Thermal decomposition of cyclotriborazane

Article abstract of DOI:10.1016/j.tca.2007.03.006

Cyclotriborazane (CTB), B₃N₃H₁₂, is a crystalline white solid, which decomposes above 400 K to hydrogen and a few other products, depending on the reaction conditions. In this work we present investigations of the thermal decomposition of both the neat compound and CTB dissolved in diglyme and tetraglyme. Several thermophysical and analytical methods, such as differential scanning calorimetry (DSC), thermogravimetry (TG), mass spectroscopy (QMS), and ¹¹B nuclear magnetic resonance spectroscopy (NMR) have been used for this investigation. The decomposition of the neat substance releases 3.1 mol H₂/mol CTB and leads to a polymeric products and borazine. In open vessels, sublimation as a competing process also occurs. The enthalpy of the decomposition process (Δ_RH_s) has been determined as Δ_RH_s = -34.0 ± 2.9 kJ/mol. In contrast to the thermal decomposition of the pure substance, the decomposition in polyethers, such as diglyme and tetraglyme, leads above 370 K to borazine and small amounts of soluble oligomeric borazine species. Also BH₃ group containing species are occurring as intermediates. In these systems no precipitation was detected. DSC measurements show for the decomposition in solution several strong exothermic effects. The overall decomposition enthalpy in diglyme is given by Δ_RH_d = -32.0 ± 2.8 kJ/mol and in tetraglyme by Δ_RH_t = -48.0 ± 4.7 kJ/mol. The enthalpy of solution of cyclotriborazane was determined in diglyme and in tetraglyme with the values Δ_DH_d = -2.1 ± 0.2 kJ/mol and Δ_DH_t = -4.6 ± 0.5 kJ/mol, respectively.

Full text of DOI:10.1016/j.tca.2007.03.006

Thermochimica Acta 457 (2007) 103–108
Thermal decomposition of cyclotriborazane
∗
R. Schellenberg , J. Kriehme, G. Wolf
Technische Universit a¨ t Bergakademie Freiberg, Institute of Physical Chemistry, Leipziger Str. 29, 09599 Freiberg, Germany
Received 19 December 2006; received in revised form 7 March 2007; accepted 7 March 2007
Available online 13 March 2007
Abstract
Cyclotriborazane (CTB), B
3
3
N H12, is a crystalline white solid, which decomposes above 400 K to hydrogen and a few other products, depending
on the reaction conditions. In this work we present investigations of the thermal decomposition of both the neat compound and CTB dissolved in
diglyme and tetraglyme. Several thermophysical and analytical methods, such as differential scanning calorimetry (DSC), thermogravimetry (TG),
1
1
mass spectroscopy (QMS), and B nuclear magnetic resonance spectroscopy (NMR) have been used for this investigation. The decomposition
of the neat substance releases 3.1 mol H /mol CTB and leads to a polymeric products and borazine. In open vessels, sublimation as a competing
process also occurs. The enthalpy of the decomposition process (ꢀ ) has been determined as ꢀ
= −34.0 ± 2.9 kJ/mol.
In contrast to the thermal decomposition of the pure substance, the decomposition in polyethers, such as diglyme and tetraglyme, leads above
70 K to borazine and small amounts of soluble oligomeric borazine species. Also BH group containing species are occurring as intermediates.
In these systems no precipitation was detected. DSC measurements show for the decomposition in solution several strong exothermic effects. The
overalldecomposition enthalpyindiglymeisgivenby ꢀ
2
R
H
s
R s
H
3
3
R
H
d
= −32.0 ± 2.8 kJ/molandin tetraglyme by ꢀ
R
H
t
= −48.0 ± 4.7 kJ/mol. The enthalpy of
solution of cyclotriborazane was determined in diglyme and in tetraglyme with the values ꢀ
D
H
d
= −2.1 ± 0.2 kJ/mol and ꢀ = −4.6 ± 0.5 kJ/mol,
D t
H
respectively.
©
2007 Elsevier B.V. All rights reserved.
Keywords: Cyclotriborazane; Borazine; Thermal decomposition; Hydrogen storage; DSC; 11B NMR
1
. Introduction
[4,5,12,13]. Several groups try to develop processes that allow
to convert the polymeric residues back into ammonia borane
[14,15]. But, beside the unfavorable thermodynamics, the rehy-
drogenation of the products poses a particular challenge because
of the inertness of the polymeric compounds. The polymers
formed are insoluble in organic solvents and a significant effort
is needed for the digestion of the material.
Therefore, it was our interest, to avoid the formation of
any polymeric products during the hydrogen release to get an
improvement in the reactivity. From this point of view, cyclotri-
borazane (“inorganic cyclohexane”) seems to be an interesting
alternative hydrogen generating material, which partly results in
borazine (“inorganic benzene”) during the release of hydrogen
[16,17]. Although from cyclotriborazane only 7 wt.% of usable
hydrogen are released, which is only half of the amount of what
can be obtained from ammonia borane, its complete conversion
to the reactive borazine would improve the chemical situation
in comparison to the inert polymer strongly.
In several papers thermal dehydrogenation reactions of com-
pounds, containing only boron, nitrogen, and hydrogen, called
BNH-compounds”, were reported [1–5]. These compounds are
“
among others useful as hydrogen generating material [6–8].
It is known that, depending on the decomposition conditions,
different products besides hydrogen are formed [1,5,9,10].
The best known species of these hydrogen generating BNH
compounds is ammonia borane, BH3NH3. During its thermal
decomposition, ammonia borane releases hydrogen in two steps.
The first one occurs at 340 K resulting in the production of poly-
meric (BH2NH2)n, followed by the conversion to (BHNH)n at
3
80 K [5,11]. A complete decomposition of ammonia borane
to (BHNH)n provides 13 wt.% usable hydrogen. Unfortunately,
thethermaldecomposition leads mainly to inertpolymeric solids
∗
Different groups calculated a free enthalpy for the dehy-
drogenation in the range of −150 to −155 kJ/mol [18,19] for
borazine and the polymeric products and therefore assumed the
Corresponding author. Tel.: +49 3731 39 2137.
E-mail address: Rene.Schellenberg@chemie.tu-freiberg.de
(
R. Schellenberg).
0
040-6031/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.tca.2007.03.006

1
04
R. Schellenberg et al. / Thermochimica Acta 457 (2007) 103–108
impossibility to reload these products with molecular hydrogen
directly. However, an indirect hydrogenation is not excluded.
Thereforeitisessential, thattheproducthasstillahighreactivity,
like borazine.
were carried out, using the Siemens D5000 Diffractometer.
NMR spectra were obtained by using the 400 MHz spectrometer
DPX-400 from Bruker. Samples were run in 10 mm tubes, each
containing a 2 mm diameter reference tube, filled with 0.19 M
BF3·OEt2 in CDCl3 (0 ppm) as external standard. For quanti-
tative NMR measurements, different diglyme and tetraglyme
solutions with known concentrations of cyclotriborazane and
borazine were prepared and used to compare the peak areas
of each signal. For further confirmation the BF3·OEt2 signal
areas were used for integration likewise. Both methods gave
consistent results. Because the nature of the other species is not
known in detail, it was not possible to create a similar reference
for the interpretation of their signals. Thus, only the quanti-
tative analysis of borazine and cyclotriborazane was carried
out.
Another interesting point is, that the decomposition tem-
perature of cyclotriborazane is moderate enough to perform
the reaction in the dissolved state as well [16]. Because of its
ability to stabilize BNH compounds and to support reactions
with hydrogen transfer in general, glymes in particular need
to be considered as favorable solvents for this type of reac-
tion [20]. The experiment can be easily performed in solution
because cyclotriborazane and borazine show both sufficient sol-
ubility in glymes [10,16]. Also, many reactions with hydrogen
transfer work better in glymes, than in other organic sol-
vents [20,21] and decomposition experiments with ammonia
borane and the currently most commonly used synthesis of
borazine [22] has been already successfully performed in these
solvents.
2.1. Decomposition of the solid cyclotriborazane
The dehydrogenation of ammonia borane in diglyme to
borazine had been described in Ref. [10] and the occurrence
of cyclotriborazane had been mentioned as an intermediate,
but further detailed research of this reaction system, especially
with respect to thermo-physical quantities, such as reaction
and solution enthalpies was done so far only theoretically
(a) In a typical experiment a 25 ml Schlenk vessel was
connected to a cooling trap (195 K). 0.67 g (7.7 mmol)
cyclotriborazane were heated up to 470 K and the sample
was held for additional 6 h at this temperature before it was
slowly cooled down. The residues in the vessel as well as
the residues in the cooling trap were analyzed by XRD and
11
[
8,23–25]. It can be assumed that if the presence of the sol-
B NMR.
vent controls the type of products formed the knowledge of the
thermodynamic data in conjunction with structural information
should support the basic mechanistic ideas. A special interest
lies in the stability of the BN-ring with and without “solvent
support”.
(b) A second kind of heating experiment was performed, using
open vessels of a DTA/TG apparatus. Two to three mil-
ligrams of cyclotriborazane were heated up to 560 K and all
volatile components were transferred to a mass spectrometer
by a helium gas stream.
(
c) For the DSC measurements, approximately 20 mg
(0.23 mmol) of cyclotriborazane in Hastelloy vessels were
heated up to 530 K with a heating rate of 1 K/min. During
some measurements, a pressure detector was used. After
completion of the heating program, the vessels were cooled
2
. Experimental
All synthesis and investigations were performed under argon
atmosphere. All precautions in respect of air and moisture sensi-
tive reagents have been taken. All chemicals used were obtained
from Aldrich. Solvents were dried over sodium and the used
sodium borohydride pellets were pulverized under inert gas
shortly before use.
1
1
down and the products were analyzed by XRD and
NMR.
B
2.2. Decomposition in glymes
CyclotriborazanewassynthesizedasreportedinRef. [16]and
was further purified by vacuum sublimation. The products were
analyzed by XRD, 11B, and H NMR and the obtained data show
very good coincidence with the literature values [1,2,10,16].
All DSC measurements were performed using a differential
scanning calorimeter, type C-80 from Setaram. In some mea-
surements, the used Hastelloy C-276 high pressure vessels have
been connected to a pressure detector PA-21S/80400.3 from
Keller. For the determination of the enthalpy of solution, reversal
mixing vessels were used.
(a) A DSC vessel was filled with 0.5 ml of an 0.3 M solution of
cyclotriborazane in diglyme and heated up to 490 K. After
the subsequent cooling to room temperature, the products
were analyzed by 11B NMR. Further decomposition exper-
iments were performed with various heating rates between
0.05 and 1 K/min. Similar procedures were also carried out
with tetraglyme as the solvent.
1
(b) In another kind of experiments, the prepared solution was
quenched during the reaction at specific temperatures and
11
DTA/TG/QMS investigations were carried out at the
B spectra of the quenched system were taken. Likewise
“
Fraunhofer Institut f u¨ r keramische Technologien und Sinter-
werkstoffe, Dresden” with a differential thermo analyzer STA
29 from Netzsch. The Netzsch 403-3 ceramic blend was cou-
pled to a Balzers QMG 421 quadrupole mass spectrometer.
All enthalpies determined are integral molar and referred to
procedures, using tetraglyme as solvent were carried out.
(c) For the enthalpy of solution in diglyme and tetraglyme,
about 2 mg of cyclotriborazane were dissolved in 2 ml glyme
(0.01 M solution). For each value about five measurements
were taken. Measurements in normal air and under argon
atmosphere were used to determine the air sensitivity of the
samples.
4
1
mol pure cyclotriborazane. For elemental analyzes a CHN O-
Rapid from Heraeus was used. X-ray diffraction measurements

R. Schellenberg et al. / Thermochimica Acta 457 (2007) 103–108
105
Fig. 2. Mass spectrum of the volatile compounds of the thermal decomposition
of cyclotriborazane, recorded at 428 K: the most intensive mass numbers are
m/z = 17 (ammonia) and m/z = 27 (diborane and BNH2). In addition to these
signals, the ones of borazine (m/z = 80) and cyclotriborazane (m/z = 86) and its
fragmentation products were detected.
Fig. 1. Undecoupled 11B spectrum of the products collected by the cooling trap
after the thermal decomposition of neat cyclotriborazane. The spectrum shows
three different signals: a dublett at 30.8 ppm, J = 137.7 Hz (Lit.: 29.1–30.5 ppm,
J = 133–139 Hz [10]) resulting from borazine, a quartet at −19.7 ppm (J = 95 Hz)
caused by unknown compounds containing a BH3 group, a dublett at −22.2 ppm
(
J = 98.4 Hz) indicating the presence of the BH group in another unidentified
equivalently and reached their maxima at the same tem-
perature. This behavior makes it hard to assign the release
of the detected species to the different thermal effects (see
Section 2.1c). The mass spectrum shows the occurrence of
large amounts of ammonia, diborane, borazine, and sub-
limed cyclotriborazane (Fig. 2).
compound. The chemical shifts are referenced to BF3·OEt2.
3
. Results
3
.1. Decomposition of the solid cyclotriborazane
(c) The DSC measurements show at least three different
exothermic effects above 420 K taken at an heating rate of
(
a) The thermal decomposition in Schlenk vessels starts,
depending on the heating rate, at about 420 K and an exten-
sive evolution of hydrogen and some other products that
were eliminated from the gas phase by condensation was
observed. At a temperature of about 430 K the sublima-
tion of cyclotriborazane as a competitive effect was noticed.
Therefore the not all of cyclotriborazane used was finally
decomposed.
1
K/min (Fig. 3).
The first two DSC signals vary in their intensities, but
the third one was always constant. The integration of the
third DSC signal typically yielded a value varying from −16
to −17 kJ/mol. All effects together reach a value of about
−
42 to −47 kJ/mol which corresponds to an enthalpy of
decomposition ꢀRHs = −34.0 ± 2.9 kJ/mol after the pres-
sure correction was applied. The steep pressure rise beyond
The end of the reaction in these experiments was indi-
cated, by the end of any gas evolution. An amorphous and
insoluble white solid remained in the vessel. Elemental anal-
ysis showed that the remaining hydrogen content is 6.4 wt.%
which corresponds to a formula of [BNH1.7]x. The solid
residue possesses most likely a similar polymeric struc-
ture, as the ones reported by several other groups for the
decomposition of ammonia borane and related compounds
450 K indicates a strong formation of gaseous products.
Analyzing the overall pressure increase results in an amount
of 3.1 mol per mol cyclotriborazane of gaseous species
[
5,26].
The cooling trap gave 0.3 ml of a white clouded liquid.
B NMR of the filtered liquid indicates the presence of
1
1
borazine and slight amounts of impurities (Fig. 1). Assum-
ing, that borazine is the only liquid species, the yield is about
4
0 mol% in respect to the amount of cyclotriborazane used.
(
b) In open vessels, the sublimation of cyclotriborazane
increases strongly. TG data showed a weight loss of
about 85 wt.% and the DTA plots obtained show a strong
endothermic effect, superposing the decomposition effect
of cyclotriborazane.
In a measurement of the mass spectra as a function of the
temperature, the intensity of all signals present increased
Fig. 3. Thermal decomposition of cyclotriborazane: plotted is the DSC signal
and corresponding pressure development.

1
06
R. Schellenberg et al. / Thermochimica Acta 457 (2007) 103–108
Fig. 4. DSC plots of the thermal decomposition of cyclotriborazane in diglyme
and tetraglyme: these experiments were done with a heating rate of 0.05 K/min.
Both plots show a remarkable analogy in their occurring effects. The signals
are shifting to higher temperatures with increasing heating rates. For the onset
temperatures of the main effect a shift from 375 to 400 K was found, when the
heating rate is increased to 1 K/min.
Fig. 5. 11B spectrum of the thermal decomposition products of cyclotriborazane:
The spectrum shows no signal of cyclotriborazane (−11.2 ppm, J = 104 Hz [10])
anymore, but the one of the product borazine (30.9 ppm). The broad signal
between 20 and 30 ppm is caused by the reaction of borazine with itself to
oligomeric species. No precipitation or clouding occurred.
released. The only non-gaseous product obtained was an
amorphous white solid. Elemental analysis showed that
it possesses a hydrogen content of 4.8 wt.% which cor-
responds to the formula [BNH1.3]x. No borazine or other
species were detected during this type of experiment.
oligomeric species was borazine. No precipitation or cloud-
ing effects occured during the heating process (Fig. 5).
Quenchedreactionsatabout410 Kalsoshowedtheoccur-
renceofBH speciesasanintermediatewithachemicalshift
3
of about −22 ppm. This shift is similar to what was reported
for ammonia borane (−22.6 ppm [27]). Beside this com-
pound, a location at the end of any open BN chain would
result in a similar chemical shift and needs to be consid-
ered as well. The occurrence of a boron containing species
besides cyclotriborazane and borazine indicates a more dif-
ficult reaction scheme than a simple dehydrogenation, but
we did not succeed in assigning possible reactions steps to
the various peaks in the DSC plots.
3
.2. Decomposition in glymes
(
a) The thermal decomposition in diglyme and tetraglyme are
similar in respect to their thermal effects (see Fig. 4). Both
show an endothermal effect, occurring between 310 and
3
30 K, depending on the heating rate. Various heating and
cooling cycles at different rates indicate a phase transition to
be responsible for this effect. However, it was not possible to
determine the nature of this effect in detail. The 11B NMR
measurements showed no change in the electronic shield-
ing around the boron nuclei. Another DSC signal, which
was not caused by a reversible effect, occurred at 350 K. It
corresponds to a value of about −2 to −3 kJ/mol.
(c) The thermal effect during the dissolving of cyclotribo-
razane in diglyme is smaller (ꢀDHd = −2.1 ± 0.2 kJ/mol)
than in tetraglyme (ꢀDHt = −4.6 ± 0.5 kJ/mol). Remark-
ably, noticeable higher effects were observed in air than in
argon (−2.8 ± 0.1 kJ/mol in diglyme and −7.7 ± 1.1 kJ/mol
in tetraglyme). During the experiments in air the thermal
effects were more pronounced and lasted longer than under
argon and resulted in a slight clouding.
Right after a small precursory exothermal effect start-
ing at about 370 K, at least two overlapping signals,
indicating the main reaction, were detected, whose onset
temperatures at the given rate of 0.05 K/min are about
4. Discussion
3
80 K. These reactions could not be separated by vary-
ing the heating regime because of their similar onset
temperatures and overlapping shapes. After pressure
correction, the overall enthalpy for these effects was
The thermal decomposition of solid cyclotriborazane takes
place in a sequence of at least three exothermal steps as was
confirmed by our calorimetric measurements. The main prod-
ucts are polymeric species with a formula of about [BNH2]x,
but generally the non-polymeric, liquid cyclic trimer [BNH2]3
– borazine – is also created in significant amounts. The yield of
40% is lower than it was reported in Ref. [16], but this might be
an effect of the different set up and the partial sublimation of the
cyclotriborazane out of the heated zone. The mass spectra of the
gas phase during this reaction show, amongst others, high con-
centrations of ammonia, diborane, and monomeric BNH2. Parts
determined to be ꢀRH = −32.0 ± 2.8 kJ/mol in diglyme
d
and ꢀRHt = −48.0 ± 4.7 kJ/mol in tetraglyme. During both
reactions a pressure increase, proportional to the heat
release, was found.
(
b) 11B spectra of the products after the calorimetric exper-
iments in diglyme as well as in tetraglyme showed that
cyclotriborazane during the process was completely decom-
posed. The main product beside small amounts of soluble

R. Schellenberg et al. / Thermochimica Acta 457 (2007) 103–108
107
of the spectra resemble the one of cyclotriborazane, mentioned
in Ref. [10], but other mass numbers in the spectra indicate also
the formation of borazine, which would be the preferred prod-
uct of the reaction. It is also apparent, that besides the cracking
pattern of cyclotriborazane and borazine additional peaks are
found that most likely result form compounds created by the
ring opening of the cyclotriborazane during the dehydrogena-
tion. In contrast to these results, no borazine was found during
and after the C80 measurements. Most likely, the vessel material,
Hastelloy C-276, an alloy of nickel, molybdenum, chromium,
iron, and tungsten, supports the dehydrogenation of borazine,
which was probably formed during the reaction. Further hydro-
gen release by borazine leads to polymeric solid products as
describedinRef. [28]. Similarreactionsinthepresenceofmetals
are reported [29]. The amounts of the gaseous products released,
calculated from the pressure measurements, and the elemental
analyses of the solid residues, both showed that the hydrogen
evolution agrees to the formula [BNH1.3]x for the residue. This
formula corresponds to a lower amount of hydrogen than the one
that was found in the experiments conducted in the glass appara-
tuses and thus would be consistent with the view that additional
hydrogen is released from polymerizing borazine.
The measured enthalpies of solution under argon atmosphere
are exothermal and indicate attractive interactions between
cyclotriborazane and the glymes. It is conceivable, that the
oxygen atoms of the glyme molecules coordinate around
the nitrogen–hydrogen groups (in analogy to ammonium 18-
crown-6 etherate [30]) by forming hydrogen bonds. The more
exothermic value in tetraglyme compared to diglyme is proba-
bly a result of better solvation, because of the hydrogen bonds
that are formed between the two additional oxygen atoms in the
solvent molecule and cyclotriborazane. Solutions of cyclotrib-
orazane in glymes, if stored under argon, are stable for a long
period of time. A 0.8 M solution of cyclotriborazane has been
stored for 3 weeks without clouding. 11B spectra taken after
this period indicated no additional signals than the one of the
pure substance. In contrast to that, the dissolving effects under
air were stronger and a beginning of clouding indicated that a
further reaction took place. This reaction is not ascribable to
autoxidation or to simple self-assembling of cyclotriborazane,
because of the stability of the solution under argon, as long
as no moisture is present. Instead, the cyclotriborazane, if dis-
solved in organic liquids, is very sensitive towards moisture in
normal air and reacts easily—in contrast to the neat substance
which is not even attacked by cold water. The inertness of the
pure cyclotriborazane against water has been reported in Refs.
comparable to the one of borazine, a yield of 70% borazine can
be calculated. In Ref. [15] it has been reported, that polyborazy-
lene can be produced by further dehydrogenation of borazine.
Although we did not find any evidence, that under these condi-
tions this reaction is exclusively responsible for the formation
of these species, we found, that it is possible to suppress the
amount of oligomers formed by lowering the cyclotriborazane
concentration and the temperature during the reaction.
Very similar results gave the experiments we performed with
tetraglyme as solvent and therefore we conclude that similar
results would be obtained in other glymes as well. Obviously,
the number of collisions of the BNH molecules with each other,
which are necessary for polymerization process to occur, are
efficiently suppressed by the presence of the glyme molecules,
which may be due to the strong interaction of the solvent
molecules with the BNH compounds. In this way, glymes sta-
bilize them—especially the reactive borazine, which has a high
tendency to undergo self-polymerization. It even seems to stabi-
lize the unstable open chained intermediates, like the BH3 group
that was detected in some 11B spectra suggests. Remarkably,
the hydrogen release temperature is much lower, i.e. by 30 K
when heated with 1 K/min, than the one for the neat substance.
Obviously the glyme molecules support the decomposition of
cyclotriborazane in a similar way, as they facilitate hydrogena-
tions of organic compounds by supporting the hydrogen transfer
[20]. For example they could promote the formation of the BH3
species. The BH3 occurrence shows, that the dehydrogenation
is not just a simple intramolecular elimination of hydrogen,
but rather a complex reaction system. Some recently performed
screeningexperimentsshowedthattheoccurrenceofthisspecies
has a significant influence on the decay rate of cyclotriborazane,
but more research is needed to answer the question about the
role of the BH3 species.
The enthalpies determined by the DSC experiments show a
stronger exothermic reaction in solution than it does in the solid
state. Unfortunately, the enthalpy of reaction of −34.0 kJ/mol
for the decomposition of the solid cyclotriborazane includes not
only the contribution of the reaction to borazine but also the
contribution by the hydrogen releasing polymerization. It can
be expected, that the polymerization of borazine is exothermic
like dehydrogenation reactions of other BNH compounds and
that therefore the enthalpy for the technically preferred reaction
to borazine is less exothermic.
In Ref. [19] the reaction enthalpy B3N3H12 (s) → B3N3H
6
(l) + H2 (g) has been predicted to be approximately −4.2 kJ/mol
at 298 K. If these results describe the situation at 298 K correctly,
one needs to conclude from our experiments that a change of the
reaction to a more exothermic behavior at higher temperatures,
i.e. above 400 K, occurs.
[
2,16] and explained to be a wetting effect. A similar behavior
was recognized in earlier experiments with several BNH com-
pounds which decomposed in the presence of ordinarily moist
air if they were dissolved in organic solvents. In contrast to that,
solid BNH compounds are fairly stable in moist air. Passivation
effects due to the formation of various hydrolysis products are
likely.
The thermal decomposition in diglyme does not lead to insol-
uble components. Only the desired product borazine and soluble
oligomers, polyborazylene [17,31], were obtained. Assuming,
the NMR sensitivity in our measurements of the oligomer is
In solution, the enthalpy values we measured show only
small differences in comparison to the solid decomposition. The
slightly higher values in tetraglyme are based on the influence of
the solvent and its ability to coordinate around the BNH species.
Because of its less exothermic behavior, weaker solvation, and
consequently its less negative free enthalpy, diglyme might be
the more favorable solvent in respect to possible rehydrogena-
tion procedures.

1
08
R. Schellenberg et al. / Thermochimica Acta 457 (2007) 103–108
In this paper we investigated the thermal decomposition
References
of cyclotriborazane to borazine in solid and in liquid phase.
Undoubtedly, the dehydrogenation of cyclotriborazane at least
in glymes is not just a simple intermolecular hydrogen elim-
ination, unlikely the corresponding organic compounds, but a
reaction that includes several decomposition pathways as can
be seen from the presence of the unexpected BH3 species; a
topic that needs to be considered and investigated in the future.
In general, the comparison with experiments of the hydrogen
decomposition in the gas phase would be of great interest, since
they would allow to identify the influence of the solvent deci-
sively. On the one hand, the calculated enthalpy values could
be confirmed and on the other hand the formation of oligomers
could be suppressed by lowering the cyclotriborazane partial
pressure in a similar way as it was done in solution. For the
possible use of a heterogeneous catalyst, the suppression of the
formation of the oligomers is important since they would adhere
to the surface of the catalyst and cause its inactivation.
[
1] K.W. B o¨ ddeker, S.G. Shore, R.K. Bunting, J. Am. Chem. Soc. 88 (1966)
396–4401.
4
[
[
2] S.G. Shore, C.W. Hickam, Inorg. Chem. 2 (1963) 638–640.
3] D. Kim, K. Moon, J. Kho, J. Economy, C. Gervais, F. Babonneau, Polym.
Adv. Technol. 10 (1999) 702–712.
[4] M. Hu, R. Geanangel, W. Westland, Thermochim. Acta 23 (1978) 249–
55.
2
[
5] F. Baitalow, J. Baumann, G. Wolf, K. Jaenicke-R o¨ ßler, G. Leitner, Ther-
mochim. Acta 391 (2002) 159–168.
[
[
6] W. Grochala, P. Edwards, Chem. Rev. 104 (2004) 1283–1315.
7] J. Baumann, F. Baitalow, G. Wolf, Thermochim. Acta 430 (2005) 9–
14.
[
[
8] M. Gutowski, T. Autrey, Prepr. Pap. - Am. Chem. Soc., Div. Fuel Chem.
4
9 (2004) 275–276.
9] T. Briggs, W. Gwinn, W. Jolly, L. Thorne, J. Am. Chem. Soc. 100 (1978)
762–7763.
7
[
[
[
[
10] J.S. Wang, R.A. Geanangel, Inorg. Chim. Acta 148 (1988) 185–
190.
11] G. Wolf, J. Baumann, F. Baitalow, F.P. Hoffmann, Thermochim. Acta 343
(
2000) 19–25.
12] V. Sorokin, B. Vesnina, N. Klimova, Zh. Neorg. Khim. 8 (1963) 66–
8.
13] J. Carpenter, B. Ault, Chem. Phys. Lett. 197 (1992) 171–173.
5
. Conclusion
6
It has been shown, that the thermal decomposition of neat
[14] S. Hausdorf, G. Wolf, Diploma Thesis, Institut f u¨ r Physikalische Chemie,
TU Bergakademie Freiberg, 2006, pp. 67–72.
cyclotriborazane mainly results in the formation of polymeric
products, having the average formula [BNH2]x, but depending
on the conditions, a formation of borazine in substantial amounts
is also possible. The polymeric residues can be avoided by using
glymes as a solvent. The influence of the solvent on the reaction
is best seen by the lowered temperatures at wich the hydrogen
release takes place. An approximately 70% yield of borazine
had been obtained. NMR, QMS, and DSC experiments show,
that the decomposition is not a simple dehydrogenation, but a
more complex reaction system and an open chained, BH3 group
containing intermediate was detected.
Comparing the experimentally determined values of the reac-
tion enthalpies to calculations of the decomposition without
solvent, significant differences are found, which can be predomi-
nately attributed not only to the reaction conditions present, such
as the temperature and the concentration, but also to the different
interactions of the dissolved species with the glyme molecules.
[
15] N. Mohajeri, A. Raissi, Material Research Society Meeting, Proceedings
Spring, 2005.
[
[
16] G.H. Dahl, R. Scheaffer, J. Am. Chem. Soc. 83 (1961) 3032–3034.
17] C.A. Jaska, K. Temple, A.J. Lough, I. Manners, J. Am. Chem. Soc. 125
(2003) 9424–9434.
[18] T. Heine, G. Seifert, R. Barthel, A. Grigas, T. Lorenz, Prepr. Pap. - Am.
Chem. Soc., Div. Fuel Chem. 51 (2006) 449–450.
[
19] D. Camaioni, T. Autrey, Prepr. Pap. - Am. Chem. Soc., Div. Fuel. Chem.
1 (2006) 568–570.
20] U. P a¨ tzold, Dissertation, Bergakademie Freiberg, 1994, pp. 37–41.
5
[
[21] V. Wege, J. Rechner, E. Zirngiebl, German Patent Application,
DE19,833,095 (2000).
[
[
22] T. Wideman, L.G. Sneddon, Inorg. Chem. 34 (1995) 1002–1003.
23] A.K. Phukan, R.P. Kalagi, S.R. Gadre, E.D. Jemmis, Inorg. Chem. 43
(
2004) 5824–5832.
[24] H. Hirao, H. Fujimoto, J. Phys. Chem. A 104 (2000) 6649–6655.
[25] T.M. Gilbert, Organomet. Chem. 17 (1998) 5513–5520.
[26] E.R. Lory, R.F. Porter, J. Am. Chem. Soc. 95 (1973) 1766–1770.
[
[
27] C.W. Heitsch, Inorg. Chem. 4 (1965) 1019–1024.
28] T. Wideman, P.J. Fazen, A.T. Lynch, K. Su, E.E. Remsen, L.G. Sneddon,
T. Chen, R.T. Paine, Inorg. Synth. 32 (1998) 232–242.
Acknowledgements
[
[
29] E. Wiberg, Naturwissenschaften 35 (1948), 182–188, 212–218.
30] Y.L. Ha, A.K. Chakraborty, J. Phys. Chem. 96 (1992) 6410–6417.
We thank Klaus Jaenicke-Roessler (Institut f u¨ r Keramik und
Sinterwerkstoffe, Dresden) for performing the DTA/TG/MS
experiments and for further support.
[31] P.J. Fazen, J.S. Beck, A.T. Lynch, E.E. Remsen, L.G. Sneddon, Chem.
Mater. 7 (1995) 1942–1956.