Ru-(phosphine-oxazoline) complexes as effective, industrially viable catalysts for the enantioselective hydrogenation of aryl ketones

Article abstract of DOI:10.1002/adsc.200505246

Complexes prepared in situ from RuCl₂(PPh₃) ₃ and chiral phosphine-oxazoline ligands are effective catalysts for the hydrogenation of various aryl ketones with ees up to 99% and substrate to catalyst ratios of 10,000-50,000; the reaction tolerates high substrate concentrations and a pilot process has been developed for the hydrogenation of 3,5-bistrifluoromethyl acetophenone.

Full text of DOI:10.1002/adsc.200505246

COMMUNICATIONS
DOI: 10.1002/adsc.200505246
Ru-(Phosphine-Oxazoline) Complexes as Effective, Industrially
Viable Catalysts for the Enantioselective Hydrogenation of Aryl
Ketones
a,
a
a
b
´ ´
Frederic Naud, * Christophe Malan, Felix Spindler, Carsten Rüggeberg,
Andreas T. Schmidt,b Hans-Ulrich Blasera
Solvias AG, P. O. Box, 4002 Basel, Switzerland
a
Fax: (þ41)-61-686-6311; e-mail: frederic.naud@solvias.com
b
Rohner AG, Gempenstr. 6, 4133 Pratteln, Switzerland
Received: June 15, 2005; Accepted: October 13, 2005
Supporting Information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
¼
Abstract: Complexes prepared in situ from
RuCl2(PPh3)3 and chiral phosphine-oxazoline ligands
are effective catalysts for the hydrogenation of vari-
ous aryl ketones with ees up to 99% and substrate to
catalyst ratios of 10,000–50,000; the reaction toler-
ates high substrate concentrations and a pilot process
has been developed for the hydrogenation of 3,5-bis-
trifluoromethyl acetophenone.
Ru N without coordinating the carbonyl group to the
Ru center.[3]
Chiral oxazoline-based ligands have attracted a lot of
attention in the past decade due to their ease of prepara-
tion, their modularity and their favorable coordinating
properties towards a large variety of metals. Indeed,
high levels of enantioselectivities have been obtained
for various catalytic reactions.[5] Of special interest to
us were several Ru-phosphine-oxazoline complexes
(see Figure 1) described by the groups of Helmchen,[6]
Sammakia[7] and Uemura[8] which are able to transfer
hydrogenate aryl ketones with ees>95% using i-
PrOHas reducing agent. Although attractive for its sim-
plicity and high enantioselectivity, this method has a lim-
ited synthetic potential due to low ketone concentra-
tions, low turnover numbers (TON¼1000) and erosion
Keywords: aryl ketone hydrogenation; enantioselec-
tive hydrogenation; industrial catalysis; Ru-(phos-
phine-oxazoline) complexes
The enantioselective reduction of aryl ketones is an im- of ee at long reaction times.
portant transformation both from an academic/syn-
Due to their very high enantioselectivity we were
thetic as well as an industrial point of view.[1] Besides bi- particularly interested in the ferrocene-based Ru
ocatalytic and hydride reduction methods, two differ- complexes which can also be created in situ from
ent effective hydrogenation methodologies have been RuCl2(PPh3)3 with somewhat lower enantioselectivities
developed: i) transfer hydrogenations, where the re- but better activities.[7] During our studies with several
ducing agent is a hydrogen donor such as formic acid aryl ketones under typical transfer hydrogenation con-
or isopropanol,[2] and ii) the catalytic addition of gas- ditions we surprisingly found that catalyst activity (ex-
eous H2 to the ArCO double bond.[3] Interestingly, pressed as turnover frequency, TOF, hꢀ1) and TONs in-
with few exceptions the two methods require different creased significantly when the reaction was carried out
types of catalysts.[4] While a few Ru complexes are able
¼
to reduce ArC O groups with hydrogen as well as hy-
drogen donors, there is no catalytic system with high
ees and high activity for both methods. In addition,
an examination of systems able to catalyze aryl ketone
reductions shows that most of the effective catalysts for
hydrogenation as well as transfer hydrogenation have
ꢀ
at least one ligand with an N Hfunction. Noyori has
convincingly explained this observation by postulating
¼
ꢀ
ꢀ ꢀ
the direct addition of the C O bond to a H Ru N H
moiety to give the reduced substrate H C O Hand
ꢀ ꢀ ꢀ
Figure 1. Structures of chiral ligands and Ru-complexes.
Adv. Synth. Catal. 2006, 348, 47 – 50
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
47

´ ´
Frederic Naud et al.
COMMUNICATIONS
Figure 3. Reaction conditions for hydrogenation.
Table 1. Effect of ligand and ketone structure on ee.[a]
1
2
3
4
5
A
C
D
E
F
G
H
J
98.1
95.8
98.5
97.2
95.0
92.9
86.7
96.9
94.7
93.6
92.9
90.2
94.3
93.2
72.6
90.7
95.8
94.9
96.6
96.4
94.1
91.9
36.7
95.0
95.0
95.6
97.5
95.7
84.4
88.4
82.4
95.3
98.4
98.3
99.3
98.0
96.2
93.7
91.6
96.2
Figure 2. Structure of tested ligands and aryl ketones.
[a]
Reaction conditions: 0.01 mmol ligand and RuCl2(PPh3)3;
2 mmol substrate; 2 mL toluene; 1 mL 1 molar aqueous
NaOH; 80 bar; room temperature; reaction time 1 h.
under hydrogen pressure while ees were affected only
slightly. We therefore investigated the effect of ligand
and ketone structure as well as of other reaction param-
eters on catalyst performance in more detail. Here we that the same holds true for the nature of the phosphine
describe a new technology which allows the hydrogena- substituents. In general, the enantioselectivities are bet-
tion of various aryl ketones with ees up to 99% without ter than 90% ee and in some cases are >98% (bold). For
erosion at substrate to catalyst (S/C) ratios of 10,000– most substrates ee values as well as the conversions vary
50,000, tolerates high substrate concentrations and has only a few percent when R (compare Awith C and D) or
been scaled up to the pilot level.
Ar are changed (compare A with E–G). Interestingly,
Since enantioselective catalysts often have high spe- the replacement of the ferrocenyl moiety by an aryl or
cificity, we not only prepared a number of different li- heteroaryl group also has a surprisingly small effect on
gands but tested these systematically on several aryl ke- the catalyst performance. With the exception of the
tones. Phosphine-oxazolines are modular ligands which commercially available ligand H which gives significant-
are stable towards hydrogenation and can easily be var- ly lower ees, ligand J as well as two other heteroaryl
ied both on the phosphine and the oxazolines moieties. based ligands (results see Supporting Information)
Besides the ferrocenyl-based derivatives, we also pre- have few values below 90% (underlined italic). Howev-
pared and tested ligands with an aryl and heteroaryl er, as Figure 4 clearly shows, the conversion after 1 h
backbone. Selected ligands as well as the investigated with these ligands was consistently lower than with li-
ketones are depicted in Figure 2.
gands A–G which in almost all cases was close to
The test hydrogenation reactions were carried out 100%. Regarding the effect of the substrate structure
with catalysts prepared in situ from RuCl2(PPh3)3 and we cannot see a clear trend for the activity of the cata-
the appropriate ligand under the conditions shown in lysts but it seems that substituents at the aryl ring gener-
Figure 3. Instead of the usual i-PrOH/i-PrOK solution ally lower the enantioselectivity anywhere from a few
we used a technically more practical toluene/aqueous percent for ligands A–E to up to 60% for the other li-
NaOHmixture which gives a similar catalyst perform-
gands. Substrate 6 (ee 97.2%) and 3-acetyl thiophene 9
ance. Under these reaction conditions, GLC analysis (ee 93.9%) were also tested using ligand A but in both
did not show any by-products due to base-catalyzed con- cases, catalyst activity was very low. Changing the meth-
densation. A hydrogen pressure of 80 bar was used in yl group in acetophenone to Et or even i-Pr (substrates 5
most tests but it was later shown that pressures as low and 6, respectively) does not affect enantioselectivity
as 20 bar afford the same catalytic performance.
very much but, as expected, conversions after 1 h are sig-
The influence of ligand and ketone structure on cata- nificantly lower.
lyst performance is summarized in Table 1 (ee) and Fig-
This new hydrogenation catalyst containing ligand A
ure 4 (conversion after 1 h) and needs few comments. As was also applied under industrially more relevant condi-
already observed for the transfer hydrogenation,[7,8] the tions. Three model reactions (substrates 1, 3 and 7) were
performance of the Ru complexes does not depend very carried out with S/C ratios of 10,000 to 50,000. The se-
much on the substituent R of the oxazolidine. We find lected results presented in Table 2 show that technically
48
asc.wiley-vch.de
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 47 – 50

Catalysts for the Enantioselective Hydrogenation of Aryl Ketones
COMMUNICATIONS
er bases were not suitable), temperature (only 85% ee at
758C, see entry 8), isolated vs. in situ formed catalyst (an
induction period was often observed for the isolated
complex but otherwise with the same catalyst perform-
ance) and purity of the substrate which turned out to
be the most important parameter affecting TON and
TOF. S/Cs of up to 50,000 were possible but required
very long reaction times and gave somewhat lower ees
(see entry 10). The optimized process was run twice
with 140 kg of substrate in a 4 m3 hydrogenation reactor
with an S/C ratio of 20,000 at 20 bar and 258C. The ee
values were 95 and 96%, respectively, and full conver-
sion was reached after 15 h.
In conclusion, we have shown that Ru-(phosphine-ox-
azoline) complexes are not only highly selective transfer
hydrogenation catalysts but are also very effective for
the hydrogenation of aryl ketones using H2 with ees of
95–99% and TONs of up to 50,000. This compares
well with several recently published catalysts for aryl ke-
tone hydrogenation such as Ru/phanephos/dpen[10] or
Ru/p-phos/dpen[11] which are also able to hydrogenate
Figure 4. Effect of ligand and ketone structure on conversion
(reaction conditions see Table 1).
feasible TONs and TOFs can be obtained with several various aryl ketones with high ees as well as high
substrates. The reaction can be run at very high concen- TONs and TOFs.
trations (entries 2 and 3), even though reaction times in-
In contrast to the catalyst system described by Noyori
crease significantly, there is no ee erosion (compare en- or the cited analogues thereof, the Ru-(phosphine-oxa-
ꢀ
tries 1 and 3). The isolated [RuCl2(A)(PPh3)] complex zoline) complexes do not contain an N Hgroup. Since
gives the same performance as the in situ prepared cata- the hydrogenation of the oxazoline is very unlikely un-
lyst (compare entries 5 and 6).
Furthermore, starting with substrate 8 we developed a ing reaction of the C O bond with an H Ru N Hmoi-
technical process for the production of 3,5-(CF3)2-phe- ety cannot be operative. While there is no experimental
der these conditions, this means that the pathway involv-
¼
ꢀ
ꢀ ꢀ
¼
nyl ethanol, a chiral building block for an NK-1 receptor evidence, a plausible alternative is a classical C O coor-
antagonist.[9] Details will be published elsewhere, here dination-insertion mechanism.[12] The function of the in-
we will only briefly describe our most important find- dispensable strong base could be to assist the heterolytic
ings. Since a limited ligand screening showed again activation of H2 typical for ruthenium catalysts. The ef-
only small variations in ee (91–95%), ligand Awas chos- fect of the substrate and ligand structure on the catalyst
en for further development. The following parameters performance is relatively small both for the various
were investigated and optimized: Nature and quantity phosphine-oxazolines as well as for the aryl alkyl ke-
of the base (aqueous NaOHwas clearly superior, weak-
tones. Since analogues of RuCl2(PPh3)3 or of isolated
Table 2. Effect of various parameters on the performance of A/RuCl2(PPh3)3.
Entry Substrate P [bar] S/C
Time [h] Conversion [%] ee [%] TOF [hꢀ1
]
Comments
1
2
3
4
1
1
1
3
7
7
8
8
8
8
8
80
80
80
20
20
20
20
20
20
20
20
10,000
10,000
50,000 78
20,000
20,000 19
20,000 19
500
500
10,000 2.3
50,000 92
20,000 15
1
6
98
98
99
92
100
99
100
100
100
99
98.5
98.5
99.0
96.2
96.0
97.4
95.0
85.0
96.0
94.3
96.0
9,800
1,633
635
18,400
1,050
1,040
500
500
4,500
500
50 mmol/18 mL toluene
50 mmol/no toluene
250 mmol/2 mL toluene
1
5
6[a]
7[a]
8[a, b]
9[a]
1
1
2.34 mmol
2.34 mmol
586 mmol/1.44 L toluene/60 mL NaOH
3.75 mol/0.35 L toluene/150 mL NaOH
547 mol; 4000-L reactor
10[a]
11
100
1,660
Reaction conditions: In situ catalyst with ligand A and RuCl2(PPh3)3; 100 mmol substrate; 9–12 mL toluene; 1 mL 1 molar
aqueous NaOH; 25 8C.
[a]
Isolated RuCl2(A)(PPh3) complex.
At 758C.
[b]
Adv. Synth. Catal. 2006, 348, 47 – 50
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
asc.wiley-vch.de
49

´ ´
Frederic Naud et al.
COMMUNICATIONS
[RuCl2(PN)(PPh3)] complexes with phosphines other References
than PPh3 are not easily accessible, the effect of this aux-
[1] a) T. Ohkuma, M. Kitamura, R. Noyori in: Catalytic
Asymmetric Synthesis, 2nd edn, (Ed.: I. Ojima), Wiley-
VCH, Weinheim, 2000, p. 1; b) T. Ohkuma, R. Noyori
in E. N. Jacobsen, H. Yamamoto, A. Pfaltz (Eds.), Com-
prehensive Asymmetric Catalysis, Springer, Berlin, 1999,
p 199.
[2] a) R. Noyori, S. Hashiguchi, Acc. Chem. Res. 1997, 30,
97; b) M. J. Palmer, M. Wills, Tetrahedron: Asymmetry
1999, 10, 2045; c) G. Zassinovich, G. Mestroni, S. Gladi-
ali, Chem. Rev. 1992, 92, 1051.
iliary ligand has not yet been investigated.
To the best of our knowledge this is the first system
which can both transfer hydrogenate and hydrogenate
aryl ketones with high ee and high efficiency and also
the first industrial catalytic process with an oxazoline
based ligand.
[3] R. Noyori, T. Ohkuma, Angew. Chem. Int. Ed. 2001, 40,
40.
Experimental Section
Ligands A–D were prepared as described in ref.[8b] and Jas de-
scribed in ref.[13] For the preparation of ligands E–G see Sup-
porting Information, ligand H is available from Strem. The cat-
alyst precursor RuCl2(PPh3) was prepared according to ref.[8a]
[4] V. Rautenstrauch, X. Hoang-Cong, R. Churlaud, K. Ab-
dur-Rashid„ R. H. Morris, Chem. Eur. J. 2003, 9, 4954.
[5] a) H. A. McManus, P. J. Guiry, Chem. Rev. 2004, 104,
4151; b) O. B. Sutcliffe, M. R. Bryce, Tetrahedron: Asym-
metry 2003, 14, 2297; c) G. Helmchen, A. Pfaltz, Acc.
Chem. Res. 2000, 33, 336.
[6] T. Langer, G. Helmchen, Tetrahedron Lett. 1996, 37,
1381.
[7] T. Sammakia, E. L. Stangeland, J. Org. Chem. 1997, 62,
6104.
[8] a) Y. Nishibayashi, I. Takei, S. Uemura, M. Hidai, Orga-
nometallics 1999, 18, 2291; b) Y. Arikawa, M. Ueoka, K.
Matoba, Y. Nishibayashi, M. Hidai, S. Uemura, J. Orga-
nomet. Chem. 1999, 572, 163.
[9] K. B. Hansen, J. R. Chilenski, R. Desmond, P. N. Devine,
E. J. J. Grabowski, R. Heid, M. Kubryk, D. J. Mathre,
R.Varsolona, Tetrahedron: Asymmetry 2003, 14, 3581.
[10] D. Chaplin, P. Harrison, J. P. Henschke, J. C. Lennon, G.
Meek, P. Moran, C. J. Pilkington, J. A. Ramsden, S. Wat-
kins, A. Zanotti-Gerosa, Org. Process Res. Devel. 2003,
7, 89.
General Hydrogenation Procedure
The catalyst is prepared in situ by allowing 0.1 mmol of ligand
and 0.1 mmol of [RuCl2(PPh3)3] in 20 mL of toluene to react for
2 h under reflux. The catalyst solution is transferred to the in-
ertized mini-autoclave and the substrate and the base are sub-
sequently added. The autoclave is sealed and pressurized with
hydrogen to the desired pressure. The reaction is started by
switching on the stirrer. When the hydrogenation time has
elapsed, the stirrer is switched off and the autoclave is ventilat-
ed. Conversion and ee are determined via GC analysis on a
Beta-Dex 110 column (for details see Supplementary Informa-
tion).
[11] J. Wu, J.-X. Ji, R. Guo, C.-H. Yeung, A. S. C. Chan,
Chem. Eur. J. 2003, 9, 2963.
Acknowledgements
[12] S. E. Clapham, A. Hadzovic, R. H. Morris, Coord. Chem.
Rev. 2004, 248, 2201.
[13] L. Tietze, J. K. Lohmann, Synlett 2002, 12, 2083.
We thank Andrea Holderer, Oanh Mai-Huynh, Stefan Modin,
Thomas Stebler and Horatio Tagliente for skillful laboratory
work, and Prof. L. Tietze for ligand samples.
50
asc.wiley-vch.de
ꢀ 2006 Wiley-VCHVerlag GmbH& Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 47 – 50