Catalytic CO2 hydrogenation to formate by a ruthenium pincer complex

Article abstract of DOI:10.1021/cs400609u

This paper reports the hydrogenation of carbon dioxide to formate catalyzed by the Ru pincer complex Ru(PNN)CO(H) (PNN = 6-(di-tert- butylphosphinomethylene)-2-(N,N-diethylaminomethyl)-1,6-dihydropyridine). Stoichiometric studies are presented that support the feasibility of the individual steps in a proposed catalytic cycle for this transformation. The influence of base and solvent on catalyst performance is explored. Overall, under optimized conditions (using diglyme as the solvent and potassium carbonate as the base) up to 23,000 turnovers of formate and a turnover frequency of up to 2,200 h^-1 can be achieved.

Full text of DOI:10.1021/cs400609u

Research Article
pubs.acs.org/acscatalysis
Catalytic CO Hydrogenation to Formate by a Ruthenium Pincer
2
Complex
Chelsea A. Huff and Melanie S. Sanford*
Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States
*
S Supporting Information
ABSTRACT: This paper reports the hydrogenation of carbon
dioxide to formate catalyzed by the Ru pincer complex
Ru(PNN)CO(H) (PNN = 6-(di-tert-butylphosphinomethy-
lene)-2-(N,N-diethylaminomethyl)-1,6-dihydropyridine). Stoi-
chiometric studies are presented that support the feasibility of
the individual steps in a proposed catalytic cycle for this
transformation. The influence of base and solvent on catalyst
performance is explored. Overall, under optimized conditions
(
using diglyme as the solvent and potassium carbonate as the base) up to 23,000 turnovers of formate and a turnover frequency
−1
of up to 2,200 h can be achieved.
KEYWORDS: hydrogenation, formic acid, homogeneous catalysis, pincer ligands, ruthenium
INTRODUCTION
Scheme 1. Hydrogenation of CO
Catalyst 1 and (b) Ir Catalyst 2
to Formate with (a) Ru
2
■
Formic acid (FA) is a commodity chemical that is widely used
in agriculture and in the textile and leather industries. The
1
current commercial production of FA involves carbonylation of
methanol to yield methyl formate, followed by hydrolysis of
1
methyl formate to produce FA and regenerate MeOH. An
attractive alternative route to FA would involve the direct
hydrogenation of CO . Indeed, significant effort has been
2
directed toward the development of catalysts for CO
2
2
−4
2−11
hydrogenation to FA,
formate salts,
forma-
Two of the most
active homogeneous CO hydrogenation catalysts reported to
5
−9,12−14
5−7,12−16
mides,
and formate esters.
2
17
date are Jessop’s Ru(PMe ) (Cl)(OAc) complex 1 and
3
4
N o z a k i ’ s I r ( P N P ) ( H ) 3 [ P N P
=
2 , 6 - b i s -
(
(
diisopropylphosphinomethylene)pyridine] pincer catalyst 2
1
8
Scheme 1). Both complexes provide formate from CO in
2
4
yields representing >10 catalytic turnovers and with turnover
frequencies in excess of 10 h .
Scheme 2. Cascade Catalytic Sequence for the
4
−1
15
Hydrogenation of CO to CH OH
2
3
We recently reported the catalytic reduction of CO to
2
methanol by a cascade sequence that begins with the
15
conversion of CO to FA (Scheme 2, step a). This step is
2
followed by Lewis-acid-catalyzed esterification to provide
methyl formate (step b), with subsequent ester hydrogenation
to liberate 2 equiv of CH OH (step c). We originally utilized
3
Jessop’s catalyst 1 for the CO hydrogenation step, in concert
2
with Milstein’s Ru pincer complex Ru(PNN)(CO)(H)
1
9−21
(3)
as the catalyst for ester hydrogenation. However,
since 3 closely resembles Nozaki’s Ir pincer catalyst 2, we
hypothesized that it might also be capable of catalyzing CO2
hydrogenation. Herein, we demonstrate that 3 is an effective
catalyst for the conversion of CO to formate in the presence of
2
an appropriate base. Furthermore, we provide insights into
divergent mechanistic pathways by which catalyst 3 reacts with
CO and H .
Received: July 26, 2013
Revised: September 5, 2013
Published: September 25, 2013
2
2
©
2013 American Chemical Society
2412
dx.doi.org/10.1021/cs400609u | ACS Catal. 2013, 3, 2412−2416

ACS Catalysis
RESULTS AND DISCUSSION
Research Article
■
Proposed Catalytic Cycle. By analogy to iridium catalyst
1
8
2,
a possible catalytic cycle for reducing CO to formate at
2
complex 3 (Scheme 3) would involve (i) heterolytic cleavage of
Scheme 3. Possible Catalytic Cycle for CO Hydrogenation
2
to Formate by Complex 3
Figure 1. ORTEP diagram (50% probability level) of the molecular
drawing of 5. The packing solvent (benzene) as well as all H atoms
(
other than the Ru−H and H−COO) have been omitted for clarity.
Selected bond lengths (Å) and angles (deg): Ru1−H1 = 1.45(3),
Ru1−N1 = 2.0983(18), Ru1−N2 = 2.2535(18), Ru1−O2 =
2
.2497(16), Ru1−C1 = 1.834(2), Ru1−P1 = 2.2626(5), O2−C2 =
.203(3), O3−C2 = 1.283(3); H1−Ru1−O2 = 169.2(10); P1−Ru1−
1
N2 = 158.83(5); N1−Ru1−O2 = 82.67(7), N2−Ru1−O2 = 83.88(6).
t
H to form ruthenium dihydride 4, (ii) insertion of CO to
Scheme 5. Deprotonation of 5 by KO Bu to Form 3 and
2
2
generate formate complex 5, and (iii) deprotonation of the
pincer ligand of 5 with concomitant release of formate to
HCOOK
22−24
complete the catalytic cycle.
Step i of this cycle is well
precedented, and has been studied in detail by Milstein and co-
2
0
workers. In contrast, the feasibility of steps ii and iii has not
yet been established for this ruthenium system.
We began our investigation by treating a solution of 3 in
anisole-d with a 4: 1 mixture of H and CO (Scheme 4). This
8
2
2
of the proposed 3-catalyzed hydrogenation of CO to formate.
To combine these steps to achieve catalysis, we initially treated
2
Scheme 4. Reaction of 3 with CO and H
2
2
3
with 10 bar CO and 30 bar H in the presence of 1200 equiv
2 2
t
of KO Bu in anisole at room temperature, conditions analogous
to those of the stoichiometric reactions in Schemes 4 and 5
(
Table 1, entry 1). The turnover number (TON) for this
reaction was determined based on the yield of formate after 4 h
as a preliminary estimate of catalyst reactivity (see Supporting
1
Information, Figure S4 for a representative H NMR
spectrum). Catalysis was sluggish under these initial conditions,
resulted in the conversion of 3 to formate complex 5 in 88%
NMR yield after 24 h at room temperature (see Supporting
Information, Figure S1), thus establishing the feasibility of steps
i and ii of the catalytic cycle proposed in Scheme 3. The
identity of 5 was confirmed by independent synthesis, and this
complex was characterized using standard one- and two-
a
Table 1. Hydrogenation of CO to Formate Catalyzed by 3
2
entry
solvent
temp (°C)
base
TON formate
25
dimensional NMR spectroscopic techniques as well as X-ray
t
1
2
3
4
anisole
25
120
120
120
120
120
120
120
120
120
200
200
KO Bu
7
270
2
6
crystallography (Figure 1). The CO adduct 6 (formed by the
t
2
anisole
KO Bu
27−29
direct reaction of 3 with CO , step ia of Scheme 3)
was
t
2
diglyme
diglyme
diglyme
diglyme
diglyme
diglyme
diglyme
diglyme
diglyme
diglyme
KO Bu
800
detected as a minor side product in this reaction (12% yield,
vide infra for further discussion).
K CO3
1,100
510
2
b
5
KOH
We next sought to identify conditions for promoting the final
step (iii) of the proposed catalytic cycle. Based on precedent by
6
7
8
K PO4
900
3
KHCO3
550
18
Nozaki with Ir catalyst 2, we hypothesized that a strong base
could deprotonate the pincer ligand of 5 and induce formate
release. Indeed, the treatment of a solution of 5 in anisole-d8
NEt3
<5
b
9
K CO3
660
2
b
b
b
,
c
10
11
12
K CO3
1,400
9,000
23,000
2
t
with 1 equiv of KO Bu resulted in quantitative formation of 3
K CO3
2
1
and potassium formate as determined by H NMR spectro-
,
d
K CO3
2
scopic analysis (Scheme 5 and Supporting Information, Figures
a
S2−S3).
Conditions: CO (10 bar), H (30 bar), 3 (0.554 μmol, 1 equiv), base
(0.6648 mmol, 1200 equiv), solvent (2 mL), 4 h. With 100,000 equiv
2
2
b
Catalytic Hydrogenation of CO . The results in Schemes
and 5 demonstrate the feasibility of all three individual steps
2
c d
4
of K CO , 5 mL of diglyme. 24 h. 48 h.
2 3
2
413
dx.doi.org/10.1021/cs400609u | ACS Catal. 2013, 3, 2412−2416

ACS Catalysis
Research Article
and only 7 turnovers were observed after 4 h (max possible
TON = 1200 based on equiv of base). However, raising the
temperature to 120 °C resulted in a dramatic improvement in
the performance of catalyst 3, as it afforded 270 turnovers
under otherwise analogous conditions (entry 2). We reasoned
that replacing the nonpolar solvent anisole with diglyme should
deactivation pathway. To test this possibility, we also examined
the reactivity of 6 as a catalyst for CO hydrogenation under
2
our standard reaction conditions. Unexpectedly, we found that
this Ru−CO adduct afforded a TON of formate comparable to
2
that of 3 and 5 after 4 h at 120 °C (Table 2, entry 2).
The observed catalytic activity of 6 can be explained by at
t
30
better solubilize KO Bu. This change resulted in a 3-fold
improvement in TON (to 800) after the same period of time
least two mechanistic possibilities. A first is that CO binding at
2
6 (step ia of Scheme 3) could be reversible at the elevated
temperatures used for catalysis. This would enable the
(
entry 3).
A variety of different bases were next evaluated for the
regeneration of 3, which could then participate in CO
hydrogenation. Alternatively, 6 could potentially be capable of
directly catalyzing CO hydrogenation. As outlined in Scheme
2
reaction in diglyme. K CO , KOH, K PO , and KHCO were
2
3
3
4
3
all effective in promoting this transformation, with TONs
2
31
6
, a possible mechanism for this latter transformation could
ranging from 510 to 1,100 (entries 4−7). The best base for
this reaction was K CO , yielding a TON of 1,100 and a TOF
2
3
−1
32,33
of 1,600 h .
These data stand in contrast to Nozaki’s
Scheme 6. Possible Catalytic Cycle for CO
at Complex 6
Hydrogenation
2
results with Ir pincer complex 2. In that system, moving from
KOH to a weaker base like K PO resulted in a significant
3
4
1
8
decrease in the TON (>7-fold). The diversity of bases
effective in the reaction of 3 suggests that deprotonation of
intermediate 5 may be more facile than deprotonation of the
analogous Ir−formate intermediate. Notably, however, the
neutral amine base NEt , which has been frequently employed
3
5
−9,12−14,17
in CO hydrogenation reactions,
resulted in <5
2
turnovers in this system (entry 8).
Because the best results were obtained using K CO as the
2
3
base (entry 4), we next examined the effect of increasing the
equivalents of K CO , which would correspondingly increase
2
3
the theoretical maximum TON of 3. The use of 100,000 equiv
of K CO at 120 °C provided 660 turnovers after 4 h (entry 9)
2
3
and 1,400 turnovers after 24 h. Notably, based on the
stoichiometry of K CO , the maximum possible turnovers is
2
3
3
2
2
00,000. Furthermore, increasing the temperature to 200 °C
involve deprotonation of 6 to generate the unsaturated complex
7 (step i), followed by H heterolysis (step ii), CO insertion
provided 9,000 and 23,000 turnovers after 4 and 48 h,
respectively, with a TOF of 2,200 h at this temperature.
−1
33
2
2
34
(
step iii), and base-promoted product release (step iv).
To test the first possibility (reversible formation of 3 from 6
Preliminary Mechanistic Considerations. As discussed
above, we originally envisioned the catalytic cycle in Scheme 3
as a plausible pathway for this transformation. If this
mechanism is operative, formate complex 5 should display
similar catalytic activity as 3. Indeed, under otherwise identical
conditions, this catalyst provided a comparable TON after 4 h
13
at elevated temperatures), we treated a CO -labeled sample of
6
2
12
with 1 bar of CO at 120 °C. After 4 h, >98% exchange was
2
observed (Supporting Information, Figure S8), indicating
essentially complete reversibility under our standard catalysis
conditions (Table 3, entry 1). Notably, the extent of exchange
decreased sharply with temperature. At 100 °C, <45% exchange
was observed after 4 h, and minimal (3%) exchange was
detected after 4 h at 70 °C (entries 2−3). To probe the
(
compare Table 1, entry 4 and Table 2, entry 1) and reaction
rate profiles (see Supporting Information, Figures S5 and S6).
The mechanism as drawn in Scheme 3 implicates Ru-CO2
complex 6 as an off-cycle side product. Previous studies from
our group showed that the formation of 6 is irreversible at
relevance of this reversibility to CO hydrogenation catalysis,
2
27
we examined the 6-catalyzed hydrogenation of CO to formate
2
room temperature, suggesting it may serve as a catalyst
at 70 °C (slow exchange conditions). As shown in Table 2,
Table 2. Complexes 3, 5, and 6 as Catalysts for CO
Hydrogenation
2
a
Table 3. Quantification of Reversible Binding of CO at
2
1
3
a
6- C at Varied Temperatures
entry
[Ru]
temp (°C)
TON formate
1
2
3
4
5
5
6
6
3
5
120
120
70
1,000
1,100
<5
13
entry
temp (°C)
%6- C remaining
70
340
1
2
3
120
100
70
1.2
54
70
430
97
a
Conditions: CO (10 bar); H (30 bar); Ru catalyst 3, 5, or 6 (0.554
2
2
a
13
μmol, 1 equiv), K CO (0.6648 mmol, 1200 equiv), diglyme (2 mL), 4
Conditions: CO (1 bar), 6- C (1.8 mg, 3.63 μmol), anisole (0.45
2
3
2
h.
mL), 4 h.
2
414
dx.doi.org/10.1021/cs400609u | ACS Catal. 2013, 3, 2412−2416

ACS Catalysis
Research Article
entry 3, less than 5 turnovers were observed after 4 h. In
contrast, 3 and 5 provided 340 and 430 turnovers under these
conditions at 70 °C (entries 4−5). The efficiency of 6 as a
catalyst for hydrogenation at higher temperatures (fast
exchange conditions), but not at lower temperatures (slow
exchange conditions), suggests that the reversible binding of
ASSOCIATED CONTENT
Supporting Information
Crystallographic data in CIF format. Further details are given in
Figures S1−S7 and Tables S1−S4. This material is available free
of charge via the Internet at http://pubs.acs.org.
■
*
S
CO is likely relevant to catalysis by 6 at elevated temperatures.
2
AUTHOR INFORMATION
Corresponding Author
E-mail: mssanfor@umich.edu. Phone: 734 615 0451.
Notes
■
*
Importantly, the reversible formation of 3 from 6 under the
conditions for catalysis does not rule out the possibility of
direct CO2 hydrogenation at 6 (catalytic cycle shown in
Scheme 6). To explore this latter possibility, we first examined
t
The authors declare no competing financial interest.
the stoichiometric reaction of 6 with 1 equiv of KO Bu in
dimethylsulfoxide (DMSO) at 25 °C. After 5 min, a color
change from yellow to bright orange was observed,
accompanied by the complete conversion of 6 to a new Ru−
H species, 7 (Scheme 7). This complex proved challenging to
ACKNOWLEDGMENTS
■
This work was supported by the U.S. National Science
Foundation under the CCI Center for Enabling New
Technologies Through Catalysis (CENTC) Phase II Renewal,
CHE-1205189. C.A.H. was supported by a NSF Graduate
Research Fellowship and by a Rackham Merit Fellowship. We
also acknowledge funding from NSF Grant CHE-0840456 for
X-ray instrumentation and the contribution of Dr. Jeff Kampf in
collecting and refining crystallographic data. Finally, we thank
Dr. Alex Miller (UNC) and Dr. Karen Goldberg (UW) for
helpful discussions.
Scheme 7. Formation of Anionic Ru Complex 7 by
Deprotonation of 6
REFERENCES
■
(
1) Reutemann, W.; Kieczka, H. Formic Acid. In Ullmann’s
Encyclopedia of Industrial Chemistry; Wiley-VCH: Weinheim, Germany,
35
isolate in high purity, as it is extremely moisture sensitive;
however, an in situ-generated sample of 7 was fully
characterized by H and C NMR spectroscopy.
2011.
1
13
(2) Schaub, T.; Paciello, R. A. Angew. Chem., Int. Ed. 2011, 50, 7278.
(3) Wesselbaum, S.; Hintermair, U.; Leitner, W. Angew. Chem., Int.
Ed. 2012, 51, 8585.
To probe whether 7 can participate in steps ii and iii of the
catalytic cycle proposed in Scheme 6, a sample of 7 was
35
(4) Zhang, Z.; Xie, Y.; Li, W.; Hu, S.; Song, J.; Jiang, T.; Han, B.
heated in DMSO-d in the presence of 1 bar of a 4:1 mixture of
6
Angew. Chem., Int. Ed. 2008, 47, 1127.
H :CO at 120 °C for 1 h in the absence of exogenous base
2
2
(
(
(
5) Leitner, W. Angew. Chem., Int. Ed. 1995, 34, 2207.
6) Jessop, P. G.; Ikariya, T.; Noyori, R. Chem. Rev. 1995, 95, 259.
7) Jessop, P. G. Homogeneous Hydrogenation of Carbon Dioxide.
(
Scheme 8). Under these conditions, 15% yield of HCOOK
Scheme 8. Stoichiometric Conversion of CO to Formate
2
In Handbook of Homogeneous Hydrogenation; Wiley-VCH: Weinheim,
Germany, 2007; p 489.
with 7
(
8) Jessop, P. G.; Joo, F.; Tai, C.-C. Coord. Chem. Rev. 2004, 248,
425.
9) Wang, W.; Wang, S.; Ma, X.; Gong, J. Chem. Soc. Rev. 2011, 40,
703.
10) Jeletic, M. S.; Mock, M. T.; Appel, A. M.; Linehan, J. C. J. Am.
Chem. Soc. 2013, 135 (31), 11533−11536.
11) Langer, R.; Disken-Posner, Y.; Leitus, G.; Shimon, L. J. W.; Ben-
David, Y.; Milstein, D. Angew. Chem., Int. Ed. 2011, 50, 9948.
12) Federsel, C.; Ziebart, C.; Jackstell, R.; Baumann, W.; Beller, M.
Chem.Eur. J. 2012, 18, 72.
13) Federsel, C.; Boddien, A.; Jackstell, R.; Jennerjahn, R.; Dyson, P.
J.; Scopelliti, R.; Laurenczy, G.; Beller, M. Angew. Chem., Int. Ed. 2010,
9, 9777.
14) Ziebart, C.; Federsel, C.; Anbarasan, P.; Jackstell, R.; Baumann,
2
(
3
(
(
(
3
6
was detected. This result suggests that the 6-catalyzed
(
hydrogenation of CO (Scheme 6) is a potentially viable
route to formate, albeit a likely minor pathway relative to that
depicted in Scheme 3.
In summary, this paper has shown that Ru(PNN)(CO)(H)
3), a known ester hydrogenation catalyst, can also catalyze the
hydrogenation of CO to formate in the presence of a base. The
transformation is proposed to proceed through a mechanism
involving (i) heterolytic cleavage of H at 3 to form a Ru−
dihydride species, (ii) CO insertion to generate a Ru−formate
complex, and (iii) base-promoted release of formate. The
feasibility of each of these proposed mechanistic steps has been
demonstrated through stoichiometric studies of organometallic
intermediates. Ongoing work in this area aims to exploit this
newfound reactivity of 3, in combination with its ester
hydrogenation activity, to accomplish 3-catalyzed production
of MeOH from CO2.
2
4
(
W.; Spannenberg, A.; Beller, M. J. Am. Chem. Soc. 2012, 134, 20701.
(15) Huff, C. A.; Sanford, M. S. J. Am. Chem. Soc. 2011, 133, 18122.
(16) Wesselbaum, S.; vom Stein, T.; Klankermayer, J.; Leitner, W.
Angew. Chem., Int. Ed. 2012, 51, 7499.
(
2
(
17) Munshi, P.; Main, A. D.; Linehan, J. C.; Tai, C.-C.; Jessop, P. G.
J. Am. Chem. Soc. 2002, 124, 7963.
18) Tanaka, R.; Yamashita, M.; Nozaki, K. J. Am. Chem. Soc. 2009,
31, 14168.
19) PNN = 6-(di-tert-butylphosphinomethylene)-2-(N,N-diethyla-
minomethyl)-1,6-dihydropyridine).
20) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. J. Am. Chem.
2
2
(
1
(
(
Soc. 2005, 127, 10840.
(21) Zhang, J.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2006, 45, 1113.
2
415
dx.doi.org/10.1021/cs400609u | ACS Catal. 2013, 3, 2412−2416

ACS Catalysis
Research Article
(
22) Computational studies on similar systems suggest against a
mechanism involving dissociation of one arm of the pincer ligand. See
also refs 23 and 24. However, we can not rule out hemilability of a
pincer ligand arm in the current system.
(
23) For a computational study of CO hydrogenation catalyzed by 1
2
see: Li, J.; Yoshizawa, K. Bull. Chem. Soc. Jpn. 2011, 84, 1039.
24) For a computational study of the hydrogenation of dimethyl
carbonate catalyzed by 3 see: Yang, X. ACS Catal. 2012, 2, 964.
25) See the Supporting Information for full details on the synthesis
and characterization of 5.
26) For an analogous FePNP complex see: Zell, T.; Butschke, B.;
Ben-David, Y.; Milstein, D. Chem.Eur. J. 2013, 19, 8068.
27) For details on the synthesis and characterization of 6, see: Huff,
C. A.; Kampf, J. W.; Sanford, M. S. Organometallics 2012, 31, 4643.
28) For binding of CO at a related RuPNP complex see: Vogt, M.;
(
(
(
(
(
2
Gargir, M.; Iron, M. A.; Diskin-Posner, Y.; Ben-David, Y.; Milstein, D.
Chem.Eur. J. 2012, 18, 9194.
(
29) For binding of other carbonyl compounds at 3 see: Huff, C. A.;
Kampf, J. W.; Sanford, M. S. Chem. Commun. 2013, 49, 7147.
30) Szmant, H. H. Organic Building Blocks of the Chemical Industry;
Wiley: New York, 1989; p 222.
31) Control reactions were peformed to probe the direct
hydrogenation of K CO and KHCO to HCOOK catalyzed by 3 in
(
(
2
3
3
the absence of CO . TONs of <5 for K CO and 65 for KHCO were
2
2
3
3
observed under the standard conditions “a” described in Table 1.
(
32) This value assumes that both K CO and KHCO can act as
2 3 3
bases in the reaction.
(
33) In all cases, the TOF was calculated using the number of
turnovers of formate detected at ∼35% conversion. The time required
to reach ∼35% conversion varied with the catalyst and conditions.
Importantly, 35% conversion represents 35% of the maximum
observed conversion rather than 35% of the maximum possible
conversion.
(
34) Alternative structures for 8 and 9 involve dissociation of the
hemilabile NEt arm, where the anionic oxygen tethered to the ligand
2
could then coordinate to the Ru metal center. Our data do not allow
us to distinguish these structures from those proposed in Scheme 6.
(
35) The product was isolated in 89% yield, but the isolated material
contained ∼10% of starting material 6 (presumably generated by
protonation).
(
36) We believe that the low yield is due to the water sensitivity of 7.
2
416
dx.doi.org/10.1021/cs400609u | ACS Catal. 2013, 3, 2412−2416