Synchrotron powder diffraction in a systematic study of 4′-[2-(tosylamino)benzylideneamino]-2,3-benzo-15-crown-5 complexes

Article abstract of DOI:10.1107/S0108768107003898

The crystal structures of two compounds, CuL 2 and LiNCS·HL [HL = 4′-[2-(tosylamino)benzylideneamino]-2,3-benzo-15-crown-5], have been determined from synchrotron powder diffraction data. Both compounds crystallize in the monoclinic space group P2 1/c and with one molecule in the asymmetric unit. In CuL 2 the four N atoms of two bidentate L ligands coordinate the Cu^II ion in a distorted tetrahedral geometry with Cu - N distances of 1.98 (5)-2.05 (5) A, while two O atoms from two sulfoxide groups complete the distorted octahedral Cu coordination [Cu - O 2.64 (4), 2.74 (4) A]. In LiNCS·HL, lithium is coordinated by all five ether O atoms with Li - O distances of 2.03 (3)-2.50 (3) A and an N atom from the thiocyanate moiety [Li - N 1.98 (3) A] in a distorted pentagonal pyramidal geometry. Preliminary potentiometric selectivity measurements for ion-selective electrodes (ISEs) based on CuL 2 and ZnL 2 demonstrated significant differences in their selectivity. In order to find a possible reason for this, theoretical calculations at the DFT (B3LYP) level were performed. These calculations used the crystal structures of CuL 2, LiNCS·HL, ZnL 2 and HL as input geometries for the minimum energy optimization in vacuo. The results indicate that in ML 2 complexes (M = Cu, Zn) the electronic structure of the metal ion determines the spatial orientation of benzo-15-crown-5 macrocycles, and their different orientation in CuL 2 and ZnL 2 results in different potentiometric selectivities of ISEs based on these compounds. International Union of Crystallography 2007.

Full text of DOI:10.1107/S0108768107003898

research papers
Acta Crystallographica Section B
Structural
Science
Synchrotron powder diffraction in a systematic
study of 40-[2-(tosylamino)benzylideneamino]-2,3-
benzo-15-crown-5 complexes
ISSN 0108-7681
Andrey V. Dorokhov,a
The crystal structures of two compounds, CuL2 and
Received 12 July 2006
Accepted 24 January 2007
LiNCSꢀHL [HL = 40-[2-(tosylamino)benzylideneamino]-2,3-
Dmitrii Yu. Chernyshov,b
Anatolii S. Burlov,c Alexandr D.
Garnovskii,c Irina S. Ivanova,d
Elena N. Pyatova,d Aslan Yu.
Tsivadze,a Leonid A. Aslanove
benzo-15-crown-5], have been determined from synchrotron
powder diffraction data. Both compounds crystallize in the
monoclinic space group P21/c and with one molecule in the
asymmetric unit. In CuL2 the four N atoms of two bidentate L
ligands coordinate the CuII ion in a distorted tetrahedral
and Vladimir V. Chernysheva,e
*
˚
geometry with Cu—N distances of 1.98 (5)–2.05 (5) A, while
two O atoms from two sulfoxide groups complete the distorted
aA. N. Frumkin Institute of Physical Chemistry
and Electrochemistry, Leninsky prospect 31,
119991 Moscow GSP-1, Russia, bESRF, BP 220,
F-38043, Grenoble CEDEX, France, cResearch
Institute of Physical and Organic Chemistry,
Rostov State University, ul. Stachki, 194/2,
344090 Rostov-on-Don, Russia, dN. S.
Kurnakov Institute of General and Inorganic
Chemistry, Leninsky prospect 31, 119991
Moscow GSP-1, Russia, and eDepartment of
Chemistry, Moscow State University, 119992
Moscow, Russia
octahedral Cu coordination [Cu—O 2.64 (4), 2.74 (4) A]. In
LiNCSꢀHL, lithium is coordinated by all five ether O atoms
˚
˚
with Li—O distances of 2.03 (3)–2.50 (3) A and an N atom
˚
from the thiocyanate moiety [Li—N 1.98 (3) A] in a distorted
pentagonal pyramidal geometry. Preliminary potentiometric
selectivity measurements for ion-selective electrodes (ISEs)
based on CuL2 and ZnL2 demonstrated significant differences
in their selectivity. In order to find a possible reason for this,
theoretical calculations at the DFT (B3LYP) level were
performed. These calculations used the crystal structures of
CuL2, LiNCSꢀHL, ZnL2 and HL as input geometries for the
minimum energy optimization in vacuo. The results indicate
that in ML2 complexes (M = Cu, Zn) the electronic structure
of the metal ion determines the spatial orientation of benzo-
15-crown-5 macrocycles, and their different orientation in
CuL2 and ZnL2 results in different potentiometric selectivities
of ISEs based on these compounds.
Correspondence e-mail:
vladimir@struct.chem.msu.ru
1. Introduction
In recent years macrocyclic polyethers have attracted much
attention in many areas of science and technology (Izatt &
Christensen, 1978; Inoue & Gokel, 1990; Fedorova et al., 2001;
Bren, 2001; Tsivadze, 2004). Very recent investigations have
shown that crown ethers can effectively be used in pharma-
ceutical applications of chiral capillary electrophoresis as
chiral selectors (Amini, 2001; Ha et al., 2006), for the trans-
portation of a protein from an aqueous phase to ionic liquids
(Shimojo et al., 2006), as supramolecular optical chemosensors
(Bell & Hext, 2004) and in many other areas (Gokel et al.,
2004; Cook et al., 2005; Mazik et al., 2006). One of the
successful and well known applications of crown ethers in
analytical chemistry is ion transfer and ion recognition,
particularly as neutral carriers in ion-selective electrodes (ISE;
¨
Bakker et al., 1997; Buhlmann et al., 1998). To improve or
enhance the ion selectivity of crown ethers, a number of novel
macrocycles have been designed and synthesized in recent
years (Liu et al., 2000, and references therein; Gokel et al.,
2004, and references therein). The synthesis of new ligands
# 2007 International Union of Crystallography
Printed in Singapore – all rights reserved
402 doi:10.1107/S0108768107003898
Acta Cryst. (2007). B63, 402–410

research papers
having two or more donor centers has attracted special
attention (Van Veggel et al., 1994; Gonzalez-Lorenzo et al.,
2003; Sousa et al., 2004), which will allow the creation of new
polyfunctional materials.
Of specific interest to our group is the design and synthesis
of azomethine derivatives of benzo-15-crown-5 ether and their
complexes along with the structural characterization of the
compounds obtained (Minacheva, Ivanova, Dorokhov et al.,
2004; Minacheva, Ivanova, Pyatova et al., 2004; Ivanova et al.,
2005; Minacheva et al., 2006). Recently, we reported the
synthesis and solid-state structures of the 40-[2-(tosylamino)-
benzylideneamino]-2,3-benzo-15-crown-5
(HL)
ligand
2.2. Potentiometric measurements
(Minacheva, Ivanova, Dorokhov et al., 2004) and its ZnL2
complex (Minacheva et al., 2006). We have also synthesized
the CuL2 and LiNCSꢀHL complexes, which did not yield single
crystals suitable for X-ray crystal structure determination but
were obtained only as crystalline powders. Preliminary
potentiometric selectivity measurements for the ion-selective
electrodes (ISEs) based on HL, ZnL2 (Ivanova et al., 2007a)
and CuL2 (Ivanova et al., 2007b) demonstrated significant
differences in their selectivity
2.2.1. Membranes and ISEs preparation. The neutral carrier
(HL, ZnL2 or CuL2), plasticizer (o-nitrophenyloctyl ether,
Fluka) and high molecular-weight polyvinylchloride (PVC;
Aldrich) were dissolved in the appropriate volume of cyclo-
hexanone and mechanically stirred. All membrane cocktails
were cast in glass rings placed on glass plates. Solvent from
PVC membrane was allowed to evaporate for a week at room
temperature. The thickness of the resulting membranes was
about 0.2–0.3 mm. Small disks were punched from the cast
membranes and mounted in electrode bodies.
2.2.2. Measurements. The electrochemical properties of the
membranes were investigated in the conventional configura-
tion. The external reference electrode was a double-junction
Ag/AgCl OP-8020 P (Radelkis). The electrochemical poten-
tial was measured using a pH/ion analyzer OP-300 (Radelkis,
Hungary). The selectivity coefficients were determined by the
separate solution method (IUPAC, 1976) using 0.01 mol lꢁ1
solutions of metal chlorides or nitrates.
HL : Pb2þ>Agþ>Kþ>Rbþ>Cu2þ>Naþ ’ Csþ
’ Sr2þ>NHþ4
ZnL2 : Agþ>Pb2þ>Kþ>Cu2þ>Rbþ>Csþ ’ Ba2þ
’ Naþ ’ NHþ4
CuL2 : Cu2þ>Kþ>NH4þ>Naþ>Rbþ ’ Csþ>Pb2þ
>Ba2þ>Agþ:
A possible reason for the different selectivity of these three
compounds may be the different geometry of the molecules
and the conformation of the macrocycles. This encouraged us
to solve the crystal structures of the complexes in the hope of
finding a correlation between their structure and potentio-
metric selectivity. Here we present the crystal structure of the
CuL2 complex determined by powder diffraction methods
(David et al., 2002; Baerlocher & McCusker, 2004). The crystal
structure of the LiNCSꢀHL complex was also determined to
estimate the conformational flexibility of the HL ligand in the
case of the cation binding to the macrocycle.
2.3. DFT calculations
Density functional theory (DFT) calculations for the free
HL ligand and the three complexes ZnL2, CuL2 and
LiNCSꢀHL in vacuo were performed using the PRIRODA
program (Laikov, 1997) employing the B3LYP exchange-
correlation function (Becke, 1988; Lee et al., 1988). For
representation of the Kohn–Sham one-electron wavefunc-
tions, sets of contracted Gaussian-type functions were used;
the contracted patterns were (311/1) for H, (51111/31) for Li,
(611111/411/11) for C, N and O, (6111111111/5111111/11) for
S,
(711111111111/511111111/5111)
for
Cu,
and
(7111111111111/5111111111/5111) for Zn. For the CuL2
system, only the configuration S = 1/2 has been taken into
account. The geometry optimization was performed using the
2. Experimental
2.1. Synthesis
quasi-Newton method to the final gradient of
3
ꢂ
10ꢁ5 Hartree Aꢁ1. The in vacuo optimized Cartesian coordi-
nates obtained for the systems described in this paper are
given in the supplementary materials.
˚
N-(40-Benzo-15-crown-5)-2-(amino-N-tosyl)-phenylaldi-
mine (HL) and its complexes were prepared as depicted below
according to the published procedures for HL (Minacheva,
Ivanova, Dorokhov et al., 2004), CuL2 and ZnL2 (Minacheva et
al., 2006), and LiNCSꢀHL (Ivanova et al., 2005; also see the
supplementary materials1 for details of the synthesis).
2.4. Crystal structure determination
2.4.1. Data collection and indexing. Preliminary powder
diffraction measurements were carried out on a laboratory
Guinier–Huber camera. The laboratory patterns revealed a
good degree of crystallinity for both compounds, but did not
1 Supplementary data for this paper are available from the IUCr electronic
archives (Reference: KD5008). Services for accessing these data are described
at the back of the journal.
ꢃ
Acta Cryst. (2007). B63, 402–410
Andrey V. Dorokhov et al.
Synchrotron powder diffraction 403

research papers
Table 1
Experimental details.
CuL2
LiNCSꢀHL
HL†
ZnL2‡
Crystal data
Chemical formula
Mr
C56H62CuN4O14S2
1142.79
C29H32LiN3O7S2
605.66
C28H32N2O7S
540.63
C56H62N4O14S2Zn
1144.62
Cell setting, space group
Temperature (K)
Monoclinic, P21/c
295 (2)
19.0356 (6), 17.4332 (6),
17.4180 (6)
113.824 (3)
53
5287.7 (3)
4
1.436
Synchrotron
0.30
Cylinder (particle
morphology: no specific
habit), black
15 ꢂ 0.3 ꢂ 0.3
101.3 (2)
Monoclinic, P21/c
295 (2)
9.2918 (3), 23.0117 (7),
15.2798 (5)
108.063 (2)
65
3106.11 (17)
4
1.295
Synchrotron
0.12
Cylinder (particle
morphology: no specific
habit), yellow
15 ꢂ 0.3 ꢂ 0.3
101.3 (2)
Monoclinic, P21/c
295 (2)
12.781 (1), 25.017 (9),
8.859 (1)
108.20 (2)
–
2691 (1)
4
1.334
Monoclinic, C2/c
295 (2)
23.139 (2), 19.931 (2),
14.604 (2)
121.20 (1)
–
5760.9 (5)
4
1.320
˚
a, b, c (A)
ꢁ (ꢄ)
M20
V (A )
3
˚
Z
Dx (Mg m–3)
Radiation type
ꢃ (mm–1)
Mo Kꢂ
–
–
Mo Kꢂ
–
–
Specimen form, colour
Specimen size (mm)
Specimen preparation pres-
sure (kPa)
–
–
–
–
Specimen preparation
temperature (K)
295 (2)
295 (2)
–
–
Data collection
Diffractometer
Data collection method
ESRF powder diffractometer
Specimen mounting:
specimen was sealed in a
0.3 mm diameter borosili-
cate glass capillary; mode:
transmission; scan method:
continuous
ESRF powder diffractometer
Specimen mounting:
specimen was sealed in a
0.3 mm diameter borosili-
cate glass capillary; mode:
transmission; scan method:
continuous
–
–
–
–
Absorption correction
2ꢀ (ꢄ)
None
2ꢀmin = 1.017, 2ꢀmax = 22.014,
increment = 0.003
None
2ꢀmin = 3.024, 2ꢀmax = 23.024,
increment = 0.004
–
–
2ꢀmin = 0, 2ꢀmax = 46
2ꢀmin = 0, 2ꢀmax = 50
Refinement
Refinement on
Rp§
Inet
Inet
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.0462
0.0447
0.0627
0.0593
0.0481
1.30
0.0435
0.0368
0.0545
0.0461
0.0274
1.98
Rwp
Rexp
Goodness-of-fit
Wavelength of incident
0.51966 (1)
0.52000 (1)
˚
radiation (A)
Excluded region(s)
Profile function
None
None
–
–
–
–
Split-type pseudo-Voigt
(Toraya, 1986)
277
Not refined
Based on measured s.u.’s
0.027
March–Dollase texture
correction (Dollase, 1986).
Direction of preferred
orientation: 100, texture
parameter r = 0.94 (2)
Split-type pseudo-Voigt
(Toraya, 1986)
213
Not refined
Based on measured s.u.’s
0.035
March–Dollase texture
correction (Dollase, 1986).
Direction of preferred
orientation: 001, texture
parameter r = 0.910 (9)
No. of parameters
H-atom treatment
Weighting scheme
(ꢀ/ꢄ)max
Preferred orientation correc-
tion
–
–
–
–
–
–
–
–
–
–
Computer programs used: MRIA (Zlokazov & Chernyshev, 1992), PLATON (Spek, 2003). † Minacheva, Ivanova, Dorokhov et al. (2004 ). ‡ Minacheva et al. (2006). § Rp, Rwp
and Rexp are defined according to Young & Wiles (1982). The results of the final bond-restrained Rietveld refinement are given in the first row of each pair of rows and the results of the
Pawley fit are given in the second row of each pair.
allow indexing. High-resolution synchrotron powder diffrac-
tion data were collected at the powder diffraction station
BM01B (Swiss–Norwegian beamlines at the European
Synchrotron Radiation Facility). The powder was placed in a
0.3 mm glass capillary and diffraction patterns were measured
in Debye–Scherrer geometry at room temperature. Owing to
the choice of wavelength, the low-angle region 2ꢀ < 3ꢄ for
LiNCSꢀHL was not accessible, so the d spacings of the two first
˚
peaks (12.270 and 11.501 A) were taken from a laboratory
pattern for indexing purposes. The monoclinic unit-cell
dimensions of LiNCSꢀHL and CuL2 were determined using
three indexing programs: TREOR90 (Werner et al., 1985), ITO
ꢃ
404 Andrey V. Dorokhov et al.
Synchrotron powder diffraction
Acta Cryst. (2007). B63, 402–410

research papers
(Visser, 1969) and AUTOX (Zlokazov, 1992, 1995). Based on
systematic extinctions the space group for both compounds
was determined to be P21/c. The unit-cell parameters and
space groups were further tested using Pawley’s fit (Pawley,
1981) and confirmed by crystal structure solution. The crys-
tallographic data for LiNCSꢀHL and CuL2 are given in Table
1.
2.4.2. Structure solution.. The crystal structures have been
solved by a combination of simulated annealing (Zhukov et al.,
2001) and grid-search (Chernyshev & Schenk, 1998) techni-
ques; detailed methodology is described elsewhere (Interna-
tional Union of Crystallography Newsletter, 2004).
LiNCSꢀHL: The initial molecular model for L without H
atoms was taken from the single-crystal structure of HL
(Minacheva, Ivanova, Dorokhov et al., 2004; CSD refcode
DANREQ). However, the conformation of the macrocycle
was changed to a more ‘planar’ configuration to be in accor-
dance with that observed in the related structure found in the
Cambridge Structural Database (CSD, Version 5.27; Allen,
2002), where the 15-crown-5 macrocycle binds the Li+ cation
(Nurtaeva & Holt, 1999; CSD refcode BEQREU). The only
torsional degree of freedom allowed for L at the beginning of
the structure solution process was the rotation of the tosyl
group around the N9—S8 bond (Fig. 1). The geometrical
parameters of the LiNCS moiety were taken from the litera-
ture (Dale et al., 1987; CSD refcode FOBHEJ) and fixed. The
simulated annealing was used in a preliminary search for
possible solutions; 12 degrees of freedom – three translational,
three rotational and one torsional parameters for L, three
translational and two rotational parameters for the LiNCS
were varied and 300 Xobs (Chernyshev & Schenk, 1998) low-
angle values were used. After each simulated annealing run,
the obtained orientations of L and LiNCS were fixed and used
search techniques were independently used in a search for the
position and orientation of L by varying nine degrees of
freedom, and both gave only one solution. The correctness of
the solution was later checked by the restraint-free Rietveld
refinement without H atoms; in spite of some distortions of the
bond lengths and angles the results were the same.
Cu L2: The initial molecular model without H atoms was
taken from the single-crystal structure of ZnL2 (Minacheva et
al., 2006). Three internal degrees of freedom were allowed for
CuL2; rotation of the tosyl group around the N9A—S8A and
N9B—S8B bonds, respectively, and limited variation of the
dihedral angle between the N9A/Cu1/N17A and N9B/Cu1/
N17B planes (Fig. 2). Following the aforementioned case of
LiNCSꢀHL, the simulated annealing and grid-search proce-
dures were tested by varying nine degrees of freedom and
using 300 Xobs low-angle values. However, no acceptable
solutions were obtained. Therefore, the strategy of structure
solution was changed and concentrated first on the location of
the Cu atom, using the grid-search technique applied to the
whole powder pattern and further tested by the Rietveld
refinement. The unique position of Cu was found and fixed,
then the orientations of the ligands with four additional
internal degrees of freedom; namely, limited rotations around
the C16A—N17A, C2A—S8A, C16B—N17B and C2B—S8B
bonds (Fig. 2) were determined by simulated annealing using
300 Xobs values. The conformations of two 15-crown-5
macrocycles were established at the stage of structural
refinement. Attempts to reveal the correctness of the solution
by the restraint-free Rietveld refinement were useless, because
of a large number of independent non-H atoms and weak
counting statistics for CuL2, compared with the case of
LiNCSꢀHL. In addition, a specific relationship between the
unit-cell dimensions (asin ꢁ ’ b ’ c) leads to severe over-
lapping of the peaks. As a result, many of the bond lengths and
angles appeared to be chemically unreasonable in the
restraint-free Rietveld refinement, although the positions of
Cu and two S atoms were the same.
˚
in a translational grid search with 0.4 A grids. This combina-
tion of simulated annealing and grid-search techniques was
applied several times and allowed the finding of the preferred
position and orientation for the LiNCS moiety, which were
further fixed. Two more internal degrees of freedom for L
were added – the limited rotations around the C16—N17 and
C2—S8 bonds (Fig. 1). The simulated annealing and grid-
Figure 2
Figure 1
View of CuL2 with the atom-numbering scheme. The numbering of C
atoms in the ligands A and B corresponds to that in LiNCSꢀHL. H atoms
have been omitted for clarity.
View of LiNCSꢀHL with the atom-numbering scheme. H atoms have been
omitted for clarity.
ꢃ
Acta Cryst. (2007). B63, 402–410
Andrey V. Dorokhov et al.
Synchrotron powder diffraction 405

research papers
Table 2
Selected bond distances (A) and angles ( ) for CuL2.
ligands L. However, taking into consideration the relatively
˚
short distances of Cu—O24A [2.74 (4) A] and Cu—O24B
ꢄ
˚
˚
[2.64 (4) A], one can see that O24A and O24B complete a
distorted octahedral coordination sphere (Fig. 2). The Cu—N
bond lengths have normal values and are in the range
Cu—N9A
1.98 (5)
2.05 (5)
2.74 (4)
1.46 (9)
1.29 (8)
Cu—N9B
1.98 (5)
2.03 (5)
2.64 (4)
1.49 (8)
1.36 (8)
Cu—N17A
Cu—O24A
N17A—C18A
N17A—C16A
Cu—N17B
Cu—O24B
N17B—C18B
N17B—C16B
˚
1.98 (5)—2.05 (5) A. The N—Cu—N tetrahedral angles are in
the range 87 (2)–143 (2)ꢄ, while in the ZnL2 complex the N—
Zn—N angles are in the range 95.2 (4)–128.5 (6)ꢄ (Minacheva
et al., 2006). Considerable distortion of the CuN4 tetrahedron
is caused mainly by the formation of six-membered metallo-
cycles with chelate angles N9X—Cu—N17X (X = A or B) of
94 (2) and 87 (2)ꢄ for the ligands A and B, respectively. The
mean planes of two chelate rings, Cu/N9A/C10A/C15A/C16A/
N17A and Cu/N9B/C10B/C15B/C16B/N17B, make a dihedral
angle of 72.9 (1)ꢄ.
N9A—Cu—N17A
N9A—Cu—N17B
N9A—Cu—N9B
O24A—Cu—N9A
O24A—Cu—N17A
O24A—Cu—N9B
O24A—Cu—N17B
O24A—Cu—O24B
94 (2)
111 (2)
143 (2)
59.2 (15)
152.3 (18)
97.8 (19)
75.0 (18)
86.7 (11)
N9B—Cu—N17B
N9B—Cu—N17A
N17A—Cu—N17B
O24B—Cu—N9B
O24B—Cu—N17B
O24B—Cu—N9A
O24B—Cu—N17A
87 (2)
103 (2)
124 (2)
61.5 (17)
141.0 (16)
87.0 (18)
87.0 (17)
The conformation of the macrocyclic ring in a crown ether
can be described as a sequence of conformations of ethylene
glycol units: O—CH2—CH2—O. Each can be written as a
three-letter word with the letters T, C, G or S corresponding to
2.4.3. Rietveld refinement. In the final bond-restrained
Rietveld refinements the patterns were fitted with the program
MRIA (Zlokazov & Chernyshev, 1992) using a split-type
pseudo-Voigt peak-profile function (Toraya, 1986) and taking
into account anisotropic line-broadening (Popa, 1998) and the
March–Dollase texture formalism (Dollase, 1986). The refined
values of the texture parameter r and the directions of the
preferred orientations were 0.910 (9) and [001] for LiNCSꢀHL,
and 0.94 (2) and [100] for CuL2, respectively.
LiNCSꢀHL: Four isotropic displacement parameters were
refined: two for the S8 and S39 atoms, respectively, one
common parameter for the C40, N41 and Li42 atoms, and one
common parameter for the remaining non-H atoms.
Restraints were applied to the intramolecular bond lengths
˚
and contacts (< 2.6 A) in the HL and LiNCS moieties. The
strength of the restraints was a function of interatomic
separation and, for intramolecular bond lengths, corresponded
˚
to an r.m.s. deviation of 0.03 A. Additional restraints were
applied to the planarity of the following fragments (Fig. 1):
(i) C1–C7, S8 (eight atoms),
(ii) N9, C10–C16 (eight atoms) and
(iii) N17, C18–C23, O26, O27 (nine atoms), with the
˚
maximal allowed deviation from the mean plane 0.1 A.
CuL2: Five isotropic displacement parameters were refined
– three for Cu1, S8A and S8B atoms, respectively, and two
common parameters for the remaining non-H atoms in each
ligand. The restraints applied to each ligand were the same as
described above for LiNCSꢀHL.
H atoms were positioned geometrically with C—H 0.93–
˚
0.98 and N—H 0.86 A, and not refined. The solid-state
structures of LiNCSꢀHL and CuL2 are shown in Figs. 1 and 2,
respectively. The diffraction profiles after the final bond-
restrained Rietveld refinement are presented in Fig. 3.
3. Discussion
3.1. Molecular structure of CuL2
A molecule of CuL2 is shown in Fig. 2. Selected bond
distances and angles are given in Table 2. The CuII ion is
tetrahedrally coordinated by four N atoms of two bidentate
Figure 3
Rietveld plots for LiNCSꢀHL (top) and CuL2 (bottom).
ꢃ
406 Andrey V. Dorokhov et al.
Synchrotron powder diffraction
Acta Cryst. (2007). B63, 402–410

research papers
al., 2002) and Dx = 1.28 Mg mꢁ3 for M = Co (Castro et al.,
2001).
Table 3
Selected bond distances (A) and angles ( ) for LiNCSꢀHL.
ꢄ
˚
Li42—O26
Li42—O27
Li42—O30
N17—C18
2.24 (3)
2.03 (3)
2.27 (3)
1.471 (18)
Li42—O33
Li42—O36
Li42—N41
N17—C16
2.50 (3)
2.44 (3)
1.98 (3)
1.338 (18)
3.2. LiNCSꢀHL structure
The molecular structure of LiNCSꢀHL is shown in Fig. 1.
Selected bond distances and angles are given in Table 3.
Lithium is coordinated by all five ether O atoms and the
thiocyanate N atom in a distorted pentagonal pyramidal
C20—O26—Li42
C21—O27—Li42
C29—O30—Li42
C32—O33—Li42
C35—O36—Li42
C40—N41—Li42
108.2 (12)
115.1 (13)
102.5 (9)
116.2 (11)
120.1 (10)
165.3 (14)
C38—O26—Li42
C28—O27—Li42
C31—O30—Li42
C34—O33—Li42
C37—O36—Li42
123.9 (10)
118.5 (11)
119.1 (10)
122.6 (12)
110.3 (9)
˚
geometry. The Li42 ion is 0.38 A above the mean plane
formed by O26/O27/O30/O33/O36 with an r.m.s. deviation of
˚
0.06 A. The planar configuration of these O atoms is typical
the trans, cis, gauche or skew torsional conformations,
respectively. In these notations, the conformations of macro-
cycles in CuL2 and ZnL2 are described (the first word corre-
sponds to the conformation of the O—CPh—CPh—O unit)
for b15c5 complexes with lithium and sodium salts (Kireeva et
al., 1991). The Li—O distances are in the range 2.03 (3)–
˚
˚
2.50 (3) A [average bond distance 2.30 (18) A], which is
normal for the distances between the lithium and ether O
atoms (Boulatov et al., 1999, and references therein). A search
of the Cambridge Structural Database (Version 5.27; Allen,
2002) revealed only two crystal structures of Li complexes
with b15c5 and its derivatives, namely bis(aqua(benzo-15-
crown-5)-lithium)(benzo-15-crown-5))hexaiodo-tetracopper
(Nurtaeva & Holt, 1999; CSD refcode BEQREU) and the
toluene and tetrahydrofuran solvates of bis-(ꢃ4-bromo)-(12)-
mercurocarborand-4)bis(aqua-(benzo-15-crown-5)lithium)-
benzo-15-crown-5) (Yang et al., 1994; CSD refcode LICHUA).
Both compounds contain two types of 15c5 macrocycles, one
of which binds to the lithium ion and another one which is
free. The average Li—O distances in the b15c5 macrocycles
CuL2ðAÞ : TCT TGT SST SGT GGT;
CuL2ðBÞ : TCT TGS SST SGT TGT;
ZnL2 : TCT TGS STS SSG TGT:
In contrast to ZnL2, the CuL2 complex does not exhibit C2
symmetry. As a result, the macrocycles in A and B have
different conformations; however, the latter are normal for
benzo-15-crown-5 (b15c5) derivatives as follows from the
inspection of 123 crystal structures containing the b15c5
macrocycle in a free state (CSD Version 5.27; Allen, 2002).
The most striking feature of the CuL2 crystal structure is its
more dense packing compared with ZnL2. The differences in
˚
with the bonded lithium are 2.31 and 2.25 A in BEQREU and
3
˚
the unit-cell volumes, 5287.7 (3) and 5760.9 (5) A , lead to
)
LICHUA, respectively. In LiNCSꢀHL the O—Li42—O angles
between the neighbouring O atoms are in the range 63.5 (8)–
76.3 (9)ꢄ, while the ideal angle is 72ꢄ.
The conformations of macrocycles in LiNCSꢀHL and HL
(Minacheva, Ivanova, Dorokhov et al., 2004) can be described
as
significantly different values of Dx (1.435 and 1.334 Mg mꢁ3
for CuL2 and ZnL2, respectively. PLATON (Spek, 2003)
analysis of the two crystal structures shows the presence of
3
voids with the volume of 150 A in ZnL2 centered at (0.5,0,0)
˚
and symmetry-related positions, and no voids in CuL2.
However, this variation in Dx values for identical complexes is
not exceptional. An illustrative example is bis((N-(4-methyl-
phenyl)sulfonyl-1H-amido-4-isopropyl-2-phenyl)-2-oxazo-
line)-M(II), which has Dx = 1.35 Mg mꢁ3 for M = Zn (Castro et
LiNCS ꢀ HL : TCT TGT TGT TGT TGT;
HL : GCT TGT TGT TGS STT:
Figure 4
The comparative plots of (a) ZnL2, (b) CuL2 and (c) LiNCSꢀHL geometries after B3LYP optimization (red) and as taken from the crystal structure
(black). This figure is in colour in the electronic version of this paper.
ꢃ
Acta Cryst. (2007). B63, 402–410
Andrey V. Dorokhov et al.
Synchrotron powder diffraction 407

research papers
Table 4
Table 5
Potentiometric selectivity constants for HL, ZnL2 and CuL2.
ꢄ
˚
A comparison of selected geometrical parameters (A, ) of X-ray (first
row of each line) and DFToptimized (second row of each line) complexes
of ML2 (M = Zn, Cu).
pot
lg k
Pb2þ=ion
Interfering ions
HL
ZnL2
CuL2
ZnL2
CuL2, ligand A
CuL2, ligand B
NHþ4
Na+
K+
ꢁ2.9
ꢁ2.8
ꢁ1.5
ꢁ1.6
ꢁ2.8
ꢁ2.0
ꢁ2.4
ꢁ1.8
ꢁ2.3
ꢁ1.6
ꢁ2.3
ꢁ0.3
ꢁ3.8
ꢁ3.8
ꢁ2.5
ꢁ3.2
ꢁ3.5
ꢁ3.3
ꢁ2.9
ꢁ3.3
ꢁ2.6
ꢁ2.0
ꢁ2.2
+0.8
+1.9
ꢁ0.8
+2.1
ꢁ0.8
ꢁ0.8
ꢁ0.2
ꢁ0.4
ꢁ0.2
ꢁ0.1
+3.3
ꢁ0.2
ꢁ0.7
M—N9
2.00 (1)
2.05
2.05 (1)
2.10
2.74 (1)
2.77
164.0 (3)
127
1.98 (5)
2.03
2.05 (5)
2.07
2.74 (4)
2.86
136 (7)
134
1.98 (5)
2.03
2.03 (5)
2.07
2.64 (4)
2.86
130 (6)
133
M—N17
Rb+
Cs+
M—O24
Mg2+
Ca2+
Sr2+
Ba2+
Cu2+
Zn2+
Ag+
C16 N17—C18—C19
The symmetry of the b15c5 macrocycle in LiNCSꢀHL is higher
than that in the free HL ligand, as follows from the confor-
mations of ethylene glycol units, all of which except one
(OCPhCPhO) transform to TGT. The same conformation of
b15c5 was observed in its complexes with Li and Na (Kireeva
et al., 1991).
conformation is energetically unfavourable in vacuo, as
follows from the geometries of the DFT optimized complexes
ZnL2 and CuL2, but in the crystal (where weak intermolecular
forces may influence molecular geometry) we observe just this
unfavourable conformation, which is, however, favourable for
the ꢅꢀ ꢀ ꢀꢅ stacking formation. Indeed, in the crystal packing of
ZnL2 ꢅꢀ ꢀ ꢀꢅ interactions exist which link the molecules into
3.3. Comparison of CuL2 and ZnL2 complexes
A comparison of CuL2 and ZnL2 shows different confor-
mations of the crowns in each molecule in the crystalline state.
Even the two crowns in two ligands of CuL2 have different
conformations. The data suggest that the differences in ion
selectivity for CuL2 and ZnL2 may be related to the hole size,
i.e. when the crown ether’s interior cavity (‘hole’) is about the
same size as a given cation, binding will be optimal. However,
it should be noted that the idea of hole size is not a principle
that can be applied universally (Gokel et al., 2004). Moreover,
the X-ray structure is essentially static; the crystal packing
effects may significantly influence the molecular geometry,
while complexation, or ion binding, typically occurs in a
solvent.
˚
ribbons with the short distance of 3.650 (5) A between the
centroids of the benzene rings C10–C15 and C18–C23 from
the neighbouring molecules (Fig. 5). Further packing of these
ribbons with the protruding macrocycles in the crystal leads to
a porous packing with considerable voids; this explains the
existence of voids with a volume of 150 A in ZnL2 crystal
packing.
3
˚
The essential difference in the values of the torsion angle
N—C—C in the crystal and in vacuo optimized ZnL2
C
indicates a weak interaction of the b15c5 macrocycle with the
rest part of the molecule in the ML2 complex; this allows easy
rotation of b15c5 around the N17—C18 bond. Another
To understand why the conformations of CuL2 and ZnL2 in
the solid state are different and to estimate their probable
geometries in a solution, we carried out a computational
investigation of these complexes by DFT using the X-ray
structures as input geometries. The minimum-energy geome-
tries calculated for ZnL2 and CuL2 are shown in Fig. 4 along
with their X-ray geometries. Both calculated structures show
some elongation of the M—N coordinating bond lengths
(Table 4). In ZnL2 the DFT optimization basically preserves
the C2 symmetry of the molecule, but causes a significant
rotation of the b15c5 moiety around the N17—C18 bond. The
torsion angle C N—C—C changes from 164.0 (7)ꢄ in the X-
ray structure to 127ꢄ in the complex optimized in vacuo. In the
CuL2 complex there are no changes in this torsion angle in
either ligands within experimental error (Table 4) and the
values of this angle in both ligands of the DFT optimized
complex (133 and 134ꢄ) are close to that in the ZnL2 complex
(127ꢄ).
Figure 5
In the crystal structure of ZnL2 the high value of the torsion
angle C N—C—C indicates that two benzene rings (C10–
C15 and C18–C23) are nearly parallel in each ligand. This
The ꢅꢀ ꢀ ꢀꢅ interactions (dotted lines) in the crystal structure of ZnL2. The
distances between the centroids of benzene rings from the neighbouring
˚
molecules are 3.650 (5) A.
ꢃ
408 Andrey V. Dorokhov et al.
Synchrotron powder diffraction
Acta Cryst. (2007). B63, 402–410

research papers
important conclusion is that the macrocycle conformation is
almost independent of the nature of the central M ion. The
central M ion defines the geometry of that part of the ligand
which coordinates it, namely, the 2-(amino-N-tosyl)phenyl-
aldimine (L0). The geometries of ML02 fragments are qualita-
tively the same in X-ray and in vacuo optimized systems of
ZnL2 and CuL2 (Fig. 4).
The Cu/N9/C10–C16/N17 fragment is distorted from
planarity in both ligands of the DFT-optimized CuL2 complex
– the maximal deviations from the mean plane in ligand A are
considered to be flexible and capable of orienting three of its
five donor groups in space. However, as soon as the macro-
cycle binds an ion, its conformation becomes more rigid and
can be predicted by DFT calculations with a good precision, as
has been shown by Platas-Iglesias and co-workers (Platas-
´
Iglesias, Vaiana et al., 2005; Platas-Iglesias, Esteban-Gomez et
al., 2005). To be sure that this statement is valid for our
compounds, we carried out a DFT geometry optimization of
the LiNCSꢀHL complex, where the macrocycle binds a lithium
ion. A comparison of the X-ray and in vacuo optimized
geometries of LiNCSꢀHL (Fig. 4c) revealed that the macro-
cycle conformation and the positions of Li and coordinating N
atoms suffer only very small variations, in spite of the
considerable displacements of the tosyl group and the S atom
from the thiocyanate moiety.
˚
ꢁ0.13 and 0.18 A for the Cu and N9A atoms, respectively, and
˚
in ligand B the same deviations are ꢁ0.16 and 0.19 A for the
Cu and N9B atoms, respectively. These two mean planes form
a dihedral angle of 65ꢄ. The same atomic fragment in DFT-
optimized ZnL2, Zn/N9/C10–C16/N17, is almost planar – the
maximal deviations from the mean planes in the two ligands
˚
do not exceed 0.04 A. However, the most important distinc-
tion from the CuL2 complex is the almost perpendicular
orientation of the two mean planes, with a dihedral angle of
89ꢄ. The value of this dihedral angle defines the dimensions of
the ‘access’ area, where an ion has the opportunity to bind to
the macrocycle without steric obstacles (caused by the C10–
C15 benzene ring and O24 atom from another ligand) – the
3.4. Conformational flexibility of the ligand HL
The conclusions that the macrocycle in the ML2 complex
should be considered flexible and that its conformation is
almost independent of the nature of the central M atom have
stimulated us to optimize the geometries of several HL ligands
in vacuo to compare their conformations and relative energies.
The five input geometries of HL ligands were taken from the
crystal structures of HL, LiNCSꢀHL, ZnL2 (one independent
ligand) and CuL2 (two independent ligands). The in vacuo
optimized geometries were compared to the input X-ray
values in Fig. 6. The differences in the energies for the five
optimized geometries do not exceed 16 kJ molꢁ1, allowing
various conformations of the macrocycle and giving no
preference to any in particular.
closer to 90ꢄ the larger the ‘access’ area. If this assumption is
pot
Pb2þ=ion
valid, the potentiometric selectivity constants k
for
ZnL2 are to correspond with those for HL, which has a
maximal possible ‘access’ area, and experimental measure-
ments of transport properties in PVC membranes show this
tendency (Table 5).
The variety of b15c5 conformations observed in the ligands
of both complexes in the solid state and optimized in vacuo
clearly shows that in ML2 solution, the macrocycle should be
Figure 6
The comparative plots of HL geometries after B3LYP optimization (red) and as taken from the crystal structure (black) of (a) HL, (b) LiNCSꢀHL, (c)
CuL2 ligand A, (d) CuL2 ligand B and (e) ZnL2. This figure is in colour in the electronic version of this paper.
ꢃ
Acta Cryst. (2007). B63, 402–410
Andrey V. Dorokhov et al.
Synchrotron powder diffraction 409

research papers
Inoue, Y. & Gokel, G. W. (1990). Editors. Cation Binding by
Macrocycles. NY: Marcell Dekker.
International Union of Crystallography Newsletter (2004). Vol. 31,
pp. 5–15. NY: IUCr.
IUPAC Recommendation for Nomenclature of Ion Selective
Electrodes (1976). Pure Appl. Chem. 48, 127.
Ivanova, I. S., Dorokhov, A. V., Pyatova, E. N., Bicherov, A. V.,
Burlov, A. S., Garnovskii, A. D. & Tsivadze, A. Yu. (2005). Russ. J.
Coord. Chem. 31, 483–488.
Ivanova, I. S., Dorokhov, A. V., Pyatova, E. N., Burlov, A. S.,
Garnovskii, A. D. & Tsivadze, A. Yu. (2007a). Russ. J. Coord.
Chem. In the press.
Ivanova, I. S., Dorokhov, A.V., Pyatova, E. N., Burlov, A. S.,
Garnovskii, A. D. & Tsivadze, A. Yu. (2007b). Russ. J. Coord.
Chem. In the press.
Izatt, R. M. & Christensen, J. J. (1978). Editors. Synthetic Multidentate
Macrocyclic Compounds. NY: Academic Press.
Kireeva, I. K., Generalova, N. B., Trofimov, V. A. & Tsivadze, A. Yu.
(1991). Russ. J. Inorg. Chem. 36, 830–838.
4. Conclusion
A comparison of X-ray and DFT optimized geometries of two
complexes, CuL2 and ZnL2, allowed us to identify the rela-
tively rigid ML02 fragments [L0 = 2-(amino-N-tosyl)phenyl-
aldimine] in the molecules. The geometry of this ML02 frag-
ment depends on the electronic structure of the metal ion M,
which determines the geometry of the MN4 coordination
tetrahedron and the spatial orientation of b15c5 macrocycles
in ML2 complexes. The differences in spatial orientations of
b15c5 macrocycles in CuL2 and ZnL2 complexes qualitatively
explain the difference in their potentiometric selectivities. The
results allow further computational study of various ML2
complexes with two bonded ions – the rigidity of the ML20
fragment and the conformation of the macrocycle with the
bonded ion provide a good basis for such time-consuming ab
initio calculations.
Laikov, D. N. (1997). Chem. Phys. Lett. 281, 151–156.
Lee, C. T., Yang, W. T. & Parr, R. G. (1988). Phys. Rev. B, 37, 785–789.
Liu, Y., Zhang, H.-Y., Chen, L.-X., He, X.-W., Wada, T. & Inoue, Y.
(2000). J. Org. Chem. 65, 2870–2874.
Mazik, M., Kuschel, M. & Sicking, W. (2006). Org. Lett. 8, 855–
858.
The authors thank the staff of the SNBL for their friendly
support during the synchrotron experiments and Professor
A. V. Yatsenko for his kind assistance with DFT calculations.
Minacheva, L. Kh., Ivanova, I. S., Dorokhov, A. V., Bicherov, A. V.,
Burlov, A. S., Garnovskii, A. D., Sergienko, V. S. & Tsivadze, A. Yu.
(2004). Dokl. Chem. 398, 179–184.
References
Minacheva, L. Kh., Ivanova, I. S., Dorokhov, A. V., Burlov, A. S.,
Garnovskii, A. D., Sergienko, V. S. & Tsivadze, A. Yu. (2006). Russ.
J. Coord. Chem. 32, 166–172.
Minacheva, L. Kh., Ivanova, I. S., Pyatova E. N., Dorokhov, A. V.,
Bicherov, A. V., Burlov, A. S., Garnovskii, A. D., Sergienko, V. S. &
Tsivadze, A. Yu. (2004). Dokl. Chem. 395, 68–73.
Nurtaeva, A. & Holt, E. M. (1999). Acta Cryst. C55, 1453–1457.
Pawley, G. S. (1981). J. Appl. Cryst. 14, 357–361.
Allen, F. H. (2002). Acta Cryst. B58, 380–388.
Amini, A. (2001). Electrophoresis, 22, 3107–3130.
Baerlocher, Ch. & McCusker, L. B. (2004). Editors. Z. Kristallogr.
219, 782–901.
¨
Bakker, E., Buhlmann, P. & Pretsch, E. (1997). Chem. Rev. 97, 3083–
3132.
Becke, A. D. (1988). Phys. Rev. A, 38, 3098–3100.
Bell, T. W. & Hext, N. M. (2004). Chem. Soc. Rev. 33, 589–598.
Boulatov, R., Du, B., Meyers, E. A. & Shore, S. G. (1999). Inorg.
Chem. 38, 4554–4558.
´
´
´
Platas-Iglesias, C., Esteban-Gomez, D., Enrıquez-Perez, T., Avecilla,
F., de Blas, A. & Rodriguez-Blas, T. (2005). Inorg. Chem. 44, 2224–
2233.
Bren, V. A. (2001). Russ. Chem. Rev. 50, 1017–1034.
¨
´
Platas-Iglesias, C., Vaiana, L., Esteban-Gomez, D., Avecilla, F., Real,
Buhlmann, P., Pretsch, E. & Bakker, E. (1998). Chem. Rev. 98, 1593–
1687.
J. A., de Blas, A. & Rodriguez-Blas, T. (2005). Inorg. Chem. 44,
9704–9713.
Popa, N. C. (1998). J. Appl. Cryst. 31, 176–180.
Shimojo, K., Nakashima, K., Kamiya, N. & Goto, M. (2006).
Biomacromolecules, 7, 2–5.
Castro, J., Cabaleiro, S., Perez-Lourido, P., Romero, J., Garcia-
Vazquez, J. A. & Sousa, A. (2001). Polyhedron, 20, 2329–2337.
Castro, J., Cabaleiro, S., Perez-Lourido, P., Romero, J., Garcia-
Vazquez, J. A. & Sousa, A. (2002). Z. Anorg. Allg. Chem. 628,
1210–1217.
Chernyshev, V. V. & Schenk, H. (1998). Z. Kristallogr. 213, 1–3.
Cook, H. A., Klampfl, C. W. & Buchberger, W. (2005). J. Chromatogr.
A, 1085, 164–169.
Dale, J., Eggestad, J., Fredriksen, S. B. & Groth, P. (1987). Chem.
Commun. pp. 1391–1392.
David, W. I. F., Shankland, K., McCusker, L. B. & Baerlocher, Ch.
(2002). Editors. Structure Determination from Powder Diffraction
Data. New York: Oxford University Press.
Sousa, C., Gameiro, P., Freire, P. & de Castro, B. (2004). Polyhedron,
23, 1401–1408.
Spek, A. L. (2003). J. Appl. Cryst. 36, 7–13.
Toraya, H. (1986). J. Appl. Cryst. 19, 440–447.
Tsivadze, A. Yu. (2004). Russ. Chem. Rev. 73, 5–23.
Van Veggel, F. C. J. M., Verboom, W. & Reinhoudt, D. N. (1994).
Chem. Rev. 94, 279–299.
Visser, J. W. (1969). J. Appl. Cryst. 2, 89–95.
Werner, P.-E., Eriksson, L. & Westdahl, M. (1985). J. Appl. Cryst. 18,
367–370.
Dollase, W. A. (1986). J. Appl. Cryst. 19, 267–272.
Fedorova, O. A., Gromov, S. P. & Alfimov, M. V. (2001). Russ. Chem.
Bull. 50, 1970–1983.
Yang, X., Knobler, C. B., Zheng, Z. & Hawthorne, M. F. (1994). J. Am.
Chem. Soc. 116, 7142–7159.
Young, R. A. & Wiles, D. B. (1982). J. Appl. Cryst. 15, 430–438.
Zhukov, S. G., Chernyshev, V. V., Babaev, E. V., Sonneveld, E. J. &
Schenk, H. (2001). Z. Kristallogr. 216, 5–9.
Zlokazov, V. B. (1992). J. Appl. Cryst. 25, 69–72.
Zlokazov, V. B. (1995). Comput. Phys. Commun. 85, 414–422.
Zlokazov, V. B. & Chernyshev, V. V. (1992). J. Appl. Cryst. 25, 447–
451.
Gokel, G. W., Leevy, W. M. & Weber, M. E. (2004). Chem. Rev. 104,
2723–2750.
Gonzalez-Lorenzo, M., Platas-Iglesias, C., Avecilla, F., Geraldes,
¨
C. F. G. C., Imbert, D., Bunzli, J.-C. G., De Blas, A. & Rodriguez-
Blas, T. (2003). Inorg. Chem. 42, 6946–6954.
Ha, P. T., Hoogmartens, J. & Van Schepdael, A. (2006). J. Pharm.
Biomed. Anal. 41, 1–11.
ꢃ
410 Andrey V. Dorokhov et al.
Synchrotron powder diffraction
Acta Cryst. (2007). B63, 402–410