Direct reductive amination versus hydrogenation of intermediates - A comparison

Article abstract of DOI:10.1002/adsc.200303208

The direct reductive amination of acetophenone with benzylamine or piperidine was studied in comparison with the hydrogenation of possible intermediates like a corresponding imine or enamine. No common features in terms of productivity and stereo-control (in the case of chiral catalysts) have been found for both processes. Hence evaluation of efficient, selective and enantioselective catalysts for direct reductive animation appears to be a separate task.

Full text of DOI:10.1002/adsc.200303208

FULL PAPERS
Direct Reductive Amination versus Hydrogenation of
Intermediates A Comparison
Vitali I. Tararov,a,* Renat Kadyrov,b Thomas H. Riermeier,b Christine Fischer,a
Armin Bˆrnera, c*
a
Leibniz-Institut f¸r Organische Katalyse e.V., Buchbinderstr. 5/6, 18055 Rostock, Germany
Degussa AG, Projekthaus Katalyse, Industriepark Hˆchst, G 830, 65926 Frankfurt/Main, Germany
Fachbereich Chemie der Universit‰t Rostock, A.-Einstein-Str. 3a, 18059 Rostock, Germany
b
c
Fax: (þ49)-381-4669324, e-mail: vitali.tararov@ifok.uni-rostock.de, armin.boerner@ifok.uni-rostock.de
Received: November 17, 2003; Accepted: March 11, 2004
Dedicated to Professor Reinhard Schmutzler on the occasion of his 70th birthday.
Abstract: The direct reductive amination of acetophe- cient, selective and enantioselective catalysts for di-
none with benzylamine or piperidine was studied in rect reductive amination appears to be a separate
comparison with the hydrogenation of possible inter- task.
mediates like a corresponding imine or enamine. No
common features in terms of productivity and stereo- Keywords: amination, enantioselective reduction, ho-
control (in the case of chiral catalysts) have been mogeneous catalysis, hydrogen, hydrogenation, rhodi-
found for both processes. Hence evaluation of effi- um, P ligands
Introduction
Recently, the Novartis group reported the first example
of an enantioselective DRA as part of the total synthesis of
The direct reductive amination (DRA) of aldehydes and the grass herbicide Metolachlor.[4a] In comparison with the
ketones is an important reaction in organic chemistry hydrogenation of a corresponding imine, representing a
with a great potential for application in industry.[1] Par- potentialintermediate of the reductiveamination, they ob-
ticularly interesting are those DRA processes which em- served the same configuration and nearly the same degree
ploy catalytically activated dihydrogen. The decisive ad- of enantioselectivity in the product. However, in all trials
vantage of such hydrogenation reactions to date main- the best productivity of the catalyst remained 100 times
ly carried out with heterogeneous catalysts[2] over the lower than in the reduction of the imine.
use of other reducing agents, e.g., boranes,[1a] consists
To the best of our knowledge no other literature re-
in the avoidance of any waste production. Undoubtedly ports concerning such comparisons are available. We be-
such environmentally friendly processes belong to lieve that such a comparison must be important since it
™green chemistry∫. Recently, evidence was given that can help to select the optimal strategy for the achieve-
DRA can also be successfully catalyzed under smooth ment of high chemical yields and, in the case of an asym-
conditions by homogeneous metal complexes.[3] More- metric synthesis, to produce highest ees.
over, we and others have provided proof that even an
asymmetric version of DRA using homogeneous chiral
Ir(I),[4] Rh(I)[5] or Ru(II)[6] catalysts bearing chiral P-li-
gands is possible, hence the potential of this methodolo-
gy has been multiplied.
From the practical point of view DRA reactions could
be superior to the hydrogenation of appropriate inter-
mediates, e.g., imines or enamines (indirect reductive
amination, IRA) since the first step of the synthesis
and the isolation of the unsaturated N-substrate is
spared (Scheme 1). But this superiority may be ques-
tionable because of the changing performance of the
same catalyst in DRA and in the corresponding IRA re-
actions.
Scheme 1. General approaches for reductive aminations.
Adv. Synth. Catal. 2004, 346, 561 565
DOI: 10.1002/adsc.200303208
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
561

Vitali I. Tararov et al.
FULL PAPERS
A few years ago we investigated the Rh(I)-catalyzed
enantioselective hydrogenation of imine 1[7] and enam-
ine 2[8] derived from acetophenone and benzylamine
and piperidine, respectively. Both substrates can be con-
sidered as probable intermediates of a DRA reaction.
Figure 1. Influence of the temperature on conversion and
amine selectivity of the DRA of PhCOMe with BnNH2 {con-
ditions: 5 mmol PhCOMe, 5 mmol BnNH2, 0.01 mmol
[Rh(dppb)COD]BF4, 10 mL MeOH, 50 bar initial H2 pres-
sure (measured at RT), exposition time 9 h}.
In order to clarify the relation of DRA to these hydro-
genations, we have now investigated the Rh(I)-cata- tophenone (5 mmol) and BnNH2 in MeOH (10 mL) at
lyzed reductive amination of acetophenone with benzyl- 50 bar H2 pressure in the presence of 0.2 mol % of
amine and piperidine, respectively. Herein these results [Rh(dppb)COD]BF4 gave poor conversion (ca. 10%)
will be detailed and compared with the data mentioned and low selectivity in favour of the production of the
above.
amine 3 (ca. 9%). Obviously, the formation of an inter-
mediate, presumably 1, is the rate-determining step in
this DRA. We were not able to detect by 1H NMR spec-
troscopy even traces of imine 1 (as well as other N-con-
taining intermediates) in the model system in MeOH-d4
Results and Discussion
The reductive amination of acetophenone can be sche- at room temperature. In contrast, hydrogenation of
matically represented as shown in Scheme 2.
imine 1 under the same conditions and catalyst loading
was rather fast. Complete conversion of imine 1 and ex-
clusive formation of amine 3 was observed within 1 h.
In order to improve the efficiency and the selectivity
of the DRA process we studied the effect of the temper-
ature. The data are presented in Figure 1. As can be seen
in the range of 70 1008C there is almost no dependence
of the conversion on the temperature. But the increase
in the temperature has a strong effect on the selectivity.
Between 90 1008C the formation of the amine is fav-
oured over the production of the undesired alcohol.
Scheme 2. Reductive amination of acetophenone.
At
1008C
two
other
achiral
catalysts,
Usually the main side reaction in DRA is the forma- [Rh(dppp)COD]BF4 [dppp¼1,3-bis(diphenylphospha-
tion of an alcohol. This side reaction imposes additional nyl)propane] and [Rh(dpoe)COD]BF4 (dpoe¼1,2-di-
demands on the catalyst performance. Therefore, in phenylphosphanyloxyethane), were tested. The com-
DRA processes not only the productivity (and the enan- parison of conversion and selectivity of these three cat-
tioselectivity in the case of the asymmetric version) but alysts is illustrated in Figure 2.
also the selectivity in favour of the formation of the
The data depicted in Figure 2 clearly show that the ac-
amine is important.[3a] The selectivity can be expressed tivity of [Rh(dppp)COD]BF4 bearing a 6-membered
as the simple [amine]/[alcohol] ratio or as a quota of chelate ring is approximately the same as that of
an amine, [amine]/[amine]þ[alcohol]. Unfortunately [Rh(dppb)COD]BF4 representing a 7-membered che-
this characteristic of the catalysts is still not yet accepted late. When the diphosphine dppb was substituted by
generally, thus making it impossible to compare differ- the diphosphinite ligand dpoe also forming a seven-
ent catalysts.
membered chelate ring the activity of the catalyst was
significantly increased. The selectivity is dependent on
the structure of the ligand in the order
[Rh(dppp)COD]BF4 <[Rh(dppb)COD]BF4 <[Rh(dpoe)
COD]BF4. This order corresponds to the decrease of the
steric hindrance of the ligands.[9] Moreover, also the
Direct Reductive Amination of Acetophenone with
Benzylamine
First, in order to find appropriate conditions for DRA, Lewis basicity of ligands forming 7-membered chelates
the achiral precatalyst [Rh(dppb)COD]BF4 [dppb¼ might play an important role.
1,4-bis(diphenylphosphanyl)butane
(formula
see
For asymmetric DRA, we tested as a precatalyst
Fig. 2); COD¼1,5-cyclooctadiene] was tested. At ambi- {Rh[(R,R)-bdpch]COD}BF4 {for the structure of the li-
ent temperature, hydrogenation of a 1:1 mixture of ace- gand (R,R)-bdpch,[10] see Table 3}. This diphosphinite
562
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 561 565

Direct Reductive Amination versus Hydrogenation of Intermediates
FULL PAPERS
cohol 5 is slightly dependent on the temperature. Inter-
estingly, the enantioselectivity of amine 3 varies dra-
matically with the temperature. Thus, the configuration
changes from (R) at 508C to (S) at 80 1008C. At 1008C
the selectivity and enantioselectivity do not depend on
the catalyst loading which is in the range of 0.2
0.8 mol %. Only an increase of the conversion to 60%
and 84% was observed with 0.4 mol % and
0.8 mol %, respectively.
In comparison to the DRA of PhCOMe with BnNH2,
the asymmetric hydrogenation of imine 1 which can be
regarded as a possible intermediate of this transforma-
tion occurred at room temperature and 50 bar initial
H2 pressure with {Rh[(R,R)-bdpch]COD}BF4 as a pre-
catalyst (0.01 mmol) in MeOH and gave amine (R)-3
in 71% ee.[7] Five mmol of imine 1 were quantitatively
reduced under these conditions in 10 mL of MeOH
within 5 h. At 1008C and 10 h reaction time the conver-
sion of imine 1 dropped to 88%. Simultaneously, the ee
decreased to 12% but nevertheless the (R)-configura-
tion in the product amine 3 remained.
Figure 2. Comparison of the performance of different cata-
lysts in the DRA of PhCOMe with BuNH2. 1:
[Rh(dppp)COD]BF4; 2: [Rh(dppb)COD]BF4; 3: [Rh(dpoe)-
COD]BF4 at 1008C (for other conditions see Fig. 1).
The inversion of the configuration of amine 3 with the
temperature produced in DRA is difficult to rationalize.
Taking into account that the configuration and the ee of
alcohol 5 is not dependent on the temperature, it is rea-
sonable to assume that the catalytically active Rh(I) spe-
cies is quite the same in the temperature range of 50
1008C. Hence we can assume that imine 1 is not an inter-
mediate in the DRA considered herein and that other
species produced from PhCOMe and BnNH2 are hydro-
genated at elevated temperatures. Another possibility
Figure 3. Influence of the temperature on the catalytic per-
formance of {Rh[(R,R)-bdpch]COD}BF4 in the asymmetric
DRA of PhCOMe with BnNH2 (for conditions see Fig. 1).
Positive ee values correspond to the (R)-configuration and
negative to the (S)-configuration.
complex exhibited the highest enantioselectivity among which could not be experimentally confirmed up to
others Rh precatalysts in the hydrogenation of the imine now is the temperature dependent (Z)/(E) isomeriza-
1.[7] The performance of this precatalyst is represented in tion of imine 1.
Figure 3.
Three other catalysts have been tested in the reductive
All parameters are dependent on the temperature. amination of PhCOMe with BnNH2. Relevant data are
Although conversions measured at 808C and 1008C collected in Table 1.
are fast and quite the same, the selectivity of DRA is
As clearly seen again, a diphosphinite complex derived
higher at 1008C. The same trend can be deduced from Kbþ-OH as ligand shows enhanced activity and selec-
from Figure 1. The enantioselectivity induced in the al- tivity in comparison with two diphosphine complexes.
Table 1. Comparison of the catalytic performance in asymmetric DRA of PhCOMe with BnNH2 and hydrogenation of imine 1
with precatalysts of the type [Rh(Ligand)COD]BF4.[a]
Run
Ligand
Conversion of
PhCOMe [%]
Selectivity
[%]
ee of amine 3
[%]
ee of amine 3
[%] from
imine 1 (RT)
1
2
3
(R,R)-DIOP[b]
32.2
59.3
69.4
32
43
66
2 (R)[e]
6 (R)[f]
19 (R)[h]
35 (R)
6[c]
K
bþ-OH[d]
20 (R)[g]
28 (S)[h]
[a]
Structure of the ligands, see Table 3; conditions: 5 mmol PhCOMe, 5 mmol BnNH2, 0.01 mmol [Rh(Ligand)COD]BF4,
10 mL MeOH, 50 bar initial H2 pressure (measured at RT), 1008C, exposition time 8 h.
[b]
[c]
[d]
Ref.[11]
Ref.[12]
Ref.[13]
[e-g]
[h]
[e]
[f]
[g]
ee (configuration) of alcohol 5: 4% (R); 15% (R); 5% (R).
Ref.[7]
Adv. Synth. Catal. 2004, 346, 561 565
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
563

Vitali I. Tararov et al.
FULL PAPERS
With the latter the enantioselectivity of amine 3 produc- properties of the ligands (Table 3, compare runs 1 and
tion in the direct process issignificantly reduced in compar- 2 and 3 5). Although the catalysts tested display moder-
ison with the hydrogenation of imine 1. In the case of the ate enantioselectivity in the hydrogenation of enamine 2
K
bþ-OH catalyst the configuration of the amine 3 is oppo- they failed completely in the corresponding DRA reac-
site in the IRA.
tion. Meanwhile, precatalysts Kbþ and Kbþ-OH display
moderate enantioselectivity in the undesired reduction
of PhCOMe affording the alcohol.
Direct Reductive Amination of Acetophenone with
Piperidine
In the DRA of PhCOMe with piperidine (Scheme 2),
enamine 2 can be assumed as an intermediate for the for-
mation of amine 4. The reduction of this substrate in
MeOH at 50 bar H2 initial pressure with
[Rh(dppb)COD]BF4 as a precatalyst occurred at room
temperature within less than 1 h with quantitative for-
mation of amine 4 [enamine 2 can be hydrogenated
also at 1 atm H2 pressure with a variety of other Rh(I)
precatalysts].[8]
Under the same conditions the hydrogenation of a 1:1
mixture of MeCOPh (5 mmol) and piperidine (5 mmol)
gave only 25% of conversion of ketone and alcohol 5 as
the main product after 20 h (Table 2, run 1). Raising the
temperature increased the yield of amine 4 in DRA
(runs 2 4) but, nevertheless, the formation of alcohol
5 is significant. It is interesting to note that an increase
in the temperature from 80 to 1008C does not have a
pronounced influence on conversion and selectivity
(runs 3 and 4). In contrast to the DRAwith benzylamine
in case of piperidine the use of the diphosphinite preca-
talyst does not improve the selectivity although the con-
version increases (run 5). In general, the efficiency of
DRA is not comparable with the hydrogenation of en-
amine 2.
Conclusions
In comparison with the hydrogenation of selected isolat-
ed N-intermediates in the corresponding DRA reac-
tions, homogeneous Rh(I) precatalysts display signifi-
cantly lower activity. For a successful production of an
amine by DRA, elevated temperatures are required.
In addition, in DRA the side reaction giving rise to the
undesired alcohol is significant thus complicating the
purification of the amine. It should be noted that inves-
tigations on the DRA of benzaldehyde with pyrrolidine
at room temperature revealed that water might not play
a significant role in the course of the reaction.[1b] In con-
trast to the results of the Novartis group observed in the
Metolachlor synthesis mediated by an Ir(I) catalyst,[4a]
we found no parallel in the degree and the sign of the
asymmetric induction in the hydrogenation of single in-
termediates and the corresponding DRA. Hence evalu-
ation of efficient selective and enantioselective catalysts
for DRA seems to be a separate task. Obviously in some
cases first isolation of intermediates, e.g., imines, enam-
ines or even sometimes N,O-acetals[15] and subsequent
hydrogenation is preferred over DRA. On the other
hand, the DRA approach presents the only possibility
to produce amines in those cases when intermediates
are not stable as has been proven in the reductive amina-
tion of a-keto acids.[5] In these investigations striking dif-
ferences in the enantioselectivities in dependence on the
nature of the prochiral substrate were noted.
Only poor results were obtained in the asymmetric
version of this DRA reaction. The comparison with en-
amine 2 hydrogenation is given in Table 3.
As clearly shown, diphosphine complexes as catalysts
in DRA display lower activity in comparison with di-
phosphinite complexes (Table 3, compare runs 1 and 2
and 3 5). This behaviour in DRAwith piperidine is sim-
ilar to the DRA with BnNH2. But the selectivity in the
former process does not correlate with the electronic
Experimental Section
MeCOPh, BnNH2 and piperidine were distilled and kept under
an Ar atmosphere. All hydrogenations were carried out ac-
cording to the protocol detailed in Ref.[1b] After the hydrogena-
tion the reactions mixtures were analyzed by NMR, HPLC and
GC on chiral columns. The conversions and selectivities were
estimated on the basis of 1H NMR spectra. The signals with fol-
lowing chemical shifts (in ppm, measured in CDCl3 relative to
TMS) were evaluated: 1.24 (d, J¼6.5 Hz; CH3 group of amine
3), 1.33 (d, J¼6.3 Hz; CH3 group of alcohol 5), 1.35 (d, J¼
6.7 Hz; CH3 group of amine 4), 2.18 (s, CH3 group of imine
1), 2.48 (s, CH3 group of MeCOPh). The ee of amine 3 was de-
termined by HPLC on Chiralcel OD-H column (eluent: hex-
ane). The ees of amine 4 and alcohol 5 were determined by
GC on CP Chirasildex-CB.
Table 2. Direct reductive amination of PhCOMe with piper-
idine in 1: 1 ratio applying precatalysts of the type [Rh(Li-
gand)COD]BF4.[a]
Run Ligand Time Temp. Conversion of Selectivity
[h]
[8C]
PhCOMe [%] [%]
1
2
3
4
5
dppb
dppb
dppb
dppb
dpoe
20
23
10
10
10
25
50
80
100
100
25
46
41
39
86
9
17
36
39
34
[a]
For conditions see Table 1.
564
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 561 565

Direct Reductive Amination versus Hydrogenation of Intermediates
FULL PAPERS
Table 3. Comparison of DRA of PhCOMe with piperidine and hydrogenation of enamine 2 applying precatalysts of the type
[Rh(Ligand)COD]BF4.[a]
Run No
Ligand
Conversion of
PhCOMe [%]
Selectivity
[%]
ee of amine 4
[%]
ee of amine 4
[%] from
enamine 2a
1
2
3
4
5
(R,R)-DIOP
6
55
62
74
88
89
42
53
63
32
34
0[c]
39 (R)[h]
45 (R)[h]
32 (R)[h]
25 (S)[i]
36 (S)[h]
0[d]
(R,R)-bdpch
6 (R)[e]
0[f]
[b]
Kbþ
K
bþ-OH
4(R)[g]
[a]
Prepared at RT.
Ref.[14]
[b]
[c-g]
[h]
[c]
[d]
[e]
[f]
[g]
ee (configuration) of alcohol 5: 0%, 3% (R); 0%; 31% (S); 24% (S).
Reduced at 1 bar H2, see also Ref.[8]
Reduced at 50 bar H2.
[i]
2000, 41, 6583 6588; c) G. D. Williams, R. A. Pike, C. E.
Wade, M. Wills, Org. Lett. 2003, 5, 4227 4230.
[4] a) H.-U. Blaser, H.-P. Buser, H.-P. Jalett, B. Pugin, F.
Spindler, Synlett 1999, 867 868; b) Y. Chi, Y.-G. Zhou,
X. Zhang J. Org. Chem. 2003, 68, 4120 4122.
[5] R. Kadyrov, T. H. Riermeier, U. Dingerdissen, V. Tarar-
ov, A. Bˆrner, J. Org. Chem. 2003, 68, 4067 4070.
[6] R. Kadyrov, T. H. Riermeier, Angew. Chem. 2003, 115,
5630 5632; Angew. Chem. Int. Ed. 2003, 42, 5472 5474.
[7] V. I. Tararov, R. Kadyrov, T. H. Riermeier, J. Holz, A.
Bˆrner, Tetrahedron: Asymmetry 1999, 10, 4009 4015.
[8] V. I. Tararov, R. Kadyrov, T. H. Riermeier, J. Holz, A.
Bˆrner, Tetrahedron Lett. 2000, 41, 2351 2355.
[9] K. Angermund, W. Baumann, E. Dinjus, R. Fornika, H.
Gˆrls, M. Kessler, C. Kr¸ger, W. Leitner, F. Lutz,
Chem. Eur. J. 1997, 3, 755 764.
Acknowledgements
Financial support from Degussa AG, Projekthaus Katalyse
(Frankfurt, Germany), the European Commission (Descartes
Prize 2001) and the Fonds der Chemischen Industrie is grateful-
ly acknowledged.
References
[1] a) A. F. Abdel-Magid, K. G. Carson, B. D. Harris, C. A.
Maryanoff, R. D. Shah, J. Org. Chem. 1996, 61, 3849
3862 and references cited therein; b) V. I. Tararov, R. Ka-
dyrov, T. H. Riermeier, A. Bˆrner, Adv. Synth. Catal.
2002, 344, 200 208; c) V. I. Tararov, R. Kadyrov, T. H.
Riermeier, U. Dingerdissen, A. Bˆrner, Org. Prep.
Proc. Int. 2004, in press.
[2] V. A. Tarasevich, N. G. Kozlov, Russ. Chem. Rev. 1999,
68, 55 72; P. N. Rylander, Hydrogenation Methods; Aca-
demic Press: New York, 1985, pp. 82 93; S. Nishimura,
Handbook ofHeterogeneous Catalytic Hydrogenation
for Organic Synthesis, Wiley: New York, 2001.
[10] M. Tanaka, I. Ogata, Chem. Commun. 1975, 735 736.
[11] H. B. Kagan, T.-P. Dang, J. Am. Chem. Soc. 1972, 94,
6429 6433.
[12] U. Berens, D. Leckel, S. Oepen, J. Org. Chem. 1995, 60,
8204 8208.
[13] R. Selke, M. Ohff, A. Riepe, Tetrahedron 1996, 52,
15079 15102.
[3] a) V. I. Tararov, R. Kadyrov, T. H. Riermeier, A. Bˆrner,
[14] R. Selke, H. Pracejus, J. Mol. Catal. 1986, 37, 213 225.
[15] V. I. Tararov, R. Kadyrov, A. Monsees, T. H. Riermeier,
A. Bˆrner, Adv. Synth. Catal. 2003, 345, 239 245.
Chem. Commun. 2000, 1867 1868; b) R. Margalef-Cata-
¡
¬
la, C. Claver, P. Salagre, E. Fernandez, Tetrahedron Lett.
Adv. Synth. Catal. 2004, 346, 561 565
asc.wiley-vch.de
¹ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
565