Evidence for direct hydride delivery from formic acid in transfer hydrogenation

Article abstract of DOI:10.1002/(SICI)1521-3765(19990802)5:8<2237::AID-CHEM2237>3.0.CO;2-O

The regioselectivity of hydrometalation in transfer hydrogenation was investigated by tracing the fate of the deuterium label in both the formyl and carboxyl positions of formic acid during the isomerization of a cis enol ether. The deuterium from the for

Full text of DOI:10.1002/(SICI)1521-3765(19990802)5:8<2237::AID-CHEM2237>3.0.CO;2-O

FULL PAPER
Evidence for Direct Hydride Delivery from Formic Acid in Transfer
Hydrogenation
Jinquan Yu and Jonathan B. Spencer*[a]
Abstract: The regioselectivity of hydro-
metalation in transfer hydrogenation
was investigated by tracing the fate of
the deuterium label in both the formyl
and carboxyl positions of formic acid
during the isomerization of a cis enol
ether. The deuterium from the formyl
position is incorporated exclusively into
the electron-deficient center of the dou-
ble bond; this indicates that hydride
hydrogen suggests that the identities of
the two hydrogens are not scrambled
during transfer hydrogenation. We pro-
pose that the delivery of the formyl
hydrogen atom is concerted with decar-
boxylation rather than proceeding
through the formation of a metal dihy-
drido species.
attack occurs during transfer hydroge-
nation. The fact that the formyl hydro-
gen is more reactive than the carboxyl
Keywords: carbon dioxide ´ formic
acid
´ heterogeneous catalysis ´
hydrogenations ´ hydrogen transfer
involved palladium diformate (3) as a key intermediate
(Scheme 2).[16] This intermediate could collapse to form the
dihydrido metal species 4, which then reacts with the
unsaturated compound. The surface-bound metal center
represented by the dihydrido metal species 4 should be
equivalent to that generated from hydrogen gas with the
catalyst, and should therefore have the same reactivity.
Introduction
Transfer hydrogenation using formic acid as the hydrogen
donor is a valuable alternative to traditional hydrogenation
with hydrogen gas.[1±10] The reverse reaction, the reduction of
carbon dioxide to formic acid, is of great economic and
ecological importance.[11±14] The development of heterogene-
ous catalysts for this reversible reaction is of particular
interest for industrial processes.
O
O
There has been much interest in how the two different
hydrogen atoms from formic acid are delivered to the
substrate. On the basis of results from a previous study using
a homogeneous catalyst, it was proposed that dihydrido metal
species 2, formed by the loss of carbon dioxide from a formato
hydrido metal complex 1, is the active species that delivers the
two hydrogen atoms to the double bond (Scheme 1).[15]
D
D
Pd + 2 DCOOH
+
H2
Pd
O
O
3
D
Pd
+
2 CO2
O
D
O
H
H
M
M
+ CO2
M + HCOOH*
4
H*
H*
Scheme 2. Proposed pairwise mechanism for heterogeneous transfer
hydrogenation involving palladium diformate.
2
1
Scheme 1. Proposed mechanism for homogeneous transfer hydrogenation
involving formato hydrido metal complex.
Results and Discussion
For heterogeneous transfer hydrogenation an alternative
mechanism was suggested on the basis of labeling studies that
We recently developed a model system to determine the
electronic properties of the metal ± hydrogen bond.[17, 18] Dur-
ing isomerization of cis alkenes to trans alkenes with
deuterium gas and palladium on carbon, deuterium was
incorporated regioselectively, depending on the polarization
of the double bond (Scheme 3). This established that the
[a] Dr J. B. Spencer, J. Yu
University of Cambridge, Department of Chemistry
Lensfield Road, Cambridge, CB2 1EW (UK)
Fax: (‡44)1223-336362
E-mail: jbs20@cam.ac.uk
Chem. Eur. J. 1999, 5, No. 8
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2237 $ 17.50+.50/0
2237

J. B. Spencer, J. Yu
FULL PAPER
Mode (a)
In fact, when deuterated formic acid was used, deuterium
was incorporated only one way round (Table 1), indicating a
different mechanism of hydrometalation. The dramatic shift
from mode (b) to mode (a) indicates that the hydrogen
transferred has the electronic nature of a hydride ion. This
suggests that the hydrogen does not come from the metal
surface (dihydrido palladium species 4). Instead, it probably
comes directly from the formyl position of intermediate 3
(Scheme 2), as decarboxylation of the formate on the metal
surface would be predicted to deliver a hydride to the
electron-deficient carbon of the polarized double bond
(Scheme 4).
OMe
D
H
Ph
H
D
δ− δ+
H
H
Ph
Rotation
OMe
Ph
Pd
Ph
Pd
OMe
OMe
H
D
Pdδ+- Dδ−
5
Mode (b)
H
Ph
H
Ph
D
H
δ− δ+
H
Rotation
Ph
OMe
Ph
OMe
Pd
D
OMe
Pd
OMe
D
H
5
D
δ+- Pdδ−
Scheme 3. Amphipolar nature of palladium hydrogen bond.
Ph
OMe
δ− δ+
H
H
palladium ± hydrogen bond is amphipolar, and that the mode of
Ph
OMe
+ CO2
hydrometalation [mode (a) Pdd‡ Hd or mode (b) Pdd Hd‡
is governed by the electronic dipole of the substrate.
]
X
Pd
D
X
Pd
D
O
O
With this new approach, the mode of hydrometalation
during hydrogenation with hydrogen gas and transfer hydro-
genation with TEAF (HCOOH/NEt3 ˆ 5:2) can be compared
directly.[10, 19, 20] Since electron-donating ligands such as quino-
line and triphenylphosphine have been shown to promote
mode (a),[21] the triethylamine used in transfer hydrogenation
may have a similar effect. To ensure a clear comparison,
isomerization of the cis isomer of the enol ether 5 with
deuterium gas was carried out in triethylamine and trifluoro-
acetic acid (similar pKa to formic acid) to mimic the
conditions used in transfer hydrogenation (formic acid/
triethylamine). The result shows that hydrometalation takes
place by both mode (a) and mode (b), and to the same extent
(Table 1). If transfer hydrogenation involves the same species
4 as hydrogenation with hydrogen gas then the same result
should be obtained if deuterated formic acid and palladium on
carbon are used to isomerize the cis isomer of 5.
Rotation
Ph
H
D
OMe
OMe
D
β-elimination
H
Ph
+
X
H
Pd
H
X
Pd
X = HCOO-
Scheme 4. Delivery of a hydride by decarboxylation of 3.
To further investigate the different reactivities of the two
hydrogen atoms of formic acid, differentially labeled formic
acid was used (DCOOH and HCOOD). When the cis enol
ether was isomerized with DCOOH and palladium on carbon,
deuterium was solely incorporated into the position next to
the methoxy group (Table 1). The reverse experiment with
HCOOD gave virtually no deuterium in the isomerized trans
enol ether (Table 1). These combined results show that the
hydrogen delivered to the cis enol ether came only from the
formyl position of formic acid and provide strong evidence
that this hydrogen is transferred directly from the formate on
the metal surface.
The selective incorporation of deuterium to the position
remote to the phenyl group of the enol ether would be
consistent with addition occurring by direct transfer of either
a hydride or a hydrogen atom from the formyl position. In
selenium chemistry a radical will add exclusively remote to
the phenyl group of cinnamate.[16±18] If the hydrogen from the
formyl position is transferred as a radical then the deuterium
should be incorporated predominantly remote to the phenyl
group of methyl cinnamate.
Table 1. Relative deuterium incorporation into 5.
Hydrogen
source
Relative deuterium
distribution [%] [a]
Ph
D
H
Ph
H
D
OMe
OMe
5
5
45
55
D2
DCOOD
4
96
With methyl cinnamate (6), deuterium was added mainly
remote to the ester group, a result that is not consistent with a
radical pathway (Table 2). When the cis isomer of substrate 7
was isomerized under the same conditions as 6, deuterium was
also mainly incorporated remote to the ester group (Table 2).
This suggests that the formyl hydrogen is transferred as a
hydride in Michael-type addition to the a,b-unsaturated
carbonyl system since the methoxy group on the aromatic
ring would be expected to promote radical addition further.[22]
2
0
98
0
DCOOH
HCOOD
[a] Errors for the data are Æ5%. The level of deuterium
incorporation is only ca. 65% rather than 100% as
would have been predicted if the final step in the
mechanism proceeded by a syn b-elimination.[12]
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2238 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8
2238

Transfer Hydrogenation
2237±2240
1
Table 2. Relative deuterium incorporation into 6 and 7.
HCOOD, 99% atom excess as determined by H NMR spectroscopy and
mass spectrometry. DCOOD and Pd/C (10%) were purchased from
1
Aldrich. H NMR spectra were recorded on Bruker DPX250 or DRX400
instruments. High-resolution mass spectra were recorded on Kratos MS890
mass spectrometers. Distribution of the deuterium across the double bond
of the trans alkene was determined by means of 1H NMR spectroscopy as
previously reported.[11]
Hydrogenation using deuterium gas
Isomerization of the cis isomer of 5: Pd/C (3 mg,10% Pd) was added to a
mixture of CF3COOH (228 mg, 2.0 mmol), Et3N (1.5 g), and cis isomer of 5
(134 mg,1 mmol). The mixture was degassed and refilled with D2 through a
balloon. The reaction mixture was stirred for 2 h at room temperature
under 1 atm of D2. The reaction mixture was then filtered to remove the
catalyst and evaporated to give the product residue (128 mg) which was
determined by 1H NMR spectroscopy to consist of a mixture of 45%
starting material, 45% reduced product, and 10% of 5. The pure form of 5
was obtained by preparative TLC, eluting with hexane; the product was
[a] Measurement of deuterium incorporation and errors in data are the
same as detailed in Table 1.
1
then evaporated to dryness and analyzed by H NMR spectroscopy (data
for both labeled and unlabeled 5 are given) and MS. Rf ˆ 0.10; 1H NMR
(250 MHz, CDCl3): d ˆ 7.60, 7.30, 7.15 (5H; ArH), 7.06 (d, 3J(H, H) ˆ
3
3
13.0 Hz, 1 H), 7.06 (t, J(H, D) ˆ 3.5 Hz, 1 H), 5.83 (d, J(H, H) ˆ 13.0 Hz,
1 H), 5.81 (t, 3J(H, D) ˆ 3.5 Hz, 1 H), 3.72 (s, 3 H; OCH3); HRMS (EI)
The conclusion that a radical pathway is not operating is
further supported by a radical trap experiment[24, 25] with
chrysanthemyl alcohol (8) (Scheme 5). Cyclopropyl groups
are used as indicators of radical species in many systems,
because they readily open if a radical is generated on an
‡
2
m/z: [M ] calcd for C9H9 HO 135.0795, found 135.0802.
Transfer hydrogenation
General procedure for isomerization of cis isomer of 5: Pd/C (18 mg, 10%
Pd) was added to a mixture of DCOOH (235 mg, 5.0 mmol), Et3N (2.8 g),
and cis isomer of 5 (67 mg, 0.5 mmol).
The mixture was vigorously stirred
under a blanket of argon for 1.5 h at
room temperature. The reaction mix-
CH2OH
ture was then filtered to remove the
catalyst and evaporated to give the
product residue (60 mg) which was
determined by H NMR spectroscopy
CH2OH
1
Pd/C
9
to consist of a mixture of 50% starting
material, 41% reduced product, and
HCOOH/NEt3
9% of 5. The pure form of 5 was
obtained by preparative TLC, eluting
CH2OH
8
CH2OH
with hexane; the product was then
+
evaporated to dryness and analyzed by
1H NMR spectroscopy (data for both
labeled and unlabeled 5 are given) and
MS. Rf ˆ 0.10; 1H NMR (250 MHz,
CDCl3): d ˆ 7.60, 7.30, 7.15 (5H; ArH),
10
11
Scheme 5. Radical trap experiment using chrysanthemyl alcohol.
7.06 (d, 3J(H, H) ˆ 13.0 Hz, 1 H), 5.83
3
3
(d, J(H, H) ˆ 13.0 Hz, 1 H), 5.81 (t, J(H, D) ˆ 3.5 Hz, 1 H), 3.72 (s, 3 H,
adjacent carbon. Chrysanthemyl alcohol, which contains a
cyclopropyl group next to a double bond, was therefore
hydrogenated to see if the ring could be opened by a radical to
give products 10 and 11. Only compound 9 was identified
during the reaction, thus providing additional evidence that a
radical is not involved.
This study establishes that the formyl hydrogen is more
reactive than the carboxyl hydrogen. The regioselectivity of
hydrometalation with polarized alkenes suggests that hydro-
gen acts as a hydride ion. This strongly supports a mechanism
involving direct hydride transfer from the formyl position of
formic acid (Scheme 4).
‡
2
OCH3); HRMS (EI) m/z: [M ] calcd for C9H9 HO 135.0795, found
135.0789.
Isomerization of the cis isomer of 6: Pd/C (18 mg, 10% Pd) was added to a
mixture of DCOOH (235 mg, 5.0 mmol), Et3N (2.8 g), and cis isomer of 6
(81 mg, 0.5 mmol). The mixture was vigorously stirred under a blanket of
argon for 1 h at room temperature. The reaction mixture was then filtered
to remove the catalyst and evaporated to give the product residue (72 mg)
which was determined by 1H NMR spectroscopy to consist of a mixture of
58% starting material, 32% reduced product, and 10% of 6. The pure form
of 6 was obtained by preparative TLC, eluting with petroleum ether (15:1);
the product was then evaporated to dryness and analyzed by 1H NMR
spectroscopy (data for both labeled and unlabeled are given) and MS. Rf ˆ
0.19; 1H NMR (400 MHz, CDCl3): d ˆ 7.70 (d, 3J(H, H) ˆ 15.9 Hz, 1 H),
7.69 (t, 3J(H, D) ˆ 2.8 Hz, 1 H), 7.60, 7.40 ± 7.18 (5H; ArH), 6.45 (d,
3J(H, H) ˆ 15.9 Hz, 1 H), 6.44 (t, 3J(H, D) ˆ 2.8 Hz, 1 H), 3.81 (s, 3 H;
‡
2
OCH3); HRMS (EI) m/z: [M ] calcd for C10H9 HO2, 163.0744; found
163.0750.
Experimental Section
Isomerization of the cis isomer of 7: Pd/C (18 mg, 10% Pd) was added to a
mixture of DCOOH (235 mg, 5.0 mmol), Et3N (2.8 g), and cis isomer of 7
(96 mg, 0.5 mmol). The mixture was vigorously stirred under a blanket of
argon for 1.5 h at room temperature. The reaction mixture was then filtered
to remove the catalyst and evaporated to give the product residue (85 mg),
which was determined by 1H NMR spectroscopy to consist of a mixture of
52% starting material, 40% reduced product, and 8% of 7. The pure form
The cis isomers of 5 and 8 were commercial products from Aldrich (5 was
further purified by preparative TLC, eluting with hexane). The cis isomers
of 6 and 7 were prepared as previously reported.[28] Labeled formic acids
were prepared by treating sodium formate (either DCOONa or HCOONa)
with 1N HCl or DCl, extracting with diethyl ether and drying over
anhydrous sodium sulfate. Distillation of the extracts gave DCOOH or
Chem. Eur. J. 1999, 5, No. 8
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2239 $ 17.50+.50/0
2239

J. B. Spencer, J. Yu
FULL PAPER
of 7 was obtained by preparative TLC, eluting with petroleum ether (10:2);
the product was then evaporated to dryness and analyzed by 1H NMR
spectroscopy (data for both labeled and unlabeled 7 are given) and MS.
[7] B. T. Khai, A. Arcelli, J. Org. Chem. 1989, 54, 949 ± 953.
[8] H. Wiener, J. Blum, Y. Sasson, J. Org. Chem. 1991,56, 4481 ±
4486.
[9] G. Zassinovich, G. Mestroni, S. Gladiali, Chem. Rev. 1992, 92, 1051 ±
1069.
[10] R. Noyori, S. Hashiguchi, Acc. Chem. Res. 1997, 30, 97 ± 102.
[11] E. T. Sundquist, Science. 1993, 259, 934 ± 940.
[12] P. G. Jessop, I. Takao, R. Noyori, Nature 1994, 368, 231 ± 233.
[13] P. J. Jessop, I. Takao, R. Noyori, Chem. Rev. 1995, 95, 259 ± 272.
[14] W. Leitner, Angew. Chem. 1995, 107, 2391 ± 2405; Angew. Chem. Int.
Ed. Engl. 1995, 34, 2207 ± 2221.
1
Rf ˆ 0.12; H NMR (400 MHz, CDCl3): d ˆ 7.48, 6.90 (4H; ArH), 7.65 (d,
3J(H, H) ˆ 15.9 Hz, 1 H), 7.63 (t, 3J(H, D) ˆ 2.8 Hz, 1 H), 6.32 (d,
3J(H, H) ˆ 15.9 Hz, 1 H), 6.31 (t, 3J(H, D) ˆ 2.8 Hz, 1 H), 3.84 (s, 3 H;
OCH3), 3.79 (s, 3 H; OCH3) HRMS (EI) m/z: [M ] calcd for C11H112HO3
‡
193.08499, found 193.08490.
Transfer hydrogenation of chrysanthemyl alcohol 8: Pd/C (18 mg, 10% Pd)
was added to a mixture of DCOOH (235 mg, 5.0 mmol), Et3N (2.8 g), and
chrysanthemyl alcohol 8 (77 mg, 0.5 mmol). The mixture was vigorously
stirred under a blanket of argon for 4 h at room temperature. The reaction
mixture was then filtered to remove the catalyst and evaporated to give the
[15] W. Leitner, J. M. Brown, H. Brunner, J. Am. Chem. Soc. 1993, 115,
152 ± 159.
1
[16] J. Yu, J. B. Spencer, Chem. Commun. 1998, 1935 ± 1936.
[17] J. Yu, J. B. Spencer, J. Am. Chem. Soc. 1997, 119, 5257 ± 5258.
[18] J. Yu, J. B. Spencer, J. Org. Chem. 1997, 62, 8618 ± 8619.
[19] B. T. Khai, A. Arcelli, Tetrahedron Lett. 1985, 26, 3365 ± 3368.
[20] H. Brunner, W. Leitner, Angew. Chem. 1988, 100, 1231 ± 1232; Angew.
Chem. Int. Ed. Engl. 1988, 27, 1180 ± 1181.
product residue (70 mg) which was determined by H NMR spectroscopy
to consist of mainly starting material, i.e. no new double bond was formed.
Acknowledgments
[21] J. Yu, J. B. Spencer, Chem. Commun. 1998, 1103 ± 1104.
[22] D. H. Wadsworth, M. R. Detty, J. Org. Chem. 1980, 45, 4611 ± 4615.
[23] J. K. Kochi in Free Radicals, Vol. 2 (Ed.: G. A. Olah), John Wiley &
Sons, New York, 1982, pp. 283 ± 288.
We thank the Royal Society for the award of a Royal Society University
Research Fellowship to JBS.
[24] C. Leory, D. Peeters, C. J. Wilante, J. Mol. Struct. 1982, 88, 217 ± 233.
[25] D. C. Nonhedel, Chem. Soc. Rev. 1993, 347 ± 359.
[26] C. J. Suckling, Angew. Chem. 1988, 100, 555 ± 570; Angew. Chem. Int.
Ed. Engl. 1988, 27, 537 ± 552.
[27] E. N. Jacobsen, L. Deng, Y. Furukawa, L. E. Martinez, Tetrahedron
1994, 50, 4331 ± 4334.
[1] R. A. W. Johnstone, A. H. Wilby, I. D. Entwistle, Chem. Rev. 1985, 85,
129 ± 170.
[2] M. E. Volꢁpin, V. P. Kukolev, V. O. Chernyshev, I. S. Kolomnikov,
Tetrahedron Lett. 1971, 46, 4435 ± 4438.
[3] K. Nanjo, K. Suzuki, M. Sekiya, Chem. Lett. 1976, 1169 ± 1172.
[4] J. Blum, I. Pribar, H. Alper, J. Mol. Catal. 1986, 37, 359 ± 367.
[5] B. T. Khai, A. Arcelli, J. Organomet. Chem. 1986, 309, C63-C66.
[6] S. Ram, R. E. Ehrenkaufer, Synthesis. 1988, 91 ± 94.
Received: January 14, 1999 [F1547]
2240
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
0947-6539/99/0508-2240 $ 17.50+.50/0
Chem. Eur. J. 1999, 5, No. 8