Enantioselective homogeneous hydrogenation of monosubstituted pyridines and furans

Article abstract of DOI:10.1007/s007060070013

The first case of an enantioselective hydrogenation of monosubstituted pyridines and furans with homogeneous rhodium diphosphine catalysts with low but significant enantioselectivities and catalyst activities is reported. Best enantioselectivities (ees of 24-27%) were obtained for the hydrogenation of 2-and 3-pyridine carboxylic acid ethyl ester and 2-furan carboxylic acid with catalysts prepared in situ from [Rh(nbd)₂]BF₄ and the chiral ligands diop, binap, or ferrocenyl diphosphines of the josiphos type. Turnover numbers (ton) were in the order of 10-20, turnover frequencies (tof) usually 1-2 h^-1. Diphosphines giving 6-or 7-ring chelates led to higher ees than 1,2-diphosphines; otherwise, no clear correlation between ligand properties and catalytic performance was found. In some experiments black precipitates were observed at the end of the reaction, indicating the decomposition of the homogeneous catalysts for certain ligand/metal/ substrate combinations.

Full text of DOI:10.1007/s007060070013

Monatshefte fuÈr Chemie 131, 1335±1343 (2000)
Enantioselective Homogeneous Hydrogenation
of Monosubstituted Pyridines and Furans
Ã
and Hans-Ulrich Blaser
Martin Studer , Christina Wedemeyer-Exl, Felix Spindler,
Solvias AG, CH-4002 Basel, Switzerland
Summary. The ®rst case of an enantioselective hydrogenation of monosubstituted pyridines and
furans with homogeneous rhodium diphosphine catalysts with low but signi®cant enantioselectivities
and catalyst activities is reported. Best enantioselectivities (ees of 24±27%) were obtained for the
hydrogenation of 2- and 3-pyridine carboxylic acid ethyl ester and 2-furan carboxylic acid with
catalysts prepared in situ from [Rh(nbd)2]BF4 and the chiral ligands diop, binap, or ferrocenyl
diphosphines of the josiphos type. Turnover numbers (ton) were in the order of 10±20, turnover
frequencies (tof ) usually 1±2 h 1. Diphosphines giving 6- or 7-ring chelates led to higher ees than
1,2-diphosphines; otherwise, no clear correlation between ligand properties and catalytic
performance was found. In some experiments black precipitates were observed at the end of the
reaction, indicating the decomposition of the homogeneous catalysts for certain ligand/metal/
substrate combinations.
Keywords. Homogeneous catalysis; Enantioselective hydrogenation; Heterocyclic arenes; Rh dip-
hosphine complexes.
Introduction
The enantioselective synthesis of chiral compounds containing substituted
cyclohexanes, piperidines, or piperazines is of interest due to their biological
activity. Classical ways to synthesizie such ring systems enantioselectively are
asymmetric Diels-Alder reactions [1], the use of starting materials from the chiral
pool [2], or enzymatic methods [3]. A catalytic two-step procedure for the synthesis
of chiral piperazines starting from the corresponding pyrazine has been described
by Funchs and Roduit [4] who reported on the enantioselective reduction of
tetrahydropyrazine (prepared by Pd/C-catalyzed hydrogenation) by a homogeneous
Rh diphosphine complex with ees of 40±97%. We have described a similar approach
for the enantioselective synthesis of ethyl nipecotinate using cinchona-modi®ed
heterogeneous Pt catalysts for the second step, but ees were rather low [5]. A one-
step stereoselective catalytic hydrogenation of the respective aromatic compound
would be a very attractive method since many substituted arenes are readily
available. However, up to now the hydrogenation of monosubstituted pyrazines is the
Ã
Corresponding author

1336
M. Studer et al.
only example of such a reaction [6]. In a patent application, piperazine-2-carboxylic
acid t-butylesters and t-butylamides have been claimed to give ees up to 78% with
homogeneous Rh complexes of ferrocenyl diphosphine ligands, whereas the
corresponding methyl ester only gave 4% ee. All other efforts to hydrogenate
aromatic rings enantioselectively have resulted in ees below 6% [7±9]. A few cases
of a diastereoselective hydrogenation of aromatic rings with chiral substituents have
also been reported with moderate to good stereoselectivities [10±12]; in most cases,
heterogeneous catalysts were used. In this paper we describe our attempts to develop
homogeneous catalysts for the enantioselective hydrogenation of monosubstituted
pyridines and furans.
Results and Discussion
Selection of a suitable metal precursor
A homogeneous catalyst not suf®ciently stabilized by its ligands and/or the substrate
will decompose to give black precipitates that are often active with respect to ring
hydrogenation resulting in racemic products, since the chiral ligand is no longer
coordinated to the metal. Because the stability of a complex depends on the metal
precursor, we ®rst tested various precursor metal salts with the achiral diphosphine
1,2-bis(diphenylphosphino) ethane (dppe) in the hydrogenation of picolinic acid
ethyl ester (1, Scheme 1). [Rh(nbd)2]BF4, [Rh(nbd)2Cl]2, [Ir(cod)( py)2]PF6, RuCl2,
Pt(cod)Cl2, and [Ir(cod)Cl]2 were chosen; the best activity and a homogenous
solution at the end of the reaction was obtained with [Rh(nbd)2]BF4. [Rh(nbd)2Cl]2
used in Ref. [6] also gave high activity, but in most of the cases the mixtures were
heterogeneous at the end of the reaction. All other precursors gave heterogeneous
precipitates and/or were considerably less active. The successful combination
[Rh(nbd)2]BF4/dppe showed also reasonable conversions and yields for substrates
2±6; only for 4 and sometimes 2, the mono-unsaturated intermediate was observed as
a byproduct. For these reasons, 1 was chosen for our further experiments unless
otherwise indicated.
Scheme 1

Hydrogenation of Pyridines and Furans
1337
Screening of ligands and reaction parameters
Representatives of important ligand families (Scheme 2) were selected for screening,
using catalysts prepared in situ from [Rh(nbd)2]BF4 in the appropriate solvent.
Selected results are presented in Tables 1±4 and Fig. 1. Since the test reactions with
dppe had shown a rather low catalytic activity, the reactions were generally run at
60ꢀC and 100 bar hydrogen pressure with 5 mol% of catalyst (s/c ratio ˆ 20). The
effect of these reaction parameters was investigated with 3 (Table 3).
The most extensive ligand and solvent testing was carried out with the 2- and 3-
pyridinecarboxylic acid ethyl esters 1 and 2. In 24 experiments with 1, all ligands
were investigated in several solvents (EtOH, MeOH, toluene, and THF);
representative results are summarized in Table 1. Best ees were obtained in ethanol
with binap, diop, and xyliphos with moderate to high conversions and very high
chemoselectivity. Results in MeOH were comparable; for all other ligand-solvent
combinations, ees and quite often also catalytic activities were much lower and/or
metal precipitation was observed. For the 3-isomer 2 14 experiments were carried
Scheme 2
Table 1. Ligand and solvent screening for the hydrogenation of 1
Ligand
Solvent
Time
h
Conv.
%
Yield
%
ee
Comments
%
diop
EtOH
EtOH
EtOH
THF
20
21
20
20
20
20
20
20
4
41
100
23
41
96
27
25
20
13
9
binap
xyliphos
prophos
22
100
100
92
100
97
With [Rh(nbd)2Cl]2
Precipitate
bdpp
EtOH
EtOH
EtOH
EtOH
Toluene
chiraphos
PPF-P(t-Bu)2
Cy2PF-PCy2
PPF-P(t-Bu)2
90
9
Precipitate
29
28
6
96
96
5
100
100
4
Precipitate
Reaction conditions: [Rh(nbd)2]BF4, 60ꢀC, 100 bar, s/c ˆ 20

1338
M. Studer et al.
Table 2. Ligand screening for the hydrogenation of 2
Ligand
Solvent
Time
h
Conv.
%
Yield
%
Part.a
%
ee
Comments
%
bdpp
EtOH
EtOH
EtOH
EtOH
THF
19
20
17
19
18
85
100
100
75
45
52
44
22
64
40
45
56
53
33
17
12
9
diop
binap
Precipitate
chiraphos
prophos
8
100
7
With [Rh(nbd)2Cl]2,
precipitate
xyliphos
EtOH
20
64
31
33
6
a
Reaction conditions: [Rh(nbd)2]BF4, 60ꢀC, 100 bar s/c ˆ 20; Partially hydrogenated product
Table 3. Ligand screening and effect of p, T, and s/c for the hydrogenation of 3 and 4
Ligand
Solvent
Time
h
Yield
%
ee
Comments
%
3
3
3
3
3
3
3
3
3
3
Cy2PF-PPh2
Cy2PF-PPh2
Cy2PF-PPh2
Cy2PF-PPh2
Cy2PF-PPh2
Cy2PF-PPh2
PPF-P(t-Bu)2
diop
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
20
19
20
20
20
20
20
20
19
14
52
26
25
25
23
20
17
17
15
10
10
s/c ˆ 100
100
19
sc ˆ 5
room temperature
10 bar
16
3.6
58
room temperature/10 bar
28
Cy2PF-PCy2
PPF-P(t-Bu)2
100
54
MeOH/Toluene 20
With [Rh(nbd)2Cl]2,
room temperature,
precipitate
4
4
Cy2PF-PCy2
PPF-P(t-Bu)2
MeOH
MeOH
20
20
8
5
17
6
Reaction conditions: [Rh(nbd)]2]BF4, 60ꢀC, 100 bar, s/c ˆ 20
out; the best results are shown in Table 2. Here, ees were generally lower, and best
ees were observed with bdpp and diop, whereas binap and the ferrocenyl
diphosphines were much less effective. In all cases, signi®cant amounts of the
stabilized 1,4,5,6-tetrahydro intermediate (column ``part.'' in Table 2) were observed.
For the 2-picolinic acid 3 fewer ligands were screened, but the effect of varying
p, T and s/c as well as of some additives was investigated. ees and activities were
similar to those of the corresponding ester 1. Quite surprisingly, Cy2PF-PPh2 turned
out to be the most selective ligand by far. Changing the s/c ratio only affected the
catalyst activity but not the ee, whereas lowering pressure and temperature lowered
the activity strongly and the enantioselectivity somewhat. Addition of triethyl amine
or tri¯uoroacetic acid signi®cantly decreased the ees without affecting activity very
much (results not shown). The hydrogenation of nicotinic acid (4) turned out to be

Hydrogenation of Pyridines and Furans
1339
Table 4. Ligand screening for the hydrogenation of 5 and 6
Ligand
Solvent
Time
h
Yield
%
ee
Comments
%
5
5
5
5
6
6
6
6
binap
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
MeOH
19
20
20
20
20
20
19
19
91
98
98
93
3
7
5
Precipitate
Precipitate
Precipitate
prophos
chiraphos
diop
4
1
PPF-P(t-Bu)2
diop
24
4
4
Cy2PF-PCy2
xyliphos
100
85
1
0
Reaction conditions: [Rh(nbd)2]BF4, 60ꢀC, 100 bar, s/c ˆ 20
very dif®cult, and due to the formation of oligomeric side products, very low yields
of nipecotinic acid resulted. In some cases, the tetrahydro intermediate was also
observed. Best ees (17%) were observed with Cy2PF-PCy2.
The activity of the different Rh-ligand complexes for the hydrogenation of the
two furan derivatives 5 and 6 was investigated only brie¯y. Whereas 5 was reduced
with high conversions with most ligands, this was mainly due to decomposition of
the catalysts to black, heterogeneous Rh metal deposits and subsequent hetero-
geneous hydrogenation with low ees. In contrast, results were just the opposite with 6
where very low conversions but ees up to 24% were obtained using PPF-P(t-Bu)2,
whereas-somewhat surprisingly-other ferrocenyl diphosphines gave high conversion
but no induction, despite the fact that the reaction mixture remained homogeneous.
Comparison of Results
Generally, heterogeneous catalysts are much better suited for the hydrogenation of
aromatic rings [13], but as already pointed out, up to now ees with chirally modi®ed
heterogeneous catalysts were below 6% [7±9]. For the homogenous hydrogenation of
aromatic compounds only very few results have been reported and so far none of the
catalytic systems has turned out useful for preparative applications [14±19] because
of low activity and productivity. Of particular importance are catalysts based on
allyl-cobalt complexes developed by Mutterties [14, 15], the C5Me5-Rh catalysts of
‡
Maitlis [19], and ± for our purposes ± Rh-dppe complexes as studied by Halpern
[17]. The patent results with the homogeneous Rh complexes described by Fuchs [6]
had not yet appeared when we began our study. Since homogeneous catalysts are
generally more promising for enantioselective transformations than heterogeneous
catalysts [20], we decided to investigate the enantioselective hydrogenation of
heteroaromatic compounds. The reason for this choice was that these are usually
easier to reduce than the analogous homoaromatic derivatives [13]. As catalysts we
chose compounds formed in situ from several noble metal precursors and a series of
chiral diphosphines (see Scheme 2) with a good track record for the enantioselective
hydrogenation of C=C, C=O, and C=N bonds [21]. These ligands cover a broad

1340
M. Studer et al.
Fig. 1. Best ees for various ligand-substrate combinations; ligands are sorted by increasing chelate
size and decreasing basicity
range of properties concerning chelate size (5±7), ¯exibility (diop vs. binap), and
steric and electronic properties (ferrocenyl diphosphines [22]).
The results presented in the preceding section and summarized in Fig. 1 show
that none of the 9 ligands gave more than 24±27% ee in our attempts to hydrogenate
1±6. This compares well to the methyl ester of piperazine-2-carboxylic acid which
gave an ee of only 3.6% [6]. It remains to be seen if the corresponding t-butyl esters
and amides of pyridines 1 and 2 give much higher ees as claimed for the
corresponding pyrazines (78% ee for the corresponding t-butyl ester [6]). In
addition to the low ees, tons and tofs were also rather low, similar to those reported
in Ref. [6]. In short, we have by no means developed a synthetically useful method.
On the other hand, we have shown that Rh diphosphine complexes are indeed able
to hydrogenate monosubstituted pyridines and furans at relatively low temperatures.
In addition, moderate but signi®cant and reproducible ees have been observed
for most of the substrates tested. The chiral induction indicates that the main
reaction path is indeed homogeneous. A more detailed comparison of ees and
conversions for the 9 ligands tested is presented in Fig. 2. Ligands are grouped
for increasing chelate size (prophos, chiraphos, bdpp, diop, and binap) and
decreasing basicity for the 4 josiphos type ligands [22]. Our results indicate that
chelate sizes of 6 or 7 are preferable. A comparison of the results for binap and
diop shows that ligand ¯exibility is not an issue. The same holds for ligand
basicity: Each of the 4 ferrocenyl diphosphines led to the highest ee for a particular
substrate.

Hydrogenation of Pyridines and Furans
1341
Fig. 2. Best ees (®rst row, scale 30%) and conversions (second row, scale 100%) for various ligand-
substrate combinations; ligands are sorted by increasing chelate size and decreasing basicity; black
columns indicate Rh precipitate
Experimental
All regents and solvents were obtained from commerical suppliers and used without further
1
puri®cation. All products showed H and 13C NMR results corresponding to literature values.
Typical hydrogenation experiment
200 mg 1 (1.3 mmol) were degassed in a Schlenk tube at high vacuum and put under argon. After the
addition of 10 cm3 freshly distilled absolute ethanol, the substrate solution was transferred under
argon to a 50 cm3 autoclave (magnetically stirred, equipped with baf¯es, thermostated, 60ꢀC).
24.7 mg [Rh(nbd)2BF4 (0.066 mmol) and a diphosphine ligand (0.066 mmol) were dissolved in
10 cm3 EtOH with the same technique and transferred to the autoclave (20 cm3 total EtOH volume).

1342
M. Studer et al.
The hydrogenation was run at 60ꢀC and 100 bar hydrogen pressure for 22 h. With dppe as ligand,
100% conversion was obtained. The reaction product contained 96% of the pipecolic acid ethyl ester.
The conversion was determined by GLC (see below), and the ee was determined by NMR after
derivatization with Mosher's acid chloride (see below). All other substrates were hydrogenated in the
same apparatus under similar conditions.
Conversion was determined by 1H NMR (2, 3, 4, and 6) or GLC analysis (1, 2 [5], and 5, Varian
Star 3700 with a packed OV 101 column (2 m) and a FID detector, injector temperature 250ꢀC,
detector temperature 280ꢀC). For 1, the temperature program was 100±250ꢀC, 10ꢀC/min, and the
retention times were 6.5 min for 1 and 5.0 min for the product. For 5, the temperature program was
80±200ꢀC, 5ꢀC/min, and the retention times were 3.7 min for 5 and 4.5 min for the product. The ees
for the hydrogenation products of 1±4 were always determined after conversion to the corresponding
Mosher amides (see Ref. [5] for details with 2 and 4).
ee determination for the hydrogenation product of 1 [23]
10 mg hydrogenation product (0.064 mmol) were dissolved in 1 cm3 absolute CH2Cl2 and treated
with 26 mm3 Et3N (0.19 mmol) and 14.5 mm3 (R)-(±)-Mosher's acid chloride (0.076 mmol). After
12 h at room temperature the mixture was ®ltered over cotton, evaporated, and analyzed by 1H NMR
spectroscopy. The two diastereomic protons at C-2 of the ester showed doublets at 5.55 (L-ester) and
5.35 ppm (D-ester). The ee of the hydrogenation product of 3 was determined with the same method
after esteri®cation with EtOH.
ee determination for the hydrogenation product of 5
210 mg hydrogenation product (0.098 mmol) were dissolved in 1 cm3 absolute CH2Cl2. After the
addition of 40.0 mm3 Et3N (0.29 mmol), 1 crystal of N,N-dimethylaminopyridine, and 11.1 mm3
Ac2O (0.12 mmol), the mixture was reacted 3 h at room temperature and analyzed by GLC (Carlo
Erba GC 6000 with an FID detector and a 30 m ꢀ-Dex 110 column supplied by Supelco, helium
carrier, 95ꢀC isotherm, injector temperature 250ꢀC, detector temperature 270ꢀC). Enantiomer a was
observed after 19.3 min, enantiomer b after 19.7 min.
ee determination for the hydrogenation product of 6
10 cm3 absolute MeOH were cooled with an ice bath to 4ꢀC. 1 cm3 thionyl chloride (12.8 mmol) was
added dropwise under stirring. The ice bath was removed, and the mixture was stirred for 15 min and
then cooled to 4ꢀC again. After the addition of 100 mg of the hydrogenation product (0.79 mmol), the
ice bath was removed, and the reaction mixtures was stirred for 12 h at room temperature. The ee of
the methyl ester was determined on the same system as for 5. Enantiomer a was observed after
9.9 min, enantiomer b after 10.2 min.
Acknowledgments
The authors would like to thank S. Burkhardt for technical support.
References
[1] Thom C, Kocienski P, Jarowicki K (1993) Synthesis 475
[2] Hanessian S (1983) Total Synthesis of Natural Products: The ``Chiron'' Approach. Pergamon,
Oxford
[3] Kazlauskas RJ, Weiss¯och ANE, Rappaport AT, Cuccia LA (1991) J Org Chem 56: 2656

Hydrogenation of Pyridines and Furans
1343
[4] Fuchs R, Roduit JP (1996) EP 0744401A2 961127 (assigned to Lonza AG)
È
[5] Blaser HU, Honig H, Studer M, Wedemeyer-Exl C (1999) J Mol Catal A: Chemical 139: 253
[6] Fuchs R (1997) EP 0803502 A2 (assigned to Lonza AG)
[7] Laine RM, Hum G, Wood BJ, Dawson M (1981) Stud Surf Sci Catal 7: 1478
È
[8] Exl C, Ferstl E, Honig H, Rogi-Kohlenprath R (1995) Chirality 7: 211
[9] Nasar K, Fache F, Lemaire M, Beziat JC, Besson M, Gallezot P (1994) J Mol Catal 87: 107
[10] Besson M, Blanc B, Champelet M, Gallezot P, Nasar K, Pinel C (1997) J Catal 170: 254
[11] Besson M, Gallezot P, Neto S, Pinel C (1998) Chem Comm 1431
[12] Ranade VS, Consiglio G, Prins R (1998) Catal Lett 25: 71
[13] Rylander PN (1985) Hydrogenation Methods, Academic Press, London
[14] Muetterties EL, Bleeke JR (1979) Acc Chem Res 12: 324
[15] Stuhl LS, Rakowski DuBois M, Hirsekorn FJ, Bleeke JR, Stevens AE, Muetterties EL (1978) J
Am Chem Soc 100: 2405
[16] Rajca I (1981) Polish Journal of Chemistry 55: 775
[17] Landis CR, Halpern J (1983) Organometallics 2: 840
[18] Rothwell IP (1997) Chem Comm 1331
[19] Maitlis PM (1978) Acc Chem Res 11: 301
[20] Jacobsen EN, Pfaltz A, Yamamoto H (eds) (1999) Comprehensive Asymmetric Catalysis.
Springer, Berlin
[21] Blaser HU, Pugin B, Spindler F (1996) Enantioselective Synthesis. In: Cornils B, Herrmann WA
(eds) Applied Homogeneous Catalysis by Organometallic Complexes. Verlag Chemie, Wein-
heim, p 992
[22] Togni A, Breutel C, Schnyder A, Spindler F, Landert H, Tijani A (1994) J Amer Chem Soc 116:
4062
[23] Ng-Young-Chen, MC, Serreqi AN, Huang Q, Kazlauskas R (1994) J Org Chem 2075
Received April 5, 2000. Accepted (revised) May 2, 2000