Syntheses, structures, and polymorphism of β-diketonato complexes - Co(thd)3

Article abstract of DOI:10.1002/zaac.200700462

Oxidation of Co(thd)₂ dissolved in different solvents has been investigated in air and oxygen atmosphere. In oxygen atmosphere and at the boiling point of the solvents this treatment leads to oxidation of Co ^II to Co^III, but also to degradation of some of the thd ligands and formation of a new mixed-ligand complex. Three pure-cultivated crystalline Co(thd)₃ phases are reported: 1 (room-temperature phase), 2 (low-temperature phase), and 3 (metastable phase) and in addition there exists an amorphous Co(thd)₃ phase (4) with approximate composition Co(thd)₃·xH(thd); x = 0.06. Reaction of metal(II) oxides (MO, M = Mn, Fe, and Co) with H(thd) under air or O₂ atmosphere is an easy direct route to M(thd)₃ complexes. Structure determinations are reported for Co(thd)₃ (1-3) based on single-crystal X-ray diffraction data. Modification 1 crystallizes in space group P3c1 with a = b = 18.8100(10), c = 18.815(2) A at 295 K; R(wR2) = 0.180, modification 2 in space group C2/c with a = 28.007(12), b = 18.482(8), c = 21.356(9) A, β = 97.999(5)° at 100 K; R(wR2) = 0.211, and modification 3 in space group Pnma with a = 19.2394(15), b = 18.8795(15), c = 10.7808(8) A at 100 K; R(wR2) = 0.193. The molecular structures of 1-3 all comprise a central Co atom octahedrally co-ordinated by the ketonato O atoms of three thd ligands. The transformation between modifications 1 and 2 is of a fully reversible second-order character. Modifications 1 and 3 are, on the other hand, related by a quasi-reversible cycle. Heat treatment (specifically sublimation) of 1 leads to 3 whereas re-crystallization or prolonged storage at room temperature is required to regenerate 1. Co(thd)₃ has sufficient thermal stability to permit sublimation without degradation. The various forms of Co(thd) ₃ are all diamagnetic, viz. a confirmation of the Co^III valence state.

Full text of DOI:10.1002/zaac.200700462

DOI: 10.1002/zaac.200700462
Syntheses, Structures, and Polymorphism of β-Diketonato Complexes ؊ Co(thd)3
˚
M. A. K. Ahmed, Helmer Fjellvag, Arne Kjekshus*, and Pascal D. C. Dietzel
Oslo/Norway, Department of Chemistry, University of Oslo
Received June 21st, 2007; accepted September 21st, 2007.
Dedicated to Professor Rudolf Hoppe on the Occasion of his 85th Birthday
˚
Abstract. Oxidation of Co(thd)2 dissolved in different solvents has
been investigated in air and oxygen atmosphere. In oxygen atmos-
phere and at the boiling point of the solvents this treatment leads
to oxidation of CoII to CoIII, but also to degradation of some of
the thd ligands and formation of a new mixed-ligand complex.
Three pure-cultivated crystalline Co(thd)3 phases are reported: 1
(room-temperature phase), 2 (low-temperature phase), and 3 (meta-
stable phase) and in addition there exists an amorphous Co(thd)3
phase (4) with approximate composition Co(thd)3 ·xH(thd); x ϭ
0.06. Reaction of metal(II) oxides (MO, M ϭ Mn, Fe, and Co)
with H(thd) under air or O2 atmosphere is an easy direct route
to M(thd)3 complexes. Structure determinations are reported for
Co(thd)3 (1Ϫ3) based on single-crystal X-ray diffraction data.
c ϭ 21.356(9) A, β ϭ 97.999(5)° at 100 K; R(wR2) ϭ0.211, and
modification 3 in space group Pnma with a ϭ 19.2394(15), b ϭ
˚
18.8795(15), c ϭ 10.7808(8) A at 100 K; R(wR2) ϭ 0.193. The mo-
lecular structures of 1Ϫ3 all comprise a central Co atom octa-
hedrally co-ordinated by the ketonato O atoms of three thd ligands.
The transformation between modifications 1 and 2 is of a fully
reversible second-order character. Modifications 1 and 3 are, on
the other hand, related by a quasi-reversible cycle. Heat treatment
(specifically sublimation) of 1 leads to 3 whereas re-crystallization
or prolonged storage at room temperature is required to regenerate
1. Co(thd)3 has sufficient thermal stability to permit sublimation
without degradation. The various forms of Co(thd)3 are all dia-
magnetic, viz. a confirmation of the CoIII valence state.
¯
Modification 1 crystallizes in space group P3c1 with a ϭ b ϭ
˚
18.8100(10), c ϭ 18.815(2) A at 295 K; R(wR2) ϭ 0.180, modifi-
Keywords: Co(thd)3; Crystal structure; Polymorphism; Oxidation;
cation 2 in space group C2/c with a ϭ 28.007(12), b ϭ 18.482(8),
Thermal stability; Diamagnetism
1 Introduction
large amounts of information have been collected. However,
there are still holes in our knowledge and new questions
keep on popping up. The present report concerns Co(thd)3.
As seen from the compilation in Fig. 1, this complex be-
longs to the category for which structural information is
lacking and one of the main intentions of the present ongo-
ing project has been to contribute to the mending of the
incompletenesses in Fig. 1. For natural reasons there are
more structural information available for the M(acac)2 and
M(acac)3 series with the simpler acac ligand than for the
homologous thd complexes.
For more than a century [1] it has been commonly recog-
nized that metal-organic complexes with β-diketonato li-
gands such as (acac)Ϫ [ϭ (C5H7O2)Ϫ; derivative of
H(acac) ϭ C5H8O2 ϭ acetylacetone ϭ pentane-2,4-dione]
and (thd)Ϫ [ϭ (C11H19O2)Ϫ; derivative of H(thd) ϭ
C11H20O2 ϭ 2,2,6,6-tetramethylheptane-3,5-dione; alterna-
tively dpm for dipivaloylmethane] have remarkable thermal
stability. More recently their suitable sublimation properties
have given them a unique niche for utilization as precursors
in the growth of thin films.
A closer inspection of the structural data behind the ref-
erences in Fig. 1 unveils a couple of notable features. First,
there appear apparently unmotivated breaks in the
structural systematics both along columns and rows of the
matrix in Fig. 1. Second, several complexes appear with
more than one modification (see the number in the upper
left corner of the boxes in Fig. 1). However, these at first
glance, only trend-breaking traits are probably to a large
extent intimately interwoven. This is nicely illustrated in a
most interesting, recent article by Geremia and Demitri [30].
This remarkable report accounts for a total of five structur-
ally distinct modifications of Mn(acac)3 which are inter-
related as low-/high-temperature modifications and meta-
stable variants obtained by re-crystallization from different
solvents. With this valuable erudition in mind we ap-
proached the targeted structural investigation of Co(thd)3.
The structural arrangement of these β-diketonato com-
plexes has naturally occupied the scientific community over
the century-long period [1Ϫ3]. In fact, the first crystal
structure determinations of the M(acac)3 complexes with
M ϭ Sc, Fe, Ga, and In were performed [3] already in the
infancy of the X-ray diffraction.
The interest in the structural aspects of the metal β-dike-
tonato complexes has been maintained over the years and
* Prof. Dr. A. Kjekshus
Department of Chemistry, University of Oslo
P.O.Box 1033, Blindern
N-0315 Oslo / Norway
Tel.: ϩ47-22855560; Fax: ϩ47-22855441
E-mail: arne.kjekshus@kjemi.uio.no
Z. Anorg. Allg. Chem. 2008, 634, 247Ϫ254
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
247

˚
M. A. K. Ahmed, H. Fjellvag, A. Kjekshus, P. D. C. Dietzel
both reactant and solvent. [0.0028 mol of Co(thd)2 and ca. 30 mL
H(thd); 5 hours (h) at b. p., cooled to room temperature (r. t.),
filtered, vacuum dried at 60 °C, and finally re-crystallized from
EtOH; dark-green crystalline product of 1, yield ca. 95 %.] (ii) Hex-
ane used (with strict measures to avoid fire ignition) as solvent.
[0.0064 mol of Co(thd)2 was refluxed in ca. 200 mL of hexane; 3
days (d) at b. p., otherwise as specified above, yield ca. 75 %.] (iii)
EtOH (absolute) used as solvent. [0.0064 mol of Co(thd)2 and ca.
150 mL EtOH; 5 h at b. p., filtered off, cooled to r. t., left for open-
to-air evaporation until dryness, re-dissolved in EtOH, the entire
procedure repeated twice, final product treated as for procedure i,
yield ca. 45 %.]
Fig.
1 Accessibility of crystal structure data for M(acac)2,
M(thd)2, M(acac)3, andM(thd)3 complexes of M ϭ ScϪZn as indi-
cated by the references in appropriate boxes. Shaded boxes mark
complexes that are likely to be non-existent for valence reasons.
Empty boxes refer to complexes that either have not been synthesi-
zed or not properly structure determined. The number in the upper
left corner of the boxes mediates how many crystallographically
distinct modifications that have been reported for the particular
complex.
As a somewhat modified variant of the above procedure i, CoO
was introduced as starting material. This series of syntheses was
soon extended to include the monoxides of M ϭ Mn, Fe, and Ni
for which we report successful synthesis of M(thd)3 from MnO,
FeO, and CoO. First 0.015 mol of MO was refluxed with 30 mL of
H(thd) for 24 h in O2 atmosphere. The reaction was then stopped
and the mixture was cooled to r. t. The liquid phase was evaporated
until dryness under vacuum, the thus formed M(thd)3 was ex-
tracted by acetone, and afterwards this solvent was evaporated at
r. t., either open-to-air or under vacuum. The products were finally
purified by sublimation and re-crystallization from EtOH.
Attention has also been paid to the preparation of
Co(thd)3 along different routes, e.g., by reacting CoIII salts
with H(thd) under basic conditions [40, 41] or from CoII
salts and H(thd) with H2O2 as oxidizing agent [42Ϫ46]. We
also report on the preparation of the M(thd)3 complexes of
M ϭ Mn, Fe, and Co by direct reaction of MII oxides,
H(thd), and O2 (g).
Elemental analysis for various samples of the different modifi-
cations of Co(thd)3 (C33H57O6Co) gave virtually identical results
for all samples tested; typical: C 65.58 (calc. 65.11), H 9.52 (9.44),
O 15.46 (15.77), Co 9.48 (9.68) %.
2 Experimental Section
Reactants and solvents
CoCl2 ·6H2O (Fluka, p.a., Ն98 %), MnO (Aldrich, 99 %), FeO
(Aldrich, 99.9 %), CoO (Sigma-Aldrich, 99.99ϩ %), NiO (L. Light,
99.95 %), H(thd) (Aldrich, purum, Ն98 %), absolute ethanol
(Arcus, prima), methanol (Merck, p.a.), hexane (Fluka, p.a.,
Ն99.5 %), heptane (Fluka, p.a., Ն99.5 %), and H2O2 [Prolabo
(BHD), ϳ33 %] were used as reactants and/or solvents without
further purification.
Powder X-ray diffraction (PXD)
All samples were characterized by PXD at 22 °C with a Siemens
D5000 diffractometer (capillary geometry) using monochromatic
˚
CuKα1 radiation (λ ϭ 1.540598 A) from an incident-beam Ge
monochromator. The detector was a Brown PSD. The diffraction
patterns were collected over the 2θ range 3Ϫ90° and auto-indexed
with help of the DICOVL [47] and TREOR [48] program packages
and unit-cell dimensions (calibrated to Si standard) were obtained
by least-squares refinements using the METRIC program [49].
Syntheses
All syntheses were performed in round-bottomed flasks equipped
with reflux condenser and magnetic stirrer. The reaction mixtures
were refluxed at the boiling point (b. p.) for different periods of
time and exposed to oxygen or inert atmosphere as specified below.
Unless otherwise mentioned all solid products were filtered through
a sinter-glass funnel and washed. The products were then dried
under vacuum and finally cleaned by sublimation (Büchi-type
B-580 sublimation apparatus at ca. 0.5 mbar pressure) and/or re-
crystallized from various solvents. The phase purity of all final
products was ascertained by powder X-ray diffraction (PXD).
Single-crystal X-ray diffraction (SXD) analysis
Crystals were mounted on thin glass fibers on brass pins. Intensity
data were collected at 295 and 100 K (MoKα radiation) on a
Bruker D8 Apex II diffractometer equipped with an Oxford Cryos-
ystems Cryostream, Plus device. The data were integrated with
SAINT [50] and corrected for absorption using SADABS [51].
Structures were solved using direct methods with the program
SIR2004 [52] or SHELXS [53] and refined using full-matrix least
squares against ΗFΗ2 with the SHELXL [54] program suite. All non-
hydrogen atoms were refined allowing for anisotropic displacement
parameters. Hydrogen atoms were assigned to idealized positions
and refined with isotropic thermal parameters proportional to the
thermal parameter of the atom to which they are attached. The r.
t. modification 1 undergoes a phase transition upon cooling to
100 K. The space group of the low-temperature modification 2
Several batches of Co(thd)3 were prepared according to different
procedures. (In the following text the various modifications of a
given complex are usually referred to by bold-faced numbers.) The
room-temperature stable modification of Co(thd)3 (1) was obtained
phase pure (overall yield ca. 25 %) by oxidizing CoCl2 ·6H2O with
aqueous solutions of H2O2 in the presence of H(thd) as described
in Ref. [42].
¯
(C2/c) is a subgroup of the space group of 1 (P3c1). Due to dis-
order of the molecules in 2, restraints idealizing the tert-butyl
groups were applied in the structural refinement of this modifi-
cation. The crystal structure of the metastable modification 3 of
Co(thd)3 suffers even more from disorder of the tert-butyl groups.
Modification 1 was also prepared by oxidation of Co(thd)2 with
O2 from air or oxygen flasks (with control experiments under inert
atmospheres for prolonged reaction periods): (i) H(thd) used as
248
www.zaac.wiley-vch.de
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2008, 247Ϫ254

Syntheses, Structures, and Polymorphism of β-Diketonato Complexes
Crystal data, data collection information, unit-cell dimensions, and
R factors for the reported structures are listed in Table 1. The good
agreement between the observed and calculated PXD patterns for
modifications 1 and 3 in Fig. 2 provides additional evidence for the
correctness of structures discussed in Sect. 3.7.
Elemental analysis
All elemental analyses were determined by the standard combus-
tion technique at Ilse Beetz.
Thermal analysis
Table 1 Crystallographic data for Co(thd)3, modifications 1Ϫ3.
Thermogravimetric (TG) analysis was performed with a Perkin-
Elmer TGA7 and DTA system in N2 atmosphere. Quartz con-
tainers were used as sample holders. The heating rate was 5 °C
minϪ1, the temperature interval covered was 40Ϫ700 °C, and the
sample mass was 15Ϫ30 mg.
Modification
1
2
3
Formula
C33H57O6Co C33H57O6Co C33H57O6Co
608.72
295
M / g molϪ1
T / K
608.72
100
608.72
100
Crystal system
Space group
trigonal
P3c1
18.8100(10)
18.8100(10)
18.815(2)
90
90
120
5765.2(8)
6
1.052
monoclinic
C2/c
28.007(12)
18.482(8)
21.356(9)
90
97.999(5)
90
10947(8)
12
orthorhombic
Pnma
19.2394(15)
18.8795(15)
10.7808(9)
90
90
90
3915.9(5)
4
1.033
¯
˚
a / A
Magnetic susceptibility
˚
b / A
˚
c / A
Magnetic susceptibility measurements were performed (tempera-
ture range 2Ϫ300 K, magnetic field 1 kOe) with a superconducting
quantum-interference-device instrument (Magnetic Property Meas-
urement System; Quantum Design). The sample mass was
15Ϫ30 mg and the samples were contained in gelatine holders dur-
ing the measurements.
α / °
β / °
γ / °
3
˚
V / A
Z
D / g cmϪ3
1.108
0.51
µ / mmϪ1
0.48
0.47
2θ range / °
4.4Ϫ36
41164
0.078
3497
1551
238 / 6
0.050
0.180
0.99
Ϫ0.62/0.33
4.6Ϫ36.2
29174
0.121
6843
3815
569 / 0
0.069
0.211
1.02
Ϫ0.45/0.72
4.3Ϫ39.4
12422
0.037
1824
1276
329 / 47
0.059
0.193
1.07
Ϫ0.17/0.28
No. reflections measured
Rint
3 Results and Discussion
3.1 Oxidation of CoII to CoIII
No. unique reflections
No. observed reflections
No. parameters / restraints
R1 [I>2σ(I)]
wR2 (all data)
GOF
Solutions of Co(thd)2 in organic solvents such as EtOH
have dark-red to purple color. On exposure to air at r. t.
(e.g., during slow evaporation) the solution gradually be-
comes dark-green and finally deposition of a dark solid
product which proved to be a mixture of 1 and an unidenti-
fied colorless solid phase. The yield of 1 depends on the
exposure time and the solvent used. The observed darken-
ing of the solution is due to the formation of 1 which in
concentrated solution appears as almost black. [Deoxygen-
ated ethanolic solutions of Co(thd)2 remain unaffected even
after prolonged refluxing at b. p. for more than a week.]
The formation of 1 accordingly implies oxidation of CoII to
CoIII by the oxygen dissolved in the solvent. The instability
of the oxidation state CoII in solution is well recognized [42,
44, 55].
Ϫ3
˚
Final difference peaks/eA
For reactions performed in O2 atmosphere it is found that
virtually complete oxidation of CoII to CoIII had occurred.
However, solid products thus obtained with EtOH as sol-
vent proved to be contaminated with a new mixed-ligand
complex. Structure determination (which will be reported
separately) has shown that this complex has the compo-
sition Co3(thd)3(OEt)4(tert-BuCOO). Some thd groups
have been oxidized and degraded to smaller species like tert-
BuCOO (C5H9O2) which in turn enters as ligand in the
complex.
3.2 Role of H(thd) as reactant and solvent during
oxidation of Co(thd)2
Fig. 2 Observed (black) and calculated (grey) PXD patterns for
Co(thd)3, modifications 1 and 3. Top: 1 (295 K; trigonal); bottom:
3 (295 K; orthorhombic). The calculated PXD pattern for modifi-
cation 3 (metastable) is based on atomic co-ordinates derived from
SXD data at 100 K. The experimental patterns contain reflections
(marked asterisks) from Si standard.
When H(thd) was used as both reactant and solvent during
oxidation experiments with Co(thd)2 it was found that the
solid product obtained after evaporation of the solvent at r.
Z. Anorg. Allg. Chem. 2008, 247Ϫ254
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
249

˚
M. A. K. Ahmed, H. Fjellvag, A. Kjekshus, P. D. C. Dietzel
t. consisted of 1 as the only crystalline phase. However, TG
investigation suggested that the product also contained at
least one other phase [the solvated variant (4) of 1 discussed
below] which accordingly had to be amorphous. For reac-
tions performed in pure O2 atmosphere under refluxing
conditions and short reaction periods (only a few hours) the
PXD characterization showed that a new, third crystalline
modification of Co(thd)3 (3; metastable phase, see Sect. 3.5)
was contained in the product in addition to 1 and 4. Heat
treatment of product mixtures of 1, 3, and 4 at 125Ϫ130 °C
converted the entire batches into Co(thd)3 (according to TG
and elemental analyses), but PXD examination unveiled
mixtures of 1 and 3. The composition of the amorphous
phase 4 is stipulated as Co(thd)3 ·xH(thd); x ഠ 0.06 (x
roughly estimated from the TG data); the possibility of a
hydrated variant was ruled out by a series of experiments in
which appropriate amounts of H2O were added to reaction
mixtures at various stages of the process. Thus there exist at
least three distinct crystalline phases with the composition
Co(thd)3 (modifications 1Ϫ3). 2 is a low-temperature vari-
ant of 1, and 3 is metastable to 1 as demonstrated by the
finding that 3 is converted to 1 by re-crystallization from
hexane or EtOH or storing at r. t. for a prolonged period
of time. In addition to modifications 1Ϫ3 comes then the
just mentioned amorphous solvated form (4).
However, for refluxing at the b. p. the rates have increased
to moderate even in air and the yields have increased corre-
spondingly. The reaction is faster in pure oxygen atmos-
phere and the yields have become quite appreciable for syn-
thesis at b. p.: 78 % for Mn(thd)3; 60 % for Fe(thd)3; and
85 % for Co(thd)3 (note: the yield specifications refer to
amounts obtained after sublimation cleaning). The con-
clusion is accordingly that these monoxides can be oxidized
directly to M(thd)3 by oxygen in the presence of H(thd).
The yields are good and the only real disadvantage with
this method is that it involves the rather expensive chemical
ingredient H(thd). The complete elimination of the last
traces of the H(thd) solvent by evaporation is also a chal-
lenge, but this can be overcome by a final re-crystalliza-
tion treatment.
A similar procedure for preparation of the corresponding
acac complexes has been patented by Chaudhari et al. [56],
but their recipes involve the use of the more powerful oxid-
izing agent H2O2.
Attempts to prepare Ni(thd)3 with the same method have
so far been in vain. There are certain indications of onset
of an analogous reaction by the formation of a thin creamy
brownish surface cover on the reaction mixture [NiO,
H(thd), and O2(solv.)], but this step could not be advanced
into amounts which could be pursued by characterization
methods.
3.3 Oxidation of Co(thd)2 in hexane or heptane
3.5 Sublimation and thermal stability
With hexane as solvent instead of EtOH, (the rest of the
procedure kept unchanged) no solid product was obtained
on filtration. The solid products obtained after evaporation
of the solvent were either pure Co(thd)2 (preparation in in-
ert atmosphere and drying under vacuum) or mixtures of
Co(thd)2, 1, and an unidentified colorless solid (the same
procedure but open-to-air evaporation at r. t.). When the
experiments were performed in O2 atmosphere mixtures of
1 and 3 or 3 and 4 were obtained.
On the other hand, when heptane was used instead of
hexane somewhat different results were obtained under O2
atmosphere. Mixtures of 1 and 3 (yield ca. 40 %) together
with an amorphous solid (recovered after filtration of the
solution) were obtained on evaporation of the solvent under
vacuum. Elemental analyses of the amorphous phase gave
Modifications 1 and 3 can be subjected to total subli-
mation, (ca. 90 °C and ca. 0.5 mbar; according to TG: onset
at ca. 150Ϫ160 °C, completed at ca. 260Ϫ280 °C under N2
atmosphere, 5 °C minϪ1). The solid product, obtained on
sublimation of 1 is 3 according to PXD. Hence, sublimation
treatment of 1 leads to the metastable modification 3 of
Co(thd)3. The subsequent conversion of 3 back to 1 provide
an additional proof for 3 as a proper (but metastable) modi-
fication of Co(thd)3. Crystals of 3 are obtained by sub-
jecting 1 to sublimation at 90Ϫ100 °C. If only powder speci-
mens of 3 are needed we recommend heat treatment of
Co(thd)3 in evacuated sealed tubes at 100Ϫ110 °C followed
by quenching to r. t. Through the just described cycle pro-
cess the conversion of modification 1 to 3 and back again
to 1 becomes quasi reversible.
an empirical composition close to C7H12O4Co
(RCOO)Co(OOCRЈ) (R CH3CH2CH2 and RЈ
CH3CH2). This shows that some degradation and/or oxi-
dation of thd groups has occurred as consequences of the
comparatively high b. p. (98Ϫ99 °C) of heptane.
ϭ
ϭ
The TG analyses of phase 4 were based on samples where
this phase was mixed with 1 and/or 3 (phase 4 was never
prepared phase pure in this study). The TG curves for such
mixtures showed an initial 2Ϫ3 % weight loss between ca.
95 and 115 °C, whereas the rest of the TG curves followed
the course for complexes 1 and 3.
ϭ
Slight temperature shifts (at most ca. 10 °C) seen between
the TG curves for phases 1, 3, and 4 (above ca. 115 °C) can
easily be attributed to variations in sample characteristics
(e.g., particle size).
3.4 Reaction between the monoxides of Mn, Fe, and
Co and H(thd) in O2 atmosphere
The monoxides of Mn, Fe, and Co react with H(thd) in O2
atmosphere to form M(thd)3 in different yields and with
reaction rate depending on temperature, oxygen partial
pressure, and the particular MO reactant under investi-
gation. At r. t. there is almost neglectable reaction in air.
3.6 Magnetic properties
The various variants of Co(thd)3 (1, 2, 3, and 4) exhibit the
unequivocal characteristics of diamagnetism, viz. a small
250
www.zaac.wiley-vch.de
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2008, 247Ϫ254

Syntheses, Structures, and Polymorphism of β-Diketonato Complexes
negative [χg ഠ Ϫ0.16 ϫ 10Ϫ6 e.m.u./g; as-observed average
for all variants] virtually temperature-independent magnetic
susceptibility. These findings are in full accordance with the
expectations for proper CoIII complexes.
Fig. 3 X-ray molecular structures of Co(thd)3 modifications 1Ϫ3. (a) The two crystallographically independent molecules of 1 at 295 K.
Atomic labeling schemes are indicated, hydrogen atoms and labeling of the tert-butyl groups are omitted for clarity. Thermal displacement
ellipsoids are drawn at the 50 % probability level. (b) The two crystallographically independent molecules of 2 at 100 K. Atomic labeling
schemes are indicated, hydrogen atoms and labeling of the tert-butyl groups are omitted for clarity. Thermal displacement ellipsoids are
drawn at the 50 % probability level. (c) The carbon atoms (atomic labeling scheme indicated) of the two enantiomeric forms of modification
3 at 100 K. Light grey shading and black toning are used to emphasize the disorder. It is renounced to show thermal displacement
parameters for the carbon atoms and the hydrogen atoms of the tert-butyl groups are omitted for clarity.
Z. Anorg. Allg. Chem. 2008, 247Ϫ254
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
251

˚
M. A. K. Ahmed, H. Fjellvag, A. Kjekshus, P. D. C. Dietzel
˚
Table 2 Selected bond lengths/A and angles/° for Co(thd)3
Table 2 (continued)
a. Modification 1 (295 K)
c. Modification 3 (100 K, metastable)
C1ϪC2
C2ϪO1
1.541 (5)
1.272 (4)
1.377 (4)
1.377 (4)
1.526 (5)
1.273 (4)
1.398 (4)
175.05 (13)
89.43 (13)
96.26 (13)
87.26 (10)
87.61 (11)
C9ϪC10
C10ϪO3
C10ϪC11
Co1ϪO1
Co2ϪO3
Co2ϪO2
1.390 (4)
1.274 (4)
1.519 (5)
1.876 (2)
1.864 (2)
1.868 (2)
C1ϪO1
C1ϪC2
C2ϪC3
C3ϪO2
C4ϪO2
C4ϪC5
C5ϪC6
C6ϪO3
1.17 (2)
1.44 (2)
1.40 (2)
1.29 (2)
1.17 (2)
1.35 (2)
1.41 (2)
1.14 (2)
1.14 (2)
124.0 (2)
122.0 (2)
123.0 (2)
1310.0 (2)
121.0 (2)
121.4 (13)
126.0 (2)
124.0 (2)
119.7 (13)
88.6 (4)
87.5 (2)
87.5 (2)
C7ϪC8
C8ϪC9
1.36 (2)
1.46 (2)
1.172 (14)
1.855 (5)
1.855 (5)
1.857 (6)
1.864 (8)
1.867 (5)
1.867 (5)
92.3 (2)
92.3 (2)
179.7 (3)
179.5 (2)
91.9 (3)
92.7 (2)
87.5 (3)
91.9 (3)
179.5 (2)
92.7 (2)
87.5 (3)
87.6 (4)
C2ϪC3
C9ϪO4
C3ϪC2i
Co1ϪO2i
C7ϪC8
C8ϪO2
C8ϪC9
Co1ϪO2
Co1ϪO4
Co1ϪO3
O1iiϪCo1ϪO1
O1ϪCo1ϪO1iii
O1ϪCo1ϪO1i
O1ϪCo1ϪO1iv
O3vϪCo2ϪO3
O3ϪCo2ϪO2vi
O3ϪCo2ϪO2v
O3ϪCo2ϪO2
O2viϪCo2ϪO2
89.21 (11)
175.45 (10)
95.51 (10)
87.85 (11)
Co1ϪO1
C7ϪO1
Co1ϪO1i
O1ϪC1ϪC2
C3ϪC2ϪC1
O2ϪC3ϪC2
O2ϪC4ϪC5
C4ϪC5ϪC6
O3ϪC6ϪC5
O1ϪC7ϪC8
C7ϪC8ϪC9
O4ϪC9ϪC8
O2iϪCo1ϪO2
O2iϪCo1ϪO4
O2ϪCo1ϪO4
O2iϪCo1ϪO3
O2ϪCo1ϪO3
O4ϪCo1ϪO3
O2iϪCo1ϪO1
O2ϪCo1ϪO1
O4ϪCo1ϪO1
O3ϪCo1ϪO1
O2iϪCo1ϪO1i
O2ϪCo1ϪO1i
O4ϪCo1ϪO1i
O3ϪCo1ϪO1i
O1ϪCo1ϪO1i
Symmetry codes: (i) y, x, 1/2Ϫz; (ii) Ϫx, Ϫxϩy, 1/2Ϫz; (iii) xϪy, Ϫy, 1/2Ϫz;
(iv) Ϫy, xϪy, z; (v) 1Ϫxϩy, 1Ϫx, z; (vi) 1Ϫy, xϪy, z.
b. Modification 2 (100 K)
C1ϪC2
C2ϪC8
C2ϪC3
C2ϪC9
C3ϪO1
C3ϪC4
C4ϪC5
C5ϪO2
C5ϪC6
C6ϪC7
C6ϪC10
1.522 (10)
1.518 (11)
1.527 (10)
1.554 (11)
1.252 (8)
1.411 (10)
1.399 (10)
1.272 (8)
1.532 (10)
1.508 (11)
1.519 (11)
1.534 (10)
1.498 (10)
1.533 (10)
1.534 (10)
1.570 (11)
1.270 (7)
1.378 (9)
1.397 (9)
1.269 (7)
1.517 (9)
1.462 (10)
1.500 (11)
1.532 (11)
1.486 (10)
1.494 (9)
1.524 (9)
1.565 (11)
1.278 (8)
1.409 (9)
1.398 (9)
96.2 (2)
C27ϪO6
1.264 (8)
1.532 (9)
1.529 (10)
1.533 (9)
1.540 (9)
1.526 (9)
1.520 (9)
1.522 (9)
1.535 (9)
1.280 (7)
1.398 (9)
1.264 (7)
1.527 (9)
1.521 (9)
1.538 (9)
1.543 (9)
1.525 (9)
1.503 (9)
1.541 (9)
1.545 (10)
1.265 (7)
1.402 (8)
1.858 (4)
1.862 (4)
1.868 (4)
1.872 (4)
1.876 (4)
1.881 (4)
1.865 (4)
1.872 (4)
1.877 (4)
95.3 (2)
C27ϪC28
C28ϪC29
C28ϪC32
C28ϪC33
C35ϪC36
C36ϪC42
C36ϪC43
C36ϪC37
C37ϪO7
Symmetry codes: (i) x, 1/2Ϫy, z.
is virtually identical with CoϪOket. bond distances in the
range 1.86Ϫ1.88 A (Table 2) and Oket.ϪCoϪOket.bond
˚
C38ϪC39
C39ϪO8
C6ϪC11
angles from 86.8 to 96.2°. The refinement reliability factors
[wR2 (all data), see Table 1] came out somewhat larger than
desirable, but we attribute this grievance to imperfections
of the simple model presently used to account for structural
disorder and thermal vibrations. This imperfection is, e.g.,
particularly conspicuous in the large scatter of the bond
distances within the ketonato rings of complex 3 (Table 2c).
The X-ray molecular structures of modifications 1Ϫ3 of
Co(thd)3 match closely those in corresponding parts of
Co(acac)3 [36]. The actual crystal structures of Co(thd)3
and Co(acac)3 differ, however, with regard to the packing
of the molecular species. This distinction clearly originates
from the space demands of the larger and more bulky tert-
butyl groups of the thd ligands.
There are two possible chiral isomers for the Co(thd)3
molecules, a left helical Λ form and a right helical ∆ form.
Since 1 and 2 crystallize in centrosymmetric space groups
their asymmetric unit contains two crystallographically in-
dependent Co(thd)3 molecules which correspond to the two
possible isomers. In modification 1 at 295 K the majority of
the tert-butyl groups are rotationally disordered (Fig. 3a),
whereas they are well localized in modification 2 at 100 K
(Fig. 3b). The involved transformation is concomitant with
the change of space group.
C12ϪC13
C13ϪC20
C13ϪC14
C13ϪC19
C14ϪO3
C14ϪC15
C15ϪC16
C16ϪO4
C39ϪC40
C40ϪC45
C40ϪC41
C40ϪC44
C46ϪC47
C47ϪC51
C47ϪC48
C47ϪC50
C48ϪO9
C48ϪC49
Co1ϪO5
Co1ϪO6
Co1ϪO2
Co1ϪO3
C16ϪC17
C17ϪC18
C17ϪC21
C17ϪC22
C23ϪC24
C24ϪC31
C24ϪC25
C24ϪC30
C25ϪO5
C25ϪC26
C26ϪC27
O5ϪCo1ϪO6
O5ϪCo1ϪO2
O6ϪCo1ϪO2
O5ϪCo1ϪO3
O6ϪCo1ϪO3
O2ϪCo1ϪO3
O5ϪCo1ϪO4
O6ϪCo1ϪO4
O2ϪCo1ϪO4
O3ϪCo1ϪO4
O5ϪCo1ϪO1
O6ϪCo1ϪO1
Co1ϪO4
Co1ϪO1
Co2ϪO8
Co2ϪO9
Co2ϪO7
O2ϪCo1ϪO1
O3ϪCo1ϪO1
O4ϪCo1ϪO1
O8ϪCo2ϪO8i
O8ϪCo2ϪO9
O8ϪCo2ϪO9i
O9ϪCo2ϪO9i
O8ϪCo2ϪO7i
O9ϪCo2ϪO7i
O8ϪCo2ϪO7
O9ϪCo2ϪO7
176.2 (2)
86.8 (2)
90.1 (2)
87.3 (2)
87.8 (2)
87.8 (2)
175.0 (2)
89.4 (2)
95.7 (2)
87.1 (2)
174.7 (2)
88.7 (2)
89.6 (3)
175.6 (2)
87.4 (2)
95.8 (3)
95.6 (2)
87.6 (2)
88.2 (2)
88.9 (2)
88.6 (29)
Symmetry codes: (i) 2Ϫx, y, 1/2Ϫz.
Instead of occupying two distinct positions in the crystal
structure, the isomeric Co(thd)3 molecules are located in a
disordered fashion on the same position in the metastable
modification 3 where they are transformed into each other
by a mirror plane (Fig. 3c). The tert-butyl groups of the two
enantiomers here superimpose on approximately the same
position in a manner which facilitates the packing of such
disordered molecules.
3.7 Crystal structures
The crystal structures of 1 (295 K), 2 (100 K), and 3 (100 K)
comprize packed mononuclear Co(thd)3 molecules. The in-
ner core of these complex units consists of Co surrounded
by six oxygen atoms (from three different bi-dentate thd
ligands) in somewhat deformed octahedral configurations.
The atomic arrangement at the Co(Oket.)6 cores (see Fig. 3)
As mentioned the space groups of the structures of modi-
fications 1 and 2 are related as super- versus sub-group
252
www.zaac.wiley-vch.de
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2008, 247Ϫ254

Syntheses, Structures, and Polymorphism of β-Diketonato Complexes
Acknowledgements. This project has received financial support from
the Research Council of Norway.
symmetry, and the unit-cell volume per formula unit in-
creases from 2 at 100 K to 1 at 295 K is some 5.1 %. The
most easily recognized distinction between these structures
is the increased thermal movements of the atoms, in par-
ticular the onset of rotational disorder of the tert-butyl
groups, on going from 100 to 295 K. This is probably the
driving force of the change in crystal symmetry. Attempts
have been made to determine the transformation tempera-
ture by differential scanning calorimetry (DSC), but these
efforts have hitherto been in vain. Nevertheless, we feel con-
fident that the transition is to be classified as second order.
The relation between 1 and 3 is, as pointed out in Sect.
3.5, not governed by a reversible transformation process al-
though it is also for these modifications of Co(thd)3 pos-
sible to identify corresponding unit-cell directions. As a
proper metastable phase 3 exhibits a less efficient packing
of the complex molecules than 1 [the unit-cell volume per
formula unit is 1.9 % larger in 3 at 100 K than 1 at 295 K;
noting also that V(1) / V(3) ഠ 3/2].
References
[1] G. T. Morgan, H. W. Moss, J. Chem. Soc. Trans. 1914, 105,
189.
[2] G. T. Morgan, H. D. K. Drew, J. Chem. Soc. Trans. 1921,
119, 1058.
[3] W. T. Astbury, Proc. Roy. Soc. London 1926, A112, 448.
[4] F. A. Cotton, C. E. Rice, G. W. Rice, Inorg. Chim. Acta 1977,
24, 231.
[5] S. Shibata, S. Onuma, H. Inoue, Inorg. Chem. 1985, 24, 1723.
[6] F. A. Cotton, G. W. Rice, New J. Chem. 1977, 1, 301.
[7] S. Shibata, S. Onuma, A. Iwase, H. Inoue, Inorg. Chim. Acta
1977, 25, 33.
[8] F. A. Cotton, R. C. Elder, Inorg. Chem. 1965, 4, 1145.
[9] J. Burgess, J. Fawcett, D. R. Russell, S. R. Gilani, Acta Crys-
tallogr. 2000, C56, 649.
[10] G. J. Bullen, R. Mason, P. Pauling, Inorg. Chem. 1965, 4, 456.
[11] M. B. Hursthouse, M. A. Laffey, P. T. Moore, D. B. New, P.
R. Raithby, P. Thornton, Dalton Trans. 1982, 307.
[12] Z. A. Starikova, E. A. Shugam, J. Struct. Chem. 1969, 10, 267.
[13] P. C. Lebrun, W. D. Lyon, H. A. Kuska, J. Crystallogr. Spec-
trosc. Res. 1986, 16, 889.
[14] M. J. Bennett, F. A. Cotton, R. Eiss, Acta Crystallogr. 1968,
B24, 904.
[15] F. A. Cotton, C. Y. Liu, C. A. Murillo, X. Wang, Inorg. Chem.
2003, 42, 4619.
When we wanted to continue our bond-valence analyses
with the structural data for the Co(thd)3 modifications we
experienced an unexpected discrepancy. Applying the
CoIIIϪO bond-valence parameter of Brese and O’Keeffe
˚
[57] (1.70 A) to the mutually consistent interatomic CoϪO
distances in the various modifications of Co(thd)3 one ob-
tains bond valences in the range of 0.613Ϫ0.652 and some
25 % too larger bond-valence sums for Co (3.73Ϫ3.83). The
ketonato oxygen atoms, on the other hand, obtain normal
bond-valence sums of around two. So the approach which
worked quite satisfactory for various thd complexes of VIV
[58] and CoII [16] gives unsatisfactory valence states for Co
in Co(thd)3. The unavoidable conclusion appears therefore
to be that the tabulated bond-valence parameter for
CoIIIϪO is incorrect.
˚
[16] M. A. K. Ahmed, H. Fjellvag, A. Kjekshus, P. D. C. Dietzel,
Z. Anorg. Allg. Chem. 2007, 633, 1371.
[17] F. A. Cotton, J. J. Wise, Inorg. Chem. 1966, 5, 1200.
[18] W. H. Watson, W. W. Holley, Croat. Chem. Acta 1984, 57, 467.
[19] S. Sans-Lenain, A. Gleizes, Inorg. Chim. Acta 1993, 211, 67.
[20] S. Patnaik, T. N. G. Row, L. Raghunathan, A. Devi, J. Gos-
wami, S. A. Shivashankar, S. Chandrasekaran, W. T. Robin-
son, Acta Crystallogr. 1996, C52, 891.
[21] F. A. Cotton, J. S. Wood, Inorg. Chem. 1964, 3, 245.
[22] T. J. Anderson, M. A. Neuman, G. A. Melson, Inorg. Chem.
1973, 12, 927.
[23] S. S. Yun, I.-H. Suh, S.-S. Choi, T.-H. Kim, S. Lee, J. Coord.
Chem. 1999, 47, 315.
[24] B. Morosin, H. Montgomery, Acta Crystallogr. 1969, B25,
1354.
An appreciable change of the CoIIIϪO bond-valence par-
˚
ameter to 1.61 A would indeed give the expected valence of
3 and thus remove the discrepancy. However, the tabulated
bond-valence parameters in Ref. [57] and parallel sources
are extracted from large collections of selected structural
˚
data. Hence, unless the tabulated value of 1.70 A represents
a trivial misprint, one would be confronted with a much
larger problem if one without further ado alters the param-
[25] F. S. Ruiz, S. M. Carrera, S. Garcia-Blanco, An. Fis. 1970,
66, 309.
˚
eter to say, the mentioned value 1.61 A. However, the rescue
[26] C. A. L. Filgueiras, A. Horn, R. A. Howie, J. M. S. Skakle, J.
L. Wardell, Acta Crystallogr. 2001, E57, m157.
[27] B. Morosin, Acta Crystallogr. 1965, 19, 131.
[28] J. P. Fackler, A. Avdeef, Inorg. Chem. 1974, 13, 1864.
[29] B. R. Stults, R. S. Marianelli, V. W. Day, Inorg. Chem. 1979,
18, 1853.
[30] S. Geremia, N. Demitri, J. Chem. Educ. 2005, 82, 460.
[31] R. B. Roof, Acta Crystallogr. 1956, 9, 781.
[32] J. Iball, C. H. Morgan, Acta Crystallogr. 1967, 23, 239.
[33] V. M. Padmanabhan, Proc. Indian Acad. Sci. 1958, A47, 329.
[34] B. Morosin, J. R. Brathovde, Acta Crystallogr. 1964, 17, 705.
[35] P. K. Hon, C. E. Peluger, J. Coord. Chem. 1973, 3, 67.
[36] G. J. Kruger, E. C. Reynhardt, Acta Crystallogr. 1974, B30,
822.
for the bond-valence concept appears paradoxically to be
that it is too simple. It is well recognized that bond dis-
tances depend not only on overall valence, but vary also
with chemical bond character, spin state, and co-ordination
environment (viz. co-ordination number as well as co-ordi-
nation geometry). According to such a more varied picture
of bond valence a given valence state may be allowed to
take several parameter values.
Available supporting information: CCDC 650381Ϫ650383 contain
the supplementary crystallographic data for this paper. These
data can be obtained free of charge via www.ccdc.cam.ac.uk/
data_request/cif, by emailing data_request@ccdc.cam.ac.uk, or by
contacting The Cambridge Crystallographic Data Centre, 12,
Union Road, Cambridge CB2 1EZ, UK; fax: ϩ44 1223 336033.
[37] V. L. Lisoivan, S. A. Gromilov, Izv. Akad. Nauk SSSR, Ser.
Khim. 1987, 2098.
Z. Anorg. Allg. Chem. 2008, 247Ϫ254
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
253

˚
M. A. K. Ahmed, H. Fjellvag, A. Kjekshus, P. D. C. Dietzel
[38] P. Magnus, A. H. Payne, M. J. Waring, D. A. Scott, V. Lynch,
Tetrahedron Lett. 2000, 41, 9725.
[49] WIN-INDEX Professional Powder Indexing, Version 3.0, Sie-
mens AG Analytical System AUT, Karlsruhe, 1996.
[50] Area-Detector Integration Software, V7.12A; Bruker-Nonius,
Inc.: WI, 2004.
[39] I. A. Baidina, P. A. Stabnikov, V. I. Alekseev, I. K. Igumenov,
S. V. Borisov, Zh. Strukt. Khim. 1986, 27, 102.
[40] E. W. Berg, N. M. Herrera, Anal. Chim. Acta 1972, 60, 117.
[41] R. J. York, W. D. Bonds, B. P. Cotsoradis, R. D. Archer, Inorg.
Chem. 1969, 8, 789.
[42] G. S. Hammond, D. C. Nonhebel, C.-H. S. Wu, Inorg. Chem.
1963, 2, 73.
[43] K. Ueno, H. Kobayashi, I. Yoshida, Mem. Fac. Eng. Kyushu
Univ. 1978, 38, 83.
[51] Bruker-Nonius, Inc. Madison, WI, 2004.
[52] M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini, G. L.
Cascarano, L. D. Caro, C. Giacovazzo, G. Polidori, R.
Spagna, J. Appl. Crystallogr. 2005, 38, 381.
[53] G. M. Sheldrick, SHELXS97 Ϫ Program for Crystal Structure
Solution. Institut für Anorganische Chemie der Universität
Göttingen, Germany, 1997.
[44] I. Yoshida, H. Kobayashi, K. Ueno, Bull. Chem. Soc. Japan
1972, 45, 2768.
[45] S. M. Schildcrout, Inorg. Chem. 1980, 19, 224.
[46] Y. Sakai, K. Endo, H. Sano, Int. J. Appl. Radiat. Isot. 1981,
32, 435.
[47] A. Boultif, D. J. Louer, J. Appl. Crystallogr. 1991, 24, 987.
[48] P.-E. Werner, Program TREOR, Version 5, University of
Stockholm, Stockholm, 1988; P.-E. Werner, L. Eriksson, M.
Westdahl, J. Appl. Crystallogr. 1985, 18, 367.
[54] G. M. Sheldrick, SHELXL97 Ϫ Program for Crystal Structure
Refinement, Institut für Anorganische Chemie der Universität
Göttingen, Germany, 1997.
[55] A. Flamini, V. Fares, A. Pifferi, Eur. J. Inorg. Chem. 2000, 537.
[56] M. K. Chaudhari, S. K. Dehury, S. S. Dhar, U. Bora, B. M.
Choudary, L. K. Mannepalli, US Patent 20040127690, 2004.
[57] N. E. Brese, M. O’Keeffe, Acta Crystallogr. 1991, B47, 192.
˚
[58] M. A. K. Ahmed, H. Fjellvag, B. Klewe, A. Kjekshus, Z.
Anorg. Allg. Chem. 2004, 630, 2311.
254
www.zaac.wiley-vch.de
 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2008, 247Ϫ254