Asymmetric Carboxycyanation of Aldehydes by Cooperative AlF/Onium Salt Catalysts: from Cyanoformate to KCN as Cyanide Source

Article abstract of DOI:10.1002/chem.201804388

Asymmetric 1,2-additions of cyanide yield enantioenriched cyanohydrins as versatile chiral building blocks. Next to HCN, volatile organic cyanide sources are usually used. Among them, cyanoformates are more attractive on technical scale than TMSCN for cos

Full text of DOI:10.1002/chem.201804388

A Journal of
Accepted Article
Title: Asymmetric Carboxycyanation of Aldehydes by Cooperative AlF-/
Onium Salt Catalysts: from Cyanoformate to KCN as Cyanide
Source
Authors: René Peters, Johannes Kästner, and Daniel Brodbeck
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.201804388
Link to VoR: http://dx.doi.org/10.1002/chem.201804388
Supported by

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
Asymmetric Carboxycyanation of Aldehydes by Cooperative AlF /
Onium Salt Catalysts: from Cyanoformate to KCN as Cyanide
Source
Daniel Brodbeck,[a] Sonia Álvarez-Barcia,[b] Jan Meisner,[b] Florian Broghammer,[a] Julian Klepp,[a]
Delphine Garnier,[a] Wolfgang Frey,[a] Johannes Kästner,*[b] and René Peters*[a]
Abstract:
Asymmetric
1,2-additions
of
cyanide
yield
sensitive technical scale applications the use of TMSCN is too
expensive in terms of follow-up costs.[8] In this respect,
cyanoformates appear to be much more attractive as cyanide
source. As an additional benefit, the carboxy groups are not
necessarily protecting groups only,[9] but they also served for
elegant further transformations like [3,3]-rearrangements and
enantiospecific allylic substitutions.[10] Nevertheless, the most
attractive cyanide source for a user should be inorganic
cyanide salts like KCN, which combines a low price and low
volatility. To the best of our knowledge, highly enantioselective
catalytic carboxycyanations of aldehydes using KCN as the
enantioenriched cyanohydrins as versatile chiral building blocks.
Next to HCN, volatile organic cyanide sources are usually used.
Among them cyanoformates are more attractive on technical scale
than TMSCN for cost reasons, but catalytic productivity is usually
lower. Here we describe the development of a new strategy for
cyanations by which this activity disadvantage is overcome. A
Lewis acidic Al center cooperates with an aprotic onium moiety
within a bifunctional AlF-salen complex which is remarkably
robust. This allowed for unprecedented TONs of up to 104. DFT
studies suggest an unexpected unique trimolecular pathway in
which the ammonium bound cyanide attacks the aldehyde which
itself is activated by the carbonyl group of the cyanoformate
exclusive
cyanide
source
have,
in
contrast
to
acylcyanations,[11] not been reported to date.
binding to the Al center. In addition,
a
novel practical
A general issue for carboxycyanations has been a significantly
lower catalytic activity compared to cyanosilylations,[12,13] thus
requiring catalyst loadings in a range of 1-10 mol% for the best
reported catalyst systems.[2a] This might be partly explained by
catalyst inhibition by the Lewis basic carboxy protective groups.
In the development of these latter methods cooperative
catalysts emerged which often provided significantly higher
enantioselectivity and catalytic activity compared to more
traditional catalysts lacking cooperative activation modes.[14-16]
Recently, we reported a new strategy for the carboxycyanation
of aldehydes to permit unprecedented TONs.[17] In this
approach a Lewis acidic Al center cooperates with an internal
ammonium salt moiety.[18] The idea behind was that a loosely
bound cyanide should be highly reactive in the 1,2-addition
step to a Lewis-acid-activated aldehyde.[18] We found that very
high catalyst efficiency was enabled by a considerably Lewis-
acidic Al–F unit which at the same time leads to an
extraordinarily stable catalyst. In this full paper we provide full
account of this synthetic development.
carboxycyanation method was developed which makes use of
KCN as the sole cyanide source. The use of a pyrocarbonate as
carboxylating reagent provided the best results.
Introduction
Nucleophilic cyanations of aldehydes are among the most
important 1,2-addition reactions and are widely employed both
in academic laboratories as well as for technical scale
production processes due to the synthetic versatility of chiral
cyanohydrine derivatives.[1-3] Industrial scale applications
usually make use of enzyme catalyzed additions of hydrogen
cyanide.[4,5] Nevertheless, employing HCN bears serious risks
due to the combination of its high volatility and toxicity. In
academic studies less volatile cyanation reagents such as
trimethylsilyl cyanide and ethyl cyanoformate are usually
preferred.[2,6,7] These reagents also offer the advantage that O-
Detailed computational studies provided insight into possible
reaction mechanisms and revealed that besides the initially
anticipated catalytic product formation pathway an unexpected
trimolecular pathway might play a role. Furthermore, we
protected cyanohydrin products less easily undergo
racemization, because the addition reaction becomes
irreversible.[Fehler! Textmarke nicht definiert.] However, for many cost-
a
present the development and application of
operationally attractive method that forms
a
new,
the
carboxycyanation product with KCN as the sole cyanide source.
[a]
[b]
M.Sc. D. Brodbeck, Dr. F. Broghammer, M.Sc. J. Klepp, Dr. D.
Garnier, Dr. W. Frey, Prof. Dr. R. Peters
Universität Stuttgart, Institut für Organische Chemie
Pfaffenwaldring 55, D-70569 Stuttgart, Germany
E-mail: rene.peters@oc.uni-stuttgart.de
Dr. Sonia Álvarez-Barcia, Dr. J. Meisner, Prof. Dr. J. Kästner
Universität Stuttgart, Institut für Theoretische Chemie
Pfaffenwaldring 55, D-70569 Stuttgart, Germany
E-mail: kaestner@theochem.uni-stuttgart.de
Results and Discussion
1. Development of a Highly Active Carboxycyanation
Catalyst using a Cyanoformate as Cyanide Source and
Mechanistic Studies
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.201804388
Chemistry - A European Journal
FULL PAPER
The Al-salen catalysts were formed in situ in our early studies
using Me3Al as Al-source and ligands such as L1, which were
equipped with two ammonium bromide side chains. The
solubility of these catalysts is problematic at low temperatures.
In CH2Cl2 at 50 °C activity and enantioselectivity were low
(Table 1, entry 1). Because (1) we were speculating that a
competition between both ammonium moieties might take
place leading to different enantioface differentiations in the 1,2-
addition step and (2) in order to increase the solubility of the
complexes, we were then studying catalysts bearing only one
ammonium bromide salt side chain.[18d] By that change the
solubility issue was successfully addressed. However, the
activity was still not useful and the enantioselectivity only
marginally improved (entry 2).
Because a cyanide/bromide exchange would be necessary in
the ammonium salt moiety to allow for the targeted mechanism,
the presence of KCN salt as additive was studied and allowed
for quantitative yields and a significant improvement of the
enantioselectivity (entry 3).
Table 1. Development of a potent cooperative bifunctional Lewis-acid / onium salt catalyst for the carboxycyanation of aldehydes.
ligand
(x)
yield
[%]a)
ee
NR2
X
n
R1
No.
R2Al-Y
equiv. 2a
y
[%]b)
3
1
L1 (5.0)
4
4
4
4
4
4
4
2
4
2
1
1
1
0
0
9
12
15
39
53
85
66
88
90
12
90
92
93
69
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlMe
Me2AlF
NMe2Bn
NMe2Bn
NMe2Bn
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
NEt2Me
Br
Br
Br
Br
1
1
1
1
1
2
3
3
3
3
3
3
3
CH2NMe2Bn/Br
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
2
L2 (5.0)
L2 (5.0)
L3 (5.0)
L4 (5.0)
L5 (5.0)
L6 (5.0)
L6 (5.0)
L7 (5.0)
L7 (5.0)
L7 (1.0)
L7 (0.1)
L7 (1.0)
20
3
2
100
100
100
63
4
5
2

2
BF4
BF4
BF4
BF4



6
2
7
2
100
100
99
8c)
1





9
2
F3CSO3
F3CSO3
F3CSO3
F3CSO3
F3CSO3
10c),d)
11c),d)
12 c),d),e)
13c),d)
1
100
100
100
Me2AlF
0.1
0.1
0.1
Me2AlF
12
Me2AlCl
a)
b)
c)
d)
Determined by 1H-NMR of the crude product using an internal standard. Determined by HPLC. The reaction was performed in CHCl3. Preformed,
isolated catalyst was used. e) Reaction time: 48 h.
Further rises in enantioselectivity were achieved by (1)
switching to a diethylmethylammonium side arm (entry 4) and
(2) by changing to a BF4 counterion (entry 5). The last result
catalyst decomposition via β-elimination in the ammonium side
arm, forming a resonance stabilized styrene derivative, seems
likely. With n = 3 ligand/catalyst decompositions are probably
less problematic and despite the high flexibility of the onium
arm the attained enantioselectivity was high.

appeared logical, because cyanide might more easily replace a

weakly binding counterion like BF4 .
In addition, different (CH2)n-linkers between the salen core and
the onium moieties were used with n = 1-3 (entries 5-7). The
best results in terms of enantioselectivity (ee = 88%) were
obtained for n = 3 (ligand L6, entry 7). In contrast, for n = 2 the
enantioselectivity was significantly lower (ligand L5, entry 6). A
Investigation of various solvents using the catalyst formed in-
situ from L6 and Me3Al revealed CHCl3 to be superior (entry 9).
In this case, the relative amounts of 2a and KCN could be
decreased while still allowing a quantitative yield and an
improved enantioselectivity (entry 8).
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
Surprising results were then obtained with complexes of this
catalytic activity, Me2AlF[20] was prepared and used for
complexation to provide the monomeric complex {AlF}OTf.
The solid state structure of the closely related {Al–F}PF6 could
be determined by X-ray single crystal structure analysis.[19]
Two different species with different configurations at the
stereogenic Al centers were identified in the solid state. Only
one of them is depicted in Figure 1, bottom. Two different
species were also present in chlorinated solvents at room
temperature as determined by NMR. The 19F NMR spectra
provided 2 signals of the Al coordinated F atoms. The
type carrying a different ‘non-nucleophilic’ counterion X


Replacing BF4 by triflate (F3CSO3 , OTf) in ligand L7 led to
almost racemic product (entry 9).
To find out more about the influence and role of the ‘non-
nucleophilic’ counterions X like BF4 and triflate, spectroscopic
studies were performed. They revealed that the less selective
catalyst generated from the triflate containing ligand L7 and
Me3Al was essentially one defined species. In contrast, the

much more selective catalyst formed from the BF4 containing
1
ligand L6 and Me3Al proved to be a complex mixture consisting
of a number of species as judged by 1H/19F NMR. We
hypothesized that an {AlF} catalyst might have been formed in
this case which might be much more active. The presumed
necessary methyl/fluoride exchange could be supported by
HR-MS data. In addition, crystallization attempts allowed to
generate single crystals of one of these species, which could
be analyzed by X-ray crystallography. This revealed a dimeric
aggregate with an anionic FAlFAlF axis (Figure 1, top).[19]
detection of a coalescence in H, 13C and 19F NMR at higher
temperatures, leading to one set of signals, is a strong
indication of interconversion between the two isomers that is
slower than NMR scale. While the two {Al–F} 19F signals merge
directly above room temperature (Tcoalescence (19F) = 303 K) into
one broad, temperature-dependent signal, coalescence in 1H-
NMR was observed at much higher temperature (Tcoalescence
(1H) = 363 K).
PGSE (Pulsed Gradient Spin Echo) NMR experiments[21]
performed on both {Al2F3} and {Al–F}OTf indicate that {Al2F3}
retains (almost entirely) a dimeric structure in CHCl3, whereas
{Al–F}OTf is monomeric (the Supporting Information contains
the details and results of the NMR experiments).
As hypothesized, the {Al–F}OTf catalyst showed a much
higher activity than the catalysts prepared from Me3Al (entries
10-12). For that reason, it was possible to (1) reduce the KCN
amounts from stoichiometric to catalytic (10 mol%), (2) to avoid
any excess of 2a, and (3) to reduce the catalyst loading from 5
to 0.1 mol%. Under these conditions the product was still
formed in quantitative yield and the ee value was improved to
93% (entry 12. The fluoride ligand effect is unique. For
instance, the AlCl counterpart to {Al–F}OTf was found to be
much less active and enantioselective (entry 13).
Using {Al–F}OTf the reaction scope was then investigated. A
catalyst loading of 0.1 mol% was used if not indicated
otherwise (Table 2).[22] In general, it can be stated that the
catalyst performance is not very sensitive towards electronic
effects. Aromatic aldehydes with σ- and -donors were well
tolerated (entries 2, 5-10, 12-13). Donor substituents in ortho-
positions resulted in a small decrease of the ee value, whereas
the reactivity was useful to high (entries 7, 12). Nearly
quantitative yields were obtained for substrates equipped with
- or -acceptor substituents (entries 11, 13-18) as well as with
2-furylcarbaldehyde (entry 19).
The reaction is not limited to aromatic aldehydes. In particular,
for enals high yields and enantioselectivities were attained. (E)-
Cinnamaldehyde gave product 3p in >99% yield and with 96%
ee (entry 20). A p-OMe group as a -donor substituent only
slightly reduced the reactivity (entry 21). An aromatic
substituent in -position is not crucial. -Acidic enals equipped
with alkyl residues at the -C atom work as well. Even the slim
crotonaldehyde 1v, in which steric interaction with the catalysts
might be reduced to a minimum, is well accepted (entry 22).
Figure 1. X-ray single crystal structure analyses of {Al2F3}BF4 (top)
containing a FAlFAlF axis and {AlF}PF6 (bottom). H atoms, included
solvent molecules and a second monomer per unit cell (for b), see SI) have
been omitted for clarity.
To get a monomeric Al-salen complex containing an AlF bond,
which would offer a free coordination site necessary for
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
Table 2. Investigation of the substrate scope.
For crotonaldehyde we studied if a further cutback of the
catalyst amount might be still feasible. In fact, the catalyst
performance was not disturbed by a catalyst loading of just
0.01 mol% of {Al–F}OTf. In that case, a highly enantioenriched
product was isolated in quantitative yield, being equivalent to a
turnover number of 10000 (entry 23).
Also, other enolizable substrates such as acetaldehyde 1w,
propanal 1x, butanal 1y and dihydrocinnamaldehyde 1z
(entries 24-27), as well as non-enolizable bulky aliphatic
substrates like pivaldehyde 1za (entry 28) are well
accommodated. Particularly noteworthy is the ee value of 80%
attained with the very volatile acetaldehyde 1w, since ee
values reported in literature for any kind of asymmetric 1,2-
cyanide addition reaction using this substrate were to the best
of our knowledge ≤ 45% so far.[23]
A gram scale experiment using cinnamaldehyde 1t was
conducted to demonstrate the preparative value of {Al–F}OTf.
The reaction was otherwise performed under the conditions
described in Table 2, entry 20. 3t was isolated in quantitative
yield and with 96% ee. To ascertain a high efficiency, it is
crucial to vigorously and continuously stir the reaction mixture.
We ascribe this to the phase transfer catalysis (PTC) character
of this reaction, in which solid KCN needs to deliver cyanide
counterions to form the active species of the catalyst. Another
practical benefit is the ease of recycling {Al–F}OTf which was
recovered unchanged in 72% yield by simple precipitation from
the reaction mixture using 1u as substrate as proven by 1H-
/19F-NMR & HR-MS. It was reused under identical conditions
giving 3u in quantitative yield and with 95% ee.
Another five reactions were performed on a 200 mg scale
under the same conditions. The results were similar to those
on the 0.1 mmol scale reported in Table 2 [1a (72 h): 100%,
93% ee; 1d (72 h): 98%, 91% ee; 1f (48 h): 100%, 90% ee; 1t
(72 h): 100%, 96% ee; 1u (72 h): 92%, 96% ee].
Control experiments were performed with the simple chiral Al-
salen complex K1 which is not equipped with an appended
ammonium moiety (Table 3). Previously, Jacobsen et al. have
shown that salen Al-complexes are very efficient catalysts in
the cyanation of other substrates than aldehydes.[24]
1/3
catalyst
loading x
[mol%]
t
yield
ee
No.
R
[h] [%]a)
[%]b)
1
a
b
c
d
e
f
2-naphthyl-
6-MeO-2-naphthyl-
1-naphthyl-
Ph-
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.5
0.1
0.5
0.1
0.1
0.5
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.01
0.1
0.1
0.1
0.1
0.1
24 >99
93
93
79
91
88
90
82
85
86
93
92
84
96
89
79
89
78
80
82
96
97
94
93
80
81
78
90
78
2
24
24
24
72
48
48
72
72
72
92
80
99
85
98
99
51
83
78
3
4
5
4-Me-C6H4-
3-Me-C6H4-
2-Me-C6H4-
4-tBu-C6H4-
3,4-Me2-C6H3-
4-MeO-C6H4-
3-MeO-C6H4-
2-MeO-C6H4-
3,4-(MeO)2-C6H3-
4-Cl-C6H4-
6
7
g
h
i
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 c)
23 d)
24 c)
25 c)
26 c)
27 c)
28 c)
j
k
l
24 >99
48
72
61
97
m
n
o
p
q
r
48 >99
24 93
48 >99
3-Cl-C6H4-
4-F-C6H4-
3-F-C6H4-
24
48
48
98
99
98
3-MeO2C-C6H4-
2-furyl-
Table 3. Control experiments.
s
t
(E)-Ph-CH=CH-
(E)-4-MeO-C6H4-CH=CH-
(E)-Me-CH=CH-
(E)-Me-CH=CH-
Me-
48 >99
u
v
v
w
x
y
z
48
72
90
96
72 >99
y (equiv.) additive (mol%) yield [%]a) ee [%]b)
48 >99
No.
1
x
5
equiv. 2
Et-
72
72
72
72
99
89
99
99
2.0
1.0
1.0
0.5
0.2
0.2
-
12
31
23
17
2
nPr-
2
0.1
-
[Et4N]OTf (0.1)
[Et4N]OTf (0.1)
PhCH2CH2-
tBu
3
-
za
a)
Determined by 1H-NMR of the crude product using an internal standard.
b) Determined by HPLC.
a)
b)
Yield of isolated product after column chromatography. Determined by
HPLC. c) Reaction in CH2Cl2/CHCl3 (1:1) at 80 °C. d) 0.05 equiv. of KCN.
Even at high catalyst loadings K1 gave very small amounts of
3a with low ee, despite working with an excess of 2 and
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
relatively large amounts of KCN (entry 1). In the presence of
[Et4N]OTf as catalytic additive the reactivity could be increased,
but the product was nearly racemic, while the catalytic activity
was much lower than with the bifunctional {Al–F}OTf (entry 2).
[Et4N]OTf also shows some activity in the absence of K1 (entry
3). The control experiments thus support the importance of the
ammonium moiety in the dual activation catalyst {Al–F}OTf.
In Scheme 1 the initially expected mechanism is shown. KCN
is probably improving the reactivity by generation of
{AlF}CN.[25] The aldehyde substrate 1 could be activated by
coordination to the Al center, facilitating a quasi-intramolecular
attack of cyanide. The absence of a non-linear effect (see the
Supporting Information) is in agreement with a mechanism in
which only one catalyst molecule is involved. Product release
and catalyst regeneration would be achieved by carboxylation
with 2.
situation around the aluminum center for X = F, Cl and Me
(Figure 2). The calculated AlF distance of 1.69 Å is in very
good agreement with the solid state structure (1.688-1.702 Å).
The IBO analysis shows that the AlF -bond is strongly
polarized towards the
F atom and only 0.23 of the
corresponding bond charge (12%) is located at the Al atom.
For Cl and Me values of 0.38 and 0.50, respectively, were
calculated. As a consequence, the central metal in the AlF
complex possesses a higher partial charge (1.26) than in the
Cl and Me complexes (1.09 and 1.11, respectively),
confirming a higher Lewis acidity. The same trend was found
by a Mulliken population analysis as shown in the Supporting
Information. The AlF bond thus has a stronger ionic character
than in the AlCl and AlMe bonds. This is in agreement with
the lower Wiberg bond order of 0.58 for the AlF bond
compared to 0.77 and 0.74 for the AlCl and AlC bonds,
respectively.
Figure 2. IBOs for the AlX bonds of complexes K2: top/right: AlF;
bottom/left: AlCl; bottom/right: AlMe.
Scheme 1. Initially expected simplified mechanism.
The mechanism was investigated by density functional theory
simulations to provide atomistic understanding of the reaction.
One of the two diastereomers of the catalyst AlF{PF6} which
were identified by X-ray single crystal analysis (Figure 1,
bottom) was used In combination with benzaldehyde 1d as
substrate. Scheme 2 provides a summary of the most probable
reaction paths identified in the simulations (main pathways A
and B). The intermediates are labeled consecutively with the
respective letter. The catalyst without substrates or reactants is
labeled K and the product is labeled P. Path B is equivalent to
the initially assumed simplified mechanism depicted in Scheme
1.[17] Path A is an alternative possible mechanism which was
found during the computational studies. Relative free energies
at 223 K (50 ºC) of all intermediates are given in Scheme 2.
Additional data (e.g., tables of bond lengths) are provided in
the Supporting Information.
The unique performance of {AlF}OTf is probably also a
consequence of its increased stability and Lewis acidity
compared to other Al-salen complexes. For AlF bonds bond
energies of 659-672 kJ/mol have been reported.[26] They are
thus among the most stable -bonds. As a consequence,
{AlF}OTf exhibits a high robustness. It is remarkably stable
against water, against air over several months at room
temperature and melted under air without decomposition as
1
judged by H/19F NMR at 221 °C. As described above and in
contrast to the related {AlCl} and {AlMe} catalysts it could
be recovered unchanged after the title reaction.
2. Computational Investigations
The AlF bond was studied by means of density functional
theory (DFT)[27] and Intrinsic Bond Orbitals (IBOs)[28] which lead
to useful Lewis structures and are appropriate to interpret
quantum chemical calculations.[29] Structurally simplified Al-
salen complexes K2 were investigated to clarify the binding
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
Scheme 2. Proposed mechanisms for the carboxycyanation of aldehydes (paths B (top) and A (bottom)). Relative free energies G(223 K) in kcal/mol with
respect to structure K are shown in brackets. Important atom numbers for the discussion are depicted in blue.
The energetic profile of the alternative path A is shown in Figure
3 while the calculated 3D structures of intermediates A.2-A.5 are
shown in Figure 4.
Figure 3. Energy profile of reaction path A. Relative free energies G(223 K)
in kcal/mol with respect to structure K are given. The red line indicates that a
barrierless step is assumed since no bond breaking/formation takes place.
In this path EtO(CO)CN and not the aldehyde coordinates to the
catalyst, forming structure A.2, in which the carbonyl oxygen
weakly interacts with the Al-center (d(Al[1]-O[8]) = 2.97 Å).
Cyanide is simultaneously stabilized by the onium moiety (d(N-
Figure 4. Geometries of the intermediates of path A. Hydrogen atoms are
omitted for clarity. Distances are indicated in Å.
N) = 4.17 Å). In the next step, the aldehyde weakly coordinates
with its carbonyl oxygen to the carbonyl carbon of EtO(CO)CN to
form structure A.3. At the same time, a weak coordination of the
cyanide anion, which is held in close proximity to the aldehyde
by the onium moiety takes place (d(N[3]-C[5]) = 2.72 Å). Since no
bonds are broken or formed, it is assumed that this step
possesses a negligible barrier, the increase in the free energy is
of purely entropic origin. The distance d(Al[1]-O[8]) = 2.32 Å in
By overcoming a barrier of 10.2 kcal/mol during the elimination
structure A.3 is rather large. In the barrierless and exothermic
In this concerted trimolecular reaction, the aldehyde is cyanated
(d(C[4]-C[5]) = 1.49 Å in A.4) and the adduct simultaneously
undergoes a protection reaction with EtO(CO)CN which is
facilitated by the catalyst. In A.4, the resulting product is strongly
bound to the catalyst (d(Al[1]-O[8]) = 1.93 Å).
of a cyanide anion from the tetrahedral intermediate (d(C[7]-
(20.9 kcal/mol) transition to the tetrahedral intermediate A.4, the
C[9]) = 2.76 Å), structure A.5 (d(Al[1]-O[8]) = 4.11 Å, d(C[4]-
attack of the cyanide at the aldehyde creates the stereocenter.
C[5]) = 1.48 Å) is formed, in which the product P is only weakly
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
bound to the Al-center. Product release requires a free energy of
9.8 kcal/mol.
In path B (Scheme 2, top and Figures 5 and 6), the aldehyde
initially weakly coordinates to the Al-center, forming structure
B.2 (d(Al[1]-O[8]) = 2.90 Å).
moieties for anionic reaction intermediates in the
copolymerization of CO2 and propylene oxide to form
polycarbonates has already been postulated in the literature.[30]
The authors proposed that a growing, anionic carbonate chain is
cleaved from the catalyst’s Co center after each propagation
step and is held in constant proximity of the Lewis acid by
Coulomb interactions with an onium moiety. In our system, a
favored cleavage of the anionic addition product and
stabilization at the positively charged onium moiety would in
principle be possible. However, no such structure could be
optimized, even when EtO(CO)CN is in close proximity to the Al-
center, as in structure B.4. The association of EtO(CO)CN to the
adduct to form B.4 is favored and enthalpically barrierless. The
rising free energy is of purely entropic origin. The nucleophilic
attack of the addition product at EtO(CO)CN yields structure B.5
which is 4.8 kcal/mol higher in energy via an endergonic step
with a calculated reaction barrier of 12.5 kcal/mol. Thus, the
formation of B.5 from B.4 and B.3 constitutes the rate-
determining step, with an estimated barrier of 18.7 kcal/mol. In
B.5, the tetrahedral intermediate leading to product P is already
detached from the Al-center (d(Al[1]-O[6]) = 4.71 Å, d(Al[1]-O[8]) =
4.41 Å), which explains the increase in energy despite the
formation of a -bond. By elimination of a cyanide anion the final
product P is formed. A possible reaction path connecting B.5 to
A.4 was identified using the NEB approach (nudged elastic
band), exhibiting an estimated barrier of 9.8 kcal/mol relative to
B.5.
Figure 5. Geometries of the intermediates of path B. Hydrogen atoms are
omitted for clarity. Distances are indicated in Å.
In summary, the computational investigations yield the following
information. In pathway A, the estimated rate-limiting barrier
(10.2 kcal/mol) is significantly lower than the one for pathway B
(18.7 kcal/mol). However, the enantioface differentiation in
pathway A is influenced only in the periphery by the sterical and
electrostatic constraints of the enantioselective catalyst. It is
possible that in complex A.2 a chiral pocket is created via the
catalyst’s core, the onium moiety and EtO(CO)CN (see Figure 4).
The substrate may preferably enter this pocket in only one
orientation, which allows for an enantioselective reaction
(structure A.3). Surprisingly, for pathway B no structures could
be optimized in which the catalyst-internal onium moiety favors
the cleavage of the anionic addition product from the Al-center
by Coulomb interactions. It is unclear how stable the
deprotonated cyanohydrin is at low temperatures and if the latter
would rapidly decay to the aldehyde and a cyanide anion after
detachment from the catalyst. The control experiments (Table 3),
however, showed that the presence of an internal onium moiety
is essential for both high activity and selectivity in the presented
carboxycyanation reactions. The absence of an NLE is in
agreement with both paths A and B.
Figure 6. Energy profile of reaction path B. Relative free energies G(223 K)
in kcal/mol with respect to structure K are given. The red line indicates that a
barrierless step is assumed since no bond breaking/formation takes place. A
connection between B.5 and A.4 is also shown.
The activated aldehyde is subsequently attacked by a cyanide
anion which is stabilized at the onium moiety, forming structure
B.3 by overcoming a barrier of 8.4 kcal/mol. In structure B.3, the
tetrahedral intermediate is coordinating to the Al-center via the
3. Development of a Synthetic Protocol with KCN as the
Sole Cyanide Source
As shown above (see subchapter 1.), the presence of solid KCN
in catalytic amounts was essential for a high catalytic efficiency
in the first-generation carboxycyanation protocol. It is likely that
oxygen with
a bond length d(Al[1]-O[6]) = 1.91 Å. For a
subsequent reaction with EtO(CO)CN, the cyanation product
might have to leave the metal center. It is questionable how
stable the anionic addition product is at low temperatures and
how fast it will undergo the reverse reaction to form the aldehyde
via elimination of a cyanide anion. A stabilizing role of onium
the ammonium moiety acts as
a phase transfer catalyst
transporting cyanide anions into solution. Nevertheless, this
anion exchange could not be confirmed spectroscopically. Even
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
using K13CN for the anion exchange, a cyanide anion was not
detected by 13C NMR. Also, by IR this exchange could not be
verified. Probably the equilibrium between the corresponding
ammonium triflate and ammonium cyanide is far on the side of
ammonium triflate.[31] For that reason, the triflate containing
catalyst could be recovered above in pure form after its use.
On the other hand, the fact that excellent results were only
obtained in the presence of KCN pointed to the possibility of
employing KCN as the exclusive cyanide source, thus avoiding
any volatile and thus particularly hazardous organic cyanation
reagents.[11] In addition, the use of KCN would be economically
attractive. Table 4 summarizes the development of such a
protocol.
conditions of entry 1 with 0.1 mol% of {Al-F}OTf the product was
formed in nearly quantitative yield and with promising
enantioselectivity (ee
=
86%, entry 3). Slightly better
enantioselectivity was attained with 1.0 mol% of catalyst (ee =
90%, entry 4). Avoiding an excess of both the carboxylating
reagent and KCN had a negative influence on activity and
enantioselectivity (entry 5). On the other hand, an almost
quantitative yield and high enantioselectivity could also be
achieved with just 1.1 equiv. of KCN at 40 °C (entry 6). The
highest ee value of 92% was obtained at 60 °C with 1.5 equiv.
of KCN and 4.0 equiv. of 2c (entry 7). Interestingly, using these
conditions but NaCN instead, only trace amounts of product
were formed. Noteworthy from a practical point of view is also
the result shown in entry 8 with an {Al-F}OTf catalyst in which
the salen core and the ammonium moiety are linked by only one
methylene group (rather than three in the standard catalyst),
because this catalyst is more rapidly and easily synthesized.
The conditions of Table 4, entry 7 were then selected to
investigate the substrate scope (Table 5).
Table 4. Development of a protocol with KCN as the sole cyanide source.
Table 5. Investigation of the substrate scope of the second-generation
carboxycyanation protocol.
2
T
t
yield ee
No.
1
X
x
y
n
(equiv.)
[°C] [h] [%]a) [%]b)
2b
(2.0)
Cl
2.0
4.0
2.0
2.0
1.0
1.1
1.5
1.5
0.1
0.1
0.1
1.0
1.0
1.0
1.0
1.0
3
3
3
3
3
3
3
1
72 44
72 75
48 96
72 92
48 64
72 99
69
40
86
90
83
91
50
50
50
50
50
40
60
60
2b
(2.0)
2
Cl
1/3
T
[°C]
yield
[%]a)
ee
No.
R
2c
(2.0)
3
O(CO)OEt
O(CO)OEt
O(CO)OEt
O(CO)OEt
O(CO)OEt
O(CO)OEt
[%]b)
1
a
b
d
e
f
2-naphthyl-
6-MeO-2-naphthyl-
Ph-
96
93
93
93
88
90
91
82
90
93
85
90
96
89
80
91
60
60
60
80
60
60
60
60
60
60
60
60
60
60
2c
(2.0)
4
2
2c
(1.0)
5
3c)
4c)
5
92
2c
(2.0)
6
4-Me-C6H4-
3-Me-C6H4-
2-Me-C6H4-
3,4-Me2-C6H3-
4-MeO-C6H4-
3-MeO-C6H4-
2-MeO-C6H4-
3,4-(MeO)2-C6H3-
4-Cl-C6H4-
85
98
2c
(4.0)
72 100 92
7
6
g
i
>99
98
2c
(4.0)
8
72 89
87
7
8
j
99
a)
Determined by 1H-NMR of the crude product using an internal standard.
b) Determined by HPLC.
9
k
l
>99
>99
90
10
11
12
13
14
Initial studies were performed with chloroformate 2b. In the
presence of 0.1 mol% of the standard {Al-F}OTf catalyst and an
excess of KCN and 2b (each 2.0 equiv.) the product was slowly
formed at 50 °C with moderate enantioselectivity (entry 1). A
larger KCN excess led to a higher yield at the cost of an erosion
of enantioselectivity (entry 2). Much more interesting results
were obtained with diethyl pyrocarbonate 2c, which delivers the
same stable products as with the previous protocol. Under the
m
n
o
p
>99
>99
>99
3-Cl-C6H4-
4-F-C6H4-
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
recovery and reuse. This appears to be the first application of a
structurally defined Al catalyst containing the extraordinarily
strong AlF -bond in asymmetric catalysis.
15
r
s
s’
t
3-MeO2C-C6H4-
2-furyl-
92
93
82
90
80
92
93
95
94
55
69
60
80
60
60
60
60
60
60
80
16 c)
17
DFT studies also suggest that the initially assumed mechanism,
in which the aldehyde substrate is activated by the Al center and
subsequently attacked by an ammonium bound cyanide ion
might be more difficult to realize than an alternative trimolecular
pathway, in which the ammonium-bound cyanide attacks the
aldehyde which itself is activated by the carbonyl group of the
cyanoformate, whereas the latter is activated by the Al center.
The originally assumed pathway seems to suffer from an
energetically difficult carboxylation step.
2-thiophenyl-
91
18
(E)-Ph-CH=CH-
(E)-4-MeO-C6H4-CH=CH-
(E)-Me-CH=CH-
(E)-Me-CH=CH-
Me-
>99
99
19
u
v
v
w
z
20 d)
21 e)
22
90
81
>99
95
Finally,
a
very practical carboxycyanation protocol was
23 c)
PhCH2CH2-
developed, which avoids the necessity of volatile organic and
thus more dangerous cyanation reagents, making use of KCN
as the exclusive cyanide source. This reagent, an inexpensive
bulk chemical in salt form with low volatility, is very attractive for
a)
b)
Yield of isolated product after column chromatography.
HPLC. Reaction in CH2Cl2/CHCl3 (1:1). 0.5 mol% of catalyst was used.
0.1 mol% of catalyst was used.
Determined by
e)
c)
d)
large scale applications. The use of
a pyrocarbonate as
Overall, it can be stated that the results and tendencies are
similar to the first-generation approach presented above, albeit
the activity is somewhat lower than using ethyl cyanoformate. It
carboxylation reagent provided the best results which are
comparable to the first-generation approach in terms of yield and
enantioselectivity.

is possible that EtOCO2 and/or EtO (formed by decarboxylation
temporarily inhibit the Al center. In addition KCN is an almost
insoluble solid, whereas the cyanoformate is completely dissolved.
Naphthyl derivatives (entries 1-2), benzaldehyde (entry 3),
aromatic aldehydes carrying alkyl groups as -donors (mono
and disubstituted, entries 4-7, respectively), -donors (entries 8,
10, 11), -acceptors (entries 9, 11-15) and -acceptors (entry
15) were all well accepted. As before, ortho-substituted
substrates yielded slightly lower enantioselectivities (entries 6,
10). Also, heteroaromatic substrates were tolerated (entries 16-
17) and even higher enantioselectivities could be attained than
with the first-generation protocol. With enals like trans-
cinnamaldehyde or -crotonaldehyde a similar efficiency was
found for both protocols (entries 18-21). For crotonaldehyde we
demonstrated that a high catalytic efficiency is maintained with
reduced catalyst loadings. For instance, with a reduced catalyst
loading of 0.1 mol% a yield of 81% and 94% ee was obtained
(entry 21).
Experimental Section
Computational Details. Geometries were optimized with density
functional theory (DFT) using the functional M06-2X[32] and the basis set
def2-SVP.[33] Frequencies were calculated at the same level. At fixed
geometries, the energy was calculated using DLPNO-CCSD(T)/cc-
pVTZ[34] and the solvation was accounted for using the COSMO solvation
model[35] with  = 8.93 (DCM as solvent) in combination with the
augmented def2-SVPD basis set.[36] All geometry optimizations and the
free-energy calculations using the rigid rotor, harmonic oscillator
approximation at 223 K were performed in DL-FIND[37] in Chemshell,[38]
DFT calculations were performed in Turbomole,[39] CCSD(T) calculations
in Orca.[40] More details to the computational protocol are given in the
Supporting Information.
General Procedure for the Formation and Isolation of AlF-Salen-
Catalysts (GP A)
In contrast to aromatic aldehydes and enals, aliphatic substrates
were less well accommodated in terms of enantioselectivity,
whereas the product yields were still high to excellent (entries
22-24).
Under a nitrogen atmosphere a Schlenk tube was charged with the
respective ligand L (1.0 equiv.), which was dissolved in CHCl3. A solution
of dimethylaluminum fluoride[20] in toluene (2.53 M, 1.0 equiv.) was added
dropwise and the resulting solution was stirred for 18 h at room
temperature. The solvent was removed under reduced pressure and the
residue was washed several times with n-pentane or an n-
pentane/CH2Cl2 mixture.
Conclusions
General Procedure for the In-Situ-Formation of Al-Salen-Catalysts
(GP B)
We have developed an exceptionally active and broadly
applicable bifunctional dual activation catalyst for the
asymmetric carboxycyanation of aldehydes. This catalyst
allowed for unprecedented turnover numbers of up to 104.
Control experiments suggest that a Lewis acidic AlF center
cooperates with an internal ammonium salt moiety. The
necessity of the fluoride ligand for exceptional activity is
explained by an increased Lewis acidity as found by DFT
Under a nitrogen atmosphere a Schlenk tube was charged with the
respective ligand L (1.0 equiv.), which was dissolved in the respective
solvent solvent. A solution of the respective aluminum source (1.0 equiv.,
in the same solvent) was added dropwise and the resulting solution was
stirred for 3 h at room temperature. The in situ formed catalysts were
directly used without further purification.
studies and
a
noteworthy catalyst stability, which is
unprecedented for Al-salen complexes and allowed for catalyst
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
General Procedure for the Asymmetric Carboxycyanation of Alde-
hydes Using EtO(CO)CN (GP C)
[1]
[2]
Early study: M. Hayashi, Y. Miyamoto, T. Inoue, N. Oguni, J. Org.
Chem. 1993, 58, 1515.
Reviews: a) N. Kurono, T. Ohkuma, ACS Catal. 2016, 6, 989; b) T.
Ishikawa in Comprehensive Chirality, Eds.: K. Maruoka, M. Shibasaki,
Elsevier, Amsterdam, 2012; Vol. 6, pp 194−213; c) W. Wang, X. Liu, L.
Lin, X. Feng, Eur. J. Org. Chem. 2010, 4751; d) M. North, D. L. Usanov,
C. Young, Chem. Rev. 2008, 108, 5146; e) T. R. J. Achard, L. A.
Clutterbuck, M. North, Synlett 2005, 1828; f) J.-M. Brunel, I. P. Holmes,
Angew. Chem. Int. Ed. 2004, 43, 2752.
Catalyst {AlF}OTf (2.05 mM in 1,2-dichloroethane, 0.1 mol%, 0.10 mol,
0.08 mg, 49 L) and KCN (0.1 equiv., 0.01 mmol, 0.7 mg) were placed in
a Schlenk tube under a nitrogen atmosphere. CHCl3 (0.2 mL) was added
and the mixture was stirred for 10 min at room temperature. Then the
solution was cooled to 50 °C. The respective aldehyde 1 (1.0 equiv.,
0.10 mmol) was added as a solid or via syringe and flushed down with
CHCl3 (0.2 mL). EtO(CO)CN 2a (1.0 equiv., 0.10 mmol, 10.0 mg, 10 L)
was added via syringe and the mixture was stirred for 24 h at 50 °C.
The reaction mixture was filtered over a short pad of silica gel (petroleum
ether/ethyl acetate 9:1) and the solvent was removed under reduced
pressure to yield the respective cyanation product 3.
[3]
[4]
Early studies using enzymes: S. Förster, J. Roos, F. Effenberger, H.
Wajant, A. Sprauer Angew. Chem. Int. Ed. Engl. 1994, 33, 1555.
P. Poechlauer, W. Skrang, M. Wubbolts, in Asymmetric Catalysis on
Industrial Scale, H. U. Blaser. E. Schmidt (Eds.), Wiley-VCH, 2004,
151-164.
[5]
Non-enzymatic catalytic asymmetric hydrocyanations: a) K. Tanaka, A.
Mori, S. Inoue, J. Org. Chem. 1990, 55, 181; b) N. Kurono, T.
Yoshikawa, M. Yamasaki, T. Ohkuma, Org. Lett. 2011, 13, 1254.
Early asymmetric carboxycyanation with ethyl cyanoformate: S.-K.
Tian, L. Deng, J. Am. Chem. Soc. 2001, 123, 6195.
General Procedure for the Asymmetric Carboxycyanation of Alde-
hydes Using KCN and (EtO(CO))2O (GP D)
[6]
[7]
Catalyst {AlF}OTf (1.0 mol%, 4.00 mol, 3.3 mg) and KCN (1.5 equiv.,
0.60 mmol, 39.1 mg) were placed in a Schlenk tube under a nitrogen
atmosphere. CHCl3 (0.2 mL) was added and the mixture was stirred for
10 min at room temperature. Then the solution was cooled to 60 °C.
The respective aldehyde 1 (1.0 equiv., 0.40 mmol) was added as a solid
or via syringe and flushed down with CHCl3 (0.2 mL). (EtO(CO))2O 2c
(4.0 equiv., 1.60 mmol, 267.5 mg, 243 L) was added via syringe and the
mixture was stirred for 72 h at 60 °C. The reaction mixture was filtered
over a short pad of silica gel (DCM) and the solvent was removed under
reduced pressure to yield the respective cyanation product 3.
An interesting alternative is the cyanophosphorylation: A. Baeza, J.
Casas, C. Nájera, J. M. Sansano, J. M. Saá, Angew. Chem. Int. Ed.
2003, 42, 3143.
[8]
[9]
a) Y. N. Belokon, A. J. Blacker, P. Carta, L. A. Clutterbuck, M. North,
Tetrahedron 2004, 60, 10433; b) T. Kull, R. Peters, Adv. Synth. Catal.
2007, 349, 1647.
For the formation of β-amino alcohols by reductive removal of the
ethylcarboxy protecting group, see e.g.: a) A. Sadhukhan, M. K.
Choudhary, N.-u. H. Khan, R. I. Kureshy, S. H. R. Abdi, H. C. Bajaj,
ChemCatChem 2013, 5, 1441; for reductive or hydrolytic removal of
carboxy protecting groups, see: b) A. Baeza, J. Casas, C. Nájera, J. M.
Sansano, J. M. Saá, Eur. J. Org. Chem. 2006, 1949.
Preparation of (R,E)-1-Cyano-3-phenylallyl Ethyl Carbonate 3t with
KCN and EtO(CO)CN on a 1 g Scale (GP C)
[10] a) [3,3]-rearrangements: J. Tian, N. Yamagiwa, S. Matsunaga, M.
Shibasaki, Org. Lett. 2003, 5, 3021; b) allylic substitutions: A. Baeza, J.
Casas, C. Nájera, J. M. Sansano, J. Org. Chem. 2006, 71, 3837.
Compound 3t was prepared according to GP5 using trans-
cinnamaldehyde 1t (1.0 equiv., 7.34 mmol, 1.00 g, 0.95 mL), CHCl3
(29.4 mL), catalyst {AlF}OTf (0.1 mol%, 7.34 mol, 5.96 mg), KCN
(0.1 equiv., 0.73 mmol, 47.80 mg) and EtO(CO)CN 2a (1.0 equiv.,
7.34 mmol, 734.67 mg, 733 L) at 50 °C during a reaction time of 72 h
to yield 3t (7.32 mmol, 1.69 g, 100%, 96% ee) as a colorless oil.
[11] High TONs could be achieved in carbocyanations with NaCN and
acetic anhydride: a) Z. Zhang, Z. Wang, R. Zhang, K. Ding, Angew.
Chem. Int. Ed. 2010, 49, 6746; selected publications describing the use
of Ac2O, cyanide salts and salen catalysts: b) Y. N. Belokon, A. V.
Gutnov, M. A. Moskalenko, L. V. Yashkina, D. E. Lesovoy, N. S.
Ikonnikov, V. S. Larichev, M. North, Chem. Commun. 2002, 244; c) Y.
N. Belokon, P. Carta, A. V. Gutnov, V. Maleev, M. A. Moskalenko, L. V.
Yashkina, N. S. Ikonnikov, N. V. Voskoboev, V. N. Khrustalev, M.
North, Helv. Chim. Acta 2002, 85, 3301; d) W. Huang, Y. Song, J.
Wang, G. Cao, Z. Zheng, Tetrahedron 2004, 60, 10469; e) W. Huang,
Y. Song, C. Bai, G. Cao, Z. Zheng, Tetrahedron Lett. 2004, 45, 4763; f)
N.-u. H. Khan, A. Sadhukhan, N. C. Maity, R. I. Kureshy, S. H.R. A., S.
Saravanan, H. C. Bajaj, Tetrahedron 2011, 67, 7073.
Catalyst Recycling
The reaction mixture was transferred to a flask filled with 100 mL n-
pentane at 50 °C. The resulting mixture was filtered over cotton and
further washed with n-pentane. Flushing down with DCM and
evaporation of the solvent yielded {AlF}OTf (72%) in unchanged form
(characterized again by 1H and 19F NMR, HR-MS).
[12] Highly active catalysts (loadings
≤ 0.2 mol%) for the highly
enantioselective cyanosilylations: a) Y. N. Belokon, S. Caveda-Cepas,
B. Green, N. S. Ikonnikov, V. N. Khrustalev, V. S. Larichev, M. A.
Moscalenko, M. North, C. Orizu, V. I. Tararov, M. Tasinazzo, G. I.
Timofeeva, L. V. Yashkina, J. Am. Chem. Soc. 1999, 121, 3968; b) Y.
N. Belokon, M. North, T. Parsons, Org. Lett. 2000, 2, 1617; c) F.-X.
Chen, H. Zhou, X. Liu, B. Qin, X. Feng, G. Zhang, Y. Jiang, Chem. Eur.
J. 2004, 10, 4790; d) N. Kurono, K. Arai, M. Uemura, T. Ohkuma,
Angew. Chem. Int. Ed. 2008, 47, 6643; e) M. North, M. Omedes-Pujol,
Tetrahedron Lett. 2009, 50, 4452; f) ref. 11a; g) N. Kurono, M. Uemura,
T. Ohkuma, Eur. J. Org. Chem. 2010, 1455; h) T. Ohkuma, N. Kurono,
Synlett 2012, 1865; i) M. Uemura, N. Kurono, Y. Sakai, T. Ohkuma,
Adv. Synth. Catal. 2012, 354, 2023.
Acknowledgements
This work was funded by the Deutsche Forschungsgemeinschaft
(DFG, German Research Foundation)
– project number
404194277 (Peters) and the European research council (ERC)
project number 646717 (Kästner). The Carl-Zeiss-Stiftung is
gratefully acknowledged for a Ph.D. fellowship to D.B. We thank
Dr. José Cabrera Crespo for initial studies on aldehyde
cyanations.
[13] Selected further examples with salen catalysts: a) C. Lv, C.-X. Miao, D.
Xu, S. Wang, C. Xia, W. Sun, Cat. Commun. 2012, 27, 138; b) S. Liang,
X. R. Bu, J. Org. Chem 2002, 67, 2702; c) M. E. Silva Serra, D.
Murtinho, A. Goth, A. M. D'A. Rocha Gonsalves, P. E. Abreu, A. A. C. C.
Pais, Chirality 2010, 22, 425.
Keywords: AlF -bond • bifunctional catalysis • nitriles • salen •
trimolecular reaction
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
[14] Selected studies: a) J. Tian, N. Yamagiwa, S. Matsunaga, M.
Shibasaki, Angew. Chem. Int. Ed. 2002, 41, 3636; b) Y. N. Belokon, A.
J. Blacker, L. A. Clutterbuck, M. North, Org. Lett. 2003, 5, 4505; c) J.
Casas, A. Baeza, J. M. Sansano, C. Nájera, J. M. Saá, Tetrahedron:
Asymmetry 2003, 14, 197; d) N. Yamagiwa, J. Tian, S. Matsunaga, M.
Shibasaki, J. Am. Chem. Soc. 2005, 127, 3413; e) S. Lundgren, E.
Wingstrand, M. Penhoat, C. Moberg, J. Am. Chem. Soc. 2005, 127,
11592; f) Y. N. Belokon, E. Ishibashi, H. Nomura, M. North, Chem.
Commun. 2006, 1775; g) S. Gou, X. Chen, Y. Xiong, X. Feng, J. Org.
Chem. 2006, 71, 5732; h) S. Gou, J. Wang, X. Liu, W. Wang, F.-X.
Chen, X. Feng, Adv. Synth. Catal. 2007, 349, 343; i) Y. N. Belokon, W.
Clegg, R. W. Harrington, E. Ishibashi, H. Nomura, M. North,
Tetrahedron 2007, 63, 9724; j) with α-ketonitriles: S. Lundgren, E.
Wingstrand, C. Moberg, Adv. Synth. Catal. 2007, 349, 364; k) ref. 10; l)
J. Wang, W. Wang, W. Li, X. Hu, K. Shen, C. Tan, X. Liu, X. Feng,
Chem. Eur. J. 2009, 15, 11642; m) N. H. Khan, A. Sadhukhan, N. C.
Maity, R. I. Kureshy, S. H.R. Abdi, S. Saravanan, H. C. Bajaj,
Tetrahedron 2011, 67, 7073; n) E. Wingstrand, C. Moberg, Synlett 2010,
355.
(priority rules). In addition, for 3a the (R)-configuration was confirmed
by X-ray crystal structure analysis. CCDC-1859428 contains the
supplementary crystallographic data for compound 3a. These data can
be obtained free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[23] H. C. Aspinall, N. Greeves, P. M. Smith, Tetrahedron Lett. 1999, 40,
1763.
[24] Early applications of homochiral Al-salen complexes: a) imine
substrates: M. S. Sigman, E. N. Jacobsen, J. Am. Chem. Soc. 1998,
120, 5315; b) conjugate addition to unsaturated imides: G. M. Sammis,
E. N. Jacobsen, J. Am. Chem. Soc. 2003, 125, 4442; for a similar
application, see: c) J. K. Myers, E. N. Jacobsen, J. Am. Chem. Soc.
1999, 121, 8959.
[25] Reviews on asymmetric onium salt catalysis: a) S. Shirakawa, K.
Maruoka, Angew. Chem. Int. Ed. 2013, 52, 4312; b) K. Maruoka, Org.
Proc. Res. Dev. 2008, 12, 679.
[26] a) R. T. Sanderson, Polar Covalence, Academic Press, New York,
1983, p. 155; b) L. Pauling, The Nature of The Chemical Bond, Cornell
University Press, Ithaca, New York, 1960, p. 85.
[15] Interesting results were also obtained with chiral bisammonium salts: R.
Chinchilla, C. Nájera, F. J. Ortega, S. Tarí, Tetrahedron: Asymmetry
2009, 20, 2279.
[27] A detailed description of the computational details is given in the SI.
[28] G. Knizia, J. Chem. Theory Comput. 2013, 9, 4834.
[29] a) G. Knizia, J. E. M. N. Klein, Angew. Chem. Int. Ed. 2015, 54, 5518;
b) for an application by our group, see: M. Ayerbe Garcia, W. Frey, M.
Ringenberg, M. Schwilk, R. Peters, Chem. Commun. 2015, 51, 16806.
[30] E. K. Noh, S. J. Na, S. S, S.-W. Kim, B. Y. Lee, J. Am. Chem. Soc.
2007, 129, 8082.
[16] R. Peters (Ed.), Cooperative Catalysis – Designing Efficient Catalysts
for Synthesis, Wiley-VCH, Weinheim, Germany, 2015.
[17] D. Brodbeck, F. Broghammer, J. Meisner, J. Klepp, D. Garnier, W. Frey,
J. Kästner, R. Peters, Angew. Chem. Int. Ed. 2017, 56, 4056.
[18] Our research group has introduced the strategy of cooperative
asymmetric intramolecular Lewis acid/aprotic onium salt catalysis: a) T.
Kull, R. Peters, Angew. Chem. Int. Ed. 2008, 47, 5461; b) T. Kull, J.
Cabrera, R. Peters, Chem. Eur. J. 2010, 16, 9132; c) P. Meier, F.
Broghammer, K. Latendorf, G. Rauhut, R. Peters, Molecules 2012, 17,
7121; d) F. Broghammer, D. Brodbeck, T. Junge, R. Peters, Chem.
Commun. 2017, 53, 1156-; related catalyst systems: e) M. Mechler, R.
Peters, Angew. Chem. Int. Ed. 2015, 54, 10303; this concept is also
very useful in macromolecular chemistry see, e.g.: f) M. Winnacker, S.
[31] We have intensively tried to install cyanide as counterion of the
catalyst, but surprisingly failed in all synthetic attempts so far.
[32] Y. Zhao, D. Truhlar, Theor. Chem. Acc. 2008, 120, 215.
[33] a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297;
b) F. Weigend, Phys. Chem. Chem. Phys. 2006, 8, 1057.
[34] a) C. Riplinger, B. Sandhoefer, A. Hansen, F. Neese, J. Chem. Phys.
2013, 139, 134101; b) T. H. Kendall, R. A., Dunning, R. J. Harrison, J.
Chem. Phys. 1992, 96, 6796.
[35] A. Klamt, G. Schüürmann, J. Chem. Soc., Perkin Trans. 2 1993, 799.
[36] D. Rappoport, F. Furche, J. Chem. Phys. 2010, 133, 134105 (1–11).
[37] J. Kästner, J. M. Carr, T. W. Keal, W. Thiel, A. Wander, P. Sherwood, J.
Phys. Chem. A 2009, 113, 11856.
Vagin, B. Rieger, in: Cooperative Catalysis
– Designing Efficient
Catalysts for Synthesis, R. Peters (Ed.), Wiley-VCH, Weinheim, 2015;
g) for other applications, see: K. Ohmatsu, S. Kawai, N. Imagawa, T.
Ooi, ACS Catal. 2014, 4, 4304.
[38] a) S. Metz, J. Kästner, A. A. Sokol, T. W. Keal, P. Sherwood, Wiley
Interdiscip. Rev. Comput. Mol. Sci. 2014, 4, 101; b) “ChemShell, a
[19] CCDC-1522966 and CCDC-1522967 contain the supplementary
crystallographic data for compounds {Al2F3}BF4 and {AlF}PF6,
respectively. These data can be obtained free of charge from The
Computational
Chemistry
Shell.,”
can
be
found
under
http://www.chemshell.org, n.d.
Cambridge
Crystallographic
Data
Centre
via
[39] “TURBOMOLE V7 2015, a development of the University of Karlsruhe
and Forschungszentrum Karlsruhe GmbH, 1989-2007, TURBOMOLE
GmbH, since 2007; available from http://www.turbomole.com.,” can be
found under http://www.turbomole.com, n.d.
www.ccdc.cam.ac.uk/data_request/cif.
[20] Me2AlF: K. Ziegler, R. Köster, Justus Liebigs Ann. Chem. 1957, 608, 1.
[21] T. Augenstein, F. Dorner, K. Reiter, H. E. Wagner, D. Garnier, W.
Klopper, F. Breher, Chem. Eur. J. 2016, 22, 7935.
[40] a) F. Neese, Wiley Interdiscip. Rev. Comput. Mol. Sci. 2012, 2, 73; b) F.
Neese, Wiley Interdiscip. Rev. Comput. Mol. Sci. 2017, 8, e1327 (1–6).
[22] The (R)-configuration was determined by comparison to published
optical rotation data (see SI) for compounds 3a, 3c-g, 3j-l, 3n, 3p, 3t,
3v, 3x, 3z and 3za. 3s and 3s’ are (S)-configured as determined by
comparison to published optical rotation data for formal reasons
This article is protected by copyright. All rights reserved.

10.1002/chem.201804388
Chemistry - A European Journal
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Trimolecular? The evolution of a
robust, highly active catalyst for
asymmetric aldehyde cyanations is
reported, which in the final synthetic
protocol uses inorganic KCN instead
of volatile organic cyanation
Daniel Brodbeck, Sonia Álvarez-
Barcia, Jan Meisner, Florian
Broghammer, Julian Klepp, Delphine
Garnier, Wolfgang Frey, Johannes
Kästner,* and René Peters*
Page No. – Page No.
reagents. DFT studies for the initial
protocol using a cyanoformate
reagent suggest a unique
Asymmetric Carboxycyanation of
Aldehydes by Cooperative AlF-
/Onium Salt Catalysts: from
Cyanoformate to KCN as Cyanide
Source
trimolecular reaction pathway.
This article is protected by copyright. All rights reserved.