Ridge-tile-like chiral topology: Synthesis, resolution, and complete chiroptical characterization of enantiomers of edge-sharing binuclear square planar complexes of Ni(II) bearing achiral ligands

Article abstract of DOI:10.1021/ja103296g

Binuclear square planar Ni(II) complexes are described, formed by two tridentate ligands with two imine-nitrogens coordinating two nickel atoms. Such complexes are synthetically readily available with great structural variety and present new types of ridge-tile-like chiral compounds that are reasonably stable in the appropriate bent conformation. Enantiomerically pure samples of these compounds have been obtained for the first time using HPLC with a chiral stationary phase. Absolute configurations and chiroptical properties are fully characterized by ECD, VCD, ORD spectroscopy, and theoretical calculations. These new compounds with ridge-tile-like chiral topology are configurationally reasonably stable [ΔG^? = 121.4 kJ mol^-1, t_1/2 = 14.9 h (78 °C, ethanol)], and therefore their chemistry, physical properties, and applications can be systematically studied.

Full text of DOI:10.1021/ja103296g

Published on Web 07/13/2010
Ridge-Tile-like Chiral Topology: Synthesis, Resolution, and
Complete Chiroptical Characterization of Enantiomers of
Edge-Sharing Binuclear Square Planar Complexes of Ni(II)
Bearing Achiral Ligands
Vadim A. Soloshonok,*,† Taizo Ono,‡ Hisanori Ueki,§ Nicolas Vanthuyne,|
Teodor Silviu Balaban,|,⊥ Jochen Bu¨rck,# Heike Fliegl,⊥ Wim Klopper,⊥
Jean-Vale`re Naubron,| Tam T. T. Bui,∇ Alex F. Drake,∇ and Christian Roussel*,|
Institute of Biorganic Chemistry and Petrochemistry, National Academy of Sciences of the
Ukraine, Murmanska Street 1, KyiV-94 02660, Ukraine, National Institute of AdVanced
Industrial Science and Technology, Nagoya, Japan, National Institute for Materials Sciences,
Tsukuba, Japan, Chirosciences, ISM2, UniVersite´ Paul Ce´zanne Aix-Marseille III, Case A62
AVenue Escadrille Normandie-Niemen, 13397 Marseille CEDEX 20, France, Institut fu¨r
Nanotechnologie and Center for Functional Nanostrutures at the Karlsruhe Institute of
Technology, Karlsruhe, Germany, Institut fu¨r Biologische Grenzfla¨chen, Karlsruhe Institute of
Technology, Karlsruhe, Germany, and Biomolecular Spectroscopy Centre, King’s College
London, The Wolfson Wing, Hodgkin Building, Guy’s Campus, London SE1 1UL, United
Kingdom
Received April 19, 2010; E-mail: christian.roussel@univ-cezanne.fr; vadimsoloshonok@gmail.com
Abstract: Binuclear square planar Ni(II) complexes are described, formed by two tridentate ligands with
two imine-nitrogens coordinating two nickel atoms. Such complexes are synthetically readily available with
great structural variety and present new types of ridge-tile-like chiral compounds that are reasonably stable
in the appropriate “bent” conformation. Enantiomerically pure samples of these compounds have been
obtained for the first time using HPLC with a chiral stationary phase. Absolute configurations and chiroptical
properties are fully characterized by ECD, VCD, ORD spectroscopy, and theoretical calculations. These
new compounds with ridge-tile-like chiral topology are configurationally reasonably stable [∆Gq ) 121.4 kJ
mol-1, t1/2 ) 14.9 h (78 °C, ethanol)], and therefore their chemistry, physical properties, and applications
can be systematically studied.
Introduction
A bent conformation adopted by two adjacent square planar
coordinated M(II) atoms and the appropriate arrangement of
achiral ligands can lead to chiral objects. As illustrated in Figure
1, when the two M(II) metals are the same and the two
connecting ligands X are identical, an inventory based upon
achiral and chiral topologies can be drawn up of the stereo-
chemical situations that may be encountered depending on the
nature of the four remaining achiral ligands (A, B, C, and D).
When A * B * C * D, chiral objects are obtained whatever
the combination or location of the four achiral ligands. In the
case A * B * C ) D ≡ A * B * C2, there are three possible
chiral arrangements. For A ) B * C ) D ≡ A2 * C2, two of
the three possible arrangements are achiral and one is chiral.
For A * B ) C ) D ≡ A * C3, the single arrangement is
chiral. Finally, when the four ligands are identical, the complex
Figure 1. Achiral and chiral topologies resulting from bent edge-sharing
binuclear square planar complexes of a d8 transition metal when both two
metals and the edge ligands are identical and the four remaining ligands A,
B, C, and D are achiral. 1, 2, and 3 refer to the class of compounds
synthesized and described in the present study.
† National Academy of Sciences of the Ukraine.
‡ National Institute of Advanced Industrial Science and Technology.
§ National Institute for Materials Sciences.
| Universite´ Paul Ce´zanne Aix-Marseille III.
⊥ Institut fu¨r Nanotechnologie and Center for Functional Nanostrutures
at the Karlsruhe Institute of Technology.
is achiral. Note the achiral compounds in Figure 1 contain a
mirror plane running through the edge-sharing axis and/or the
two metal centers. If the two metals or the connecting ligands
(X, X) were different, many more possibilities would be
# Karlsruhe Institute of Technology.
∇ King’s College London.
9
10.1021/ja103296g  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 10477–10483 10477

A R T I C L E S
Soloshonok et al.
Scheme 1. Edge-Sharing Square Planar Binuclear Complexes of
Ni(II) Bearing Achiral Ligands: 1, 2, and 3 (Bn ) Benzyl)a
Scheme 2. Preparation of the Racemic Compounds 2 (a), 1 and 3
(b)a
a Compound 4 presents additional chirality from the proline-based chiral
ligands.
obtained, leading to a very attractive but as yet unexplored
chirality domain.
The “ridge-tile-like” or “roof-type” complexes belong to the
known class of edge-sharing binuclear square planar complexes
of d8 transition metal ions.1 However, the chirality of this type
of compound has been virtually ignored because the chiral bent
conformation has been said to be highly unstable,2 undergoing
rapid inversion, similar to the chirality of trisubstituted nitrogen.
To the best of our knowledge, there are no reported examples
of the isolation and chiroptical characterization of the enanti-
omers of this type of chiral compound when the metals bear
achiral ligands.
Here, the first description is reported of the synthesis,
isolation, and full chiroptical characterization of the enantiomers
of two types of edge-sharing binuclear square planar complexes
of nickel(II) with chirality provided solely by a stable bent
conformation and a suitable arrangement of achiral ligands. Two
types of metal complex are considered represented by the
compounds 1, 2, and 3 indicated in Figure 1 and illustrated in
Scheme 1.
a Product proportions are given in parentheses.
in solution in various solvents. Doubt was cast on the stability
of the edge-sharing configuration of the Ni(II) complexes of
this type in general, compounds 1, 2, and 3 in particular, in
which a higher flexibility of the less rigid ligands could be
expected.
In this Article, the existence of chiral topologies resulting
from bent edge-sharing binuclear square planar complexes of a
d8 transition metal has been unambiguous confirmed. The
absolute configurations of the corresponding enantiomerically
pure compounds have been assigned. These results pave the
way to a systematic study of this fascinating and previously
unexplored class of chiral structures.
In a preliminary communication,3 compound 4 (Scheme 1),
which possesses two optically pure (S)-proline derived ligands,
was shown to lead to unequally populated and isolatable
diastereomeric forms. These diastereomers are hereafter called
minor-4 and major-4.
The existence of the diastereomers was the first physical proof
of the chirality introduced by the bent conformation of the
binuclear square planar nickel(II) complex in addition to the
configurationally stable chirality center of the proline ligands.
However, some degree of epimerization of the diastereomeric
derivatives 4 was observed during storage in the solid state or
Results and Discussion
Synthesis. The racemic compound 2 (Scheme 2a) was
produced in a manner similar to the synthesis of the diastere-
omeric derivatives 4.3
(1) (a) Aullo´n, G.; Ujaque, G.; Lledos, A.; Alvarez, S.; Alemany, P. Inorg.
Chem. 1998, 37, 804–813. (b) Aullo´n, G.; Lledos, A.; Alvarez, S.
Inorg. Chem. 2000, 39, 906–916. (c) Aullo´n, G.; Ujaque, G.; Lledos,
A.; Alvarez, S. Chem.-Eur. J. 1999, 5, 1391–1410. (d) Marshall, W. J.;
Aullo´n, G.; Alvarez, S.; Dobbs, K. D.; Grushin, V. V. Eur. J. Inorg.
Chem. 2006, 3340–3345. (e) Mizuta, T.; Aoki, S.; Nakayama, K.;
Miyoshi, K. Inorg. Chem. 1999, 38, 4361–4364. (f) Aullo´n, G.;
Alvarez, S. Inorg. Chem. 2001, 40, 4937–4946. (g) Lasri, J.;
Kopylovich, M. N.; Guedes da Silva, M. F. C.; Charmier, M. A. J.;
Pombeiro, A. J. L. Chem.-Eur. J. 2008, 14, 9312–9322. (h) Azerraf,
C.; Cohen, S.; Gelman, D. Inorg. Chem. 2006, 45, 7010–7017.
(2) For examples of roof shape binuclear Ni complexes with additional
chirality coming from chiral ligand, see: (a) Brunner, H.; Dormeier,
S.; Grau, I.; Zabel, M. Eur. J. Inorg. Chem. 2002, 2603–2613. (b)
Duran, J.; Polo, A.; Real, J.; Benet-Buchholz, J.; Poater, A.; Sola, M.
Eur. J. Inorg. Chem. 2003, 4147–4151.
Exposure of a solution of the n-dibutylamine containing
complex 5 in acetonitrile to the action of strong base resulted
in quantitative formation of the binuclear product 2, which was
additionally purified by column chromatography (Scheme 2-a).
The synthesis of compounds 1 and 3 was achieved by an
interesting scrambling experiment. When a mixture of diben-
zylamine 6 and piperidine 7 containing complexes was submit-
ted to the same reaction conditions, the following three products
were obtained: the corresponding dibenzylamine 1 and piperi-
dine 9 derived dimers as expected, but also the scrambled
product 3 (Scheme 2b). While detailed mechanistic rationale
and the generality of this rare nitrogen reductive dealkylation
(3) Soloshonok, V. A.; Ueki, H. J. Am. Chem. Soc. 2007, 129, 2426–
2427.
9
10478 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Ridge-Tile-like Complexes
A R T I C L E S
mization can be a serious issue; a noticeable epimerization on
storage has already been reported for 4.3
The configurational stability of the enantiomers of 1, 2, and
3 is hopefully related to the strength of the coordination of the
nickel with the nitrogen of the imine and/or the nickel with the
nitrogen of the tertiary amine. The three structures under study
and the diastereomeric pair 4 all include relatively basic tertiary
amines. Therefore, the configurational stability should be high
enough to permit a separation of the corresponding enantiomers
using liquid chromatography with a chiral stationary phase.
Screening of various CSPs (see the Supporting Information)
showed that the (S,S)-Whelk-O1 column from Regis was
sufficiently efficient in separating the enantiomers of compounds
1, 2, and 3 as well as the diastereomers of 4 using a mobile
phase composed of a mixture of hexane and an alcohol (EtOH
or 2-PrOH): hexane/EtOH (1:1 v:v) for 1, hexane/2-PrOH (8:2
v/v) for 2, and pure EtOH for 3. The peaks of the two
enantiomers were consistently sharp with baseline separation
yielding good resolution values: 2.55 for 1, 4.52 for 2, and 3.53
for 3. The chromatograms did not show any plateau, a diagnostic
for exchange between the enantiomers on the column.6 The
configurational stability of the enantiomers at room temperature
was further confirmed during the check of their optical purity
after semipreparative separation and careful evaporation of the
solvent. Online CD detection at 390 nm showed that the first
eluted enantiomers for compounds 1, 2, and 3 all gave a (+)
CD390 sign. The CD390 sign of the first eluted diastereomer (less
stable minor-4) of 4 is (-), indicative of a possible inversion
of elution order between the diastereomers of 4 and the
enantiomers of 1, 2, and 3 on the Whelk column. ECD and
VCD spectrocopies (see below) prove that the first eluted
diastereomer (less stable minor-4) of 4 has an opposite joining-
edge absolute configuration to the first eluted enantiomers in
the series 1, 2, 3. These results show once again that liquid
chromatography with chiral stationary phases is a very powerful
method to separate enantiomers or diastereomers, but they also
show that chiral HPLC cannot be used safely to assign the
absolute configuration based upon elution order alone even on
a given chiral stationary phase.7 Semipreparative separations
Figure 2. Absolute configuration nomenclature and selected enantiomer
structures from the X-ray data of racemic 1, 2, and 3. The dotted lines
were added to highlight the roof-tile shape of the edge-sharing square planar
binuclear complexes of Ni(II).
reaction still await more comprehensive studies, this result
strongly supports the in situ formation of the intermediates 8
and 8′ (Scheme 2b), which can undergo further homo- or
heterodimerization leading to symmetric or scrambled products.
The racemic products 1 and 3 were selected for this study
and purified to analytical purity via column chromatography.
The structures of compounds 1, 2, and 3 were confirmed by
single crystal X-ray analysis.
Resolution of Enantiomers. The X-ray data of compounds 1,
2, and 3 showed that two enantiomers are observed in the solid
state. While the X-ray data provide useful starting geometries
of the enantiomers, which were used for further structure
optimization, they do not shed light on the configurational
stability of this type of chiral compound in solution. Figure 2
defines the absolute configuration nomenclature and illustrates
the enantiomers (P)-1, (P)-2, and (P)-3 as they appeared in the
X-ray data. (Note: In the absence of established rules for the
absolute stereochemistry nomenclature in this type of metal
complex, it was now considered far more convenient and
straightforward to use axial chirality with the chirality axis
represented by the edge shared by the two square planar
complexes. Unfortunately, this is an inverse nomenclature for
the (M) and (P) forms as compared to the previous article, which
was based upon helical chirality for the diastereomers of
compound 4. In the current Article and the Supporting Informa-
tion, the X-ray structures of the assigned forms are reported to
illustrate the absolute configuration without ambiguity.)
(5) For some recent semi-preparative separations of metal complexes on
chiral support, see: (a) Norel, L.; Rudolph, M.; Vanthuyne, N.;
Williams, J. A. G.; Lescop, C.; Roussel, C.; Autschbach, J.; Crassous,
J.; Reau, R. Angew. Chem., Int. Ed. 2010, 49, 99–102. (b) Setsune,
J.-I.; Tsukajima, A.; Okasaki, N.; Lintuluoto, J. M.; Lintuluoto, M.
Angew. Chem., Int. Ed. 2009, 48, 771–775. (c) Lai, R.; Chanon, F.;
Roussel, C.; Sanz, M.; Vanthuyne, N.; Daran, J.-C. J. Organomet.
Chem. 2008, 693, 23–32. (d) Rang, A.; Engeser, M.; Maier, N. M.;
Nieger, M.; Lindner, W.; Schalley, C. A. Chem.-Eur. J. 2008, 14,
3855–3859. (e) Sun, P.; Krishnan, A.; Yadav, A.; Singh, S.; Mac-
Donnell, F. M.; Armstrong, D. W. Inorg. Chem. 2007, 46, 10312–
10320. (f) Warnke, M. M.; Cotton, F. A.; Armstrong, D. W. Chirality
2007, 19, 179–183. (g) Chen, X.; Okamoto, Y.; Yano, T.; Otsuki, J.
J. Sep. Sci. 2007, 30, 713–716. (h) Lassen, P. R.; Guy, L.; Karame,
I.; Roisnel, T.; Vanthuyne, N.; Roussel, C.; Cao, X.; Lombardi, R.;
Crassous, J.; Freedman, T. B.; Nafie, L. A. Inorg. Chem. 2006, 45,
10230–10239. (i) Paisner, S. N.; Bergman, R. G. J. Organomet. Chem.
2001, 621, 242–245. (j) Ciclosi, M.; Lloret, J.; Estevan, F.; Lahuerta,
P.; Sanau, M.; Pe´rez-Prieto, J. Angew. Chem., Int. Ed. 2006, 45, 6742–
6744.
To study the unique chiroptical properties and assign the
absolute configurations of the enantiomers with intact chirality,
they need to be isolated. Provided the racemization of the
isolated enantiomer is limited or nonexistent,4,5 the method of
choice for the enantiomer resolution of compounds 1, 2, and 3
is HPLC with chiral stationary phases (CSP). However, race-
(6) For exchange processes on the chiral column, see: (a) Schurig, V.
J. Chromatogr., A 2009, 1216, 1723–1736. (b) Wolf, C. Chem. Soc.
ReV. 2005, 34, 595–608. (c) Trapp, O. Anal. Chem. 2006, 78, 189–
198. (d) Trapp, O. Chirality 2006, 18, 489–497. (e) Vanthuyne, N.;
Andreoli, F.; Fernandez, S.; Roman, M.; Roussel, C. Lett. Org. Chem.
2005, 2, 433–443. (f) Roussel, C.; Vanthuyne, N.; Bouchekara, M.;
Djafri, A.; Elguero, J.; Alkorta, I. J. Org. Chem. 2008, 73, 403–411.
(7) Review: Roussel, C.; Del Rio, A.; Pierrot-Sanders, J.; Piras, P.;
Vanthuyne, N. J. Chromatogr., A 2004, 1037, 311–3.
(4) Piras, P.; Roussel, C. J. Pharm. Biomed. Anal. 2008, 46, 839–847.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10479

A R T I C L E S
Soloshonok et al.
Figure 3. ECD spectra of the enantiomers of 1-3 (green, first eluted
enantiomers; red, second eluted) and minor-4 first eluted diastereomer (order
of elution on the (S,S)-Whelk column, 2 mg/10 mL dry CH2Cl2).
Figure 4. ORD spectra of the enantiomers of 1-3 in CHCl3: green, first
eluted enantiomers; red, second eluted (order of elution on the (S,S)-Whelk)
column. Data were smoothed with a window factor of 8 for better
presentation and normalized to A390 ) 1.0.
of the enantiomers of compounds 1, 2, and 3 were successfully
performed on the same chiral support, with resulting ee > 97%.
Chiroptical Properties of the Enantiomers of 1, 2, and 3.
Electronic circular dichroism spectra (ECD) of the enantiomers
of 1, 2, and 3 were recorded in dry CH2Cl2. The spectra are
presented in Figure 3 together with the ECD spectra of the
minor-4 diastereomer.
means a very low concentration would be required in the
conventional 1.0 dm path length cell used in polarimetry to
ensure that enough light is transmitted for detection. A positive
CD maximum is observed near 589 nm. This will have an
associated ORD bisignate Cotton effect with a zero (or close
to zero) crossover at ∼589 nm, giving at best a relatively low
optical rotation. Throughout a series of compounds of this type,
the crossover wavelength may vary, and hence sign will be
uncertain. In this series of compounds, a considered wavelength
needs to be chosen for the (+)/(-) nomenclature; λ ) 589 nm
is a bad choice. Reviewing the ORD spectra, we recommend
that enantiomers are labeled according to the sign of their optical
rotation at 365 nm, that is, (+)365 or (-)365, provided the
concentration is extremely low. This should be readily measur-
able on a conventional polarimeter using a mercury lamp.
Pathlengths shorter than 1 dm would be preferred. Even better
would be recording optical rotation at 650 nm and labeling
enantiomers as (+)650 or (-)650. The 650 nm optical rotation is
relatively clear-cut and corresponds to the longest wavelength
transition ORD Cotton effect maximum associated with the 590
nm CD. Unfortunately, this is not a measurement that could be
made with a conventional polarimeter. It is worth noting that
the sign of the optical rotation is opposite at 365 and 650 nm.
The great changes in optical rotation both in intensity and in
sign with wavelength have practical issues when an online
polarimeter is used to monitor chiral HPLC. For example, the
observed sign using a JASCO OR-1590 online polarimeter
detector results from the summation of contributions from
several spectral components across the whole range of wave-
lengths from 350 to 900 nm in the mobile phase.8 This technical
issue has little consequence for the large majority of optically
pure compounds, which present a simple ORD curve. However,
The spectra are very similar showing that the substituent
changes in the series 1, 2, and 3 do not affect the general shape
of the ECD. Moreover, the CD intensities are almost identical
for λ > 350 nm, some minor changes being observed for λ <
350 nm. The ECD spectra of the enantiomers of 1, 2, and 3 are
strikingly similar to those previously reported for the two
isolated optically pure diastereomers of binuclear Ni(II) com-
plexes 4 in which two additional stereogenic centers were
introduced through two N-benzyl-(S)-proline residues.3 The ECD
spectra of the second eluted enantiomers of 1, 2, and 3 are
strikingly similar to those of the first eluted minor diastereomer
of 4 (minor-4) (red traces in Figure 3). The main contributions
to the ECD spectra of the enantiomers of 1-3 and the
diastereomers of 4 derive from the chiral core of the binuclear
complex structure with only relatively minor contributions from
the remote (chiral) substituents attached to the tertiary-amine.
The ORD curves, in the wavelength region 820-220 nm,
for the enantiomers of 1-3 recorded in CHCl3 are reported in
Figure 4. Optical rotation is dependent on enantiomeric purity,
concentration, and path length. To better ensure enantiomer
comparison ensuring any variation is due only to enantiomeric
purity, the ORD curves are reproduced here normalized to A390
) 1.0. A 10 mm cell path length was employed. The ORD
curves of 1, 2, and 3 are very similar, and the spectra of the
enantiomers effectively mirror each other. Again, the substit-
uents on the tertiary amine have a relatively negligible effect
on the chiroptical properties.
Enantiomers are frequently labeled as (+) or (-) based upon
optical rotation data, normally at 589 nm. This is not trivial for
the metal complexes described here. Absorption at 589 nm
(8) Roussel, C.; Vanthuyne, N.; Serradeil-Albalat, M.; Vallejos, J.-C.
J. Chromatogr., A 2003, 995, 79–85.
9
10480 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Ridge-Tile-like Complexes
A R T I C L E S
Figure 6. Display of two optimized conformations of (P)-2, which differ
in the arrangement of the butyl groups.
Figure 5. From top to bottom, experimental VCD spectra of 1-3 (green,
first eluted enantiomers; red, second eluted) and major-4 (second eluted)
(lower TLC spot more stable, green). All spectra were recorded in CD2Cl2
with a concentration of 0.025 mol L-1 (order of elution on the (S,S)-Whelk
column).
in the rare cases of multiple Cotton effects, as in this study, the
optical rotation magnitude may be relatively weak due to sign
cancellation, and the overall sign resulting from the summation
over the whole range of wavelengths could be opposite to the
one measured at 589 nm. The first eluted enantiomer for
compounds 1, 2, and 3 gave a (-) sign according to JASCO
OR-1590 online polarimeter. The observed sign could well be
different using other commercially available online polarimeters,
particularly those based upon single wavelength detection.
At this point of the discussion, according to ECD and ORD
spectra, it can be safely stated that the second eluted enantiomers
of complexes 1, 2, and 3 have the same absolute configuration
at the edge-sharing binuclear Ni(II) square planar stereogenic
element. This configuration is the same as the first eluted
diastereomer of 4 (minor-4).
The VCD spectra of the enantiomers of 1-3 were recorded
in CD2Cl2. Figure 5 shows that like the ECD and ORD, a
consistent set of spectra are obtained across the series 1-3. The
VCD spectra of the major diastereomer of 4 (major-4, second
eluted on the (S,S)-Whelk column) were also recorded and found
to be similar to those obtained for the first eluted enantiomers
of 1-3 (Figure 5, green curves).
As expected, the VCD spectra are more sensitive than ECD
to substituent change on the tertiary amine. However, four bands
situated at 1441, 1473, 1553, and 1591 cm-1 are excellent
configuration markers. These VCD spectral components are all
positive for the first eluted enantiomer in the series 1-3 and
the second eluted enantiomer in major-4, confirming that they
share the same absolute configuration at the binuclear Ni(II)
complex fragment.
Absolute Configuration Determination. The assignment of
absolute configuration was based upon calculated ECD and VCD
spectra as compared to the experimental spectra. Compound 2
was selected for the calculations bearing in mind that relative
configurations across the 1, 2, and 3 series are based upon ECD,
ORD, and VCD spectra. Calculations of the predicted ECD and
VCD spectra were performed on two fully optimized conforma-
tions of (P)-2: A-2 (starting from the X-ray conformation) and
B-2 (C2-symmetry), which differ in the arrangement of the butyl
groups (Figure 6). For the ECD spectra, all calculations were
performed with the TURBOMOLE9 program using density
Figure 7. Comparison of experimental spectra (CH2Cl2) for second eluted
2 and calculated ECD spectra obtained for the optimized A-2 conformation
of (P)-2. The black bars correspond to calculated rotatory strengths.
functional theory (DFT). The B3LYP functional10 was used
together with the def2-SVP basis set.11
The ECD spectra were calculated for the two geometries A-2
and B-2 using time-dependent density functional theory (TD-
DFT).12 For all calculations, fine quadrature grids of size m4
were employed.13 For a comparison between theoretical results
and the experimental ∆ε values, the calculated ECD spectra
have been modeled with Gaussian functions, using a root-mean-
square width of 0.14 eV. It is worth noting that conformations
A-2 and B-2 resulted in almost identical calculated ECD spectra.
A clear assignment of the electronic transitions involved is
difficult because the orbitals are not purely ligand or metal
centered but delocalized over the phenyl ligands, the nitrogens,
and the Ni atoms (see the Supporting Information). Bearing in
mind that the calculations simulate the gas phase and that the
calculated results have not been shifted, Figure 7 shows that
the agreement between calculated and experimental spectra is
relatively good, allowing a safe assignment of the absolute
configuration.
(10) (a) Slater, J. C. Phys. ReV. 1951, 81, 385–390. (b) Vosko, S. H.; Wilk,
L.; Nusair, M. Can. J. Phys. 1980, 58, 1200–1211. (c) Lee, C.; Yang,
W.; Parr, R. G. Phys. ReV. 1988, B37, 785–789. (d) Becke, A. D.
Phys. ReV. 1988, A38, 3098–3100. (e) Becke, A. D. J. Chem. Phys.
1993, 98, 5648–5652. (f) Stephens, P. J.; Devlin, F. J.; Chablowski,
C. F.; Frisch, M. J. J. Phys. Chem. 1994, 98, 11623–11627.
(11) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297–
3305.
(12) (a) Furche, F.; Rappoport, D. Density functional methods for excited
states: equilibrium structure and electronic spectra. In Computational
Photochemistry; Olivucci, M., Ed.; Elsevier: Amsterdam, 2005; Vol.
16, Chapter III. (b) Bauernschmitt, R.; Ahlrichs, R. Chem. Phys. Lett.
1996, 256, 454–464.
(9) Ahlrichs, R.; Ba¨r, M.; Horn, H.; Ko¨lmel, C. Chem. Phys. Lett. 1989,
162, 165–169.
(13) Treutler, O.; Ahlrichs, R. J. Chem. Phys. 1995, 102, 346–354.
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10481

A R T I C L E S
Soloshonok et al.
Figure 9. X-ray structure of the minor (P,S,S)-4 showing the inclusion of
CH2Cl2 in accordance with the experimental protocol.
Figure 8. Comparison of experimental VCD spectra (CD2Cl2) for the
second eluted 2 (in red) and calculated VCD spectra obtained for the
optimized A-2 and B-2 conformation of (P)-2. The bands labeled with “*”
are robust configuration markers (see text and the Supporting Information).
In conclusion, all of the (S,S)-Whelk second eluted enanti-
omers of compounds 1-3 present the same (P) absolute
configuration depicted in Figure 2, and minor-4 ((S,S)-Whelk
first eluted) is (P,S,S).
For the VCD spectra, all calculations were performed using
Gaussian software14 to yield the optimized A-2 and B-2
conformations of (P)-2. Calculated VCDs were obtained using
B3LYP functionals associated with the 6-31G(d,p) basis set for
H, C, N, and O atoms. For Ni atoms, the MDF10 Stuttgart/
Dresden ECP for core electrons and the associated basis set for
valence electrons were used. Frequencies were scaled by a factor
of 0.9627. The calculated VCD were very similar using A-2 or
B-2 optimized geometries (Figure 8). In contrast to the ECD
transitions in which the orbitals of Ni atoms were contributing,
the calculated VCD frequencies in the range 1800-1000 cm-1
generally do not involve the Ni atoms except for the stretching
of the Ni-N-CdO bonds at 1210 cm-1 (observed at 1223
cm-1), which is mixed with the CH bending mode of the
bridging phenyl groups. The four characteristic VCD bands are
derived, respectively, from the antisymmetrical stretching of the
two CdN bonds and the stretching of the CdC bond of the
free phenyl moieties (calculated 1607 cm-1, observed 1591
cm-1), the antisymmetrical stretching of the CdC bond of the
bridging phenyl moieties (calculated 1551 cm-1, observed 1553
cm-1), the stretching of the CdC bond of the bridging phenyl
moieties with the CH2 and CH3 bending of the butyl groups
(calculated 1460 cm-1, observed 1473 cm-1), and the rocking
Figure 10. X-ray structure of the major (M,S,S)-4 showing the inclusion
of Et2O and CHCl3 in accordance with the experimental protocol.
was recorded, the Ni atoms are not directly involved. A good
fit was observed between the experimental and calculated spectra
(Figure 8).
The VCD study indicates that all of the second eluted
enantiomers of compounds 1-3 present the same (P) absolute
configuration depicted in Figure 8 and defined in Figure 2. The
absolute configuration of the major-4 second elution component
is (M,S,S).
The absolute configurations of the diastereomers minor-4 and
major-4 have been assigned by comparison of their ECD and
VCD spectra with those predicted and observed for the pure
enantiomers of 1, 2, and 3. In the preliminary communication,3
the Supporting Information records that the crystal of the minor
diastereomer form of 4 (minor-4) used for X-ray crystallography
was grown by a vapor diffusion method using CH2Cl2 and
MeOH as solvents; the crystal of the major form (major-4) was
grown by a vapor diffusion method using CHCl3 and Et2O as
solvents. The X-ray structures of minor-4 and major-4 are
reported in Figures 9 and 10; they are unambiguously identified
thanks to the cocrystallization of CH2Cl2 with minor-4 (Figure
9) and both CHCl3 and Et2O with major-4 (Figure 10),
respectively. The observed configurations are in perfect agree-
ment with those derived from a comparison of ECD and VCD
spectra.
Racemization. The minor (P,S,S)-4 compound, after isolation,
showed a tendency to give the major (M,S,S)-4 compound on
standing for weeks at room temperature. The qualitative
observation of an epimerization process prompted a quantitative
study of the configurational stability of the isolated enantiomers
of 1, 2, and 3. Ethanol at reflux was chosen as the solvent to
follow the racemization kinetics. The enantiomers of 2 and 3
racemized very cleanly. Data analysis based upon a first-order
of the CH bonds of the phenyl moieties (calculated 1435 cm-1
,
observed 1441 cm-1). These bands are particularly well
reproduced (“*” in Figure 8); they are negative for the second
eluted enantiomers of 1, 2, and 3. Analysis of the origin of these
bands showed that they constitute a signature of the bridging
phenyl groups and the imine ligands independently of the
substituents on the tertiary amine and therefore observed in all
of the enantiomers 1-3, as well as in the corresponding proline
diastereomers 4. In the range of frequencies for which the VCD
(14) , Frisch, M. J.; et al. Gaussian 03, revision E.01; Gaussian, Inc.:
Wallingford, CT, 2004.
9
10482 J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010

Ridge-Tile-like Complexes
A R T I C L E S
chiral topology, are configurationally reasonably stable and
therefore do offer new practical applications of this previously
unstudied source of chirality.
Conclusions
Chromatography with a chiral stationary phase has led to the
first isolation of the enantiomers of two classes of stable “roof
shape” bimetallic chiral complexes constructed around two
nickel atoms. Off-line measurements of ECD and VCD spectra
and associated DFT calculations on complex 2 have enabled
an unambiguous assignment of the absolute configuration of
the isolated enantiomers.8 A remarkable similarity was observed
in the ECD, ORD, and VCD spectra of the enantiomers of 1-3.
The second eluted enantiomers on the (S,S)-Whelk column of
complexes 1-3, which are characterized by a (+) sign using
online JASCO polarimetric detection and a (-) sign using online
CD detection at 390 nm, have a (P) absolute configuration. By
comparison, the (M) absolute configuration is assigned to the
first eluted enantiomers [(S,S)-Whelk column] of 1-3, which
are characterized by (+) CD signs at 390 and 590 nm, (-) sign
using on-line JASCO polarimetric detection and optical rotations
Figure 11. Enantiomerization kinetics of 2 in ethanol.
kinetic model gave a straight line from which the rate constant
of enantiomerization can be extracted15 (Figure 11). The
participation of ethanol as a possible nickel ligand is not seen
in these measurements because the ethanol is in very large
excess, giving rise to pseudo first-order kinetics. The observed
barriers of enantiomerization are almost identical for 2 and 3
(∆Gq ) 121.4 kJ mol-1, t1/2 ) 14.9 h (78 °C, ethanol)).
For compound 1, the estimated barrier is slightly higher than
in 2 and 3.16 The required prolonged heating to achieve
racemization of compound 1 results in extensive parallel
chemical decomposition while refluxing in ethanol. A more
systematic study of the solvent effect and of the role of the pKa
and steric bulk of the liganding amine on the barrier is in
progress, and initial results show that these chiral compounds
are not chemically inert.17 However, the stability data do suggest
that compounds 1, 2, and 3, possessing the new ridge-tile-like
with (-)365 and (+)650
.
Complexes 1-3 present the first examples of chirality due
solely to the bent shape of the binuclear square planar Ni(II)
complexes; there is no additional element of chirality brought
by the ligands that are achiral. This new type of ridge-tile-like
chiral topology has been demonstrated to be configurationally
quite stable, opening a door for systematic studies and practical
applications.
Acknowledgment. We thank the CRCMM and the Spectropoˆle
Marseille for computer time and use of the VCD facility, respec-
tively. The Deutsche Forschungsgemeinschaft is thanked for its
support of the Center for Functional Nanostructures at the Karlsruhe
Institute of Technology, which permitted the collaboration between
the groups of Wim Klopper (theory), Anne Ulrich (CD Spectros-
copy, project E1.2), and Silviu Balaban (project C3.5). A.F.D. and
T.T.T.B. thank the Wellcome Foundation and Applied Photophysics
for their support.
(15) (a) Wolf, C. Racemization, enantiomerization and diastereomerization.
Dynamic Stereochemistry of Chiral Compounds, Principles and
Applications; RCS Publishing: Cambridge, UK, 2008; Chapter III. (b)
Maskill, H. The Physical Basis of Organic Chemistry; Oxford
University Press: Oxford, 1985.
(16) A barrier of 1 (∆Gq ) ca. 124 kJ mol-1) was estimated from three
available experimental points, which were perfectly aligned. This is
not sufficient for a precise determination of the barrier; however, it
provides evidence that the barrier is qualitatively higher in 1 than in
2 and 3.
Supporting Information Available: Synthesis of 1, 2, 3, and
4 (minor and major), chiral chromatography, optical rotation,
ECD, IR, and VCD, Cartesian coordinates of conformations A-2
and B-2, crystallographic information, kinetics of enantiomer-
ization, and complete ref 14. This material is available free of
charge via the Internet at http://pubs.acs.org.
(17) In a very preliminary experiment, the replacement of the tertiary amine
in 1, 2, or 3 by a less basic 2-carboxy-pyridine ligand results in a
much lower barrier to racemization, giving rise to a plateau during
room temperature chromatography on a chiral stationary phase (data
not shown) and precluding isolation at room temperature of the
enantiomers (t1/2 ) 11 min at 15 °C in EtOH). The tridendate nature
of the ligand and the high basicity and steric bulk of the tertiary amine
are the basis for the reasonable configurational stability observed in
1, 2, 3, and 4.
JA103296G
9
J. AM. CHEM. SOC. VOL. 132, NO. 30, 2010 10483