Ligand Effect in Alkali-Metal-Catalyzed Transfer Hydrogenation of Ketones

Article abstract of DOI:10.1002/chem.201902240

This work unveils the reactivity patterns, as well as ligand and additive effect on alkali-metal-base-catalyzed transfer hydrogenation of ketones. Crucially to this reactivity is the presence of a Lewis acid (alkali cation), as opposed to a simple base effect. With aryl ketones, the observed reactivity order is Na⁺>Li⁺>K⁺, whereas for aliphatic substrates it follows the expected Lewis acidity, Li⁺>Na⁺>K⁺. Importantly, the reactivity pattern can be drastically changed by adding ligands and additives. Kinetic, labelling, and competition experiments as well as DFT calculations suggested that the reaction proceeds through a concerted direct hydride-transfer mechanism, originally suggested by Woodward. The lithium cation was found to be intrinsically more active than heavier congeners, but in the case of aryl ketones a decrease in reaction rate was observed at ≈40 percent conversion with lithium cations. Noncovalent-interaction analysis revealed that this deceleration effect originated from specific noncovalent interactions between the aryl moiety of 1-phenylethanol and the carbonyl group of acetophenone, which stabilize the product in the coordination sphere of lithium and thus poison the catalyst. The ligand/additive effect is a complicated phenomenon that includes a combination of several factors, such as the decrease of activation energy by ligation (confirmed by distortion/interaction calculations of N,N,N’,N’-tetramethylethylenediamine, TMEDA) and the change in relative stabilization of reagents and substrates in the solution and the coordination sphere of the metal. Finally, we observed that lithium-base-catalyzed transfer hydrogenation can be further facilitated by the addition of an inexpensive and benign reagent, LiCl, which likely operates by re-initiating the reaction on a new lithium center.

Full text of DOI:10.1002/chem.201902240

A Journal of
Accepted Article
Title: Ligand Effect in Alkali Metal-Catalyzed Transfer Hydrogenation of
Ketones Iryna D. Alshakova, Hayden C. Foy, Travis Dudding, and
Georgii I. Nikonov*[a]
Authors: Georgii Nikonov, Iryna Alshakova, Hayden Foy, and Travis
Dudding
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.201902240
Link to VoR: http://dx.doi.org/10.1002/chem.201902240
Supported by

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
Ligand Effect in Alkali Metal-Catalyzed Transfer Hydrogenation of
Ketones
Iryna D. Alshakova, Hayden C. Foy, Travis Dudding, and Georgii I. Nikonov*[a]
Abstract: This paper unveils the reactivity patterns, as well as ligand
and additive effect on alkali metal base catalyzed transfer
hydrogenation of ketones. Crucially to this reactivity is the presence
of a Lewis acid (alkali cation), as opposed to a simple base effect.
With aryl ketones, the observed reactivity order is Na+ > Li+ > K+,
whereas for aliphatic substrates it follows the expected Lewis acidity,
Li+ > Na+ > K+. Importantly, the reactivity pattern can be drastically
changed by adding ligands and additives. Kinetic, labelling, and
competition experiments and DFT calculations suggested that the
reaction proceeds via a concerted direct hydride transfer mechanism,
originally suggested by Woodward. Lithium cation was found to be
intrinsically more active than heavier congeners, but in the case of aryl
ketones a decrease in reaction rate was observed at ~40% conversion
with lithium cations. Non-covalent interaction analysis revealed that
this deceleration effect originated from specific non-covalent
interactions between the aryl moiety of 1-phenylethanol and the
carbonyl group of acetophenone, which stabilize the product in the
coordination sphere of lithium and thus poison the catalyst. The
ligand/additive effect is a complicated phenomenon that includes a
combination of several factors, such as the decrease of activation
energy by ligation (confirmed by D/I calculations of a diamine,
TMEDA) and the change in relative stabilization of reagents and
substrates in the solution and the coordination sphere of the metal.
Finally, we observed that lithium base catalyzed transfer
hydrogenation can be further facilitated by the addition of an
inexpensive and benign reagent, LiCl, which likely operates by re-
initiating the reaction on a new lithium center.
process included the application of other Group 3 and 13
centers[7] and the use of tailored ligand platforms in the traditional
aluminum catalysis to achieve enantioselective reduction[8] and/or
increased activity.[9]
Chart 1. Suggested transition states in aluminum- (1) and alkali metal-catalyzed
(2) MPV processes.
Although alkali metals are not generally considered as catalytic
centers, but merely as innocent counter-cations, their role in
catalysis is not negligible. In fact, the oldest alkali metal-mediated
reaction, albeit usually not considered as such, is the classical
Cannizzaro reaction,[10] i.e. disproportionation of aldehydes under
the action of alkali metal bases, e.g. KOH. In 1945, Woodward et
al. reported that alkali metal bases can catalyze the MPV
reduction and Oppenauer oxidation,[11] which was successfully
applied to the synthesis of quininone.[12] Mechanistic studies by
von Doering and Aschner showed that radical intermediates were
not formed in this reaction and that the hydride is transferred from
the carbon atom of the carbinol group in essentially a symmetrical
transition state 2.[13] Despite this finding, the role of alkali metals
in catalysis had been largely neglected[14] until Bäckvall et al.
discovered the accelerating effect of bases in metal-mediated
transfer hydrogenation (TH).[15] After that, the application of
excess alkali metal bases (in the form of hydroxides, alkoxides,
carbonates, or phosphates) as promoters in catalytic TH became
very common.[16] Nevertheless, the idea that alkali metals
themselves can be the catalytic centers had been dormant until
2004, when Crabtree et al. observed that carbonates M2CO3 (M
= Rb, Cs) catalyzed transfer hydrogenation of 2-naphthaldehyde
with 2-propanol.[17] Alkali metal catalysis was rediscovered again
in 2007, when Adolfsson et al. reported reduction of an array of
aryl and alkyl ketones in isopropanol mediated by lithium
isopropoxide, albeit at elevated temperatures (180 °C).[18] Since
these earlier reports a variety of alkali metals (Li-Cs), bases
(hydroxides, alkoxides, carbonates, phosphates) and solvents
(isopropanol, ethanol) have been studied in catalytic transfer
hydrogenation[19] and in the closely related alkylation of alcohols
and ketones.[20],[21] Ouali et al. provided convincing evidence that
transition metal contaminants are not responsible for the
catalysis and also made an interesting observation that the
activity follows th e order Li < K < Na, which was explained by the
balancing effect of alkali metal Lewis acidity on the carbonyl
activation and product decomplexation steps.[19a] Two questions
remain unanswered. First, can we improve the efficiency of this
catalytic system by employing ligands in the same manner as
Introduction
The discovery by Meerwein and Schmidt[1] and by Verley[2] of
aluminum-catalyzed transfer hydrogenation of aldehydes by
primary alcohols was a major milestone in the reduction chemistry
of the 20th century. The further finding by Ponndorf[3] that
secondary alcohols can be conveniently used for reduction of
ketones and aldehydes laid grounds for the wide application of
the MPV process in both industrial and academic settings up to
the point when soluble main group hydrides were introduced as
an alternative in mid 50-s. Labelling,[4] kinetic,[5] and
computational[6] studies supported the notion that the reaction
proceeds via a cyclic transition state 1, with the rate determining
step being the transfer of hydride from the carbon atom bearing
the OAl(OR)2 group (Chart 1). Further advances in the MPV
[a]
Dr. I. D. Alshakova, H. C. Foy, Prof. T. Dudding, and Prof. G. I.
Nikonov*
Chemistry Department
Brock University
1812 Sir Isaac Brock Way, St. Catharines, L2S 3A1, Ontario,
Canada
E-mail: gnikonov@brocku.ca
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
used in transition metal catalysis? And, in particular, can an
enantioselective reduction be accomplished? And second, what
is the mechanism of alkali metal mediated transfer
hydrogenation? The following report provides some answers to
both these questions. Thus, we found evidence that lithium cation
is intrinsically more active in promoting transfer hydrogenation
and that its activity in the TH of acetophenone can be strongly
affected by both ligands and additives. We further provide
combined experimental and computational data explaining the
unique role of lithium and a rationalization of the ligand effect.
Experimental Results
Figure 1. Kinetic profiles for the TH of acetophenone in isopropanol catalyzed
by LiOiPr, NaOiPr, and KOiPr.
Effect of ligands on alkali metal-mediated transfer
hydrogenation. We commenced our investigation with checking
whether a strong, metal-free base can alone mediate catalytic
reduction. To this end, acetophenone, the quintessential
substrate for transfer hydrogenation, was employed using
isopropanol with heating in the presence of 10 mol% phosphorus
ylides,
such
as
methylene(triphenyl)phosphorane
returned
and
zero
phenylmethylene(triphenyl)phosphorane,
conversion.[22] However, addition of 10 mol% LiCl to a mixture of
10% Ph3P=CH2 and acetophenone in isopropanol results in 68%
conversion after reflux for 8 hours. Carrying out the reaction in the
presence of catalytic LiOiPr (10 mol%) under or without hydrogen
atmosphere (1 atm) showed the same efficiency.
We then evaluated the relative activity of alkali metal cations (Li+,
Na+, and K+). Transfer hydrogenation of acetophenone with
isopropanol catalyzed by 10 mol% MOiPr (M= Li, Na, K) was
again chosen as the model system. The kinetic profiles, presented
in Figure 1, show that the conversion increases in the order
Li+<K+<Na+, which largely agrees with the catalytic activity
documented for other alkali metal bases.[19a] Since transfer
hydrogenation is an equilibrium process, the maximum
conversion of about 90% is achieved for NaOiPr after about 8 h,
whereas for KOiPr the curve is much less steep but gives a high
conversion of 54 % (80% after 20 h). In contrast, a very different
behavior was observed for LiOiPr. The initial reaction is very fast,
reaching 60% after 2 h vs 47% for NaOiPr, but then the reduction
slows down and shows saturation behavior at about 68%
conversion. These data clearly indicate that the lithium cation is
intrinsically more active than its heavier congeners but is also
passivated by the product of this reaction, PhCH2OH.
To test this hypothesis, experiments we repeated with a substrate
more resembling the HOiPr/acetone redox pair, that is
cyclohexanone. To this effect, the kinetic profiles shown in Figure
2 conclusively prove that Li+ was the most active, reaching 94%
conversion after only 2 hours. The activity now decreases in the
order Li+>Na+>K+, consistent with the decreasing Lewis acidity of
the alkali cation. The observation that cyclohexanone is more
active than acetophenone is quite remarkable and counterintuitive,
given the fact that dialkyl-substituted ketones have lower
oxidation potential.[23] However, the abnormal oxidation behavior
of cyclohexanone has been previously noted.[24]
Figure 2. Kinetic profiles for the TH of cyclohexanone in isopropanol catalyzed
by LiOiPr, NaOiPr, and KOiPr.
The important conclusion from these two sets of experiments is
that the catalytic activity may depend not only on the substrate
and reductant but also on compounds being added to the mixture
or formed during the reaction. This idea prompted us to
investigate systematically the ligand effect on alkali metal
catalyzed transfer hydrogenation. Reduction of acetophenone in
isopropanol was again chosen as the model system, while the
efficiency of additives was gauged by the time required to reach
equilibrium. Gratifyingly, addition of simple chelating diamines,
such as ethylenediamine and TMEDA, significantly improves the
conversion after 12 hours (Table 1, entries 2 and 3 vs entry 1).
Both a chelating diether, such as DME, and even a monoligating
ether, 1,4-dimethoxybenzene, were more effective, as equilibrium
was reached in a much shorter time, 8 h (entries 4 and 5).
Unexpectedly, soft donors, such as the chelating diphosphine
dppe (entry 6) and monoligating phosphines (entries 7 and 8)
turned out to be even stronger promoters, whereas the N-
heterocyclic carbene IMes (entry 9) was just a bit weaker. The
highest activity was achieved using the chelating DalPhos ligand
3 featuring both hard amino and soft phosphine sites (entry 10).
Interestingly, using an equimolar amount (10 mol%) of 12-crown-
4 still had a beneficial effect on catalysis (entry 11) and, in fact,
this cyclic polyether was more effective that diamines (entries 2
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
and 3) as equilibrium was reached in a shorter time. Though,
using two equivalents of this crown ether per lithium resulted in
sequestering of the cation, likely in the form of a sandwich, and
partial inhibition of catalysis (entry 12).
Table 2. The effect of chiral ligands (10 mol%) on the TH of acetophenone
catalyzed by 10 mol% LiOiPr.[a]
Entry
Ligand
Time
Conversion[b]
1
2
3
4
5
(-)-sparteine
PyBox
7h
86%[c]
85%[c]
85%[c]
85%[c]
NR
4h
(S)-DTBM-SEGPHOS
(S)-BINAP
8h
10h
12h
(S)-BINOL
Table 1. TH of acetophenone with 10 mol% LiOiPr in the presence of various
ligands (10 mol%).[a]
[a] Reaction conditions: acetophenone (50 L), LiOiPr (2.8 mg), ligand
(0.043 mmol), and 2-propanol (1.5 mL), 100oC; [b] conversions were
determined by NMR analysis; [c] no asymmetric induction.
Entry
Ligand
Time
Conversion[b]
To elucidate, whether the lack of enantioselectivity was due to fast
racemization of the chiral product, racemization of (R)-1-
phenylethanol with 10 mol% LiOiPr as catalyst and 10 mol% of
acetophenone as hydrogen acceptor was studied (Scheme 1).
Emerging from this study was a very slow decrease from 98% ee
to 77% ee after 24 hours, thus further substantiating the point that
the TH at lithium center is impeded by 1-phenylethanol.
1
-
12h
12h
12h
8h
68%
83%
82%
83%
84%
86%
84%
85%
84%
86%
84%
51%
2
ethylenediamine
TMEDA
3
4
DME
5
1,4-dimethoxybenzene
dppe
8h
6
7h
7
PPh3
7h
8
PEt3
7h
9
IMes
10h
4h
Scheme 1. Racemization of (R)-1-phenylethanol in the presence of LiOiPr in 2-
propanol.
10
11
12
DalPhos
12-Crown-4 (10%)
12-Crown-4 (20%)
10h
12h
Examples of synthetic application. To demonstrate the
synthetic utility of this catalytic system, we screened reduction of
several substrates (Scheme 2). Acetophenone, benzophenone,
p-chloroacetophenone, and p-cyanoacetophenone were reduced
much faster in the presence of 10% dppe, whereas the ligand
addition had no effect on the reduction of aliphatic ketones,
cyclohexanone and methyl tert-butyl ketone. Surprisingly, the TH
of p-methoxyacetophenone was also insensitive to the addition of
dppe.
[a] Reaction conditions: acetophenone (50 L), LiOiPr (2.8 mg), ligand
(0.043 mmol), and 2-propanol (1.5 mL), 100oC; [b] conversions were
determined by NMR analysis.
Encouraged by the discovery of the ligand effect, we moved on to
address the question whether enantioselective reduction can be
accomplished. Several chiral ligands were tried, and the results
are shown in Table 2. Unfortunately, in neither case was any
asymmetric induction obtained as evinced by the Feringa’s chiral
test.[25] For example, (-)-sparteine, a naturally occurring alkaloid
that is a common chiral inducer for asymmetric lithiation
reactions,[26] failed to bring about any asymmetric induction in this
transfer hydrogenation (Table 2, entry 1). In terms of efficiency,
chelating nitrogen-based ligands, (-)-sparteine and 2,6-bis[(4R)-
(+)-isopropyl-2-oxazolin-2-yl]pyridine (PyBox), performed the
best, reaching equilibrium in 7 and 4 hours, respectively (entries
1 and 2). Conversely, no reaction took place in the case of a
chelating diphenol ligand, such as BINOL (entry 5), likely because
of the increased stability of its dianionic form and hence
decreased concentration of the reactive isopropoxide in solution.
The possibility of phosphorus coordination to lithium was probed
6
by 31P NMR and Li NMR spectroscopy. Addition of LiOiPr to a
solution of dppe in isopropanol did not result in any change of the
chemical shift of phosphine. Likewise, no coupling to 31P was
observed in the lithium spectrum, suggesting that in the case of
dppe coordination to lithium cation is minimal and the activating
effect should have a different origin (vide infra).
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
2
Toluene
50%
50%
50%
50%
5%
10 h
10h
10h
10h
10h
10h
10h
10h
10h
82%
81%
82%
82%
70%
81%
74%
72%
39%
3
Mesitylene
Hexamethylbenzene
THF
4
5
6
Toluene
7
Toluene
10%
100%
300%
1500%
8
Toluene
9
Toluene
10
Toluene
[a] Reaction conditions: acetophenone (50 L), LiOiPr (2.8 mg), ligand (0.043
mmol), and 2-propanol (1.5 mL), 100oC; [b] conversions were determined by
NMR analysis.
Scheme 2. Ligand-assisted transfer hydrogenation of ketones. [a] No ligand
was added.
Mechanistic studies. The above data clearly show that alkali
metal cations play the key role in the catalytic transfer
hydrogenation and that ligands and additives can have significant
impact on their performance, which is particularly noticeable in the
case of lithium. For the traditional MVP reactions, three
mechanistic pathways were considered: the hydridic route based
on formation of a metal-hydride, a radical route, and a direct H-
transfer from alkoxide to carbonyl via a six-membered transition
state (most common).[6-7] For the alkali metal catalyzed reaction,
the hydridic route can be reliably ruled out as the formation of MH
species in alcoholic solutions is highly unlikely and because
added hydrogen, the likely product of a reaction between transient
MH and isopropanol, had no impact on catalysis. On the other
hand, under water- and alcohol-free conditions, alkali metal
hydrides can indeed become catalytically relevant.[21]
Additive effect on alkali metal-mediated transfer
hydrogenation. We were intrigued that quite a vast diversity of
ligands containing potential oxygen, amine, phosphine, and
aromatic binding sites can significantly enhance the catalytic
activity of LiOiPr. The activating effect of phosphine in the
absence of apparent Li-P interactions in NMR (vide supra) was
equally puzzling. To understand whether the mere presence of a
simple functional group, such as arene ring or oxygen atom, can
have a beneficial effect on catalysis, we decided to investigate the
effect of aromatic and ethereal additives on the lithium-catalyzed
TH.
To our surprise, addition of benzene and methyl substituted
benzenes also resulted in the enhancement of catalytic activity of
LiOiPr. Thus, the presence of 0.5 equivalents (relatively to the
substrate)
of
benzene,
toluene,
mesitylene,
or
To get a further insight into the mechanism of catalytic action,
kinetic studies under pseudo-first order conditions were
performed by using large excess of isopropanol (10-25
equivalents). Cyclohexanone was chosen as the model substrate
because its transfer hydrogenation can be considered as a
virtually irreversible process at the start of the reaction (up to
approximately 20% conversion). In all cases, first order kinetics in
the substrate was observed. The dependence of the reaction rate
on the catalyst was established by studying the variation of LiOiPr
loading from 2 to 10 mol% in 1 mL of 2-propanol. A perfect linear
plot of the effective reaction rate vs the amount of base was
obtained (Figure S1), which shows that the reaction is also first
order in the alkali metal catalyst. Further variation of the amount
of 2-propanol at a fixed LiOiPr load (4 mol %), established that the
reaction is also first order in the reducing agent to give an overall
second order reaction, with the kinetic law being rate =
k[catal][ketone][alcohol].
hexamethylbenzene allowed the reaction to reach 81-83%
conversion in 10 h (Table 3, entries 1-4), as compared to the
maximum 68% conversion observed in the absence of these
additives. A similar enhancement effect was found upon addition
of THF (Table 3, entry 5). The dependence of the catalytic activity
on toluene loading was studied next (entries, 2, 6-10), which
revealed increasing the amount of toluene up to 0.5 equivalent led
to a steady advancement of the reaction to 82% conversion in 10
hours (Table 3, entry 2). However, addition of a larger amount of
toluene (1 - 15 equivalents, entries 8-10) decreased the
conversion down to 39% with a 15-fold excess of toluene after 10
hours (entry 10). Even more surprisingly, addition of toluene to
the sodium or potassium isopropoxide catalyzed TH had zero
effect on the reaction rate, underlying the unique role of lithium in
this catalysis.
The effect of temperature was studied next by measuring the rate
of reaction between 70C and 90C (Figure S3), which following
use of the Van't Hoff equation (1) provided a temperature
coefficient of 2.7.
Table 3. TH of acetophenone with LiOiPr in the presence of various additives.[a]
Entry
1
Additive
Load
50%
Time
10 h
Conversion[b]
83%
푇
−푇
1
2
Benzene
10
푟2 = 푟 훾
(1)
1
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
Linearization of data in Arrhenius and Eyring coordinates (Figures
S4 and S5), based on equations 2 and 3, respectively, allowed for
the activation energy, enthalpy, and entropy be determined (Table
4). The relatively low enthalpy of activation and negative entropy
of activation point to an organized transition state, which is
consistent with a six-membered transition state commonly
accepted for the aluminum-catalyzed MPV reaction.[6]
−퐸
푎
푅푇
푘 = 퐴푒
(2)
(3)
≠
−퐺
푘 = 푇 푒
푘
푏
푅푇
ℎ
Table 4. Kinetic parameters found for the TH of cyclohexanone.[a]
Parameter
Value
Uncertainty
0.5 KJ/mol
1.2 sec-1
Ea
103.0 KJ/mol
26.0 sec-1
100.1 KJ/mol
-38.7 J/mol
Scheme 3. a) Direct hydride transfer mechanism and b) single electron transfer
mechanism for the lithium-catalyzed transfer hydrogenation of cyclopropyl
phenyl ketone.
A
H
S
0.6 KJ/mol
1.6 J/mol
The effect of substitution in the phenyl ring of acetophenone was
probed then and the resulting Hammett plot is presented in Figure
3. The positive slope shows that the reaction accelerates when
electron-withdrawing substituents in the para position of the
phenyl ring are present. This implies that a negative charge is
developed at the carbonyl group during the rate determining step
(RDS), which is more effectively stabilized by electron-
withdrawing rather than electron-donating groups. This result is
consistent with the hydride transfer to the substrate in the RDS.
[a] Reaction conditions: cyclohexanone (25.5 µL), LiOiPr (1.1 mg), and 2-
propanol (1.0 mL), 70-90oC.
We then probed the possibility of a radical mechanism by studying
the TH of cyclopropyl phenyl ketone, employing the cyclopropyl
group as a radical probe (a “radical clock”). Although earlier
studies by von Doering and Aschner ruled out a radical
mechanism,[13] some later work supported the possibility of one
electron transfers to substrates prone to stabilize radicals. For
example, the ketyl radical were detected in the MPV reduction of
benzophenone.[27] Cyclopropyl phenyl ketone would give different
products, depending on the mechanism of the reduction process.
When an MPV-like transfer of hydrogens occurs, only the C=O
bond undergoes transformation to the hydroxyl functionality
(Scheme 3a). But if a radical is generated during the reaction, it
can cause intramolecular isomerization, which would be triggered
by the steric strain of the cyclopropyl ring (Scheme 3b). In the
case of cyclopropyl phenyl ketone reduction under the alkali metal
catalyzed TH, only cyclopropyl(phenyl)methanol was observed,
suggesting an MPV mechanism and not a single electron transfer.
Further insight into the mechanism of transfer hydrogenation of
ketone was provided by kinetic isotope effect measurements
carried out in isopropanol with 10 mol% lithium isopropoxide and
10% TMEDA. Benzophenone, a non-enolizable ketone, was
chosen as a substrate to avoid any side effects, which can be
caused by enolization. Comparison of the rates obtained in
HOCHMe2 and DOCHMe2 fetched a small KIE of 1.1, indicating
that proton transfer is not involved in the rate determining step
(RDS). On the other hand, a significant KIE of 3.6 was obtained
for the reaction carried out in the fully deuterated isopropanol
DOCD(CD3)2, which is consistent with the C-D bond cleavage in
the RDS. This result corroborates further the suggestion that
reaction proceeds via direct hydrogen transfer from the alkoxide
to carbonyl.
Figure 3. Hammett plot for lithium cation-catalyzed transfer hydrogenation of
acetophenones in isopropanol.
Taken together, these kinetic data underpin the Woodward’s
proposal[12] that alkali base-catalyzed transfer hydrogenation
proceeds via a six-membered cyclic transition state similar to the
conventional
mechanism
of
Meerwein–Ponndorf–Verley
reduction mediated by aluminum alkoxides.
Discussion of experimental results
The understanding of alkali metal catalyzed reduction is very
important in the context of development of more benign and
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
sustainable synthetic processes that would circumvent the use of
toxic and expensive transition metals and minimize the production
of waste. The latter aspect is of great concern in the traditional
aluminum-based MPV catalysis which usually requires
stoichiometric amounts of aluminum alkoxide or alkyl reagents.[7]
So far, the use of alkali metals in the MPV reactions has been
very limited because of the low activity and the need of using
increased amounts of the catalyst.[7a] In this study we show that
this problem can be mitigated by the application of ligands and
promoters.
Before discussing the alkali metal catalysis, one question should
be addressed: Is it possible that catalysis is triggered by traces of
transition metals? The detailed study by Ouali et al. shows that
transition metals are not responsible for the observed base
catalysis and, in fact, their addition has a detrimental effect.[19a]
This finding may explain why large excess of alkali base is
required in some “transition metal-catalyzed” transfer
hydrogenations.
transfer interactions between the aromatic rings in the
intermediates [(PhMeC=O)Li(HOCHMePh)2(OCHMePh)] and
[Li(HOCHMePh)3(OCHMePh)]. Indirectly supporting this idea is
the observation that Li-catalyzed reduction noticeably slows down
at about 40% conversion, which corresponds to four molecules of
the product per the alkali metal ion (at the 10% catalyst load). If
this hypothesis is correct, the stabilizing aromatic interactions are
weakened in the case of sodium and potassium because their
larger size places the aromatic groups of ligated 1-phenylethanol
farther away. This stabilizing effect should be absent in the case
of cyclohexanone which is reduced quicker on the lithium center
because of its high Lewis acidity and stronger activation of the
carbonyl function.
The inhibiting influence of 1-phenylethanol was studied next. TH
of acetophenone was performed with addition of 40 mol% of 1-
phenylethanol before the reaction was launched at reflux. The
process was much slower comparing to the reaction without
additional 1-phenylethanol (Figure S6).
The main puzzle of the alkali metal-catalyzed MPV reduction is
that the reactivity order in the transfer hydrogenation of
acetophenone, the most common model substrate for the TH, is
Na > K > Li. This order is counterintuitive because MPV requires
the presence of a strong Lewis acid to polarize and activate the
C=O bond, whereas the order of Lewis acidity of alkali metals is
Li > Na > K. Ouali et al. explained this reactivity order by the need
to balance the substrate activation step with the rate of product
de-coordination, which requires a weaker Lewis acid, so that the
maximum activity is observed with sodium.[19a] However, the
ligand exchange for alkali metal ions is known to be very fast.[7a]
The lack of reactivity by the application of ylide H2C=PPh3 alone
versus the productive catalysis in the case of a combined action
of ylide and LiCl illustrates the need of a Lewis acid. So, how can
one explain the abnormal reactivity order for alkali metals? The
change of the reactivity order to the expected Li > Na> K in the
case of cyclohexanone shows conclusively that the reactivity is
substrate-dependent and therefore the abnormal behavior should
be caused by the presence of the arene ring in acetophenone.
A seemingly obvious explanation is that the change of reactivity
is caused by specific interactions between the alkali metal ion and
the aromatic ring. Indeed, alkali cation- interactions are very well
established, so that alkali cations can be solvated by aromatic
molecules through interaction with -electrons.[28] However, Kochi
et al. conclusively demonstrated that the strength of alkali
metal/aromatic interactions increases down Group 1, with the
sodium cation (the smallest studied) showing no sign of Na+…-
interactions.[29] Therefore, it is unlikely to play any major role in
the lithium catalysis, and if this effect were operating, the activity
in the TH should have changed monotonously down the group.
Since Lewis acidity of the lithium cation is not a decisive factor in
impeding the reduction of acetophenone, we envisaged that it
could be caused by the different stabilization of the product, 1-
phenylethanol, in solution and in complex. Because 1-
phenylethanol is a relatively large molecule, it disrupts the
hydrogen bonding network of the solvent (isopropanol). On the
other hand, placing two or more molecules of 1-phenylethanol in
the coordination sphere of a lithium ion may allow for additional
stabilization through the - stacking interactions or charge-
The ligand effect in the case of acetophenone is then explained
by a dual phenomenon. First, strong ligands, and in particular
chelating ligands, can coordinate to lithium, thus preventing the
accumulation of 1-phenylethanol in the coordination sphere of the
cation. The enhanced activity of soft ligands, e.g. carbene, versus
hard nitrogen- and oxygen-based donors is likely caused by the
same reason: the former type of ligands are stabilized worse by
the polar media than the latter and tend to bind the lithium cation
better. Second, the need to stabilize the product, 1-phenylethanol,
in solution is nicely illustrated by the effect of aromatic additives.
Small amounts of toluene and other aromatics (and likely Ph-
containing ligands, such as dppe) solvate 1-phenylethanol more
effectively than isopropanol, likely by means of - interactions
and help the product leave the coordination sphere of lithium, thus
freeing the catalyst.
Another way to break the proposed arene/arene -interactions is
to increase the temperature. This allows for the need of very high
temperature (180 °C) in the lithium isopropoxide-catalyzed
reduction of ketones reported by Adolfsson et al.[18] At such
temperatures, nearly quantitative formation of targeted alcohols is
achieved.
To test further our hypothesis, we carried out reduction of
acetophenone in the presence of 10 mol% lithium isopropoxide
and a stoichiometric amount of LiCl. Fast reduction was observed,
reaching 79% conversion after only 2 h versus 60% conversion in
the absence of this additive (the equilibrium value was
accomplished within 3 h). The tendency for acceleration of
transfer hydrogenation of other aromatic substrates (Scheme 4)
was similar to the one with the addition of dppe. And no reaction
rate improvement was observed for cyclohexanone.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
DFT calculations
To shed additional light on the mechanism of these alkali-metal
catalyzed transfer hydrogenations, DFT calculations exploring a
reaction pathway involving direct H-transfer via a cyclic six-
membered transition state were performed. Emerging from these
calculations were pathways initiating from tetra-coordinated alkali
metal complexes 4Li,Na,K (Figure 4). From this complex, hydride
transfer by cyclic transition states TS1Li,Na,K ensues with modest
activation barriers of 12.5 kcal mol-1, 10.1 kcal mol-1, and 11.2 kcal
mol-1, respectively, to afford complex 5Li,Na,K. Notably, these
calculated barriers are in line with the observed Na+ > K+ > Li+
order of reactivity found experimentally for the ligand/additive-free
catalytic processes (vide supra). The salient features of these
transition states included nearly equivalent bond forming
C(1)•••H(2) and bond breaking C(3)•••H(2) distances of 1.33-1.37
Scheme 4. Reaction conditions: 10 mol% LiOiPr, 1eq. LiCl in isopropanol (1 ml).
[a] No LiCl was added.
Figure 4. Energy profile corresponding to concerted model for H-transfer from isopropoxy groups to acetophenone with the incorporation of TMEDA ligand.
Calculated relative free energies (kcal mol-1) at SMD(Isopropanol)/B3LYP-D3//6-31+G(d,p) level of theory are shown for transitions states with TMEDA ligand and
with isopropanol ligand.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
Table 5. Bond making, bond breaking and metal cation heteroatom coordination distances of TS1Li,Na,K and TS2Li,Na,K computed at SMD(Isopropanol)/B3LYP-
D3//6-31+G(d,p) level of theory. All bond distances are reported in ångströms.
Structure
TS1Li
TS1Na
TS1K
Structure
TS2Li
TS2Na
TS2K
C(1)•••H(2)
C(3)•••H(2)
O(5)•••M+
O(4)•••M+
L(6)•••M+
L(7)•••M+
1.33
1.35
1.89
1.89
2.14
2.11
1.34
1.35
2.26
2.24
2.47
2.46
1.34
1.37
2.63
2.75
2.94
2.96
C(1)•••H(2)
C(3)•••H(2)
O(5)•••M+
O(4)•••M+
L(6)•••M+
L(7)•••M+
1.33
1.36
1.90
1.90
2.00
1.96
1.34
1.37
2.25
2.28
2.33
2.32
1.35
1.36
2.68
2.71
2.75
2.79
Å with little dependence on the alkali metal cation (Table 5). In
contrast, there was visible elongation of the alkali metal-ligand
bonds (to L(6) and L(7)) from 1.96 Å to 2.79 Å upon descending
the group I series. Likewise, the distances between the metal
cation and carbonyl oxygen O(4) of the substrate increased from
1.90 Å to 2.71 Å.
Intrigued by the role of supporting ligands in these reductions, we
next examined the effect of exchanging the coordinated
isopropanols for a chelating TMEDA ligand, which notably
resulted in more stable complexes. Thus, substitution of two
molecules of isopropanol in complexes 4M results in complexes
6M (M= Li, Na, K), with respective relative free energies of -50.87,
-50.95 and -49.10 kcal mol-1. Like in the case of isopropanol
ligand, reduction of acetophenone in the presence of TMEDA
complexes exhibited qualitatively similar geometrical trends with
respect to the cyclic six-membered transition state subassembly
TS2K (O(5)-C(1)-C(3)-O(4) = 11.7̊ vs 27.4̊ ) ascribed to the larger ionic
radius of the potassium cation allowing for greater flexibility in the
cyclic six-membered transition state fragment. However, in TS2K
the larger dihedral angle creates steric contacts as seen by an
H•••H distance of 2.37 Å, resulting in transition state
destabilization. Further, the relative free energies of transition
states TS1Li,Na,K and TS2Li,Na,K revealed the later series of first-
order saddle points were energetically more stable (see SI).
To better understand the origin of this intriguing divergence in
reactivity, a distortion/interaction (D/I) analysis was applied. In this
context, D/I analysis is a useful tool for analyzing activation
barriers in terms of the energy required to distort the reactants to
their transition state geometries and the affinities to bring together
these fragments, which in turn provides insights into the factors
controlling reactivity.[30] The energy, namely the distortion energy
or activation strain, constitutes the major component of the
activation energy. To overcome the distortion energy leading to
the products, there is the requirement for strong bonding
interactions between the two reactants, referred to as the
interaction energy.
(Table 5, right). For instance, in transition states TS2Li,Na,K
,
derived from complex 6Li,Na,K and affording complex 7Li,Na,K, the
bonds from the acetophenone and isopropoxide oxygen atoms to
the alkali metal cations, O(4)•••M+ and O(5)•••M+ (M+ = Li+, Na+,
K+) were comparable for each cation (Figure 4). Further, there
was a steady elongation from 1.89 Å to 2.75 Å in going from
lithium to potassium (Figure 5). Meanwhile, the alkali metal cation
Turning to the case at hand, the D/I analysis for
TS1Li,Na,K/TS2Li,Na,K was broken down as follows: equation 4
describes the distortion energy (Edist) arising from ligand
to TMEDA nitrogen atom bond distances, N(6)•••M+ and N(7)•••M+, (TMEDA or HOiPr) association to the acetone-acetophenone
varied from 1.96 Å to 2.79 Å. By the same token, the bond forming
C(1)•••H(2) and bond breaking C(3)•••H(2) distances were nearly
equivalent 1.33-1.37 Å, and very close to the distances observed
with isopropanol ligands. Despite these structural similarities
between the isopropanol and TMEDA series, the G≠ values for
the hydride transfer transition states progressively increased
down group I in the order 10.5 kcal mol-1, 11.1 kcal mol-1, and 12.4
kcal mol-1, in excellent agreement with our experimental
observation. The lower activation barrier in the presence of
complex 8Li,Na,K as found in their transition state geometries. This
can be further broken down into three components, E1 is the
distortion imposed by ligand dissociation from the precomplex
4Li,Na,K or 6Li,Na,K (Figure 5). The next components, E2 and E3,
represent the strain incurred by perturbing the acetone-
acetophenone metal complex and the ligand fragments,
respectively, from their minimized geometries to the transition
state geometries. Equation 5 describes the relationship between
the distortion energy Edist and the interaction energy Eint of the
TMEDA relative to isopropanol ligands in the case of lithium cation, ligand chelation to the acetone-acetophenone complex to give the
i.e., G≠ = 2.0 kcal mol-1, is noteworthy. Contributing to this
difference, in part, were unfavorable eclipsing interactions as
seen from the O(5)C(1)•••C(3)O(4) dihedral angles of TS2Li and
transition state structure. The interplay of the distortion energy
and interaction energy results in the activation energy Eact
.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
∆퐸푑ꢀ푠푡 = ∆퐸1 + ∆퐸2 + ∆퐸3
∆퐸푎푐푡 = ∆퐸푑ꢀ푠푡 + ∆퐸ꢀ푛푡
(4)
(5)
were 16.3 kcal mol-1, 13.5 kcal mol-1, and 14.3 kcal mol-1, while
those for TS1Li,Na,K were 17.0 kcal mol-1, 14.2 kcal mol-1, and 13.9
kcal mol-1, respectively. By comparison, the distortion energies
E1 were similar for both lithium and sodium cation-based
systems, whereas with potassium the energy was ~5.0 kcal mol-1
lower.
Figure 6 shows the distortion-interaction analysis for transition
states TS1Li,Na,K and TS2Li,Na,K. Evident from this analysis were
similar Edist energies for each metal cation. What is more, the
ligand distortion energy E3, irrespective of isopropanol or
TMEDA as the ligand, was consistently smaller than the distortion
energy E2 for the acetone-acetophenone fragment or E1.
Further, E2 was greater for lithium relative to sodium or
potassium cation for both isopropanol and TMEDA coordinated
complexes. For example, the distortion energies E2 for TS2Li,Na,K
The interaction energy Eint of TS1Li was lower than TS2Li by 2.2
kcal mol-1, while for sodium and potassium cation the interaction
energies decreased chromatically. Ultimately, this manifests in a
lower activation barrier (Eact) for the reduction of acetophenone
in the presence of lithium cation and ligand TMEDA, thus
correlating with the experimentally observed rate acceleration in
the presence of an additive. Conversely, an analogous ligand-
based trend on Eint energies (isopropanol vs TMEDA) was not
observed for sodium or potassium cation catalysis, which is again
consistent with the experiment. From these energetics it is clear
that the lower activation barrier for lithium cation catalysis derives
in large part from larger interaction energies relative to sodium or
potassium cation catalysis, which is likely caused by the higher
charge density of the smaller lithium cation. It is also meaningful
to note that the observed rate acceleration conferred by ligand
TMEDA relative to isopropanol is accounted for by the 2.2 kcal
mol-1 difference in Eint values.
As for the deceleration of lithium cation catalysis in the absence
of additional ligands, we attribute this to specific interligand
interactions between 1-phenylethanol and acetophenone.
Consistent with this was the fact that Li-catalyzed reductions
noticeably slowed down at about 40% conversion, corresponding
to about four molecules of the product per the alkali metal ion (at
the 10% catalyst load), while Na+ and K+ catalyzed reductions
slowed gradually. This divergency is ascribed to a decreased rate
of product de-coordination and/or ligand exchange at the metal
center of Li+ complexes relative to Na+ and K+ complexes. To
probe this hypothesis, non-covalent interaction (NCI) plots were
computed of complexes 9Li,Na,K corresponding to the lowest
energy structures of product coordination (Figure 7). Clear from
these plots was greater phenylethanol coordination to the alkali
metal ion and greater shielding of the metal cation in the case of
9Li relative to 9Na.
[31] Contributing to this in part were non-covalent
interactions between the aryl moiety of 1-phenylethanol and the
carbonyl group of acetophenone in Li+ complex 9Li. Conversely,
in the cases of Na+ and K+ complexes, 9Na,K, because of their
larger ionic radius the aromatic groups reside farther away from
the alkali metal resulting in attenuated -orbital and CH/
interactions between coordinated phenylethanol molecules (see
SI for select interaction distances).
Figure 5. Distortion/interaction analysis for TS1Li,Na,K/TS2Li,Na,K. Activation
energy (black arrow); interaction energy (red arrow); distortion energy (E1) for
ligand dissociation from precomplex 4Li,Na,K/6Li,Na,K (green arrow); distortion
energy (E2) of the acetone-acetophenone metal complex 8Li,Na,K (blue arrow);
distortion energy (E3) of the ligand (orange arrow). Calculated energies are
shown in kcal mol-1. M+ = Li+, Na+, K+.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
Figure 6. D/I analysis for transition states TS1Li,Na,K and TS2Li,Na,K. Activation energy (black arrow); interaction energy (red arrow); distortion energy (E1) for ligand
dissociation from the precomplex (green arrow); distortion energy (E2) of the acetone-acetophenone metal complex (blue arrow); distortion energy (E3) of the
ligand (orange arrow). Calculated energies are shown in kcal mol-1.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
Figure 7. Non-covalent interaction (NCI) plots for structures 9Li,Na,K displaying interactions between the aryl moiety of 1-phenylethanol and the carbonyl group of
acetophenone.
specific non-covalent interactions, such as charge transfer,
between the aryl moiety of 1-phenylethanol product and the
carbonyl group of acetophenone, which hampers product
dissociation from the catalytic center. This observed
Conclusions
ligand/additive effect is likely a combination of several factors,
including a change in activity of alkali metal cations and the
relative stabilization of reagents and substrates in solution and the
coordination sphere of the metal. For the case of isopropanol and
diamine ligands, the D/I calculations revealed a noticeable
decrease of activation energy of ligation by a diamine (TMEDA),
while in the cases of sodium and potassium cations there was
only a marginal impact. Other additives can likely stabilize better
the product in solution by solvation (e.g. excess toluene) or initiate
a new reaction center (the case of LiCl).
In conclusion, our combined experimental and computational
study of the alkali metal base-catalyzed transfer hydrogenation of
ketones revealed an important dependency on the alkali metal
cation employed and offered insights into the mechanism of these
reactions. In particular, the counterintuitive reactivity order for
catalytic reduction of acetophenone, Na+ > Li+ > K+, can be
changed to the expected reactivity pattern Li+ > Na+ > K+ by
adding ligands and additives. These additives accelerate the
transfer hydrogenation significantly in the case of aromatic
substrates, but not for aliphatic ketones, which usually react fast
and show the expected decrease of activity down Group 1. The
reaction can be also easily accelerated by adding excess of a
cheap and benign reagent, LiCl. Kinetic, labelling, and
competition experiments, supported by DFT calculations, point to
Author Contributions
a
concerted direct hydride transfer mechanism, originally
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the
manuscript.
suggested by Woodward, as the principle reaction pathway. The
experimentally observed deceleration of lithium cation-catalyzed
reaction at ~40% conversion was explained by the presence of
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
[16] D. Wang, D. Astruc, Chem. Rev. 2015, 115, 6621-6686.
Funding Sources
[17] a) J. R. Miecznikowski, R. H. Crabtree, Organometallics 2004, 23, 629-
631; b) M. D. le Page, D. Poon, B. R. James, in Catal. Org. React., Vol.
89, Dekker, 2003, p. 61.
[18] J. Ekstrom, J. Wettergren, H. Adolfsson, Adv. Synth. Catal. 2007, 349,
1609-1613.
This research was supported by NSERC (Discovery Grants DG
2012-326899 and 2017-05231 to G.I.N. and 2014-04410 to T.D.)
[19] a) A. Ouali, J. P. Majoral, A. M. Caminade, M. Taillefer, ChemCatChem
2009, 1, 504-509; b) V. Polshettiwar, R. S. Varma, Green Chem. 2009,
11, 1313-1316; c) D. Wang, C. Deraedt, J. Ruiz, D. Astruc, J. Mol. Catal.
A: Chem. 2015, 400, 14-21; d) R. Radhakrishan, D. M. Do, S. Jaenicke,
Y. Sasson, G. K. Chuah, ACS Catal. 2011, 1, 1631-1636.
[20] a) Q. Xu, J. H. Chen, Q. Liu, Adv. Synth. Catal. 2013, 355, 697-704; b)
Y. F. Liang, X. F. Zhou, S. Y. Tang, Y. B. Huang, Y. S. Feng, H. J. Xu,
RSC Adv. 2013, 3, 7739-7742; c) L. J. Allen, R. H. Crabtree, Green
Chem. 2010, 12, 1362-1364.
[21] The experimental part of this work was close to completion when
Stephan et al. reported the first example of alkali metal catalyzed
hydrogenation: a) M. T. Xu, A. R. Jupp, Z. W. Qu, D. W. Stephan, Angew.
Chem. Int. Edit. 2018, 57, 11050-11054; for a related alkaline earth metal
catalyzed hydrogenation, see: b) H. Bauer, M. Alonso, C. Farber, H.
Elsen, J. Pahl, A. Causero, G. Ballmann, F. De Proft, S. Harder, Nat.
Catal. 2018, 1, 40-47; for hydrogenation of aldimines and ketimines
catalyzed by a lithium amide, see: c) D. C. Elliott, A. Marti, P. Mauleon,
A. Pfaltz, Chem. Eur. J. 2019, 25, 1918-1922.
Acknowledgements
We thank SHARCNET (Shared Hierarchical Academic Research
Computing Network: www.sharcnet.ca) and Compute/Calcul
Canada for computing resources.
Keywords: lithium • transfer hydrogenation • ligands • DFT
calculation
[1]
H. Meerwein, R. Schmidt, Justus Liebigs Ann. Chem. 1925, 444, 221-
238.
[22] pKa for [Ph3PCH3]+ and isopropanol in DMSO is 22.4 and 29.3,
respectively.
[23] For a vast series of oxidation potentials of carbonyl compounds, see: H.
Adkins, R. M. Elofson, A. G. Rossow, C. C. Robinson, J. Am. Chem. Soc.
1949, 71, 3622-3629.
[24] F. W. Cox, H. Adkins, J. Amer. Chem. Soc. 1939, 61, 3364-3370.
[25] B. L. Feringa, A. Smaardijk, H. Wynberg, J. Am. Chem. Soc. 1985, 107,
4798-4799.
[26] For example: a) D. J. Gallagher, H. Du, S. A. Long, P. Beak, J. Am. Chem.
Soc. 1996, 118, 11391-11398; b) W.-P. Deng, V. Snieckus, C. Metallinos,
in Chiral Ferrocenes in Asymmetric Catalysis: Synthesis and
Applications (Eds.: L.-X. Dai, X.-L. Hou), Wiley-VCH Verlag GmbH & Co.
KGaA, Weinheim, Germany, 2010, pp. 15-53.
[27] E. C. Ashby, J. N. Argyropoulos, J. Org. Chem. 1986, 51, 472-476.
[28] A. S. Mahadevi, G. N. Sastry, Chem. Rev. 2013, 113, 2100-2138.
[29] G. K. Fukin, S. V. Linderman, J. K. Kochi, J. Am. Chem. Soc. 2002, 124,
8329-8336.
[30] a) D. H. Ess, K. N. Houk, J. Am. Chem. Soc. 2007, 129, 10646; b) D. H.
Ess, K. N. Houk, J. Am. Chem. Soc. 2008, 130, 10187-10198. c) F. M.
Bickelhaupt, K. N. Houk, Angew. Chem. Int. Ed. 2017, 56, 10070-10086.
[31] a) Schrödinger Release 2018-4: Jaguar, Schrödinger, LLC, New York,
NY, 2018; b) E. R. Johnson, S. Keinan, P. Mori-Sanchez, J. Contreras-
Garcia, A. J. Cohen, W. T. Yang, J. Am. Chem. Soc. 2010, 132, 6498-
6506.
[2]
[3]
[4]
A. Verley, Bull. Soc. Chim. Fr. 1925, 37, 537-542.
H. W. Knipping, W. Ponndorf, H-S Z Physiol. Chem. 1926, 160, 25-60.
E. D. Williams, K. A. Krieger, A. R. Day, J. Am. Chem. Soc. 1953, 75,
2404-2407.
[5]
[6]
V. J. Shiner, D. Whittaker, J. Am. Chem. Soc. 1969, 91, 394-398.
R. Cohen, C. R. Graves, S. B. T. Nguyen, J. M. L. Martin, M. A. Ratner,
J. Am. Chem. Soc. 2004, 126, 14796-14803.
[7]
a) C. F. de Graauw, J. A. Peters, H. van Bekkum, J. Huskens, Synthesis-
Stuttgart 1994, 1007-1017; b) J. S. Cha, Org. Proc. Res. Dev. 2006, 10,
1032-1053; c) C. R. Graves, E. J. Campbell, S. T. Nguyen, Tetrahedron
Asym. 2005, 16, 3460-3468.
[8]
[9]
For example: E. J. Campbell, H. Y. Zhou, S. T. Nguyen, Abstr. Pap. Am.
Chem. S. 2002, 224, U134-U134.
a) Y. C. Liu, B. T. Ko, B. H. Huang, C. C. Lin, Organometallics 2002, 21,
2066-2069; b) B. McNerney, B. Whittlesey, D. B. Cordes, C. Krempner,
Chem. Eur. J. 2014, 20, 14959-14964.
[10] a) S. Cannizzaro, Justus Liebigs Ann. Chem. 1853, 88, 129-130; b) L.
Hammett, Physical Organic Chemistry, McGraw-Hill, New York, 1940.
[11] R. V. Oppenauer, Recl. Trav. Chim. Pays-Bas 1937, 56, 137-144.
[12] R. B. Woodward, N. L. Wendler, F. J. Brutschy, J. Am. Chem. Soc. 1945,
67, 1425-1429.
[13] W. van Doering, T. C. Aschner, J. Am. Chem. Soc. 1953, 75, 393-397.
[14] For example, de Graauw, Peters, van Bekkum, and Huskens concluded
that alkali metal alkoxides cannot be used as catalysts in the MPV
reduction because of low charge density of sodium and pottassium and
low coordination number of lithium, see ref 7a.
[15] R. L. Chowdhury, J. E. Backvall, J. Chem. Soc. Chem. Comm. 1991,
1063-1064.
This article is protected by copyright. All rights reserved.

10.1002/chem.201902240
Chemistry - A European Journal
FULL PAPER
FULL PAPER
This paper unveils the reactivity
patterns, as well as ligand and additive
effect on alkali metal base catalyzed
transfer hydrogenation of ketones.
Kinetic, labelling, and competition
experiments and DFT calculations
suggested that the reaction proceeds
via a concerted direct hydride transfer
mechanism.
Iryna D. Alshakova, Hayden C. Foy,
Travis Dudding, and Georgii I. Nikonov*
Page No. – Page No.
Ligand Effect in Alkali Metal-
Catalyzed Transfer Hydrogenation of
Ketones
This article is protected by copyright. All rights reserved.