Direct observation of deuterium migration in crystalline-state reaction by single-crystal neutron diffraction. II. 3-1 photoisomerization of a cobaloxime complex

Article abstract of DOI:10.1107/S0108768199012008

Single crystal neutron diffraction analysis of photo-exposed (3-cyanopropyl-d₂ ^α,α)-[(R)-1-phenylethylamine-d ₁₁]bis(dimethylglyoximato-d₁₄)cobalt(III) was carried out in order to clarify the mechanism of the cr

Full text of DOI:10.1107/S0108768199012008

research papers
Direct observation of deuterium migration in
Acta Crystallographica Section B
Structural
Science
crystalline-state reaction by single-crystal neutron
diffraction. II. 3±1 Photoisomerization of a cobal-
oxime complex
ISSN 0108-7681
Takashi Ohhara,a Jun Harada,a,b
Yuji Ohashi,a* Ichiro Tanaka,c
Shintaro Kumazawac and Nobuo
Niimurac
Single crystal neutron diffraction analysis of photo-exposed
(3-cyanopropyl-d2ꢀ,ꢀ)-[(R)-1-phenylethylamine-d11]bis(di-
Received 25 June 1999
Accepted 17 September 1999
methylglyoximato-d14)cobalt(III) was carried out in order to
clarify the mechanism of the crystalline-state photoisomeriza-
tion of the 3-cyanopropyl group bonded to the Co atom in
some cobaloxime complexes. Before irradiation the two H
atoms bonded to the C1 atom of the 3-cyanopropyl group were
exchanged with the D atoms such as ÐCH2CH2CD2CN. On
exposure to a xenon lamp, the cell dimensions of the crystal
were gradually changed. After 7 d exposure the change
became insigni®cantly small. The structure was analyzed by
neutron diffraction. The 3-cyanopropyl group was trans-
formed to the 1-cyanopropyl group such as ÐCD(CN)C-
(H1/2,D1/2)2CH3 with retention of the single-crystal form. This
indicates that one of the D atoms bonded to C1 migrates to
either position bonded to C2. The other atoms of the complex
remained unchanged. These results indicate that photoisomer-
ization proceeded in two steps: the 3-cyanopropyl group was
isomerized to the 2-cyanopropyl group in the ®rst place and
then the 2-cyanopropyl group was transformed to the 1-
cyanopropyl group. Moreover, it was made clear that the
second-step isomerization was irreversible, since one of the D
atoms was retained. The disordered structure at C2 is
estimated to be caused by the interconversion between the
1-cyanopropyl group produced and its dehydrogenated ole®n
after the photoisomerization.
aDepartment of Chemistry and Materials
Science, Tokyo Institute of Technology, O-
okayama, Meguro-ku, Tokyo 152-8551, Japan,
bCREST Fellow, Japan Science and Technology
Corporation, Kawaguchi, Saitama 332-0012,
Japan, and cAdvanced Science Research Center,
Japan Atomic Energy Research Institute (JAERI),
Tokai, Ibaraki 319-1195, Japan
Correspondence e-mail:
yohashi@chem.titech.ac.jp
1. Introduction
Since the photoracemization of a cobaloxime complex was
found to proceed with retention of the single crystal form
(Ohashi & Sasada, 1977), such crystalline-state reactions have
been highly regarded because their reaction processes at the
initial, intermediate and ®nal stages have been analyzed by
single-crystal X-ray diffraction. Several reactive groups such
as 1-cyanoethyl (Ohashi, 1988), 1-methoxycarbonylethyl
(Sekine, Saitoh et al., 1998), 1,2-bis(methoxycarbonyl)ethyl
(Ohashi et al., 1995) and 1,2-bis(ethoxycarbonyl)ethyl (Sato &
Ohashi, 1999) bonded to the Co atoms in some cobaloxime
complexes have been found to be racemized on exposure to a
xenon lamp without destroying the single-crystal form.
Moreover, the 2-cyanoethyl group bonded to the Co atom in
some cobaloxime complex crystals was found to be isomerized
to the 1-cyanoethyl group with retention of the single-crystal
form by irradiation with the xenon lamp (Sekine et al.,
1997a,b). The rate and selectivity of the reactions have been
explained on the basis of the crystal structure, i.e. the reaction
cavity (Ohashi et al., 1981). However, the reaction mechanisms
of CoÐC bond cleavage by photo-irradiation, the rotation or
# 2000 International Union of Crystallography
Printed in Great Britain ± all rights reserved
ꢀ
Acta Cryst. (2000). B56, 245±253
Ohhara et al. Deuterium migration 245

research papers
translation of the alkyl radical produced and the recombina-
tion of the CoÐC bond in the process of racemization or
isomerization have been obscure.
The single-crystal neutron diffraction analysis, on the other
hand, can distinguish the H atom from the D atom, because
the H atom has negative neutron scattering length, while the
D atom has a positive one (Koester, 1977). This suggests that if
the H atoms in speci®c positions of the reactive group are
exchanged with the D atom and the crystallinity was kept in
the reaction, the reaction process will be made clear since the
migration of the D atom, i.e. the hydrogen-transfer process, is
easily observed by neutron diffraction analysis. In the previous
paper the racemization process of the 1-cyanoethyl group was
analyzed by neutron diffraction and successfully elucidated
(Ohgo et al., 1997, part I of this series; Ohhara et al., 1998).
Recently, a novel solid-state reaction, 3±1 photoisomeriza-
tion, of the 3-cyanopropyl group bonded to the Co atom in
Figure 1
Structural change in the 3-cyanopropyl group on exposure to a xenon
lamp.
impossible to distinguish the two mechanisms only by the
X-ray crystal structure analyses before and after the reaction.
In order to clarify the mechanism, we intended to analyze
the structures before and after reaction by neutron diffraction
in addition to X-ray diffraction, since the hydrogen migrates to
the neighboring C atoms during the process of the reaction.
Although the single-crystal neutron diffraction technique is a
powerful method to analyze the mechanism of hydrogen
migration, several dif®culties should be overcome. In the ®rst
place, a considerably large (ꢁ8 mm3) single crystal is neces-
sary for neutron diffraction. Next, the visible light should
penetrate into such a large crystal and the photoreaction
should occur inside the crystal without deteriorating the
crystallinity. More than 20 complexes were prepared changing
the axial base ligand and the crystals were exposed to the
xenon lamp. It was found after many trials that the crystals of
the complex with (R)-1-phenylethylamine as an axial base
ligand were large enough for neutron diffraction and that the
photoisomerization proceeded ꢁ90% without deteriorating
the crystallinity in a large enough crystal for neutron diffrac-
tion.
In order to obtain the precise intensity data by neutron
diffraction, however, the number of H atoms should be
decreased since the H-atom incoherent scattering increases
the background of the Bragg re¯ection. For the analysis of the
disordered structure, data collection with precise intensity
measurements is essential. All the H atoms of the equatorial
ligands, bis(dimethylglyoximato) moiety, and the axial base
ligand, (R)-1-phenylethylamine, were replaced with D atoms
(III). For the 3-cyanopropyl group, two H atoms bonded to the
some cobaloxime complex crystals was found by Ohgo and co-
workers [Kurashima et al., 1995; (I)]. When (R)-1-phenyl-
ethylamine was used as an axial base ligand (1a), the
isomerization proceeded with retention of the single-crystal
form (Sekine, Yoshiike et al., 1998). The structures which were
analyzed by X-rays before and after irradiation (Fig. 1)
suggest two possible mechanisms. One is the two-step
isomerization as shown in (IIa). This scheme may require a
large amount of atomic motion of the 3-cyanopropyl group in
the crystal lattice and it seemed dif®cult to retain the single-
crystal form. Another mechanism is the direct translation of
the cyano group (Bury et al., 1982), as shown in (IIb). It is
ꢀ
246 Ohhara et al. Deuterium migration
Acta Cryst. (2000). B56, 245±253

research papers
C1 atom neighboring the cyano group were replaced with D
atoms. At least one of the D atoms should migrate to the other
C atoms, since the C1 atom of the 3-cyanopropyl group bonds
with the Co atom after irradiation. This paper reports the
structural change during photoisomerization and proposes the
mechanism of the isomerization.
2.1.2. Dimethylglyoxime-d7 (2). Compound (2) was
prepared according to the reported method (Arduengo III et
al., 1994; Zav'yalov & Ezhova, 1984), as shown in (IVa). The
methyl H atoms of butane-2,3-dione were exchanged with
deuterium using D2O and D2SO4 catalysts for seven cycles.
For each cycle the amounts of D2O and D2SO4 were adjusted,
depending on the amount of butanedione used in the cycle.
For the ®rst cycle, butane-2,3-dione (100 g), D2O (200 ml) and
D2SO4 (2 ml) were heated at 368 K for 12 h. The butane-2,3-
dione was isolated by distillation under atmospheric pressure
at 351±353 K and was separated from D2O, which codistilled.
The butane-2,3-dione thus isolated was used without further
puri®cation in the next cycle. After the ®nal cycle the butane-
2,3-dione-d6 produced was dried over anhydrous magnesium
sulfate. The yield after seven cycles was 19.5 g of 99.1 atom
%D butane-2,3-dione-d6. The butane-2,3-dione-d6 (17.67 g)
prepared was re¯uxed with HClÁH2NOH (142.1 g), KOH
(85% pure, 93.7 g) and D2O (400 ml) at 368±373 K for 90 min.
The white precipitate of (2) was ®ltered and washed with
NaHCO3(aq.) and water. The yield was 18.4 g of 98.5 atom
%D.
2.1.3. Acetophenone-d8 (3). Compound (3) was prepared
according to the usual Friedel±Crafts reaction (Mross &
Zundel, 1968) between benzene-d6 (Aldrich, 99.6 atom %D)
and acetyl chloride-d3 (Aldrich, >99 atom %D).
2.1.4. Acetophenone oxime-d8 (4). To a solution of
hydroxylamine hydrochloride (39.66 g, 0.57 mol) and potas-
sium hydroxide (25.7 g, 0.46 mol) in D2O (190 ml) was added
acetophenone-d8 (14.13 g, 0.11 mol). The mixture was heated
at 363±373 K for 3.5 h and then cooled to room temperature.
The aqueous solution was
2. Experimental
2.1. Preparation of complexes
2.1.1.
(3-Cyanopropyl-d2a,a)[(R)-1-phenylethylamine-
d11]cobaloxime-d14 (1b). Compound (1b) was prepared
from CoCl2Á6H2O, dimethylglyoxime-d6 (2), (R)-1-phenyl-
ꢀ,ꢀ
ethylamine-d11 (3) and 3-iodobutylonitrile-d2 (4) using a
method similar to that reported by Ohgo & Takeuchi (1985).
To the mixture of CoCl2Á6H2O (3.57 g, 15 mmol) and (2)
(3.66 g, 30 mmol) dissolved in 30 ml of methanol was added
NaOH (1.2 g, 15 mmol) dissolved in 5 ml of water under an Ar
atmosphere with stirring. Pyridine (1.21 ml, 15 mmol), (4)
(2.95 g, 15 mmol) and 0.57 g of NaBH4 dissolved in 5 ml of
water were then added successively, and the mixture was
stirred for 20 min. The addition of 100 ml of water and
®ltration
afforded
3.15 g
of
(3-cyanopropyl-
d2ꢀ,ꢀ)(pyridine)cobaloxime-d14. This cobaloxime complex was
then dissolved in 30 ml of MeOD and the mixture was stirred
with 10 g of Dowex 50-80X (50±100 mesh, H form) and 5 ml of
D2O overnight. To the ®ltrate was added equimolar (3) and
the mixture was evaporated to afford powder of (1b) (3.34 g,
yield 95.0%). The starting materials, (2), (3) and (4), were
prepared as shown in (IV). The details are given below.
extracted with dichloromethane.
The organic layer was washed
with aqueous sodium bicarbo-
nate and then with water. The
organic layer was then dried
over Na2SO4 and the solvent
was removed to yield aceto-
phenone oxime-d8 as a white
solid (14.92 g, 95%).
2.1.5. 1-Phenylethylamine-d9
(5). Compound (5) was prepared
according to the method
reported (Kano et al., 1980),
under an Ar atmosphere. To an
ice-cooled stirred mixture of
titanium(IV) chloride (13 ml,
0.12 mol), sodium borodeu-
teride (9.90 g, 0.24 mol; Aldrich,
98 atom %D) and anhydrous
1,2-dimethoxyethane (220 ml)
was added dropwise a solution
of acetophenone oxime-d8
(8.00 g, 0.05586 mol) in anhy-
drous
1,2-dimethoxyethane
(50 ml). The mixture was stirred
for 14 h at room temperature
ꢀ
Acta Cryst. (2000). B56, 245±253
Ohhara et al. Deuterium migration 247

research papers
Table 1
Experimental details.
X-ray: initial structure
X-ray: irradiated structure
Neutron: irradiated structure
Crystal data
Chemical formula
[Co(C4D7N2O2)2-
(C4H4D2N)(C8D11N)]
504
Monoclinic
P21
9.6772 (1)
13.8988 (2)
8.9432 (1)
96.322 (1)
1195.56 (3)
2
[Co(C4D7N2O2)2-
(C4H4D2N)(C8D11N)]
504
Monoclinic
P21
9.1579 (1)
13.9694 (2)
9.2757 (1)
99.954 (1)
1168.78 (2)
2
[Co(C4D7N2O2)2-
(C4H4D2N)(C8D11N)]
504
Monoclinic
P21
9.1579 (1)
13.9694 (2)
9.2757 (1)
99.954 (1)
1168.78 (2)
2
1.359
Neutron
1.060
Chemical formula weight
Cell setting
Space group
Ê
a (A)
Ê
b (A)
Ê
c (A)
ꢁ (?)
3
Ê
V (A )
Z
Dx (Mg m
3
)
Radiation type
1.329
1.359
Mo Kꢀ
0.71073
512
Mo Kꢀ
0.71073
512
Ê
Wavelength (A)
No. of re¯ections for cell para-
meters
±
ꢂ range (ꢂ)
3±30
0.754
504
293 (2)
Plate
2±30
0.771
504
293 (2)
Plate
±
0.132
504
293 (2)
Plate
1
)
ꢃ (mm
F(000)
Temperature (K)
Crystal form
Crystal size (mm)
Crystal color
0.4 Â 0.3 Â 0.15
0.5 Â 0.3 Â 0.1
3.5 Â 3.5 Â 0.7
Red
Red
Red
Data collection
Diffractometer
Data collection method
Absorption correction
Siemens SMART CCD
! scan
Empirical (SADABS; (Sheldrick,
Siemens SMART CCD
! scan
Empirical (SADABS; (Sheldrick,
BIX-I
! scan
None
1996)
0.696
0.801
14 136
7077
6573
I > 2ꢄ(I)
0.0146
30.52
13 ! h ! 13
19 ! k ! 19
12 ! l ! 12
1996)
0.461
0.753
10 177
6655
5273
I > 2ꢄ(I)
0.0407
30.75
12 ! h ! 13
19 ! k ! 18
13 ! l ! 7
Tmin
Tmax
±
±
2540
1509
1497
I > 2ꢄ(I)
0.0676
43.37
No. of measured re¯ections
No. of independent re¯ections
No. of observed re¯ections
Criterion for observed re¯ections
Rint
ꢂmax
Range of h, k, l
11 ! h ! 9
14 ! k ! 1
11 ! l ! 10
Re®nement
Re®nement on
F2
F2
F2
R(F) [F2 > 2ꢄ(F2)]
wR(F2) [all re¯ections]
S
0.0262
0.0701
1.007
7077
0.0471
0.1177
1.072
6655
0.1065
0.2853
1.107
1509
No. of re¯ections used in re®ne-
ment
No. of parameters used
H-atom treatment
341
Mixture of independent and
constrained re®nement
281
Mixture of independent and riding
model
422
All H- and D-atom parameters
re®ned
w = 1/[ꢄ2(Fo + (0.0505P)2],
w = 1/[ꢄ2(Fo + (0.0501P)2],
w = 1/[ꢄ2(Fo + (0.2366P)2 +
2
2
2
Weighting scheme
where P = (Fo2 + 2Fc2)/3
where P = (Fo2 + 2Fc2)/3
0.144P], where P =
(Fo2 + 2Fc2)/3
0.024
0.146
0.109
(Á/ꢄ)max
Áꢅmax (e A )
Áꢅmin (e A )
Extinction method
Source of atomic scattering factors
0.001
0.273
0.240
0.004
0.842
1.189
None
-3
Ê
Ê
-3
None
None
International Tables for Crystallo-
graphy (1992, Vol. C, Tables
4.2.6.8 and 6.1.1.4)
International Tables for Crystallo-
graphy (1992, Vol. C, Tables
4.2.6.8 and 6.1.1.4)
International Tables for Crystallo-
graphy (1992, Vol. C, Tables
4.2.6.8 and 6.1.1.4)
Computer programs
Data collection
Cell re®nement
Data reduction
Structure solution
SMART (Siemens, 1995)
SMART (Siemens, 1995)
SAINT (Siemens, 1995)
SIR97 (Altomare et al., 1999)
SMART (Siemens, 1995)
SMART (Siemens, 1995)
SAINT (Siemens, 1995)
SIR97 (Altomare et al., 1999)
Local program
Local program
Local program
Initial structure determined
by X-rays
SHELXL93 (Sheldrick, 1993)
Structure re®nement
SHELXL97 (Sheldrick, 1997)
SHELXL97 (Sheldrick, 1997)
ꢀ
248 Ohhara et al. Deuterium migration
Acta Cryst. (2000). B56, 245±253

research papers
ꢀ,ꢀ
sure (354 K, 5 mmHg) to yield 4-iodobutyronitrile-d2 as a
pale yellow liquid (18.96 g, 76%).
and then the reaction was quenched by the addition of water
with ice-cooling. After basi®cation with 28% aqueous
ammonia, the reaction mixture was ®ltered through Celite
with suction and washed with ether. After removal of the
organic solvent from the ®ltrate, dichloromethane was added
to the residue. The organic layer was separated and the water
layer was extracted with dichloromethane. The combined
extract was washed with saturated aqueous sodium chloride
and dried over Na2SO4. Removal of the solvent yielded 1-
phenylethylamine-d9 as a colorless liquid (6.32 g, 87%).
2.1.6. (R)-1-Phenylethylamine-d9 (6). Optical resolution of
the 1-phenylethylamine-d9 was carried out according to the
published procedure (Theilacker & Winkler, 1954).
2.2. Recrystallization of (1b) and exposure to visible light
Compound (1b) (1.4 g) was solved in a mixed solution of
MeOD (19 ml) and D2O (30 ml) and left for 2 weeks at room
temperature. Then large crystals (ꢁ3.5 Â 3.5 Â 0.7 mm), large
enough for single-crystal neutron diffraction, were obtained.
Some of the large crystals with analogous size and morphology
were irradiated with a Xe lamp (USHIO SUPER BRIGHT
152S) for 7 d.
2.1.7. 4-Chlorobutyronitrile-2,2-d2 (7). To a mixture of 4-
chlorobutyronitrile (20.14 g, 0.19 mol) and D2O (120 ml;
Aldrich, 99.9 atom %D) was added a D2O solution of NaOD
(40 wt %, 2.0 ml; Aldrich, 99.9 atom %D). The mixture was
vigorously stirred under re¯ux for 2 h and then cooled to room
temperature. The aqueous mixture was extracted with
dichloromethane. The extract was concentrated and to the
residue was added D2O (120 ml) and a D2O solution of NaOD
(40 wt. %, 2.0 ml). The mixture was vigorously stirred under
re¯ux for 2 h and then cooled to room temperature. The
aqueous mixture was extracted with dichloromethane.
Removal of the solvent yielded 4-chlorobutyronitrile-2,2-d2
2.3. Determination of the crystal structure by X-rays before
and after irradiation
A crystal of (1b) before irradiation was mounted on a
Siemens Smart CCD diffractometer and the three-dimen-
sional intensity data were collected using monochromated Mo
Kꢀ radiation. A small size crystal cut from the irradiated large
crystal was mounted on a Siemens Smart CCD diffractometer.
The intensity measurement was carried out in the same way as
that before irradiation. The crystal structures before and after
irradiation were re®ned with the program SHELXL97
(Sheldrick, 1997). The crystal data and measurement condi-
tions are shown in Table 1.1
(>99 atom %D) as a yellow liquid (13.39 g, 65%).
a,a
2.1.8. 4-Iodobutyronitrile-d2
(8). Compound (8) was
prepared according to the reported method (Edwards III &
Glenn, 1976). To a powder of sodium iodide (38.0 g, 0.25 mol)
was added acetone (115 ml) at 273±278 K with stirring. To the
2.4. Single-crystal neutron diffraction measurement
ꢀ,ꢀ
One of the irradiated crystals, 3.5 Â 3.5 Â 0.7 mm, was ®xed
on an aluminium pin by halocarbon grease (MOLYKOTE HP-
300 grease) and the pin was mounted on the BIX-I neutron
diffractometer set up at the JRR-3M reactor at the Japanese
Atomic Energy Research Institute (JAERI). The diffract-
ometer has a three-axes goniometer and two conventional gas-
®lled proportional area detectors which are two-dimensionally
movable (active area 25 Â 25 cm; pixel size 2 Â 2 mm; solid
angle 24ꢂ in one side; Niimura, Tanaka, Minezaki et al., 1995).
The neutron beam, monochromated by a bent Si perfect
crystal (Niimura, Tanaka, Karasawa & Minakawa, 1995), has a
mixture was added 4-chlorobutyronitrile-d2
(13.39 g,
0.13 mol) at room temperature. The mixture was heated under
re¯ux for 48 h and then cooled to room temperature. After
removal of the solvent, dichloromethane and water were
added to the residue. The organic layer was separated and the
water layer was extracted with dichloromethane. The
combined organic layer was dried over MgSO4 and the solvent
was removed. The residue was distilled under reduced pres-
Ê
wavelength of 1.06 A. The intensity data were collected at 16
area-detector positions by a ! step-scan method (0.2ꢂ per
step) at room temperature. The maximum and minimum d-
Ê
spacing values were 9.12 and 0.77 A, respectively. A total of
1509 independent re¯ections were observed. Cell parameters
were ®xed to the values determined by X-rays and the U
matrix was re®ned for each area-detector position by using
some intense re¯ections [I > 10ꢄ(I)]. The re¯ection peaks
were picked up with the threshold I > 5ꢄ(I) by the PKSK
program (local program for BIX-I) and were integrated by the
INTX program (local program). The Lorentz correction was
carried out and then the data were modi®ed to SHELXL
format.
1 Supplementary data for this paper are available from the IUCr electronic
archives (Reference: SE0292). Services for accessing these data are described
at the back of the journal.
Figure 2
Crystal structure of (1b) before irradiation, viewed along the a axis.
ꢀ
Acta Cryst. (2000). B56, 245±253
Ohhara et al. Deuterium migration 249

research papers
2.5. Structure refinement by neutron diffraction data
imperfect H±D exchange cancelled the positive or negative
peaks. The details of the re®nement are given in Table 1.
Re®nement of the structure was carried out with the
program SHELXL93 (Sheldrick, 1993). The positional para-
meters of non-H and non-D atoms, except those of the original
3-cyanopropyl and 1-cyanopropyl groups produced, were ®xed
to those obtained by X-ray analysis. The positional parameters
of the H and D atoms of the 3-cyanopropyl and 1-cyanopropyl
groups, the displacement parameters of all the atoms and the
occupancy factors of the disordered groups were re®ned. The
displacement parameters of the ordered atoms were re®ned
anisotropically and those for the disordered atoms were
performed isotropically. The D atoms between OÁ Á ÁO in the
cobaloxime plane were not assigned probably because the
3. Description of the structures
3.1. X-ray analyses before and after irradiation
Fig. 2 shows the crystal structure viewed along the a axis
before irradiation, which is determined by X-rays. The mole-
cules are arranged as a ribbon along the 21 axis (b axis). The
molecular structure with the numbering of the atoms is shown
in Fig. 3. The crystal was irradiated with the xenon lamp for
7 d. The cell dimensions changed with retention of the single-
crystal form, as given in Table 1. The crystal structure deter-
mined by X-rays is shown in Fig. 4. Although the structures of
the equatorial ligands and axial base ligand are approximately
the same as those before irradiation, most of the 3-cyano-
propyl groups were transformed to the (S)-1-cyanopropyl
group. The other residual peaks were assigned to the (R)-1-
cyanopropyl group and the original 3-cyanopropyl group. The
occupancy factors of (S)- and (R)-1-cyanopropyl and 3-
cyanopropyl groups are 0.83, 0.06, and 0.11, respectively. The
molecular structure with the (S)-1-cyanopropyl group is
shown in Fig. 5. The structures of the minor parts, (R)-1-
cyanopropyl and 3-cyanopropyl groups, are shown in Figs. 6(a)
and (b), respectively.
3.2. Structure after irradiation by neutron diffraction
Fig. 7 shows the major part of the molecule structure with
the (S)-1-cyanopropyl group. The (S)-1-cyanopropyl group
produced has a D atom bonded to the C1 atom, D17, and
another D atom bonded to the C2 atom, D19A and D19B,
disordered. The ratio of D/H at D17 is 100%, whereas the D/H
ratios at D19A and D19B are 0.5. In addition, since the
occupancy factors of all D atoms of the axial amine are almost
Figure 3
Molecular structure of (1b) before irradiation with atomic numbering.
The thermal ellipsoids are drawn at the 50% probability level.
Figure 4
Crystal structure of (1b) after 7 d irradiation, viewed along the a axis.
Only the (S)-1-cyanopropyl group among the disordered groups is shown
for clarity.
Figure 5
Molecular structure of the (S)-1-cyanopropyl group produced, with atom
numbering. The thermal ellipsoids are drawn at the 50% probability level.
ꢀ
250 Ohhara et al. Deuterium migration
Acta Cryst. (2000). B56, 245±253

research papers
100%, there is no intermolecular H/D exchange between the
3-cyanopropyl group and the neighboring axial amine ligand.
These results clearly indicate that the reaction proceeds in two
after reaction occupies a space similar to that before the
reaction. In the original 3-cyanopropyl group observed as a
minor part, no exchange between H and D atoms was
observed.
4. Discussion
4.1. Intermediate structure by molecular mechanics
Neutron diffraction revealed that the isomerization
proceeds in two steps. It seems important to estimate the
structure of the intermediate 2-cyanopropyl group, which has
S and R con®gurations. Since the 2-cyanopropyl group is
produced in the original crystalline lattice by photoirradiation,
the 2-cyanopropyl group with a more stable con®guration
would be produced. In order to estimate the energy of the 2-
cyanopropyl group produced in the original lattice, a mini
crystal lattice model was applied (Zimmerman & Zhu, 1995;
Zimmerman & Sebek, 1997). A mini crystal lattice composed
of the 36 original 3-cyanopropyl complexes was computa-
tionally generated by referring to the known crystal structure
of the 3-cyanopropyl complex. Then, a central complex was
replaced with a (R)- or (S)-2-cyanopropyl complex. The
energy-minimized calculation was performed with the Cerius2
program (Molecular Simulations Inc., 1994). A universal force
®eld (Rappe et al., 1992; Castonguay & Rappe, 1992; Rappe &
Colwell, 1993) was adopted. Although the positions of all the
atoms belonging to the 2-cyanopropyl group of the central
molecule and all the H atoms of the surrounding molecules
were re®ned (Buttar et al., 1998), the other atoms were ®xed.
The calculated energy for the (R)-2-cyanopropyl complex is
4509.10 Â 103 kJ mol 1, whereas that for the (S)-2-cyano-
propyl complex is 4548.72 Â 103 kJ mol 1. The (R)-2-cyano-
steps, 3±2 and 2±1, as shown in (V). Moreover, the second 2±1
step should be irreversible, because the D17 atom would be
partly exchanged with the H atom, H19A or H19B if the
second step were reversible. Since Ohgo et al. suggested from
the NMR measurement that the ®rst step is reversible
(Kurashima et al., 1995), the 1-cyanopropyl group is more
stable than the 2-cyanopropyl and 3-cyanopropyl groups, i.e. 1-
cyanopropyl >> 2-cyanopropyl > 3-cyanopropyl. Although the
atomic motion of the cyanopropyl group seems to be consid-
erably large, the crystallinity may be kept since the molecule
1
propyl complex is more stable by 39.62 Â 103 kJ mol than
Figure 6
(a) The disordered structures of (S)- and (R)-1-cyanopropyl groups, with
atom numbering. The atoms labeled B belong to the minor (R)-1-
cyanopropyl group. (b) The disordered structure of (S)-1-cyanopropyl
and 3-cyanopropyl groups, with atom numbering. The atoms labeled C
belong to the original 3-cyanopropyl group.
Figure 7
Molecular structure of (1b) analyzed by neutron diffraction. The thermal
ellipsoids are drawn at the 30% probability level. The unshaded thermal
ellipsoids are non-H or -D atoms.
ꢀ
Acta Cryst. (2000). B56, 245±253
Ohhara et al. Deuterium migration 251

research papers
the (S)-2-cyanopropyl complex. As shown in (VI), the (R)-2-
cyanopropyl complex is estimated to be produced in the
intermediate step.
When the central molecule was replaced with the original 3-
cyanopropyl complex or the complex product (S)-1-cyano-
propyl, the calculated energies are smaller by 28.03 kJ mol
1
and larger by 14.23 kJ mol 1 for the 3-cyanopropyl and (S)-1-
cyanopropyl complexes, respectively, than that for the (R)-2-
cyanopropyl complex. Since the 1-cyanopropyl group is more
stable electronically than the 2-cyanopropyl group, as
described in x3.2, this may explain the reason why the inter-
mediate 2-cyanopropyl group is undetectable during the
reaction.
4.2. Olefin formation
The above discussion is also supported by the cavity shape.
Fig. 8 shows the reaction cavity for the 3-cyanopropyl group in
the original crystal structure. The 2-cyanopropyl group
produced with R or S con®guration with minimum energy is
drawn in the cavity. The 2-cyanopropyl group with R con®g-
uration is better accommodated in the cavity than its enan-
tiomer.
It seems adequate to assume that the absolute con®guration
of the C2 atom of the (S)-1-cyanopropyl complex produced is
R, as shown in (VI), since the Co atom would be replaced with
the D atom. However, the con®guration of the C2 atom of the
analyzed structure by neutron diffraction is racemic.
This may be explained by the mechanism of ole®n forma-
tion. As shown in (VII), the CoÐC bond would be cleaved by
continuous exposure to the xenon lamp and the ole®n and
cobalt hydride, {CD(CN) = CHCH3} + CoÐD or {CD(CN) =
CDCH3} + CoÐH, would be produced. Then the ole®n and
cobalt hydride may return to the 1-cyanopropyl cobalt. In
these equilibrium steps, the H and D atoms are easily
exchanged with each other. It is proposed that the disordered
structure at the C2 atom is caused by the equilibrium steps, as
shown in (VII), after the 1-cyanopropyl group was produced.
4.3. Asymmetric induction
In the 2±1 isomerization, most of the 2-cyanopropyl groups
were transformed to the (S)-1-cyanopropyl group. Fig. 9 shows
the reaction cavity for the 3-cyanopropyl group in the original
crystal. The 1-cyanopropyl groups produced with S and R
con®gurations, which are the same as those in Fig. 6(a), are
also shown. The S con®guration is better accommodated than
the R con®guration. Fig. 10 shows the cavity for the 1-cyano-
propyl group in the crystal after irradiation. The shape of the
cavity is very similar to that before irradiation and the
1-cyanopropyl group with S con®guration is suited to this
cavity. Fig. 10 explains why the crystallinity is kept in the
isomerization and why the S con®guration is preferably
produced.
5. Summary
From the single-crystal neutron diffraction analysis of the
photoirradiated crystal of (1b), the following mechanism has
been established:
(i) the photoisomerization of the 3-cyanopropyl group
occurs in the two-step, 3±2 and 2±1 isomerization,
(ii) the 2±1 isomerization is irreversible,
Figure 8
Reaction cavity for the 3-cyanopropyl group in the original crystal
accommodating (a) the (R)-2-cyanopropyl group and (b) the (S)-2-
cyanopropyl group. The R isomer ®ts the cavity better.
ꢀ
252 Ohhara et al. Deuterium migration
Acta Cryst. (2000). B56, 245±253

research papers
References
Altomare, A., Burla, M. C., Camalli, M., Cascarano, G. L.,
Giacovazzo, C., Guagliardi, A., Moliterni, A. G. G., Polidori, G.
& Spagna, R. (1999). J. Appl. Cryst. 32, 115±119.
Arduengo III, A. J., Dias, H. V. R., Dixon, D. A., Harlow, R. L.,
Klooster, W. T. & Koetzle, T. F. (1994). J. Am. Chem. Soc. 116,
6812±6822.
Bury, A., Bougeard, P., Corker, S. J., Johnson, M. D. & Perlmann, M.
(1982). J. Chem. Soc. Perkin Trans. 2, pp. 1367±1372.
Buttar, D., Charlton, M. H., Docherty, R. & Starbuck, J. (1998). J.
Chem. Soc. Perkin Trans. 2, pp. 763±772.
Castonguay, L. A. & Rappe, A. K. (1992). J. Am. Chem. Soc. 114,
5832±5842.
Edwards III, W. B. & Glenn, D. F. (1976). J. Label. Compd
Radiopharm. 12, 145±151.
Kano, S., Tanaka, Y., Sugino, E. & Hibino, S. (1980). Synthesis, pp.
695±697.
Koester, L. (1977). Springer Tracts in Modern Physics, Vol. 80,
Neutron Physics. New York: Springer-Verlag.
Kurashima, F., Takatsu, N., Ishida, I., Arai, Y., Takeuchi, S. & Ohgo, Y.
(1995). Annual Meeting of the Chemical Society of Japan, Kyoto,
Abstract II, p. 290.
Molecular Simulations Inc. (1994). Cerius2. Release 1.6. Burlington,
Massachusetts.
Mross, W. D. & Zundel, G. (1968). Chem. Ber. 101, 2865±2869.
Niimura, N., Tanaka, I., Karasawa, Y. & Minakawa, N. (1995). Physica
B, 213, 929±928.
Niimura, N., Tanaka, I., Minezaki, Y., Karasawa, Y., Tanaka, I., Miki,
K., Sato, M., Hidaka, M., Minakawa, N. & Morii, Y. (1995). Physica
B, 213, 786±789.
Figure 9
Stereoviews of the reaction cavity for the 3-cyanopropyl group in the
original crystal accommodating (a) the (S)-1-cyanopropyl group
produced and (b) the (R)-1-cyanopropyl group.
Ohashi, Y. (1988). Acc. Chem. Res. 21, 268±274.
Ohashi, Y., Sakai, Y., Sekine, A., Arai, Y., Ohgo, Y., Kamiya, N. &
Iwasaki, H. (1995). Bull. Chem. Soc. Jpn, 68, 2517±2525.
Ohashi, Y. & Sasada, Y. (1977). Nature, 267, 142±144.
Ohashi, Y., Yanagi, K., Kurihara, T., Sasada, Y. & Ohgo, Y. (1981). J.
Am. Chem. Soc. 103, 5805±5821.
Ohgo, Y., Ohashi, Y., Klooster, W. T. & Koetzle, T. F. (1997).
Enantiomer, 2, 241±248.
Ohgo, Y. & Takeuchi, S. (1985). J. Chem. Soc. Chem. Commun. pp.
21±22.
Ohhara, T., Uekusa, H., Ohashi, Y., Tanaka, I., Kumazawa, S. &
Niimura, N. (1998). Chem. Lett. pp. 365±366.
Rappe, A. K., Casewit, C. J., Colwell, K. S., Goddard-III, W. A. &
Skiff, W. M. (1992). J. Am. Chem. Soc. 114, 10024±10035.
Rappe, A. K. & Colwell, K. S. (1993). Inorg. Chem. 32, 3438±3450.
Sato, H. & Ohashi, Y. (1999). Bull. Chem. Soc. Jpn, 72, 367±375.
Sekine, A., Saitoh, M., Hashizume, D., Uekusa, H., Ohashi, Y., Arai,
Y. & Ohgo, Y. (1998). Enantiomer, 3, 159±168.
Sekine, A., Tatsuki, H. & Ohashi, Y. (1997a). J. Organomet. Chem. pp.
536±537.
Figure 10
Reaction cavity for the 1-cyanopropyl group after irradiation accom-
modating the (S)-1-cyanopropyl group.
Sekine, A., Tatsuki, H. & Ohashi, Y. (1997b). J. Organomet. Chem. pp.
389±398.
(iii) the equilibrium between the 1-cyanopropyl cobalt and
the propionitrile + cobalt hydride causes the disordered
structure at the C2 atom after the 1-cyanopropyl group is
produced, and
(iv) the chirality of the 1-cyanopropyl group depends on the
asymmetric environment surrounding the cyanopropyl group
before irradiation.
The single-crystal neutron diffraction provides important
information on the reaction mechanism. We believe that this
method will play an important role in the ®eld of crystal
chemistry.
Sekine, A., Yoshiike, M., Ohashi, Y., Ishida, K., Arai, Y. & Ohgo, Y.
(1998). Mol. Cryst. Liq. Cryst. 313, 321±326.
È
Sheldrick, G. M. (1993). SHELXL93. University of Gottingen,
Germany.
Sheldrick, G. M. (1996). SADABS. University of Gottingen, Germany.
È
Sheldrick, G. M. (1997). SHELXL97. University of Gottingen,
Germany.
È
Siemens (1995). SMART and SAINT. Siemens Analytical X-ray
Instruments Inc., Madison, Wisconsin, USA.
Theilacker, W. & Winkler, H. G. (1954). Chem. Ber. 87, 690±691.
Zav'yalov, S. I. & Ezhova, G. I. (1984). Bull. Acad. Sci. USSR Div.
Chem. Sci. p. 142.
Zimmerman, H. E. & Sebek, P. (1997). J. Am. Chem. Soc. 119, 3677±
3690.
This work was partly supported by CREST from Japan
Science and Technology Corporation (JST).
Zimmerman, H. E. & Zhu, Z. (1995). J. Am. Chem. Soc. 117, 5245±
5262.
ꢀ
Acta Cryst. (2000). B56, 245±253
Ohhara et al. Deuterium migration 253