Engineering the Morphology and Particle Size of High Energetic Compounds Using Drop-by-Drop and Drop-to-Drop Solvent-Antisolvent Interaction Methods.

Morphology-controlled precipitation of three powerful organic high energetic compounds (HECs) viz. cyclotrimethylenetrinitramine (RDX), octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX), and 2-methyl-1,3,5-trinitrobenzene (TNT) was achieved by two different processes, namely, drop-by-drop (DBD) and drop-to-drop (DTD) solvent-antisolvent interaction methods. Effect of different experimental parameters on the mean size and morphology of the prepared submicron-sized particles of HECs was investigated thoroughly. The DBD method favors the formation of nanosized particles of RDX and TNT at lower concentrations (5 mM). However, a significant increase in the mean particle size occurred at higher concentrations (25 and 50 mM). Formation of facetted crystals of RDX, HMX, and nanorods of TNT was observed at higher concentrations because of the interaction of crystal facets with the antisolvent. Relatively, smaller sized, spherical particles of RDX and HMX could be prepared through the DTD method even at higher concentrations (25 mM). The DTD method is a continuous process and hence is a facile method for industrial applications. X-ray diffraction and Fourier transform infrared spectroscopy studies revealed that RDX, HMX and TNT were precipitated in their most stable polymorphic forms α, β, and monoclinic, respectively. Differential scanning calorimetry showed that the thermal response of the nano-HECs was similar to the respective raw-HECs. A slight decrease in crystallinity and the melting point was observed because of the decrease in the mean particle size.

Introduction

High energetic compounds (HECs) have significant applications in military munitions as well as in solid propellants. HECs often have the contradictory requirements of properties, that is, having less sensitivity toward external stimuli such as shock, impact, and friction while possessing high energy density. The physical and chemical properties of explosive compounds affect their sensitivity toward external stimuli.[1,2] Accordingly, there are two fundamental ways to reduce the sensitivity of explosives. First, new HEC molecules having lower sensitivity could be designed and synthesized with the help of molecular modeling. The second strategy is to reduce the sensitivity of existing HECs through crystal engineering. Although there have been many attempts to synthesize new chemical entities to meet the stringent requirements of HECs, limited success could only be achieved through the molecular modeling routes. This is the major reason for cyclotrimethylenetrinitramine (RDX) and octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX) being widely used today, even though they were developed many decades ago.[3−5] Powerful explosives such as CL-20 (hexanitrohexaazaisowurtzitane) and octanitrocubane have much higher energy density than RDX and HMX.[4,6] However, their sensitivity to accidental stimuli is a matter of grave concern.

Recently, it has been realized that physical properties such as crystal size, shape, morphology, purity, inclusions, and crystal defects can be altered to improve the performance of existing explosives.[1] With the advent of nanotechnology, it has been realized that particle size reduction is an effective way of reducing sensitivity while enhancing performance. Novel behavior in deflagration to detonation transition was observed with submicron (SM) particles.[7] A few studies have indicated that the particle size of explosives has an influence on impact sensitivity and maximum energy output from a detonation.[8] When explosives were nanosized, the crystal defects, voids, and solvent inclusion also become less, resulting in lower sensitivity.[9,10] Many reports on the relationship between the sensitivity and crystal size showed that smaller explosive particles have less sensitivity, and this effect is still valid for the sub-micrometer scale RDX particles.[11,12] The shape of the particles also significantly influences the sensitivity. For example, RDX particles having a spheroidal shape with a smoothened surface have less shock sensitivity than facetted particles.[13]

Akkbarzade et al. calculated inter- and intramolecular energies and sublimation enthalpies per molecule of HMX crystals of different sizes through molecular dynamic simulation to investigate the stability of nanoscale high energetic compounds (HECs). Results revealed that sublimation enthalpies increased with the increase of the particle size.[14] Around 10 °C decrease in thermal decomposition temperature of nano-TATB was observed when compared with micro TATB.[15] Wang et al. have reduced the particle size of RDX successfully from 87 μm to 270 nm by the milling method and analyzed the factors affecting the safety of explosives by combining Khasanov’s and Merzhanov’s models to establish the relationship between the mean particle size, critical temperature, and critical size of hot spots.[16] They concluded that the mean particle size does not affect sensitivity, unless the mean particle size is lower than 400 nm. Their study also revealed that nanoscale level materials had lower sensitivity than the raw materials because of the small hot-spot size and high critical hot-spot temperature.[16] Guo et al. prepared CL20 nanoparticles with a particle size of 200 nm by utilizing ball-milling technology.[17] Liu et al. used bidirectional grinding mill technology to convert micron-sized particles of RDX, HMX, and CL-20 to nanosized particles.[9] All the nanosized particles showed lower friction and impact sensitivities than the bulk-sized particles.

Plastic-bonded explosives which are prepared utilizing nano-HEMs also showed decrease in shock sensitivity and enhancement in detonation velocities.[9] Han et al. have prepared nano-HMX through mechanical trituration process and utilized them in composite modified double base propellant containing RDX. The nano-HMX showed lower impact and friction sensitivities compared with raw HMX while the performance of propellant improved 30% compared to raw-HMX.[18] Ma et al. prepared TATB-coated HMX crystals of size ranging from 10 to 25 μm through spray drying process using Estane 5703 as a binder. The prepared core–shell crystals showed improved performance of explosion and superior mechanical properties.[19] Compared with core–shell crystals, HMX and TATB nanoparticles showed relatively poor impact and friction sensitivities because of critical hot-spot formation.[19] Thus, preparation of micrometer or sub-micrometer-sized solid particles is of great interest in explosives. In the past several decades, there were many unintended accidents by explosives such as USS Forrestal fire in the year 1967,[20] Camp Doha, Kuwait fire in 1991,[21] Texas city disaster in 1947,[22] explosions in Toulouse, France in 2001,[23] and Ryongchon, South Korea in 2004.[24] It is thus imperative to lower the sensitivity of explosives to accidental stimuli. The above-mentioned studies clearly established the possibility of lowering the sensitivity of conventional explosive compounds while maintaining or augmenting their performance by particle size reduction.[15]

Some of the methods of preparing SM-sized HECs include rapid crystallization from solvent by the addition of antisolvent,[25,26] sol–gel method,[27,28] rapid expansion of supercritical solution,[29,30] mechanical milling,[9,31] and aero-sol method.[32] As HECs are very sensitive to mechanical stimuli, some of these methods such as mechanical milling could be hazardous. Moreover, modulation of the particle shape remains as an open challenge during these processes. Precipitation of the organic solute by the interaction with an antisolvent is the basis for most of the methods for making nanoparticles of organic compounds.[33] Instantaneous precipitation occurs by a rapid desolvation of the hydrophobic active ingredient in the antisolvent medium, when water is used as an antisolvent for organic compounds having poor aqueous solubility.[25,26]

It is possible to prepare nanosized particles of simple organic compounds by finding the suitable solvent–antisolvent pairs and experimental conditions to enhance the rate of supersaturation. We have recently developed the evaporation-assisted solvent–antisolvent interaction (EASAI) method by enhancing the rate of supersaturation.[34,35] Nanoparticles of HECs (RDX and HMX) were prepared by the EASAI method using acetone as solvent at 70 °C.[36] This method is not only limited to HECs but also useful for the nanoformulation of different simple organic compounds. Nanocrystals of various poorly water-soluble drugs could be prepared using the EASAI method to enhance their solubility and hence the bioavailability.[37−39] However, the EASAI method is essentially a batch process, and it is challenging to increase the solid loading in the antisolvent. In the present work, we explored the feasibility of converting the batch process into a continuous process. Additionally, we have also explored the possibility of increasing the solid loading while tuning the particle morphology. The solvent was added to antisolvent in two different ways. Drop-by-drop (DBD) addition of the solvent into an antisolvent helped us to increase the solid loading, whereas drop-to-drop (DTD) mixing of the solvent and antisolvent yielded SM-sized particles continuously in flow. The methods were used to prepare SM-sized particles of HECs such as RDX, HMX, and 2-methyl-1,3,5-trinitrobenzene (TNT), and the results are presented here.

Results and Discussion

Mixing the solvent with the antisolvent using MITOS pressure-pump in both modes adopted, that is, DBD addition and DTD mixing, resulted in the formation of SM-sized particles of the respective HECs. However, the objective of each mode of mixing was different. In the conventional EASAI method, a single spray of the solution onto the surface of the antisolvent was used to yield very fine nanoparticles.[38,40] However, increasing the solid loading remains a challenge to be addressed, as increasing the concentration or the ratio of solvent to antisolvent results in relatively larger particles. In the DBD method, the solution was added DBD to the antisolvent, using microfluidics MITOS pressure-pump at the optimized experimental conditions of the EASAI method. However, in the DTD method, a drop of the solvent was added to a drop of the antisolvent, resulting in a final drop where supersaturation followed by nucleation and growth of nanoparticles occurred. To investigate the suitable interaction condition of the solvent with the antisolvent, we have thoroughly characterized all the prepared nano-HECs. There are many different parameters that affected the mean particle size and shape. To prepare SM-HEC particles with spherical morphology, smaller particle size, and narrow size distribution, we had investigated the effect of different experimental parameters systematically. The mean size of the particles that was formed under each condition was monitored using dynamic light scattering (DLS). The effect of different parameters is discussed in the following sections.

Effect of Concentration

A series of experiments were performed by using solutions of different concentrations (5, 25, and 50 mM) to investigate the effect of concentration on the particle size and morphology. The particle size distribution profiles as per the DLS data are given in Figures S1 and S2 (the Supporting Information), respectively, for DBD and DTD methods. From Table , it is clear that the mean particle size of the SM-HEC particles increase significantly as the concentration of HECs in solution increases in the DBD method. Such an increase in the size of the particles with concentration was observed in our previous studies also.[35] This can be explained based on continued growth of the nuclei because of the availability of enough growth species after the nucleation at high concentrations. However, there was only relatively lesser increase in the mean particle size with increasing solution concentration using the DTD mixing method (Table ). This may be due to the confinement of the particle growth in the combined drop whose size is a constant, irrespective of the concentration. Moreover, the solvent to antisolvent interfacial area also remains the same in DBD method leading to the formation of more nuclei with increasing concentration. Hence, relatively lesser increase in mean particle size was observed in the DTD method compared with the DBD method where chance of heterogeneous nucleation was also there due to the presence of particles that are grown from the initially added drops.

Table 1

Variation in the Mean Size of the Particles of HECs That Were Prepared Using the DBD Addition Method from Acetone Solutions of Different Concentrationsa

	mean particle size (d/μm)
	5 mM		25 mM		50 mM
name of HEC	DLS	FESEM	DLS	FESEM	DLS	FESEM
RDX	1.11 ± 0.10	732 ± 34 nm	4.12 ± 1.00	8.40 ± 1.50	5.38 ± 1.64	11.70 ± 2.10
HMX	5.11 ± 0.90	5.30 ± 1.19	5.81 ± 1.28	6.80 ± 1.80	6.20 ± 1.17	9.4 ± 1.90
TNT	114 ± 07 nm	45 ± 10 nm	1.24 ± 0.01	900 ± 45 nm	5.46 ± 0.86	NA

25 mL of antisolvent, 300 mb pressure of solution, antisolvent temperature 70 °C, duration of addition 300 s, rate of magnetic stirring 1200 rpm.

Table 2

Variation in the Mean Size of the Particles of HECs That Were Prepared Using the DTD Addition Method from Acetone Solutions of Two Different Concentrationsa

	mean particle size (d/nm)
	5 mM		25 mM
name of the HEC	DLS	FESEM	DLS	FESEM
SM-RDX	450 ± 80	270 ± 56	623 ± 81	320 ± 27
SM-HMX	519 ± 64	280 ± 38	806 ± 60	335 ± 41
SM-TNT	1.73 ± 0.09 μm	Sb600 + Ss450, Wj750 nm	1.97 ± 0.10 μm	Sb700 + Ss500, Wj800 nm

100 mb solvent, 600 mb antisolvent, Sb: size of the bigger particles; Ss: size of the smaller particles; Wj: width of junction.

Representative field emission scanning electron microscopy (FESEM) images of the HEC particles that were prepared using DBD method from solutions of various concentrations are presented in Figure . Change in the size as well as the shape of the HEC particles with the change in experimental conditions is evident from Figure . RDX and TNT nanoparticles having a spherical morphology were observed when the concentration of their respective solutions was 5 mM while facetted crystals of relatively larger particles of HMX were observed at the same concentration. At higher concentrations (25 and 50 mM), RDX particles also got precipitated as facetted crystals like HMX, whereas TNT particles remained spherical even at 25 mM, and then, rod-shaped crystals were formed at 50 mM. Thus, the mean particle size and shape also depends on the organic compound being crystallized.

Figure 1

FESEM images of the SM particles of HECs that were prepared by DBD solvent–antisolvent interaction using different concentrations.

The DTD method seems to be unique, as monodispersed particles of the HECs having spherical morphology could be prepared using this method as shown in Figure . RDX and HMX yielded well-dispersed spherical particles, whereas TNT yielded twinned spherical particles. Formation of spherical particles clearly indicates the confinement of nanoparticle growth within the combined droplets of the solution and antisolvent. The twinned spherical morphology of TNT particles points toward the unique progression of diffusion of the solvent and antisolvent from the interface of each drops. Variation in morphology of the particles of different HECs under similar experimental conditions clearly point out the effect of solubility, crystal structure, and precipitation kinetics on the morphology.

Figure 2

FESEM images of the SM particles of HECs that were prepared by DTD solvent–antisolvent interaction method using different concentrations.

Controlling the shape of particles in the crystallization process is still a challenge because the shape of the particles not only depends on internal defects but also on the external particle growth conditions. Experimental conditions such as the crystallization method, temperature, supersaturation, solvent, and additives are known to affect particle morphologies.[41−43] Computational simulations were applied to predict the particle morphology using growth mechanism such as Bravais, Friedel, Donnay, and Harker rules,[44] Monte Carlo simulation,[45] attached energy model and its modification,[46] the occupancy model,[47] the spiral growth model,[48] and two-dimensional nucleation model.[49]

Recently, Liu et al. studied the effect of the solvent (acetone) at the five growth faces of RDX particles through computational simulation using molecular dynamics method along with quantum chemistry calculations.[50] It was reported that morphology of RDX particles is dominated by five growth faces viz. (111), (200), (020), (002), and (210). The binding affinity of the solvent (acetone) varies with the crystal faces in the following order: (002) ≈ (210) > (200) > (020) > (111). These results also showed good agreement with the experimental RDX morphology grown in acetone.[50] Experimental studies conducted by Van der Heijden et al. revealed that the change in morphology of particles of RDX, HMX, and CL-20 was due to the interaction between solvent and growth faces, variations in supersaturation, and presence of impurities.[12] They concluded that with the increase of supersaturation above 40%, particles with random orientations were formed.[12] Thus, in the present case, variation in crystal morphologies of RDX, HMX, and TNT just by varying their concentrations in acetone, under identical conditions may be due to the interaction between the growing particles and the medium. In the DBD method, there is a possibility of heterogeneous nucleation taking place. This could also affect the morphology when compared with the DTD method as well as the EASAI method. The nucleation and growth of nuclei in DBD, DTD, and EASAI are shown in Scheme . Homogeneous nucleation led to the formation of spherical RDX particles at the same concentration (25 mM) using the DTD method as well as EASAI method.[35]

Scheme 1

Schematic Diagram Showing the Comparison of Nucleation and Growth during EASAI, DBD, and DTD Methods of Precipitation of the HECs

Solubility is another major parameter affecting the final morphology of particles of HECs during crystallization in the DBD method. The solubility of HECs in water is the deciding factor for the supersaturation rate. Solubility of HECs in water at 20 °C is in the following order: TNT (94.0–100.7 mg/L) > RDX (35.1–38.9 mg/mL) > HMX (3.68–3.97 mg/mL).[51] However, there is a chance of enhancement in solubility of the solute in the presence of a mixture of solvents. Kim, D.-Y. and Kim, K.-J. studied the solubility of RDX in binary solvent mixture of acetone and water.[52] Compared with mixtures of other solvents with water, RDX showed the lowest solubility in the acetone and water mixture. The solubility increased with increasing temperature and decreased with the increasing water content in the mixture.[52] At fixed concentration of HECs, that is, 5 mM, the mean particle size of HECs also followed the following trend: SM-HMX > SM-RDX > SM-TNT. Thus, it is clear that, at fixed concentration of compounds, higher solubility of TNT in the antisolvent resulted in the formation of smaller sized particles with a spherical morphology.

In addition, with the increase of concentration, an increase in mean particle size and formation of facetted particles were observed for RDX and HMX. However, for TNT, as the concentration increases from 5 to 25 mM, only a slight increase of the mean particle size was observed without any change in the morphology of its particles. However, at very high concentration, that is 50 mM, very long rod-shaped particles of TNT were formed. The variation in the morphology of the HECs at high concentration can be explained based on theoretical and experimental studies reported so far on the effect of solvent and interaction of solvent with particle growth faces.[12,50] It is reported that stronger interaction suppresses the growth of the respective face, whereas the faces which are having poor interactions will grow well. In the DBD method, as well as in the standard EASAI method, acetone gets quickly evaporated from the surface of the hot antisolvent. However, very small spherical RDX and HMX nanoparticles were formed at the same concentration (25 mM) using EASAI method, whereas large facetted particles were obtained using the DBD method.[35] This is clearly due to the heterogeneous nucleation and growth in the DBD method when compared with the EASAI method.

From Figures and 2, it is clear that the DTD solvent–antisolvent interaction method is suitable for the preparation of SM-RDX and SM-HMX particles with a spherical morphology with the mean particle size below 500 nm. It was interesting to compare the data of raw-HECs with SM-HEC particles prepared through the DTD solvent–antisolvent interaction method. We had characterized it thoroughly and discussed in following sections.

Apart from tuning the particle size and shape of the crystals of HECs, it is also critically important to selectively precipitate the preferred polymorphic form of the HEC. This is because, it is well established that the HECs exist in various polymorphic forms and their sensitivity and energetic performance varies with the change in polymorphic form. For example, α-form of RDX and β-form of HMX are known to be the most stable polymorphic forms and hence are preferred for application in munitions. Moreover, solvent inclusion in the crystals is a possibility while the particles are prepared using solvent–antisolvent interaction method. Hence, detailed characterization of the prepared particles was undertaken to ascertain the crystal structure, chemical compositions, and thermal properties in comparison with the raw-particles of the HECs.

Fourier Transform Infrared

Fourier transform infrared (FTIR) spectra of raw-HECs along with the prepared SM-HECs are shown in Figure a. It is clear from Figure a that the SM-HECs have IR bands similar to the raw-HECs. The major bands for the RDX samples were 1592 cm–1 (νs NO2), 1270 cm–1 (νs NO2 + ν N–N), 1039 cm–1 (ring stretching bands), 945 and 783 cm–1 (δ NO2 and γ NO2), and 604 cm–1 (τ + γ NO2).[35] The major IR bands for HMX samples were assigned as follows: 1564 cm–1 (νs NO2), 1145 cm–1 (νs NO2, ν ring), 964 and 946 cm–1 (ring stretching bands), 830 cm–1 and 761 of (δ and γ NO2), and 625 and 600 cm–1 (t + γ NO2).[53] The major IR bands for TNT were observed at 3100, 750 cm–1 corresponding to C–H stretch vibrations of aromatic ring and C–N–O bending. The peaks at 1355 and 1540 cm–1 correspond to C–NO2 symmetric and asymmetric stretching vibrations, respectively. The presence of characteristic peaks at 736 and 754 cm–1 in Figure a confirmed the formation of the α-polymorphic form of RDX.[54] The absence of any peak in the range of 700–750 cm–1 confirmed the formation of β-polymorph of HMX.[55]

Figure 3

(a) FTIR spectra and (b) XRD patterns of the prepared SM-HEC particles along with the corresponding raw-HECs.

Powder X-ray Diffraction (XRD)

RDX is well known to have at least five different polymorphs: α, β, γ, δ, and ε. The α-form (orthorhombic, a = 1.3182, b = 1.1574, and c = 1.0709 nm) is the most stable one at room temperature.[56,57] HMX may exist as four different polymorphs: α (orthorhombic), β (monoclinic), γ (monoclinic), and δ (hexagonal) phase. Among these, the β-HMX (monoclinic, a = 0.65370, b = 1.10540 and c = 0.87018 nm) is highly dense and is the most stable one at room temperature.[58,59] The TNT has two polymorphs, monoclinic and orthorhombic forms. The monoclinic form is the most stable form of TNT at room temperature.[60]

X-ray diffraction (XRD) patterns for the SM-HECs and the respective raw-HEC are shown in Figure b. The XRD pattern of raw-HECs matched well with the SM-HECs. This clearly indicates that HEC particles made by DBD and DTD solvent–antisolvent interaction methods are also crystalline. The XRD patterns of RDX, HMX, and TNT samples showed similar patterns corresponding to α-RDX, β-HMX, and monoclinic TNT, respectively.[57,61,62] The peaks at 2θ for the RDX samples at 13.1°, 13.4°, 15.4°, 17.4°, 17.8°, 20.4°, 22.0°, 25.4°, and 29.3° confirmed that the RDX samples were in α-polymorphic form.[35] XRD peaks for the HMX samples at 2θ, 14.7°, 16.0°, 23.0°, 26.1°, 29.6°, and 31.9°, confirmed that the HMX samples were in β-polymorphic form.[53] Monoclinic form of TNT was confirmed from the diffraction peaks at 2θ: 17.7°, 25.2°, 26.6°, 29.2°, and 32.5°.[60] However, the relative intensities of various peaks in XRD patterns were different. Particle size reduction could be the reason for the observed variation in XRD pattern.[63] Variation in the relative intensities of different peaks in the XRD can also be due to a preferred orientation of the particles when preparing the samples for XRD. For instance, in case of the rod-shaped TNT crystals, one could expect an alignment of the particles, giving rise to a preferred orientation. This generally results in several strong peaks in the XRD, whereas other peaks that should normally be present are much smaller or even absent.

Differential Scanning Calorimetry

Differential scanning calorimetry (DSC) thermal curves for raw and SM-HECs are shown in Figure , and the observed thermal phenomena are summarized in Table . The overall thermal response of the SM-HECs was very much similar to the raw-HECs.[3,5] RDX melts initially and then decompose to give gaseous products. An endothermic peak corresponding to melting of RDX was observed at around 204 °C.[64] A typical endothermic peak corresponding to solid-state phase transition from β to α phase (monoclinic to hexagonal) was observed in HMX samples.[3,66−70] This was followed by an exothermic thermal decomposition after 270 °C.[65,68] An endothermic peak corresponding to melting of TNT was observed at around 81 °C followed by another endotherm around 230 °C corresponding to evaporation. Thermal behavior of the TNT matched well with literature reports.[60,69] The SM-RDX showed a considerable decrease in heat of decomposition compared with raw-RDX, whereas the heats of decomposition for SM-HMX and raw-HMX were almost similar. This is because the heat of decomposition of energetic compounds such as RDX is an intricate interplay of various thermal processes such as melting, evaporation, and thermolysis in gaseous and condensed phases.[3,70] The heat of decomposition of RDX can be as high as 2250 J/g and will vary depending upon the experimental conditions as well as the nature of samples, even though the chemical structure is same.[70] In the present study, the SM-RDX particles melt at slightly lower temperature (205.3 °C) than bulk-sized raw-RDX particles (206.1 J/g). The heat of melting for the SM-RDX was higher (145 J/g) than the raw-RDX particles (113.7 J/g). The apparent heat of decomposition of RDX is the net value of heats of thermolysis (exothermic) and evaporation (endothermic). Thus, the observed lower heat of decomposition of SM-RDX was due to the relative higher contribution from evaporation along with thermolysis when compared to raw-RDX.

Figure 4

DSC thermal curves for SM-HECs along with raw-HECs.

Table 3

Phonological DSC Data of Nano-HECs Along With Raw-HECsa

	raw-RDX	SM-RDX	raw-HMX	SM-HMX	raw-TNT	SM-TNT
Tendo/°C	206.1	205.3			82.5	82.0
ΔHendo/(J/g)	113.7	145.0			67.8	98.4
Texo/°C	232.5	228.8	277.6	277.8
ΔHexo/(J/g)	931.6	602.8	1331.2	1280.8
Tpt/°C			190.9	189.0

Tendo: endothermic peak temperature in °C; Texo: exothermic peak temperature in °C; ΔHendo: endothermic peak enthalpy in J/g; ΔHexo: exothermic peak enthalpy in J/g; Tpt: phase transition temperature for HMX.

The present DTD method for preparing SM particles of high energetic particles is versatile and unique as it is one of the rarest continuous methods for the purpose. A comparison of the various methods that are reported in the literature for the preparation of SM particles of RDX is given in Table T1 (the Supporting Information). The supercritical processing methods have so far achieved only limited success in commercial application.[71] The DTD method stands out because of its ability to prepare monodispersed particles with a spherical morphology.

Conclusions

We have successfully prepared the SM particles of HECs through DBD and DTD solvent–antisolvent interaction methods. It was found that the DTD antisolvent method is suitable for the preparation of SM-RDX and SM-HMX with controlling the size below 500 nm and spherical morphology with uniform size distribution. A comparison of different methods of performing solvent–antisolvent precipitation clearly confirmed that nucleation and growth can be manipulated to control the particle size and shape of HEC particles. Homogeneous nucleation with arrested growth yields smaller spherical particles, whereas heterogeneous nucleation and uncontrolled growth tends to yield larger particles. The prepared SM-sized particles of RDX and HMX were having spherical morphology with a mean particle size below 500 nm. However, TNT particles were formed in different morphologies such as spherical, rod, and dumbbell at different experimental conditions. The prepared SM-HECs were formed in their most stable polymorphic forms. SM-RDX, SM-HMX, and SM-TNT were in α, β, and monoclinic forms respectively. In conclusion, the DTD reprecipitation method can be used to continuously prepare particles of HECs having a mean particle size below 500 nm with a spherical morphology without affecting the chemical and crystal nature of the HECs.

Materials and Methods

Materials

Powerful HECs, RDX, HMX, and TNT were prepared at TBRL, Chandigarh.[72] Acetone was purchased from Merck India and used as received. Ultrapure water (18.2 MΩ cm) from a double-stage water purifier (ELGA PURELAB Option-R7) was used throughout. A syringe filter with 0.22 μm pore size was purchased from Millipore, USA. The microfluidics experimental setup with accessories were purchased from Syrris Ltd., India. A magnetic stirrer with a hot plate was purchased from IKA Pvt. Ltd., India.

Preparation of SM-HECs

Stock solutions of HECs (50 mM) in acetone were prepared by dissolving the accurately weighted amount of the respective HECs. Acetone was used as the solvent because of good solubility of RDX, HMX, and TNT and good miscibility with the antisolvent (water). Water was used as the antisolvent because of poor solubility of HECs in it. The solutions of lower concentration (25 and 5 mM) were prepared through dilution of the stock solution (50 mM). The solution of HECs was always filtered before use, using a syringe filter of pore size 0.22 μm to ensure that no particle was present in it. Simple experimental setups of both the processes are presented in Figure . The acetone solution of the HECs and the antisolvent were filled in small glass vials (10 mL) of two different MITOS pressure pumps and sealed them with a lid. The opening in the lid allows the liquids to be pumped out through a fluorinated ethylene propylene (FEP) tube (0.8 mm o.d. and 0.1 mm i.d.) as shown above in Figure .

Figure 5

Schematic experimental setup of DBD (left) and DTD (right) solvent–antisolvent interaction methods.

In the DBD solvent–antisolvent interaction method, the antisolvent (i.e., water, 25 mL) was taken in a conical flask (100 mL) and heated on a heating plate to 70 °C while being magnetically stirred for ensuring uniform heating. The solution of HECs was added DBD into the hot antisolvent, using the microfluidic MITOS pressure-pump by applying pressure (300 mb) under magnetic stirring of 1200 rpm. The solution was added DBD continuously up to 300 s into the antisolvent under stirring. On the contrary, in the DTD solvent–antisolvent interaction method, the outer end of FEP tubes of both the pumps were kept very close to each other in such a way that each drop of the solution of HECs and antisolvent get mixed at the tube opening as showed in Figure . The two drops of solvent and antisolvent get mixed and merged into a single droplet while dropping to the collection vessel (conical flask, 100 mL). The particle formation takes place in the merged droplets and can be continuously collected. The pressure of solvent and antisolvent were 100 and 600 mb, respectively. The particles were collected by centrifugation at 22 000 rpm. The collected samples were used for further characterization.

Mean Particle Size and Morphology

Primary characterization of the mean particle size (z-average diameter, d/nm) and polydispersity index of the precipitated particles was done using photon correlation spectroscopy or DLS technique (Zetasizer Nano ZS, Malvern Instrument Ltd., UK) at 25 °C. The instrument had a 4 mW He–Ne laser operating at a wavelength of 633 nm and incorporates noninvasive backscatter optics. All measurements were made at a detection angle of 173°. DLS measurement of samples were done using disposable polystyrene cuvettes of 3 mL by taking the as-prepared particle suspension in the antisolvent (water, 2 mL). Each measurement was done with an incubation time of 120 s. Each experiment data was recorded five times, and the mean was reported.

Accurate particle size and morphology of the prepared particles were observed using field emission scanning electron microscopy (FESEM, FEI, NOVA NANOSEM 450 model). The suspension of HEC particles in water was drop-coated on a clean silica wafer and air-dried for 48 h. The dried sample containing silica wafer was kept on a clean aluminum stub that is covered with carbon tape. The sample was subsequently sputter-coated with gold at 20 mA for 2 nm thickness before the FESEM observations. The mean particle size of more than 300 particles from different FESEM images that were taken from different regions of the sample was calculated in each experimental condition using ImageJ software.

FTIR Spectroscopy

The samples were properly grounded with KBr powder and then pressed to obtain suitably sized pellets for recording FTIR spectra. Pure KBr pellet was used for background correction. FTIR spectroscopy was performed using the Agilent Technologies Cary 660 FTIR spectrometer. FTIR spectra of raw and processed HECs were recorded from 4000 to 500 cm–1 frequency range with a resolution of 4 cm–1 and 16 scans.

Powder XRD

The sample was placed on a quartz sample holder and scanned in the 2θ range of 5–90° at a scan rate of 2° min–1 with a step size of 0.02°. The XRD measurements were performed on a SmartLab X-ray diffractometer (Rigaku, Japan) using Cu Kα radiation as an X-ray source (λ = 0.15418 nm) at room temperature. The voltage and current applied were 45 kV and 100 mA, respectively.

Thermogravimetric Analysis (TGA)-DSC

A small amount of the sample (2–3 mg) was taken in a standard alumina pan with an alumina lid with a pin hole at the middle. An empty crucible was used as the reference. Thermogravimetric Analysis (TGA)-DSC analyses were carried out by using a NETZSCH STA 449 F1 Jupiter instrument. The samples were heated from room temperature to 300 °C at a heating rate of 5 °C min–1 under nitrogen atmosphere with a flow rate of 60 mL min–1 protective and 40 mL min–1 purge gases.