Molecular properties of a bis(ketoiminato)-bis(tricarbonyliron) complex obtained by symmetric cleavage from acetophenone azine. Idealized symmetry in anti and syn isomers studied by X-ray diffraction, NMR and density functional theory

Article abstract of DOI:10.1016/S0022-2860(01)00604-4

Distinct anti and syn isomers of the bis(μ₂-acetophenoniminato)-bis(tricarbonyliron) obtained from iron dodecacarbonyl and 1,4-dimethyl-1,4-diphenyl-2,3-diazabuta-1,3-diene by symmetric cleavage of the azine have been compared in terms of molecular deviations from symmetry. The study was carried out by the analysis of intra- and intermolecular close contacts in crystals, quantum chemical DFT calculations for the isolated molecules and by NMR in solution. In the crystalline state the intramolecular contacts and also the symmetry perturbations were more strongly expressed in the syn form as compared with anti, and the same relation was perceived in the DFT-optimized single molecules. However, in solution only symmetric conformations were observed for both isomers anti and syn by ¹H and ¹³C NMR at room temperature and at -70°C.

Full text of DOI:10.1016/S0022-2860(01)00604-4

Journal of Molecular Structure 597 ꢀ2001) 211±221
www.elsevier.com/locate/molstruc
Molecular properties of a bisꢀketoiminato)-bisꢀtricarbonyliron)
complex obtained by symmetric cleavage from acetophenone
azine. Idealized symmetry in anti and syn isomers studied by
X-ray diffraction, NMR and density functional theory
a,
Andrzej Zimniak , Grzegorz Bakalarski
b
*
aFaculty of Pharmacy, Medical Academy of Warsaw, Banacha 1, 02-097 Warsaw, Poland
bInterdisciplinary Centre for Mathematical and Computational Modelling, Warsaw University, 02-106 Warsaw, Pawinskiego 5A, Poland
Received 22 May 2000; revised 26 March 2001; accepted 30 March 2001
Abstract
Distinct anti and syn isomers of the bisꢀm2-acetophenoniminato)-bisꢀtricarbonyliron) obtained from iron dodecacarbonyl and
1,4-dimethyl-1,4-diphenyl-2,3-diazabuta-1,3-diene by symmetric cleavage of the azine have been compared in terms of
molecular deviations from symmetry. The study was carried out by the analysis of intra- and intermolecular close contacts
in crystals, quantum chemical DFT calculations for the isolated molecules and by NMR in solution. In the crystalline state the
intramolecular contacts and also the symmetry perturbations were more strongly expressed in the syn form as compared with
anti, and the same relation was perceived in the DFT-optimized single molecules. However, in solution only symmetric
conformations were observed for both isomers anti and syn by 1H and 13C NMR at room temperature and at 2708C.
q 2001 Elsevier Science B.V. All rights reserved.
Keywords: Iron complexes; Azines; Bisꢀtricarbonyliron); Structure; Symmetry; Close contacts; DFT calculations; NMR; Bisꢀm2-acetopheno-
niminato)-bisꢀtricarbonyliron)
1. Introduction
with ortho-proton rearrangement to the azomethine
carbon and the metallization of the phenyl ring [1,2].
In complexes obtained from ketazines ꢀFig. 1)
the iron atoms form bridges between new chemi-
cally identical independent ligands, in which the
CyNdouble bonds are not engaged in the
complexation ꢀ[3], structure determination for
R1 ˆ R2 ˆ p-CH3Ph). In case when R1 ± R2 struc-
tural isomers could be expected, but conforma-
tional selectivity has been reported for the complex
when R1 ˆ H and R2 ˆ CH3; which was isolated only
as isomer syn [4], however, this compound was not
directly obtained from acetaldazine.
In the reaction of iron dodecacarbonyl with ketazine
R1R2CyN 2 NyCR1R2 system bisꢀm2-ketoiminato)-
bisꢀtricarbonyliron) complexes are produced ꢀschematic
structure is shown in Fig. 1), formed by cleavage of
the N±N bond. In contrast, in the complexation of
similar benzaldazine system RꢀH)CyN±NyCꢀH)R,
where R ˆ Ph or para-substituted Ph, no cleavage of
the molecule was observed, and the reaction proceeds
* Corresponding author. Fax: 148-228231487.
E-mail address: axzimmi@farm.amwaw.edu.pl ꢀA. Zimniak).
0022-2860/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-2860ꢀ01)00604-4

212
A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
Fig. 1. Schematic structure of bis-iminato complexes obtained from azines in the reaction with iron dodecacarbonyl.
Nevertheless, from acetophenone azine where
1,4-Dimethyl-1,4-diphenyl-2,3-diazabuta-1,3-diene
ꢀsynthesis described in Ref. [6]) ꢀ1 mmol) and triiron-
dodecacarbonyl ꢀ3 mmol) were re¯uxed in isooctane
under nitrogen, chromatographed on Kieselgel 60 and
recrystallized from n-hexane [7]. The separation of
anti and syn isomers of the bis-ketoiminato complex
was accomplished by fractional recrystallization also
from n-hexane. The reagents used were pure and have
been purchased from Aldrich.
The X-ray data were collected on a Siemens P3
four-circle diffractometer at rt. The detailed crystal-
lographic data, parameters for data collection, re®ne-
ment procedures and parameters of molecular
structures of bis-ketoiminato anti I and syn II isomers
are described in Ref. [5]. The structural parameters
needed were acquired from the above X-ray data,
however, C±H bond lengths were normalized to the
values typical for neutron diffraction, i.e. to 1.083 for
Car±H in the phenyl rings and 1.059 for Csp3±H in the
R1 ˆ Ph and R2 ˆ CH3; two distinct anti and syn
structural isomers of the bis-ketoiminato complex
were obtained by one of us [5], showing in the crystal-
line state deviations from symmetry different for each
isomer.1 For the form anti, a quasi-two-fold axis
perpendicular to the Fe±Fe bond has been observed
ꢀC2 symmetry point group), whereas the form syn
shows a quasi-symmetry plane ꢀCs symmetry point
group). The aim of the work presented was the quanti-
tative estimation of deviations from symmetry in
compared isomers, the analysis of close contacts,
and the correlation of experimental and calculated
molecular properties obtained by X-ray structural
analysis and by DFT optimization, respectively.
Furthermore, it was taken into consideration that the
deviations from symmetry could be maintained in
solutions, and this was examined by NMR.
Ê
methyl groups ꢀin A) [8].
The 1H and 13C NMR spectra were recorded
with a Varian Gemini BB 200 MHz instrument in
d-chloroform.
2. Experimental and methods
Abbreviations used: NMR, nuclear magnetic reson-
ance; d-chloroform, deuterated chloroform; X-ray,
crystallographic structural analysis; DFT, density
functional theory; BLYP, nonlocal correlation func-
tional with gradient corrections to exchange; DNP,
double numerical basis set with polarization functions;
RMSD, root mean square displacement; vdW, van der
Waals; L±J, Lennard±Jones; rt, room temperature.
Complexation reactions with iron dodecacarbonyl
were carried out according to published methods.
The DFT calculations have been carried out by use
of dmol program [9]. BLYP functional [10,11] was
used in all the calculations along with DNP basis.
Electron density was analyzed by using Hirshfeld
charges [12]. The optimized structures and calculated
parameters are marked by asterisks ꢀ p ).
RMSD values were calculated for superimposed
symmetrical moieties using standard equation [13]
v
iˆ1
u
N
0 2
0 2
0 2
X
u
t
ꢀxi 2 xi † 1 ꢀyi 2 yi † 1 ꢀzi 2 zi †
1
Deviations from symmetry have been also studied by us in
RMSD ˆ
N
conformers of bisꢀ1,1-diphenylallenylidene)-octacarbonyltriiron
obtained by symmetric cleavage from tetraphenylhexapentaene
Ph2CyCyCyCyCyCPh2 ꢀwill be published separately).
where the superimposition is aimed at aligning N

A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
213
Fig. 2. View of the molecules I and II showing the atom numbering scheme. Close contacts are expressed as dashed and dotted lines, indicating
i2x
intra- and intermolecular interactions, respectively. Symbols
distinguish external molecules.

214
A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
atoms in two acetophenoniminato ligands of the same
molecule of the complex and xi, yi, zi represent the
3.1. Deviations from symmetry, close contacts and
repulsive forces in the solid state
spatial coordinates of atom i in one ligand while x0 ,
i
y0 , z0 represent the spatial coordinates of the atom i0
The study has been accomplished by use of the
structural X-ray data for I and II published earlier
by one of us [5]. All positions of atoms considered
were re®ned and the distances are given with the
proper margins of error.
In order to relate quantitatively the deviations from
symmetry in experimental structures of compared
isomers, the RMSD values have been calculated for
superimposed molecular moieties in the appropriate
operations of symmetry.
i
i
re¯ected in the operation of symmetry ꢀtwo-fold axis
in I and symmetry plane in II) in the other ligand.
In the analysis of close contacts the van der Waals
contact radii according to Bondi [14] have been
accepted: ꢀC) 1.70, ꢀH) 1.20, ꢀO) 1.52, ꢀN) 1.55 ꢀin
Ê
A). Distances between atoms separated by less than
four bonds were neglected.
Van der Waals repulsive interactions have been
approximated using Lennard±Jones ꢀ6±12) equation
for nonbond potential Eij [15]
Respectively, for I and II, results for all atoms are
0.1774 and 0.3286, and in case when hydrogen atoms
Ê
were omitted 0.1435 and 0.2228 ꢀin A). The values
"
!
! #
12
6
Rij
rij
Rij
rij
Eij ˆ eij
22
limited to the central region of the molecule, including
carbonyliron moiety, CyN, methyl groups and
quaternary aromatic carbons, are 0.0581 and 0.1022,
accordingly. These results clearly point to lower
symmetry of II with respect to I. Moreover, the
RMSD values falling on one pair of atoms are nearly
twice as high when all atoms are considered with
respect to the data for the central region, i.e. 0.0341
and 0.0155 for I, whereas the values for II are 0.0632
and 0.0273, accordingly. Thus, as expected, the
conformational irregularities are mainly placed in
the peripheral areas of the molecule.
In the analysis of molecular spacing the contact was
accepted as close when the interatomic distance was
smaller than the sum of van der Waals radii according
to Bondi [14] for the examined atoms ꢀsee Section 2
for details). Although there are no well-accepted
values for atomic contact radii, the Bondi's values
are acknowledged as one of the standards. It should
be stressed that in this work most critical is the esti-
mation of differences between two quasi-symmetric
ligands and not the determination of absolute values.
In the speci®cation of intramolecular contacts
several restrictions have been imposed. In case when
the difference between compared intramolecular
distances in two ligands was smaller than the three-
fold value of maximal error, this result was omitted.
Such limitation was essential for C±H bond lengths
where eij is the potential well depth for pair of atoms i
p
and atom j.
and j ꢀequal to eiiejj), Rij is the sum of vdW radii
ꢀRij ˆ Ri 1 Rj† and rij is the distance between atom i
Parameters eii and Ri for atoms C, Nand O were
taken directly from SYBYL force-®eld [16] ꢀfor these
atoms SYBYL vdW radii are equal to Bondi's radii).
For H atom parameter eH was also gained from
SYBYL force-®eld whereas RH decreased from 1.5
Ê
to 1.2 A, which is a value suggested by Bondi. This
modi®cation results from an observation, that the
value for RH in many force-®elds seems to be over-
estimated. Recent reparameterizations [17,18] based
on accurate quantum mechanical calculations tend to
lower its value down to the Bondi's standard.
3. Results and discussion
Intramolecular parameters taken into consideration
in both quasi-symmetrical ligands have been re¯ected
from one ligand to another according to the operation
of symmetry appropriate to the geometry of each of
compared isomers, i.e. two-fold axis was applied in I
ꢀanti) and a symmetry plane in II ꢀsyn), and the sum of
the differences in values obtained in this manner for
each isomer was analyzed. It should be stressed that
the intra- and intermolecular contacts of the O´´´H
type observed in isomers I and II cannot be classi®ed
as hydrogen bonds. Structures of I and II showing the
atom numbering scheme are presented in Fig. 2.
Ê
bearing relatively high errors of 0.03±0.04 A, while
errors of C±C bonds lengths usually did not exceed
Ê
0.004 A. Moreover, though iron atoms possibly
are involved in the intramolecular contacts, such

Table 1
Intramolecular close contacts and respective repulsive forces for I and II. Atom numbering scheme is given in Fig. 2, respective locations in ligands 1 and 2 are re¯ected by
molecular symmetry operations. Distances marked by ꢀ p ) result from DFT-calculations.
Distance D1p
Ê
ꢀA) calc.
Location in
ligand 2
Force F2
Ê
ꢀkJ/mol A)
Distance D2p
Ê
ꢀA) calc.
uD1p 2 D2pu
Ê
ꢀA) calc.
uD1 2 D2u
Ê
ꢀA)
uF1 2 F2u
Ê
ꢀkJ/mol A)
Ê
Distance D1 ꢀA)
Ê
Distance D2 ꢀA)
Location in
ligand 1
Force F1
Ê
ꢀkJ/mol A)
from X-ray
from X-ray
I
Cꢀ1b)±Cꢀ27)
Cꢀ1b)±Cꢀ22)
Cꢀ2b)±Hꢀ21c)
3.269 ꢀ12)
3.355 ꢀ3)
2.697 ꢀ35)
0.553
3.332
3.415
2.745
Cꢀ2b)±Cꢀ17)
Cꢀ2b)±Cꢀ12)
Cꢀ1b)±Hꢀ11c)
3.346 ꢀ17)
3.285 ꢀ22)
2.698 ꢀ32)
0.178
3.416
3.356
2.704
0.084
0.059
0.077
0.070
0.375
0.317
0.144
Omitteda
0.461
Omitteda
±
±
±
P
P
P
uD1p 2 D2pu
uD1 2 D2u
ˆ 0:147
uF1 2 F2u
ˆ 0:692
ˆ 0:143
II
Cꢀ1b)±Cꢀ21)
Cꢀ1b)±Hꢀ21b)
Cꢀ2a)±Cꢀ23)
Cꢀ2a)±Hꢀ23)
Cꢀ2b)±Cꢀ22)
Cꢀ2b)±Cꢀ23)
Cꢀ21)±Hꢀ27)
Hꢀ27)±Hꢀ21a)
Oꢀ2a)±Hꢀ23)
Cꢀ27)±Hꢀ21a)
3.379 ꢀ4)
2.472 ꢀ20)
3.330 ꢀ4)
2.710 ꢀ30)
3.238 ꢀ4)
3.437 ꢀ4)
3.088 ꢀ30)
2.519 ꢀ40)
2.667 ꢀ30)
2.881 ꢀ30)
0.062
3.449
2.537
3.412
2.792
3.351
3.521
3.127
2.551
2.770
2.917
Cꢀ1b)±Cꢀ11)
Cꢀ1b)±Hꢀ11c)
Cꢀ2c)±Cꢀ13)
Cꢀ2c)±Hꢀ13)
Cꢀ2b)±Cꢀ12)
Cꢀ2b)±Cꢀ13)
Cꢀ11)±Hꢀ17)
Hꢀ17)±Hꢀ11a)
Oꢀ2c)±Hꢀ13)
Cꢀ17)±Hꢀ11a)
3.305 ꢀ4)
2.558 ꢀ30)
3.431 ꢀ4)
2.566 ꢀ30)
3.407 ꢀ4)
3.290 ꢀ4)
2.826 ꢀ30)
2.341 ꢀ40)
2.708 ꢀ30)
2.928 ꢀ30)
0.358
3.422
2.672
3.481
2.639
3.493
3.354
2.835
2.328
2.752
2.938
0.027
0.135
0.069
0.153
0.142
0.167
0.292
0.223
±
0.074
0.086
0.101
0.144
0.169
0.147
0.262
0.178
±
0.296
2.567
0.243
2.028
0.757
0.434
0.234
0.166
±
5.706
3.139
No repulsionb
0.243
0.936
2.964
No repulsionb
0.434
0.757
No repulsionb
No repulsionb
No repulsionb
Omitteda
Omitteda
0.234
0.166
Omitteda
Omitteda,b
±
±
±
P
P
P
uD1p 2 D2pu
uD1 2 D2u
ˆ 1:161
uF1 2 F2u
ˆ 6:725
ˆ 1:208
a
The difference between corresponding distances in compared ligands was smaller than three-fold maximal error.
This distance exceeds the sum of vdW radii.
b

216
A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
interactions were neglected because the vdW radii of
six-coordinated iron atoms are not determined. Intra-
molecular short distances for both isomers meeting
above requirements are listed in Table 1.
calculated repulsive forces were summarized separately
for each moiety, including acetophenoniminato ligand
and the carbonyl groups, re¯ected by appropriate
operation of symmetry, and excluding the chain
Oꢀ1b)±Cꢀ1b)±Feꢀ1)±Feꢀ2)±Cꢀ2b)±Oꢀ2b) parting the
molecule ꢀresults are shown in Table 2). In rough
estimation, these values can be assumed as the extent
of deformation of each ligand caused by external
repulsions. For I the results for moieties 1 and 2 are
In the present analysis not only distances expres-
sing close contacts have been considered but also the
resulting repulsive forces were calculated according
to the L±J potential [15]. As one can see from Table 1,
in particular cases only one of the two compared
distances, quasi-symmetrically located in different
ligands, was identi®ed as short contact, while the
other in the pair exceeded the sum of vdW radii. In
such instances both distances were analyzed, but the
repulsive force corresponding to the latter interaction
was assumed as zero.
In I only two pairs of symmetrically arranged close
intramolecular contacts were considered, located
between carbonyl groups and adjacent phenyl rings
ꢀFig. 2). Carbonyl carbon atoms and also the quaternary
and ortho aromatic carbons are engaged into this
interactions. In contrast, in the molecule of II alto-
gether 14 interactions were noted, situated between
following groups: carbonyl±methyl, carbonyl±phenyl
and methyl±phenyl. Involved are carbon atoms from
carbonyls located between the phenyl rings or
between the methyl groups, quaternary carbon of the
phenyl ring, protons and carbons in positions ortho,
and also protons and carbons of the methyl substitu-
ents. As shown in Table 1, where all contacts are
speci®ed, the sum of differences ꢀabsolute values)
as follows ꢀquantities representing intermolecular
interactions are marked by primes): F10 3.306
P
P
ꢀ®ve contacts), F20 1.660 ꢀ®ve contacts) and the
P
P
0
difference u F10 2 F2 u is 1.646 ꢀin kJ/mol A).
Ê
P
The respective values for II are: F10 0.597 ꢀ®ve
P
P
contacts), F20 0.017 ꢀone contact), and u F10 2
P
F20u is 0.580. Thus, the external repulsive forces
are de®nitely more strongly expressed in I as
compared with II and also the difference in strength
of deforming interactions acting on each moiety is
considerably higher in I.
Based on the above results one can conclude that
from among repulsive interactions rather the intra-
molecular repulsions than the network stresses con-
tribute to higher deviations from symmetry, observed
in the molecule of II as compared with I.
3.2. DFT calculations
Full geometry optimization by the DFT method has
been performed in order to compare the electronic
energies for the isolated molecules of isomers I and
II and to note the differences between experimental
and theoretical conformations, which to some extent
could arise from the absence of the network inter-
molecular contacts in the calculated structures. To
start conformations those known from the X-ray
analysis were used [5].
The energies for Ip and IIp ꢀsymbols related to the
calculated quantities are marked by asterisks) are
practically equal, differing by 3.8 kJ/mol. As effect
of optimization no substantial changes in the overall
molecular geometry of I and II have been observed
ꢀselected bond lengths, valence and dihedral angles of
experimental and theoretical structures are compared
in Table 3). Contrarily, only limited modi®cations in
conformations were noted, e.g. after mathematical
superimposition of the structures determined by
X-ray and the calculated ones the positions of
between accordingly located analyzed distances
uD1 2 D2u is 0.147 in I, whereas in II it increases
by nearly an order of magnitude to 1.161 ꢀinA).
P
Ê
More adequate extent of steric crowding than only
close contacts are the resulting repulsive forces ꢀTable
1). Outstanding strong interactions of this type were
noted in II between the following atoms: Cꢀ1b)±
Hꢀ11c), Cꢀ1b)±Hꢀ21b) and Cꢀ2c)±Hꢀ13). The absol-
ute values of differences of respective forces localized
in both ligands were summarized giving uF1 2 F2u
0.692 in I, whereas in II the result was 10-fold as high
P
Ê
reaching 6.725 ꢀin kJ/mol A). Similar difference was
observed for close contacts. The noted inequality in
distribution of internal stresses in II could be consid-
ered as one of the possible reasons of deviations from
symmetry.
In the dislocation of intermolecular close contacts,
as expected, no symmetry was observed, thus, the

Table 2
Ê
Intermolecular and respective intermolecular repulsive forces for I and II. The O±H distances are given with maximal errors of 0.04 A. Close contacts were analyzed separately for
each quasi-symmetric moiety of the molecule exclusive of interactions with atoms Oꢀ1b)±Cꢀ1b)±Feꢀ1)±Feꢀ2)±Cꢀ2b)±Oꢀ2b) laying on an idealized symmetry plane in II and with
x
ix
v
Ê
respective atoms in I. ꢀFollowing contacts in I were excluded: Oꢀ1b)±Hꢀ11a) , 2.714; Oꢀ1b)±Hꢀ25) , 2.715; Oꢀ2b)±Hꢀ14) , 2.663 ꢀin A). No intermolecular contacts with atoms
laying on symmetry plane were found in II.) Symbols i2x distinguish external molecules.
0
0
Ê
Dist. ꢀA) from X-ray
Ê
Forces F1 ꢀkJ/mol A)
Ê
Dist. ꢀA) from X-ray
Ê
Forces F2 ꢀkJ/mol A)
Locations in moiety 1
Locations in moiety 2
I
Oꢀ1a)±Hꢀ24)i
Hꢀ11a)±Oꢀ1b)vii
Hꢀ11b)±Oꢀ1c)iii
Hꢀ14)±Oꢀ2b)v
Hꢀ16)±Hꢀ16)vi
2.550
2.714
2.582
2.663
2.172
0.958
0.018
0.681
0.203
1.446
Oꢀ1c)±Hꢀ11b)ii
Oꢀ1c)±Hꢀ25)viii
Hꢀ24)±Oꢀ1a)iv
Hꢀ25)±Oꢀ1b)iv
Hꢀ25)±Oꢀ1c)viii
2.582
2.719
2.550
2.715
2.719
0.681
0.003
0.958
0.015
0.003
P
P
P
F10 ˆ 3.306
F20 ˆ 1:660
P
P
u
F10 2 F20u ˆ 1:646
II
Oꢀ1a)±Hꢀ24)iii
Oꢀ1a)±Hꢀ27)v
Hꢀ23)±Cꢀ13)ii
Hꢀ24)±Oꢀ1a)iv
Hꢀ27)±Oꢀ1a)vi
2.649
2.712
2.893
2.649
2.712
0.267
0.023
Cꢀ13)±Hꢀ23)i
2.893
0.017
0.017
0.267
0.023
F10 ˆ 0.597
P
F20 ˆ 0:017
P
P
u
F10 2 F20u ˆ 0:580

218
A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
Table 3
p
p
Ê
Selected bond lengths ꢀA), valence and dihedral angles ꢀ8). Experimental data ꢀX-ray [5]) are given for I and II, theoretical for I and II . The
DFT-optimized structures are marked by ꢀ p )
I
Ip
II
IIp
Bond lengths
Feꢀ1)±Feꢀ2)
Feꢀ1)±Cꢀ1a)
Feꢀ1)±Cꢀ1b)
Feꢀ1)±Cꢀ1c)
Feꢀ2)±Cꢀ2a)
Feꢀ2)±Cꢀ2b)
Feꢀ2)±Cꢀ2c)
Feꢀ1)±Nꢀ1)
Feꢀ1)±Nꢀ2)
Feꢀ2)±Nꢀ1)
Feꢀ2)±Nꢀ2)
Nꢀ1)±Cꢀ10)
Nꢀ2)±Cꢀ20)
Cꢀ10)±Cꢀ11)
Cꢀ10)±Cꢀ12)
Cꢀ20)±Cꢀ21)
Cꢀ20)±Cꢀ22)
2.4137 ꢀ7)
1.784 ꢀ3)
1.805 ꢀ3)
1.780 ꢀ3)
1.784 ꢀ3)
1.800 ꢀ3)
1.791 ꢀ4)
1.918 ꢀ2)
1.925 ꢀ2)
1.918 ꢀ2)
1.915 ꢀ2)
1.256 ꢀ3)
1.256 ꢀ3)
1.500 ꢀ3)
1.489 ꢀ3)
1.503 ꢀ4)
1.491 ꢀ3)
2.5014
1.812
1.813
1.818
1.814
1.817
1.818
1.962
1.984
1.976
1.961
1.282
1.282
1.517
1.505
1.517
1.504
2.3990 ꢀ6)
1.790 ꢀ3)
1.804 ꢀ3)
1.786 ꢀ3)
1.783 ꢀ3)
1.828 ꢀ3)
1.778 ꢀ3)
1.924 ꢀ2)
2.919 ꢀ2)
1.924 ꢀ2)
1.925 ꢀ2)
1.263 ꢀ3)
1.254 ꢀ3)
1.501 ꢀ3)
1.491 ꢀ3)
1.499 ꢀ3)
1.491 ꢀ3)
2.4810
1.814
1.809
1.815
1.809
1.818
1.807
1.989
1.978
1.982
1.974
1.287
1.285
1.523
1.505
1.520
1.507
Bond angles
Nꢀ1)±Feꢀ1)±Nꢀ2)
Nꢀ1)±Feꢀ2)±Nꢀ2)
76.48 ꢀ8)
76.71 ꢀ7)
51.01 ꢀ5)
50.88 ꢀ5)
51.02 ꢀ5)
51.23 ꢀ5)
77.98 ꢀ7)
77.89 ꢀ7)
141.6 ꢀ2)
140.4 ꢀ2)
140.4 ꢀ2)
141.6 ꢀ2)
122.7 ꢀ2)
120.8 ꢀ2)
116.5 ꢀ2)
122.5 ꢀ3)
120.9 ꢀ2)
116.5 ꢀ2)
0.6 ꢀ2)
75.93
76.16
50.80
50.24
50.32
51.07
78.88
78.69
141.2
139.9
140.0
141.2
123.2
120.5
116.3
123.4
120.4
116.3
0.5
76.23 ꢀ8)
76.10 ꢀ8)
51.42 ꢀ5)
51.50 ꢀ5)
51.44 ꢀ5)
51.28 ꢀ5)
77.14 ꢀ7)
77.22 ꢀ7)
139.2 ꢀ2)
143.3 ꢀ2)
142.4 ꢀ2)
140.3 ꢀ2)
121.3 ꢀ2)
122.3 ꢀ2)
116.5 ꢀ2)
122.7 ꢀ2)
121.1 ꢀ2)
116.2 ꢀ2)
0.8 ꢀ2)
75.63
75.87
51.21
51.05
51.47
51.18
77.38
77.76
139.5
142.9
141.6
140.6
121.9
122.8
115.3
122.8
121.3
115.9
0.5
Nꢀ1)±Feꢀ1)±Feꢀ2)
Nꢀ2)±Feꢀ1)±Feꢀ2)
Nꢀ1)±Feꢀ2)±Feꢀ1)
Nꢀ2)±Feꢀ2)±Feꢀ1)
Feꢀ1)±Nꢀ1)±Feꢀ2)
Feꢀ1)±Nꢀ2)±Feꢀ2)
Cꢀ10)±Nꢀ1)±Feꢀ1)
Cꢀ10)±Nꢀ1)±Feꢀ2)
Cꢀ20)±Nꢀ2)±Feꢀ1)
Cꢀ20)±Nꢀ2)±Feꢀ2)
Nꢀ1)±Cꢀ10)±Cꢀ11)
Nꢀ1)±Cꢀ10)±Cꢀ12)
Cꢀ11)±Cꢀ10)±Cꢀ12)
Nꢀ2)±Cꢀ20)±Cꢀ21)
Nꢀ2)±Cꢀ20)±Cꢀ22)
Cꢀ21)±Cꢀ20)±Cꢀ22)
Cꢀ1)±Feꢀ1)±Feꢀ2)±Cꢀ2)a
Nꢀ1)±Cꢀ10)±Cꢀ12)±Cꢀ17)
Nꢀ2)±Cꢀ20)±Cꢀ22)±Cꢀ27)
2 85.6 ꢀ2)
2 81.2 ꢀ3)
2 86.4
2 83.4
128.2 ꢀ3)
2 112.0 ꢀ3)
128.1
2 113.3
a
Average calculated from sum of absolute values of dihedral angles de®ned by the carbonyls in facing ꢀsyn-periplanar) locations.
according atoms in the central core nearly overlapped.
Thus, for isomer anti the angle between the bonds
ꢀFe±CO) and respective ꢀFe±CO)p was 0.98 ꢀaverage
for three results), whereas analogous value for the
isomer syn was 2.18. Similarly, the difference
between experimental and calculated dihedral angles
Cꢀ11)±Cꢀ10)±Cꢀ12)±Cꢀ13) ꢀfor numbering of atoms
see Fig. 2) was 0.28 for anti, and 0.68 for syn.
The RMSD values, re¯ecting well the extent of
deviations from symmetry, were calculated for the
optimized molecules. Results for isomer anti and
syn are, respectively, as follows ꢀfor comparison, in

A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
219
Table 4
Comparison of Hirshfeld charges ꢀin e) for pairs of atoms de®ned by the operations of symmetry in optimized isomers Ip and IIp. Charges on
atoms Oꢀ1b)±Cꢀ1b)±Feꢀ1)±Feꢀ2)±Cꢀ2b)±Oꢀ2b) laying on an idealized symmetry plane in IIp and respective atoms in Ip were omitted
Pairs of atoms
Charges in Ip
uDuꢀIp
)
Charges in IIp
20.1263
uDuꢀIp)
Nꢀ1±2)
20.1251
0.0978
20.1253
0.0979
0.0002
0.0001
0.0002
0.0001
0.0003
20.1256
0.0007
Cꢀ1a±2a)
Oꢀ1a±2a)
Cꢀ1c±2c)
Oꢀ1c±2c)
Cꢀ1a±1c)
Oꢀ1a±1c)
Cꢀ2a±2c)
Oꢀ2a±2c)
Cꢀ10±20)
Cꢀ11±21)
Cꢀ12±22)
Cꢀ13±23)
Hꢀ13±23)
Cꢀ14±24)
Hꢀ14±24)
Cꢀ15±25)
Hꢀ15±25)
Cꢀ16±26)
Hꢀ16±26)
Cꢀ17±27)
Hꢀ17±27)
Hꢀ11a±21a)
Hꢀ11b±21c)
Hꢀ11c±21b)
Hꢀ11b±21b)
Hꢀ11a±21a)
Hꢀ11c±21c)
20.1112
0.0999
20.1078
20.1110
0.1000
20.1081
0.0995
20.1090
0.0991
0.0991
20.1094
0.1002
0.0004
0.0004
0.0011
0.0008
0.0010
0.0012
0.0001
0.0014
0.0023
0.0006
0.0006
0.0001
0.0000
0.0017
0.0008
0.0005
0.0019
0.0036
0.0010
0.0030
20.1080
0.0848
20.1088
0.0858
0.0850
20.0949
0.0000
0.0848
20.0949
0.0001
0.0002
0.0000
0.0001
0.0006
0.0006
0.0000
0.0000
0.0006
0.0003
0.0005
0.0004
0.0001
0.0005
20.0961
0.0004
20.0949
0.0003
20.0412
0.0434
20.0406
0.0440
20.0372
0.0364
20.0386
0.0387
20.0377
0.0466
20.0377
0.0466
20.0357
0.0493
20.0363
0.0487
20.0385
0.0473
20.0379
0.0476
20.0365
0.0476
20.0366
0.0476
20.0373
0.0475
20.0368
0.0479
20.0393
0.0461
20.0376
0.0469
20.0392
0.0456
0.0391
20.0393
0.0426
0.0475
20.0388
0.0445
0.0511
0.0451
0.0471
0.0461
0.0381
0.0351
0.0485
0.0492
0.0382
0.0496
0.0478
0.0394
0.0011
0.0014
0.0012
P
P
p
p
uDu
ˆ 0:0085
uDu
ˆ 0:0232
ꢀI †
ꢀII †
brackets the appropriate RMSD values for X-ray
experimentally determined structures are given): for
all atoms 0.1712p ꢀ0.1774) and 0.3235p ꢀ0.3286); after
excluding the hydrogen atoms 0.1411p ꢀ0.1435) and
0.2087p ꢀ0.2228); and for the central region including
Furthermore, the results related to calculated and
experimental short contacts, given by expressions
P
P
uD1p 2 D2pu and uD1 2 D2u ꢀdata is given in
Table 1, explanations of symbols see pt. 1.3.) are
0.143p and 0.147 for anti, respectively, and 1.208p
and 1.161 for syn. In general, most of the close
contacts observed by X-ray maintained after optimi-
zation their shorter distances than the sum of vdW
radii, although elongation also has been noted
ꢀTable 1). Since the calculations have been accomp-
lished for isolated molecules, these results con®rm the
earlier statement that the contribution of intramol-
ecular repulsions to the observed deviations from
symmetry is more signi®cant as compared with inter-
molecular forces.
Fe2ꢀCO)6, CyN , CH and aromatic CIV carbons
3
p
p
Ê
0.0604 ꢀ0.0581) and 0.0703 ꢀ0.1022) ꢀin A). These
results con®rm, that the symmerization of the mol-
ecule as a result of optimization was observed neither
for anti nor for syn, moreover, the magnitude of the
deviations from symmetry remained practically
unchanged after calculations. In fact, the lower
symmetry of II as compared with I was maintained
in the optimized isolated molecule. In both experi-
mental and calculated structures the conformational
irregularities are mainly placed in the peripheral
molecular regions.
Quasi-symmetrical ligands might be also compared
in terms of charges localized on respective atoms.

220
A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
Table 5
NMR shifts ꢀin ppm, d, spectra taken in d-chloroform) for I and II at different temperatures. Sharp singlets have been observed, unless otherwise
noted
Isomer/nuclei
CH3
Aromatic CH
Aromatic CIV
145.65a
CyN
I/13C
32.87a
33.30b
125.59a
128.41a
125.30b
128.32b
144.96b
145.25b
177.32a
176.68b
177.42b
I/1H
2.40a
2.38b
Multiplet
Multiplet
II/13C
32.83a
33.22b
125.32a
128.30a
128.41a
125.82b
128.32b
128.32b
145.77a
177.69a
II/1H
2.50a
2.46b
Multiplet
Multiplet
a
At 258C.
At 2708C.
b
Hirshfeld charges calculated for Ip and IIp are
collected in Table 4 ꢀatoms laying on the idealized
mirror plane of isomer syn and respective atoms in
anti were omitted). The absolute values of differences
in charges for atoms re¯ected in appropriate operation
of symmetry were summarized, giving 0.0085 for anti
and 0.0232 ꢀin e) for syn, what points on perceptibly
less regular dislocation of electron density in the
isomer syn. The differences in respective charges
also testify a relevant perturbations in symmetry in
the regions of methyl groups in both isomers, and
around the ortho positions in the phenyl rings in
isomer syn.
According to the NMR analysis, both isomers I and
II generally preserved high symmetry in solution,
although the multiplets of aromatic protons were not
regular in the spectra of both compounds. The simpli-
®cation of the sequence of aromatic carbon signals
was higher in I as compared with II, but in general
no spectral differentiation of ketoiminato ligands was
observed in both isomers. One can conclude, that in
the range of temperature from 25 to 2708C fast
conformational transitions resulted in symmetrization
of ligands and this was perceived in both isomers anti
and syn.
4. Conclusions
3.3. NMR investigation
Considering results obtained in the analysis of close
contacts an attempt to distinguish the ketoiminato
ligands by the NMR techniques in solution seemed
promising, especially regarding to more spatially
crowded isomer syn. As observed earlier [5], the
methyl groups of isomers I and II give at room
From among the two structural isomers of bisꢀm2-
acetophenoniminato)-bisꢀtricarbonyliron), the syn form
discloses in the solid state de®nitely higher deviations
from symmetry, as observed in comparison with
several experimental and calculated molecular para-
meters. Moreover, in syn the repulsing forces origi-
nating from intramolecular close contacts are
substantially higher and their distribution in related
ligands is less symmetric, whereas the intermolecular
repulsions are generally weaker in syn and also the
difference in their strength regarding each quasi-
symmetric part of the molecule is lower with respect
to anti. Thus, it might be supposed that the intra-
molecular repulsive interactions contribute more to
the deviations from symmetry in investigated isomers
with reference to similar intermolecular effects.
1
temperature in H NMR sharp singlets at 2.40 and
2.50 ppm, respectively, thus, no differentiation of
ligands was noted.
In this work the 13C resonances of the methyl
groups at rt were obtained at 32.87 and 32.83 ppm
for I and II, accordingly. Furthermore, the low-
1
temperature H and 13C NMR study was performed
at 2708C, but neither the splitting nor broadening of
the signals occurred, and the chemical shifts remained
nearly unchanged ꢀthe data is given in Table 5).

A. Zimniak, G. Bakalarski / Journal of Molecular Structure 597 42001) 211±221
221
The latter assumption was con®rmed by the DFT-
optimization of the isolated molecules. The overall
conformations of theoretical and X-ray crystal struc-
tures remained analogous and also the deviations from
symmetry were in general maintained in both isomers.
These differences could be quantitatively estimated
by root mean square displacement method and also
by comparison of calculated Hirshfeld charges in both
twin-ligands. Despite of higher steric crowding in syn
the calculated energies of both isomers are practically
equal, which allows the conclusion that the intra-
molecular repulsing forces are generally weak.
sity of Technology for cooperation. Computer resources
were provided by ICM, Warsaw University. Access
to Poland Countrywide License for MSI software
sponsored by State Committee for Scienti®c Research
is also acknowledged.
References
[1] P.E. Baike, O.S. Mills, Chem. Commun. ꢀ1966) 707.
[2] J. Zachara, A. Zimniak, Acta Cryst. C54 ꢀ1998) 353.
[3] D. Bright, O.S. Mills, Chem. Commun. ꢀ1967) 245.
[4] G. Gervasio, P.L. Stanghellini, R. Rossetti, Acta Cryst. B37
ꢀ1981) 1198.
An approach to distinguish the quasi-symmetric
1
ketoiminato ligands have been accomplished by H
[5] A. Zimniak, J. Zachara, J. Organomet. Chem. 533 ꢀ1997) 45.
[6] J. Elquero, R. Jacquier, C. Marzin, Bull. Soc. Chim. Fr. ꢀ1968)
713.
and 13C NMR in solution at room temperature and
also at 2708C, but the differentiation could be
observed neither in anti nor in syn, thus, a perfect
molecular symmetrization of each isomer proceeds
as effect of fast conformational transitions, taking
place also in syn despite the close molecular packing.
Conclusion can be made that in the studied complex
the deviations from symmetry originating primarily
from the intramolecular interactions are stabilized
by the network in solids and can also be analyzed
by quantum chemical calculations. Moreover, the
conformational analogy of structures determined
experimentally and optimized validates the adequacy
of the DFT method for structural analysis of this type
of organoiron compounds.
[7] A. Zimniak, Z. Buczkowski, Bull. Acad. Pol. Chim. 29 ꢀ1981)
231.
[8] International Tables for Crystallography, V. C., Kluwer
Academic Publishers, Dordchect, 1995, p. 696.
[9] DMOL 96.0, May 1998, San Diego, Molecular Simulations
Inc., 1998.
[10] C. Lee, W. Yang, R.G. Parr, Phys. Rev. B 37 ꢀ1988) 785.
[11] A.D. Becke, Phys. Rev. A 38 ꢀ1988) 3098.
[12] F.L. Hirshfeld, Theor. Chim. Acta B 44 ꢀ1977) 129.
[13] InsightII User Guide, December 1998, San Diego: Molecular
Simulation Inc., 1998.
[14] A. Bondi, J. Phys. Chem. 68 ꢀ1964) 441.
[15] SYBYL Theory Manual, Version 6.0, Tripos Associates, St.
Louis, MO, 1992.
[16] M. Clark, R.D. Cramer III, N. Van Opdenbosch, J. Comp.
Chem. 10 ꢀ1989) 982.
[17] D.L. Veenstra, D.M. Ferguson, P.A. Kollman, J. Comp.
Chem. 13 ꢀ1992) 971.
Acknowledgements
[18] N.L. Allinger, K.A. Durkin, J. Comp. Chem. 21 ꢀ2000)
1229.
We thank Dr Janusz Zachara from Warsaw Univer-