Solid state-specific and chiral lattice-controlled asymmetric photoisomerization of 3-substituted propyl cobalt complexes, and direct observation of the intermediate complex

Article abstract of DOI:10.1016/j.jorganchem.2005.12.067

Solid state-specific and chiral lattice-controlled asymmetric photoisomerization of 3-cyanopropyl cobaloxime complexes coordinated with chiral axial ligand, 1a-h, was found to occur with moderate to relatively high enantioselectivities (～91%ee), even though the reaction proceeds through radical species. The enantioselectivities at a lower temperature (-78 °C) are extremely enhanced as compared to those at room temperature, in most cases. The configuration of the major enantiomer of the isomerized product is predictable from the shape of the reaction cavity, drawn based on the crystal structure of the starting material. The structure (including absolute configuration) of the intermediate 2-cyanopropyl complex was directly observed by X-ray crystallographic analyses of a single-crystal-to-single-crystal reaction of (S)-1-cyclohexylethylamine-coordinated 3-cyanopropyl cobaloxime, 1e.

Full text of DOI:10.1016/j.jorganchem.2005.12.067

Journal of Organometallic Chemistry 691 (2006) 2319–2326
www.elsevier.com/locate/jorganchem
Solid state-specific and chiral lattice-controlled asymmetric
photoisomerization of 3-substituted propyl cobalt complexes,
and direct observation of the intermediate complex
*
Yoshiaki Ohgo , Kanako Ishida, Yoshito Hiraga, Yoshifusa Arai, Seiji Takeuchi
Department of Applied Life Sciences, Niigata University of Pharmacy and Applied Life Sciences, Higashijima 265-1, Niigata 956-8603, Japan
Received 29 October 2005; received in revised form 23 December 2005; accepted 23 December 2005
Available online 23 February 2006
Abstract
Solid state-specific and chiral lattice-controlled asymmetric photoisomerization of 3-cyanopropyl cobaloxime complexes coordinated
with chiral axial ligand, 1a–h, was found to occur with moderate to relatively high enantioselectivities (ꢀ91%ee), even though the reac-
tion proceeds through radical species. The enantioselectivities at a lower temperature (ꢁ78 ꢀC) are extremely enhanced as compared to
those at room temperature, in most cases.
The configuration of the major enantiomer of the isomerized product is predictable from the shape of the reaction cavity, drawn based
on the crystal structure of the starting material. The structure (including absolute configuration) of the intermediate 2-cyanopropyl com-
plex was directly observed by X-ray crystallographic analyses of a single-crystal-to-single-crystal reaction of (S)-1-cyclohexylethylamine-
coordinated 3-cyanopropyl cobaloxime, 1e.
ꢁ 2006 Elsevier B.V. All rights reserved.
Keywords: Asymmetric; Photoisomerization; Chiral lattice; Lattice-controlled; Solid-state reaction; Crystalline-state reaction; Alkyl cobaloxime; X-ray
structure analysis
1. Introduction
(c) isomerized to 1-substituted propyl (a) via 2-substituted
propyl complexes (b) irreversibly by visible light irradiation
in the solid state [5]. The reaction does not proceed in solu-
tion and is therefore ‘‘solid state-specific’’ (see Scheme 1).
Here, if the crystal lattice is chiral, asymmetric isomeriza-
tion is expected to occur. Thus, several 3-cyanopropyl cob-
aloxime complexes coordinated with chiral axial bases were
prepared and solid-state photoisomerization was examined.
Solid-state (solvent-free) organic and organometallic
reactions have attracted much attention, not only from
purely chemical interest, but also from the viewpoint of
‘‘green and sustainable chemistry’’ [1]. Since a reactive
group of a molecule is surrounded and constrained by
other molecules in the solid-state, the reaction proceeds
with high selectivity, or a unique reaction can take place
which does not proceed in a solution state.
We have previously reported solid state-specific and irre-
versible photoisomerization [2] and chiral lattice-controlled
asymmetric (b ! a) photoisomerization of 2-substituted
ethyl cobaloxime complexes [3,4].
2. Results and discussion
2.1. Preparation and solid-state photoreaction
We prepared several 3-cyanopropyl cobaloxime
complexes (1a–h) coordinated with chiral axial ligand (a:
(R)-2-aminobutanol, b: (S)-phenylalaninol, c: (R)-1-phenyl-
ethylamine, d: (R)-2-phenylglycinol, e: (S)-1-cyclohexylethyl-
amine, f: (S)-1-benzyl-3-aminopyrrolidine, g: (S)-3-amino-
In an attempt to develop a novel and analogous system,
we found that 3-substituted propyl cobaloxime complexes
*
Corresponding author. Tel./fax: +81 250 25 5123.
E-mail address: ohgo@niigata-pharm.ac.jp (Y. Ohgo).
0022-328X/$ - see front matter ꢁ 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.12.067

2320
Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
Scheme 1.
pyrrolidine, h: (1S,2S)-2-amino-1-phenyl-1,3-propanediol),
which serves as a chiral handle for forming the chiral crys-
tal-lattice, by ligand displacement of aqua, benzylamine or
aniline-coordinated 3-cyanopropylcobaloximes with a–h,
and the structures were characterized by IR and NMR
spectra.
Powdered samples of these complexes 1a–h were sus-
pended in nujol and irradiated with a Wacom solar simula-
tor (light source of a 500 W Xe lamp with flux density:
100 mW/cm2) for a certain period of time (shown in
Fig. 1). The chiral base was displaced by an achiral ligand,
dimethylphenylphosphine (p) (Scheme 2), and the sample
was analyzed by HPLC using a chiral column (Chiralcel
OD–H), to obtain the ratio of dimethylphenylphosphine-
coordinated c complex (1p), (S) and (R)-isomers of b (2p)
and a (3p) complexes. As a representative example, the
time courses of the ratios of complexes, 1c, and the isomer-
ized products (2c and 3c), and the time courses of the
enantioselectivities of each isomerized product (2c and
3c) are shown in Figs. 1 and 2, respectively.
decrease. Finally, the 1-cyanopropyl complex (a-complex)
3c appeared, and the ratio increased rapidly with time.
After irradiation for 8 h, the ratio of the a-complex reached
to 98.1% (at this stage the ratio of b and c complexes are
0% and 1.9%, respectively). This pattern clearly shows
the reaction to be a typical consecutive one. The first-order
rate constant (k1 = 1.0 · 10ꢁ3 sꢁ1) of the first step was
obtained from the time course of 1c at the early stage.
Using the first-order rate constant, initial concentration
(in ratio) of 1c (100%) and concentration (in the ratio) of
2c at time t, we can calculate the rate constant of the sec-
ond step (k2) to be 1.6 · 10ꢁ2 sꢁ1 [5].
The enantioselectivity of b-complex reached 13.8%ee at
60 min and tended to decrease gradually thereafter. The
enantioselectivity of a-complex reached 71.2%ee (maxi-
mum) at 240 min and thereafter did not decrease notably.
Other complexes coordinated with other chiral amines
were also examined, and the rate constants, maximum
enantioselectivities, and the configuration of the major
enantiomer produced in each reaction at room temperature
are summarized in Table 1.
The configuration of the major enantiomer of the a
complex 3c produced was assumed from the sign of the
optical rotation of the final product 3p [6]. The configura-
tion of the major enantiomer produced in the other reac-
tion was inferred from the correlation of the HPLC data
((R)-enantiomer of 3p travels faster than the (S)-enantio-
mer on HPLC (Chiralcel OD–H) using ethanol/hexane
(5/95) as the solvent system). The configuration (at the car-
bon bonded to cobalt) of the major enantiomer of the b-
complex 2e produced, was assigned by direct observation
of the intermediate b-complex by X-ray crystallographic
analyses, as described later, and the configuration of the
major enantiomer of b-complexes produced in each reac-
tion was assigned by the correlation between the HPLC
data of 2p produced from each reaction and that of 2p pro-
duced from the reaction of 1e.
As shown in Fig. 1, the ratio of the starting material 1c
decreased very rapidly with time, and shortly after the ini-
tiation of the reaction, the ratio of 2-cyanopropyl complex
2c increased for a short period of time and then began to
Much higher enantioselectivity was expected to be
obtained at a lower temperature due to contraction of
the crystal lattice. So, we carried out the same reaction at
ꢁ78 ꢀC. The results are summarized in Table 1. As
expected, the enantioselectivity at ꢁ78 ꢀC was remarkably
enhanced as compared with that at room temperature, in
most cases examined. Especially, in the case of (S)-phenyla-
laninol-coordinated complex (1b), a dramatic enhancement
in enantioselectivity was observed: The enantioselectivity in
Fig. 1. The time courses of complexes 1c (m),2c ( ), and 3c (d).
n

Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
2321
Scheme 2.
The enantioselectivity at a lower temperature (ꢁ78 ꢀC)
is extremely enhanced as compared to that at room
temperature.
It is noteworthy that the enantioselectivity (of 2e) of the
first step in the reaction of 1e is exceptionally high and
moreover, there is a period in which the ratio of b-isomer
is relatively high. These advantages made it possible to
observe an intermediate state directly and also to deter-
mine the configuration of the major enantiomer (at the
carbon bonded to cobalt) of the intermediate b-complex
by X-ray analyses.
2.2. X-ray studies
2.2.1. Mechanism of enantioselection
In order to elucidate the mechanism of enantioselection,
we investigated the X-ray crystallographic analysis of 1e.
The crystal structure and molecular structure of 1e at
ꢁ80 ꢀC are shown in Figs. 3 and 4, respectively. Based
on the crystal structure, we can describe the reaction cavity
where the reactive group can move freely [7].
Fig. 2. The time courses of enantioselectivities for 2c (n) and 3c (d). 2c
was not detectable at 14400 s. and 28800 s.
Fig. 5(a) and (b) are the top view (viewed along the nor-
mal to the inplane ligand) and a side view of the cavity (the
projection from the side (N(1) and N(4))), respectively.
The nitrogen atom of CN of the b complex is considered
to be accommodated almost in the same region as that of
the c-complex, because the methylene group (at b) should
work as an adjuster, and the larger methyl group is prefer-
entially located in the wider space (in the direction toward
N(4)), and the small hydrogen is also located in the narrow
space. Consequently, the major enantiomer of the b com-
plex is expected to have (R)-configuration. However, for
the a carbon to bind to cobalt, to give the a-complex,
rod-like C(a)–C„N group demand a long-shaped space,
and to be nearly parallel to the planar inplane ligand so
a-complex (3b) at ꢁ78 ꢀC reached 90.9%ee (R), while that
at room temperature was 34%ee (R). In the reaction of
1b and 1c, the enantioselectivity of the b-complex (2b and
2c) was also extremely enhanced at ꢁ78 ꢀC, as compared
with those at room temperature (from 23%ee (R)(2b) at
r.t. to 63.8%ee (R)(2b) at ꢁ78 ꢀC; from 13.8%ee (S)(2c)
at r.t. to 60.7%ee (R)(2c) at ꢁ78 ꢀC).
Asymmetric induction was found to occur in every case
here examined. A relatively high degree of enantioselectiv-
ities was attained, even though the reaction proceeded
through radical species. Generally, the reaction rate in
the second step (b ! a) is much faster than that in the first
step (c ! b).

2322
Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
Table 1
Rate constants and enantioselectivities in the solid-state photoisomerization of chiral base-coordinated 3-cyanopropyl cobaloximes
Substrate (1)
Rate constant (sꢁ1
r.t.
)
Enantioselectivity (%ee)a (config.)b
Code
B*
r.t.
ꢁ78 ꢀC
k1
k2
2
3
2
3
la
(R)-2-Aminobutanol
(S)-Phenylalaninol
(R)-1-Phenylethylamine
(R)-2-Phenylglycinol
(S)-1-Cyclohexylethylamine
(S)-1-Benzyl-3-amino-pyrrolidine
(S)-3-Aminopyrrolidine
9.9 · 10ꢁ5
4.2 · 10ꢁ5
1.0 · 10ꢁ3
4.0 · 10ꢁ5
9.5 · 10ꢁ5
1.6 · 10ꢁ4
6.7 · 10ꢁ5
5.5 · 10ꢁ5
1.2 · 10ꢁ3
4.9 · 10ꢁ4
1.6 · 10ꢁ2
5.8 · 10ꢁ4
6.8 · 10ꢁ4
3.1 · 10ꢁ3
1.5 · 10ꢁ3
1.2 · 10ꢁ4
25.5(R)
23.1(R)
13.8(S)
5.4(S)
64.1(R)
26.0(R)
2.5(R)
28.4(S)
34.0(R)
71.2(S)
51.4(R)
10.2(S)
11.3(R)
2.9(R)
17.5(R)
63.8(R)
60.7(R)
41.2(S)
90.9(R)
52.1(S)
1b
1c
1d
1e
1f
1g
1h
a
65.4(R)
50.8(R)
36.0(S)
44.9(R)
(1S,2S)-2-Amino-1-phenyl-1,3-propanediol
31.8(R)
32.6(S)
19.7(R)
25.4(S)
Maximum enantioselectivity.
Configuration of the major enantiomer.
b
(a)
Fig. 3. Crystal structure of 1e.
(b)
Fig. 5. Top view (a) and side view (b) of reaction cavity and alkyl group
drawn based on the crystal structure of 1e.
that the cyano group cannot make a parallel movement
toward the in-plane ligand from the original position,
because there is not enough room in this region. Conse-
quently, the cyano group must make fairly large move-
ments. C(a)–C„N should be directed toward C(4) for
Fig. 4. Molecular structure of 1e. Hydrogen atoms are omitted for clarity.

Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
2323
meeting the long-shaped and rod-like steric demand. In this
situation, the ethyl group is favored to locate in the space
toward O(2) and N(2), then, the configuration becomes
S, which is consistent with the result (see Table 1). The con-
figuration of the major enantiomer of 3c is also predictable
from the shape of the reaction cavity of 1c, drawn based on
the crystal structure [8].
Thus, the configuration of the major enantiomer (of a
and b isomers) produced, was found to be predictable from
the shape of the reaction cavity, drawn based on the crystal
structure of the starting material.
2.2.2. Direct observation of intermediate b-complex in a
single crystal-to-single crystal reaction
We have previously found and reported many examples
of crystalline state racemization of chiral alkyl coordinated
cobaloxime complexes in which the crystal lattice does not
collapse during the reaction [7,9]. We expected that the
present reaction would also proceed without degradation
of crystallinity.
In order to prove the hypothetical movement of alkyl in
the process of the isomerization of 1e, and the inferred
mechanism for the configurational reversal of the major
enantiomer between the first and second steps, we tried
to observe directly the movement of the alkyl group with
the reaction time by X-ray crystallographic analysis.
The crystallographic data and final R factors obtained
by crystal structure analyses of samples irradiated for 10
and 20 h at ꢁ78 ꢀC indicated that the crystals still main-
tained good crystallinity, enough to enable crystal structure
analysis even after irradiation.
Fig. 6. Molecular structure of 1e (after 20 h of irradiation). Hydrogen
atoms are omitted for clarity. 2-Cyanopropyl, C(182)–C(17)–C(192)–
C(202)–N(62), appears clearly in occupancy factor of 36%. The figure
shows that 2-cyanopropyl has (R)-configuration.
lower temperature (ꢁ78 ꢀC) is extremely decreased in the
isomerization of 1e. Further investigation along this line
is now being undertaken.
Molecular structure and reaction cavity obtained by
structural analysis at irradiation time (20 h) are shown in
Figs. 6 and 7.
3. Experimental
3.1. General
After 10 h of irradiation, the 2-cyanopropyl (the inter-
mediate b-complex) appeared, though the occupancy was
only 15%. The intermediate b-complex was clearly
observed in an occupancy factor of 36% in the molecular
structure obtained (by structural analysis) after 20 h of
irradiation. The cavity at the same stage as the molecular
structure shows that the cyano nitrogen of the b-complex
produced is situated in almost the same region as in the ori-
ginal, as predicted previously. C(a)–C„N turned nearly
vertical toward the in-plane ligand, and the methyl group
of b-cyanopropyl is located in the region directed toward
N(4) from the cobalt atom, as expected previously. Thus,
the configuration of the major enantiomer of b-complex
(2e) was proved to be R which was consistent with the pre-
dicted one.
At present, we could not observe 3e in the single-crystal-
to-single-crystal reaction of 1e. Two factors, described
below, are considered to operate synergetically, which
would be the reason why we could not observe 3e in the
reaction using single crystal at ꢁ78 ꢀC: (1) The reaction
rates in the crystalline-state are usually 1 order lower than
those in the solid-state reactions using a finely powdered
sample. (2) The reaction rate of the second step (k2) at a
1H NMR spectra were recorded on JEOL JNM- a400
and BRUKER BIOSPIN AVANCE DPX250 spectrome-
ters, using TMS as the internal standard. IR spectra were
recorded on a JASCO FT/IR-460 PLUS spectrometer.
Optical rotations were measured on a Perkin–Elmer 241
polarimeter. A WACOM Solar Simulator WX-105H (Xe
lamp with Flux density: 100 mW/cm2) was used as a light
source for photoreactions. Microanalyses were performed
on a Perkin–Elmer 240-002 and CE INSTRUMENTS
EA1110.
3.2. Preparation of substrates (1a–h)
Substrates (1a–h) were prepared basically according to
the method described in a previous paper [5], except for
using chiral base in axial ligand displacement.
A typical procedure of ligand displacement: 1.347 g
(2.99 mmol) of (aniline)(3-cyanopropyl)bis(dimethylglyoxi-
mato)cobalt(III) was dissolved in methanol (40 ml). To the
solution 5 g of an ion exchange resin (Dowex 50W-X2) and
water (10 ml) were added. The reaction mixture was stirred
atroomtemperature.Thedegreeofaxialliganddisplacement

2324
Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
CDCl3): d 3.63(dd, 1H, J = 3.8 and 11.4 Hz, –CH–O of
(a)
aminobutanol), 3.22(dd, 1H, J = 6.1 and 11.4 Hz, –CH–O
of aminobutanol), 2.34(m, 1H, –NH), 2.25(s, 6H, CH3
of dimethylglyoxime), 2.24(s, 6H, CH3 of dim-
ethylglyoxime), 2.23(m, 1H, N–CH– of aminobutanol),
2.21(broad, 1H, –OH of aminobutanol), 2.18(t, 2H,
J = 7.2 Hz, –CH2–CN), 1.67(m, 1H, –NH of aminobuta-
nol), 1.43–1.14(m, 6H, –C–CH2–C of aminobutanol, Co–
CH2–, and Co–C–CH2–), 0.77(t, 3H, J = 7.5 Hz, –C–CH3
of aminobutanol). IR(KBr): 3459.7, 3293.3, 3244.7,
2238.0, 1569.8, 1228.4, and 1069.8 cmꢁ1. Anal. Calc. for
C16H31CoN6O5: C, 43.05; H, 7.00; N, 18.83. Found: C,
42.77; H, 7.29; N 18.53%. 1b: 1H NMR(250 MHz, CDCl3):
d 7.30(m, 3H, aromatic), 7.02(m, 2H, aromatic), 3.56(dd,
1H, J = ꢀ 1.0 and 10.0 Hz, –CH–O of phenylalaninol),
3.34(dd, 1H, J = 2.8 and 10.0 Hz, –CH–O of phenylalani-
nol), 2.60(s, 3H, N–CH–CH2– of phenylalaninol), 2.30(m,
1H, –NH of phenylalaninol), 2.19(s, 6H, CH3 of dim-
ethylglyoxime), 2.16(m, 2H, –CH2–CN), 2.14(s, 6H, CH3
of dimethylglyoxime), 2.00(broad, 1H, –OH of phenylalan-
inol), 1.71(m, 1H, –NH of phenylalaninol), 1.33(m, 2H,
Co–CH2–), 1.14 (m, 2H, Co–C–CH2–). IR(KBr): 3331.4,
3292.9, 3244.2, 2237.0, 1559.2, 1222.7, 1085.2, 746.8, and
704.9 cmꢁ1. Anal. Calc. for C21H33CoN6O5: C, 49.61; H,
6.54; N, 16.53. Found: C, 49.89; H, 6.70; N, 16.23%. 1c:
1H NMR(400 MHz, CDCl3): d 7.29(m, 3H, aromatic),
7.08(m, 2H, aromatic), 3.69(tq, 1H, J = 6.3 and 6.3 Hz,
–N–CH– of phenylethylamine), 2.44(m, 1H, –NH of phen-
ylethylamine), 2.156(t, 2H, J = 7.2 Hz, –CH2–CN),
2.151(s, 6H, CH3 of dimethylglyoxime), 2.149(s, 6H, CH3
of dimethylglyoxime), 1.72(m, 1H, –NH of phenylethyl-
amine), 1.35(m, 2H, Co–CH2–), 1.26(d, 3H, J = 6.8 Hz,
CH3 of phenylethylamine), 1.16 (m, 2H, Co–C–CH2–).
IR(KBr): 3321.8, 3273.1, 2246.7, 1564.5, 1235.2, 1077.1,
(b)
Fig. 7. Top view (a) and side view (b) of reaction cavity and alkyl groups
drawn based on the crystal structure of 1e (after 20 h of irradiation).
Hydrogen atoms of the alkyl groups are omitted for clarity. Black line and
red line are carbon skeletons of the original 3-cyanopropyl and the
isomerized 2-cyanopropyl group, respectively. Cyano nitrogen of each
alkyl is colored with blue. (For interpretation of the references to colour in
this figure legend, the reader is referred to the Web version of this article).
769.5, 761.3 and 702.4 cmꢁ1
.
Anal. Calc. for
C20H31CoN6O4: C, 50.21; H, 6.53; N, 17.57. Found: C,
50.48; H, 6.79; N, 17.60%. 1d: 1H NMR(250 MHz,
CDCl3): d 7.30(m, 3H, aromatic), 7.08(m, 2H, aromatic),
3.78(m, 1H, N–CH– of phenylglycinol), 3.62(dd, 1H,
J = 3.9 and 11.5 Hz, –CH–O– of phenylglycinol),
3.15(dd, 1H, J = 9.8 and 11.5 Hz, –CH–O– of phenylgly-
cinol), 2.76(d, 1H, J = 11.1 Hz, –NH of phenylglycinol),
2.48(t, 1H, J = 10.9 Hz, –NH of phenylglycinol),
2.42(broad, 1H, –OH of phenylglycinol), 2.14(t, 2H,
J = 7.0 Hz, –CH2–CN), 2.01(s, 6H, CH3 of dim-
ethylglyoxime), 2.00(s, 6H, CH3 of dimethylglyoxime),
1.30(m, 2H, Co–CH2–), 1.13(m, 2H, Co–C–CH2–).
IR(KBr): 3314.6, 3247.1, 2238.0, 1559.2, 1230.8, 1088.1,
1056.3, 760.8, and 703.4 cmꢁ1. Anal. Calc. for C20H31-
CoN6O5: C, 48.59; H, 6.32; N, 17.00. Found: C, 48.57;
H, 6.46; N, 16.87%. 1e: 1H NMR(250 MHz, CDCl3):
d2.25(s, 12H, CH3 of dimethylglyoxime), 2.17(t, 2H, J =
7.0 Hz, –CH2–CN), 2.15–0.65(m, 13H, –C6H11 and –NH2
of cyclohexylethylamine), 2.08(m, 1H, N–CH– of cyclo-
hexylethylamine), 1.34(m, 2H, Co–CH2–), 1.15(m, 2H,
Co–C–CH2–), 0.86(d, 3H, J = 6.6 Hz, N–C–CH3 of cyclo-
hexylethylamine). IR(KBr): 3314.6, 3252.8, 2922.6, 2239.0,
with aqua was monitored by TLC. After completion of the
ligand displacement, the reaction mixture was filtered.
0.39 ml (1.02 eq) of (R)-1-phenylethylamine was added to
the filtrate with stirring. Ligand exchange reaction of aqua
with the amine was monitored by TLC. After completion
of the reaction, the solution was evaporated to dryness
under a reduced pressure. The crude product was recrystal-
lized from ethyl acetate (70 ml) and hexane (140 ml) to give
0.111 g of orange crystals of 1c. From the filtrate, another
0.61 g of crystals was obtained. The filtrate was evaporated
to dryness under a reduced pressure to give 0.498 g of crys-
tals from which 0.096 g of crystals (1c) was also obtained
by the same recrystallization procedure as above. Total
yield: 0.818 g (57.1%).
1H NMR, IR, and microanalyses data of several key
substrates are described as follows. 1a: 1H NMR(400 MHz,

Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
2325
1563.0, 1235.7, and 1093.4 cmꢁ1
.
Anal. Calc. for
3.3.1. Preparation and 1H NMR, IR and microanalyses data
for 1p, 2p, and 3p
C20H37CoN6O4: C, 49.58; H, 7.70; N, 17.35. Found: C,
49.41; H, 7.64; N, 17.21%. 1f: 1H NMR(400 MHz, CDCl3):
d 7.28(m, 3H, aromatic), 7.21(m, 2H, aromatic), 3.55(d,
1H, J = 12.8 Hz, N–CH–Ph of 1-benzyl-3-aminopyrroli-
dine), 3.45(d, 1H, J = 12.8 Hz, N–CH–Ph of 1-benzyl-3-
aminopyrrolidine), 2.92(m, 1H, H-3 of 1-benzyl-3-amino-
pyrrolidine), 2.79(dt, 1H, J = 3.3 and 8.7 Hz, H-5 of 1-ben-
zyl-3-aminopyrrolidine), 2.37(m, 1H, H-2 of 1-benzyl-
3-aminopyrrolidine), 2.24(dd, 1H, J = 6.0 and 10.4 Hz,
H-20 of 1-benzyl-3-aminopyrrolidine), 2.20(s, 6H, CH3 of
dimethylglyoxime), 2.19(s, 6H, CH3 of dimethylglyoxime),
2.15(t, 2H, J = 7.1 Hz, –CH2–CN), 2.15–2.00(m, 3H, –NH,
H-4, and H-50 of 1-benzyl-3-aminopyrrolidine), 1.93(m,
1H, –NH of 1-benzyl-3-aminopyrrolidine), 1.34(m, 1H, H-40
of 1-benzyl-3-aminopyrrolidine), 1.31(m, 2H, Co–CH2–),
1.15(m, 2H, Co–C–CH2–). IR(KBr): 3479.9, 3270.2,
3242.7, 2247.2, 1565.4, 1235.2, 1088.1, 755.0, and 706.8
cmꢁ1. Anal. Calc. for C23H36CoN7O4 Æ H2O: C, 50.09; H,
6.94; N, 17.78. Found: C, 50.19; H, 6.87; N, 17.79%.
Dimethylphenylphosphine-coordinated 3-cyanopropyl,
2-cyanopropyl and 1-cyanopropyl cobaloxime complexes
(1p, 2p and 3p) used as the authentic samples for the quan-
titative analysis were prepared by ligand displacement of
the corresponding aniline- or benzylamine-coordinated
cobaloxime complexes with dimethylphenylphosphine
according to the same procedure described in the previous
Section 3.2 and also in a previous report [5]. 1H NMR, IR
and microanalysis data are described as follows. 1p: H
NMR(400 MHz, CDCl3): d 7.36(m, 3H, aromatic), 7.10
(m, 2H, aromatic), 2.19(t, 2H, J = 7.4 Hz, –CH2–CN),
1.94(d, 12H, JH–P = 3.4 Hz, CH3 of dimethylglyoxime),
1.53(m, 2H, Co–CH2–), 1.39(d, 6H, JH–P = 9.1 Hz, CH3
of PMe2Ph), 1.28(m, 2H, Co–C–CH2–). IR(KBr): 2902.8,
2239.0, 1558.7, 1234.2, 1091.0, 743.4, and 691.8 cmꢁ1
.
Anal. Calc. for C20H31CoN5O4P: C, 48.49; H, 6.31; N,
14.14. Found: C, 48.61; H, 6.43; N, 13.89%. 2p: 1H
NMR(400 MHz, CDCl3):
d 7.36(m, 3H, aromatic),
7.09(m, 2H, aromatic), 2.45(ddd, 1H, J = 4.4 and 6.6 Hz,
J = 17.1 Hz, –CH–CN), 1.99(ddd, 1H, J = 2.9 and
3.3. A typical procedure for the photoreactions and the
subsequent analytical consequences
9.3 Hz, J = 17.1 Hz, –CH–CN), 1.95(d, 6H, JH–P
2.7 Hz, CH3 of dimethylglyoxime), 1.94(d, 6H, JH–P
=
=
In a typical experiment, 51.4 mg of crystals, 1c, was sus-
pended in 2.5 ml of nujol and powdered by a grinder for
10 min. The suspended sample was transferred into a
petri-dish with 8.8 cm-diameter and a trace amount of
the residue left on the dish was washed two times with
another 2 ml of nujol, and was transferred into the same
petri-dish. The sample suspended in nujol was irradiated
with a solar simulator (Flux density: 100 mW/cm2) for
28800 s (8 h). The reaction products were filtered, and the
nujol was washed out with hexane to give 39.5 mg of crys-
tals. To crystals were added 30 ml of methanol, 0.5 ml of
6 N HCl, and 0.018 ml (1.5 eq) of dimethylphenylphos-
phine. The resulting solution was stirred and the reaction
was monitored by TLC. After the completion of the ligand
displacement, the reaction mixture was neutralized by a
NaHCO3 solution and was evaporated to dryness under
a reduced pressure. To the residue water was added and
the mixture was extracted with methylene chloride. The
methylene chloride solution was applied to a silica-gel col-
umn and developed with a mixed solvent [ethyl acetate/
hexane (1/1)]. The eluate was concentrated to dryness
under a reduced pressure to give 40 mg of crystalline solid.
The sample was analyzed by HPLC using Chiralcel OD–H
(solvent system: ethanol/hexane (5/95)). The data showed
the ratio (and enantioselectivity) of 1p:2p:3p to be
1.93:0:98.07 (68.7%ee, rich in (S)-enantiomer), respectively.
[a]589 ꢁ31.5ꢀ (c 0.213, CHCl3, 15.5 ꢀC), [a]578 ꢁ32.9ꢀ (c
0.213, CHCl3, 15.5 ꢀC). The HPLC data for the sample,
obtained by treating the reaction products after 14400 s
(4 h) of irradiation with the same procedure as above,
showed the ratio (and enantioselectivity) of 1p:2p:3p
to be 2.49:0:97.51 (71.2%ee, rich in (S)-enantiomer),
respectively.
3.2 Hz, CH3 of dimethylglyoxime), 1.37(d, 6H,
JH–P = 9.3 Hz, CH3 of PMe2Ph), 1.24(m, 1H, Co–CH–),
0.81(dd, 3H, J = 7.1 Hz, JH–P = 7.1 Hz, Co–C–CH3).
IR(KBr): 2850.8, 2229.3, 1550.5, 1233.3, 1093.0, 908.8,
751.6, and 696.7 cmꢁ1. Anal. Calc. for C20H31CoN5O4P:
C, 48.49; H, 6.31; N, 14.14. Found: C, 48.52; H, 6.31; N,
1
14.10%. 3p: H NMR(400 MHz, CDCl3): d 7.41(m, 3H,
aromatic), 7.14(m, 2H, aromatic), 2.06(m, H, Co–CH–
CN), 2.03(d, 6H, J = 4.0 Hz, CH3 of dimethylglyoxime),
2.02(d, 6H, J = 2.6 Hz, CH3 of dimethylglyoxime),
1.420(d, 3H, J = 10.6 Hz, CH3 of PMe2Ph), 1.417(d, 3H,
J = 12.0 Hz, CH3 of PMe2Ph), 1.14–0.92(m, 2H, Co–C–
CH2–), 0.96(m, 3H, Co–C–C–CH3). IR(KBr): 2959.7,
2195.6, 1555.8, 1234.7, 1090.1, 910.2, 749.7, and
693.3 cmꢁ1. Anal. Calc. for C20H31CoN5O4P: C, 48.49;
H, 6.31; N, 14.14. Found: C, 48.40; H, 6.18; N, 14.16%.
3.4. X-ray crystallographic studies
Tables of atomic coordinates, anisotropic thermal
parameters, bond lengths and angles of 1e and 1e (after
20 h of irradiation) may be obtained free of charge from
The Director CCDC, 12 Union Road, Cambridge CB2 1
EZ, UK, on quoting the deposition numbers CCDC
286804 and 286805, the names of authors and the journal
citation (http://www.ccdc.cam.ac.uk). The procedures of
crystal structure determination for both 1e and 1e (after
20 h of irradiation) were the same. X-ray data were col-
lected at 173 K on a Rigaku AFC7R four circle diffractom-
eter [10] using Mo Ka X-ray (0.71069) source and a
graphite monochromator. The unit cell dimensions were
obtained from a least-squares fit to setting angles of 25
reflections. w scan absorption corrections were applied

2326
Y. Ohgo et al. / Journal of Organometallic Chemistry 691 (2006) 2319–2326
[11]. The structures were solved by direct methods using
SIR97 [12] for 1e and MITHRIL for 1e (after 20 h of irradi-
ation) [13] and refined by full-matrix least-square method
using shelxl97 [14]. In the final step of the refinement pro-
cedure, all non-hydrogen atoms were refined with aniso-
tropic displacement parameters for 1e. For 1e (after 20 h
of irradiation), cyanoalkyl groups bound to cobalt directly
were disordered in two parts, (C181, C191, C201 and N61)
and (C182, C192, C202 and N62) with refined occupancies
of 0.635 and 0.365. Non-hydrogen atoms in the disordered
cyanoalkyl groups were refined with isotropic displacement
parameters. The other non-hydrogen atoms were refined
with anisotropic displacement parameters. All hydrogen
atoms were placed in calculated positions and included
in the final least squares refinement with a riding model.
for Scientific Research (C) from Japan Society for the Pro-
motion of Science. The authors are grateful to Ms. Yukie
Suzuki for preparation of substrates and her technical
assistance.
References
[1] K. Tanaka, F. Toda, Chem. Rev. 100 (2000) 1025–1074.
[2] Y. Ohgo, S. Takeuchi, Chem. Commun. (1985) 21–22.
[3] Y. Ohgo, Y. Arai, M. Hagiwara, S. Takeuchi, H. Kogo, A. Sekine,
H. Uekusa, Y. Ohashi, Chem. Lett. (1994) 715–716.
[4] Y. Ohgo, M. Hagiwara, M. Shida, Y. Arai, S. Takeuchi, Mol. Cryst.
Liq. Cryst. 277 (1996) 241–246.
[5] Y. Ohgo, F. Kurashima, Y. Hiraga, K. Ishida, N. Takatsu, Y. Arai,
S. Takeuchi, Chem. Lett. (2001) 1190–1191.
[6] Y. Ohgo, S. Takeuchi, Y. Natori, J. Yoshimura, Y. Ohashi, Y.
Sasada, Bull. Chem. Soc. Jpn. 54 (1981) 3095–3099.
˚
˚
1e: C20H37CoN6O4, a = 12.521(2) A, b = 16.775(8) A, c =
[7] The reaction cavity is defined as a space limited by a concave surface
of the spheres of surrounding atoms. The radius of each sphere (of the
˚
11.323(2) A, orthorhombic P212121, Z = 4, crystal size
˚
atom) is taken to be 1.2 A greater than the van der Waals radius of
(mm) = 0.27 · 0.18 · 0.13, T = 193 ꢀK, h range for data
collection (ꢀ) = 2.71–27.49, index ranges h = ꢁ16 ! 16,
k = ꢁ21 ! 21, l = ꢁ14 ! 14, reflections collected/unique
[Rint] = 7531/5448 [0.0565], R1 = 0.0428(I > 2r(I)), R1 =
0.0902(all), wR = 0.1184(all), goodness-of-fit = 1.119,
Flack v parameter = ꢁ0.03(2). 1e (after 20 h of irradia-
the corresponding atom. See references cited below. (a) Y. Ohashi, K.
Yanagi, T. Kurihara, Y. Sasada, Y. Ohgo, J. Am. Chem. Soc. 103
(1981) 5805–5812;
(b) Y. Ohashi, A. Uchida, Y. Sasada, Y. Ohgo, Acta Crystallogr. B
39 (1983) 54.
[8] Preliminary X-ray result of 1c is briefly described in the Proceedings
of ICCOSS XIII; A. Sekine, M. Yoshiike, Y. Ohashi, K. Ishida, Y.
Arai, Y. Ohgo, Mol. Cryst. Liq. Cryst. 313 (1998) 321–326, The
detailed results will be published later.
˚
˚
tion): C20H37CoN6O4, a = 12.563(1) A, b = 16.749(2) A,
˚
c = 11.382(3) A, orthorhombic P212121, Z = 4, crystal size
(mm) = 0.25 · 0.22 · 0.22, T = 193 ꢀK, h range for data
collection (ꢀ) = 2.71–27.49, index ranges h = ꢁ9 ! 16,
k = 0 ! 21, l = 0 ! 14, reflections collected/unique
[Rint] = 3618/3466 [0.0258], R1 = 0.0339(I > 2r(I)), R1 =
0.0539(all), wR = 0.0948(all), goodness-of-fit = 1.309,
Flack v parameter = ꢁ0.00(2).
[9] Y. Ohashi, Y. Sakai, A. Sekine, Y. Arai, Y. Ohgo, N. Kamiya, H.
Iwasaki, Bull. Chem. Soc. Jpn. 68 (1995) 2517, and references cited
therein.
[10] MSC/AFC Diffractometer Control Software, Molecular Structure
Corporation, MSC, 3200 Research Forest Drive, The Woodlands, TX
77381, USA.
[11] A.C.T. North, D.C. Phillips, F.S. Mathews, Acta Cryst. A 24 (1968)
351.
[12] A. Altomare, M.C. Burla, M. Camalli, G. Cascarano, C. Giacovazzo,
A. Guagliardi, A.G.G. Moliterni, G. Polidori, R. Spagna, J. Appl.
Cryst. 32 (1999) 115.
[13] C.J. Gilmore, J. Appl. Cryst. 17 (1984) 42.
[14] G.M. Sheldrick, Shelxl97 Program for the Refinement of Crystal
Structures, University of Go¨ttingen, Germany, 1997.
Acknowledgements
This work was partly supported by Grant-in-Aids for
Scientific Research from the Ministry of Education, Sci-
ence, Sports and Culture, Japan, and also a Grant-in-Aid