CO2Hydrogenation Catalyzed by a Ruthenium Protic N-Heterocyclic Carbene Complex

M. Cecilia Johnson, Dylan Rogers, Werner Kaminsky, and Brandi M. Cossairt*

Article

CO Hydrogenation Catalyzed by a Ruthenium Protic N‑Heterocyclic

Carbene Complex

Cite This: Inorg. Chem. 2021, 60, 5996 −6003

Read Online

ACCESS

Metrics & More

Article Recommendations

sı Supporting Information

ABSTRACT: We describe the hydrogenation of CO to formate

catalyzed by a Ru(II) bis(protic N-heterocyclic carbene, p-NHC)

phosphine complex [Ru(bpy)(MeCN)(P (p-NHC) )](PF )

6 2

(

1). Under catalytic conditions (20 μmol catalyst, 20 bar CO₂,

0 bar H , 5 mL THF, 140 °C, 16 h), the activity of 1 is limited

only by the amount of K PO present in the reaction, yielding a

nearly 1:1 ratio of turnover number (TON) to equivalents of

K PO (relative to 1), with the highest TON = 8040. Additionally,

analysis of the reaction solution post-run reveals the catalyst intact

with no free ligand observed. Stoichiometric studies, including

examination of unique carbamate and hydride complexes as relevant intermediates, were carried out to probe the operative

mechanism and understand the importance of metal−ligand cooperativity in this system.

INTRODUCTION

■

Protic N-heterocyclic carbene (p-NHC) ligated organometallic

complexes are of interest in small molecule activation and

catalysis due to the potential for metal−ligand cooperativity

facilitated by the proximity of the NH wingtip to the transition

metal center. Additionally, the strong σ donor character of p-

NHCs makes them useful supporting ligands for reduction

chemistry, particularly in hydrogenative catalytic transforma-

Figure 1. Catalyst 1 and the CO adduct 2.

tions. While NHCs are a thoroughly studied ligand platform,

the catalytic properties of the protic analogues have been less

donating character and tris-chelating nature of the bis-carbene

thoroughly examined. To the best of our knowledge, there are

phosphine platform should endow this system with high

fewer than ten catalytic studies using p-NHCs, six of which are

thermal stability resulting in prolonged catalyst lifetimes.

focused on hydrogenation or dehydrogenation catalysis. Three

Previous work with this ligand platform examined stoichio-

of these involve the direct hydrogenation of imines, alkenes,

metric reactivity with CO , formate, and bicarbonate. Each of

and carbonyls, one is concerned with acceptorless dehydro-

these reagents ultimately yields a carbamate complex (2), with

the oxygen bound to the ruthenium metal center and the

carbon bound through the deprotonated NH-wingtip. When

genation and dehydrogenative coupling, and two focus on

transfer hydrogenation−dehydrogenation of carbonyls and

alkenes. In direct hydrogenation, it has been proposed that p-

formate is treated with the starting [Ru(bpy)(MeCN)(P (p-

NHCs are involved in the initial activation of H between the

NHC) )](PF ) complex, 1, the reaction proceeds through a

6 2

metal center and the deprotonated NH-wingtip, used as a

dimeric formate-bridged intermediate before liberating H gas

molecular recognition unit for substrates via hydrogen

and yielding 2.

bonding, or simply a spectator. In these three cases, high

yields are achieved, though relatively high catalyst loadings

This observed stoichiometric activation of CO , as well as

the tripodal geometry and strong σ donor character of the

^lⁱ^g^aⁿ^d^,^c^o^m^bⁱⁿ^e^m^aⁿ^y^c^o^m^m^oⁿ^f^e^a^t^u^r^e^s^o^f^s^u^c^c^e^s^s^f^u^l^C^O2

result in the highest turnover number (TON) being less than

The catalytic system studied herein was designed with two

principal concepts in mind. First, two p-NHCs are installed cis

Received: February 9, 2021

Published: March 29, 2021

to one another with nearly parallel NH-wingtips (Figure 1,

[

Ru(bpy)(L)(P (p-NHC) )](PF ) complex 1). The aim

6 2

here is to increase the number of protons proximal to the site

where hydrogenation occurs or to potentially act as a second

site for metal−ligand cooperativity. Additionally, the strong σ-

2021 American Chemical Society

996

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

hydrogenation catalysts. Leitner and co-workers reported the

Table 1. Screening Additives

ﬁrst homogeneous catalyzed direct hydrogenation of CO to

Cat

Additive (100 equiv)

TON

m e t h a n o l u s i n g

r u t h e n i u m 1 , 1 , 1 - t r i s -

^K^P^O4

(

diphenylphosphinomethyl)ethane (“triphos”) precatalyst.

DBU

The facial coordination of the triphos ligand increases the

favorability of hydride transfer to the carboxylate units, and the

high thermal stability of the ruthenium triphos complex is

necessary for catalyst longevity under the harsh conditions

required for the reaction to proceed. Additionally, several

^K^P^F6

K PO + H O

Conditions: 20 μmol catalyst, THF (5 mL), 16 h, 140 °C, CO :H

2 2

20:60 bar). Equiv relative to catalyst. Work up: centrifuge.

Experiment run in triplicate. 100 equiv of K PO and 100 equiv

(

research groups have reported CO activation between the

of H O.

metal center and the ligand framework, particularly with pincer

−14

ligands.

In these systems, CO activation via metal−ligand

cooperativity is reversible, and the so-formed CO adducts are

with CO and literature precedent, so basic conditions, as well

as Lewis acids, were used. The latter has been shown to

stabilize relevant activated CO and formate intermediates,

thus facilitating catalysis (Table 1).

With 100 equiv of K PO , a TON of 92 was observed. When

KPF was used, the TON was suppressed to 4, indicating the

important role of base for catalysis to proceed at an appreciable

rate. When 100 equiv of water was added in addition to 100

equiv of K PO , catalysis was suppressed to a TON of 37,

generally regarded as oﬀ-cycle catalytic intermediates. By

examining CO hydrogenation with our ruthenium bis-p-NHC

phosphine complex (1), which combines tridentate facial

coordination and strong σ-donating character, we aim to

understand whether the dynamic proton binding and metal−

ligand cooperation aﬀorded by the p-NHCs will be a beneﬁt

for eﬃcient catalysis.

16,17

In this work, we evaluated complex 1 as a CO hydro-

genation catalyst. We observed a linear dependence between

catalyst turnover and equivalents of base (relative to catalyst),

with turnovers of formate as high as 8040 measured using 0.2

mM of catalyst. Identiﬁcation of catalyst speciation following

catalytic runs supports the hypothesis that the tridentate and

strongly σ-donating p-NHC phosphine ligand creates a robust

and long-lived catalyst, and that catalysis is limited solely by

the consumption of K PO . Mechanistic investigations were

likely due to the equilibrium between K PO and K PO H.

The presence of water would shift the equilibrium toward

K PO H, thereby inhibiting the catalysis if it requires the

deprotonation of the ligand or an intermediate. Additionally,

the requirement for basic conditions was expected to be

necessary, since formic acid production is thermodynamically

uphill from H and CO , but the reaction becomes favorable in

the presence of base to yield formate. Amine bases also show

decreased TONs compared to K PO . This allows us to

carried out and the reactivity of 1 with H was investigated,

which in the presence of base forms a unique hydride bridged

dimer (3) with two symmetric hydrogen bonding pairs, each

involving one p-NHC from opposing monomer units. Both 2

and 3 were investigated as possible catalytic intermediates,

revealing details about the importance of metal−ligand

hypothesize that the Brønsted basic and Lewis acidic

properties of K PO are complementary in the operative

catalytic cycle.

When K PO was used as an additive in variable equivalents

relative to 1, nearly stoichiometric equivalents of product to

base were observed (Table 2 and Figure 2). When 9900 equiv

of K PO was added, 8040 equiv of formate was produced. A

cooperativity in the hydrogenation of CO .

RESULTS AND DISCUSSION

Compound 1 was synthesized using a streamlined procedure

■

control experiment revealed a slow background reaction with

K PO in the absence of catalyst, but by comparing molar

involving the direct reaction of [Ru(bpy)(C H )OTf][OTf]

equivalents of product formed to base, we observe a 13-fold

increase with catalyst (Table 2). This well-documented

^wⁱ^t^h^P⁽¹^-^N^-^e^t^h^y^l^-⁵^,⁶^-^dⁱ^m^e^t^h^y^l^b^eⁿ^zⁱ^mⁱ^d^a^z^o^l^e⁾2 (L) in N-

methyl-2-pyrrolidone (NMP, Scheme 1). The reaction

Table 2. Screening Equivalents of K PO and Control

Experiment

Scheme 1. One-Pot Synthesis of 1

equiv K PO (mmol

TON (mmol

Prod.)

Molar

[Cat]

K PO )

Ratio

(

0.70

0.92

0.98

0.81

0.07

0.2)

100

2.0)

980

19.8)

9900

9.9)

(0.14)

(1.83)

965

(19.49)

8040

(8.04)

(

d,e

.2 mM

(

proceeds ﬁrst through the coordination of the phosphine to

the ruthenium, followed by the thermally driven tautomeriza-

tion of the benzimidazoles to the desired p-NHC complex in a

(

2.0)

(0.14)

9% isolated yield. In order to probe the catalytic activity of

Conditions: catalyst 1, THF (5 mL), 140 °C, 16 h, 20 bar CO , 60

bar H . THF (7 mL). 500 μL of a 2.0 mM catalyst solution in THF.

Experiments were run in triplicate. At 0.2 mM the reaction rate

decreased and was run for 4 days. The molar ratio of formate product

for the hydrogenation of CO to formate, a series of additives

were screened at a ratio of 100 equiv relative to catalyst 1 (4

mM in THF) using 20:60 bar CO :H at 140 °C for 16 h, as

summarized in Table 1. The additives screened were tailored

to K PO additive was used as a metric to compare the role of the

based on our knowledge of the stoichiometric reactivity of 1

catalyst versus the control run in the absence of catalyst.

997

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

the phosphine of L. It is characterized by a ³¹P NMR shift of

4 ppm in CD CN (Figure S13B ). These species are depicted

in Figure 3 .

Figure 2. Linear relationship between equivalents of K PO and TON

with catalyst 1.

background reaction is attributed to formate being thermody-

namically downhill from CO and H , and under the forcing

conditions used here, the reaction is able to proceed at a

8,19

measurable rate.

Screening for higher TONs was

prohibited by the physical constraints of the reactors and the

insoluble nature of K PO and the potassium formate product.

Because the TONs correlate with the equivalents of K PO

present, we sought further evidence that the TON was being

limited solely by the consumption of K PO . Examination of

the change in pressure over the course of the reaction shows a

steady decrease in pressure corresponding to the consumption

of CO and H . After a period of time, for example, after 5 h

with 1000 equiv K PO , the change in pressure reaches a

Figure 3. (A) Major catalyst speciation products following catalysis.

plateau, indicating an abrupt end to the reaction at

approximately 1000 TON (Figure S8). This observation

suggests that the change in pressure can be used to estimate

turnover frequencies, which in the case of 1000 equiv of K PO

(B) Catalyst speciation observed under conditions with only CO or

only H₂.

When 1 is treated under standard catalytic conditions in

−

is 193 h . In addition, Na PO was screened as an additive at

THF with 100 equiv of K PO and the vessel is charged with

000 equiv relative to catalyst, and the reaction proceeds at a

only 20 bar CO , the P NMR signals observed are at 45 ppm

signiﬁcantly slower rate than in the presence of 1000 equiv

and 50, 52, and 54 ppm, the latter of which are likely due to

diﬀerent binding environments at the open coordination site of

K PO , with a TON consistent with what was seen for K PO

as assessed by the pressure change and NMR spectroscopy

1 (Figure S14 ). When this is repeated using only 60 bar H ,

the peaks observed in the P NMR are at 52 and 25 ppm,

(Table S2 ).

In order to probe the fate of 1, the reaction mixtures were

analyzed following catalytic runs and workup by centrifugation

of the reaction mixture to separate the insoluble (phosphate

conﬁ rming that the peak at 25 ppm is due to reactivity with the

H (Figure S15 ). Notably in all of these experiments, we do

not observe any free ligand or intractable solids that would

indicate ruthenium decomposition, supporting the claim of the

robust nature of 1 as a catalyst.

Deprotonation of the NH-wingtips is expected to play a

central role in catalysis through metal ligand cooperation. Two

pathways are plausible following the initial deprotonation:

either reaction with hydrogen to form a hydride complex (3)

and formate) components, which were then dissolved in D O

after the organics were decanted away. In the aqueous portion,

we routinely observed an orange precipitate settle out of

solution. This orange precipitate was assigned as 2 via NMR

spectroscopy, with a ³¹P NMR resonance at 45 ppm and a H

NMR spectrum that overlays exactly with independently

synthesized 2 (Figures S10 and S11). This assignment was

also conﬁrmed by X-ray crystallography (Figure S12). When

the organic portion was analyzed, two peaks were observed in

or reaction with CO to form the carbamate complex (2)

(Figure 4). These pathways have been investigated and will be

discussed in detail. In the p-NHC literature, there are a

number of examples of hydrogen being split between the metal

the P NMR spectrum at 25 and 50 ppm in CD Cl . We

,20

propose that the peak at 50 ppm corresponds to the starting

catalyst 1 with the open coordination site either bound by

solvent (THF) or formate (Figure S13A). We have previously

reported the independently synthesized formate-bridged dimer

center and a deprotonated NH-wingtip.

In our system, we

have previously reported the facile formation of 2 via activation

of CO between the metal center and the deprotonated NH-

wingtip, reminiscent of CO activation through metal ligand

9−13

with a P NMR shift of 52 ppm in CD Cl (Figure S13B).

cooperativity in pincer ligand chemistry.

We hypothesize that the peak at 25 ppm corresponds to a

complex with the carbenes no longer coordinated to the metal

center, Ru-PR , possibly due to non-innocent reactivity of H

When 13 equiv of K PO was added to a solution of

complex 1 in THF-d and stirred at room temperature for 1 h,

there was a complete disappearance of peaks corresponding to

the NH-wingtips, a shifting in the methyl region, and a shifting

with the p-NHCs, yielding the uncoordinated benzimidazole

arms (Figure S13A). The Ru-PR complex was previously

of the aromatic peaks. The P NMR spectrum showed a shift

from 54 to 48 ppm. Together, these results indicate complete

deprotonation of the NH-wingtips to form the doubly

isolated en route to the complete metalation of the ligand via

thermally driven tautomerization and is bound solely through

998

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

C labeled CO incorporated (Figure 5 and Figure S19 ). This

coupling constant is consistent with that of H CO Na taken

independently in D O.

Figure 5. Treatment of ¹³C labeled 2 with H gas with and without

K PO .

The observation that K PO is required for the reactivity of

with H reveals that the role of the K PO is not limited to

the deprotonation of the NH-wingtips. Once CO is bound,

the potassium could be required as a Lewis acid to (I) aid in

the dissociation of the charge-balanced potassium formate

product and/or (II) stabilize the activated CO and weaken or

break the ruthenium−oxygen bond and provide an open site

16,17

for hydrogen activation.

Alternatively, the phosphate may

be necessary to help with the activation of the H . Previous

reports have illustrated the deprotonation of the metal-bound

19,21

hydrogen directly,

or speciﬁcally in the p-NHC literature,

activation of molecular hydrogen between the metal center and

,5,20

the deprotonated NH-wingtip has been observed.

To probe the dissociation of CO₂, C labeled 2 was heated

Figure 4. Proposed catalytic pathways for the hydrogenation of CO₂

to formate by 1.

in THF-d to 105 °C and monitored by NMR spectroscopy.

After heating, a small amount of free CO at 125.8 ppm was

observed in the C NMR spectrum in addition to the peak

corresponding to 2 at 156.4 ppm, conﬁrming gradual

deprotonated species (1a). Upon heating at 100 °C for 1.5 h,

dissociation of CO at elevated temperature, though the

the P NMR spectrum shows the persistence of the peak

corresponding to the deprotonated species, suggesting that 1a

remains stable in THF with excess K PO .

equilibrium lies toward coordinated CO (Figure S20 ). When

C labeled 2 was heated only to 90 °C, no free CO was

observed by C NMR spectroscopy, indicating that the CO is

When 2 was treated in the same manner as above, little

only labile above 90 °C (Figure S22 ). Notably, when the

standard catalytic conditions were changed to decrease the

temperature from 140 to 65 °C, we saw a corresponding

decrease in TON from 92 to 4 (Scheme 2). When the run at

change was observed. With the addition of K PO there was a

disappearance of the H NMR peak corresponding to the NH-

wingtip, but no change in the remainder of the spectrum,

suggesting that there is an equilibrium interaction occurring

between the phosphate and the p-NHC. Heating at 100 °C for

.5 h also shows no signiﬁcant transformation, which is

Scheme 2. Temperature Dependence for Catalytic CO

Hydrogenation

corroborated by P NMR spectroscopy. This suggests that the

pK of the phosphate conjugate acid is close to that of the NH-

wingtip in 2.

When a sample of 2 with C labeled CO incorporated was

dissolved in THF-d and pressurized with 8 bar H , no reaction

was observed, even after heating to 100 °C for 24 h (Figure

S16 ). When this was repeated in the presence of excess K PO ,

a reaction occurred at elevated temperatures. Tracking by

NMR spectroscopy, we began to see a new peak form at 32

ppm after 1 h at 100 °C. After an extended period of time at

room temperature, we see the peak at 32 ppm persist, a gradual

loss of signal at 45 ppm (2) and the appearance of a new peak

at 51 ppm, likely corresponding to complex 1 (Figure S17).

65 °C was analyzed for catalyst speciation, a signiﬁcant amount

of 2 was identiﬁed by the characteristic P NMR peak at 45

ppm. Together, these data support the hypothesis that high

temperature is required for labilizing CO in 2, which is likely

important for catalysis to proceed. Attempts at methylation of

Additionally, by H NMR spectroscopy we see the peaks

the NH-wingtip to block the coordination of CO and directly

corresponding to the methyl groups on the benzimidazole shift

validate the role of metal ligand cooperativity in the CO₂

activation pathway observed here have been made; however,

the synthesis has so far been unsuccessful. It should be noted

that when independently synthesized 2 is used as the

precatalyst, catalytic activity is maintained (Table 3).

Figure S18 ). Simultaneously, a solid precipitated from the

solution. The precipitate was isolated and evaluated by H

NMR spectroscopy in D O. A doublet appears at 8.45 ppm

with a J = 194 Hz, which is evidence of formate with the

999

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

Table 3. Screening Catalyst Precursors

in proteins has been studied by H NMR spectroscopy, and

characteristic chemical shifts are associated with diﬀerent types

of hydrogen bonding. It is observed that symmetrical hydrogen

bonds, referred to as single-well hydrogen bonds, have

Catalyst

equiv of K PO

TON

100

1000

100

965

23,24

characteristic H NMR shifts of 20−22 ppm.

Single-well

hydrogen bonds are the strongest type of hydrogen bonds, and

the proton lies linearly and equidistant between the

heteroatoms. In strong hydrogen bonding, the bond between

hydrogen and the heteroatom is longer relative to the normal

covalent length and shorter than the normal hydrogen bonding

length. The typical heteroatom separation for N−HN in a

1000

981

^C^oⁿ^dⁱ^tⁱ^oⁿ^s^:²⁰^μ^m^o^l^c^a^t^a^l^y^s^t^,^T^H^F⁽⁵^m^L⁾^,¹⁶^h^,¹⁴⁰^°^C^,^C^O^:^H2

(

20:60 bar).

Based on these results, 2 has been shown to be catalytically

relevant. Formate has been observed in situ via NMR

spectroscopy following reactivity 2 with H in the presence

of K PO , indicating that the CO in 2 is capable of being

single-well hydrogen bond is 2.6 to 2.7 Å. In the case of 3,

the distance between both sets of nitrogen atoms that are

oriented toward one another is 2.7 Å, which is consistent with

the literature and further supports the claim of two protons

hydrogen bonded in these positions. This leads to an

incorporated into the product. The coordinated CO becomes

labile around 100 °C, which is likely a contributing factor for

the high temperatures required for catalytic activity. We can

conclude that the formation of 2 does not inhibit catalysis

under our standard conditions. Another possible catalytic route

is through reactivity ﬁrst with H . The insertion of CO into a

environment where the proton is uniquely deshielded.

Complex 3 can also be synthesized as a minor product

through reaction with NaH, as has been demonstrated in a

related system containing one p-NHC ligand. When NaH is

added incrementally to 1, the evolution of hydrogen gas is ﬁrst

observed by the appearance of a singlet at 4.5 ppm in the H

NMR spectrum in THF-d , along with a decrease in intensity

for 1 equiv NaH) and eventual disappearance (2 equiv) of the

NH-wingtip peak. The presence of 3 is ﬁrst indicated by the

peak at 20.1 ppm in the H NMR with the addition of 1 equiv

of NaH. With the addition of the second equivalent of NaH,

the hydride triplet at −9.3 ppm begins to grow in, and with the

addition of a third equivalent, the peaks at 20.1 and −9.3 ppm

persist and the P NMR peak at 38 ppm indicates that 3 is one

of three major products.

metal−hydride bond is a common pathway for the hydro-

genation of CO .

,20

Following literature precedent,

the hydride complex, 3,

was targeted by deprotonation of 1 with either stoichiometric

KO Bu or excess K PO , followed by reaction with an

atmosphere of hydrogen gas. Over the course of 10 days, the

reaction between 1a and H reached completion (Scheme 3).

A triplet in the H NMR spectrum at −9.3 ppm (with a J

coupling constant of 37 Hz) indicates the formation of a

(

hydride trans to two equivalent phosphine ligands. This is

further conﬁrmed by H{ P} NMR spectroscopy, wherein the

resonance at −9.3 ppm is observed as a singlet. This suggests

the formation of a hydride-bridged dimer. Additionally, a peak

Previous reports of reactivity between p-NHC complexes

with both H gas and NaH involve systems containing a

bidentate phosphine p-NHC ligand with Cp or Cp*

is present at 20.2 ppm in the H NMR spectrum. This strongly

,5,20

supporting ligands.

In these reports, the hydride resonance

deshielded peak is suggestive of a proton hydrogen bonded

between two nitrogen atoms of the p-NHCs (vide infra). The

four distinct peaks corresponding to the methyl groups of the

benzimidazole backbone indicate that the ligand arms are in

two distinct environments. This assignment has been

conﬁrmed by X-ray crystallography (Figure 6). In addition to

appears as a doublet with J ranging from 28.4 to 36.5 Hz,

indicating a terminal hydride trans to a phosphorus atom.

Additionally, the NH-wingtips in these systems are observed in

their typical H NMR regions. The diﬀerences in reactivity

observed in the present system, primarily the formation of a

bridging hydride and the symmetrical hydrogen bonding, may

be in part explained by the lack of steric congestion at the

metal center. The bipyridine ligand leads to a relatively ﬂat and

open face in 1, which enables the formation of the dimer 3. In

addition, the dimer is further stabilized by π−π stacking

between the two center benzimidazole rings, between the

bipyridine rings, and between the phenyl and bipyridine rings,

as is evident from the X-ray crystallography (Figures S6 and

S7 ). These stabilizing forces may contribute to the diﬀerences

in reactivity observed between the system studied presently

−

conﬁrming the hydride-centered dimer, only one PF₆

counterion was observed, indicating an overall charge of +1

for 3. This is consistent with two of the four p-NHCs being

deprotonated and is supported by the integration observed in

the H NMR for the hydride peak at −9.3 ppm (1 hydrogen)

and the peak at 20.2 ppm (2 hydrogens).

Chemical shifts around 20 ppm in diamagnetic H NMR

spectra are not common. However, it is known within the p-

NHC literature that the addition of a hydrogen bonding

acceptor causes the resonance corresponding to the NH-

,22

4,5

wingtip to shift downﬁeld.

Furthermore, hydrogen bonding

and related literature examples.

Scheme 3. Synthesis of 3

000

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

Figure 6. (A) Molecular structure of 3 with thermal ellipsoids shown at 50% probability. Solvent of crystallization (THF), PF₆⁻counterion, and all

hydrogen atoms except the metal-bound hydride and the ligand N−H protons are removed for clarity. Electron density was observed between the

associated nitrogen atoms and the disorder model was used to place the N−H protons in ideal geometry at half-site occupancy. (B) H NMR

(

THF-d , 300 MHz, 298 K) spectrum of 3, with integrations for the hydride peak at −9.3 ppm (J = 37 Hz) and the N−H protons at 20.2 ppm.

When 1 was exposed to hydrogen gas in the absence of base,

turnover. When 2 is treated with H gas in the presence of

no reaction occurred after an extended period of time (4 days)

K PO , potassium formate is formed with incorporation of the

at room temperature. Heating at 60 °C for 2 h and 20 min

coordinated CO into the product. Deprotonation of the NH-

−

showed partial decomposition of the PF counterion to

wingtips is also required for 1 to react with H gas, forming the

−

PO F , evident by P NMR spectroscopy by the presence of

hydride-bridged dimer 3. When 3 is treated with CO , there is

a triplet at −13 ppm ( J = 971 Hz), which is consistent with

evidence by H NMR spectroscopy of the formation of the

the literature. Upon cooling, this triplet disappeared, and the

proton NMR is restored to its previous state, indicating a

reversible reaction.

formate coordinated complex, in addition to the formation of

. This is consistent with the known reactivity of the formate

complex, which releases H gas to form 2, and thus indicates

^T^h^e^r^e^a^c^tⁱ^oⁿ^o^f³^aⁿ^d^C^O2 ﬁrst leads to the full

disappearance of the peak at 20.1 ppm, indicating that the

symmetrical hydrogen bond has been disrupted. Additionally,

there is a shifting of the hydride peak from −9.3 ppm to −9.7

ppm and a gradual decline in peak intensity, suggesting the

consumption of the hydride (Figures S23 and S24). After 4 h

at room temperature, there is a small peak growing in at 11.5

ppm, which is consistent with a typical NH-wingtip signal.

After 24 h, the signal for the peak at 11.5 ppm has increased

and there is a small peak at 13.2 ppm, which is consistent with

the formate bound complex previously identiﬁed by our

this is a reversible step in the catalytic cycle. Based on the

stoichiometric studies, it is observed that 2 forms far more

readily than 3, with the latter transformation reaching

completion over the course of 7 days. In addition, the

prevalence of 2 as the product in a number of stoichiometric

transformations, as well as its isolation as a dominant product

post-catalysis, supports its existence as a low energy resting

state. Based on these observations, we propose that catalysis

likely proceeds through 2 as the dominant pathway; however,

both pathways are feasible especially under the catalytic

conditions. Studying the catalytic speciation following the

conclusion of the reaction shows no signiﬁcant catalyst

degradation, with all the isolated species on or readily

accessible from the catalytic cycle and no evidence of free

ligand or ruthenium decomposition. This, along with the linear

dependence of TON on the molar equivalence of K PO ,

group. The formate bound complex has been observed in the

solid state as a formate bridged dimer. This suggests that one

possible catalytic pathway involves the insertion of CO into

the Ru−H−Ru bond. The reaction of 3 with CO led to the

formation of a precipitate. The precipitate was dissolved in

DMSO-d and characterized by NMR spectroscopy as 2, which

is consistent with our previous study of the formate dimer

suggests an incredibly robust and long-lived catalyst.

complex which is capable of liberating hydrogen gas to yield

. When 3 was isolated and used as the precatalyst, it showed

CONCLUSIONS

comparable catalytic activity to 1 (Table 3). As such, both 2

and 3 are viable precatalysts. When catalytic speciation was

evaluated following the run with 3, peaks at 38 and 24 ppm in

■

Complex 1 has been demonstrated to be a robust catalyst for

the hydrogenation of CO to formate, limited only by the

the P NMR are revealed, consistent with previously identiﬁed

catalyst speciation (Figure 3).

consumption of stoichiometric K PO , achieving a TON of

040. In our p-NHC system, stoichiometric base also

Given the insights gained from the stoichiometric studies, it

is proposed that the reaction proceeds ﬁrst through the

deprotonation of the NH-wingtips, converting 1 to 1a (Figure

contributes to the formation of 2 and 3, both of which have

been shown to be relevant to the catalytic cycle. The TONs are

only limited by the consumption of K PO , as is evident by the

). K PO is also required for the extrusion of formate from

3 4

lack of catalyst decomposition observed by studying the

catalyst speciation, as well as the linear dependence of TON on

the catalyst, likely due to the Lewis acidic properties of K .

Complex 2 is readily formed from the addition of CO to 1a

the molar equivalence of K PO . It is proposed that the strong

and is isolated from the reaction mixture following catalysis

indicating that it is likely a resting state on the catalytic cycle or

an easily accessible oﬀ cycle species. Complex 2 becomes labile

around 100 °C, which is likely a contributing factor for the

required high temperatures to achieve appreciable catalytic

σ donation of p-NHCs lends to the stability of the catalyst.

The hydrogenation of CO to formate likely proceeds though

2 as the dominant pathway, though under the catalytic

conditions both proposed pathways are likely operative.

001

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

Article

4) Chang, W.; Gong, X.; Wang, S.; Xiao, L.-P.; Song, G.

ASSOCIATED CONTENT

sı Supporting Information

The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acs.inorgchem.1c00417.

■

Acceptorless Dehydrogenation and Dehydrogenative Coupling of

Alcohols Catalysed by Protic NHC Ruthenium Complexes. Org.

Biomol. Chem. 2017, 15 (16), 3466−3471.

(5) Miranda-Soto, V.; Grotjahn, D. B.; Cooksy, A. L.; Golen, J. A.;

Moore, C. E.; Rheingold, A. L. A Labile and Catalytically Active

Imidazol-2-Yl Fragment System. Angew. Chem., Int. Ed. 2011, 50 (3),

Experimental Details, NMR spectra, and crystal structure

data (PDF )

631−635.

Accession Codes

(6) Gomez-Lopez, J. L.; Chávez, D.; Parra-Hake, M.; Royappa, A.

T.; Rheingold, A. L.; Grotjahn, D. B.; Miranda-Soto, V. Synthesis and

Reactivity of Bis(Protic N-Heterocyclic Carbene)Iridium(III) Com-

plexes. Organometallics 2016, 35 (18), 3148−3153.

CCDC 2060318 −2060319 contain the supplementary crys-

tallographic data for this paper. These data can be obtained

free of charge via www.ccdc.cam.ac.uk/data_request/cif , or by

emailing data_request@ccdc.cam.ac.uk , or by contacting The

Cambridge Crystallographic Data Centre, 12 Union Road,

Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

7) Norris, M. R.; Flowers, S. E.; Mathews, A. M.; Cossairt, B. M. H

Production Mediated by CO ₂ via Initial Reduction to Formate.

Organometallics 2016, 35 (17), 2778−2781.

(8) Wesselbaum, S.; Moha, V.; Meuresch, M.; Brosinski, S.; Thenert,

K. M.; Kothe, J.; Stein vom, T.; 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

AUTHOR INFORMATION

Corresponding Author

Brandi M. Cossairt − Department of Chemistry, University of

Washington, Seattle, Washington 98195-1700, United

States; orcid.org/0000-0002-9891-3259;

Email: cossairt@uw.edu

■

(

1), 693−704.

(

9) 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 (30), 9194−9197.

Authors

(10) Vogt, M.; Nerush, A.; Diskin-Posner, Y.; Ben-David, Y.;

M. Cecilia Johnson − Department of Chemistry, University of

Washington, Seattle, Washington 98195-1700, United States

Dylan Rogers − Department of Chemistry, University of

Washington, Seattle, Washington 98195-1700, United States

Werner Kaminsky − Department of Chemistry, University of

Washington, Seattle, Washington 98195-1700, United

States; orcid.org/0000-0002-9100-4909

Milstein, D. Reversible CO ₂ Binding Triggered by Metal-Ligand

Cooperation in a Rhenium(I) PNP Pincer-Type Complex and the

Reaction with Dihydrogen. Chem. Sci. 2014, 5 (5), 2043−2051.

(11) Zhang, Y.; MacIntosh, A. D.; Wong, J. L.; Bielinski, E. A.;

Williard, P. G.; Mercado, B. Q.; Hazari, N.; Bernskoetter, W. H. Iron

Catalyzed CO

Co-Catalysts. Chem. Sci. 2015, 6 (7), 4291−4299.

12) Huff, C. A.; Kampf, J. W.; Sanford, M. S. Role of a

Noninnocent Pincer Ligand in the Activation of CO at (PNN)Ru-

Hydrogenation to Formate Enhanced by Lewis Acid

Complete contact information is available at:

https://pubs.acs.org/10.1021/acs.inorgchem.1c00417

H)(CO). Organometallics 2012, 31 (13), 4643 −4645.

13) Huff, C. A.; Sanford, M. S. Catalytic CO Hydrogenation to

Notes

Formate by a Ruthenium Pincer Complex. ACS Catal. 2013, 3 (10),

The authors declare no competing ﬁnancial interest.

2412−2416.

(14) Gunanathan, C.; Milstein, D. Bond Activation and Catalysis by

Ruthenium Pincer Complexes. Chem. Rev. 2014, 114 (24), 12024−

2087.

15) Flowers, S. E.; Johnson, M. C.; Pitre, B. Z.; Cossairt, B. M.

Synthetic Routes to a Coordinatively Unsaturated Ruthenium

Complex Supported by a Tripodal, Protic Bis(N-Heterocyclic

Carbene) Phosphine Ligand. Dalton Trans. 2018 , 47 , 1276 −1283.

(16) Silvia, J. S.; Cummins, C. C. Binding, Release, and

Functionalization of CO2 at a Nucleophilic Oxo Anion Complex of

Titanium. Chem. Sci. 2011, 2 (8), 1474.

ACKNOWLEDGMENTS

■

(17) Heimann, J. E.; Bernskoetter, W. H.; Hazari, N. Understanding

We would like to thank the David and Lucile Packard

Foundation and the Camille and Henry Dreyfus Foundation

for funding, as well as the University of Washington Clean

Energy Institute for a Graduate Research Fellowship.

Analytical data were obtained from the CENTC Elemental

Analysis Facility at the University of Rochester. The authors

thank Dr. Rajan Paranji and Dr. Andrew Ritchhart for

assistance obtaining DOSY NMR spectra. In addition, we

thank Dr. Louise Guard, Dr. David Ung, Dr. Braden Zahora,

and Prof. Mike Heinekey for helpful discussions, and Dr.

Zuzana Culakova for assistance with the Parr reactors.

the Individual and Combined Effects of Solvent and Lewis Acid on

CO ₂ Insertion into a Metal Hydride. J. Am. Chem. Soc. 2019, 141

(

26), 10520−10529.

18) Jessop, P. G.; Ikariya, T.; Noyori, R. Homogeneous

Hydrogenation of Carbon Dioxide. Chem. Rev. 1995, 95 (2), 261−

72.

(19) Schmeier, T.; Dobereiner, G.; Crabtree, R.; Hazari, N.

REFERENCES

■

(

1) Aznarez, F.; Iglesias, M.; Hepp, A.; Veit, B.; Sanz Miguel, P. J.;

Secondary Coordination Sphere Interactions Facilitate the Insertion

Oro, L. A.; Jin, G.-X.; Hahn, F. E. Iridium(III) Complexes Bearing

Chelating Bis-NHC Ligands and Their Application in the Catalytic

Reduction of Imines: Iridium(III) Complexes Bearing Chelating Bis-

NHC Ligands and Their Application in the Catalytic Reduction of

Imines. Eur. J. Inorg. Chem. 2016, 2016 (28), 4598−4603.

Step in an Iridium(III) CO

2011, 133 (24), 9274−9277.

Reduction Catalyst. J. Am. Chem. Soc.

(20) Miranda-Soto, V.; Grotjahn, D. B.; DiPasquale, A. G.;

Rheingold, A. L. Imidazol-2-Yl Complexes of Cp *Ir as Bifunctional

Ambident Reactants. J. Am. Chem. Soc. 2008, 130 (40), 13200−

13201.

2) Meier, N.; Hahn, F. E.; Pape, T.; Siering, C.; Waldvogel, S. R.

Molecular Recognition Utilizing Complexes with NH,NR-Stabilized

Carbene Ligands. Eur. J. Inorg. Chem. 2007, 2007 (9), 1210−1214.

(21) Chinn, M. S.; Heinekey, D. M. Dihydrogen Complexes of

Ruthenium. 2. Kinetic and Thermodynamic Considerations Affecting

Product Distribution. J. Am. Chem. Soc. 1990, 112 (13), 5166 −5175.

(22) Hahn, F. E.; Naziruddin, A. R.; Hepp, A.; Pape, T. Synthesis,

Characterization, and H-Bonding Abilities of Ruthenium(II) Com-

(

3) Muhlen, C.; Linde, J.; Rakers, L.; Tan, T. T. Y.; Kampert, F.;

Glorius, F.; Hahn, F. E. Synthesis of Iron(0) Complexes Bearing

Protic NHC Ligands: Synthesis and Catalytic Activity. Organometallics

2019, 38 (12), 2417−2421.

002

Inorg. Chem. 2021, 60, 5996−6003

Inorganic Chemistry

plexes Bearing Bidentate NR,NH-Carbene/Phosphine Ligands ^†.

Article

Organometallics 2010, 29 (21), 5283 −5288.

(23) Ishikita, H.; Saito, K. Proton Transfer Reactions and Hydrogen-

Bond Networks in Protein Environments. J. R. Soc., Interface 2014, 11

91), 20130518.

24) Frey, P. A. Review: Strong Hydrogen Bonding in Molecules

(

and Enzymatic Complexes. Magn. Reson. Chem. 2001, 39 (S1), S190−

S198.

25) Christe, K.; Dixon, D.; Schrobilgen, G.; Wilson, W.

Tetrafluorophosphate Anion. J. Am. Chem. Soc. 1997, 119, 3918−392.

003