Tetrahedron Letters 58 (2017) 3330–3332

Contents lists available at ScienceDirect

Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet

Carbon dioxide hydrogenation: Efficient catalysis by an NHC-amidate

Pd(II) complex
Nima Zargari, Erica Jung, Joo-Ho Lee, Kyung Woon Jung ⇑
Loker Hydrocarbon Research Institute and Department of Chemistry, University of Southern California, Los Angeles, CA 90089-1062, United States

a r t i c l e i n f o a b s t r a c t

Article history: The utilization of carbon dioxide as a carbon source has long been a challenge in modern organic chem-
Received 20 June 2017 istry due to its low reactivity, yet high abundance. Herein we demonstrate the highly efficient hydrogena-
Revised 8 July 2017 tion of carbon dioxide into formic acid in the presence of an NHC-amidate Pd(II) complex. Excellent
Accepted 10 July 2017
turnover number was observed when the catalyst was used under heterolytic conditions. This catalytic
Available online 11 July 2017
system provides a new and efficient carbon dioxide hydrogenation method.
Ó 2017 Elsevier Ltd. All rights reserved.
Keywords:
Carbon dioxide
Hydrogenation
Formic acid
Palladium catalysis

Being an undesirable side product of the industrial revolution ondary alkyl amines, the direct carbonylation of carbon dioxide
and the driving force of the destruction of biomass, carbon dioxide may take place in order to produce formamides or methylamines.10
is a greenhouse gas which is difficult to activate and react due to its Performing the reactions in supercritical carbon dioxide signifi-
thermodynamic stability.1 Moreover, since there are exhaustible cantly increased the yields of these products, compared to reac-
resources available, it is of paramount importance to utilize carbon tions using liquid solvents.
dioxide as a C1-building block not only to reduce the amount of A plethora of studies have been found which selectively hydro-
greenhouse gasses, but also to recycle carbon dioxide as a basic genate carbon dioxide into formic acid, under varying solvents,
raw material into chemical feedstock products or alternative fuel metal catalysts, and bases. This is of great value since several stud-
sources.2 Various works have been performed in order to utilize ies have employed formic acid as a hydrogen source in direct for-
carbon dioxide by either direct hydrogenation, other activation mic acid fuel cells (DFAFC). DFAFC are considered to be an
reactions, or by direct carbon dioxide capture and storage.3 alternative to methanol and hydrogen fuel cells due to their ease
Employing N,N0 -ethylenebis(salicylimine) (Salen) and metal of refueling, efficiency, and safety. This allows DFAFC to potentially
catalysts, carbon dioxide has been copolymerized with epoxides be utilized in common consumer electronics and automobiles.11
producing polycarbonates,4 while cyclic carbonate and polycar- Notably, recent studies have employed the use of NHC ligands
bonate formation was achieved by Group 3, 4, and 5 metal in order to reduce carbon dioxide into either methanol or formic
complexes.5 acid.12 Herein, we report the development of a set of reaction con-
The most sought out reaction of carbon dioxide is the direct ditions which selectively hydrogenate carbon dioxide into formic
hydrogenation of the gas into valuable products such as methanol, acid utilizing our NHC-amidate Pd(II) complex 1 (Fig. 1).13 Our ini-
formamides, and formic acid.6 Transition metals, including early tial reaction conditions (Scheme 1) closely emulated that of Take-
transition metals, have excessively been utilized to catalyze reac- zaki et. al. studies14 where a palladium source was utilized in the
tions with carbon dioxide and attain these products.7 Using super- presence of a base, producing formic acid as the major product.
critical carbon dioxide and a ruthenium catalyst, carbon dioxide When an equal pressure (10 atm) of carbon dioxide and hydrogen
has been selectively converted into methanol.8 Other reaction con- gas were stirred at 160 °C for three hours in the presence of 1
ditions employ carbon dioxide at much lower pressures yet still (5 lmol) in an alkaline aqueous solution, 32.2 lmol (6.4 TON) of
produce methanol in the presence of copper, zinc, and aluminum formic acid were produced as the major product, while methanol
co-catalysts.9 In the presence of ammonia and primary and sec- and formaldehyde were produced in trace amounts (less than
0.5 lmol).
⇑ Corresponding author. In light of this promising result, further optimization was
E-mail address: kwjung@usc.edu (K.W. Jung).
investigated to obtain the highest yield for formic acid. Firstly,

http://dx.doi.org/10.1016/j.tetlet.2017.07.040
0040-4039/Ó 2017 Elsevier Ltd. All rights reserved.
 N. Zargari et al. / Tetrahedron Letters 58 (2017) 3330–3332 3331

Table 1
Effects of varying temperature.a

Entry Temperature (°C) Time (h) HCOOH (TON)b

1 60 3 1
2 100 3 5
3 160 3 12
4 200 3 57
Fig. 1. Structure of our NHC-amidate Pd(II) complex.
a
Reaction conditions: 1.7 mmol KOH and 5 lmol complex 1 were dissolved in
1.0 mL H2O. The reaction solution was then placed in a stainless steel reactor and
charged with 6.8 atm (800 lmol) hydrogen and 20 atm (2400 lmol) carbon dioxide
gas.
b
Yields were calculated using DMSO as an internal NMR standard.

Scheme 1. Initial reaction conditions utilizing complex 1.

since there is the possibility of over reduction of carbon dioxide to

form methanol or formaldehyde, variation of ratio between CO2
and H2 was tested. When the same equivalents (10 atm) of CO2
and H2 were used for the reaction, trace amount of methanol
(0.1 TON) and formaldehyde (0.04 TON) were formed as well as
6 TON of formic acid (Scheme 1). We found the optimal conditions
were the use of three equivalents of CO2 to H2 (Scheme 2), where
the over reduced products were not detected. Due to this high
chemoselectivity, the yield of formic acid increased, doubling to
12 TON.
With this ratio, when the reaction temperature was lowered,
the formation of formic acid was significantly decreased (Table 1,
entries 1 and 2). Only 1 TON of the acid was produced at 60 °C
and 5 TON at 100 °C. On the contrary, when the reaction tempera- Fig. 2. Time study performed under standard reaction conditions from 1 to 6 h.
ture was increased to 200 °C, the production of formic acid was
dramatically increased to 57 TON.
We then examined time dependence of this transformation Table 2
Effects of varying base.a
(Fig. 2). The reaction reached the highest turnover number after
three hours, while half of the product amount was produced in Entry Base HCOOH (TON)b MeOH (TON)b HCHO (TON)b
two hours. Having the reaction run for an additional hour (4 h) 1 – 0 0 0
drastically decreased the amount of the acid. This result might 2 LiOH 1 0 0
stem from the lower concentration of CO2 or potassium carbonate, 3 NaOH 6 1 0
4 KOH 57 0 0
which could form by the reaction of CO2 with potassium hydroxide
5 RbOH 16 1 0
(vide infra). After 4 h (4–6 h), the interconversion between formic 6 CsOH 1 0 0
acid and H2 and CO2 equilibrated to about 34 TON of formic acid. 7 Ca(OH)2 0 1 0
This interconversion was previously reported by other Pd-NHC cat- 8 AgTFA 0 0 1
9 K2CO3 3 0 0
alysts and other metal catalysts such as rhodium and ruthenium.15
10 Et3N 0 0 0
Varying the base added to the reaction mixture was then tested. 11 NH4OH 1 1 0
As expected, in the absence of any base, no reaction took place
a
Reaction conditions: 1.7 mmol of varying base and 5 lmol complex 1 were
(Table 2, entry 1). KOH not only stood out from other group one
dissolved in 1.0 mL H2O. The reaction solution was then placed in a stainless steel
alkaline bases (Table 2, entries 2–6), but also from group two reactor and charged with 6.8 atm (800 lmol) hydrogen and 20 atm (2400 lmol)
alkaline bases and organic bases as well. The major product of carbon dioxide gas. The mixture was heated at 200 °C for 3 h.
b
the reaction using calcium hydroxide was methanol in low yield Yields were calculated using DMSO as an internal NMR standard.
(Table 2, entry 7), while silver trifluoroacetate provided formalde-
hyde as the major product (Table 2, entry 8). Potassium carbonate
absence of a palladium source (Table 3, entry 1). Palladium(II) chlo-
selectively produced formic acid in moderate yields (Table 2, entry
ride produced the lowest yield of formic acid (Table 3, entry 3),
9). Amine bases such as triethylamine and ammonium hydroxide
while bis(acetonitrile)dichloropalladium (II) produced the rela-
failed to produce any significant result or produced trace amounts
tively high yield of formic acid (Table 3, entry 4). Palladium(II)
of both formic acid or methanol, respectively (Table 2, entries 10
acetate was slightly less active, producing 43 TON of formic acid
and 11).
(Table 3, entry 5).
We then compared our NHC-amidate Pd(II) complex 1 with
Furthermore, in order to improve yields, higher pressures of
other palladium sources, which are commercially available and
carbon dioxide and hydrogen were applied under homogeneous
widely used (Table 3). Critically, no reaction took place in the
and heterogeneous conditions (Table 4). Increasing the pressure
of carbon dioxide to 68 atm and hydrogen to 41 atm yielded almost
twice formic acid with trace amounts of methanol (Table 4, entry
1). These pressures were the maximal amounts that could be
applied to the reactor system, while using the same amount of
the solvent. Interestingly, when NHC-amidate Pd(II) complex 1
Scheme 2. Optimal ratio of carbon dioxide and hydrogen. was employed as a heterogeneous catalyst, a high turnover of
3332 N. Zargari et al. / Tetrahedron Letters 58 (2017) 3330–3332

Table 3 carbonate through the appearance of 13C NMR peak at

Effects of palladium catalysts.a 168.2 ppm., which was confirmed by the separate spectral analysis
Entry Pd Source HCOOH (TON)b of potassium carbonate (d 168.2 ppm) and potassium bicarbonate
1 – 0 (d 159.9 ppm). Therefore, we believe the carbon source which
2 1 57 reacts with complex 1 is carbonate and not carbon dioxide
3 PdCl2 7 (Scheme 3). As a primary catalytic complex, the formation of
4 Pd(MeCN)2Cl2 49 Pd-OCO2 (2) could be proposed through the substitution of
5 Pd(OAc)2 43
chloride by carbonate. Reduction of carbonate by H2 could lead
a
Reaction conditions: 1.7 mmol KOH and 5 lmol of a palladium source were to form formate complex 3. Then, this formate would be replaced
dissolved in 1.0 mL H2O. The reaction solution was then placed in a stainless steel by another carbonate molecule.
reactor and charged with 6.8 atm (800 lmol) hydrogen and 20 atm (2400 lmol)
In summary, we have developed a set of reaction conditions
carbon dioxide gas. The mixture was heated at 200 °C for 3 h.
b
Yields were calculated using DMSO as an internal NMR standard. which selectively hydrogenate carbon dioxide gas into formic acid,
which can be utilized in alternative fuel sources, such as DFAFC.
Notably, our NHC-amidate Pd(II) complex outcompetes other pal-
Table 4 ladium salts. In addition, the heterolytic catalysis using our NHC-
Effects of varying palladium sources under homogeneous and heterolytic conditions.a amidate Pd(II) catalyst 1 gave rise to the optimal conditions, yield-
Entry Pd Source H2O (mL) HCOOH (TON)b ing formic acid selectively and efficiently. This is of great impor-
1 1 1.0 95
tance in the sense of converting greenhouse gases into possible
2 1 0.0 150 alternative fuel sources.
3 – 0.0 0
Acknowledgment
a
Reaction conditions: 1.7 mmol KOH and 5 lmol of a palladium source were
added into a stainless steel reactor and charged with 68 atm (8000 lmol) carbon
dioxide and 41 atm (4800 lmol) hydrogen gas. The mixture was heated at 200 °C We acknowledge the Hydrocarbon Research Foundation for
for 3 h. generous financial support.
b
Yields were calculated using DMSO as an internal NMR standard.

A. Supplementary data

Supplementary data associated with this article can be found, in

the online version, at http://dx.doi.org/10.1016/j.tetlet.2017.07.
040.

References

1. Leitner W. Angew Chem Int Ed Engl. 1995;34:2207–2221.

2. Chavier C, Cantat T. ACS Catal. 2017;7:2107–2115.
3. Mikkelsen M, Jorgensen M, Krebs FC. Energy Environ Sci. 2010;3:43–81.
4. Darensbourg DJ. Chem Rev. 2007;107:2388–2410.
5. Darensbourg DJ, Holtcamp MW. Coord Chem Rev. 1996;153:155–174.
6. (a) Choudhury J. ChemCatChem. 2012;4:609–611;
(b) Jessop PG, Ikariya T, Noyori R. Chem Rev. 1995;95:259–272.
7. Grice KA. Coord Chem Rev. 2017;336:78–95.
8. Jessop PG, Ikariya T, Noyori R. Nature. 1994;368:231–233.
9. (a) Hong Z, Cao Y, Deng J, Fan K. Catal Lett. 2002;82:37–44;
(b) Saito M, Murata K. Catal Surv Asia. 2004;8:285–294;
(c) Saito M. Catal Surv Jpn. 1998;2:175–184;
(d) Saito M, Takeuchi M, Fujitani T, et al. Appl Organomet Chem.
2000;14:763–772;
(e) An X, Li J, Zuo Y, Zhang Q, Wang D, Wang J. Catal Lett. 2007;118:264–269.
10. (a) Jessop PG, Hsiao Y, Ikariya T, Noyori R. J Am Chem Soc. 1996;118:344–355;
(b) Jessop PG, Hsiao Y, Ikariya T, Noyori R. J Am Chem Soc.
1994;116:8851–8852;
(c) Vaska L, Schreiner S, Felty RA, Yu JY. J Mol Catal. 1989;52:L11–L16;
(d) Haynes P, Slaugh LH, Kohnle JF. Tetrahedron Lett. 1970;365–368.
11. (a) Rees NV, Compton RGJ. Solid State Electrocem. 2011;15:2095–2100;
(b) Uhm S, Chung ST, Lee J. J Power Sources. 2008;178:34–43;
(c) Rice C, Ha S, Masel RI, Waszczuk P, Wieckowski A, Barnard T. J Power Sources.
2002;111:83–89.
12. (a) Schmeier TJ, Dobereiner GE, Crabtree RH, Hazari N. J Am Chem Soc.
Scheme 3. Proposed mechanism of CO2 hydrogenation by 1. 2011;133:9274–9277;
(b) Riduan SN, Zhang Y, Ying JY. Angew Chem. 2009;121:3372–3375;
(c) Sanz S, Benítez M, Peris E. Organometallics. 2010;29:275–277;
formic acid (150 TON) was selectively produced in the absence of (d) Lee CH, Laitar DS, Mueller P, Sadighi JP. J Am Chem Soc.
2007;129:13802–13803.
any solvent (Table 4, entry 2). Though reported otherwise,16 no 13. Lee JH, Yoo KS, Park CP, et al. Adv Synth Catal. 2009;351:563–568.
product was formed when no catalyst was used (Table 4, entry 14. Kudo K, Sugita N, Takezaki Y. Nippon Kagaku Zasshi. 1977;3:302–309.
3). This discrepancy between prior art may be due to the differ- 15. (a) Broggi J, Jurčík V, Songis O, et al. J Am Chem Soc. 2013;135:4588–4591;
(b) Himeda Y, Miyazawa S, Hirose T. ChemSusChem. 2011;4:487–493.
ences in reaction scale or time. Comparatively, other palladium 16. Sanz S, Azua A, Peris E. Dalton Trans. 2010;39:6339–6343.
(II) sources furnished only 24 TON of formic acid under these het- 17. (a) Osadchuk I, Tamm T, Ahlquist MSG. Organometallics. 2015;34:4932–4940;
erolytic conditions. (b) Appel AM, Bercaw JE, Bocarsly AB, et al. Chem Rev. 2013;113:6621–6658;
(c) Suh HW, Schmeier TJ, Hazari N, Kemp RA, Takase MK. Organometallics.
Based on the aspects of these conditions, we considered the fol- 2012;31:8225–8236;
lowing mechanistic rationale. In an aqueous alkaline media, carbon (d) Tanaka R, Yamashita M, Chung LW, Morokuma K, Nozaki K. Organometallics.
dioxide is readily converted into potassium carbonate.17 When we 2011;30:6742–6750.
mixed CO2 and KOH in D2O, we observed the sole formation of