Full Paper

DOI: 10.1002/prep.202000187

Heat of Formation of Triazole-Based Salts: Prediction and

Experimental Validation
Julien Glorian,*[a] Kyung-Tae Han,[a] Silke Braun,[a] and Barbara Baschung[a]

Abstract: This work contributes to the growing interest in B3LYP 6-31G(d,p) level with and without corrections. Lattice
predictions linked with energetic salts. A reliable method to enthalpy estimations, based on calculated ionic volumes,
accurately compute the heat of formation of triazole-based were performed with the help of Jenkins and Gutowski
salts was investigated. Calculations were based on Born- methods. Combinations of the obtained results (gas-phase
Haber energy cycles: gas-phase enthalpy of ions and lattice enthalpy of ions and lattice enthalpy) were used in the
enthalpy were calculated separately. Ten triazole-based Born-Haber approach to predict solid phase enthalpy of for-
salts were synthesized and fully characterized. Their heat of mation of studied energetic salts. Direct comparison with
combustion was measured by bomb calorimeter. Gas-phase experimental measurements enabled the identification of
heat of formation of cations and anions were computed at the most reliable path for energetic salt standard enthalpy
four different levels of theory: B3LYP 6-31G(d,p), CBS-4M, of formation prediction.
CBS-QB3, and G4. Ionic volumes were calculated at the
Keywords: Energetic salts · Heat of formation · Computational chemistry · Bomb calorimetry

1 Introduction be more derivatized thanks to their two carbon atoms in

the ring [5]. It was shown by Schmidt et al. that substitution
New generation of energetic materials comes with an ex- does not hinder the ability of the triazolium cation to form
tended list of standards that have to be met. Although high ionic liquids [6]. Many synthesis works concerning triazole-
performances are undoubtedly required, it is nowadays de- based energetic salts were reported by Shreeve [1, 7, 8].
manded to ensure safe handling and to avoid environ- Darwich et al. synthesized and fully characterized (pu-
mental pollution and health risks. Azole-based salts (or ionic rity, molecular structure, sensitivity, and thermal stability)
liquids) are one advanced family of energetic compounds 1,2,4-triazolium-cation-based (methylguanazinium) en-
that agree with the mentioned requirements [1]. The high ergetic salts [5]. The authors calculated solid-phase heats of
nitrogen content provides to the material positive aspects: formation through the Born-Haber cycle using ions elec-
enhanced density, good oxygen balance (less carbon and tronic energies and approximated lattice enthalpies. Ions
hydrogen content), high thermal stability, and environ- electronic energies were computed using Møller-Plesset
mentally friendly (N2 as a main combustion product). Some perturbation theory truncated at the second order (MP2).
of them are greatly insensitive towards outer stimuli, their Predicted salt solid enthalpies of formation were post-proc-
impact sensitivity (IS) can reach 100 J (e. g. IS = 4.4 J for essed by EXPLO5 [9] to obtain detonation pressure and ve-
HMX and 5.4 J for RDX, measured values based on 1 re- locity. This approach is often applied in the literature con-
action in 6 trials according with the BAM procedure). For cerning energetic salts. Different levels of theory can be
the gun propellant purpose, azole-based salts combustion used to calculate ions gas-phase heats of formation: CBS-
delivers a low flame temperature, which could prevent ero- 4M [2, 10, 11], DFT-B3LYP [12, 13], MP2 [14, 15, 16, 17], G2
sion in the gun tube, and maintains a valuable force [18] and G3 [19]. A few studies report back-calculated solid
(around 1000 J g 1). In addition, they are known for their salt heat of formation from bomb calorimeter measure-
high heat of formation, which is strongly coupled with their ments [20, 21]. In most of the references, calculated or back-
high nitrogen content. The heat of formation is a key prop- calculated, the heat of formations are used in thermochem-
erty, such as density, because it enables to estimate other
properties of interest like a specific impulse, detonation ve-
[a] J. Glorian, K.-T. Han, S. Braun, B. Baschung
locity, and pressure [2, 3], and sensitivity [4].
Propellant and Energetic Materials (PEM) group
A triazole consists of a five membered heterocycle that French-German Research Institute of Saint-Louis
contains 3 nitrogen atoms with two possible isomeric con- 5 rue du Général Cassagnou, 68300 Saint-Louis, France
figurations: 1,2,3-triazole and 1,2,4-triazole. Although they *e-mail: [email protected]
have a potentially lower energy content (linked with the Supporting information for this article is available on the WWW
number of nitrogen content) than tetrazoles, triazoles can under https://doi.org/10.1002/prep.202000187

124 © 2020 Wiley-VCH GmbH Propellants Explos. Pyrotech. 2021, 46, 124–133
Heat of Formation of Triazole-Based Salts: Prediction and Experimental Validation

ical programs that predict performances such as EXPLO 5 Df H0ðsalt; 298:15 K Þ ¼ Df H0ðg; cation;298:15 K Þ þ
[9] or CHEETAH 4.0 [22]. (1)
Energetic salts synthesis and testing are costly in terms Df H0ðg; anion;298:15 K Þ DHL
of time and money. Since heat of formation is a key prop-
erty, relevant predictions seem essential to obtain con-
sistent performance properties. The accuracy of calculated Calculations were performed with the Gaussian 16® soft-
enthalpies of formation is not necessarily discussed in all ware [29].
the reported studies. Byrd and Rice investigated a variety of
methods able to predict accurately the solid phase heat of
formation of energetic salts Df H0ðsalt; 298:15 K Þ [23]. They com- 2.1 Ions Enthalpy of Formation
pared computed data with experimental values and they
obtained their best predictions using G3MP2B3 level of Anions and cations electronic energies were computed at
theory combined with lattice enthalpy calculations using four different levels of theory. For each ions, 5 to 10 distinct
the Jenkins [24, 25, 26] and Gutowski model [27]. These conformers (depending on the complexity of the molecular
models are extensively used in the community and will be structure) were investigated and the chemical structure that
described in the computational approach section. The au- ensured the lowest energy was selected.
thors also pointed out the lack of experimental data, which First, ions gas-phase standard enthalpy of formation
might mitigate conclusions on the accuracy of their meth- Df H0298:15K ðgÞ was computed by following the procedure pro-
odology. More recently, Forquet et al. evaluated calculation posed by Osmont et al. [30]. This method has been devel-
methods to determine solid phase heats of formation for oped for the prediction of thermochemical data of en-
2,2-dimethyltriazanium-cation-based energetic salts [28]. ergetic materials. It is based on a geometry optimization
Their comparison between computed and experimental and frequency calculation at the DFT B3LYP/6-31G(d,p) lev-
data led to a reliable methodology combining a CBS-QB3 el. The gas-phase enthalpy of formation is then determined
level for the gas phase enthalpy calculation of ions of inter- with the Equation (2).
est, a non-corrected B3LYP/6-31G(d,p) calculation for the
molecular volume, and the Gutowski model to obtain the !
X
0
lattice enthalpy. Df H 298:15K ðgÞ ¼ 2625:5 � E þ DHcorr þ ai c i (2)
In this paper, both experimental measurement and cal- i

culations of the heat of formation are considered. Ten tria-

zole-based-salts were synthesized and characterized. Their Gaussian 16® software [29] outputs energies in Hartree.
heats of combustion (Qcomb) were experimentally measured molecule-1. In Equation (2), 2625.5 is the conversion factor
by bomb calorimetry. Electronic energies were computed at from Hartree.molecule 1 to kJ.mol 1. E and ΔHcorr are, re-
four different level of theories to calculate ions gas-phase spectively, the electronic energy and the thermal correc-
enthalpy of formation. Molecular volumes were also com- tion. αi is the number of atom i and ci is the atomic correc-
puted and Jenkins and Gutowski models were applied to tion for atom i, more details are reported in Osmont et al.
obtain the lattice enthalpy. Calculated Df H0ðsalt; 298:15 K Þ are [30].
post-processed to determine Qcomb. Combinations between Compound methods were also performed to obtain ions
various quantum chemical methods and lattice energy gas-phase standard enthalpies of formation. Enthalpies of
models were confronted with experimental data and the formation were obtained by using an atomization energy
methodology that provided the best agreement is reported. scheme (Equation (3)).
�
X �
Df H0ðg; ion;298:15 K Þ ¼ Hðion;298:15 K Þ þ Hðatom;298:15 K Þ þ
2 Theoretical Approach X (3)
Df H0ðatom;298:15 K Þ
An equation obtained from the Born-Haber cycle approach
�
(Equation (1)) is used to estimate the solid phase standard Enthalpies of the ions (Hðion;298:15 K Þ ) were computed with
heat of formation Df H0ðsalt; 298:15 K Þ of the energetic salt [1]. It three different compound methods: CBS-4M, CBS-QB3, and
requires 3 main physicochemical properties that can be G4. While CBS-4M is routinely used by Klapötke et al.
predicted thanks to ab initio calculations: Gas-phase stan- [2, 10, 31, 32] and Hermann et al. [33], Forquet et al. [28]
dard heat of formation of the anion and the cation demonstrated the effectiveness of using the CBS-QB3 meth-
Df H0ðg; cation;298:15 K Þ and Df H0ðg; anion;298:15 K Þ and the lattice en- od. The last compound method used in this study is the G4
thalpy (DHL ). method [34]. The G4 was designed to be a highly accurate
method. However, it has a considerable calculation cost
compared to the previously mentioned methods. One of
the goal here is to evaluate the gain in accuracy when in-
creasing the calculation cost. Values for calculated atom en-

Propellants Explos. Pyrotech. 2021, 46, 124–133 © 2020 Wiley-VCH GmbH www.pep.wiley-vch.de 125
Full Paper J. Glorian, K.-T. Han, S. Braun, B. Baschung
�
thalpy (Hðatom;298:15 K Þ ) and experimental (Df H0ðatom;298:15 K Þ ) stan- Enthalpy of combustion (ΔHcomb) is then calculated by
dard enthalpy were collected from, respectively, the Com- the difference between the standard enthalpy of combus-
putational Chemistry Comparison and Benchmark DataBase tion of the products and reactants according to reaction (7).
(CCCBDB) [35] and the NIST chemistry WebBook [36]. Since N2(g) and O2(g) correspond to the most stable forms
of nitrogen and oxygen in their standard state, at the refer-
ence temperature of 298.15 K, their standard enthalpies of
2.2 Lattice Enthalpy Calculation formation are equal to zero.

Lattice enthalpy was estimated thanks to volume based cal- b

culations introduced by Mallouk et al. [37] and Jenkins et al. DHcomb ¼ aDf H0CO2ðgÞ þ Df H0H2OðlÞ Df H0ðsalt; 298:15 K Þ (8)
2
[24, 25]. The first step is to separately optimize the struc-
tures of the cation Mq + and the anion Xp . The volume of ΔHcomb is the enthalpy of reaction according to reaction (7),
ions was obtained from an isodensity surface polarized con- the heat of combustion is obtained by Qcomb =- ΔHcomb.
tinuum model (IPCM) calculation at the DFT B3LYP/6-31G(d, Df H0CO2ðgÞ and Df H0H2OðlÞ are taken from the NIST database
p) level, similar method as Forquet et al. [28]. Corrections [35], respectively, 393.52 and 285.83 kJ.mol 1.
proposed by Rice et al. [38] were evaluated. Results from
ion volume calculations were used in Equation (4).
� 3 Results and Discussion
V Mp X q � pV ðMqþ Þ þ qV ðX p Þ (4)
3.1 Volume Calculations
Volumes calculated from equation 4 were validated
thanks to density measurements, after comparison with ex- Since lattice energy predictions are based on computed
perimental data it was decided to use, or not, Rice et al. [38] volumes, energetic salts volumes should be accurately pre-
corrections. Volumes were then used in Equation (5) to ob- dicted. To validate computed volumes, density measure-
tain lattice energies. ments for each salt were conducted with a gas pycnometer.
� � Experimental salts molecular volumes could be obtained
� a thanks to V = M/1.
DUL Mp X q ; s ¼ ðpq2 þ qp2 Þ p ffi ffiffi þ b (5)
3
V Volume calculations were performed for each ion pre-
sented in Figure 1. The calculation protocol is explained in
α and β are coefficients taken from Jenkins et al. [25] and section 2 and computed ion volumes are listed in Table 1.
Gutowski et al. [39]. Lattice energy estimations are used in Calculated salts volumes are plotted against ex-
Equation (6) to obtain lattice enthalpies. perimental ones (Figure 2). Corrected and non-corrected
volumes are in good agreement with experimental meas-
� � �n � �n �� urements. Mean errors for non-corrected and corrected are,
DHL Mp X q ; s ¼ DUL þ p M 2 þq X 2 RT (6)
2 2 respectively, 4.0 % and 2.8 %.
Better accuracy was obtained by using Rice et al. correc-
nM and nX values are, for the cation and the anion, 3 for tions, they were then applied for lattice enthalpy estima-
monoatomic ions, 5 for linear polyatomic ions, and 6 for tions. Lattice enthalpy was estimated twice for each salt:
nonlinear polyatomic ions.

2.3 Heat of Combustion Calculation

Calculated heats of combustion were obtained with calcu-

lated solid-phase standard enthalpy of formation that are
based on Jenkins and Gutowski methods. During Qcomb
bomb calorimeter measurements, the studied energetic salt
burns in excess of oxygen. In this case, CHNO species com-
bustion products are liquid H2O, gaseous CO2, and N2:
� �
b c
Ca Hb Oc Nd þ aþ O ðgÞ !
4 2 2
(7)
b d
aCO2 ðgÞ þ H OðlÞ þ N2 ðgÞ
2 2 2
Figure 1. Cations (from 1 to 6) and anions (from A to D).

126 www.pep.wiley-vch.de © 2020 Wiley-VCH GmbH Propellants Explos. Pyrotech. 2021, 46, 124–133
Heat of Formation of Triazole-Based Salts: Prediction and Experimental Validation

Table 1. Calculated ion volumes with and without Rice et al. cor- Table 2. Calculated ion enthalpy of formation at different levels of
rections. theory. The results are in kJ.mol 1.

Ions Calculated Calculated volume with Ions B3LYP/ CBS-4M CBS-QB3 G4

volume (Å3) Rice et al. corrections (Å3) 6-31G (d,p)
1 111.1 104.7 1 698.4 818.4 819.5 726.5
2 96.6 91.2 2 842.1 847.5 851.9 767.9
3 108.6 101.4 3 1907.4 1914.4 1916.6 1826.7
4 79.3 74.8 4 806.6 834.4 832.3 755.9
5 98.2 92.8 5 1103.1 1121.9 1116.7 1032.9
6 140.2 132.9 6 818.6 711.4 712.9 617.4
A 117.0 115.4 A 90.4 56.5 53.1 27.2
B 121.8 120.2 B 146.7 204.4 214.0 299
C 139.0 138.4 C 17.9 52.3 61.5 150.5
D 156.8 151.4 D 84.3 124.0 152.9 249.4

each salt with Jenkins and Gutowski model and at four dif-
ferent levels of theory (Equation (1)). Based on these com-
puted enthalpies of formation, heats of combustion were
calculated and used for direct comparison with ex-
perimental heat of combustion (Figures 3 and 4). Since Gu-
towski coefficients only applies to 1 : 1 salts, the heat of
combustion for the 3.A.A. salt (Diamino-1,2,4-triazolium-bis-
(4-nitro-1,2,3-triazolate)) does not appear in Figure 3.
For each energetic salt investigated, bomb calorimeter
measurements were conducted 3 times. Since the mean ex-
perimental deviation was found to be 0.2 % at the most.
Disparities observed, between calculated and experimental
heat of combustion, are essentially caused by calculation
errors.
Figure 3 reports calculated energetic salts heat of com-
Figure 2. Non corrected vs corrected calculated salt volume plotted bustion based on a lattice enthalpy estimated with Gu-
against experimental salt volume. towski method and at four different levels of theory for ions
enthalpy of formation. A good overall agreement can be
noted.
one with Gutowski method and the other with Jenkins
method.

3.2 Calculated Heat of Combustion

As it is explained in Section 2, 4 types of calculations were

performed for ion standard enthalpy of formation. Results
are presented in Table 2. For each ion, calculated values at
different levels of theory are in the same range. However,
differences from few decades kJ mol 1 to few hundreds can
be observed from a levels of theory to another. Results
from CBS methods are quite similar. Predictions obtained
with CBS methods are 100 kJ mol 1 higher than G4 ones.
Results calculated with DFT B3LYP/6-31G(d,p) are in the
same range or higher than CBS methods calculated ones.
These differences will certainly impact the accuracy of the
final result. The level of theory used for calculations of ions
enthalpy of formation seems crucial. Figure 3. Calculated heat of combustion with lattice enthalpy esti-
Lattice enthalpy estimations and results from Table 2 mated Gutowski et al. method vs experimental heat of combustion
made possible the calculation of enthalpy of formation of (kJ.mol 1).

Propellants Explos. Pyrotech. 2021, 46, 124–133 © 2020 Wiley-VCH GmbH www.pep.wiley-vch.de 127
Full Paper J. Glorian, K.-T. Han, S. Braun, B. Baschung

deficiencies coming from the ion gas-phase enthalpy calcu-

lated with CBS methods.
It is important to mention that performing calculations
at the G4 level is more expensive in terms of calculation
time than the other tested compound methods. However,
the calculation time for each studied ion did not exceed
30 hours at the G4 level (4 cores involved), which can be
considered reasonable. CBS methods calculations and
B3LYP 6–31/G(d,p) did not exceed, respectively, 15 hours
and 45 minutes.

4 Conclusion

The synthesis of 10 triazole-based salts have been success-

fully conducted. In addition to a full characterization, their
Figure 4. Calculated heat of combustion with lattice enthalpy esti- heat of combustion has been measured for each energetic
mated Jenkins et al. method vs experimental heat of combustion salt by a bomb calorimeter.
(kJ.mol 1). Relevant predictions of heat of formation were obtained
thanks to quantum chemical calculations. The most accu-
rate calculation path involves a G4 level coupled with atom-
Figure 4 presents results obtained by applying Jenkins ization reactions for ions standard enthalpy of formation
coefficients. It can be observed that applying G4 method prediction. Lattice enthalpy were based on ionic volume
leads to more accurate results. This trend is discernible in calculation at B3LYP/6-31G(d,p) with Rice et al. corrections
Figure 4 and confirmed by the calculated mean error of for ionic volumes and Jenkins coefficients. Computed en-
2.3 % (Table 3). This result was expected since the accuracy thalpies of formation were converted to the heat of com-
of the G4 level of theory, for nitrogen-rich cations, have bustion for direct comparisons with experimental measure-
been reported by Nirwan and Ghule [40]. The other tested ments. A 2.3 % mean error was estimated following this
levels of theory (B3LYP/6-31G(d,p), CBS-4 M, and CBS-QB3), calculation path and calculation time is reasonable.
combined with Jenkins coefficients, overestimated the heat This work aims at improving predictions concerning
of combustion. azole-based salt materials. It will support future activities in-
For both methods, Gutowski and Jenkins, the worst volving triazole-based salts. In future works, it is planned to
agreement was obtained using B3LYP/6-31G(d,p), 7.2 %, check the reliability of the method for other types of en-
and 4.3 %, respectively. According to Table 3, similar mean ergetic salts such as tetrazole and/or ammonium-based
errors (3.6 % with CBS-QB3, 3.7 % with G4 to 3.8 % with CBS- salts.
4 M) are reported when using compound methods and Gu-
towski coefficient. Forquet et al. [28] and Klapötke et al. [2]
obtained their most accurate results using a CBS level com- 5 Experimental Section
bined with Gutowski method. The large differences of cal-
culated ions gas-phase enthalpy observed in Table 2 re- 5.1 General Methods
vealed the importance of the level of theory applied. Since
calculations involving a G4 level combined with Jenkins The isoperibolic Qcomb measurements were conducted using
method were found to be most accurate, it can be ex- an IKA 2000 basic bomb calorimeter. A decomposition ves-
pected that Gutowski lattice model absorbs a part of the sel is equipped with an oxygen valve, outlet valve for the
combustion gases and a holding and electrical ignition de-
vice. For the Qcomb measurement, the EM sample was press-
ed into a pellet of 1 g, placed in the decomposition vessel
Table 3. Heat of combustion calculated mean errors (%) and root
mean squarred errors (kJ.mol 1).
which was filled with 3 MPa of pure oxygen to ensure com-
plete combustion. The decomposition vessel is immersed in
Mean errors (%) RMSE (kJ.mol 1) a measuring cell filled with water at a temperature of 25 °C.
Method Jenkins Gutowski Jenkins Gutowski The measuring cell consists of an inner vessel surrounded
B3LYP 7.2 4.3 978.4 609.6 by an insulating water jacket with a fixed temperature to
6-31G(d,p) ensure constant parameters. The temperature curve of the
CBS-4 M 6.9 3.8 922.9 571.1 inner water vessel is registered during the combustion
CBS-QB3 6.5 3.6 873.4 522.8 process and the temperature increase permits the calcu-
G4 2.3 3.7 323.4 460.5 lation of heat of combustion. Results of Qcomb with Wash-

128 www.pep.wiley-vch.de © 2020 Wiley-VCH GmbH Propellants Explos. Pyrotech. 2021, 46, 124–133
Heat of Formation of Triazole-Based Salts: Prediction and Experimental Validation

burn corrections applied can be found in the supple-

H 13C 15N NMRa,IRb, DTAc, DSCd

H 13C 15N NMRa,IRb, DTAc, DSCd
H 13C 15N NMRa,IRb, DTAc, DSCd

H 13C 15N NMRa,IRb, DTAc, DSCd

mentary materials.
IR spectra were recorded by using ATR for solids on a
Bruker Tensor 27 spectrometer.
The density was measured by a pycnometer Pycnomatic
from Porotec at standard conditions (293.15 K, 105 Pa) with
helium.

H 13C NMRa,
1
H, 13C, and 15N NMR spectra were recorded with Varian

IRb, DTAc
Analysis
method
MR 400 nuclear magnetic resonance spectrometer operat-
ing at 400, 100, and 41 MHz, respectively at 25 °C. Chemical

1
shifts are reported relative to TMS for 1H and 13C NMR spec-
tra and to CH3NO2 for 15N spectra. The solvent was dimethyl

fraction purity
sulfoxide-d6. NMR spectra are given in the supplementary

Final mole
material.

> 0.99

> 0.99
> 0.99
> 0.99

> 0.95
The differential thermal analysis was measured using SII
TG/DTA7200 and the differential scanning calorimetry by
using SII X–DSC7000 (SII NanoTechnology Inc.) with a heat-
ing rate of 5 K min 1. The chemicals as well as their purities

Purification method
and suppliers are listed in Table 4.

Differential scanning calorimetry

recrystallization

recrystallization
recrystallization
recrystallization

recrystallization
5.2 Physicochemical Properties

none
none
none
none
none

none
none
The thermal stabilities of the triazole-based salts were de-
termined by differential thermal analysis (DTA) measure-

84406-63-3
ments (Table 4). Given that the salts have good thermal sta-
1455-77-2
932-64-9

584-13-4
288-88-0
108-78-1

505-71-5
CAS no.

61-82-5

67-64-1
64-17-5
bility with decomposition temperatures ranging from 187
(2.D.) to 303 °C (6.B.). The density was measured at standard
conditions (293.15 K, 105 Pa) with a helium pycnometer and

d
the values are in the range of 1.621 (2.A.) to 1.722 g cm 3 (5.

Infrared spectroscopy, c Differential thermal analysis,

fraction purity

C.). Melting temperatures are reported except for 1.B. and 6.

Initial mole

B. No melting points were observed for both ionic salts 1.B.

and 6.B. They have NTO anions in common, which, in its
> 0.98
0.95
0.98
0.99
0.99

molecular form, is known for its high melting temperature

(more than 270 °C).
Impact and friction sensitivities were measured accord-
Zundel Kohler France
Sigma-Aldrich France
Sigma-Aldrich France

Sigma-Aldrich France

ing to the standard BAM methods [41]. Most of the salts

Alfa Aesar France

Alfa Aesar France

EURENCO France

can be considered as highly insensitive to impact. The low-

est measured impact sensitivity is 10.4 J (4.C.) which is still
VWR France

higher than classical insensitive energetic materials (For

synthesis
synthesis
synthesis

synthesis

RDX and HMX, IS � 7.4 J). Studied ionic salts are highly in-
Source

sensitive to friction since the lowest value is 288 N (1.D.).

Back-calculated standard heat of formation from ex-
Table 4. Specification of Chemical Samples.

perimentally measured heat of combustion are presented.

1,3,5-triazine-2,4,6-triamine (melamine)

Reported calculated standard enthalpies are the ones com-

puted with the most accurate path (G4 calculation and Jen-
sodium-3,5-dinitro-1,2,4-triazolate

kins method) previously presented. Oxygen balance and Ni-

Nuclear magnetic resonance,

trogen content are also reported in Table 5.

3-nitro-1,2,4-triazol-5-one

3,5-diamino-1,2,4-triazole
3-amino-1,2,4-triazole
4-amino-1,2,4-triazole

4-nitro-1,2,3-triazole

ethylendinitramine

5.3 Synthesis
Chemical name

Energetic salts
1,2,4-triazole

The studied ions are presented in Figure 1 and the syn-

theses, based on the Xue et al. work were slightly modified
acetone
ethanol

[42].
a

Propellants Explos. Pyrotech. 2021, 46, 124–133 © 2020 Wiley-VCH GmbH www.pep.wiley-vch.de 129
Full Paper J. Glorian, K.-T. Han, S. Braun, B. Baschung

Table 5. Physicochemical properties.

Salt 1a OBb Tmc Tdd Ne ΔfH expf ΔfH calcg ISh FSi
1.B 1.658 � 0.005 59.3 – 238 55.0 112.5 � 0.2 67.0 > 30 360
1.C 1.734 � 0.027 43.4 204 236 54.3 133.2 � 0.3 91.3 > 24.5 360
1.D 1.635 � 0.002 61.0 134 187 50.6 20.0 � 0.1 6.5 12.3 288
2.A 1.621 � 0.001 72.7 121 226 56.6 239.6 � 0.3 235.1 > 24.5 324
2.D 1.645 � 0.0003 75.4 110 187 52.8 63.0 � 0.1 0.2 19.6 324
3.A.A 1.627 � 0.0002 61.1 114 214 55.6 355.8 � 0.3 242.7 12.3 > 360
4.C 1.699 � 0.014 42.1 186 222 49.1 215.7 � 0.2 104.0 10.4 360
5.A 1.623 � 0.001 72.7 97 214 56.6 381.0 � 0.3 501.1 > 24.5 360
5.C 1.722 � 0.002 42.8 120 215 51.9 328.7 � 0.1 391.4 22.1 360
6.B 1.675 � 0.026 68.7 – 303 54.7 85.8 � 0.1 161.2 > 30 > 360
a
Measured density (g cm 3, p0 = 105 Pa, T = 293.15 K). b Oxygen Balance (%). c Melting Point under nitrogen (°C, p0 = 105 Pa). d Decomposition
temperature (°C). e Nitrogen Content (%). f Back-calculated standard heat of formation (kJ mol 1), the uncertainties of the mean were
calculated from the uncertainty of the measurements of the heat of combustion, with benzoic acid as standard. g Calculated standard heat of
formation (kJ mol 1). h BAM test Impact sensitivity (J). i Friction sensitivity (N). Uncertainties given in the table are expanded uncertainties at
0.95 level of confidence.

1.B. 3,5-Diamino-1,2,4-triazolium-3-nitro-1,2,4-triazol- 1.D. 3,5-Diamino-1,2,4-triazolium-ethylendinitraminate

5-onate
To a suspension of 3,5-diamino-1,2,4-triazole (0.99 g, 10
3-Nitro-1,2,4-triazol-5-one (1.31 g, 10 mmol) was dissolved mmol) in 50 ml acetone a solution of ethylendinitramine
in 50 ml warm ethanol and afterwards 3,5-diamino-1,2,4-tri- (1.5 g, 10 mmol) in 40 ml acetone was added. The solvent
azole (1 g, 10 mmol) was added. The orange colored prod- was evacuated under reduced pressure. The precipitate was
uct precipitated while stirring. It was filtered off, washed recrystallized in ethanol to produce white crystals. Yield at
with ethanol and dried under reduced pressure, yield 76 % (1.9 g).
93.0 % (2.15 g). IR (ATR): 3461, 3417, 3344, 2726, 1707, 1637, IR (ATR): 3222, 3108, 1622, 1591, 1560, 1486, 1447, 1403,
1604, 1525, 1494, 1360, 1305, 1103, 1054, 1019, 988, 889, 1346, 1308, 1243, 1118, 1058, 788, 772, 596 cm 1; 1H-NMR δ:
846, 778, 761, 741, 678, 656, 616 cm 1; 1H-NMR δ: 9.06 ppm; 3.60 ppm; 13C-NMR δ: 42.4, 158.4 ppm.
13
C-NMR δ: 154.0, 156.6, 162.6 ppm; 15N-NMR δ: 22.4,
111.7, 175.9, 194.8, 199.1, 200.1, 225.8,
328.7 ppm. 2.A. 3-Amino-1,2,4-triazolium-4-nitro-1,2,3-triazolate

To a warm 100 ml acetone solution of 3-amino-1,2,4-tria-

1.C. 3,5-Diamino-1,2,4-triazolium-3,5-dinitro-1,2,4-triazo- zole (0.84 g, 0.1 mol) was added 4-nitro-1,2,3-triazole
late (1.14 g, 0.1 mol) in 40 ml acetone. The clear solution was
stirred, the solvent was evacuated to produce pale yellow
Sodium-3,5-dinitro-1,2,4-triazolate (5 g, 25 mmol) was sus- crystals. The precipitate was recrystallized in ethanol. Yield
pended in diethyl ether and 10 % aqueous HCl was drop- at 100 % (1.98 g).
ped until the solution becomes clear. The organic phase IR (ATR): 3424, 3341, 3149, 2979, 2322, 1979, 1638, 1529,
was dried and concentrated under an air stream. The con- 1494, 1384, 1231, 1035, 966, 863, 827, 747, 612 cm 1; 1H-
centrated solution was added to a suspension of 3,5-dia- NMR δ: 7.69, 8.96 ppm; 13C-NMR δ: 126.6, 146.5, 154.0,
mino-1,2,4-triazole (1 g, 10 mmol) in ethanol. The solution 157.0 ppm; 15N-NMR δ: 21.7, 37.8, 41.2, 91.9, 119.6,
was heated to boiling and the product precipitated while 178.4, 193.0, 330.3 ppm.
cooling. It was filtered off, washed with ethanol, and dried
under reduced pressure. Yellow solid product at 45 % yield
(1.04 g). IR (ATR): 3467, 3423, 3361, 3288, 3189, 2609, 1670, 2.D. 3-Amino-1,2,4-triazolium-ethylendinitraminate
1610, 1545, 1492, 1397, 1353, 1300, 1129, 1059, 998, 872,
845, 832, 796, 766, 694, 647 cm 1; 1H-NMR δ: 6.95, Ethylendinitramine (0.75 g, 5 mmol) was dissolved at room
11.23 ppm; 13C-NMR δ: 151.4, 162.9 ppm; 15N-NMR δ: 20.2, temperature in 50 ml acetone and this solution was added
50.5, 146.3, 326.5 ppm. to a boiling solution of 4-amino-1,2,4-triazole (0.84 g,
10 mmol) in 50 ml acetone. The clear solution was stirred,
the solvent was evacuated to produce white crystals at
100 % (1.59 g) yield.

130 www.pep.wiley-vch.de © 2020 Wiley-VCH GmbH Propellants Explos. Pyrotech. 2021, 46, 124–133
Heat of Formation of Triazole-Based Salts: Prediction and Experimental Validation

IR (ATR): 3404, 3213, 2926, 2724, 1632, 1592, 1536, 1446, 5.C. 4-Amino-1,2,4-triazolium-3,5-dinitro-1,2,4-triazolate
1404, 1348, 1268, 1245, 1211, 1118, 1044, 967, 876, 825,
789, 727, 596 cm 1; 1H-NMR δ: 3.60, 5.75, 7.44, 12.09 ppm; Freshly prepared 3,5-dinitro-1,2,4-triazole (~ 3 g, 19 mmol)
13
C-NMR δ: 42.3, 147.6, 157.8 ppm. was dissolved in 25 ml ethanol and 4-amino-1,2,4-triazole
(1 g, 12 mmol) was added. The solution was heated to boil-
ing and the product precipitated while cooling. The precip-
3.A.A. 3,5-Diamino-1,2,4-triazolium-bis-(4-nitro-1,2,3-triazo- itate was recrystallized in ethanol. Yellow solid, yield up to
late) 54 % (1.56 g)
IR (ATR): 3325, 3241, 3155, 2469, 1638, 1549, 1493, 1421,
To a suspension of 4-nitro-1,2,4-triazole (1.14 g, 10 mmol) in 1384, 1353, 1301, 1209, 1113, 1079, 1029, 992, 928, 871,
50 ml acetone a solution of 3,5-diamino-1,2,4-triazole (0.5 g, 843, 827, 648, 614 cm 1; 1H-NMR δ: 9.04, 9.55 ppm; 13C-NMR
5 mmol) in 40 ml acetone was added. The solvent was evac- δ: 144.4, 163.1 ppm.
uated under reduced pressure. The precipitate was re-
crystallized in ethanol to produce yellow crystals. Yield at
98 % (1.6 g). 6.B. 2,4,6-Triamino-1,3,5-triazinium-3-nitro-1,2,4-triazol-
IR (ATR): 3465, 3420, 3157, 2589, 1685, 1650, 1511, 1456, 5-onate
1411, 1373, 1286, 1206, 1136, 1063, 1028, 1001, 853, 830,
784, 757, 730, 708, 661 cm 1; 1H-NMR δ: 8.79, 9.61 ppm; 13C- 3-Nitro-1,2,4-triazol-5-one (2 g, 15 mmol) was dissolved in
NMR δ: 127.0, 153.7, 153.8 ppm; 15 N-NMR δ: 20.6, 28.5, 40 ml warm ethanol and afterwards, melamine (1.94 g,
40.0, 78.7, 196.4, 225.3, 328.3 ppm. 15 mmol) was added. The solution was heated to boiling
and the product precipitated while cooling. Yield up to
85 % (3.35 g).
4.C. 1,2,4-Triazolium-3,5-dinitro-1,2,4-triazolate IR (ATR): 3195, 1690, 1542, 1474, 1338, 1187, 1006, 829,
785, 753, 723, 680, 602 cm 1; 1H-NMR δ: 6.99 ppm; 13C-NMR
Sodium-3,5-dinitro-1,2,4-triazolate (2.8 g, 16 mmol) was sus- δ: 157.7, 163.8, 163.7 ppm ; 15N-NMR δ: 20.6, 111.8,
pended in diethyl ether and 10 % aqueous HCl was drop- 198.6, 206.4, 216.4, 294.2 ppm.
ped until the solution becomes clear. The organic phase
was dried and concentrated under an air stream. The con-
centrated solution was added to a solution of 1,2,4-triazole Acknowledgements
(1 g, 14 mmol) in ethanol. The yellow product precipitated
while stirring. It was filtered off, washed with ethanol, and This work was possible thanks to the experimental efforts con-
dried under reduce pressure. The precipitate was recrystal- ducted by Fabrice Ciszek and Bernd Wanders (ISL).
lized in ethanol. Yellow solid product at 70 % yield (2.24 g).
IR (ATR): 3165, 3128, 3063, 2902, 2788, 2576, 2516, 2476,
1955, 1669, 1580, 1542, 1500, 1457, 1414, 1394, 1359, 1314, Data Availability Statement
1305, 1281, 1256, 1182, 1128, 1111, 1062, 1030, 1001, 956,
942, 931, 881, 840, 826, 772, 664, 649, 627, 606, 533 cm 1; No data available.
1
H-NMR δ: 9.34, 15.2 ppm; 13C-NMR δ: 143.2, 162.8 ppm.

References
5.A. 4-Amino-1,2,4-triazolium-4-nitro-1,2,3-triazolate
[1] H. Gao, J. M. Shreeve, Azole-based energetic salts, Chem. Rev.
4-Amino-1,2,4-triazole (0.84 g, 10 mmol) and 4-nitro-1,2,3- 2011, 111, 7377–7436.
triazole (1.14 g, 10 mmol) was dissolved in 30 ml acetone [2] T. M. Klapötke, J. Stierstorfer, H. D. B. Jenkins, R. van Eldik, M.
respectively and were put together to form a clear solution. Schmeisser, Calculation of Some Thermodynamic Properties
The solvent was evacuated under reduce pressure. The pre- and Detonation Parameters of 1-Ethyl-3-Methyl-H-Imidazolium
cipitate was recrystallized in ethanol, pale yellow crystals at Perchlorate, [Emim][ClO4], on the Basis of CBS-4M and CHEE-
TAH Computations Supplemented by VBT Estimates, J. Inorg.
95 % (1.9 g) yield.
Gen. Chem. 2011, 637, 1308–1313.
IR (ATR): 3364, 3294, 3132, 3112, 2436, 1952, 1601, 1563, [3] H. Gao, C. Ye, C. M. Piekarski, J. M. Shreeve, Computational
1521, 1484, 1455, 1391, 1375, 1227, 1186, 1065, 1038, 1018, Characterization of Energetic Salts, J. Phys. Chem. C 2007, 111,
981, 949, 916, 880, 824, 759, 698, 682, 622, 537 cm 1; 1H- 10718–10731.
NMR δ: 8.41, 9.10 ppm; 13C-NMR δ: 126.0, 144.6, 153.9 ppm. [4] D. Mathieu, Sensitivity of Energetic Materials: Theoretical Rela-
tionships to Detonation Performance and Molecular Structure,
Ind. Eng. Chem. Res. 2017, 56, 8191–8201.
[5] C. Darwich, T. M. Klapötke, C. M. Sabaté, 1,2,4-Triazolium-cati-
on-based energetic salts, Chem. Eur. J. 2008, 14, 5756–5771.

Propellants Explos. Pyrotech. 2021, 46, 124–133 © 2020 Wiley-VCH GmbH www.pep.wiley-vch.de 131
Full Paper J. Glorian, K.-T. Han, S. Braun, B. Baschung

[6] M. W. Schmidt, M. Gordon, J. A. Boatz, Triazolium-Based En- unit) volumes, and thermochemical radii, Inorg. Chem. 1999,
ergetic Ionic Liquids, J. Phys. Chem. A 2005, 109, 7285–7295. 38, 3609–3620.
[7] Q. Zhang, J. M. Shreeve, Energetic Ionic Liquids as Explosives [25] H. D. B. Jenkins, D. Tudela, L. Glasser, Lattice potential energy
and Propellant Fuels: A New Journey of Ionic Liquid Chemistry, estimation for complex ionic salts from density measurements,
Chem. Rev. 2014, 114, 10527–10574. Inorg. Chem. 2002, 41, 2364–2367.
[8] R. P. Singh, R. D. Verma, D. T. Meshri, J. M. Shreeve, Energetic [26] H. D. B. Jenkins, L. Glasser, Volume-based thermodynamics: Es-
Nitrogen-Rich Salts, and Ionic Liquids, Angew. Chem. Int. Ed. timations for 2 : 2 salts, Inorg. Chem. 2006, 45, 1754–1756.
2006, 45, 3584–3601; Angew. Chem. 2006, 118, 3664–3682. [27] K. E. Gutowski, R. D. Rogers, D. A. Dixon, Accurate thermo-
[9] M. Suceska, EXPLO5 program, Zagreb, Croatia; 2010. chemical properties for energetic materials applications. I.
[10] N. Fischer, T. M. Klapötke, M. Reymann, J. Stierstorfer, Nitro- Heats of formation of nitrogen-containing heterocycles and
gen-Rich Salts of 1H,1’H-5,5’-Bitetrazole-1,1’-diol: Energetic Ma- energetic precursor molecules from electronic structure theo-
terials with High Thermal Stability, Eur. J. Inorg. Chem. 2013, ry, J. Phys. Chem. A 2006, 110, 11890–11897.
12, 2167–2180. [28] V. Forquet, C. Miro Sabaté, H. Chermette, G. Jacob, E. Labarthe,
[11] T. M. Klapötke, P. C. Schmid, S. Schnell, J. Stierstorfer, Thermal H. Delalu, C. Darwich, Energetic Properties of Rocket Propel-
stabilization of energetic materials by the aromatic nitrogen- lants Evaluated through the Computational Determination of
rich 4,4’,5,5’-tetraamino-3,3’-bi-1,2,4-triazolium cation, J. Mater. Heats of Formation of Nitrogen-Rich Compounds, Chem. Asian
Chem. A 2015, 3, 2658–2668. J. 2016, 11, 730–744.
[12] X. Li, S. Liu, S. Zhang, H. Wu, B. Wang, Q. Yang, Q. Wei, G. Xie, [29] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
S. Chen, S. Gao, A low sensitivity energetic salt based on fur- Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson,
azan derivative and melamine: Synthesis, structure, density H. Nakatsuji, X. Li, M. Caricato, A. V. Marenich, J. Bloino, B. G.
functional theory calculation, and physicochemical property, J. Janesko, R. Gomperts, B. Mennucci, H. P. Hratchian, J. V. Ortiz,
Chem. Eng. Data 2015, 61, 207–212. A. F. Izmaylov, J. L. Sonnenberg, D. Williams-Young, F. Ding, F.
[13] X. Zang, W. Zhu, T. Wei, C. Zhang, H. Xiao, Densities, Heats of Lipparini, F. Egidi, J. Goings, B. Peng, A. Petrone, T. Henderson,
Formation, Energetic Properties, and Thermodynamics of For- D. Ranasinghe, V. G. Zakrzewski, J. Gao, N. Rega, G. Zheng, W.
mation of Energetic Nitrogen-Rich Salts Containing Sub- Liang, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M.
stituted Protonated and Methylated Tetrazole Cations: A Com- Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, K.
putational Study, J. Phys. Chem. C 2010, 114, 13142–13152. Throssell, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M. J.
Bearpark, J. J. Heyd, E. N. Brothers, K. N. Kudin, V. N. Staroverov,
[14] H. Gao, Y. Huang, C. Ye, B. Twamley, J. M. Shreeve, The Syn-
T. A. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. P.
thesis of Di(aminoguanidine) 5-Nitroiminotetrazolate: Some Di-
Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, J. M. Mill-
protic or Monoprotic Acids as Precursors of Energetic Salts,
am, M. Klene, C. Adamo, R. Cammi, J. W. Ochterski, R. L. Martin,
Chem. Eur. J. 2008, 14, 5596–5603.
K. Morokuma, O. Farkas, J. B. Foresman, D. J. Fox, Gaussian 16,
[15] Q H. Lin, Y C. Li, Y Y. Li, Z. Wang, W. Liu, C. Qi, S P. Pang,
Revision A.03. Gaussian, Inc., Wallingford CT, 2016.
Energetic salts based on 1-amino-1,2,3-triazole and 3-methyl-
[30] A. Osmont, L. Catoire, I. Gökalp, V. Yang, Ab initio quantum
1-amino-1,2,3-triazole, J. Mater. Chem. 2012, 22, 666–674.
chemical predictions of enthalpies of formation, heat capaci-
[16] Q H. Lin, Y C. Li, C. Qi, W. Liu, Y. Wang, S P. Pang, Nitrogen-
ties, and entropies of gas-phase energetic compounds, Com-
rich salts based on 5-hydrazino-1H-tetrazole: a new family of
bust. Flame 2007, 151, 262–273.
high-density energetic materials, J. Mater. Chem. A 2013, 1,
[31] T. M. Klapötke, M. Leroux, P. C. Schmid, J. Stierstofer, Energetic
6776–6785. Materials Based on 5,5’-Diamino-4,4’-dinitramino-3,3’-bi-1,2,4-
[17] A. Hammerl, T. M. Klapötke, H. Nöth, M. Warchhold, G. Holl, M. triazole, Chem. Asian J. 2016, 11, 844–851.
Kaiser, U. Ticmanis, [N2H5] + 2[N4C-N = N-CN4]2-: A New High-Ni- [32] T. M. Klapötke, F. Mieskes, J. Stierstorfer, M. Weyrauther, Stud-
trogen High-Energetic Material, Inorg. Chem. 2001, 40, 3570– ies on Energetic Salts Based on (2,4,6-Trinitrophenyl) guani-
3575. dine Propellants, Explos. Pyrotech. 2016, 41, 217–222.
[18] Y. Huang, H. Gao, H. Twamley, J. M. Shreeve, Nitroamino Tri- [33] T. Hermann, T. M. Klapötke, B. Krumm, J. Stierstorfer, Synthesis,
azoles: Nitrogen-Rich Precursors of Stable Energetic Salts, Eur. Characterization, and Properties of Di- and Trinitromethyl-
J. Inorg. Chem. 2008, 2560–2568. 1,2,4-Oxadiazoles and Salts, Asian J. Org. Chem. 2018, 7, 739–
[19] J. C. Galvez-Ruiz, G. Holl, K. Karaghiosoff, T. M. Klapötke, K. 750.
Lohnwitz, P. Mayer, H. Noth, H. K. Polborn, C. Rohbogner, M. [34] L. A. Curtiss, P. C. Redfern, K. Raghavachari, Gaussian-4 theory,
Suter, J. Weigand, Derivatives of 1,5-Diamino-1H-tetrazole: A J. Chem. Phys. 2007, 126, 084108.
New Family of Energetic Heterocyclic-Based Salts, Inorg. Chem. [35] D. Russel, I. Johnson, NIST Computational Chemistry Compar-
2005, 44, 4237–4253. ison, and Benchmark Database. [Online] 2018. Available from:
[20] H. Xue, B. Twamley, J. Shreeve, Energetic salts of substituted http://cccbdb.nist.gov/.
1,2,4-triazolium and tetrazolium 3,5-dinitro-1,2,4-triazolates, J. [36] D. R. Burgess, NIST Chemistry WebBook. [Online] 2018. Avail-
Mater. Chem. 2005, 15, 3459–3465. able from: https://webbook.nist.gov/chemistry/.
[21] K. Karaghiosoff, T. M. Klapötke, P. Mayer, C. M. Sabaté, A. Pen- [37] T. E. Mallouk, G. L. Rosenthal, G. Müller, R. Brusasco, N. Bartlett,
ger, J. M. Welch, Salts of Methylated 5-Aminotetrazoles with Fluoride ion affinities of germanium tetrafluoride and boron
Energetic Anions, Inorg. Chem. 2008, 47, 1007–1019. trifluoride from thermodynamic and structural data for
[22] CHEETAH 2.0. Energetic Materials Center, Lawrence Livermore (SF3)2GeF6, ClO2GeF5, and ClO2BF4, Inorg. Chem. 1984, 23, 3167–
National Laboratory 1998. 3173.
[23] E. F. C. Byrd, B. M. Rice, A Comparison of Methods to Predict [38] M. B. Rice, J. H. Hare, E. F. C. Byrd, Accurate Predictions of Crys-
Solid Phase Heats of Formation of Molecular Energetic Salts, J. tal Densities Using Quantum Mechanical Molecular Volumes, J.
Phys. Chem. A 2009, 113, 345–352. Phys. Chem. A 2007, 111, 10874–10879.
[24] H. D. B. Jenkins, H. K. Roobottom, J. Passmore, L. Glasser, Rela- [39] K. E. Gutowski, R. D. Rogers, D. A. Dixon, Accurate Thermo-
tionships among ionic lattice energies, molecular (formula chemical Properties for Energetic Materials Applications. II.

132 www.pep.wiley-vch.de © 2020 Wiley-VCH GmbH Propellants Explos. Pyrotech. 2021, 46, 124–133
Heat of Formation of Triazole-Based Salts: Prediction and Experimental Validation

Heats of Formation of Imidazolium-, 1,2,4-Triazolium-, and Tet- [42] H. Xue, H. Gao, B. Twamley, J. M. Shreeve, Energetic Salts of 3-
razolium-Based Energetic Salts from Isodesmic and Lattice En- Nitro-1,2,4-triazole-5-one, 5-Nitroaminotetrazole and Other Ni-
ergy Calculations, J. Phys. Chem. B 2007, 111, 4788–4800. tro-Substituted Azoles, Chem. Mater. 2007, 19, 1731–1739
[40] A. Nirwan, V. D. Ghule, Estimation of heats of formation for ni-
trogen-rich cations using G3, G4, and G4 (MP2) theoretical
methods, Theor. Chem. Acc. 2018, 137, 115–124. Manuscript received: July 24, 2020
[41] UN Recommendations on the Transport of Dangerous Goods, Revised manuscript received: September 3, 2020
Manual of tests and Criteria, 4th ed. United Nations. 2003. Version of record online: November 24, 2020

Propellants Explos. Pyrotech. 2021, 46, 124–133 © 2020 Wiley-VCH GmbH www.pep.wiley-vch.de 133