Cyanogen formation during asymmetric cyanohydrin synthesis

Article abstract of DOI:10.1039/c001703e

During asymmetric cyanohydrin synthesis catalysed by vanadium ^Voxo(salen) complexes, the catalysts are reduced to vanadium ^IVoxo(salen) as determined by EPR spectroscopy; the reducing agent is cyanide which is oxidised to cyanogen via a non-radical mechanism.

Full text of DOI:10.1039/c001703e

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Cyanogen formation during asymmetric cyanohydrin synthesisw
Victor Chechik,*a Marco Conte,a Trevor Dransfield,a Michael North*b and
Marta Omedes-Pujolb
Received (in Cambridge, UK) 27th January 2010, Accepted 2nd March 2010
First published as an Advance Article on the web 20th March 2010
DOI: 10.1039/c001703e
During asymmetric cyanohydrin synthesis catalysed by
vanadiumVoxo(salen) complexes, the catalysts are reduced to
vanadiumIVoxo(salen) as determined by EPR spectroscopy; the
reducing agent is cyanide which is oxidised to cyanogen via a
non-radical mechanism.
1a–c were proposed to contain vanadium ions in both the +4
and +5 oxidation states. This was supported by mass spectro-
metry and by the characterisation of the corresponding
(catalytically active) heterobimetallic complexes containing
vanadiumV(salen) and titaniumIV(salen) units.10 The aim of
this study was to obtain mechanistic information about the
VIV–VV redox process in asymmetric cyanohydrin synthesis.
Complex 2 is paramagnetic and gives an EPR spectrum
consisting of eight lines. EPR spectra recorded on complexes
1a–c in dichloromethane at room temperature under an air
atmosphere showed that only a small amount (o6%) of
vanadiumIV species was present (Fig. S1).w When benzaldehyde
and TMSCNz were added sequentially to dichloromethane
solutions of complexes 1a–c, asymmetric cyanohydrin synthesis
occurred and EPR spectra (Fig. S2–S4)w recorded every five
minutes showed that the amount of vanadiumIV species
present increased rapidly and reached a steady state (12, 10
and 51% of the total vanadium content for complexes 1a–c,
respectively) before decreasing.
Having determined that vanadiumIV was indeed generated
in situ during asymmetric cyanohydrin synthesis catalysed by
complexes 1a–c, the question remained as to how it was being
formed and in particular as to the nature of the reducing
agent. Since vanadiumV species are well known oxidising
agents11 it was expected that benzaldehyde would function
as the reducing agent and this would also explain why ketones
were not substrates for asymmetric cyanohydrin synthesis
catalysed by complexes 1a–c. However, when benzaldehyde
was added to dichloromethane solutions of complexes 1a–c,
there was no increase in the amount of vanadiumIV species
present (Fig. S5–S7).w
The asymmetric addition of cyanide to aldehydes and ketones
has developed over the last 15 years into a highly studied
asymmetric carbon–carbon bond forming reaction.1 We have
reported vanadiumVoxo(salen) complexes 1 as catalysts for
asymmetric cyanohydrin synthesis2 using either trimethylsilyl
cyanide (TMSCN)3 or potassium cyanide4 as cyanide source
(Scheme 1). Other researchers have used complexes 1 within
synthetic projects,5 immobilised the complexes,1,6 utilised
alternative cyanide sources,7 or used salen ligands in other
applications such as fluorescent sensors.8
Recently,9 we showed that the activity (but not the enantio-
selectivity) of catalysts 1 depended on the nature of the
counterion, with the order of reactivity being 1a o 1b o 1c.
Based on the reaction kinetics, precursor complexes 1a–c
were proposed to form both mononuclear and dinuclear
catalytically-active complexes in situ. Fluorescence studies
along with cyclic voltametry indicated that vanadiumV
complexes 1a–c are reduced to vanadiumIV(salen) containing
species in situ. Since VO(salen) 2 is catalytically inactive4 and
an oxygen containing atmosphere is essential to avoid catalyst
deactivation, the dinuclear complexes formed from complexes
Surprisingly, when TMSCN was added to dichloromethane
solutions of complexes 1a–c, a significant increase in the
amount of vanadiumIV species was observed (Fig. 1 and Fig.
S9–S11).w This suggests that vanadiumIV is generated by the
reduction of vanadiumV complexes with TMSCN rather than
benzaldehyde. The evolution of the EPR signal with time
(Fig. 1) clearly shows that both the rate of vanadiumIV
formation and its total amount increase in the order
1a o 1b o 1c, matching the catalytic activity of the three
complexes.9 At long reaction times, a very small amount of
another vanadiumIV species was observed. The concentration
of this species increased significantly in the presence of ben-
zaldehyde (Fig. 2). In order to determine the magnetic para-
meters and hence identify the vanadiumIV species, EPR
spectra in the absence and presence of benzaldehyde (in order
to increase the signal intensity of the minor VIV species) were
recorded at 130 K in frozen dichloromethane solutions
(Fig. 3 and Fig. S12).w The principal values of A and g tensors
Scheme 1 Asymmetric cyanohydrin synthesis using catalysts 1a–c.
a Department of Chemistry, University of York, Heslington, York,
UK YO10 5DD. E-mail: vc4@york.ac.uk
b School of Chemistry and University Research Centre in Catalysis
and Intensified Processing, University of Newcastle upon Tyne,
Bedson Building, Newcastle upon Tyne, UK NE1 7RU.
E-mail: Michael.north@ncl.ac.uk; Fax: 44 (0)191 222 6929;
Tel: 44 (0)191 222 7128
w Electronic supplementary information (ESI) available: Details of
chemicals and equipment, copies of spectra and details of spin
trapping experiments. See DOI: 10.1039/c001703e
ꢀc
This journal is The Royal Society of Chemistry 2010
3372 | Chem. Commun., 2010, 46, 3372–3374

View Article Online
the observed EPR signals might be due to dinuclear
complexes,13 EPR spectra were recorded at 192 K (Fig. S8).w
At this temperature, EPR spectra are very sensitive to
molecular tumbling and dinuclear complexes should show a
significantly reduced tumbling rate. The EPR spectra recorded
at 192 K were simulated using a slow motion isotropic
rotation model with magnetic parameters from Table 1. The
rotational correlation times of complex 2 and the major and
minor components formed from complex 1c in the presence of
TMSCN were 8.4 ꢁ 10ꢂ10, 1.0 ꢁ 10ꢂ9, and 6.9 ꢁ 10ꢂ10 s,
respectively. Thus, it was concluded that the major vanadiumIV
species formed during reduction of complexes 1a–c is complex
2. The minor component showed a slightly faster tumbling rate
than that of complex 2 and is therefore likely to be a
degradation product of complex 2. This is consistent with
the deactivation of catalysts 1a–c at long reaction times and
the presence of low molecular weight fragments in the mass
spectra of deactivated catalyst.
Fig. 1 Formation of vanadiumIV species from 1a–c and TMSCN.
The reduction of vanadiumV with TMSCN to give complex
2 is not related to the trimethylsilyl group as complex 2 was
also detected by EPR when complex 1c was mixed with
potassium cyanide (Fig. S13).w These results could be
explained if complexes 1a–c react with cyanide (either free
cyanide or TMSCN) in situ to form vanadiumV cyanide
complex 1d and in the case of reactions with TMSCN do so
at different rates and to different equilibrium positions, thus
accounting for the reactivity trend in Fig. 1. Attempts to
prepare complex 1d have been unsuccessful due to its instability.
However, when complexes 1b,c were mixed with TMSCN, it
was possible to detect trimethylsilyl chloride and trimethylsilyl
isothiocyanate, respectively, by 13C NMR and GC-MS
(Fig. S14 and S15).w
Fig. 2 (a) X-band EPR spectra of 2; (b) mixture of 1c and TMSCN
after 20 minutes; and (c) mixture of 1c, TMSCN and PhCHO two days
after mixing.
Homolysis of the V–CN bond of complex 1d could
potentially generate vanadiumIVoxo(salen) 2 and a cyanide
radical. Therefore, experiments were undertaken to try to
detect the cyanide radicals. EPR spin trapping experiments
using complexes 1a–c in the presence of DMPO spin trap
showed small amount of cyanide adducts which were assigned
by comparison with genuine CN adducts (Fig. S16).w
However, EPR spectra also gave clear evidence that vanadiumV
complexes oxidised spin traps.w It is therefore likely that the
cyanide spin adducts were formed by nucleophilic addition of
cyanide anion to the oxidised spin trap (e.g., DMPO radical
cation); this is commonly known as inverse spin trapping.14
Hence, spin trapping experiments did not provide conclusive
evidence for or against cyanide radical formation. The reaction
between complex 1c and TMSCN was also carried out in the
presence of the stable free radical TEMPO and monitored by
EPR. If cyanide radicals were formed, they would be expected
to recombine with TEMPO, thus leading to a reduction of the
TEMPO EPR signal. However, only a small reduction in the
TEMPO EPR intensity was observed and no TEMPO–CN
adduct was detected by GC-MS analysis of the reaction
mixture (Fig. S17–S19).w To further test if cyanide radicals
are formed in the reaction, the reduction of vanadiumV
complexes with TMSCN was carried out in a sealed system
in the presence of styrene. Any cyanide radicals were expected
to initiate polymerisation of styrene, and as the reaction was
carried out in the presence of air, a rapid reduction in oxygen
Fig. 3 (a) EPR spectra at 130 K for mixture of 1c and TMSCN, (b)
simulation of main and (c) minor components.
Table 1 Magnetic parameters of complex 2 and vanadiumIV species
formed in the reaction of complex 1c with TMSCN
Complex gx
gy
gz
|Ax|/MHz |Ay|/MHz |Az|/MHz
2
1.985 1.981 1.955 162
1.982 1.982 1.954 155
1.984 1.982 1.959 155
1.978 1.978 1.906 113
166
163
163
113
489
484
477
432
1ca
1cb
1cc
a
b
Mixture of 1c and TMSCN 20 minutes after mixing. Main
vanadiumIV species formed in the reaction of 1c with TMSCN and
benzaldehyde two days after mixing. Minor vanadiumIV species
c
formed in the reaction of 1c with TMSCN and benzaldehyde two
days after mixing.y
obtained by the simulation of these powder pattern spectra are
shown in Table 1.
Magnetic parameters of the major species formed from
vanadiumV complexes were very similar to those of authentic
complex
complexes.12 To unambiguously determine whether or not
2
and were typical of vanadiumIVoxo(salen)
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 3372–3374 | 3373

View Article Online
complexes, this provides a convenient method for the generation
of cyanogen under non-aqueous conditions.16
The authors thank EPSRC for financial support.
Notes and references
z Caution: cyanide, cyanohydrins and cyanogen are highly toxic. See
ESIw for more detail.
y As the EPR spectra of this compound could only be recorded in
the presence of the main vanadiumIV product, accurate spectrum
simulation is impossible. The values quoted in Table 1 have been
obtained assuming axial symmetry and gave an adequate fit, see ESIw.
Scheme 2 Mechanism to account for cyanogen formation.
1 M. North, D. L. Usanov and C. Young, Chem. Rev., 2008, 108,
5146.
2 Y. N. Belokon’, M. North and T. Parsons, Org. Lett., 2000, 2,
1617.
concentration was expected. The reaction was monitored by
EPR. No evidence for oxygen consumption (which would have
resulted in the sharpening of the EPR lines of complex 2)15 or
increased viscosity of the solution (polymer formation would
have slowed down the tumbling rates of complex 2) was
detected (Fig. S20 and S21).w Finally, attempts were made to
trap cyanide radicals with electron-rich vinyl ethers, but no
evidence for cyanide adducts was detected.w Taken together,
these results strongly suggest that cyanide radicals are not
involved as discrete intermediates in the reaction mechanism.
There were a number of possible products into which
cyanide could conceivably be oxidised including cyanate,
cyanogen and carbon dioxide. Therefore, reactions were
monitored by GC-MS to identify any volatile cyanide derived
products. In order to ensure that metal-catalysed reactions do
not occur in the injection chamber of the GC apparatus, the
volatile products formed in the reaction of complex 1c with
TMSCN were carried with a stream of inert gas (N2) into a
cooled solvent trap. GC-MS analysis of the trapped products
showed the presence of cyanogen [(CN)2] (Fig. S22–S25).w
Thus, it was concluded that vanadiumVoxo(salen) complexes
oxidise cyanide to cyanogen.
3 Y. N. Belokon’, B. Green, N. S. Ikonnikov, M. North, T. Parsons
and V. I. Tararov, Tetrahedron, 2001, 57, 771; Y. N. Belokon’,
V. I. Maleev, M. North and D. L. Usanov, Chem. Commun., 2006,
4614; Y. N. Belokon’, J. Hunt and M. North, Synlett, 2008, 2150;
Y. N. Belokon’, J. Hunt and M. North, Tetrahedron: Asymmetry,
2008, 19, 2804; M. North and M. Omedes-Pujol, Tetrahedron Lett.,
2009, 50, 4452.
4 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 and M. North, Helv. Chim.
Acta, 2002, 85, 3301.
5 A. J. Blacker and I. N. Houson, WO 02/066410A1, 2002;
S. Hanesssian, G. J. Reddy and N. Chahal, Org. Lett., 2006, 8,
5477; G. C. Lloyd-Jones, P. D. Wall, J. L. Slaughter, A. J. Parker
and D. P. Laffan, Tetrahedron, 2006, 62, 11402.
6 H. Yang, L. Zhang, P. Wang, Q. Yang and C. Li, Green Chem.,
2009, 11, 257.
7 B. Kirschbaum and R. Fell, WO 03/005010A3, 2003;
A. Watanabe, K. Matsumoto, Y. Shimada and T. Katsuki, Tetra-
hedron Lett., 2004, 45, 6229; J. Takaki, H. Egami, K. Matsumoto,
B. Saito and T. Katsuki, Chem. Lett., 2008, 37, 502; N. H. Khan,
S. Agrawal, R. I. Kureshy, S. H. R. Abdi, K. J. Prathap and
R. V. Jasra, Eur. J. Org. Chem., 2008, 4511.
8 K. Dhara, K. Sarkar, P. Roy, M. Nandi, A. Bhaumik and
P. Banerjee, Tetrahedron, 2008, 64, 3153.
9 Y. N. Belokon’, W. Clegg, R. W. Harrington, V. I. Maleev,
M. North, M. Omedes-Pujol, D. L. Usanov and C. Young,
Chem.–Eur. J., 2009, 15, 2148.
10 Y. N. Belokon’, M. North, V. I. Maleev, N. V. Voskoboev,
M. A. Moskalenko, A. S. Peregudov, A. V. Dmitriev,
N. S. Ikonnikov and H. B. Kagan, Angew. Chem., Int. Ed., 2004,
43, 4085; Y. N. Belokon’, W. Clegg, R. W. Harrington, C. Young
and M. North, Tetrahedron, 2007, 63, 5287; Y. N. Belokon’,
W. Clegg, R. W. Harrington, M. North and C. Young, Inorg.
Chem., 2008, 47, 3801.
11 K. Nakajima, M. Kojima and J. Fujita, Chem. Lett., 1986, 1483;
K. Nakajima, K. Kojima, M. Kojima and J. Fujita, Bull. Chem. Soc.
Jpn., 1990, 63, 2620; H. Schmidt, M. Bashirpoor and D. Rehder,
J. Chem. Soc., Dalton Trans., 1996, 3865; P. P. Reddy, C.-Y. Chu,
D.-R. Hwang, S.-K. Wang and B.-J. Uang, Coord. Chem. Rev., 2003,
237, 257; A. T. Radosevich, C. Musich and F. D. Toste, J. Am. Chem.
Soc., 2005, 127, 1090.
12 D. M. Murphy, I. A. Fallis, E. Carter, D. J. Willock, J. Landon,
S. Van Doorslaer and E. Vinck, Phys. Chem. Chem. Phys., 2009,
11, 6757.
The generation of cyanogen in this reaction via a non-
radical mechanism is consistent with the proposed formation
of dinuclear vanadium complexes in asymmetric cyanohydrin
synthesis. A plausible mechanism is shown in Scheme 2.
Ligand exchange between 1a–c and TMSCN generates
vanadiumVoxo cyanide complex 1d which can form bimetallic
species 3. Complex 3 can eliminate cyanogen to form
bimetallic vanadiumIVoxo(salen) species 4 which can dissociate
back to mononuclear complex 2. During asymmetric cyano-
hydrin synthesis, complexes 1a–c and 2 can combine to form
bimetallic, mixed oxidation state complex 5 which is the key
complex in the previously proposed bimetallic mechanism.9
Alternatively, complex 5 could be formed by the oxidation of
complex 4 by oxygen or complexes 1a–c. Should too much of
complexes 1a–c be reduced to vanadiumIVoxo(salen) complex
2, then molecular oxygen and TMSCN can reconvert complex
2 into vanadiumVoxo(salen) complex 1d, thus ensuring that
the appropriate ratio of vanadiumIV and vanadiumV species is
maintained.
13 The coupling of VV with the electron spin in dinuclear VVOVIV
complexes is often undetectable by cw-EPR: see for example
M. Mahroof-Tahir, A. D. Keramidas, R. B. Goldfarb,
O. P. Anderson, M. M. Miller and D. C. Crans, Inorg. Chem.,
1997, 36, 1657.
14 A. Alberti and D. Macciantelli, in Electron Paramagnetic
Resonance. A Practitioner’s Toolkit, ed. M. Brustolon and
E. Giamello, Wiley, New Jersey, 2009.
15 W. K. Subczynski and H. M. Schwartz, in Biomedical EPR—Part
A: free radicals, metals, medicine, and physiology, ed. S. S. Eaton,
G. R. Eaton and L. J. Berliner, Kluwer Academic Publishers,
Amsterdam, 2005.
In conclusion it has been shown that, unexpectedly,
vanadiumVoxo(salen)
complexes
are
reduced
to
vanadiumIVoxo(salen) by cyanide rather than by benzaldehyde
during asymmetric cyanohydrin synthesis. The cyanide is
oxidised to cyanogen via a non-radical mechanism. In addition
to providing new mechanistic insights into asymmetric cyano-
hydrin synthesis and other reactions catalysed by vanadiumV
16 T. K. Brotherton and J. W. Lynn, Chem. Rev., 1959, 59, 841.
ꢀc
This journal is The Royal Society of Chemistry 2010
3374 | Chem. Commun., 2010, 46, 3372–3374