Synthesis and Thermal Behavior of Heteroleptic γ-Substituted Acetylacetonate-Alkoxides of Titanium

Article abstract of DOI:10.1002/ejic.202100209

A series of heteroleptic titanium derivatives of general formula [Ti(OiPr)₂(R-acac)₂] with acetylacetonate ligands modified in the internal (γ- or 3-) position by different substituents (R=OAc, NO₂, Me, Et, Cl, Br) has been synthesized and completely characterized by liquid multinuclear NMR and FTIR. The influence of the nature of the group on the thermal stability of the different complexes was studied by thermogravimetric analysis (TGA) and gave the following decreasing stability ranking: H22°-acac radical, which triggers the decomposition.

Full text of DOI:10.1002/ejic.202100209

Full Papers
doi.org/10.1002/ejic.202100209
Synthesis and Thermal Behavior of Heteroleptic γ-
Substituted Acetylacetonate-Alkoxides of Titanium
Diane Bijou,[a, b] Thibaut Cornier,[a] Shashank Mishra,[a] Lynda Merzoud,[c] Henry Chermette,[c]
Erwann Jeanneau,[d] William Maudez,[b] Giacomo Benvenuti,[b] and Stéphane Daniele*[a, e]
A series of heteroleptic titanium derivatives of general formula
[Ti(OiPr)2(R-acac)2] with acetylacetonate ligands modified in the
internal (γ- or 3-) position by different substituents (R=OAc,
NO2, Me, Et, Cl, Br) has been synthesized and completely
characterized by liquid multinuclear NMR and FTIR. The
influence of the nature of the group on the thermal stability of
the different complexes was studied by thermogravimetric
analysis (TGA) and gave the following decreasing stability
ranking: H<NO2 <Cl<Me<Br<Et<OAc. DFT calculations
showed that this ranking followed the same order as the
stability of the keto-enolate form compared to the β-diketonate
form. Other DFT calculations also confirmed that for these
molecules the ligands degraded sequentially and that the
singular high reactivity of the Et-acac based complex could be
°
due to the rupture of the CÀ C bond to lead to a CH2 -acac
radical, which triggers the decomposition.
Introduction
inhibition of the growth of the layer due to a very low
degradation of the precursor. Improving the quality of deposits
therefore implies the design of more reactive precursors,
especially for deposition techniques such as Chemical Beam
vapor Deposition (CBVD),[5] which does not allow the additional
addition of a reactive gas.
Titanium tetraisopropoxide [Ti(OiPr)4]m (TTIP) is one of the most
widely used titanium alkoxide as a starting reagent for gas
phase TiO2 deposition processes due to its high volatility.[1]
However, it is a thermally unstable and extremely reactive
compound that generates a lot of handling difficulties. In order
to have a monomer structure and increase the thermal stability,
complexes based on disubstituted-diketonates of the general
formula [Ti(OiPr)2(β-dik)2] (β-dik = acac, 2,4-pentanedione; thd,
2,2,6,6-tetramethyl-3,5-heptanedione) have been extensively
investigated as precursors for the production of TiO2 thin films
in the gas phase (over 200 patents mention their uses).[2,3]
Mono-substituted β-diketonate dimer complexes of the general
formula [Ti(OiPr)3(β-dik)]2 are little employed because they are
generally converted into di-substituted derivatives by a redis-
tribution reaction of the ligand.[4] Such derivatives presented as
common features high thermal stability for good transportation
but lower volatility and mainly low reactivity. Their use are
generally accompanied by a high level of carbon in the final
layer requiring the use of a reactive gas (O2, H2O …) or even an
Few studies have been reported in the literature on the
volatility and thermal behavior of [Ti(OiPr)2(β-dik)2] compounds
by varying the nature of the group at the outer position of the
β-diketonate.[6] Hence, [Ti(OiPr)2(thd)2] exhibited a better degra-
dation than [Ti(OiPr)2(acac)2] while being less volatile. Mass
spectrometry studies have shown that [Ti(OiPr)2(thd)2] thermally
degraded to Ti(thd)3 or TiO(thd)2 species demonstrating that
the nature of the β-diketonate played a major role in the
complete decomposition of the precursor[3] and that one of the
starting point of the decomposition pathway was the cleavage
of the external alkyl group (i.e. by β-elimination).[7] The internal
(γ- or 3-) position of the acac ligand has often been modified to
allow the grafting of metal alkoxide complexes onto oxide
surfaces[8,9] but data exploiting this modification to study the
thermal degradation of such complexes were scarce and dealt
only with homoleptic systems M(β-dik)3.[10–14]
Herein, we report the syntheses of a series of heteroleptic
titanium derivatives [Ti(OiPr)2(R-acac)2] with acac ligands modi-
fied with different substituents (R=OAc, NO2, Me, Et, Cl, Br) at
the γ-position and the impact on their thermal stability by
thermogravimetric analysis (TGA) and Density Functional Theory
(DFT) calculation in order to reach optimal thermal degradation
of acac-based titanium complexes. DFT has become the
theoretical quantum approach of choice to calculate structure
and properties of molecular systems, thanks to its reasonable
computational effort required. We used it for a) estimating the
keto-enol forms ratio for the differently γ-modified β-diketones
ligands, and b) performing an energy decomposition analysis
which provides the interaction energy of e.g. a ligand with
respect to the remaining part of the complex. The energy
decomposition analysis, first introduced by Ziegler and Rauk in
[a] D. Bijou, T. Cornier, Dr. S. Mishra, Prof. S. Daniele
University of Lyon 1, IRCELYON UMR-CNRS 5256,
2 avenue Albert Einstein, 69626 Villeurbanne cedex; France
E-mail: stephane.daniele@univ-lyon1.fr
[b] D. Bijou, Dr. W. Maudez, Dr. G. Benvenuti
3D-oxide
70 rue Gustave Eiffel, 01630 Saint Genis Pouilly, France
[c] Dr. L. Merzoud, Prof. H. Chermette
University of Lyon, ISA UMR-CNRS 5280,
5 rue de la Doua, 69100 Villeurbanne cedex, France
[d] Dr. E. Jeanneau
University of Lyon, ISA, Centre de diffractométrie Henry Longchamp,
5 rue de la Doua, 69100 Villeurbanne cedex, France
[e] Prof. S. Daniele
Present address: Université Claude Bernard Lyon 1,
CNRS-UMR 5218, CP2M-ESCPE Lyon,
69616, Villeurbanne cedex, France
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/ejic.202100209
Eur. J. Inorg. Chem. 2021, 1–9
1
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
Scheme 1. Synthesis of heteroleptic Ti(IV) γ-modified β-diketonatealkoxides.
°
Table 1. Selected bond lengths (Å) and angles ( ) of derivatives 1, 3, and 6.
1979,[15] has been generalized by Bickelhaupt and Baerends,[16]
but, to our knowledge, has never been applied before to
interpret the thermal behavior of precursors.
NO2-acac (1)
Me-acac (3)
Br-acac (6)
TiÀ O5 (OR)
TiÀ O6 (OR)
TiÀ O1 (acac)
TiÀ O2 (acac)
TiÀ O3 (acac)
TiÀ O4 (acac)
O1-TiÀ O2
1.777 (3)
1.772 (3)
2.016 (3)
2.083 (3)
2.002 (3)
2.076 (2)
81.18 (11)
81.83 (10)
101.54 (13)
1.785 (2)
1.814 (2)
1.981 (2)
2.090 (2)
2.023 (2)
2.001 (2)
80.76 (9)
81.38 (9)
98.55 (11)
1.796 (3)
1.792 (3)
2.016 (3)
2.067 (3)
2.059 (3)
2.006 (3)
81.08 (12)
81.30 (12)
101.76 (15)
Results and Discussion
O3-TiÀ O4
Syntheses and structures
O5-TiÀ O6
Titanium diisopropoxide bis (γ-modified-acetylacetonate) com-
plexes of the general formula [Ti(OiPr)2(R-acac)2] [R=NO2 (1),
OAc (2), Me (3), Et (4), Cl (5)] have been synthesized by reacting
two equivalents of γ-modified diketone ligands and one
equivalent of [Ti(OiPr)4]m in n-hexane at room temperature and
were obtained in quantitative yield. The compound [Ti-
(OiPr)2(Br-acac)2] (6) was obtained through direct bromination
of compound [Ti(OiPr)2(acac)2] (A) in 82% yield, using N-
bromosuccinimide (NBS). (Scheme 1) All were soluble in most
organic solvents. The liquid 1H and 13C NMR spectra of all
compounds were in perfect agreement with their raw formula
and high purity. Compounds A, 2 and 4 were yellow-orange
liquids or oils while compounds 1, 3, 5 and 6 were in yellow
solid form. The liquid compounds could not be purified by
vacuum distillation but some of the solid compounds could be
purified by crystallization in hexane or hexane/isopropanol
mixture and characterized by single crystal X-ray diffraction.
[Ti(OiPr)2(NO2-acac)2] (1) and [Ti(OiPr)2(Me-acac)2] (3) crystal-
lized in the triclinic space group P-1 and [Ti(OiPr)2(Br-acac)2] (6)
in the monoclinic space group P21/n. They all adopted a
mononuclear structure, where the 6-coordinated titanium atom
was surrounded by two isopropoxyde ligands and two
bidentate β-diketonate ligands in cis position (Figure 1) as
generally observed.[6] The angles between two neighboring
compared well with the literature values reported for related
compounds.[4,6,17] Concerning the nitro group in γ-position of
the diketone, C-NO2 bond lengths [1.442 (3) and 1.454 (3) Å]
was in the normal range for this type of compound.[16] However,
the presence of nitro group in the γ-position of the β-
diketonate moiety brought in significant intermolecular H-
bonding (3.47–3.59 Å) between oxygen atoms of NO2 group
and hydrogen atoms in α-position of β-diketonate ligand. There
was also weak interaction present between oxygen atoms of β-
diketonate moieties and central methine hydrogen atoms on
isopropoxide ligands. Compared to the plane of β-diketonate
moiety, the NO2 group was slightly tilted with the torsion angle
°
varying in the range 47.5–51.6 . For compound 3, the CÀ C bond
length involving the methyl group in the γ-position of the β-
diketone was consistent with usual CÀ C bond length. For
compound 6, the CÀ Br bond lengths [av. 1.910 Å] was in the
usual range.[11–13] Similar to [Ti(OiPr)2(Me-acac)2] (3), [Ti(OiPr)2(Br-
acac)2] (6) discrete mononuclear molecules were rather loosely
packed, and no significant intermolecular interaction was
found.
°
oxygen atoms lied in the range 80–102 , indicating distorted
Thermal behavior
octahedral geometries around metal center. The average TiÀ O
bond distances of iso-propoxide ligands [av. 1.775 (3) Å] was
shorter than those of β-diketonate ones [av. 2.044 (3) Å].
(Table 1) Overall, these bond lengths and bond angles
All compounds were tested in TGA under argon atmosphere
between 20 and 600 C with a temperature ramp of 5 Cmin .
The curves are summarized in Figure 2 and compared to
À 1
°
°
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
2
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
compound A. All these compounds were found to be non-
volatile under reduced pressure, so the weight losses observed
can be considered as resulting only from thermal degradation
and therefore dependent from the stability of the compounds.
In general, all the new compounds showed an onset of weight
loss at lower temperatures than the reference compound
suggesting lower thermal stability induced by the functionaliza-
tion of the internal position of the acac ligand. Based on the
temperature of onset of weight loss, the following decreasing
°
°
stability ranking was established: H (A) 172 C>NO2 (1) 162 C>
°
°
°
°
Cl (5) 159 C>Me (3) 151 C>Br (6) 141 C>Et (4) 108 C>OAc
°
(2) 92 C (Table 3).
Another relevant parameter to check was the difference
between the experimental weight loss and the theoretical one
to obtain TiO2. The results are summarized in Table 2 and
showed that the smallest differences were observed for the
alkyl groups. In particular, the presence of the ethyl group in
compound [Ti(OiPr)2(Et-acac)2] (4) was accompanied by a
degradation rate very close to that of TiO2, making it possible to
envisage a more complete degradation than for the other
precursors. In comparison, the compound [Ti(OiPr)2(OAc-acac)2]
(2), which started to degrade at the lowest temperature, yielded
one of the highest residues far from the theoretical one to
obtain TiO2.
Theoretical calculations
In order to explain the evolution of the thermal stability of the
precursors observed by the TGA, we sought to correlate the
effect of the different substituents of the acac ligand on
thermodynamic stability by theoretical calculations. The
Brønsted acid-base reactions between Ti(OR)4 and 2 equivalents
of γ-modified β-diketone can be summarized as below.
In this case, Brønsted acidities of γ-modified β-diketone
ligands are key thermodynamic parameters to obtain stable
final [Ti(OiPr)2(R-acac)2] complexes. Considering the pKa of 16.5
isopropanol in water, the more acidic the R-acacH ligand, the
more thermodynamically favorable the reaction will be. β-
Diketones can exhibit two forms (β-diketone and keto-enol) and
only this latter form can undergo a Brønsted acid-base reaction.
It was therefore crucial to evaluate the proportion between
these two forms before determining their Brønsted acidity. The
Figure 1. ORTEP drawings of (A) [Ti(OiPr)2(NO2-acac)2] (1), (B) [Ti(OiPr)2(Me-
acac)2] (3) and (C) [Ti(OiPr)2(Br-acac)2] (6) with 50% probability ellipsoids
(Hydrogen atoms are omitted for clarity).
Table 2. TGA data vs acac functionalization.
R group
H (A)
NO2 (1)
Cl (5)
Me (3)
Br (6)
Et (4)
OAc (2)
°
Onset temperature weight loss [ C]
Residues [%]
Theoretical loss for TiO2 [%]
Δ (Exp-Theo)
172
30.4
22.0
8.4
162
159
151
141
108
92
34.7
17.5
17.2
37.7
18.0
19.7
30.6
20.4
10.2
37.6
15.3
22.3
20.6
19.0
1.6%
37.7
16.5
21.2
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
3
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
°
Figure 2. TGA Patterns (20–600 C) of derivatives A and 1–6.
~
°
ratio between the diketone and keto-enol forms was estimated
by DFT calculations to better understand β-diketones stability
and to anticipate their reactivity. DFT calculations by B3LYP
functional were performed and have consisted first of optimiz-
ing the structure of the β-diketones and keto-enol forms and
compared the Gr values to deduce which form was predom-
inant. Table 3 summarizes results obtained for the calculations
on the acacH ligand and γ-modified analogues and classified
°
according to decreasing ΔGr values.
For the ligand OAc-acacH, an optimized structure compris-
ing a hydrogen bond (oxygen-hydrogen distance=2.28 Å<
2.5 Å) between the hydrogen in the γ-position and the oxygen
of the ketone of the -OAc group (Figure 3) has been calculated
and has demonstrated a strong stabilization of the β-diketone
form to the detriment of the keto-enol one.
~
°
then of calculating the Gibbs free enthalpy Gr between both
forms (Figure S1). For this tautomeric equilibrium, we did not
try to find an intermediate state or activation energy, we simply
Except for compounds A and 6, it should be noted that the
order of the Gibbs free enthalpy of the ligands follows the same
order as that of the onset degradation temperatures and
therefore thermal stability. The more the equilibrium is shifted
towards the keto-enol form the more thermally stable the
compound [Ti(OiPr)2(R-acac)2] is. This observation suggests that
DFT calculations on β-diketone ligands could predict the
thermal stability of such heteroleptic complexes. Indeed, the
ligands in the major diketone form (OAc, Et) being less acidic,
the ligand-metal bonds are weaker, resulting in less stable
molecular complexes. On the other hand, although the two
Table 3. Ranking of γ-modified β-diketones ligands and keto-enol/dike-
°
tone ratio versus ΔGr .
°
ΔG [kcal/mol]
Ratio [%]
Ketoenol
Diketone
NO2-acacH
H-acacH
Cl-acacH
Me-acacH
Br-acacH
Et-acacH
4.87
2.81
2.71
1.65
1.25
0.46
3.30
À 2.94
100
99.1
99.0
94.1
89.1
68.5
99.6
0.7
–
0.9
1.0
5.9
10.9
31.5
0.4
99.3
Normal
+H bond
OAc-acacH
Figure 3. Optimized structures of OAc-acacH ligand with hydrogen bond (Grey ball=carbon, red ball=oxygen, white ball=hydrogen).
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
4
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
compounds, [Ti(OiPr)2(OAc-acac)2] (2) and [Ti(OiPr)2(Et-acac)2]
In addition, the fact that the [Ti(OiPr)2(R-acac)2] [R=OAc (2),
Et (4)] complexes behaved singularly by yielding the lowest
onset degradation temperatures would indicate that these
complexes probably did not decompose by overall removal of
ligands, but probably by partial decomposition of ligands.
Calculations of the binding energy of different fragments in
relation to the rest of the complex showed that the departures
of the OAc and Me groups were favoured (Table 5). However,
(4), had close onset degradation temperatures, their weight
losses were very different from the theoretical percentage to
have TiO2, far for 2 and very close for 4 (Table 1). We were
therefore interested in the influence of the fragmentation of
these heteroleptic compounds in order to identify a key
parameter allowing near-complete degradation.
To investigate the decomposition mechanism involved in
the TGA experiment, DFT calculations have been performed on
several [Ti(OiPr)2(R-acac)2] complexes [R=NO2 (1), OAc (2), Et
(4)]. When the complex dissociated, the expected fragments
that were eliminated may be isopropoxyl (OiPr) and γ-modified
acetylacetonate (R-acac). If a decomposition into neutral frag-
ments was considered, the binding energy of Ti with the four
°
producing an acac-CH2 radical more reactive/less stable than
°
acac one, could be at the origin of the almost complete
degradation of [Ti(OiPr)2(Et-acac)2] (lowest residue content)
compared to others complexes of this study.
fragments was À 21.66 eV (499 kcal/mol), whereas if ionic frag- Conclusions
ments (Ti4+ and negatively charged ligand fragments) were
considered, the binding energy of Ti4+ with the four fragments
was À 102.0 eV. This clearly indicated that due to the electro-
static contributions, decomposition of the complex did not lead
to charged fragments. The calculation energies are shown in
Table 4 for [Ti(OiPr)2(R-acac)2] complexes [R=NO2 (1), OAc (2),
Et (4)].
The results summarized in Table 4 demonstrated that the
differences were minimal, so that the two fragments OiPr and
R-acac could be expected to be removed with the same ease.
Moreover, for the three complexes, the fragmentation energies
were very close, which did not explain the differences in
thermal stability observed in the TGA, or even contradicted
them, as it was lower for the complex 1 based on the NO2-acac
ligand. Complete fragmentation scheme and DFT calculations
of [Ti(OiPr)2(R-acac)2] complexes [R=NO2 (1), OAc (2)] are
shown in Figure S1 and Table S2. It should be mentioned that
such calculations assumed that the experiments involved rapid
degradation of the complex, i.e. not without significant time
between the removal of the first and second ligand. Con-
Six new titanium alkoxide complexes di-substituted with γ-
modified β-diketonates were synthesized and thoroughly
characterized. None of them showed volatility, and TGA
demonstrated that thermal stability was modulated by the
nature of the γ-position group. These results were consistent
with DFT calculations, which demonstrated that improved
thermal stability of the Ti(OR)2(R-acac)2 complexes could be
correlated with the greater stability of the γ-modified ketoeno-
late isomer compared to that of β-diketonates. DFT calculations
also showed that thermal decomposition should occur by
sequential fragmentation. An alkyl group at the γ-position such
as ethyl was found to enhance the reactivity of this type of
molecule, giving the lowest residue content close to TiO2. It
could be envisaged that this reactivity (and hence degradation)
could be improved by the introduction of a group that could
allow a primary degradation step such as CÀ C cleavage to give
a reactive radical.
sequently, all calculations were made against a geometry of the Experimental Section
initial neutral complex. If the experiment assumed that the
degradation of the complex was a slow process, it would be
necessary to re-optimize the eliminated fragments, a process
that would lead to modified energies, but not to significantly
modified trends.
Synthesis
All manipulations using metal alkoxides were carried out under
inert gas using a glove box (N2) or standard Schlenk techniques
(Ar). [Ti(OiPr)4]m was purchased from STREM Chemicals and
distillated under vacuum prior being used. N-bromosuccinimide
(C4H4NO2Br, NBS) and acetylacetone (2,4-pentadione, acacH) were
purchased from Acros organics. 3-ethyl-2,4-pentanedione (Et-acacH)
and 3-chloroacetylacetone (Cl-acacH) were supplied from Merck
and used as received. γ-modified acacH ligands (NO2-acacH,
CH3COO-acacH, Me-acacH) were synthesized through modified
procedures reported in the literature and thoroughly characterized
by multinuclear NMR and FTIR (See Supplementary Materials).
Anhydrous n-hexane was obtained using a solvent drier, SPS
(solvent purification system) from MBRAUN and stored under argon
over 4 Å molecular sieves. Dichloromethane (CH2Cl2) was dried over
MgSO4, filtered and stored under argon over 4 Å molecular sieves.
Isopropanol was refluxed and distillated on [Al(OiPr)3]4 and then
stored under argon over 4 Å molecular sieves. Deuterated chloro-
form was stored in a Rotaflow flask over 4 Å molecular sieves.
Table 4. Energies (eV) for the first fragmentations of [Ti(OiPr)2(R-acac)2]
complexes [R=NO2 (1), OAc (2), Et (4)].
Fragmentation reaction
R=NO2
[eV]
R=OAc
[eV]
R=Et
[eV]
Loss of R-acac
Loss of OiPr
4.96
5.23
5.06
5.14
5.46
5.23
Table 5. Energies (eV) for the partial fragmentations of [Ti(OiPr)2(R-acac)2]
complexes [R=OAc (3), Et (5)].
R=OAc
R=Et
Fragments
Energy (eV)
Ac
4.23
OAc
1.20
Me
5.23
Et
5.90
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
5
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
Preparation of [Ti(OiPr)2(acac)2] (A)
The synthesis was done by
Preparation of [Ti(OiPr)2(Et-acac)2] (4)
a
modified protocol based on
The same protocol as described above with 1.06 g of Et-acacH
(8.3 mmol) and 1.18 g of [Ti(OiPr)4]m (4.2 mmol) gave, after removal
of all volatiles, a viscous orange-brown liquid. Purification by
vacuum distillation was attempted but product entirely decom-
reference.[18] 3.9 g of [Ti(OiPr)4]m (13.7 mmol) were diluted in n-
hexane (65 mL) then 2.75 g of acacH (27.5 mmol) was added slowly.
The mixture was let under stirring overnight. All volatiles were
removed under vacuum, the brown-orange liquid obtained was
1
3
posed. Yield: 93% (m=1.65 g). H NMR (CDCl3, δ ppm) 0.99 [t, J=
3
3
°
then purified by vacuum distillation (1 mbar, oil bath at 150 C) and
6.5 Hz, 6H, CH2CH3], 1.11 [d, J=6.3 Hz, 6H, OCH(CH3)2], 1.19 [d, J=
6.3 Hz, 6H, OCH(CH3)2], 1.96 [s, 6H, COCH3], 2.04 [s, 6H, COCH3],
2.22–2.28 [q, 4H, CH2CH3], 4.69 [m, 2H, OCH(CH3)2]. 13C NMR (CDCl3,
δ ppm) 15.2 [CH2CH3], 22.4 [OCH(CH3)2], 25.3, 26.6 [COCH3], 76.3
[OCH(CH3)2], 114.4 [COC(Me)CO], 185.4, 190.6 [CH3CO]. FT-IR (Nujol,
cmÀ 1) 2959, 2922, 2855 [ν CÀ H]; 1605, 1582, 1527 [ν C=O], 1380,
1359; 1325, 1122, 1011 [ν CÀ O], 993, 928; 664, 623 [ν TiÀ O].
1
gave a bright orange liquid (A). Yield: 80% (m=4.00 g) . H NMR
3
(CDCl3, δ ppm) 1.12–1.18 [d, J=6.4 Hz, 12H, OCH(CH3)2], 1.88, 1.99
[s, 12H, COCH3], 4.74 [m, 2H, OCH(CH3)2], 5.47 [s, 2H, COCHCO]. 13C
NMR (CDCl3, δ ppm) 25.3 [OCH(CH3)2], 26.1, 27.1 [COCH3], 78.3 [OCH
(CH3)2], 102.8 [COCHCO], 187.3, 191.4 [CH3COCH]. FT-IR (Nujol,
cmÀ 1) 2964, 2925, 2859 [ν CÀ H]; 1608, 1587, 1524 [ν C=O], 1383,
1359; 1326, 1125, 1012 [ν CÀ O], 993, 929; 665, 619 [ν TiÀ O].
Preparation of [Ti(OiPr)2(Cl-acac)2] (5)
Preparation of [Ti(OiPr)2(NO2-acac)2] (1)
The same protocol as described above with 1.79 g of Cl-acacH
(13.3 mmol) and 1.78 g of [Ti(OiPr)4]m (6.3 mmol) gave, after removal
of all volatiles, a yellow powder. MP (ATG-DSC)=65.5 C. Yield
The same protocol as described above with 1.51 g (10.4 mmol) of
NO2-acacacH and 1.48 g of [Ti(OiPr)4]m (5.2 mmol) gave, after
removal of all volatiles, an orange-yellow viscous liquid. Purification
by vacuum distillation was attempted, but the product completely
decomposed. The product crystallized in an isopropanol/n-hexane
°
>99% (m=2.71 g). 1H NMR (CDCl3, δ ppm) 1.17 [d, 3J=6.3 Hz,
12H, OCH(CH3)2], 2.18 [s, 6H, COCH3], 2.27 [s, 6H, COCH3], 4.74 [m,
2H, OCH(CH3)2]. 13C NMR (CDCl3, δ ppm) 25.3 [OCH(CH3)2], 26.1
[COCH3], 79.9 [OCH(CH3)2], 110.5 [COC(Cl)CO], 186.3, 188.6 [CH3CO].
FT-IR (Nujol, cmÀ 1) 2955, 2921, 2851 [ν CÀ H]; 1603, 1579, 1524 [ν
C=O], 1449, 1362; 1331, 1162, 1126, 1003 [ν CÀ O], 851 [ν CÀ Cl]; 620,
559 [ν TiÀ O].
°
mixture at about 2 C to give light yellow needle crystals. MP (ATG-
DSC)=57 C. Yield >99% (m=2.35 g). 1H NMR (CDCl3, δ ppm)
°
3
1.17–1.26 [d, J=6.4 Hz, 12H, OCH(CH3)2], 2.23 [s, 6H, COCH3], 2.36
[s, 6H, COCH3], 4.76 [m, 2H, OCH(CH3)2]. 13C NMR (CDCl3, δ ppm)
24.9, 25.1 [OCH(CH3)2], 25.6, 25.9, 26.6, 29.9 [COCH3], 82.2, 83.9 [OCH
(CH3)2], 139.4 [COC(NO2)CO], 187.3, 189.6 [CH3CO]. FT-IR (Nujol,
cmÀ 1) 2967, 2927, 2861 [ν CÀ H]; 1608, 1588, 1528 [ν C=O], 1415,
1346 [ν C-NO2]; 1287, 1161, 1117[ν CÀ O]; 926, 824; 684, 667, 624
[νTiÀ O].
Preparation of [Ti(OiPr)2(Br-acac)2] (6)
The synthesis was done following a new protocol. 0.67 g of
Ti(OiPr)2(acac)2 (1.8 mmol) was diluted in dichloromethane (56 mL)
°
cooled at around 0 C. Then 0.72 g of dry N-bromosuccinimide
(4.1 mmol) in 15 mL of dichloromethane was added very slowly
Preparation of [Ti(OiPr)2(OAc-acac)2] (2)
°
(10 min addition) and at 0 C. The mixture was let under stirring
The same protocol as described above with 1.02 g of OAc-acacH
(6.5 mmol) and 0.89 g of [Ti(OiPr)4]m (3.1 mmol) gave, after removal
of all volatiles, an orange-brown viscous oil. Purification by vacuum
distillation was attempted but product entirely decomposed. Yield:
overnight. After removal of all volatiles, an orange brown powder
was obtained. Attempts to remove succinimide by washing with n-
hexane and filtering failed. Product crystallized in isopropanol/n-
°
hexane mixture at around 2 C to give square light-yellow crystals.
1
3
°
99% (m=1.48 g). H NMR (CDCl3, δ ppm) 1.18 [d, J=6.5 Hz, 12H,
OCH(CH3)2], 1.90 [s, 6H, COCH3], 1.97 [s, 6H, COCH3], 2.20 [s, 3H,
OOCCH3], 4.77 [m, 2H, OCH(CH3)2]. 13C NMR (CDCl3, δ ppm) 24.7,
25.3 [OCH(CH3)2], 27.7 [COCH3], 30.1 [OOCCH3], 82.5, 83.6 [OCH
(CH3)2], 129.5 [COC(OAc)CO], 187.3, 189.6 [CH3CO], 199.5 [OOCCH3].
FT-IR (Nujol, cmÀ 1) 2962, 2922, 2857 [ν CÀ H]; 1756, 1720 [ν COO],
1599, 1585, 1525 [ν C=O]; 1415, 1346, 1287, 1161, 1117, 1001 [ν
CÀ O]; 926, 824; 684, 667, 624 [ν TiÀ O].
MP (ATG-DSC)=75 C. Yield: 82% (m=0.77 g). 1H NMR (CDCl3, δ
3
ppm) 1.17 [d, J=6.3 Hz, 12H, OCH(CH3)2], 2.26 [s, 6H, COCH3], 2.35
[s, 6H, COCH3], 4.74 [m, 2H, OCH(CH3)2]. 13C NMR (CDCl3, δ ppm)
26.2 [OCH(CH3)2], 26.9 [COCH3], 81.2 [OCH(CH3)2], 112.2 [COC(Cl)CO],
187.3, 189.1 [CH3CO]. FT-IR (Nujol, cmÀ 1) 2954, 2920, 2856 [ν CÀ H];
1601, 1575, 1524 [ν C=O], 1448, 1359; 1330, 1159, 1124, 1010 [ν
CÀ O], 701 [ν CÀ Br]; 622, 562 [ν TiÀ O].
Characterizations
Preparation of [Ti(OiPr)2(Me-acac)2] (3)
NMR spectra were recorded on Bruker AVANCE III 400 MHz with a
BBFO probe (Z gradient) using Topspin software. All chemical shifts
were measured relatively to the deuterated solvents in case of 13C
NMR (in CDCl3: δ=77.23 ppm triplet), or to the residual protic
The same protocol as described above with 1.05g of Me-acacH
(9.2mmol) and 1.31g of [Ti(OiPr)4]m (4.6mmol) gave, after removal
of all volatiles, an orange powder. Product crystallized in n-hexane
1
°
at around 2 C to give square light-yellow crystals. MP (ATG-DSC)=
solvent for H NMR (in CDCl3: δ=7.24 ppm singlet). FT-IR spectra
1
°
71 C. Yield: 99% (m=1.80g). H NMR (CDCl3, δ ppm) 1.14–1.15 [d,
were recorded from 4000 to 400 cmÀ 1 at room temperature using
Bruker Vector 22 and Opus software. Operation chamber is
maintained under N2 with constant gas flow. Samples were
prepared under argon flow between KBr windows with Nujol (dried
over 3 Å molecular sieves).
3J=6.4 Hz, 12H, OCH(CH3)2], 1.86 [s, 6H, CCH3], 1.99 [s, 6H, COCH3],
2.06 [s, 6H, COCH3], 4.73 [m, 2H, OCH(CH3)2]. 13C NMR (CDCl3, δ ppm)
24.9, 25.1 [OCH(CH3)2], 25.6, 25.9, 26.6, 29.9 [COCH3], 82.2, 83.9 [OCH
(CH3)2], 139.4 [COC(NO2)CO], 187.3, 189.6 [CH3CO]. FT-IR (Nujol,
cmÀ 1) 2959, 2922, 2855 [ν CÀ H]; 1605, 1582, 1527 [ν C=O], 1380,
1359; 1325, 1122, 1011 [ν CÀ O], 993, 928; 664, 623 [ν TiÀ O].
TGA data were collected on a TGA/DSC 1 thermal analysis MX1
from Mettler Toledo, Stare system, gas controller GC200. In order to
prepare samples, the following procedure was used: 100μL
aluminum crucibles (top and bottom) are weighed together on
high precision balance (10À 3 mg) then they are brought into the
glovebox. Into the glovebox (under N2), 5–10 mg of product were
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
6
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
placed into the crucible bottom, the cap is added and then the
performed with TZP basis set[28] and PBE exchange-correlation
functional.[31] DFT calculations for fragmentation of the Ti(OR)2(R-
acac)2 derivatives have then been performed according to the
following steps:
entire crucible is sealed using a special baler from Mettler Toledo
and finally samples are brought out of the glovebox. Before
analysis, samples are weighed once again on high precision
balance. Sample were pierced just before analysis; the method uses
as carrier gas argon with a flow of 30cc per minute with a program
1. Geometry optimization of the whole complex through Kohn-
Sham calculations with PBE exchange-correlation functional and
TZP basis set, with ADF software.[28]
2. Calculation of the electronic structure of the (separated) four
ligands in their geometry within the complex, and the free Ti
atom (unrestricted calculation)
3. Calculation of the electronic structure of the whole complex
from its fragments, namely the four ligands and the Ti atom
4. Calculation of the electronic structure of fragments consisting of
the complex stripped from one ligand (either R-acac or OiPr) (in
their geometry within the complex) (unrestricted calculation)
5. Calculations of the electronic structure of the entire complex
°
°
beginning with 4 minutes at 25 C and then a heating rate of 5 C
°
per minute until reaching 600 C.
Suitable single crystals of derivatives [Ti(OiPr)2(NO2-acac)2] (1),
[Ti(OiPr)2(CH3-acac)2] (3) and [Ti(OiPr)2(Br-acac)2] (6) (Table 6) were
mounted on a Gemini kappa-geometry diffractometer (Agilent
Technologies UK Ltd) equipped with an Atlas CCD detector and
using Mo radiation (λ=0.71073 Å). Intensities were collected at
150 K by means of the CrysalisPro software. Reflection indexing,
unit-cell parameters refinement, Lorentz-polarization correction,
peak integration, and background determination were carried out
with the CrysalisPro software.[19] An analytical absorption correction
was applied using the modeled faces of the crystal.[20] The resulting
set of hkl was used for structure solution and refinement. The
structures were solved by direct methods with SIR97[21] and the
least-square refinement on F2 was achieved with the CRYSTALS
software.[22] All non-hydrogen atoms were refined anisotropically.
The hydrogen atoms were all located in a difference map, but those
attached to carbon atoms were repositioned geometrically. The H
atoms were initially refined with soft restraints on the bond lengths
and angles to regularize their geometry (CÀ H in the range 0.93–
0.98, NÀ H in the range 0.86–0.89 and OÀ H=0.82 (Å) and Uiso(H) (in
the range 1.2–.5 times Ueq of the parent atom), after which the
positions were refined with riding constraints.
from
a ligand fragment and the corresponding stripped
complex. These two calculations provide the binding energies
of the fragments that are first eliminated in the experiment. The
Gibbs free energies are not calculated but can be estimated as
slightly lower due to the increase in entropy during dissociation.
6. The process is further pursued by a calculation of the electronic
structure of fragments consisting of the complex undergoing a
second fragmentation from one ligand (either R-acac or OiPr) (in
their geometry within the complex).
7. Calculations of the electronic structure of the complex ampu-
tated with a ligand fragment and the corresponding complex
amputated twice. These four calculations provide the binding
energies of the second fragments that are eliminated during the
experiment. These calculations are reported in the Supplemen-
tary Information file.
DFT calculations
Supporting Information (see footnote on the first page of this
article): Synthesis and 1H, 13C NMR and FTIR characterizations of the
synthesized ligand.
DFT calculations for diketone/keto-enol ratio determinations have
been performed using Gaussian09 package,[23] the hybrid B3LYP
functional[24,25] without symmetry restrictions and chloroform as
solvent.[26] An energy decomposition analysis was performed, a
procedure which provided the interaction energy of a compound
with respect to its constituting fragments. This technique, which
was implemented in ADF code,[27,28] has been developed by
Baerends, Ziegler and Frenking.[29,30] Calculations have been
Deposition Numbers 2018641 (for 1), 2018642 (for 3), and 2018643
(for 6) contain the supplementary crystallographic data for this
paper. These data are provided free of charge by the joint
Cambridge Crystallographic Data Centre and Fachinformationszen-
trum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/
structures.
Table 6. Crystallographic data of derivatives 1, 3 and 6.
[Ti(OiPr)2(R-acac)2]
R=NO2 (1)
R=CH3 (3)
R=Br (6)
Empirical formula
Formula weight (gmolÀ 1
Crystal system
Space group
a (Å)
C16H26N2O10Ti
454.29
Triclinic
�
P
7.3757 (10)
11.1868 (16)
13.9854 (18)
94.076 (11)
100.886 (11)
108.497 (13)
1064.0 (3)
2
C18H30O6Ti
390.33
Triclinic
�
P
9.3260 (12)
10.3893 (17)
11.9249 (14)
81.287 (12)
85.906 (10)
70.706 (13)
1077.7 (3)
2
C16H26Br2O6Ti
522.09
Monoclinic
P21/n
11.751 (2)
14.156 (2)
13.174 (2)
)
b (Å)
c (Å)
°
α ( )
°
β ( )
105.671 (19)
°
γ ( )
V (Å3)
Z
2110.0 (6)
4
4.22
μ (mmÀ 1
)
0.46
0.42
Temperature (K)
Measured reflections
100
13917
200
18774
150
17612
Independent reflections (Rint
Data/restrains/parameters
Goodness of fit
)
5153 (0.056)
5138/12/272
0.97
0.061
0.146
5398 (0.049)
5389/0/226
0.98
0.059
0.127
5227 (0.052)
5216/0/227
0.98
0.048
0.094
R[F2 >2σ (F2)]
wR(F2)
Residual electron
À 0.98 to 0.87
À 0.72 to 0.70
À 1.13 to 1.16
density (e.ÅÀ 3
)
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
7
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

Full Papers
doi.org/10.1002/ejic.202100209
[15] T. Ziegler, A. Rauk, Inorg. Chem. 1979, 18, 1558.
Declaration of Competing Interest
[16] F. M. Bickelhaupt, E. J. Baerends, Rev. Comput. Chem. 2000, 15, 1.
[17] G. A. Seisenbaeva, S. Gohil, V. G. Kessler, M. Andrieux, C. Legros, P. Ribot,
M. Brunet, E. Scheid, Appl. Surf. Sci. 2011 257(6), 2281.
[18] J. D. St. Clair, N. A. Carlson, Patent US6562990B1 2003.
[19] T. S. Lewkebandara, P. J. McKarns, B. S. Haggerty, G. P. A. Yap, A. L.
Rheingold, C. H. Winter, Polyhedron 1998, 17(1), 1.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to
influence the work reported in this paper.
[20] R. C. Clark, J. S. Reid, Acta Crystallogr. Sect. A 1995, 51(6), 887.
[21] A. Altomare, C. Giacovazzo, A. Guagliardi, A. G. G. Moliterni, R. Rizzi, J.
Appl. Crystallogr. 1999, 32(5), 963.
[22] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J. Watkin, J.
Appl. Crystallogr. 2003, 36(6), 1487.
Acknowledgements
[23] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H.
Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G.
Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J.
Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven,
J. A. Montgomery Jr., J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E.
Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K.
Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi,
N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C.
Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin,
R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G.
Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D.
Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, D. J. Fox,
Gaussian09, Gaussian, Inc., Wallingford CT, 2009.
The authors gratefully acknowledge the GENCI/CINES (Project
cpt2130) for HPC resources/computer time and ANRT (CIFRE
°
contract n 2014/0840) for financial support. D.B. thanks Dr
Frédéric Guégan, Dr Valérian Forquet and Dr Walid Lamine for
supervising B3LYP calculations.
Conflict of Interest
The authors declare no conflict of interest.
[24] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785.
[25] A. D. Becke, J. Chem. Phys. 1993, 98, 5648.
[26] C. Reichardt, T. Welton, Solvents and Solvent Effects in Organic Chemistry
(Wiley-VCH Verlag GmbH & Co. KGaA, ( 2010) 65.
Keywords: Density functional calculations · β-Diketonates ·
Ligand design · Thermal behavior · Titanium
[27] G. te Velde, F. M. Bickelhaupt, E. J. Baerends, C. Fonseca Guerra, S. J. A.
van Gisbergen, J. G. Snijders, T. Ziegler, J. Comput. Chem. 2001, 22, 931.
[28] E. J. Baerends, T. Ziegler, A. J. Atkins, J. Autschbach, D. Bashford, A.
Bérces, F. M. Bickelhaupt, C. Bo, P. M. Boerrigter, L. Cavallo, D. P. Chong,
D. V. Chulhai, L. Deng, R. M. Dickson, J. M. Dieterich, D. E. Ellis, M.
van Faassen, L. Fan, T. H. Fischer, C. Fonseca Guerra, M. Franchini, A.
Ghysels, A. Giammona, S. J. A. van Gisbergen, A. W. Götz, J. A. Groene-
veld, O. V. Gritsenko, M. Grüning, S. Gusarov, F. E. Harris, P.
van den Hoek, C. R. Jacob, H. Jacobsen, L. Jensen, J. W. Kaminski, G.
van Kessel, F. Kootstra, A. Kovalenko, M. V. Krykunov, E. van Lenthe,
D. A. McCormack, A. Michalak, M. Mitoraj, S. M. Morton, J. Neugebauer,
V. P. Nicu, L. Noodleman, V. P. Osinga, S. Patchkovskii, M. Pavanello,
C. A. Peeples, P. H. T. Philipsen, D. Post, C. C. Pye, W. Ravenek, J. I.
Rodriguez, P. Ros, R. Rüger, P. R. T. Schipper, H. van Schoot, G.
Schreckenbach, J. S. Seldenthuis, M. Seth, J. G. Snijders, M. Solà, M.
Swart, D. Swerhone, G. te Velde, P. Vernooijs, L. Versluis, L. Visscher, O.
Visser, F. Wang, T. A.Wesolowski, E. M. van Wezenbeek, G. Wiesenekker,
S. K. Wolff, T. K. Woo, A. L. Yakovlev, ADF2017.01, SCM, Theoretical
Chemistry, Vrije Universiteit, Amsterdam, The Netherlands, http://
www.scm.com.
[1] a) M. E. Dufond, M. W. Diouf, C. Badie, C. Laffon, P. Parent, D. Ferry, D.
Grosso, J. C. S. Kools, S. D. Elliott, L. Santinacci, Chem. Mater. 2020, 32,
1393; b) J. Xu, H. Nagasawa, M. Kanezashi, T. Tsuru, ACS Appl. Mater.
Interfaces 2018, 10, 42657.
[2] A. Möllmann, D. Gedamu, P. Vivo, R. Frohnhoven, D. Stadler, T. Fischer, I.
Ka, M. Steinhorst, R. Nechache, F. Rosei, S. G. Cloutier, T. Kirchartz, S.
Mathur, Adv. Eng. Mater. 2019, 21, 1801196.
[3] A. E. Turgambaeva, V. V. Krisyuk, S. V. Sysoev, I. K. Igumenov, Chem. Vap.
Dep. 2001, 7(3), 121.
[4] R. J. Errington, J. Ridland, W. Clegg, R. A. Coxall, J. M. Sherwood,
Polyhedron 1998, 17 (5–6), 659.
[5] a) E. Wagner, C. S. Sandu, S. Harada, C. Pellodi, M. Jobin, P. Muralt, G.
Benvenuti, ACS Comb. Sci. 2016, 18(3), 154; b) C.S Sandu, E. Wagner, S.
Harada, G. Benvenuti, W. Maudez, M. Jobin, C. Pellodi, P. Muralt, Thin
Solid Films 2016, 615, 265.
[6] R. G. Gordon, Patent WO 98/46617 1998.
[7] R. G. Gordon, S. Barry, R. N. R. Broomhall-Dillard, D. J. Teff, Adv. Mater.
Opt. Electron. 2000, 10, 201.
[29] T. Ziegler, A. Rauk, Theor. Chim. Acta 1977, 46, 1.
[30] M. von Hopffgarten, G. Frenking, Wiley Interdiscip. Rev.: Comput. Mol. Sci.
2012, 2, 43.
[8] W. R. McNamara, R. C. Snoeberger, G. Li, J. M. Schleicher, C. W. Cady, M.
Poyatos, C. A. Schmuttenmaer, R. H. Crabtree, G. W. Brudvig, V. S. Batista,
J. Am. Chem. Soc. 2008, 130, 14329.
[9] B. Yan, B. Zhou, J. Photochem. Photobiol. A 2008, 195, 314.
[10] K. J. Eisentraut, R. E. Sievers, J. Inorg. Nucl. Chem. 1967, 29(8), 1931.
[11] V. G. Isakova, I. A. Baidina, N. B. Morozova, I. K. Igumenov, V. B. Rybakov,
J. Struct. Chem. 1999, 40(2), 276.
[31] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77, 3865.
[12] V. G. Isakova, I. A. Baidina, N. B. Morozova, I. K. Igumenov, Polyhedron
2000 19 (9), 1097.
[13] T. S. Knowles, B. J. Howlin, J. R. Jones, D. C. Povey, C. A. Amodio,
Polyhedron 1993, 12(24), 2921.
[14] D. H. Johnston, J. T. Brangham, C. D. Rapp, Acta Crystallogr. Sect. E 2012,
68(3), m312.
Manuscript received: March 15, 2021
Revised manuscript received: April 6, 2021
Accepted manuscript online: April 8, 2021
Eur. J. Inorg. Chem. 2021, 1–9
www.eurjic.org
8
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

FULL PAPERS
D. Bijou, T. Cornier, Dr. S. Mishra,
Dr. L. Merzoud, Prof. H. Chermette,
Dr. E. Jeanneau, Dr. W. Maudez,
Dr. G. Benvenuti, Prof. S. Daniele*
1 – 9
Synthesis and Thermal Behavior
of Heteroleptic γ-Substituted Ace-
tylacetonate-Alkoxides of
Titanium
The thermal behavior (stability and
decomposition) of a series of hetero-
leptic titanium derivatives [Ti(OiPr)2(R-
acac)2] with acetylacetonate ligands
modified in the internal (γ- or 3-)
position by different substituents
(R=OAc, NO2, Me, Et, Cl, Br) was
studied by thermogravimetric analysis
and DFT calculations.