand-Cooperation
Mediated by an Iridium Pincer Complex. J. Am. Chem. Soc. 2016,
138, 6445−6454.
(35) Vogt, M.; Nerush, A.; Diskin-Posner, Y.; Ben-David, Y.;
Milstein, D. Reversible CO2 binding triggered by metal−ligand
cooperation in a rhenium(I) PNP pincer-type complex and the
reaction with dihydrogen. Chem. Sci. 2014, 5, 2043−2051.
(36) Vogt, M.; Rivada-Wheelaghan, O.; Iron, M. A.; Leitus, G.;
Diskin-Posner, Y.; Shimon, L. J. W.; Ben-David, Y.; Milstein, D.
Anionic Nickel(II) Complexes with Doubly Deprotonated PNP
Pincer-Type Ligands and Their Reactivity toward CO2. Organo-
metallics 2013, 32, 300−308.
(37) Huff, C. A.; Kampf, J. W.; Sanford, M. S. Reversible carbon−
carbon bond formation between carbonyl compounds and a
ruthenium pincer complex. Chem. Commun. 2013, 49, 7147−7149.
(38) Mukherjee, A.; Nerush, A.; Leitus, G.; Shimon, L. J. W.; Ben
David, Y.; Espinosa Jalapa, N. A.; Milstein, D. Manganese-Catalyzed
Environmentally Benign Dehydrogenative Coupling of Alcohols and
Amines to Form Aldimines and H2: A Catalytic and Mechanistic
Study. J. Am. Chem. Soc. 2016, 138, 4298−4301.
(39) Montag, M.; Zhang, J.; Milstein, D. Aldehyde Binding through
Reversible C−C Coupling with the Pincer Ligand upon Alcohol
Dehydrogenation by a PNP−Ruthenium Catalyst. J. Am. Chem. Soc.
2012, 134, 10325−10328.
(40) Filonenko, G. A.; Hensen, E. J. M.; Pidko, E. A. Mechanism of
CO2 hydrogenation to formates by homogeneous Ru-PNP pincer
catalyst: from a theoretical description to performance optimization.
Catal. Sci. Technol. 2014, 4, 3474−3485.
(58) Gunanathan, C.; Milstein, D. Bond Activation and Catalysis by
Ruthenium Pincer Complexes. Chem. Rev. 2014, 114, 12024−12087.
(59) Khusnutdinova, J. R.; Milstein, D. Metal−Ligand Cooperation.
Angew. Chem., Int. Ed. 2015, 54, 12236−12273.
(41) Bertini, F.; Glatz, M.; Gorgas, N.; Stoger, B.; Peruzzini, M.;
Veiros, L. F.; Kirchner, K.; Gonsalvi, L. Carbon dioxide hydrogenation
catalysed by well-defined Mn(I) PNP pincer hydride complexes.
Chem. Sci. 2017, 8, 5024−5029.
(42) Rønne, M. H.; Cho, D.; Madsen, M. R.; Jakobsen, J. B.; Eom,
́
S.; Escoudé, E.; Hammershøj, H. C. D.; Nielsen, D. U.; Pedersen, S.
U.; Baik, M.-H.; Skrydstrup, T.; Daasbjerg, K. Ligand-Controlled
Product Selectivity in Electrochemical Carbon Dioxide Reduction
Using Manganese Bipyridine Catalysts. J. Am. Chem. Soc. 2020, 142,
4265−4275.
(43) Léval, A.; Agapova, A.; Steinlechner, C.; Alberico, E.; Junge, H.;
Beller, M. Hydrogen production from formic acid catalyzed by a
phosphine free manganese complex: investigation and mechanistic
insights. Green Chem. 2020, 22, 913−920.
(44) Tondreau, A. M.; Boncella, J. M. 1,2-Addition of Formic or
Oxalic Acid to −N{CH2CH2(PiPr2)}2-Supported Mn(I) Dicarbonyl
8369
https://doi.org/10.1021/acscatal.1c01709
ACS Catal. 2021, 11, 8358−8369