New heterometallic pivalates with FeIII and ZnII ions: Synthesis, structures, magnetic, thermal properties

Article abstract of DOI:10.1016/j.poly.2017.08.028

Techniques for synthesizing Zn(II)-Fe(III) heterometallic molecular complexes [Fe₂Zn₄O₂(Piv)₁₀] (1), [Fe₄Zn₃O₃(Piv)₁₂(H₂O)]·1.5Et₂O (2) and [Fe₃Zn₂O₂(Piv)₉(1,10-phen)] (3) (where HPiv is pivalic acid, 1,10-phen – C₁₂H₈N₂) have been developed. It has been found that recrystallization of the solid product formed from reaction mixture of Fe(NO₃)₃, Zn(NO₃)₂, KOH and HPiv (in the 3:2:13:34 ratio) in water (?Reagent A?) from acetone results in hexanuclear complex 1, whereas thermolysis of ?Reagent A? at 130 °C in air followed by crystallization of the thermolysis products from ether makes it possible to isolate complex 2. It has been found that dissolution of thermolyzed ?Reagent A? in MeCN in the presence of 1,10-phenanthroline leads to formation of complex 3. The structures of complexes 1–3 were determined by single crystal X-ray diffraction analysis. The doublet M?ssbauer spectra of complexes 1 and 2 correspond to high-spin Fe³⁺ ions in an octahedral environment of oxygen atoms, while the non-equivalence of the environment of iron ions in compound 3 (FeO₆ and FeO₄N₂) is reflected by two observed doublets with different intensities. Based on the results of magnetic data simulation and quantum-chemical calculations, it has been shown that antiferromagnetic exchange coupling exists in complexes 1 and 2 (J₁₂ = ?3.05 cm^?1 for 1 and J₁₂ = ?0.4, J₁₃ = ?13.5, J₂₃ = ?3.6, J₂₄ = ?27.2 and J₃₄ = ?30.1 cm^?1 for 2). The value of the exchange integral between pairs of metal ions correlates with the Fe-O distances and the Fe-O-Fe angles in the bridging oxo ligands. Processes of thermal destruction have been studied for complexes 1 and 2 in argon atmosphere in the temperature range of 25–500 °C.

Full text of DOI:10.1016/j.poly.2017.08.028

Accepted Manuscript
New heterometallic pivalates with FeIII and ZnII ions: synthesis, structures, mag-
netic, thermal properties
Irina A. Lutsenko, Mikhail A. Kiskin, Nikolay N. Efimov, Elena A. Ugolkova,
Yurii V. Maksimov, Vladimir K. Imshennik, Alexander S. Goloveshkin, Andrey
V. Khoroshilov, Anton S. Lytvynenko, Aleksey A. Sidorov, Igor L. Eremenko
PII:
S0277-5387(17)30559-4
DOI:
http://dx.doi.org/10.1016/j.poly.2017.08.028
POLY 12792
Reference:
Polyhedron
To appear in:
Received Date:
Revised Date:
Accepted Date:
10 July 2017
17 August 2017
23 August 2017
Please cite this article as: I.A. Lutsenko, M.A. Kiskin, N.N. Efimov, E.A. Ugolkova, Y.V. Maksimov, V.K.
Imshennik, A.S. Goloveshkin, A.V. Khoroshilov, A.S. Lytvynenko, A.A. Sidorov, I.L. Eremenko, New
heterometallic pivalates with FeIII and ZnII ions: synthesis, structures, magnetic, thermal properties, Polyhedron
(2017), doi: http://dx.doi.org/10.1016/j.poly.2017.08.028
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

New heterometallic pivalates with FeIII and ZnII ions: synthesis,
structures, magnetic and thermal properties
Irina А. Lutsenkoa,*, Мikhail А. Kiskina, Nikolay N. Efimova, Elena A. Ugolkovaa,
Yurii V. Maksimov b, Vladimir K. Imshennik b, Alexander S. Goloveshkinc, Andrey
V. Khoroshilova, Anton S. Lytvynenkod, Аleksey А. Sidorova, Igor L. Eremenkoa,c
a N.S. Kurnakov Institute of General and Inorganic Chemistry of the Russian Academy of
Sciences, Leninsky Prosp. 31, 119991 GSP-1, Moscow, Russian Federation
Fax (495) 952 1279. E-mail:irinalu05@rambler.ru
b Semenov Institute of Chemical Physics, Russian Academy of Sciences, Kosygina 4, 119991
Moscow, Russian Federation
с A.N. Nesmeyanov Institute of Organoelement Compounds, Russian Academy of Sciences,
Vavilova St. 28, 119991 Moscow, Russian Federation
d
L.V. Pisarzhevskii Institute of Physical Chemistry of the National Academy of Sciences of
Ukraine, Prospekt Nauki 31, 03028 Kiev, Ukraine
*irinalu05@rambler.ru
ABSTRACT
Techniques for synthesizing Zn(II)-Fe(III) heterometallic molecular complexes
[Fe2Zn4O2(Piv)10] (1), [Fe4Zn3O3(Piv)12(H2O)]·1.5Et2O (2) and [Fe3Zn2O2(Piv)9(1,10-phen)] (3)
(where HPiv is pivalic acid, 1,10-phen - C12H8N2) have been developed. It has been found that
recrystallization of the solid product formed from reaction mixture of Fe(NO3)3, Zn(NO3)2, KOH
and HPiv (in the 3 : 2 : 13 : 34 ratio) in water («Reagent A») from acetone results in hexanuclear
complex 1, whereas thermolysis of «Reagent A» at 130 °C in air followed by crystallization of
the thermolysis products from ether makes it possible to isolate complex 2. It has been found that
dissolution of thermolyzed «Reagent A» in MeCN in the presence of 1,10-phenanthroline leads
to formation of complex 3. The structures of complexes 1-3 were determined by single crystal X-
ray diffraction analysis. The doublet Mössbauer spectra of complexes 1 and 2 correspond to
high-spin Fe3+ ions in an octahedral environment of oxygen atoms, while the non-equivalence of
the environment of iron ions in compound 3 (FeO6 and FeO4N2) is reflected by two observed
doublets with different intensities. Based on the results of magnetic data simulation and
quantum-chemical calculations, it has been shown that antiferromagnetic exchange coupling
1

exists in complexes 1 and 2 (J12 = -3.05 cm-1 for 1 and J12 = -0.4, J13 = -13.5, J23 = -3.6, J24 = -
27.2 and J34 = -30.1 cm-1 for 2). The value of the exchange integral between pairs of metal ions
correlates with the Fe-O distances and the Fe‒O‒Fe angles in the bridging oxo ligands. Processes
of thermal destruction have been studied for complexes 1 and 2 in argon atmosphere in the
temperature range of 25-500 °C.
Keywords: pivalate complexes iron(III) and zinc, crystal structure, magnetic properties,
mӧssbauer spectroscopy, simultaneous thermal analysis, DFT.
1. Introduction
In the past decades, special attention of researchers was focused on various polynuclear
coordination compounds, from molecules with two metal centers to multicenter systems,
including coordination polymers. Heterometallic molecules have a special place among these
objects, since, by varying the nature of metals in the metal core of these structures, it becomes
possible not only to change the geometric and electronic characteristics of molecules, but also to
control the physicochemical properties of crystals containing such molecules. Furthermore,
polynuclear complex compounds containing two or more metal ions of different nature that
manifest unique catalytic, sorption, magnetic, optical and/or electrically conductive properties
depending on their spatial and electronic structure [1-7] are promising components in the
creation of new-generation functional molecular materials or convenient precursors for directed
thermal synthesis of inorganic materials, including nanomaterials [8-12].
In this study, we focused on the development of synthesis methods of new heterometallic
{FeIII‒ZnII} carboxylate compounds, in which the presence of a magnetic center, Fe(III) (S =
5/2), ensures the manifestation of magnetic properties. We expect that varying the composition
and configuration of the {FenZnm} metal core, in which zinc(II) ions (d10) play the role of
diamagnetic diluents, would allow us to find means for controlling the magnetic characteristics
of such objects.
Despite relatively large amount of polynuclear compounds with iron and zinc ions are
known (370 structures according to CCDC data [(CSD version 5.38 updates (May, 2017)] and
the majority of them belong to organometallic and ionic compounds [13-24], we believe that the
most promising examples in terms of magnetic properties are represented by molecules in which
iron and zinc atoms are bound by carboxylate bridges, which formally allows one to change the
magnetic characteristics of such systems by varying the number of magnetic centers in the
2

molecule (e.g., FeIII ions), their positions in the metal core, as well as the composition of their
coordination sphere that, as a rule, noticeably affects the magnetic characteristics of objects. A
search for such heteronuclear {Fe-Zn} carboxylates in the CCDC database [(CSD version 5.38
updates (May, 2017)] allowed us to identify only a few compounds, viz., trinuclear complexes
with linear and triangular structures, ZnII FeII(OAc)6(py)2 [25] and ZnIIFeIII2O(O2CR)6(py)3 (R =
2
CH3 [25]; R = CCl3 [26]; R = CF3 [27]) respectively. Most likely, this result can be explained by
the small number of studies in this area. Perhaps, the low activity of researchers in this field of
chemistry is, to some extent, due to problems in the synthesis of such polynuclear molecules,
which naturally hinders the development of the chemistry and physics of heterometallic objects
of this kind.
As concerns synthetic approaches to the design of heterometallic carboxylates with ions of
iron and other metals, let us note that we have recently found new methods for the design of such
systems by combining known methods of solid state chemistry and solution methodologies. For
example, we obtained a series of heterometallic pivalates with iron and manganese ions using a
method involving, at the first stage, mild thermolysis in various atmospheres (argon, air, 90 °С)
of a solid intermediate that appears upon mixing the starting compounds, Fe(NO3)3 and
Mn(NO3)2, as well as KOH and pivalic acid in aqueous medium. It was then extracted by organic
solvents [28]. We used a similar methodology it this study. The only difference was that we used
zinc nitrate instead of manganese nitrate mentioned above, and also used 1,10-phenanthroline as
an additional substrate for thermolyzed «Reagent A» (product formed from Fe(NO3)3, Zn(NO3)2,
KOH and HPiv (in the 3 : 2 : 13 : 34 ratio) in water) in MeCN medium in one of the
experiments.
2. Experimental
2.1. Materials and methods
Commercial reagents and solvents were used for the synthesis: trimethylacetic acid (HPiv,
99%, «Merck»), KOH (analytical grade), Fe(NO3)3·9H2O (analytical grade), Zn(NO3)2·6H2O
(analytical grade), diethyl ether, acetone, 1,10-phenanthroline («Alfa Aesar»).
Attenuated total reflection infrared (ATR-IR) spectra were recorded in the 400–4000 cm-1
region using a Spectrum-65 Perkin Elmer FT- IR spectrometer. Microprobe analyses were
carried out using an Carlo Erba EA 1108 Series CHN Elemental Analyser (Center of Collective
Use of IGIC RAS. Solid); for the analysis on metals we used a mass spectrometric method with
inductively coupled plasma (ICP-MS) (Agilent 7500ce; Agilent Technologies Inc., USA).
3

2.2. Synthesis
2.2.1. Synthesis of [Fe2Zn4(O)2(Piv)10](1)
HPiv (30 g, 0.29 mol) and KОН (6 g, 0.11 mol) were dissolved in 40 ml of water at 60 °С.
Solutions of salts Fe(NO3)3·9H2O (10 g, 0.025 mol) and Zn(NO3)2·6H2O (5 g, 0.017 mol) were
added to the resulting homogeneous solution. The mixture was stirred to form an oily layer (ca.
15 min). This layer was separated from the solution by decantation and extracted with diethyl
ether. The resulting extract, after evaporation of the solvent, gave a solid mass that was washed
with MeCN (to remove pivalic acid residues), and the resulting amorphous reddish brown
powder («Reagent A») was dissolved in acetone and kept for a day at 20 °С. The precipitated
yellow crystals 1 suitable for X-ray diffraction study were separated from the solution by
decantation, washed with cold MeCN (0 °C.), and air dried. The yield was 35% (based on the
initial iron nitrate). For 1, Anal. calc. for C50H90Fe2Zn4O22: C, 42.39; H, 6.39. Found: C, 42.20;
H, 6.87; the ratio of metals according to ICP is Fe0.24Zn0.46; IR, ν/cm–1 (s = strong, m = medium,
w = weak): 2962 m, 2930 w, 2874 w, 1524 m, 1482 vs, 1422 vs, 1378 m, 1360 m, 1227 vs, 1031
w, 937 w, 898 w, 845 m, 787 m, 603 br. m, 512 br. m, 435 m.
When the powder is extracted with diethyl ether, it is possible to isolate a mixture of two
complexes, the previously reported [29] triuclear iron(III) [Fe3(O)(Piv)6(H2O)3]·Piv (brown) and
hexanuclear complex [Fe2Zn4(O)2(Piv)10] (1) (yellow). Both complexes are isolated from the
solution in the form of single crystals, which makes it possible to separate compound 1 by the
Pasteur method [30]. Recrystallization of a mixture of crystals from hexane results in the
crystallization of the main complex 1, although it contains a small amount of triangular
[Fe3(O)(Piv)6(H2O)3]·Piv (less than 4%), which is confirmed by the powder pattern (see Fig. S2)
and elemental analysis data.
A sample of «Reagent A» (2.5 g) was kept in a porcelain beaker at 130 °C in air for 120
minutes, with shaking of the contents at 30 minute intervals. The resulting red powder of
thermolysed «Reagent A» was dissolved in acetone at 40 °C. After 24 hours, yellow crystals of
crystallographic modification 1 formed. Thus, both crystallographic modifications are always
present, but in different ratios - 98% (C2/c) / 2% (P21/c) (for the crystallization of non-
thermolysed «Reagent A») and 88% (C2/c) / 12% (P21/c) (upon crystallization of the
thermolyzed «Reagent A») according to PXRD data.
2.2.2. Synthesis of [Fe4Zn3(O)3(Piv)12(H2O)]·1.5Et2O (2)
4

The thermolysed «Reagent A» was dissolved in diethyl ether (30 ml, +5 oC). After 48
hours, the resulting brown crystals of heptanuclear complex [Fe4Zn3(O)3(Piv)12(H2O)]·1.5Et2O
(2), were washed with cold acetonitrile (5 oC) and dried in air. The yield was 30%. According to
the powder pattern, the crystals were homogeneous. (Experimental and simulated powder X-ray
diffractograms are reported in Fig. S2).
For 2, Anal. calc. for C66H125Fе4Zn3O29.5: C, 43.77; H, 6.53. Found: C, 43.75; H, 6.85; the
ratio of metals according to ICP is Fe0.59Zn0.37; IR, ν/cm–1: 2962 m, 2929 w, 2870 w, 1556 br. vs,
1482 vs, 1419 vs, 1378 m, 1358 m, 1226 vs, 1032 w, 937 w, 898 w, 787 m, 586 br. m, 503 br. m,
435 m.
2.2.3. Synthesis of adduct [Fe3Zn2(O)2(Piv)9(1,10-phen)] (3)
A sample of «Reagent A» (100 mg, 0.07 mmol) thermolyzed at 130 °C and 1,10-
phenanthroline (13 mg, 0.07 mmol) were dissolved in 40 ml of acetonitrile (70 ºС, 40 min) and
the resulting brown solution was aged for two days at room temperature. The precipitated brown
crystals were separated from the solution by decantation, washed with cold MeCN (0 oC) and
dried in air. For 3, Anal. calc. for C57H89Fе3Zn2O20N2: C, 49.19; H, 6.09; N, 2.28. Found: C,
48.63; Н, 6.26; N, 1.96. IR, ν/cm–1: 2960 m, 2929 w, 2869 w, 1575 br. s, 1562 br. s, 1518 w,
1481 vs, 1417 vs, 1373 br. m, 1359 br. m, 1308 w, 1225 vs, 1142 br. w, 1107 w, 983 m, 934 w,
885 w, 870 w, 848 w, 785 w, 750 w, 727 w, 697 br. w, 646 w, 603 br. m, 498 br. m, 427 s.
2.3. Physical measurements
2.3.1. Crystal structure determination
Single crystal X-ray studies of compounds 1 – 3 were carried out on a Bruker Apex II
diffractometer equipped with a CCD detector (MoKα, λ = 0.71073 Å, graphite monochromator)
[31]. A semiempirical adjustment for absorption was introduced for all the complexes [32]. The
structures of the compounds obtained were solved by direct methods and refined in the full-
matrix least-squares anisotropic approximation using the SHELX2014 software complexes [33].
The hydrogen atoms in the ligands were calculated geometrically and refined in the «riding»
model. Determination of the tert-butyl substituents at the carboxylate group in compounds 1-3
and the carboxylate group in 3 was performed with disordering taken into account. The positions
of methyl carbon atoms in the {СМе3} moieties (the populations of the carbon atoms were 0.5 in
1, 0.689(11) in 2, 0.654(14) and 0.346(14) in 3, respectively) and one oxygen atom in {О2С}
moiety (the population of the О16 and О16А atoms in 3 were 0.48(2) and 0.52(2), respectively)
were localized from differential Fourier synthesis. Structures of complexes 1 (C2/c and P21/c)
5

and 3 contain solvent accessible voids of 338 Å3, 198 Å3 and 195 Å3, respectively, but Fourier
maps do not have intensive peaks that could be assigned to solvent molecules. Observed
hydrophobic voids are small (maximum diameters of voids are 1.7 Å for 1 (C2/c and P21/c) and
2.2 Å for 3) to include small molecules such as water (d(H2O) = 2.75 Å) and others. The
crystallographic parameters of the complexes and structure refinement details are presented in
Table 1.
Table 1
Selected crystal data and parameters for structure refinement of the title compounds 1–3
1
1
2
3
Empirical
formula
Formula
weight
Т, К
Crystal system
Space group
a (Å)
b (Å)
c (Å)
α (º)
β (º)
C50H90Fe2Zn4O22
C50H90Fe2Zn4O22 C66H125Fе4Zn3O29.5
C57H89Fе3Zn2O20N2
1416.40
1416.40
1810.16
1420.59
296(2)
monoclinic
C2/c
13.0172(7)
29.6969(16)
18.5595(10)
90.00
106.034(1)
90.00
296(2)
monoclinic
P21/c
150(2)
triclinic
P-1
296
monoclinic
P21/n
13.9388 (6)
24.7994 (11)
21.3332 (9)
90.00
97.7202 (8)
90.00
13.371(4)
22.050(6)
24.322(7)
90.00
104.746(6)
90.00
14.4767(13)
15.5740(14)
22.393(2)
84.455(2)
74.863(2)
63.7460(10)
4369.9(7)
2
γ (º)
V(Å3)
6895.4(6)
4
1.364
6935(3)
4
7307.5 (5)
4
1.291
Z
3
–
Dcalc/ (g·cm )
1.357
1.376
1
–
1.843
1.83
1.526
1.291
µ (mm )
1.4–27.5
1.3–27.1
–17→17
–28→28
–31→31
0.838/0.930
2952
0.9–29.1
2.1–26.9
θ range (º)
Range of h, k
and l
→16
→19
→17
–16
–37→38
→24
–19
–21→21
→30
–17
–23→31
→21
–17
–30
–27
0.590/0.867
2952
0.750/0.928
1902
0.668/0.915
2972
Тmin/Тmax
F(000)
Number of
parameters
Rint
368
703
927
780
0.086
1.03
0.045
0.081
1.05
0.063
0.079
1.050
0.062
0.136
0.085
1.056
0.077
0.181
GООF
R1(I > 2σ(I))
wR2
0.102
0.140
(I > 2σ(I))
2.3.2. X-ray powder diffraction
6

The powder pattern was measured on Bruker D8 Advance diffractometer with LynxEye
detector and Ge (111) monochromator, λ(CuKα1) = 1.5406 Å, θ/2θ scan from 4° to 40°, step
size 0.02°. The measurement was performed in transmission mode, the compounds were
deposited between two Mular films. The powder pattern refinement was performed in TOPAS 5
software using low-temperature single-crystal structures. The preferred orientation was described
with 4th order spherical harmonics.
2.3.3. Mӧssbauer spectra
The Mӧssbauer spectra 57Fe of 1 - 3 were obtained on a Wissel electrodynamic type
spectrometer (Germany) in the temperature range of 16-300 K using a Janis (CCS-850) helium
cryostat with a Lake Shore Cryotronics temperature controller (332). The accuracy of
maintaining the temperature was not less than 0.1 K. The activity of 57Co(Rh) as a source of
Mӧssbauer radiation was 1.1 GBk. The isomer shifts were measured from the center of the
magnetic hyperfine structure (STS) of metallic iron. The spectra were processed by standard
counting and simulation programs for the Mӧssbauer transition 3/2 → 1/2. Complex processing
of Mӧssbauer spectra was carried out by the least squares method using LRT programs -
Semenov Institute of Chemical Physics of the Russian Academy of Sciences and WINNORMOS
(Germany).
2.3.4. Magnetic properties
Magnetic susceptibility measurements of 1 and 2 were performed using a Quantum Design
PPMS-9 susceptometer. The instrument operates between 2 and 300 K for dc applied fields of
5000 Oe. The magnetic data were corrected for the sample holder and diamagnetic contribution.
2.3.5. DFT calculations
All quantum chemical calculations were provided using ORCA 4.0.0.2 software [34]
applying spin-unrestricted density functional theory (DFT) with B3LYP [35] exchange-
correlation (XC) potential (unless specified otherwise) and def2-SVP [36] or def2-TZVP [36]
basis sets. In order to reduce computational efforts, the resolution of the identity (RI) [37] and
“chain-of-spheres” (COSX) [38] approximations for Coulomb and exchange parts of the Fock
matrices were applied, as implemented in ORCA, unless specified otherwise. Def2/J [39]
auxiliary basis set was used for these approximations.
For comparison, the exchange interaction within {Fe2Zn4}Ac species (vide infra) was
also calculated using PBE0 [40], BHandhLYP [41], M06 [42], TPSSh [43,44] XC potentials
7

instead of B3LYP; as well as by using modified B3LYP functional with exact (Hartree-Fock)
exchange fraction increased to 35 %.
Geometries.
{Fe2Zn4}Ac and {Fe4Zn3}Ac. The geometries of the species describing 1 and 2,
respectively, were prepared in the following manner. Crystallographic data were used to generate
the initial geometries. As the positions of the atoms of the pivalate groups were mostly
disordered, and in order to reduce computational costs, the Piv--groups were removed and
replaced by acetate groups manually added via Avogadro [45] 1.1.1 software. The positions of
atoms of the acetate groups were optimized via molecular mechanics with UFF [46] force field
(as implemented in Avogadro software on the basis of OpenBabel [47] library) keeping the rest
of the atoms frozen. The resulted species, {Fe2Zn4}Ac and {Fe4Zn3}Ac, were used for
calculation of the exchange integrals.
{Fe2Zn4}Acopt and {Fe2Zn4}Pivopt. For comparison, we provided more advanced
procedure for preparation of the species describing 1 instead of {Fe2Zn4}Ac. {Fe2Zn4}Acopt
was prepared from {Fe2Zn4}Ac by constrained DFT optimization utilizing def2-TZVP basis set
and PBE [48,49] XC potential (only positions of atoms of the acetate groups were optimized):
pure-GGA PBE XC potential was chosen instead of hybrid XC potentials utilized in the rest of
the calculations in order to reduce computational efforts). {Fe2Zn4}Pivopt was prepared
similarily to {Fe2Zn4}Acopt, but pivalate groups instead of acetate groups were manually added
to the structure, and def2-SVP basis was used instead of def2-TZVP to reduce computational
cost.
Exchange integrals.
The exchange integrals between FeIII ions in both {Fe2Zn4}Ac and {Fe4Zn3}Ac species
were calculated via broken symmetry DFT approach, as implemented in ORCA. Fine grids and
tight convergence criteria (enforced by «Grid7 NoFinalGrid GridX9 VeryTightSCF» ORCA
instructions) were applied in order to avoid errors in J values originating from the numeric noise.
{Fe2Zn4}Ac and {Fe2Zn4}Acopt. The exchange integral between two Fe ions in
{Fe2Zn4}Ac and {Fe2Zn4}Acopt was found by converging the single point DFT calculation to
the high-spin (multiplicity equal to 11, which corresponds to MS = +5/2 for both ions) state and
then reconverting to the broken symmetry state constructed by ORCA via flipping the spins on
one of the Fe ions («BrokenSym 5,5» ORCA instruction leading to the singlet broken symmetry
state). The exchange integral was calculated from the energies of both states using the equation
by Yamaguchi et al. [50,51]:
EHS − EBS
J = −
2
2
ˆ
ˆ
− S
S
HS
BS
8

{Fe2Zn4}Pivopt. The exchange coupling parameter between two Fe ions in
{Fe2Zn4}Pivopt was found with two levels of precision:
Level 1. Same procedure as for {Fe2Zn4}Ac and {Fe2Zn4}Acopt.
Level 2. Same as level 1, but Def2-TZVP basis set used only for Fe, Zn and O atoms,
Def2-SVP used for the rest of atoms.
{Fe4Zn3}Ac. Two approaches were utilized in order to calculate the exchange integrals
between four Fe3+ ions in the {Fe4Zn3}Ac species.
Approach 1. The exchange integrals were calculated separately for each selected pair of
ions by the same approach used for {Fe2Zn4}Ac after substitution of the rest of Fe3+ ions by
diamagnetic Sc3+ ones.
Approach 2. In this approach, the DFT calculation was first converged to the high-spin
state (multiplicity equal to 21, which corresponds to MS = +5/2 for all FeIII ions). After that, 6
different broken symmetry calculations where provided, flipping different spins or pairs of spins
−5/2 states by ORCA
from MS =+5/2 to MS =
«FlipSpin» directive with appropriate FinalMs
value («5.0» if only one spin was flipped, and «0.0» if two spins were flipped). The ions were
numbered basing on their numbers in the CIF file, the exchange interactions were numbered after
the numbers of the involved ions.
Each spin was involved in a few exchange interactions, thus, flipping each spin led to
complicated change of the system’s energy. In order to take this into account, the full-matrix
spin-hamiltonian approach was utilized following the procedure similar to the previously
j>i
ˆ ˆ
j
ˆ
reported one [52]. The exchange part of the system’s spin-hamiltionian
was
H = −2 J S S
∑
ex
ij
i
i, j
decomposed over a set of basis functions constructed by taking all possible combinations of MS
projections of the spins of FeIII ions: ϕ = MS1;MS 2 ;MS 3;MS 4 . The high-spin solution
5
5
5
5
5
5
5
5
2
ˆ
corresponded to + ;+ ;+ ;+ Hex + ; + ;+ ;+
element of the spin-hamiltonial
2
2
2
2
2
2
2
full matrix, the broken-symmetry solution obtained by flipping the first spin as described above
5
5
5
5
5
5
5
5
2
ˆ
corresponded to the − ;+ ;+ ;+ Hex − ; + ;+ ;+
element and so on. The
2
2
2
2
2
2
2
analytical forms of such elements expressed via the Jij exchange integrals were generated using
Mjöllnir 0.3 software [53,54] (variable «dump_full_matrices» must be set to «1» in the config.py
file of the program in order to generate the explicit view of the eigenvectors and the matrix
elements). The system of equations resulted from these expressions and the DFT results (Table
1s, Equation 1s) was solved via Octave 4.0.0 software [55].
9

2.3.6. Thermal analysis
The thermal behavior of complexes 1 and 2 was studied using the simultaneous thermal
analysis (STA) technique for parallel recording of TG (thermogravimetry) and DSC (differential
scanning calorimetry) curves. The study was performed on an STA 449 F1 Jupiter instrument
(«NETZSCH») in Al-crucibles under a lid with a hole to ensure a vapour pressure of 1 atm
during the thermal decomposition of the samples. The rate of heating to 500 ºC was 5 ºC min-1
(1) / 10 ºC min-1 (2) under an argon atmosphere (Ar content > 99.998%, O2 < 0.0002%, N2 <
0.001%, water vapor < 0.0003%, СН4 < 0.0001%). The sample mass was in the range of 1.32–
2.88 mg. The accuracy of temperature measurements was 0.7 ºC, and that of mass changes was
1 ˟ 10-2 mg.
3. Results and discussion
3.1. Synthesis of compounds
We synthesized new {Zn-Fe} pivalates by the previously reported approach used to the
synthesis of heteronuclear molecules with iron(III) and manganese(II/III) atoms [28], as well as
iron(III) and cobalt(II) atoms [CoIIFeIII O(Piv)6(HPiv)3] [9]. The main idea here was to combine
2
synthetic methods of solid-state and solution chemistry. In fact, simple reagents, such as salts of
various metals, an alkali, and pivalic acid were dissolved in water at the initial stage. As a rule,
stirring of such systems resulted in an «organic oily layer» that was separated from the solution.
Washing this oily product with organic solvents (e.g., ether and acetonitrile) to remove the
excess pivalic acid gives solid powders that possibly consist of mixtures of pivalate complexes
or polymeric compounds. Direct recrystallization of these powders from donor organic solvents
results in small polynuclear structures with two or three metal centers [see, e.g., 28, 9]. However,
if the powders obtained are subjected to mild thermolysis (100 – 150 оС) in order to remove the
coordinated terminal molecules of HPiv (and in a general case, any other non-bridging organic
component) present in the complexes, then the coordination deficiency that appears on the metal
centers can be expected to cause a growth of the metal core and various rearrangements in it. We
observed a similar situation previously in the formation of an octanuclear cobalt cluster
[Co8O2(Piv)12] obtained by thermal treatment of tetranuclear complex [Co4(OH)2(Piv)6(EtOH)4]
at 170 oC [56] with removal of the coordinated ethanol molecules, followed by binding of two
tetranuclear moieties with liberation of two water molecules.
10

In this study, mixing Fe(III) and Zn(II) nitrates (Fe : Zn ratio = 4 : 3) in the presence of
excess KOH and HPiv by the procedures described above gave a reddish-brown powder
(designated as «Reagent A» in Experimental). Extraction of the latter with Et2O (or MeCN) at
room temperature followed by crystallization allowed us to isolate hexanuclear complex
[Fe2Zn4(O)2(Piv)10] (1) as yellow crystals. It should be noted that a complex with similar
composition was obtained previously by the stoichiometric reaction of [Fe3(O)(Piv)6(HPiv)3]·Piv
and [Zn(Piv)2]n (the reagents were taken in amounts providing the ratio Fe : Zn = 1 : 2) [57],
though its structure was not confirmed by X-ray single crystal analysis. In this study we
succeeded in growing single crystals of 1 and characterizing them by single-crystal X-ray
diffraction (Fig. 1, Table 1, two crystallographic modifications where the molecules of complex
1 had similar geometric characteristics) (see Section 2.2.1).
Subsequently, «Reagent A» was subjected to solid-phase thermolysis in order to remove
the molecules of low-boiling substances. The thermolysis was carried out in air at various
temperatures, heating rates and exposure times. The thermolysis products were monitored by
recording IR spectra (the removal of neutral pivalic acid molecules was mainly monitored by the
bands around 1702 and 1303 cm-1 corresponding to vibrations of the OH-groups in HPiv
molecules. These bands absent in the spectra of «Reagent A» after thermolysis that corresponds
to desorption of HPiv from the complex (Fig. S1a,b)). As a result, the following optimum
thermolysis parameters were found: heating to 130 ºС; heating rate 10 deg/min in air; exposure
for 90 min at the maximum temperature. (∆m = 26.5%). After thermal treatment, the product was
dissolved in Et2O and crystals of the [Fe4Zn3O3(Piv)12(H2O)]·1.5Et2O complex were isolated
from the solution (2).
According to X-ray powder diffraction data, the final reaction products contained up to
13% of admixtures that we failed to remove by recrystallization. (Experimental and simulated
powder X-ray diffractograms are shown in Fig. S2).
We also studied the reactivity of the product of thermal treatment of «Reagent A» toward
the N-donor chelating ligand, viz., 1,10-phen. As a result, extraction of a mixture of «Reagent A»
and 1,10-phenanthroline with acetonitrile gave pentanuclear complex
[Fe3Zn2O2(Piv)9(1,10-phen)] (3).
IR spectroscopic data for the starting compounds and complexes 1-3 show that they are
individual compounds (Fig. S1c-e) contained carboxylate groups. The appearance of new signals
around 983–499 cm-1 in complex 3 indicates that the 1,10-phenanthroline molecule is bound to
the metal center (Fig. S1е) [58].
3.2. Single crystal X-Ray structure of 1-3
11

Complex 1 crystallizes in monoclinic system, in groups C2/c and P21/c. Both modifications
have similar geometric characteristics (Table 1). In a crystal with C2/c group, the C2 axis passes
through the O1M and O2M atoms (Fig. 1a; the main bond lengths are listed in Table 2).
Table 2
Selected bond lengths (Å) and angles for complexes 1–3
Complex/
1
2
3
parameter
Bond, d (Å)
Zn–O(µ-O)
C2/c
P21/c
1.955(2)– 1.952(4)– 1.930(3)–1.970(3)
1.957(2) 1.971(4)
1.979(2)– 1.977(4)– 1.842(3)–2.047(3)
1.984(2) 1.984(4)
1.933(4)–1.938(4)
1.823(4)–1.959(4)
Fe–O(µ-O)
Zn–O(Piv)
Fe–O(Piv)
1.921(3)– 1.916(4)– 1.910(4)–2.022(4)
1.979(3) 1.978(5)
1.925(4)–1.981(5)
1.999(3)– 1.978(3)– 1.965(4)–2.160(4)
2.044(3)
–
1.967(1)–2.067(4)
2.144(5)–2.200(6)
2.051(4)
–
Fe–N
–
Angle, ω (deg)
Fe–µ–O–Fe
98.8(2)
98.5(2)
98.3(2)
98.4(2)
Fe1–O1M–Fe2 97.9(1)
Fe1–O2M–Fe2 96.8(1)
Fe1–O2M–Fe3 121.2(1)
Fe2–O2M–Fe3 92.9(1)
Fe2–O3M–Fe3 99.8(1)
Fe2–O3M–Fe4 128.1(2)
Fe3–O3M–Fe4 130.8(2)
Fe1–O1M–Fe2
Fe2–O1M–Fe3
Fe3–O1M–Fe1
127.6(2)
99.9(2)
120.7(2)
The metal core of the complex is a {Fe2Zn4O2} moiety (two tetrahedra with a common side
Fe1Fe1A, Fig. 1b), in which the central Fe1 and Fe1A atoms (3.0052(9) Å) are bound to each
other and to four peripheral Zn atoms (Zn…Zn 3.0425(9), 3.1080(10) Å, Zn…Fe 3.4048(7),
3.4049(7) Å) by µ4-bridging oxo groups. The metal atoms are additionally bound by ten bridging
carboxylate groups to form a hexanuclear complex with geometry similar to the known
compounds with other metal ions [MIII MII O2(O2CR)10L4] where the environment of the
2
4
peripheral metal ions is complemented by N- or O-donor molecules (L = py, H2O, EtOH, MeOH,
THF, etc.) [59,60]. In a crystal of P21/c group, a complex molecule occupies a special position
(the molecule structure is similar to the modification in space group C2/c, Fig. 1a).
12

Fig. 1. Structure of the metal-carboxylate core (a) in 1 (monoclinic modification in space group
С2/c, (tBu groups are omitted for clarity) and the {Fe2Zn4O2} metal-oxide core (b) of 1 (atoms
with an additional index A are at (1-x, -y, 0.5-z)) (Monoclinic modification in space group С2/c).
The Fe1…Fe2 interatomic distances (2.9993(14) Å), Zn…Zn (3.0686(12) and 3.0802(12)
Å) and Zn…Fe (3.1375(13)‒3.4300(13) Å) and Fe-O, Zn-O (see Table 2) are similar to those in
the analogue described above.
The coordination environment of Fe atoms is octahedral and that of zinc atoms is
tetrahedral. This is rather typical of compounds with an oxygen environment of Fe(III) ions in
carboxylate complexes and with the environment of Zn(II) ions in similar systems with similar
M-O bond lengths that are also observed in the known homometallic compounds [61,62].
Complex 2 crystallizes as a solvate with 1.5 Et2O molecules in triclinic space group P–1.
The metal-carboxylate core of 2 can formally be constructed from the metal core of 1 by
replacement of one peripheral zinc ion in it by iron to give the {Fe3Zn3O2} hexanuclear moiety
(Zn…Zn 3.0836(8) Å, Zn…Fe 3.2298(10)‒3.5694(9) Å, Fe(3)…Zn(1) 3.2023(9) Å,
Fe(3)…Fe(2) 2.9641(10), Fe(3)…Fe(1) 3.5106(10) Å). This moiety is subsequently
complemented to a heptanuclear one by binding of the central Fe(2) atom and the peripheral
Fe(3) atom with Fe(4) by µ3-O and two µ-Piv groups (Fig. 2a; the main bond lengths are given in
Table 2).
13

Fig. 2. Structure of the metal-carboxylate core (a) in 2 (the tBu groups and H atoms at Et2O
solvate molecule are omitted for clarity) and {Fe4Zn3O3} metal-oxide core (b) of 2.
Formally, the heptanuclear metal oxide core in 2 can be regarded as a combination of two
tetrahedra, {Fe2Zn2O} and {Fe3ZnO}, with a common side Fe1Fe2, and one trinuclear moiety
{Fe3O} (Fe…Fe 2.9641(10), 3.4041(10), 3.4321(10) Å, the O(3M) atom is displaced from the
Fe2Fe3Fe4 plane by 0.124(3) Å), which has a common side Fe2Fe3 with the {Fe3ZnO}
tetrahedron (Fig. 2b). As a result, the geometry of the hexanuclear metal-oxide moiety
{Fe3Zn3O2} differs from the known moieties in the other reported [MIII MII O2(O2CR)10L4]
2
4
compounds, which results in a change in the bonding of the metal atoms in {Fe3Zn3O2} by
carboxylate groups. In this case, all iron atoms in it are in an octahedral environment of oxygen
atoms: the Fe(1) atom coordinates two µ4-O and four µ-Piv groups, the Fe(2) atom coordinates
one µ3-O, two µ4-O and three µ-Piv groups, the Fe(3) coordinates one µ3-O, one µ4-O and four
µ-Piv groups, while the Fe(4) atom coordinates one µ3-O, three µ-Piv and one η2-Piv groups. All
zinc atoms are located in a tetrahedral environment of oxygen atoms: the Zn(1) atom coordinates
one µ4-O group, two µ-Piv groups, and one water molecule, while each Zn(2) and Zn(3) atom
coordinates one µ4-O and three µ-Piv groups. One of the bridging groups binds the Zn(3) atom in
the {Fe3Zn3O2} hexanuclear moiety with the Fe(4) atom, (Zn(3)…Fe(4) 3.9425(10) Å).
Additionally, hydrogen bonds are formed between an H atom of the coordinated H2O molecule
and an O(9) atom of the bridging carboxylate group, as well as the O atom of the solvate ether
molecule (O1W–H…O9: O…O 2.750 Å, H…O 1.91 Å, the O–H–O angle is 145.8o; O1W–
H…O1S: O…O 2.647 Å, H…O 1.75 Å, the O–H–O angle is 156.5o) (Fig. 2a).
14

Complex 3 crystallizes in monoclinic space group P21/n. The metal core of 3 is formally
formed upon replacement of a pair of peripheral zinc atoms in the metal core of 1 by a Fe(1,10-
Phen) moiety (Fig. 3a, Table 2).
Fig. 3. Structure of the metal-carboxylate core (a) in 3, (tBu groups are omitted for clarity) and
{Fe3Zn2O2} metal-oxide core (b) of 3.
The structure of the metal core of 3 can be represented as a triangle {Fe3(µ3-O)} (Fe…Fe
2.9659(13)‒3.3956(12) Å, the O(1M) atom is displaced from the Fe1Fe2Fe3 plane by 0.375(4)
Å) and a {Fe2Zn2O} tetrahedron with a common side Fe2Fe3 (Zn1…Zn2 3.0027(11) Å, Zn…Fe
3.2154(13)‒3.3601(14) Å). Zinc atoms are bound with each other and iron atoms (Fe(2) and
Fe(3)) by one µ4-O group and five µ-Piv groups (one of the carboxylate groups is disordered by
two positions). All iron atoms are located in an octahedral environment: the Fe(2) and Fe(3)
atoms coordinate one µ3-O, one µ4-O and four Piv groups; the Fe(2) atom coordinates one µ3-O,
two µ4-O and three µ-Piv groups, while the Fe(1) atom coordinates one µ3-O, three µ-Piv groups
and a 1,10-phen chelate molecule. The environment of each zinc atom is a trigonal bipyramid
(τ(Zn1) = 0.72, τ(Zn2) = 0.65 [63]) owing to coordination of one µ4-O and four µ-Piv groups.
3.3. The Mӧssbauer spectra of 1-3
In order to estimate the magnetic state of atoms, we studied the Mössbauer spectra of the
compounds synthesized. Fig. 4 shows the doublet spectra of paramagnetic complexes 1 - 3 at T =
300 and 16 K in zero magnetic field. The isomeric shifts (δ) and quadrupole splitting (∆) make it
possible to conclude that the complexes contain high-spin Fe3+ ions in an octahedral environment
of O atoms, which agrees with X-ray single crystal diffraction data and magnetic measurements.
There are no iron(II) ions in complexes 1-3. The isomeric shifts and quadrupole splitting in the
complexes obtained lie in broad ranges (Table 3), unlike the previously studied iron(III)
15

carboxylate complexes [28, 64-67], whose δ and ∆ values range within 0.42–0.47 and 0.62–0.68
mm/s, respectively.
Fig.4. Mӧssbauer powder spectra of compounds 1 (a, 300 K), 3 (b, 300 K) and 2 (c, 300 K; d, 16
K) at various temperatures
Table 3
Mössbauer parameters for 1 – 3
δ
∆
Г
Hin,
0.5T
A,
0.05
Compound T, K Form Fe
0.03 mm/s
300 Fe3+(paramagnetic) 0.36
300 Fe3+(paramagnetic) 0.40
0.77
0.90
1.02
0.85
1.27
0.44
0.64
0.87
0.54
0.57
-
-
1.00
1.00
6.8*
0.79
0.21
1
2
16
Fe3+(paramagnetic) 0.54
Fe3+(paramagnetic) 0.39
Fe3+(paramagnetic) 0.47
-
-
300
3
δ – isomer shift relative α-Fe;
∆ – quadrupole splitting;
Г – line width;
Hin – internal magnetic field at the 57Fe nucleus;
A – abundance.
*the value of 6.8 at 16 K is relative (the intensity at 300 K was taken to be 1).
The Mössbauer spectrum of complex 1 is an asymmetric doublet with nearly equal and
rather small line widths. Two explanations of the line asymmetry are possible: 1) the texturing
16

(spontaneous sample orientation) effect, according to which the facets of many adjacent particles
are arranged parallel to each other, i.e., preferential orientation of a set of particles relative to the
observation axis is preserved to some extent; 2) manifestation of the Goldanskii-Karyagin effect
[68], that appears due to a rather strong anisotropy of thermal vibrations of iron atoms in the
complex (or inequality of the root-mean-square amplitudes of atom displacement) in different
crystallographic directions. As a result, the corresponding probabilities of recoilless transitions
(fZ, fX, fY) also become nonequal. Therefore, the expression for the dependence of the intensities
of π- and σ-transitions (Iπ and Iσ) on the observation angle θ (i.e., the angle between the
direction of the γ-radiation beam and the crystal’s axis of symmetry) can be represented as:
Iπ = I0 3/2(1 + cos2θ) × f(θ), Iσ = I0 (1 +3/2 sin2θ) × f(θ)
Thus, the line intensity ratios Iπ/Iσ in a polycrystalline sample of compound 1 differ from 1,
though the values cos2θ = 1/3, sin2θ = 2/3 averaged over a sphere in it remain unchanged.
(Perhaps, at room temperature the amplitude of vibrations of iron ions in the direction
perpendicular to the surface exceeds the corresponding values for in-plane vibrations, or vice
versa).
A distinctive feature of the Mössbauer spectrum of 2 at Т = 300 K is that the lines are
broadened considerably (Г = 0.64 mm·s-1) (see Table 3, Fig. 4с). The following main reasons of
this phenomenon can be distinguished:
1) the scatter of values (distribution) of quadrupole splitting, with a constant isomeric shift
value, near some mean value (∆ = 0.90 mm·s-1, see Table 3), which assumes a considerable
structural disorder of the sample;
2) manifestation of paramagnetic (spin-lattice) relaxation. In this case, broadening is due to
changes in the electron spin fluctuation frequency in the region of frequencies characteristic of
57Fe Mössbauer spectroscopy (108 – 109 s-1);
3) nonequivalence of the main electron levels of iron ions in positions 1, 4 and 2, 3, which
is in correct agreement with magnetic measurement data (see Section 2.3.2, Fig. 4а). Apart from
the contribution of structural disorder, a contribution related to the decrease in the spin-spin
relaxation frequency may appear in this case, since energy is not conserved in simultaneous re-
orientation of adjacent spins, and hence «flip-flop» processes do not slow down.
A study of the structural and electronic properties of complex 2 at Т = 16 K has shown (see
Table 3, Fig. 4c) that the А value at this temperature is 6.8 times larger than the similar value at Т
= 300 K. This difference agrees with the expected behavior in the molecular crystal model
[69,70] and indicates that intermolecular vibrations make the main contribution to the
temperature dependence. Furthermore, the width of resonance absorption lines increases
considerably at T=16 K, which agrees with the assumption on the decrease in the spin-lattice
17

relaxation frequency with a decrease in temperature. It should be noted that an increase in the
measurement temperature from 16 to 300 K results in a decrease in the isomeric shift by ~ 0.14
mm/s due to the second-kind Doppler effect [64].
Unlike complexes 1 and 2, two structurally non-equivalent positions of Fe3+ ions are
observed in a polycrystalline sample of 3. The parameters of these positions differ from those of
the single positions in samples 1 and 2. In general, complexes 2 and 3 (300 K) have higher δ and
∆ values in comparison with complex 1 (see Table 3). Thus, incorporation of an additional iron
atom to the metal core of 2 or a 1,10-phenanthroline molecule in 3 (see Figs. 2, 3) affects the
changes in the electron density in the molecule, and in 3 it gives rise to two structurally
nonequivalent positions of iron, and as a result, to considerably different parameters of the local
metal environment.
3.4. Magnetic properties 1 and 2
The magnetic characteristics of compounds 1 and 2 were studied by measurements of the
effect of temperature on molar magnetic susceptibility (χМ) in the range from 4 to 300 K in a
field of Н = 0.5 kOe. As the temperature decreases from 300 to 4 K, the χMT value steadily
decreases from 7.56 cm3·K·mol-1 for 1 and 7.77 cm3·K·mol-1 for 2 to 0.33 cm3·K·mol-1 for 1 and
0.43 cm3·K·mol-1 for 2, respectively (Fig. 5). The calculated χMT values for systems of two and
four non-interacting high-spin ions S = 5/2 with g = 2 are considerably higher than the
experimental ones (8.75 and 17.5 cm3·K·mol-1, respectively) at room temperature, which, along
with the shape of the χMT(T) curve, indicates that antiferromagnetic interactions predominate in
the compound in question (Fig. 5).
18

8
6
4
2
0
1
2
Τ
Τ
Τ
Τ
χ
χ
χ
χ
0
100
200
300
T, K
Fig. 5. Temperature plots χT (a) and calculated curves (solid lines) for 1 (1) and 2 (2).
The experimental χMT(T) plot for complexes 1 and 2 were interpreted on the basis of
isotropic spin Hamiltonians for two (1) and four (2) magnetically equivalent nuclei with a spin of
S = 5/2 (the exchange coupling scheme for the {Fe4} moiety is shown in Scheme 1),
respectively.
ˆ ˆ
ˆ
ˆ
ˆ
H = −2J12S1S2 + βH(g1S1 + g2S2 )
(1)
ˆ ˆ
ˆ ˆ
ˆ ˆ
ˆ ˆ
ˆ ˆ
ˆ
ˆ
ˆ
ˆ
ˆ
H = −2J12S1S2 − 2J13S1S3 − 2J23S2S3 − 2J24S2S4 − 2J34S3S4 + βH(g1S1 + g2S2 + g3S3 + g4S4 )
(2)
Scheme 1
The calculation was carried out by full-matrix diagonalization using the Mjollnir program
[71]. The best fit of χMT vs. Т for 1 is reached with the following parameters: J12 = -3.05 cm-1, g1
19

= g2 = 1.99, concentration of paramagnetic impurity 4.2% with S = 5/2 (R2 = 8.8·10-4, R2 =
2
∑[(χMT)exp–(χMT)calc] /( χMT)exp2)). The best fit of the calculated and experimental plots of χMT
∑(
vs. Т for 2 is reached with the following parameters: J12 = -0.4 cm-1, J13 = -13.5 cm-1, J23 = -3.6
cm-1, J24 = -27.2 cm-1, J34 = -30.1 cm-1, g1 = g2 = g3 = g4 = 2, concentration of paramagnetic
impurity of 6% with S = 5/2 (R2 = 4.7·10-4).
The data obtained in this calculation was agree with the magnetic-structural correlations
reported previously [72], in which the exchange parameter J depends on the Fe-O bond length
and the Fe-O-Fe angle, where the O atom is a monoatomic bridging ligand, specifically, an oxo
group in these compounds. In complex 1, the Fe-O-Fe angles in the binuclear moiety {Fe2O2}
amount to 98.5o and 98.8(2)o and the Fe-O bond lengths are 1.979(2) and 1.984(2) Å. Based on
the results reported in [72], this assumes that the J integral value is close to -5 cm-1; in the case of
complex 1, J12 = -3.05 cm-1. In complex 2 the smallest angles (Fe1-O-Fe2 96.84(14), 97.93(14))
with similar Fe-O bond lengths of 1.992(3)-2.044(3) Å are observed in the {Fe1(µ-O)2Fe2}
moiety, so the exchange integrals in the molecule are the smallest, J12 = -0.4 cm-1. In the {Fe2(µ-
O)2Fe3} moiety, attention should be paid to the increase in one of the Fe-O-Fe angles (Fe3-
O3M-Fe2, 99.8(2)o) and shortening of the Fe-O bond lengths (1.932(3), 1.945(3) Å), which
results in a higher parameter J23 = -3.6 cm-1. In the {Fe1(µ-O)2Fe3}, {Fe2(µ-O)2Fe4} and
{Fe3(µ-O)2Fe4} moieties, the angle grows from 121.2(2)° to 130.9(2)°, while the bond lengths
decrease from 2.037(3) to 1.842(3) Å. As a result, here we observe a transition to more efficient
exchange coupling between the paramagnetic centers: J13 = -13.5 cm-1, J24 = -27.2 cm-1, J34 = -
30.1 cm-1, respectively.
Thus, the entire set of data shows that the values of exchange integrals obtained correlate
with the bond lengths and bond angles in the {Fe(µ-O)Fe} moieties and complement similar
correlations reported elsewhere [72].
3.5. DFT -calculations
Additionally to magnetochemical simulations of the experimental χT vs T dependencies
for 1 and 2 we have also provided DFT calculations of the exchange coupling parameters. The
primary target of such calculations was to avoid overparametrization issues in the case of 2
magnetochemical properties simulation. The calculations for 1 were used primarily for validation
of the approximations adopted for 2.
The exchange integral J12 within {Fe2Zn4}Ac species was calculated using broken
symmetry DFT approach utilizing a set of hybrid exchange-correlation potentials. First, using
B3LYP XC potential, the influence of basis set quality on the result was evaluated (Table 4). It
was found that improving the basis set from double-zeta (Def2-SVP) to triple-zeta (Def2-TZVP)
20

quality significantly changed the result, while further improvement to quadruple-zeta one (Def2-
QZVPP) led to minor changes (0.05 cm–1) with 6-fold increase of computation time. Thus,
further calculations were provided using def2-TZVP basis set.
Application of another hybrid XC potentials, including more theoretically advanced,
failed to overcome the B3LYP result. M06 and TPSSh XC potentials gave inferior results
comparing to B3LYP, and PBE0, BHandHLYP and modified B3LYP (with fraction of exact
exchange increased to 35 %, such approach was shown to be promising for calculation of
magnetic coupling constants for a set of mixed oxides and fluorides [73]) gave qualitatively
incorrect (wrong sign or order of magnitude) results.
Table 4
The values of the magnetic exchange integral J12 in {Fe2Zn4}Ac, {Fe2Zn4}Acopt and
{Fe2Zn4}Pivopt obtained via different approximations.
Species
XC potential
Basis set
J12, cm-1
−2.50
−2.50
−2.44
−1.74
−2.50
−2.55
−0.17
+6.63
−8.72
−1.57
+3.95
−3.05
Def2-TZVP/ Def2-SVP*
Def2-TZVP
Def2-TZVP
Def2-SVP
{Fe2Zn4}Pivopt
{Fe2Zn4}Acopt
B3LYP
Def2-TZVP
Def2-QZVPP
Def2-TZVP
Def2-TZVP
Def2-TZVP
Def2-TZVP
Def2-TZVP
PBE0
{Fe2Zn4}Ac
BhandHLYP
TPSSh
M06
Modified B3LYP
Simulated from magnetochemical data for 1
* See Experimental section 2.3.4 for details.
It was expected that replacing the pivalate groups by acetate ones would cause the
notable change in the values of the exchange coupling parameter due to redistribution of the
charge density on the bridging ligands similarly to the results previously reported by us [74].
However, such substitution had sense due to significant reduce of computational efforts. As the
results in both cases were expected to be semi-quantitative, we considered acceptable to sacrifice
the accuracy for speed in such situation. We also considered that positions of acetate groups
atoms could be optimized only at molecular mechanics level. However, in order to check how
21

significantly all our simplifications could mangle the results of the calculations for 2, in the case
of 1 we provided additional calculations with more complicated model particles. It was found
(see Table 4) that DFT optimization (PBE XC functional, Def2-TZVP basis – {Fe2Zn4}Acopt
species) of acetate group atoms slightly lower the magnitude of magnetic coupling parameter.
Surprisingly, it was also found that the replacement of acetate groups by pivalate groups
(optimized utilizing PBE XC functional and Def2-SVP basis – {Fe2Zn4}Pivopt species) leads to
virtually no change of the J value. Substitution of Def2-TZVP basis set by Def2-SVP one for
pivalate groups in {Fe2Zn4}Pivopt in the calculation of J value didn’t change the result.
Basing on the results of the exchange integral calculation for {Fe2Zn4}Ac species,
B3LYP exchange-correlation potential with Def2-TZVP basis set were utilized for calculation of
the exchange integrals within {Fe4Zn3}Ac species.
{Fe4Zn3}Ac consisted of four paramagnetic Fe3+ ions, there could potentially be six
different exchange interactions between them. Flipping spin at any ion leads to complicated
change of the system’s energy due to involvement of each ion in a few exchange interactions. In
order to extract each exchange coupling parameter, two approaches were applied:
Approach 1. Exchange coupling parameter between each pair of FeIII ions was calculated
by regular broken symmetry DFT approach after substitution of other FeIII ions by ScIII ions. ScIII
was chosen as the closest diamagnetic analogue to FeIII with equal charge. Though such
substitution implied an additional source of error, this approach was simple and robust, allowing
independent determination of each coupling parameter.
Approach 2. Within this approach, a series of broken symmetry DFT calculations was
provided, flipping spins of some (one or two) FeIII ions of {Fe4Zn3}Ac species and comparing
the resulted energies with the energy of the high-spin state of {Fe4Zn3}Ac. The energies of each
considered broken symmetry state were determined by a few exchange coupling parameters. In
order to extract individual J values, the system of equations resulted from application of full
matrix spin hamiltonian approach (see Experimental section and Supplementary Info for details)
was solved. This approach avoids additional source of error originated from substitution of FeIII
ions by ScIII ions, but it required more insight and was more technically complicated.
Additionally, if J values significantly differed (by orders of magnitude), the errors in
determination of the states’ energies could significantly impair the precision of determination of
the smallest (by magnitude) J values.
The results of both approaches (Table 5) were close, except the smallest J12 (differed by
an order of magnitude) and J14 (4-fold difference). It should be noted, that comparison with
experimental data couldn’t help to deduce the most precise approach, as the precision of
extracting the smallest J values from simulation of the magnetochemical data should be
22

considered low. However, in our opinion, the results provided by approach 1 should be more
reliable for small J values due to above mentioned reasons.
Table 5
The values (in cm-1) of the magnetic exchange integrals in {Fe4Zn3}Ac.
Exchange
integral
J12
Calculated by
approach 1
−0.7
Calculated by
approach 2
−0.07
Simulated from
magnetochemical data
−0.4
J13
−23.9
−22.9
−13.5
J23
J24
−5.1
−46.1
−6.7
−43.7
−3.6
−27.2
J34
−48.4
−0.24
−46.3
−0.81
−30.1
Considered zero
J14
3.6. Thermal analysis 1 and 2
The thermal behavior of 1 and 2 was studied by the STA method under argon with
simultaneous TG and DSC recording (Fig. 6). Samples are characterized by different thermal
stability and different course of thermal destruction. The behavior of carboxylates in the
complexes during thermolysis is determined by the type of structural organization and the
strength of bonds in the molecule. In the compounds in question, М–О, (СН3)3С–CОО, С–Н and
С–С are the main bonds. Bonds of the first two types are broken rather readily (this is confirmed
by literature data, e.g., [75,76]). Hydrogen transfer from the hydrocarbon radical to oxygen also
occurs rather readily and determines the removal of HPiv that is formed. Complex 1 is stable up
to 241 ºС, whereas the first indications of mass loss in 2 already begin at 45 ºС, which is due to
the presence of outer-sphere volatile ether molecules (mcalc.= 6%). In the temperature range of
241-375 ºС, only one stage of chemical destruction of the complex (∆m = 94.43 %) was recorded
for 1 (Fig. 6a). The intense endothermal peak with an extremum at 305 ºC corresponds to
melting followed by decomposition of the compound. For 2, a gradual decrease in the sample
mass is observed up to 225 ºС (∆m = 15.4 %, Fig. 6b).
23

Fig. 6. TG (a) and DSC (b) curves of complexes 1 (a) and 2 (b)
Previous studies on the thermolysis of carboxylate systems [77,78] indicate that diverse
low-molecular carbon-containing products (СH2O, C4H8, C5H10O, etc.) are formed in the 200-
370 ºС range, and it is the temperature range where the maximum mass loss of complexes 1 and
2 occurs. The exothermic effects with extremums at 326 ºС (1) / 325 ºС (2) correspond to
sublimation of thermolysis products (a golden-brown deposit is observed in both cases on the
crucible lids after the experiments).
According to X-ray powder diffraction and EDX data, the residue from thermal analysis of
1 and 2 is an amorphous phase that predominantly consists of γ-Fe2O3 with an insignificant
content of carbon (Fig. S4). (The Al peak is due to the material of the crucible in which the
thermolysis was performed).
The mass of the residue formed from 1 is as small as 5.6% with respect to Fe2O3, which is
about two times smaller than the expected calculated value (mcalc.=11 %). In contrast to 1, the
mass of the residue from 2 is only 4% smaller than the calculated value. (mcalc./found.=23% / 19%,
respectively). Perhaps, the volatility of 1 results from the higher content of zinc atoms in
comparison to iron in the cluster. It should be noted that the volatility of thermolysis products of
zinc complexes under conditions of thermal experiments in dynamic argon atmospheres was
observed previously [79-83]).
4. Conclusions
Thus, we have synthesized heteropolynuclear pivalate Fe-Zn complexes 1-3 and
characterized their structures. It has been found that the use of solid-phase thermolysis at one of
the stages of synthesizing heteronuclear {Fe-Zn} compounds expands the potential of varying
the metal core in the resulting pivalate complexes. According to X-ray single-crystal analysis,
24

iron atoms are in octahedral environment and zinc atoms are predominantly (except in complex
3) in tetrahedral environment of oxygen atoms. Simulation of magnetic data and quantum-
chemical calculations have shown that antiferromagnetic exchange coupling exists in complexes
1 and 2; and the values of the exchange coupling parameters correlates with the Fe-O distances
in Fe‒O‒Fe bonds in the oxo bridging moieties. Mössbauer spectroscopic data are characterized
by elevated ∆ values in 2 and 3 and provide evidence of a considerable distortion of the
octahedral environment of Fe(III) due to the addition of the seventh iron(III) ion to the metal
core in 2, or a 1,10-phenanthroline molecule in 3.
Appendix A. Supplementary material
CCDC 1560567 (1, C2/c), 1560569 (1, P21/c), 1560570 (2) and 1560568 (3) contain the
supplementary crystallographic data for compounds 1-3. These data can be obtained free of
charge via http: //www.ccdc.cam.ac.uk. Supporting information file contains PXRD (Figs. S2,
S3), EDX spectrum of the end-product after thermolysis of 2 (Fig. S4) and Table 1s with relative
DFT energies of the broken symmetry states of {Fe4Zn3}Ac.
Acknowledgements
This work was financially supported by the Russian Scientific Foundation (project 14-23-
00176). We are grateful to the User Facilities Center of Institute of General and Inorganic
Chemistry of the Russian Academy of Sciences for physicochemical measurement (single crystal
X-ray, IR-spectroscopy, magnetic measurements, elemental and termogravimetric analysis).
25

References
[1] M.M. Rashad, O.A. Fouad, Mater. Chem. Phys. 94 (2005) 365.
[2] M.K. Karunananda, F.X. Vázquez, E.E. Alp, W. Bi, S. Chattopadhyay, T. Shibatad, N.P.
Mankad, Dalton Trans. 43 (2014) 13661.
[3] M. Veith, M. Haas, V. Huch, Chem. Mater., 17 (2005) 95.
[4] R.V. Godbole, P. Rao, P.S. Alegaonkar, S. Bhagwat, Mater. Chem. Phys. 161 (2015) 135.
[5] L.W. Yeary, J.W. Moon, C.J. Rawn, L.J. Love, A.J. Rondinone, J.R. Thompson, B.C.
Chakoumakos, T.J. Phelps, J. Magn. Magn. Mater. 323 (2011) 3043.
[6] M.G. Naseri, E.B. Saion, M. Hashim, A.H. Shaari, H.A. Ahangar, Sol. State Comm. 151
(2011) 1031.
[7] F. Doungmene, P.A. Aparicio, J. Ntienoue, Ch.S. A. Mezui, P. de Oliveira, X. López, I.M.
Mbomekallé, Electrochim. Acta 125 (2014) 674.
[8] J.R. Long, Molecular Cluster Magnets / Ed. Yang P. Hong Kong: World Scientific. 2003,
291 p.
[9] K.O. Abdulwahab, M.A. Malik, O’Brien P., G.A. Timco, F. Tuna, C.A. Muryn, R.E.P.
Winpenny, R.A.D. Pattrick, V.S. Coker, E. Arenholz, Chem. Mater. 26 (2014) 999.
[10] B. Singh, J.R. Long, F. Biani, D. Gatteschi, P. Stavropoulos, J. Am. Chem. Soc. 119 (1997)
7030.
[11] N.P. Mankad, Synlett. 25 (2014) 1197.
[12] U. Jayarathne, S.R. Parmelee, N.P. Mankad, Inorg.Chem. 53 (2014) 7730.
[13] X.M. Sheng-Min Hu, Ch.-H. Tan, Ch.-J. Shen, Q.-L. Zhu, Y.-H. Wen, R.-B. Fu, T.-L.
Sheng, X.-T. Wu, Polyhedron 41 (2012) 86.
[14] B.E. Zaugg, T. Kolb, A.M. Arif, R.D. Ernst, J. Chem. Cryst. 40 (2010) 778.
[15] E.Y. Tshuva, S.J. Lippard, Chem. Rev. 104 (2004) 987.
[16] M.K. Karunananda, F.X. Vázquez, E.E. Alp, W. Bi, S. Chattopadhyay, T. Shibata, N.P.
Mankad, Dalton Trans. 43 (2014) 13661.
[17] A.M. Kluwer, R. Kapre, F. Hartl, M. Lutz, A.L. Spek, A.M. Brouwer, P.W.N.M. van
Leeuwen, J.N.H. Reek, Proc. Nat. Acad. Sci. USA 106 (2009) 10460.
[18] X.F. Liu, X.W. Xiao, J. Organomet. Chem. 696 (2011) 2767.
[19] A. Hazra, S. Bonakala, S.K. Reddy, S. Balasubramanian, T.K. Maji, Inorg. Chem. 52 (2013)
11385.
[20] A.M. Kluwer, R. Kapre, F. Hartl, M. Lutz, A.L. Spek, A.M. Brouwer, P.W.N.M.van
Leeuwen, J.N.H. Reek, Proc. Nat. Acad. Sci. USA 106 (2009), 10460.
26

[21] S. Biswas, M. Tonigold, H. Kelm, H.-J. Kruger, D. Volkmer, J. Chem. Soc. Dalton Trans.
39 (2010), 9851.
[22] M. Veith, M. Haas, V. Huch, Chem. Mater. 17 (2005) 95.
[23] R. Jovito, A. Neves, A.J. Bortoluzzi, M. Lanznaster, V. Drago, W. Haase, Inorg. Chem.
Commun. 8 (2005) 323.
[24] F.R. Xavier, A.J. Bortoluzzi, A. Neves, Chem. Biodiversity 9 (2012) 1794.
[25] B. Singh, J.R. Long, G.C. Papaefthymiou, P. Stavropoulos, J. Am. Chem. Soc. 118 (1996)
5824.
[26] C. Turta, S. Shova, D. Prodius, V. Mereacre, M. Gdaniec, Y. Simonov, J. Lipkowski, Inorg.
Chim. Acta 357 (2004) 4396.
[27] M.M. Amini, M. Yadavi, S.W. Ng, Acta Crystallogr. E 60 (2004) 495.
[28] I.А. Lutsenko, М.А. Kiskin, V.K. Imshennik, Yu.V. Maksimov, А.А. Sidorov, I.L. Eremenko,
Russ. J. Coord. Chem. 6 (2017) 348.
[29] R.D. Cannon, R.P. White, Prog. Inorg. Chem. 36 (1988) 195.
[30] W. Jolly, Preparative inorganic reactions T. I, Wiley-Interscience Publication John Wiley &
Sons, New York London Sydney Toronto, 1964, 141 p.
[31] Bruker (1997). SMART (Control) and SAINT (Integration) Software. Version 5.0: Bruker
AXS Inc., Madison, WI, 1997.
[32] G.M. Sheldrick, SADABS, Program for Scanning and Correction of Area Detector Data,
Göttingen University, Göttinngen, Germany, 2004.
[33] G.M. Sheldrick, Acta Crystallogr. A. 64 (2008) 112.
[34] F. Neese, Wiley Interdiscip. Rev. Comput. Mol. Sci. 2 (2012) 73.
[35] P.J. Stephens, F.J. Devlin, C.F. Chabalowski, M. J. Frisch, J. Phys. Chem. 98 (1994) 11623.
[36] F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 7 (2005) 3297.
[37] F. Neese, J. Comput. Chem. 24 (2003) 1740.
[38] F. Neese, F. Wennmohs, A. Hansen, U. Becker, Chem. Phys. 356 (2009) 98.
[39] F. Weigend, Phys. Chem. Chem. Phys. 8 (2006) 1057.
[40] C. Adamo, V. Barone, J. Chem. Phys. 110 (1999) 6158.
[41] A.D. Becke, J. Chem. Phys. 98 (1993) 1372.
[42] Y. Zhao, D.G. Truhlar, Theor. Chem. Acc. 120 (2008) 215.
[43] V.N. Staroverov, G.E. Scuseria, J. Tao, J.P. Perdew, J. Chem. Phys. 119 (2003) 12129.
[44] V.N. Staroverov, G.E. Scuseria, J. Tao, J.P. Perdew, J. Chem. Phys. 121 (2004) 11507.
[45] M. D. Hanwell, D. E. Curtis, D. C. Lonie, T. Vandermeersch, E. Zurek, G. R. Hutchison, J.
Cheminformatics. 4 (2012) 17.
27

[46] A.K. Rappe, C.J. Casewit, K.S. Colwell, W.A. Goddard, W.M. Skiff, J. Am. Chem. Soc.
114 (1992) 10024.
[47] N.M. O’Boyle, M. Banck, C.A. James, C. Morley, T. Vandermeersch, G.R. Hutchison, J.
Cheminformatics 3 (2011) 33.
[48] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (996) 3865.
[49] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 78 (1997) 1396.
[50] K. Yamaguchi, Y. Takahara, T. Fueno, in Appl. Quantum Chem. (Ed.: V.H. Smith), Reidel,
Dordrecht, 1986, 155 p.
[51] T. Soda, Y. Kitagawa, T. Onishi, Y. Takano, Y. Shigeta, H. Nagao, Y. Yoshioka, K.
Yamaguchi, Chem. Phys. Lett. 319 (2000) 223.
[52] A. Bencini, F. Totti, Int. J. Quantum Chem. 101 (2005) 819.
[53] A.S. Litvinenko, E.A. Mikhaleva, S.V. Kolotilov, V.V. Pavlishchuk, Theor. Exp. Chem. 46
(2011) 422.
[54] R.A. Polunin, S V. Kolotilov, M.A. Kiskin, O. Cador, E.A. Mikhalyova, A.S. Lytvynenko,
S. Golhen, L. Ouahab, V.I. Ovcharenko, I L. Eremenko, et al., Eur. J. Inorg. Chem. 2010
(2010) 5055–5057.
[55] John W. Eaton, David Bateman, Søren Hauberg, Rik Wehbring (2015). GNU Octave
version 4.0.0 manual: a high-level interactive language for numerical computations. URL
http://www.gnu.org/software/octave/doc/interpreter/
[56] A.A. Sidorov, I.G. Fomina, M.O. Poninz, G,G, Alexandrov, S.E. Nefedov, I.L. Eremenko,
I.I. Moiseev, Russ. Chem. Bull., Engl. Transl. 49 (2000) 958.
[57] K.O. Abdulwahab, M.A. Malik, P. O'Brien, G.A. Timco, F. Tuna, R.E.P. Winpenny, R.A.
D. Pattrick, V.S. Cokerd, E. Arenholz, J. Mater. Chem. C, 2 (2014) 6781.
[58] T.P. Gerasimova, S.A. Katsyuba, Dalton Trans. 42 (2013) 1787.
[59] K.S. Gavrilenko, S.V. Punin, O. Cador, S. Golhen, L. Ouahab, V.V. Pavlishchuk, Inorg.
Chem. 44 (2005) 5903.
[60] P. Kar, R. Haldar, C.J. Gómez-García, A. Ghosh, Inorg. Chem. 51 (2012) 4265.
[61] J. Kim, H. Cho, Inorg. Chem. Comm. 7 (2004) 122.
[62] D. Lee, L. Sorace, A. Caneschi, S.J. Lippard, Inorg. Chem. 40 (2001) 6774.
[63] A.W. Addison, T.N. Rao, J. Chem. Soc. Dalton Trans. 1984, 1349.
[64] S.A. Sotnik, S.V. Kolotilov, M.A. Kiskin, Zh.V. Dobrokhotova, K.S. Gavrilenko, V.M.
Novotortsev, I.L. Eremenko, V.K. Imshennik, Yu.V. Maksimov, V.V. Pavlishchuk, Russ.
Chem. Bull. 63 (2014) 862.
[65] R.V. Pound, G.A. Rebka, Phys. Rev. Lett. 4 (1960) 274.
[66] M. Takano, J. Phys. Soc. Jpn., 33 (1972) 1312.
28

[67] B. D. Rumbold, G.V.H. Wilson, J. Phys. Chem. Solids, 34 (1973) 1887.
[68] V.I. Goldanskii, H. Herber, Chemical Application of Mossbauer Spectroscopy / Academic
Press, New York: 1968, 320 р.
[69] A.J. Rein, R.H. Herber, J. Chem. Phys. 63 (1975) 1021.
[70] R.H. Herber, M.F. Leahy, Ibid. 67 (1977) 2718.
[71] R.A. Polunin, S.V. Kolotilov, M.A. Kiskin, O. Cador, E.A. Mikhalyova, A.S. Lytvynenko,
S. Golhen, L. Ouahab, V.I. Ovcharenko, I.L. Eremenko, V.M. Novotortsev, V.V.
Pavlishchuk, Eur. J. Inorg. Chem., (2010) 5055.
[72] K.J. Mitchell, K.A. Abboud, G. Christou, Inorg. Chem., 2016, 55, 6597.
[73] Y. Kitagawa, T. Matsui, N. Yasuda, H. Hatake, T. Kawakami, S. Yamanaka, M. Nihei, M.
Okumura, H. Oshio, K. Yamaguchi, Polyhedron, 66 (2013) 97.
[74] I.N. Shcherbakov, T.M. Ivanova, M.A. Kiskin, S.V. Kolotilov, V.M. Novotortsev, I.L.
Eremenko, V.A. Kogan, Inorg. Chim. Acta. 421 (2014) 507.
[75] Yu.S. Matveev, A.V. Kuchin, Russ. J. Chem. Vegetable raw materials. 2 (2001) 21.
[76] Yu.S. Matveev, T.V. Khurshkainen, A.V. Kuchin, Butlerov Comm. 6 (2002) 71.
[77] N.P. Burkovskaya, M.E. Nikiforova, M.A. Kiskin, Zh.V. Dobrokhotova, A.S. Bogomyakov,
P.S. Koroteev, V.I. Pekhnyo, A.A. Sidorov, V.M. Novotortsev, I.L. Eremenko, Polyhedron,
35 (2012) 116.
[78] Zh.V. Dobrokhotova, I.G. Fomina, M.A. Kiskin, G.B. Belov, M.A. Bykov, V.M.
Novotortsev, Russ. J. Phys. Chem. A 80 (2006) 323.
[79] I.A. Lutsenko, E.V. Korneeva, A.V. Ivanov, Russ. J. Coord. Chem. (Engl. Transl.) 42
(2016) 494.
[80] I.A. Lutsenko, A.V. Ivanov, E.V. Korneeva, A.I. Tursina, Russ. J. Coord. Chem. (Engl.
Transl.) 38 (2012) 779.
[81] A. Tavlaridis, R. Neeb, Z. Fresenius, Anal. Chem. 293 (1978) 290.
[82] M. Zupaccic, I. Turel, P. Bukovec, A.J.P. White, D.J. Williams, Croatica Chem. Acta. 74
(2001) 61.
[83] F.A. Cotton, G. Wilkinson, Advan. Inorg. Chem., Wiley and Sons, New York, 1088, 605 p.
29