Thermal decomposition and combustion of cocrystals of CL-20 and linear nitramines

Abstract

The thermal decomposition and combustion behavior of the bimolecular cocrystals of 2,4,6,8,10,12-hexanitrohexaazaisowurtzitane (CL-20) with linear nitramines 2,4-dinitro-2,4-diazapentane (DNP) and 3,5-dinitro-3,5-diazaheptane (DNG) were studied. Being more volatile than CL-20, the linear nitramines have been shown to evaporate first at heating, leaving CL-20 in the amorphous form. The loss of the crystal lattice is a reason for the increased rate of CL-20 decomposition observed, which is comparable to the decomposition rate of CL-20 in the liquid phase. Different volatilities of linear nitramines and CL-20 caused the heat release zones to be spatially separated in the combustion wave of the co-crystals. Considering the lower energetic characteristics of the linear nitramines as compared to CL-20, they play the role of the diluents in the combustion wave.

---

1. Introduction

Cocrystallization is realized by combining two or more neutral components in a defined molar ratio through non-covalent interactions (e.g. electrostatic interactions, van der Waals forces, π-effects, etc.) to form a unique crystal structure. Cocrystallization technologies are currently being pursued in pharmaceuticals¹ and agrochemicals² to enhance or alter the physico-chemical properties. Cocrystallization is becoming increasingly hot in the field of energetic materials.^3–9 It is hoped, first of all, that cocrystallization engineering can alter such properties of existing energetic molecules in cocrystals as power and safety. The evaluations of the energetic performance and safety of a large range of binary energetic cocrystals of the very powerful explosives like 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12-hexaazaisowurtzitane (CL-20) and octogen (HMX) with the less energetic and much safer explosives like 2,4,6-trinitrotoluene (TNT) and 2,4,6-triamino-1,3,5-trinitrobenzene (TATB) show that the energetic performance is diluted but the safety may be improved after cocrystallization, compared to the more energetic initial components.⁷ An increase in performance may be due to an increase in density, but numerous experiments have shown⁷ that the density of the energetic cocrystals is almost equal to the additive density of its pure components.

The morphology of a crystalline material and the particle sizes have a profound influence on its processability and hence its real utility in energetic formulations. Cocrystallization offers a mechanism by which the crystal morphology of a given compound can be changed to give more desirable properties.¹⁰

The interaction between the components in the cocrystals seriously alters the melting point of the components and perhaps may also affect their thermal decomposition. The thermal decomposition processes in TNT/CL-20 cocrystal were studied theoretically with help of reactive molecular dynamics simulations using the ReaxFF reactive force field.¹¹ It was found that the TNT/CL-20 cocrystal releases energy at a lower rate with a higher energy barrier compared to pure crystal of CL-20 and simple TNT/CL-20 mixture, but at a higher reaction rate than pure crystal of TNT. Authors of ref. 11 suggested that the formation of carbon clusters arising from TNT, a carbon-rich molecule, can confine the active fragments, formed due to CL-20 decomposition, preventing their participation in the secondary reactions.

Unfortunately, the experimental confirmation of theoretical predictions is missing: only the TNT/CL-20 cocrystal melting point is given in the literature.^3,12 Experimental studies of a related TATB/CL-20 cocrystal also did not confirm the expected increase in thermal stability: the DSC curve of the TATB/CL-20 cocrystal reveals a strong exothermic peak at 231.8 °C, attributed to the decomposition event of the cocrystal, which is distinctly lower than that of pure CL-20 (245.6 °C) or TATB (380.9 °C).¹³

Despite the active investigation of various physicochemical properties of cocrystals, no papers on combustion of cocrystals have been published until now. At the same time, in cocrystals, the components are mixed at the molecular level, thus presenting excellent models to study the combustion behavior of ideal mixtures.

In this connection, the goal of this work was to study the combustion behavior and thermal decomposition of cocrystals.

Among the energetic cocrystals, a series of cocrystals containing the high energy material CL-20 are of most interest.^4–9,14,15 Cocrystals of CL-20 with linear nitramines 2,4-dinitro-2,4-diazapentane (DNP) and 3,5-dinitro-3,5-diazaheptane (DNG) have been chosen as the objects of study.⁹ CL-20 forms molecular complex with DNP in a ratio of 2 : 1 (CLD, Fig. 1) and with DNG in the ratio 1 : 1 (CLD-2, Fig. 2).

Fig. 1 The composition of the CLD.

Fig. 2 The composition of the CLD-2.

The content of less energetic linear nitramines is not large in the complexes, which allows expecting that explosive characteristics of the molecular complexes will be only slightly different from the relative CL-20 characteristics. According to calculations, the adiabatic combustion temperature (T_f) of pure CL-20 at a pressure of 10 MPa amounts to 3345 °C. It is only 100 °C higher than T_f for the CLD cocrystal, whereas the temperature drop for the CLD-2 cocrystal (700 °C) is more considerable.

2. Experimental

2.1. Preparation

CLD and CLD-2 cocrystals were synthesized by a method published elsewhere.⁹ Cocrystallization of α-CL-20 with DNP and DNG was performed in a mixture of acetic acid esters using a vaporization method at 40–50 °C. After a portion of the solvent was removed under vacuum, the suspension was cooled to 10–15 °C, and the crystals precipitated were filtered off and washed with cold ethyl acetate.

2.2. Decomposition study

Thermal behavior of the cocrystals over a temperature range of 25–300 °C was studied using differential scanning calorimetry (DSC) and thermogravimetry (TGA) with a thermal analyzer DSC 822e Mettler Toledo. The cocrystals (1–3 mg) were placed in hermetic aluminum pans with a pinhole and scanned with a heating rate ranging from 5 to 20 °C min⁻¹ (only 10 °C min⁻¹ in the case of TGA) under nitrogen flow.

Manometric experiments were carried out in thin-walled glass manometers of the compensation type (the glass Bourdon gauge) at 200 °C. A sample of CLD-2 (∼10 mg) was loaded in a glass manometer of 10 cm³ volume. The device was evacuated to a pressure of 10⁻² mmHg, then it was sealed and immersed in a thermostat with the Wood's metal alloy. The temperature in the thermostat was maintained with an accuracy of ±0.5 °C. The pressure of gases evolved in the experiments (the accuracy of pressure measurements was ±1 mmHg) was converted to the gas volume at standard ambient temperature and pressure, and the rate of gas evolution was then evaluated. The rate constant was calculated as a ratio of the initial gas evolution rate to the final volume of gases.

2.3. Vapor pressure measurements

The glass Bourdon gauges were used to measure CLD-2 vapor pressures at different temperatures. A sample of 10 mg was loaded into the vial. The vapor pressure over CLD-2 sample at a given temperature was fixed as the gauge reading after 5 minutes of warming-up the vial in the thermostat assuming that no perceptible decomposition had time to take place. Extracting the vial from the thermostat and cooling it to the room temperature allowed assessment of whether the decomposition took place. Usually the maximum remaining pressure in the vial after cooling did not exceed 2 mmHg, while the full decomposition of the sample would result in 200–300 mmHg pressure of cooled gases.

2.4. Combustion study

Burning rates of CLD and CLD-2 cocrystals were measured in a constant-pressure window bomb of 1.5 liter volume in the 0.1–10 MPa pressure interval under nitrogen.

Samples to test were prepared as pressed cylinders of cocrystals with 3–4 mm height confined in transparent acrylic tubes of 4 mm i.d. and 6 mm o.d. Prior to pressing, the material was carefully milled in order to produce samples with a minimum possible pore size, thus minimizing the possibility of flame propagation between particles.

The pressed samples of cocrystals were ignited with help of electrically heated hot wire placed on the top of the sample. Burning propagated from the top to bottom in the stationary-state (laminar) regime. The combustion process was recorded with a video camera. The burning rate was determined by measuring the position of the flame front over the time.

The density of the pressed charges of CLD and CLD-2 cocrystals were 1.74 and 1.64 g cm⁻³, respectively. In the CLD cocrystal, both the CL-20 molecules exist in the γ-conformation.⁹ The known density of the γ-polymorph is 1.92 g cm⁻³. Taking into account this value, the density of DNP/CL-20 mixture calculated by the additive scheme of non-interacting substances is 1.84 g cm⁻³. The crystallographic density of CLD is 1.928 g cm⁻³ at 150 K.⁹ When recalculated to the room temperature on the assumption that the cocrystal has the same coefficient of volume expansion as pure CL-20 (1.5 × 10⁻⁴ K⁻¹),⁷ the density of CLD is 1.886 g cm⁻³ which is close to the additive value. Thus, the density of the pressed CLD charges was 0.95–0.92 of the theoretical maximum density.

In the CLD-2 cocrystal, the CL-20 molecule exists in the β-conformation. The crystallographic density of CLD-2 was measured as 1.750 g cm⁻³ at 200 K (ref. 9) or 1.725 g cm⁻³ if recalculated to the room temperature. Densities of CLD-2 pressed charges were around 0.95 of the theoretical maximum density.

Temperature profiles in the combustion wave were measured using fine thermocouples. The thermocouples were welded from 80%W + 20%Re and 95%W + 5%Re wires 25 μm thick followed by rolling in bands 5–7 μm thick. The thermocouple signal was recorded with a Pico ADC 212 digital oscilloscope. The thermocouple was embedded into the center of the sample. The method of sample preparation and temperature profile measurements was described in detail in ref. 16.

Thermodynamic calculations were performed with the computer simulation of chemical equilibrium code “REAL”.¹⁷

3. Results and discussion

3.1. Thermal decomposition

According to the thermogravimetric analysis,¹⁸ both linear nitramines DNP and DNG evaporate in the temperature range of 70–170 °C without any visible signs of decomposition. At the same time, according to thermocouple measurements, CL-20 is a non-volatile substance with high evaporation temperature.¹⁹ It can be expected, therefore, that the linear nitramines will evaporate first during decomposition in open conditions as well as during combustion of the cocrystals.

According to TGA data for the CLD and CLD-2 cocrystals (Fig. 3), the mass loss was 16.2 and 32.5% in the temperature range 170–200 and 170–210 °C, respectively. These mass losses agree well with the content of the linear nitramines in the cocrystals (15.7 and 30.5%). The mass loss of the second stage for both CLD and CLD-2 cocrystals proceeds at similar temperatures 210–250 °C in the form of thermal runaway. The remainder after the second-stage decomposition of CLD and CLD-2 cocrystals was found to be 12.6% and 19.7%, respectively, of the CL-20 content in these cocrystals. These figures are comparable with the residue after decomposition of pure CL-20 (9–17%).²⁰ According to earlier studies,²¹ the remainder represents a polymeric product, which contains melon-like fragments.

Fig. 3 Thermogravimetric curves of cocrystals CLD (1) and CLD-2 (2) at heating rate of 10 °C min⁻¹.

Weak endothermic signals in the area 111–114 °C for CLD (Fig. 4) and in the area 139–145 °C for CLD-2 (Fig. 5) are observed on the DSC thermograms, depending on the heating rate (5–20 °C min⁻¹). These temperatures are significantly higher than the melting points of linear nitramines DNP and DNG (54 and 76 °C,⁹ respectively). This indicates that DNP/DNG and CL-20 molecules form a well-defined cocrystal lattice instead of two different crystalline phases.

Fig. 4 Comparison of DSC curves of CLD cocrystal (1) and CL-20 (2) at heating rate 10 °C min⁻¹. (3) is enlarged fragment of DSC curve of CLD cocrystal.

Fig. 5 DSC curve of CLD-2 cocrystal at heating rate 5 °C min⁻¹ (1) and 10 °C min⁻¹ (2).

The maximum measured heat effects at this stage were 15.9 and 57.2 J per gram of the CLD and CLD-2, or 101.2 and 187.5 J per gram of the net aliphatic nitramines. For DNP, the value obtained practically coincides with the heat of melting of pure nitramine 102.3 J g⁻¹,²² indicating that only DNP has melted, while CL-20 remains in the solid state. The heat of melting of DNG in the cocrystal is also close to the heat of melting of pure DNG 163.9 J g⁻¹,²² the difference (∼7 J per gram of DNG/CL-20 cocrystal) could be attributed to the experimental error.

To prove the DSC results, the melting processes have been visualized by the hot stage microscopy. It was proved that appearance of the liquid phase was observed at temperatures of 112 °C for the DNP/CL-20 cocrystal and at ∼140 °C for the DNG/CL-20 cocrystal, a part of the samples remained in the solid state. TNT/CL-20 cocrystal was shown earlier to exhibit a similar behavior: differential scanning calorimetry revealed that the cocrystal converts at 136 °C to liquid TNT (T_m = 81–82 °C) and solid β-CL-20 (confirmed by Raman spectroscopy).³

Therefore, DNP or DNG/CL-20 cocrystals behave like solvates, which are held together with weak, van der Waals interactions. Upon heating they do not melt uniformly, decomposing into solid CL-20 and liquid linear nitramines DNP/DNG (so called incongruent melting). Despite the formation of relatively stable van der Waals complexes, the reaction enthalpy of their interaction often is small, if any. For example, the reaction enthalpy of the formation of clathrate-type solvates of fullerenes C60 with halogenated alkanes, alkenes, cycloalkanes is equal to 0–3 kJ.²³ Another example is the benzene solvate of spironolactone (1 : 1).²⁴ DSC thermogram of this solvate shows two overlapping endotherms (83.2 and 94.8 °C, total heat effect 82.5 J g⁻¹) which are associated with 17.3% TGA weight loss due to total release of solvent (15.8%). As in the case of CL-20 cocrystals with DNP/DNG, incongruent melting transition of benzene solvate of spironolactone (∼83 °C) is significantly higher than melting point of benzene (5.4 °C) and even overlaps its boiling point (80.1 °C). The total measured heat effect of fusion/evaporation of solvate (82.5 J per gram of the solvate or 521.8 J per gram of pure benzene) coincides well with the total heat of melting and evaporation of pure benzene (162 + 360.9 = 522.9 J g⁻¹). The second component of solvate, spironolactone, after melting and evaporation of benzene remains in the solid state prior to its normal fusion temperature which occurs near 208 °C with thermal effect (48.1 J per g of spironolactone) closest to the heat of parent spironolactone (46.3 J g⁻¹).

Evaporation of DNP is recorded in the DSC thermograms as an endothermic peak at 182 °C, bordering to the exothermic peak of CL-20 decomposition at 214–223 °C depending on the heating rate. The endothermic heat effect (27–29 J g⁻¹ or 173–185 J per gram of DNP) corresponds to only one third of the heat of DNP evaporation;²⁵ the remainder evaporates together with CL-20 decomposition. The maximum measured heat effect of CLD decomposition amounts to 2852 J g⁻¹. Taking into account the proportion of CL-20 and the heat loss due to DNP evaporation, it gives the CL-20 decomposition heat in the cocrystal as 3440 J g⁻¹. This value is in agreement with the value of 3000 ± 200 J g⁻¹ obtained in an earlier similar study of pure CL-20.²⁰

No endothermic peak of DNG evaporation is observed on DSC thermograms for CLD-2. The total exothermic effect of CL-20 decomposition amounts to 2269 J g⁻¹. Taking into account the heat loss due to DNG evaporation, the heat effect recalculated for the net CL-20 will yield a value 3420 J g⁻¹, thus indicating no influence of linear nitramines on the chemistry of CL-20 decomposition.

Comparison of DSC curves for CLD cocrystal and parent CL-20 (Fig. 4) allows making an important observation. The main heat release peak on decomposition of cocrystal is shifted to lower temperatures as compared to the main exotherm of CL-20.

As can be seen from Fig. 4 and 5, the heat release during decomposition of the cocrystals takes place in two stages sometimes. The second stage at 240–245 °C is consistent with the decomposition peak of pure CL-20, but the first peak is observed in a low-temperature range (214–223 °C). It has been proposed in ref. 9 that the first stage of decomposition is due to the decomposition of DNP/DNG and the interaction of the decomposition products with the initial molecules of the components. However, this contradicts the above data on the full evaporation of linear nitramines. Furthermore, the second peak substantially disappears when the heating rate increases and becomes more than 5 degrees per minute (Fig. 5). It is obvious that both peaks are caused by CL-20 decomposition only.

The two-stage CL-20 decomposition process has already been described in TGA experiments (235 and 245 °C in DTA at 10 °C min⁻¹).²⁰ It was noted that the mass loss in the first decomposition step is generally smaller than that of the second step and generally decreases as the heating rate increases.

At high heating rate the first step is not always visible clearly in DSC, but often appears as an adjoining shoulder at the main peak. According to ref. 26, the splitting of the exothermic peak (10-degree shift of the second peak to the high-temperature area) is observed when escape of gases is hindered. At heating rates more than 5 °C, the first peak becomes insignificant. It can be suggested that the first peak is caused by heat released in decomposition of the stressed cage structure of CL-20 and the second one is determined by reactions between intermediate volatile and solid decomposition products. This idea is supported by the observation of low temperatures of CL-20 combustion (∼700 °C) measured at low pressures under conditions of dynamic outflow of volatile decomposition products.^19,27 Taking the specific heat of CL-20 as 1.0 kJ kg⁻¹ K⁻¹ (0.24 cal g⁻¹ K⁻¹),²⁸ the heat effect of reactions in the condensed phase during combustion at low pressures can be estimated as 710 J g⁻¹ (170 cal g⁻¹, 74 kcal mol⁻¹), which is considerably less than the total heat effect in DSC experiments (>3000 J g⁻¹). Incidentally, the value of 74 kcal mol⁻¹ is comparable with the enthalpy of formation of CL-20 (90.2 kcal mol⁻¹).²⁹

At the same time, it should be noted that the nature of the two exotherms observed in the present work on decomposition of CL-20 cocrystals is quite different. Firstly, the temperature difference between the two peaks is significantly more (30 °C). Secondly, as the heating rate increases, the second peak vanishes rather than makes its appearance (Fig. 5).

The rate constants obtained with the help of Kissinger's method³⁰ for the first exotherm on the DSC thermograms at different heating rates (Table 1), are in a good agreement with previously published results for the decomposition of a CL-20 solution in dinitrobenzene³¹ (Fig. 6). Thus, the stability of CL-20 in cocrystals decreases markedly, with the rate constants increased more than 10 times as compared to pure CL-20.^20,31 A reason for this is likely the amorphous form of CL-20 that remains after evaporation of linear nitramines and neglects the influence of the crystal lattice.

Table 1 The rate constants obtained with the help of Kissinger's method for the first exotherm on the DSC thermograms at different heating rates

Heating rate, °C min^−1	CLD		CLD-2
	Tmax, °C	K, s^−1	Tmax, °C	k, s^−1
5	216.2	0.0090	215.3	0.0096
10	222.0	0.0177	220.8	0.0188
15	227.1	0.0260	224.6	0.0278
20	227.9	0.0345	227.1	0.0368

---

Fig. 6 Comparison of the kinetics of CL-20 decomposition in the solid state (1, isothermal conditions,³¹ 2, nonisothermal conditions²⁰) and in the dinitrobenzene solution (3 (ref. 31)) with kinetics of CL-20 decomposition in cocrystals CLD (4, nonisothermal condition, points and line) and CLD-2 (5, nonisothermal condition, points; 6, isothermal condition, point).

The second exotherm on the DSC thermograms is associated with the CL-20 decomposition in the normal crystal lattice. It increases with decreasing heating rate, suggesting that the process of formation of the CL-20 crystal lattice (crystallization) occurs during heating. In support of this conjecture one can adduce results of manometric experiment with CLD-2 carried out in thin-walled glass manometers of the compensation type at 200 °C. A sample of CLD-2 was keeping at this temperature for one hour more to follow its decomposition. The rate constant was calculated as a ratio of the gas evolution rate to a known value of the final volume of evolved gases (620 cm³ g⁻¹). The rate constant obtained appears to be in agreement with kinetic data for CL-20 decomposition in the normal crystal lattice.³¹

3.2. Vapor pressure

Fig. 7 represents the temperature dependencies of vapor pressure for CLD-2 cocrystal and CL-20. Vapor pressure of pure DNG is unknown, but, since the vapor pressure of CLD-2 cocrystal at the same temperature is about 3 orders of magnitude higher than the analogous value for CL-20, it is obvious that the gaseous phase above CLD-2 cocrystal consists mainly of DNG. The cocrystal of salicylamide with 4-acetamidobenzoic acid behaves in a similar way.³² Salicylamide vapor pressure is 3 orders of magnitude higher than the analogous value for 4-acetamidobenzoic acid at the same temperature. The sublimation experiment with this cocrystal has shown that only volatile salicylamide is sublimed.

Fig. 7 Vapor pressure of CLD-2 (1 – data obtained in Bourdon pressure gauges, point and line, 2 – thermocouple data, points) and CL-20 (3 – points and line).¹⁹

Vapor pressure above CLD-2 cocrystal was measured directly in the overall temperature interval 150–200 °C using Bourdon gauge. In this temperature interval, the temperature dependence of vapor pressure can be described as ln P (atm) = 20.74–11 287/T, and the adjusted value of the heat of vaporization is 93.8 kJ mol⁻¹ (22.4 kcal mol⁻¹). Heat of DNG sublimation can be estimated as 125.2 kJ mol⁻¹ (29.9 kcal mol⁻¹), using heat of fusion of pure DNG (31.4 kJ mol⁻¹ or 7.5 kcal mol⁻¹).²² The resulting value is in a good agreement with the values of heats of sublimation of DNP and other related linear nitramines 102.5–125.5 kJ mol⁻¹ (24.5–30 kcal mol⁻¹) determined experimentally in ref. 25. According to the dependence of DNG vapor pressure on temperature, the boiling point of DNG at atmospheric pressure is extrapolated as 270 °C. It may be noted that thermocouple measurements¹⁹ show the CL-20 sublimation temperature 410 °C at atmospheric pressure.

3.3. Combustion

It was shown in ref. 19 that the CL-20 burning rate is determined by the kinetics of decomposition in the condensed phase. In the area of 1 MPa burning rate-pressure dependence of CL-20 has a kink. There are two explanations for the appearance of this kink: 1. CL-20 melts at 1 MPa and thermo-physical parameters in the combustion wave is changed;²⁷ 2. as the pressure increases heat flux from the gas phase also increases and begins to affect the burning rate.¹⁹

Although studies on the combustion of individual linear nitramines were not conducted, the combustion mechanism of a ternary mixture of DNP with two structurally similar nitramines has been investigated.³³ It was found that combustion of the linear nitramines obeyed the model with the leading reaction in the condensed phase.

According to the above results, the volatility of aliphatic nitramines is much higher than volatility of CL-20. This means that the heat release zones of CL-20 and linear nitramines are spatially separated in the combustion wave. Taking into account much lower energetic characteristics of DNP (ΔH⁰_f= −51.5 kJ mol⁻¹,²⁵ T_f = 2030 K at 10 MPa, calculated heat of combustion in inert atmosphere 3580 J g⁻¹) and DNG (ΔH⁰_f = −135 kJ mol⁻¹, T_f = 1425 K at 10 MPa, calculated heat of combustion 2130 J g⁻¹) than energetic characteristics of CL-20 (T_f = 3615 K at 10 MPa, calculated heat of combustion 6820 J g⁻¹) and similar kinetic parameter of decomposition,³⁴ the linear nitramines will act as the diluents.

Indeed, both cocrystals burn at rates lower than the burning rate of CL-20,¹⁹ but faster than linear nitramines³³ (Fig. 8). As in the case of CL-20, the burning rate-pressure dependence of the cocrystals consists of two segments. The first section of the r_b(p) curve (at low pressures) is characterized by a low pressure exponent which, however, changes to a high one when pressure increases above 1–2 MPa.

Fig. 8 Comparison of the burning rates of CLD, CLD-2, CL-20, DNP and ternary mixture TM.³³

As can be seen from Fig. 8, an addition of aliphatic nitramines to CL-20 results in a decrease in the burning rate, which, at high pressures, is not proportional to the amount of additives. Thus, the burning rate of CLD differs slightly from the burning rate of CL-20, whereas CLD-2 demonstrates a three times drop in the burning rate. At lower pressures, a decrease in the burning rate is proportional to the amount of additives, which underlines the leading role of CL-20 in combustion of cocrystals.

The flame structure of the cocrystal at atmospheric pressure was investigated with the help of a thin thermocouple. The temperature growth in the heating wave of an energetic material obeys the exponential law. Such physical processes as melting and evaporation show themselves as visible kinks on temperature profiles. One of the most important characteristic temperatures on the temperature profile is the surface temperature (T_S) which delimits the condensed and gas phases. The surface temperature is usually controlled by evaporation and is dependent on pressure. A comparison of thermocouple profiles of CLD-2 cocrystal and CL-20 recorded under atmospheric pressure shows that the surface temperature on CLD-2 profiles is less due to DNG evaporation (Fig. 9). The DNG boiling point determined with thermocouples is 261 ± 13 °C, which is in a good agreement with the results obtained by the manometric method (Fig. 7). There are typical kinks on profiles of CLD-2 notable at 140 °C, associated with melting of the cocrystal (T_m), and at 410 °C, associated with evaporation of CL-20. Temperature oscillations visible at the profiles of the cocrystal above the surface are due to ejection of condensed CL-20 particles into the gas phase. The profiles become smooth when the temperature reaches the CL-20 sublimation point. Because of different volatilities of the components, there is a specific layer above the burning surface which contains condensed CL-20 particles and linear nitramine vapors.

Fig. 9 Comparison of the temperature profiles in the combustion wave of CLD-2 cocrystal and CL-20 at atmospheric pressure. T_S and T_m is the surface temperature and melting point, respectively.

Since the burning of CL-20 and aliphatic nitramines can be described by the same combustion model with the leading reaction in the condensed phase,^27,35,36 burning rates of the cocrystals have been calculated using this model. The question is at what temperature, evaporation temperature of linear nitramines or of CL-20, reaction proceeds. Calculations for the CLD-2 cocrystal at the surface temperature (boiling point of DNG) give velocities close to the burning rate of aliphatic nitramines, which is far from experimental results. The decomposition depth of the aliphatic nitramines in the condensed phase at the boiling temperature at the experimental burning rates of the cocrystals does not exceed 2%. Therefore, as it may be expected, the aliphatic nitramines act as inert additives, and the burning rate is determined by the kinetics of the heat release of CL-20 at its evaporation temperature, taking into account the expenses for heating and evaporation of linear nitramines.

However, the burning rates calculated from the condensed-phase model agree well with experimental data only at low pressures. Significant differences are observed at pressures above 1 MPa, at which both increasing and decreasing experimental burning rates in comparison with calculated ones are observed (Fig. 10). It was assumed in ref. 19 that a change in the CL-20 combustion law at pressures above 1 MPa is connected with increased contribution of the heat flux from the gas phase. One can assume that the impact of the linear nitramines on the heat flux from the gas phase varies widely not only with their content in the cocrystals, but with their energetic characteristics, which leads to different deviations from the model. Indeed, the thermodynamic calculation shows that adiabatic flame temperature of CLD is only 100 °C below than T_f for the CL-20, whereas the temperature drop for the CLD-2 cocrystal (700 °C) is more considerable. Thus, the study of CL-20 cocrystals combustion clearly shows that the heat flux from the gas phase has a substantial effect on the burning rate of CL-20.

Fig. 10 Comparison of calculated (dashed lines) and experimental (points) burning rates of CLD and CLD-2.

4. Conclusions

Cocrystals of CL-20 and linear nitramines break up into the constituent components when heated to the melting point (incongruent melting). In doing so, only the low-melting component passes to the liquid state, whereas CL-20 remains in the solid state. Upon further heating, the volatile linear nitramines evaporate, leaving CL-20 in the amorphous state. The lack of the crystal lattice results in increasing decomposition rate of CL-20 comparable to the rate of CL-20 decomposition in the liquid phase. Reducing the heating rate and, hence, increasing the residence time of the CL-20 at elevated temperatures leads to crystallization of the amorphous phase and appearance of the second peak of decomposition on DSC thermograms.

Different volatilities of aliphatic nitramines and CL-20 result in the fact that the heat release zone of CL-20 and linear nitramines is spatially separated in the combustion wave of the cocrystals. Taking into account much lower energetic characteristics of DNP and DNG than energetic characteristics of CL-20 and close kinetic parameter of decomposition, the linear nitramines act as the diluents during combustion. At low pressures the combustion of cocrystals of CL-20 and linear nitramines are well described by the model with the leading reaction in the condensed phase. At higher pressures, as in the case of CL-20 combustion, the heat flux from the gas phase has a substantial effect on the burning rate.

Acknowledgements

The authors are grateful to Dr. N.N. Kondakova for carrying out DSC and TGA measurements and Ms. A.V. Burzhava for help in vapour pressure measurements. The authors would like to acknowledge Dr. Ruth Doherty for help in editing and valuable remarks and discussion. The study was funded by RFBR according to the research project No 16-29-01026.

Notes and references

1. (a) M. B. Hickey, M. L. Peterson, L. A. Scoppettuolo, S. L. Morrisette, A. Vetter, H. Guzman, J. F. Remenar, Z. Zhang, M. D. Tawa, S. Haley, M. J. Zaworotko and O. Almarsson, Eur. J. Pharm. Biopharm., 2007, 67, 112–119; (b) N. Qiao, M. Li, W. Schlindwein, N. Malek, A. Davies and G. Trappitt, Int. J. Pharm., 2011, 419(1–2), 1–11; (c) R. Thakuria, A. Delori, W. Jones, M. P. Lipert, L. Roy and N. Rodríguez-Hornedo, Int. J. Pharm., 2013, 453(1), 101–125.
2. B. S. Sekhon, Int. J. Agric. Sci., 2015, 5(3), 472–475.
3. O. Bolton and A. J. Matzger, Angew. Chem., Int. Ed., 2011, 50(38), 8960–8963.
4. (a) D. I. Millar, H. E. Maynard-Casely, D. R. Allan, A. S. Cumming, A. R. Lennie, A. J. Mackay, I. D. H. Oswald, C. C. Tang and C. R. Pulham, CrystEngComm, 2012, 14(10), 3742–3749; (b) K. B. Landenberger and A. J. Matzger, Cryst. Growth Des., 2012, 12, 3603–3609; (c) J. Zhang and J. M. Shreeve, CrystEngComm, 2016, 18, 6124–6133.
5. O. Bolton, L. R. Simke, P. F. Pagoria and A. J. Matzger, Cryst. Growth Des., 2012, 12(9), 4311–4314.
6. Z. W. Yang, H. Z. Li, X. Q. Zhou, C. Zhang, H. Huang, J. Li and F. Nie, Cryst. Growth Des., 2012, 12(11), 5155–5158.
7. C. Y. Zhang, Y. F. Cao, H. Z. Li, Y. Zhou, J. H. Zhou, T. Gao, H. B. Zhang, Z. W. Yang and G. Jiang, CrystEngComm, 2013, 15(19), 4003–4014.
8. B. Gao, D. Wang, J. Zhang, Y. Hu, J. P. Shen, J. Wang, B. Huang, Z. Q. Qiao, H. Huang, N. Nie and G. C. Yang, J. Mater. Chem. A, 2014, 2(47), 19969–19974.
9. T. K. Goncharov, Z. G. Aliev, S. M. Aldoshin, D. V. Dashko, A. A. Vasil’eva, N. I. Shishov and Y. M. Milekhin, Russ. Chem. Bull., 2015, 64(2), 366–374.
10. H. J. Lloyd, C. R. Pulham and R. Doherty, Proc. Conf. “New Trends in Research of Energetic Materials”, 2015, 178–187.
11. D. Guo, Q. An, S. V. Zybin, W. A. Goddard III, F. Huang and B. Tang, J. Mater. Chem. A, 2015, 3(10), 5409–5419.
12. D. Spitzer, B. Risse, F. Schnell, V. Pichot, M. Klaumünzer and M. R. Schaefer, Continuous engineering of nano-cocrystals for medical and energetic applications, Sci. Rep., 2014, 4, 6575,  DOI:10.1038/srep06575.
13. H. Xu, X. Duan, H. Li and C. Pei, RSC Adv., 2015, 5(116), 95764–95770.
14. B. Gao, D. Wang and J. Zhang, et al., J. Mater. Chem. A, 2014, 2(47), 19969–19974.
15. S. M. Aldoshin, Z. G. Aliev, T. K. Goncharov, D. V. Korchagin, Y. M. Milekhin and N. I. Shishov, Russ. Chem. Bull., 2011, 60(7), 1394–1400.
16. A. E. Fogelzang, V. Y. Egorshev, V. P. Sinditskii and M. D. Dutov, Combust. Flame, 1991, 87, 123–135.
17. G. B. Belov, Propellants, Explos., Pyrotech., 1998, 23, 86–89.
18. R. Vijayalakshmi, N. H. Naik, G. M. Gore and A. K. Sikder, J. Energ. Mater., 2015, 33(1), 1–16.
19. (a) V. P. Sinditskii, V. Y. Egorshev, M. V. Berezin, V. V. Serushkin, Y. M. Mileykhin, S. A. Gusev and A. A. Matveev, Zh. Fiz. Khim., 2003, 22(7), 69–74; (b) V. P. Sinditskii, V. Y. Egorshev, V. V. Serushkin, A. I. Levshenkov, M. V. Berezin, S. A. Filatov and S. P. Smirnov, Thermochim. Acta, 2009, 496(1), 1–12.
20. R. Turcotte, M. Vachon, Q. S. Kwok, R. Wang and D. E. Jones, Thermochim. Acta, 2005, 433(1), 105–115.
21. D. G. Patil and T. B. Brill, Combust. Flame, 1993, 92(4), 456–458.
22. D. Spitzer, S. Braun, M. R. Schäfer and F. Ciszek, Propellants, Explos., Pyrotech., 2003, 28(2), 58–64.
23. N. V. Avramenko, M. V. Korobov, A. M. Parfenova, P. A. Dorozhko, N. A. Kiseleva and P. V. Dolgov, J. Therm. Anal. Calorim., 2006, 84(1), 259–262.
24. G. A. Neville, H. D. Becksteadlf and J. D. Cooney, Fresenius. J. Anal. Chem., 1994, 349, 746–750.
25. E. A. Miroshnichenko, T. S. Kon’kova, Y. N. Matyushin and Y. O. Inozemtsev, Russ. Chem. Bull., 2009, 58(10), 2015–2019.
26. T. Gołofit and K. Zyśk, J. Therm. Anal. Calorim., 2015, 119(3), 1931–1939.
27. V. P. Sinditskii, V. Y. Egorshev, V. V. Serushkin, A. I. Levshenkov, M. V. Berezin and S. A. Filatov, Int. J. Energ. Mater. Chem. Propul., 2010, 9(2), 147–192.
28. J. Hommel and J.-F. Trubert, Proc. 33 Inter. Ann. Conf. ICT, FRG, 2002, 10, 1–17.
29. R. L. Simpson, P. A. Urtiew, D. L. Ornellas, G. L. Moody, K. J. Scribner and D. M. Hoffman, Propellants, Explos., Pyrotech., 1997, 22(5), 249–255.
30. H. E. Kissinger, Anal. Chem., 1957, 29(11), 1702–1706.
31. B. L. Korsounskii, V. V. Nedel'ko, N. V. Chukanov, T. S. Larikova and F. Volk, Russ. Chem. Bull., 2000, 49(5), 812–818.
32. A. N. Manin, A. P. Voronin, N. G. Manin, M. V. Vener, A. V. Shishkina, A. S. Lermontov and G. L. Perlovich, J. Phys. Chem. B, 2014, 118(24), 6803–6814.
33. A. S. Ermakov, E. Y. Petrov, D. B. Vinogradov, F. Van Toai, A. P. Denisyuk and D. L. Rusin, Theor. Found. Chem. Eng., 2007, 41(5), 660–667.
34. S. Zeman, Thermochim. Acta, 1999, 333(2), 121–129.
35. Y. B. Zeldovich, Zh. Eksp. Teor. Fiz., 1942, 12(11–12), 498–524.
36. V. P. Sinditskii, V. Y. Egorshev, V. V. Serushkin and S. A. Filatov, Combust., Explos. Shock Waves, 2012, 48(1), 81–99.

---