The importance of alkali cations in the [{RuCl2(p-cymene)} 2]-pseudo-dipeptide-catalyzed enantioselective transfer hydrogenation of ketones

Article abstract of DOI:10.1002/chem.200501384

We studied the role of alkali cations in the [{RuCl₂(p-cymene)} ₂]-pseudo-dipeptide-catalyzed enantioselective transfer hydrogenation of ketones with isopropanol. Lithium salts were shown to increase the enantioselectivity of the reaction when iPrONa or iPrOK was used as the base. Similar transfer-hydrogenation systems that employ chiral amino alcohol or monotosylated diamine ligands are not affected by the addition of lithium salts. These observations have led us to propose that an alternative reaction mechanism operates in pseudo-dipeptide-based systems, in which the alkali cation is an important player in the ligand-assisted hydrogen-transfer step. DFT calculations of the proposed transition-state (TS) models involving different cations (Li⁺, Na⁺, and K⁺) confirm a considerable loosening of the TS with larger cations. This loosening may he responsible for the fewer interactions between the substrate and the catalytic complex, leading to lower enantiodifferentiation. This mechanistic hypothesis has found additional experimental support; the low ee obtained with [BnNMe ₃]OH (a large cation) as base can be dramatically improved by introducing lithium cations into the system. Also, the complexation of Na ⁺, K⁺, and Li⁺ cations by the addition of [15]crown-5 and [18]crown-6 ethers and cryptand 2.1.1 (which selectively bind to these cations and, thus, increase their bulkiness), respectively, to the reaction mixture led to a significant drop in the enantioselectivity of the reaction. The lithium effect has proved useful for enhancing the reduction of different aromatic and heteroaromatic ketones.

Full text of DOI:10.1002/chem.200501384

DOI: 10.1002/chem.200501384
The Importance of Alkali Cations in the [{RuCl2(p-cymene)}2]–Pseudo-
dipeptide-Catalyzed Enantioselective Transfer Hydrogenation of Ketones
Patrik Västilä,[a] Alexey B. Zaitsev,[a] Jenny Wettergren,[a] Timofei Privalov,[b] and
Hans Adolfsson*[a]
Abstract: We studied the role of alkali
cations in the [{RuCl2(p-cymene)}2]–
pseudo-dipeptide-catalyzed enantiose-
lective transfer hydrogenation of ke-
tones with isopropanol. Lithium salts
were shown to increase the enantiose-
lectivity of the reaction when iPrONa
or iPrOK was used as the base. Similar
transfer-hydrogenation systems that
employ chiral amino alcohol or mono-
tosylated diamine ligands are not af-
fected by the addition of lithium salts.
These observations have led us to pro-
pose that an alternative reaction mech-
anism operates in pseudo-dipeptide-
based systems, in which the alkali
cation is an important player in the
ligand-assisted hydrogen-transfer step.
DFT calculations of the proposed tran-
sition-state (TS) models involving dif-
ferent cations (Li+, Na+, and K+) con-
firm a considerable loosening of the TS
with larger cations. This loosening may
be responsible for the fewer interac-
tions between the substrate and the
catalytic complex, leading to lower
enantiodifferentiation. This mechanistic
hypothesis has found additional experi-
mental support; the low ee obtained
with [BnNMe3]OH (a large cation) as
base can be dramatically improved by
introducing lithium cations into the
system. Also, the complexation of Na+,
K+, and Li+ cations by the addition of
[15]crown-5 and [18]crown-6 ethers
and cryptand 2.1.1 (which selectively
bind to these cations and, thus, increase
their bulkiness), respectively, to the re-
action mixture led to a significant drop
in the enantioselectivity of the reac-
tion. The lithium effect has proved
useful for enhancing the reduction of
different aromatic and heteroaromatic
ketones.
Keywords: amino acids
· amino
alcohols · asymmetric catalysis ·
hydrogen transfer · ruthenium
Introduction
formations, further improvement of catalyst performance,
especially concerning enantioselectivity, is a continuing chal-
lenge. Thus, the design of new ligands as well as new metal
complexes possessing desirable catalytic properties is of
high importance. Better catalysts can be obtained by the in-
troduction of appropriate achiral promoters, which can im-
prove dramatically the characteristics of existing catalytic
systems, virtually making it possible in some cases “to make
good asymmetric catalysis perfect”.[9] Such additives often
act to deoligomerize structures to form the desired catalyst
species. The use of bases as additives can enable a more
rapid dissociation of the catalyst–product complex through
ligand-exchange reactions. Additives like molecular sieves
can, for example, remove water produced in the reaction
that might be detrimental to the catalyst. Furthermore, cer-
tain additives can be used to poison undesirable catalyst
species, thereby enhancing the performance of the desired
catalyst. Another important result of modifying a catalytic
system by the addition of achiral promoters could be a
better mechanistic understanding of the catalytic process,
with far-reaching consequences.
During the last decade, asymmetric transfer hydrogenation
has enjoyed tremendous popularity among chemists working
in the field of catalysis. This process allows for easy prepara-
tion of enantiomerically enriched alcohols[1–8] and
amines[2–4,8] from prochiral ketones and imines in the pres-
ence of asymmetric catalysts and avoids the hazardous use
of molecular hydrogen. Despite the number of efficient cat-
alysts introduced in recent years for these particular trans-
[a] P. Västilä, A. B. Zaitsev, J. Wettergren, H. Adolfsson
Department of Organic Chemistry, The Arrhenius Laboratory
Stockholm University, 10691 Stockholm (Sweden)
Fax : (+46)8-154-908
E-mail: hansa@organ.su.se
[b] T. Privalov
Department of Chemistry, Organic Chemistry
Royal Institute of Technology (KTH), 10044 Stockholm (Sweden)
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
3218
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3218 – 3225

FULL PAPER
Table 1. Effect of additives on the [{RuCl2(p-cymene)}2]–pseudo-dipep-
Previously, we reported on the [{RuCl2(p-cymene)}2]–
pseudo-dipeptide catalytic system which, in the presence of
base, allowed for asymmetric transfer hydrogenation of ke-
tones in isopropanol, yielding the corresponding alcohols
with high enantioselectivity.[10–13] The pseudo-dipeptides
were prepared in a straightforward fashion from vicinal
amino alcohols and N-protected amino acids.[14] Upon re-ex-
amination of this catalytic system, we found that the intro-
duction of an alkali cation along with the base had a direct
influence on the outcome of the reduction reaction. Here,
we present unprecedented results showing a highly synerget-
ic effect between the ruthenium catalyst and alkali cations.
In addition, we report on how the enantioselectivity of the
alcohols formed can be improved by careful tuning of the
system by using appropriate alkali cations.
tide-catalyzed transfer hydrogenation of acetophenone.[a,b]
Entry
Additive
Conversion [%][c]
ee [%][c]
1
–
83
3
30
–
92( S)
>99 (S)
94 (S)
n.d.
2Sc(OTf)
3
4
5
6
7
8
9
10
11
12LiClO
13
3 (1 mol%)
Ti(OiPr)4 (5 mol%)
CuCl2 (10 mol%)
CuI (10 mol%)
AgOTf (5 mol%)
NaCl (10 mol%)
KCl (10 mol%)
LiCl (10 mol%)
LiBr (10 mol%)
LiI (10 mol%)
–
n.d.
58
74
76
85
84
77
77
76
93 (S)
93 (S)
93 (S)
95 (S)
94 (S)
94 (S)
95 (S)
94 (S)
4 (10 mol%)
LiOAc (10 mol%)
[a] Reaction conditions: acetophenone (1 equiv, 0.2m in 2-propanol),
[{RuCl2(p-cymene)}2] (0.5 mol%), ligand (1.1 mol%), NaOiPr (5 mol%),
room temperature. [b] The reaction mixtures were analyzed after 30 min
to avoid erosion of ee, which can occur as the reaction equilibrium is ap-
proached; here, at 91–92% conversions. [c] Conversions and enantiose-
lectivities were determined by GLC.
Results and Discussion
A metal complex used in asymmetric catalysis normally ac-
quires its specific catalytic properties from a combination of
metal–ligand interactions and the chiral space that the li-
gands create around the actual center of reactivity. Other
factors affecting the overall activity and selectivity of a cata-
lytic system include solvent effects and the use of noncoor-
dinated counterions. In nature, highly selective enzymes
serve as good examples of how efficient catalysis can be per-
formed based on these simple considerations. However, to
function properly, numerous enzymes require additional co-
catalysts that assist and participate in the overall catalytic
process. This principle of adding co-catalysts or certain addi-
tives, chiral or achiral, can effectively change and improve
an existing asymmetric catalytic reaction that involves tran-
sition-metal catalysts. With this in mind, we investigated the
outcome of adding different additives to the [{RuCl2(p-
cymene)}2]–pseudo-dipeptide catalytic system.
our standard reaction setup (i.e., the reduction of acetophe-
none in 0.2m 2-propanol with 0.5 mol% of [{RuCl2(p-
cymene)}2], 1.1 mol% of the pseudo-dipeptide ligand, and
5 mol% of sodium isopropoxide). Copper salts were found
to completely inhibit the reduction reaction, and the addi-
tion of silver triflate resulted in a somewhat lower conver-
sion than that obtained under the standard conditions (en-
tries 4–6, Table 1). The addition of alkali salts gave some in-
teresting results; reactions with sodium or potassium chlo-
ride as additive gave similar results to the nonadditive reac-
tion (entries 7 and 8, Table 1), whereas the addition of
lithium chloride resulted in a higher enantioselectivity of the
product 1-phenylethanol (entry 9, Table 1). Apparently,
modification of the [{RuCl2(p-cymene)}2]–pseudo-dipeptide–
iPrONa system by addition of lithium chloride (10 mol%)
led to a slight increase in the selectivity of the reduction re-
action. Similar behavior was also observed when other lithi-
um salts were used as additives (entries 10–13, Table 1). The
enhanced enantioselectivity observed upon addition of lithi-
um salts prompted us to investigate further the origin and
nature of this effect. In the standard reaction setup, we used
sodium isopropoxide as the base; we found that replacing
the sodium base by its lithium congener resulted in the
same ee enhancement (entries 3 and 8, Table 2) as that ob-
served when LiCl was added. Thus, by employing the stereo-
chemically matching ligands 1, 3, 5, and 7 (Figure 1) in com-
bination with a lithium salt or base in the reaction, we ob-
served an average increase in enantioselectivity of 3%. In
the most favorable case, involving ligand 3, we obtained 1-
phenylethanol in high yield and with excellent enantioselec-
tivity (98% ee) after 30 min (entry 7, Table 2). The behavior
of their mismatched congeners was more complex and de-
pended on the nature of the substituents at the stereogenic
centers. Surprisingly, in the case of the mismatched ligand 6,
the ee was increased by as much as 49% upon addition of
LiCl (entry 14, Table 2).
Effect of achiral additives: The [{RuCl2(p-cymene)}2]–
pseudo-dipeptide-catalyzed enantioselective transfer hydro-
genation of ketones is highly sensitive towards the amount
of base that is added during the reduction reaction. We
found that the addition of less than three equivalents, with
respect to the catalyst (i.e., typically <3 mol%), of either
sodium hydroxide or isopropoxide to the reaction mixture
resulted in little or no conversion of the ketone to the sec-
ondary alcohol. By adding a base, we introduce an addition-
al metal ion to the system, an alkali ion, which could poten-
tially act as a Lewis acid to activate the substrate and there-
by influence the overall reactivity of the system. We were
intrigued to see whether the nature of this metal cation
could have any influence on the catalytic reduction reaction
and, therefore, we performed a number of transfer-hydroge-
nation reactions in which different metal salts were added to
the reaction mixture (Table 1).
The addition of strong Lewis acids, such as scandium tri-
flate or titanium isopropoxide, had a negative effect on the
reactivity of the system (entries 2and 3, Table 1) relative to
Chem. Eur. J. 2006, 12, 3218 – 3225
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3219

H. Adolfsson et al.
Table 2. Effect of lithium cations on the [{RuCl2(p-cymene)}2]–pseudo-di-
that alkyl substituents have a larger effect on enantioselec-
tivity than phenyl groups. This effect is particularly pro-
nounced with the second-generation pseudo-dipeptides 5
and 6, in which the hydroxy group is adjacent to one of the
stereogenic centers. In these cases, the effect of the alkyl
substituent of the amino alcohol moiety is more dominant
than the effect of the phenyl group of the amino acid. As
seen in entry 13 of Table 2, the R-configured product was
formed as the major isomer even though an l-amino acid
based ligand was used. The addition of LiCl increased the
enantioselectivity to 56% ee (entry 14, Table 2), but still the
R alcohol was favored. Apparently, the presence of lithium
cations amplifies the importance of the alcohol stereogenic
center in the enantiodifferentiating step. However, in gener-
al, a decrease in product ee can be expected upon addition
of lithium salts to reduction reactions catalyzed by rutheni-
um complexes containing mismatched second-generation li-
gands, unless these ligands are based on phenylglycine (e.g.,
ligand 6).
peptide-catalyzed transfer hydrogenation of acetophenone.[a,b]
Entry
Ligand
Base
([mol%])
Additive
([mol%])
Conver.
[%][c]
ee
[%][c]
1
2
3
4
5
6
7
8
1
1
1
2
2
3
3
3
4
4
5
5
6
6
7
7
8
8
iPrONa (5)
iPrONa (5)
iPrOLi (10)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrOLi (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
iPrONa (5)
–
83
85
86
41
41
76
88
77
23
16
81
85
31
24
87
89
24
22
92( S)
95 (S)
95 (S)
94 (R)
97 (R)
97 (S)
98 (S)
98 (S)
61 (S)
45 (S)
95 (R)
98 (R)
7 (R)
LiCl (10)
–
–
LiCl (10)
–
LiCl (10)
–
9
–
10
11
12
13
14
15
16
17
18
LiCl (10)
–
LiCl (10)
–
LiCl (10)
–
LiCl (10)
–
LiCl (10)
56 (R)
92( S)
95 (S)
61 (R)
58 (R)
[a] Reaction conditions: acetophenone (1 equiv, 0.2m in 2-propanol),
[{RuCl2(p-cymene)}2] (0.5 mol%), ligand (1.1 mol%), base (see table),
room temperature. [b] The reaction mixtures were analyzed after 30 min
to avoid erosion of ee, which can occur as the reaction equilibrium is ap-
proached; here, at 91–92% conversions. [c] Conversions and enantiose-
lectivities were determined by GLC.
Effect of lithium additives on other transfer-hydrogenation
systems: To investigate whether the addition of lithium salts
could influence other typical transfer-hydrogenation systems
we performed reactions with ruthenium catalysts derived
from some commonly used ligands (Figure 2).[2,4] In contrast
Figure 2. Amino alcohol and diamine ligands used in the ruthenium-cata-
lyzed transfer hydrogenation of ketones.
to the pseudo-dipeptide systems, we observed no effect of
adding LiCl to catalytic reactions in which the amino alco-
hols 9 and 10 or (R,R)-TsDPEN (11) were employed as li-
gands (see Supporting Information). The absence of an
alkali-cation effect on these systems suggests that the hy-
dride transfer may occur by a different mechanism to that
operating when catalysts based on pseudo-dipeptides are
used in the ketone reduction.
The importance of alkali-metal cations in the hydrogena-
tion of ketones employing molecular hydrogen and the
trans-[RuCl2{(S)-binap}{(S,S)-dpen}][15] catalyst has been
demonstrated previously (binap=1,1’-binaphthalene-2,2’-
diylbis(diphenylphosphane), dpen=1,2-diphenylethylenedia-
mine).[16] In this system, a decrease in activity in the order
K>Na~Rb>Li was observed. This was attributed to the
formation of a potassium-specific binding site, formed with
the aid of the two phenyl rings of the ligand, which helps to
position better the tBuOK molecules in the transition state,
leading to the deprotonation of the intermediate dihydrogen
complex. Previously, secondary interactions with ancillary
ligand functionalities were shown to play an important role
in some catalytic asymmetric processes, including hydroge-
nation.[17–19] However, in our case these interactions appear
Figure 1. Pseudo-dipeptide ligands used in the study of cation effects on
the [{RuCl2(p-cymene)}2]–pseudo-dipeptide-catalyzed transfer hydroge-
nation of acetophenone.
The stereochemical outcome of the transfer-hydrogena-
tion reactions performed with ruthenium–pseudo-dipeptide
catalysts is determined predominantly by the stereocenter
present in the amino acid moiety.[11,12] Therefore, it is possi-
ble to predict the absolute configuration of the product of a
given reaction by choosing an appropriate ligand; pseudo-di-
peptides based on natural l-amino acids induce S-product
formation and ligands based on d-amino acids give the R
product. However, the addition of lithium salts revealed
3220
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3218 – 3225

Transfer Hydrogenation of Ketones
FULL PAPER
to be truly critical for catalytic activity, as demonstrated
below.
on these facts, we propose that a ruthenium complex with
structure 12 is formed under the reaction conditions
(Scheme 1).[21]
Mechanistic considerations: We previously reported that the
presence of a base-stable carbamoyl group at the N terminal
and a free hydroxy functionality at the C terminal is crucial
for efficient catalysis.[11] From these observations it is evi-
dent that the Boc-NH and OH functionalities act in a highly
cooperative fashion. Complexes based on ligands lacking
either of these groups were completely ineffective in the
transfer-hydrogenation reaction. For instance, no conversion
was observed when the N-tosyl analogue of 1 was employed
as a ligand of ruthenium. Similarly poor results were ob-
tained by using the O-methylated derivative of ligand 1.[11]
In addition, when the reduction was performed with ligands
lacking a central amide functionality (no peptide bond) or
when this amide was N-alkylated, we observed no product
formation. Another factor that has a huge effect on these re-
ductions is the amount of base added to generate the active
catalyst and to start the reaction. We observed that more
than three equivalents of base (isopropoxide) are typically
needed for the reaction to occur. If the reduction was per-
formed with less than three equivalents, we observed poor
conversion to the alcohol. The pseudo-dipeptides have three
acidic positions that can undergo deprotonation relatively
easily. However, the acidities of the hydroxy and amide NH
functionalities are several orders of magnitude higher than
that of the carbamate NH. Under the reaction conditions
employed (5 mol% base), the hydroxy and amide NH func-
tionalities are the most likely to undergo deprotonation
readily. In fact, Beck and co-workers demonstrated that
treatment of a Boc-protected dipeptide with NaOMe in the
presence of [{RuCl2(p-cymene)}2] results in the formation of
a Ru–arene–dipeptide complex in which no carbamate de-
protonation occurs.[20] In accordance with this result, we
assume that only two equivalents of base are consumed in
the ligand-deprotonation reactions. Consequently, the re-
quired third equivalent of base must be involved in a differ-
ent process. If three equivalents of isopropoxide were con-
sumed in the formation of the catalyst, we would obtain in-
active anionic ruthenium complexes. As previously demon-
strated, the use of ligands containing three equally acidic
sites (e.g., the N-tosyl analogue of 1) resulted in ruthenium
complexes with no catalytic ac-
Scheme 1. Proposed catalytic intermediates in the reduction of acetophe-
none under hydrogen-transfer conditions.
This 18-electron complex must release one of its donor
atoms to form the reducing ruthenium–hydride species. Re-
lease of the weakly coordinated carbamate followed by sub-
sequent hydride formation would again generate a formally
negatively charged, catalytically inactive ruthenium com-
plex. Instead, in line with the significant effect imposed by
the alkali cation, it can be proposed that ruthenium complex
12 reacts with iPrOLi to form hydride complex 13. The sub-
strate reacts with the hydride complex 13 to generate the
desired secondary alcohol. The role of the lithium ion pres-
ent in complex 13 is to activate and direct the incoming
ketone prior to hydride transfer from the ruthenium center.
We suggest that the transfer of the hydride and alkali ion to
the substrate proceeds via a six-membered transition state,
analogous with the aluminium alkoxide mediated Meer-
wein–Pondorff–Verley reduction[22] (Figure 3a and b). Lewis
acid activation and orientation of the ketone by lithium (or
other alkali-metal)-ion coordination[23] are important for the
selective reduction process employing the [{RuCl2(p-
cymene)}2]–pseudo-dipeptide catalytic system. The increased
selectivity obtained in the presence of lithium salts com-
pared to other alkali cations can be explained in terms of
the formation of a tighter transition state with the smaller
lithium ion (Figure 3c). Consequently, the close participa-
tion of larger-sized alkali cations in the transition state
should be limited, leading to lower reactivity and selectivity.
The observation that the product configuration is predomi-
nantly determined by the stereogenic center in the amino
tivity. To study further the
nature of the catalyst, we per-
formed reduction reactions
using mixtures of ligands 2 and
ent-2. The absence of a nonlin-
ear relationship between the
enantiomeric purities of the
ligand and product in the
transfer hydrogenation of ace-
Figure 3. a) Cyclic transition state for the aluminium alkoxide mediated reduction of ketones by 2-propanol
(Meerwein–Pondorff–Verley reduction). b) Proposed schematic cyclic transition state involved in the RuII-
(arene)–pseudo-dipeptide-catalyzed reduction of ketones in the presence of lithium salts. c) Possible geometry
of the transition state showing the important interactions between the ruthenium hydride, the lithium cation,
and the substrate.
tophenone by using ligands 2/
ent-2 (see Supporting Informa-
tion) is in favor of a monoligat-
ed ruthenium catalyst. Based
Chem. Eur. J. 2006, 12, 3218 – 3225
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3221

H. Adolfsson et al.
acid moiety of the ligand is explained by the presence of
this center in the N,N-chelate of 13. In addition, the chirality
of the amino acid certainly has a strong influence on the
configuration of the ruthenium stereogenic center.
process in the pseudo-dipeptide-based catalytic system.
Indeed, reduction reactions performed with [BnNMe3]OH
as base resulted in poor conversion and enantioselectivity of
the alcohol formed (entries 7 and 13, Table 3), however,
upon addition of LiCl the ee was dramatically improved (en-
tries 8 and 14, Table 3). Furthermore, the addition of macro-
cyclic ethers [15]crown-5 or [18]crown-6, which selectively
complex alkali cations and increase their effective size, to
reaction mixtures containing the bases iPrONa or tBuOK,
respectively, led to a noticeable decrease in both the conver-
sion and enantioselectivity of 1-phenylethanol (entries 2, 4,
Computational results: To further our understanding of the
effect of different-sized alkali cations on the transition-state
geometry of the hydrogen-transfer reaction, we performed
gas-phase DFT computations on the reaction systems. We
decided to perform these calculations because our recent
studies[24] demonstrated that the intimate details of complex
reaction mechanisms could be treated well by computational
methods. The computational details are described in the Ex-
perimental Section. A comprehensive account of solvent ef-
fects in the model with explicit solvent molecules and a po-
larizable continuum medium will be discussed elsewhere.
Here, we present gas-phase calculations that enable compa-
rative estimation of the geometrical parameters of the tran-
sition states involving Li+, Na+, and K+ complexes. The cal-
culated transition-state (TS) structures are presented in
Figure 4.[25]
Table 3. Effect of cation size on the [{RuCl2(p-cymene)}2]–pseudo-dipep-
tide-catalyzed transfer hydrogenation of acetophenone.[a,b]
Entry Ligand Base
([mol%])
Additive
([mol%])
Conver. ee
[%][c]
[%][c]
1
2
3
4
1
1
1
1
1
1
1
1
3
3
3
3
3
3
8
8
iPrONa (5)
iPrONa (5)
tBuOK (5)
tBuOK (5)
iPrOLi (10)
iPrOLi (10)
[BnNMe3]OH
–
83
68
70
40
86
92( S)
83 (S)
91 (S)
64 (S)
95 (S)
64 (S)
78 (S)
92 (S)
97 (S)
S)
98 (S)
14 (S)
73 (S)
98 (S)
61 (R)
48 (R)
[15]crown-5 (30)
–
[18]crown-6 (30)
–
cryptand 2.1.1 (30) 22
5
6[d]
7[e]
8[e]
9
As illustrated in Figure 4, the TS structure is looser for
–
21
21
76
5292
77
6
10
40
24
17
À
À
ions with a larger ionic radius. The Ru C and Ru ion dis-
[BnNMe3]OH LiCl (10)
tances increase as the alkali-ion radii increase,[26] resulting in
iPrONa (5)
iPrONa (5)
iPrOLi (5)
iPrOLi (5)
[BnNMe3]OH
[BnNMe3]OH LiCl (10)
iPrONa (5)
iPrONa (5)
–
À À À À
À
a larger Ru H C O ion O ring motif. In addition, in con-
trast to the results of theoretical studies by Noyori and co-
workers on the ruthenium–TsDPEN catalyst,[27] aryl–aryl
(CH–p)-directing interactions between the h6-arene ligand
and the substrate do not appear to be crucial for hydride
transfer using the ruthenium–pseudo-dipeptide catalyst.[29]
10
11
12[d]
13[e]
14[e]
15
16
[15]crown-5 (30) (
–
cryptand 2.1.1(30)
–
–
[15]crown-5 (30)
[a] Reaction conditions: acetophenone (1 equiv, 0.2m in 2-propanol),
[{RuCl2(p-cymene)}2] (0.5 mol%), ligand (1.1 mol%), base (see table),
room temperature. [b] The reaction mixtures were analyzed after 30 min
to avoid erosion of ee, which can occur as the reaction equilibrium is ap-
proached; here, at 91–92% conversions. [c] Conversions and enantiose-
lectivities were determined by GLC. [d] Cryptand 2.1.1=4,7,13,18-tet-
raoxa-1,10-diazabicyclo[8.5.5]eicosane. [e] [BnNMe3]OH was employed
as a 40% aqueous solution.
Addition of crown ethers and cryptands: The observed loos-
ening of the TS with larger alkali cations is likely to be re-
sponsible for the lower enantiocontrol observed in systems
using iPrONa or iPrOK as base. If this assumption is true, a
further increase in cation size may lead to an additional
drop in the enantioselectivity of the transfer-hydrogenation
Figure 4. DFT-optimized TS structures involving a) Li+, b) Na+, and c) K+ ions. All distances are in Š. The TS structures were characterized by one
À
À
À
imaginary frequency along the appropriate normal mode. The key geometrical parameters [Š] are: S-TS-Li: Ru C 3.231, Ru Li 3.768, O O 3.388;
À
À
À
À
À
À
S-TS-Na: Ru C 3.327, Ru Na 3.997, O O 3.903; S-TS-K: Ru C 3.474, Ru K 4.833, O O 3.987.
3222
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3218 – 3225

Transfer Hydrogenation of Ketones
FULL PAPER
Table 4. Effect of lithium chloride on the reduction of different aromatic
10, and 16, Table 3). Addition of [12]crown-4 to the system
containing iPrOLi did not affect the reaction outcome sig-
nificantly.[30] However, performing the reaction under the
same conditions in the presence of cryptand 2.1.1 led to a
drastic decrease in conversion and enantioselectivity
(entry 12, Table 3), which is most likely due to the increased
size of the cation upon coordination to the crown ethers or
the cryptand. This coordination inhibits participation of the
cations in any preorganization of the reaction partners that
lead to the product, and consequently, lower activity and
scrambling of the enantioselectivity are observed. These re-
sults provide additional evidence that the alkali cation must
play a significant intimate role in the hydride-transfer step
of the reaction when pseudo-dipeptides are used as ligands.
The use of triethylamine as base resulted in a low conver-
sion (<2%) and selectivity (<2%), and this result was not
improved by the addition of LiCl, indicating that high base
strength is important for promoting the reaction.
and heteroaromatic ketones by using ligand 1.[a]
Entry
Substrate
LiCl
[mol%]
Time
[min]
Conver.
[%][b]
ee
[%][b]
1
2(10
–
30
64
93 (S)
S)
30
79
92
3
4
–
10
12
120
0
2
S)
0
32
83 (S)
5
6
–
10
30
30
86
88
90 (S)
90 (S)
7
8
–
10
120
12
8
16 (S)
S)
Effect of lithium chloride on the reduction of other ketones:
To investigate the scope of the lithium salt effect, we studied
the effect of lithium chloride on the transfer hydrogenation
of different aromatic and heteroaromatic ketones by using
ligands 1 (Table 4) and 3 (see Supporting Information) as
matching ligands of the first- and second-generation pseudo-
dipeptides, respectively. The data presented in the tables
show that the magnitude of the effect is dependent on the
nature of the substrate. Surprisingly, an increase in ee of up
to 47% could be obtained for the transfer hydrogenation of
0
9
10
–
10
30
30
97[c]
99[c]
92( S)
96 (S)
11
19241(0
–
12
0
96 S)
S)
89 (
82
4-cyanoacetophenone by using ligand 1 upon addition of
0
99
12
LiCl (entry 8, Table 4).
Thus, the addition of lithium salts to hydrogen-transfer
catalytic systems based on pseudo-dipeptides is generally
useful for enhancing the reduction of different aromatic and
heteroaromatic ketones.
13
14
–
10
12
120
0
81 S)
84 (S)
95
15
16
–
10
30
30
97
97
91 (S)
95 (S)
Conclusion
We have demonstrated that the transfer hydrogenation of
ketones catalyzed by the [{RuCl2(p-cymene)}2]–pseudo-di-
peptide system is highly dependent on the presence of alkali
cations. The addition of crown ethers or cryptands signifi-
cantly reduces the activity and selectivity of the reduction
reaction, which implies that the alkali ion is part of the re-
ducing catalyst and that a different reaction mechanism
must operate. Furthermore, increased enantioselectivity was
observed upon addition of lithium salts to the system. By
employing DFT computations we could illustrate and vali-
date key mechanistic findings, and the lithium effect is ex-
plained in terms of a significant tightening of the transition
state that leads to the product alcohol. No alkali-ion effect
has been observed in “traditional” transfer-hydrogenation
systems (e.g., Ru–arene complexes of 1,2-amino alcohols),
revealing the uniqueness of these modular pseudo-dipep-
tides with functional groups operating in a highly coopera-
tive enzymelike mode.
17
18
–
10
30
30
3
31
87 (S)
91 (S)
19
2
–
12
0
11 S)
10S)
82
0
12
21
22
–
10
30
30
89
91
63 (S)
70 (S)
[a] Reaction conditions: acetophenone (1 equiv, 0.2m in 2-propanol),
[{RuCl2(p-cymene)}2] (0.5 mol%), ligand (1.1 mol%), iPrONa (5 mol%),
room temperature. [b] Conversions and enantioselectivities were deter-
mined by GLC. [c] Partial transesterification of the methoxycarbonyl
group with isopropanol was observed.
Chem. Eur. J. 2006, 12, 3218 – 3225
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3223

H. Adolfsson et al.
K. Hattori, I. Furokawa, Chem. Lett. 1998, 491–492; b) T. Ohta, S.
Nakahara, Y. Shigemura, K. Hattori, I. Furokawa, Appl. Organomet.
Chem. 2001, 15, 699–709; c) D. Carmona, F. J. Lahoz, R. Atencio,
L. A. Oro, M. P. Lamata, F. Viguri, E. San JosØ, C. Vega, J. Reyes, F.
Joó, . Kathó, Chem. Eur. J. 1999, 5, 1544–1564; d) . Kathó, D.
Carmona, F. Viguri, C. D. Remacha, J. Kovµcs, F. Joó, L. A. Oro, J.
Organomet. Chem. 2000, 593–594, 209–306; e) D. Carmona, M. P.
Lamata, F. Viguri, I. Dobrinovich, F. J. Lahoz, L. A. Oro, Adv.
Synth. Catal. 2002, 344, 499–502; f) D. Carmona, M. P. Lamata,
L. A. Oro, Eur. J. Inorg. Chem. 2002, 2239–2251; g) J. W. Faller,
A. R. Lavoie, Organometallics 2001, 20, 5245–5247; h) H. Y. Rhyoo,
Y.-A. Yoon, H.-J. Park, Y. K. Chung, Tetrahedron Lett. 2001, 42,
5045–5048; i) H. Y. Rhyoo, H.-J. Park, Y. K. Chung, Chem.
Commun. 2001, 2064–2065; j) H. Y. Rhyoo, H.-J. Park, W. H. Suh,
Y. K. Chung, Tetrahedron Lett. 2002, 43, 269–272.
Experimental Section
Computational methods—technical details: Geometry optimizations of
transition states and stable intermediates were carried out by using the
B3LYP functional[31] with the lacvp*/6-31G(d,p) basis set.[32, 33] All degrees
of freedom were optimized, and only positive vibrational frequencies
were obtained for the optimized geometries of the intermediates. The
transition states were characterized by the presence of only one imagina-
ry vibrational frequency along the appropriate normal mode. In the
second step, B3LYP energies were evaluated for the optimized geometry
by using the much larger triple-z basis set, lacv3p**+/6-311+G(d,p)
with additional diffuse and polarization functions. Control calculations
with respect to energy minimization and TS search were performed by
using the BP86 functional. No discrepancy between the results of the two
functionals was found for identical basis sets.
[15] Also known as the Noyori catalytic system. binap=1,1’-binaphtha-
lene-2,2’-diylbis(diphenylphosphane); dpen=1,2-diphenylethylene-
diamine.
[16] a) R. Hartmann, P. Chen, Angew. Chem. 2001, 113, 3693–3697;
Angew. Chem. Int. Ed. 2001, 40, 3581–3585; b) R. Hartmann, P.
Chen, Adv. Synth. Catal. 2003, 345, 1353–1359; c) similar cation ef-
fects on the catalytic activity of other RuII–diphosphine/1,2-diamine
complexes have recently been reported, see: J.-H. Xie, S. Liu, X.-H.
Huo, X. Cheng, H.-F. Duan, B.-M. Fan, L.-X. Wang, Q.-L. Zhou, J.
Org. Chem. 2005, 70, 2967–2973.
[17] M. Sawamura, Y. Ito, Chem. Rev. 1992, 92, 857–871.
[18] I. Yamada, M. Ohkouchi, M. Yamaguchi, T. Yamagishi, J. Chem.
Soc. Perkin Trans. 1 1997, 1869–1873.
[19] A. Bçrner, Eur. J. Inorg. Chem. 2001, 327–337.
[20] M. Prem, K. Polborn, W. Beck, Z. Naturforsch. B 1998, 53, 1501–
1505.
All computations were performed by using the Jaguar v4.0 and v6.0 suite
of ab initio quantum chemistry programs.[34] Frequency analysis of the TS
structures was performed by using Jaguar v6.0.
General procedure for the transfer hydrogenation of ketones by using
ligands 1–8: Ligand (0.011 mmol), [RuCl2(p-cymene)]2 (0.005 mmol), and
LiCl (0.1 mmol) (or other additives; see details in the tables) were dried
under vacuum in a dry Schlenk tube for 15 min. 2-Propanol (4.5 mL) and
a 0.01m solution of iPrONa (0.5 mL, 5 mol%) in iPrOH was added
under nitrogen. The solution was stirred for 5 min and then the ketone
(1 mmol) was added. The reaction mixture was stirred at ambient tem-
perature. Aliquots were taken after the reaction times indicated in the
tables and were then passed through a pad of silica with EtOAc as the
eluent. The resulting solutions were analyzed by GLC (CP Chirasil
DEXCB). For analytical data, see Supporting Information.
[21] Attempts to isolate and structurally characterize any ruthenium
complex(es) containing pseudo-dipeptide ligands have so far been
unsuccessful. Spectrometric studies (NMR and ESI-MS) of solutions
of the catalytically active mixture have provided no conclusive infor-
mation regarding the structure(s) of the catalyst or catalyst precur-
sor.
[22] C. F. de Graauw, J. A. Peters, H. van Bekkum, J. Huskens, Synthesis
1994, 1007–1017.
[23] Note that lithium bonds are generally considerably stronger than
their proton-centered counterparts, see: S. Scheiner in Lithium
Chemistry (Eds.: A.-M. Sapse, P. von R. Schleyer), Wiley, New York,
1995, pp. 67–87.
[24] a) S. Paavola, K. Zetterberg, T. Privalov, I. Csçregh, C. Moberg,
Adv. Synth. Catal. 2004, 346, 237–244; b) T. Privalov, C. Linde, K.
Zetterberg, C. Moberg, Organometallics 2005, 24, 885–893; c) M.
Schultz, R. Adler, W. Zierkiewicz, T. Privalov, M. Sigman, J. Am.
Chem. Soc. 2005, 127, 8499–8507; d) W. Zierkiewicz, T. Privalov,
Organometallics 2005, 24, 6019–6028.
Acknowledgements
We thank Professor Masakatsu Shibasaki of the University of Tokyo and
Professors Lena Mäler and Armando Córdova of the Stockholm Univer-
sity for valuable discussions. The Swedish Research Council and the
Wenner-Gren Foundation are thanked for financial support and a post-
doctoral fellowship (A.B.Z.), respectively.
[1] G. Zassinovich, G. Mestroni, S. Gladiali, Chem. Rev. 1992, 92, 1051–
1069.
[2] R. Noyori, S. Hashiguchi, Acc. Chem. Res. 1997, 30, 97–102.
[3] C. Bianchini, L. Glendenning, Chemtracts 1997, 10, 333–338.
[4] M. J. Palmer, M. Wills, Tetrahedron: Asymmetry 1999, 10, 2045–
2061.
[5] T. Ohkuma, R. Noyori in Comprehensive Asymmetric Catalysis I–
III, Vol. 1 (Eds.: E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer-
Verlag, Berlin, 1999, pp. 199–246.
[25] The TS structures were optimized with respect to convergence crite-
ria and the Hessian eigenvalues.
[26] These results were confirmed by the use of basis sets with additional
diffuse and polarization functions. There was a considerable differ-
ence between the structures obtained with the lacvp and lacvp*
basis sets. Upgrading the basis sets to lacvp* or lacvp*+ made little
difference.
[6] H.-U. Blaser, C. Malan, B. Pugin, F. Spindler, H. Steiner, M. Studer,
Adv. Synth. Catal. 2003, 345, 103–151.
[7] K. Everaere, A. Mortreux, J.-F. Carpentier, Adv. Synth. Catal. 2003,
345, 67–77.
[8] T. Ohkuma, R. Noyori, Compr. Asymmetric Catal. Suppl. 2004, 1,
43–53.
[9] E. M. Vogl, H. Grçger, M. Shibasaki, Angew. Chem. Int. Ed. 1999,
38, 1570–1577.
[10] I. M. Pastor, P. Västilä, H. Adolfsson, Chem. Commun. 2002, 2046–
2047.
[11] I. M. Pastor, P. Västilä, H. Adolfsson, Chem. Eur. J. 2003, 9, 4031–
4045.
[12] A. Bøgevig, I. M. Pastor, H. Adolfsson, Chem. Eur. J. 2004, 10, 294–
302.
[13] P. Västilä, J. Wettergren, H. Adolfsson, Chem. Commun. 2005,
4039–4041.
[27] a) K.-J. Haack, S. Hashiguchi, A. Fujii, T. Ikariya, R. Noyori,
Angew. Chem. 1997, 109, 297–300; Angew. Chem. Int. Ed. Engl.
1997, 36, 285–288; b) M. Yamakawa, H. Ito, R. Noyori, J. Am.
Chem. Soc. 2000, 122, 1466–1478; c) R. Noyori, M. Yamakawa, S.
Hashiguchi, J. Org. Chem. 2001, 66, 7931–7944; d) M. Yamakawa, I.
Yamada, R. Noyori, Angew. Chem. Int. Ed. 2005, 44, 1241–1244.
[28] For other recent mechanistic discussions, see: a) D. A. Alonso, P.
Brandt, S. J. M. Nordin, P. G. Andersson, J. Am. Chem. Soc. 1999,
121, 9580–9588; b) D. G. I. Petra, J. N. H. Reek, J.-W. Handgraaf,
E. J. Meijer, P. Dierkes, P. C. J. Kamer, J. Brussee, H. E. Schoemak-
er, P. W. N. M. van Leeuwen, Chem. Eur. J. 2000, 6, 2818–2829;
c) O. Pꢁmies, J.-E. Bäckvall, Chem. Eur. J. 2001, 7, 5052–5058;
d) C. P. Casey, J. B. Johnson, J. Org. Chem. 2003, 68, 1998–2001;
[14] For other examples of amino acid based ligands in the transfer hy-
drogenation of ketones, see: a) T. Ohta, S. Nakahara, Y. Shigemura,
3224
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3218 – 3225

Transfer Hydrogenation of Ketones
FULL PAPER
e) J. B. Johnson, J.-E. Bäckvall, J. Org. Chem. 2003, 68, 7681–7684;
f) J.-W. Handgraaf, J. N. H. Reek, E. J. Meijer, Organometallics 2003,
22, 3150–3157; g) J.-W. Handgraaf, J. N. H. Reek, L. Bellarosa, F.
Zerbetto, Adv. Synth. Catal. 2005, 347, 792–802; h) P. Brandt, P.
Roth, P. G. Andersson, J. Org. Chem. 2004, 69, 4885–4890.
[31] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; b) C. Lee, W.
Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[32] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[33] a) W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys. 1972, 56,
2257–2261; b) M. M. Francl, W. J. Pietro, W. J. Hehre, J. S. Binkley,
M. S. Gordon, D. J. Defrees, J. A. Pople, J. Chem. Phys. 1982, 77,
3654–3665; c) P. C. Hariharan, J. A. Pople, Theor. Chim. Acta 1973,
28, 213–222.
[29] Experimental and theoretical studies on Ru(arene)amino alcohol-
catalyzed transfer hydrogenation of aryl ketones by Brandt et al.
showed that this type of attractive interaction, which has been as-
sumed to be dominated by electrostatic forces, is not the most im-
portant in determining the product enantioselectivity, see ref. [28].
[30] Lithium ions display a lower affinity for [12]crown-4 than the corre-
sponding sodium and potassium ions for [15]crown-5 and [18]crown-
6, respectively, see: R. M. Izatt, J. S. Bradshaw, S. A. Nielsen, J. D.
Lamb, J. J. Christensen, Chem. Rev. 1985, 85, 271–339.
[34] Jaguar versions 4.0 and 6.0, Schrçdinger, Portland, OR, 1991–2000.
Received: November 7, 2005
Published online: January 27, 2006
Chem. Eur. J. 2006, 12, 3218 – 3225
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3225