ε-CL-20 is prepared from raw CL-20 by solvent evaporation and precipitation methods. Experiments were also done using solvent evaporation coupled with in-situ ultrasonication method. Using precipitation method, ε-CL-20 is scaled up to 500 g batch. Raw CL-20 was assigned to $\alpha$ -CL-20. The chemical and polymorphic purity of prepared ε-CL-20 was found to be about 98 per cent and > 95 per cent, respectively. ε-CL-20 was obtained agglomeration free with well defined geometry in comparison with raw CL-20 and its crystal morphology is dominantly bi-pyramidal or lozenge crystal shapes. The obtained mean particle size of prepared ε-CL-20 by solvent evaporation method with and without in-situ ultrasonication and also by precipitation methods is about 30 µm - 40 µm, 150 µm - 200 µm and 150 µm - 300 µm, respectively. The measured true density of prepared ε-CL-20 by precipitation method with 100 g and 500 g batch scale using Helium gas pycnometer was 2.038 g/cm³ and 2.043 g/cm³, respectively. The lower value of calculated void percentage of ε-CL-20 (0.05-0.29%) indicate better crystal quality. Conclusively, prepared ε-CL20 has high true density with less percentage of voids, less total moisture content and free from agglomeration as compared with the starting raw CL-20 material.

Keywords:    ε-CL-20,  solvent evaporation,  precipitation,  ultrasonication,  true density 

2,4,6,8,10,12-Hexanitro-2,4,6,8,10,12-hexaazaisowurtzitane (CL-20) is one of the most potential candidate for explosive and propellant formulations, which exhibits higher oxygen balance, density and heat of formation than the conventional energetic nitramines such as RDX and HMX. It was first synthesized by Nielson¹, et al. and later it’s explosive performance was found to be approximately 14 per cent higher than that of HMX². Due to the presence of different molecular conformations and arrangements in lattice, CL-20 has been reported to exist in five different polymorphic forms. Four polymorphs (ε, β, α, and γ) exist at ambient condition and one (ζ) exists at high pressure^3-5. Due to the high energetic performance, high density, and low sensitivity compared to the other polymorphs, ε-CL-20 is more desirable for use in propellant and weapons systems.^2,3 Hence, the preparation of ε-CL-20 with good crystal quality, desired particle size and crystal habit has been a great deal of interest in the view of high explosive and solid rocket propellant based formulations. Further, these physical characteristics of CL-20 play vital role in the sensitivity of CL-20⁶. Though worldwide researchers have been reported a few crystallization methods^7,8 for the preparation of ε-CL-20, it is essential to optimize suitable crystallization process to prepare as well to improve the quality of CL-20. The preparation of the desirable polymorph is possible by ‘solvent – nonsolvent’ technique which greatly helpful to improve crystal quality as well⁹. According to Foltz⁵, et al., it was shown that the β-form of crystals could be transformed to γ-one and then finally to ε-one by adjusting solvent and temperature for re-crystallization. Correspondingly, the crystal density is also shifted from 1.98 g/cm³ (β-form) to 2.044 g/cm³ (ε-form) as with the structural transformation. Hoffman¹⁰ reported the influence of the synthetic method of CL-20 on the crystal density of CL-20. Using nuclear quadruple resonance spectroscopic technique, Caulder¹¹, et al. studied the crystal quality of ε-CL-20 obtained from different precursors and solvent systems.

In the present study, it is focused to prepare and characterization by both solvent evaporation and precipitation methods. The present evaporation method also relates for obtaining agglomeration free fine ε-CL-20 crystals by in situ use of ultrasound to crystallization solution. The obtained crystals were characterized using HPLC, FTIR, Raman, Powder XRD, SEM and DSC techniques. Particle size measurement of obtained ε-CL-20 was carried out using particle size analyzer and true density by Helium gas pycnometer.

Raw CL-20 sample was obtained from Premier Explosive’s Laboratories, India. All the solvents and allied chemicals used for the processes are of analytical grade with > 99 per cent purity. ε-CL-20 is prepared using both solvent evaporation and precipitation techniques. Raw CL-20 is dissolved in ethyl acetate solvent and then antisolvent, n-heptane is added into the solution to induce the crystallization. For the effective recovery in the drowning-out crystallization, the fraction of raw CL-20, ethyl acetate and n-heptane was applied as about from 1 : 15 : 35, respectively in evaporation method, whereas 1 : 2 : 5 ratio was used in precipitation method. Then, free flowing white shiny crystals of ε-CL-20 were filtered, washed with n-heptane and dried. The crystallization was also repeated using ultrasound of 35 kHz frequency with ultrasonic intensity level 80 per cent to study the effects of ε-CL-20 particle size and its crystal morphology. The solvent evaporation rate of was controlled approximately from 2 ml/min to 4 ml/min in solvent evaporation method. In precipitation method, the stirring speed of CL-20 solution was kept about 150 rpm - 250 rpm and antisolvent, n-heptane addition rate was in the range of 40 ml/min - 50 ml/min. The mean particle size of prepared ε-CL-20 is largely dependent on solvent evaporation rate or antisolvent addition rate. Using both solvent evaporation and precipitation method, and several batches were conducted to establish the reproducibility of the method for preparation of ε-CL-20. ε-CL-20 was finally scaled up to 500 g batch level using precipitation method.

The prepared ε-CL-20 was well characterized using HPLC, vibrational spectroscopic, X-ray powder diffraction and scanning electron microscopic techniques. The IR spectrum (4000 cm - 400 cm^-1) of the ε-CL-20 was recorded using a Thermo Nicolet, USA, make Nicolet 5700 FTIR spectrometer with spectral resolution of 4 cm^-1 and typically 64 scans were coadded to obtain good signal-to-noise ratio. The Raman spectrum (4000 cm - 100 cm^-1) of the ε-CL-20 was recorded using an In via Reflex dispersive Raman spectrometer from Renishaw, UK, with 2 scans at 2 cm^-1 spectral resolution and at 0.1 per cent power filter of 300 mW laser power (Nd YAG 785 nm laser). Powder X-ray diffraction pattern of the ε-CL-20 was recorded using a PW3040/60 X’Pert PRO X-ray diffractometer from PAN analytical, Netherlands. Morphological information of the ε-CL-20 polymorphs ware obtained with a Quanta 200 Scanning Electron Microscope (SEM) from FEI, Netherlands. All the micrographs were recorded in low vacuum mode using a Large-Field Detector (LFD). The accelerating voltage was maintained within minimum required for image development and to prevent any charging up of samples and subsequent damages the particle surface. Particle size analysis of samples was carried out using the Hydro-2000MU particle size analyzer from Malvern Instruments, UK. The analyzer is based on the principle of laser light scattering with working range of particle size from 0.2 µ to 2000 µ. The true density of prepared ε-CL-20 and raw CL-20 has been measured by Quantachrome Helium gas ultrapycnometer 1000.

Table 1 lists the recovery, chemical and polymorphic purity of prepared ε-CL-20 using solvent evaporation without and with in situ ultrasonication and precipitation methods. In most of the batches, the sample recovery of prepared ε-CL-20 is about 85-90%. HPLC analysis showed the chemical purity of the prepared ε-CL-20 was about 98 per cent. Raw CL-20 used for the preparation of ε-CL-20 show higher degree of agglomeration with mean particle size of 185 µm Fig.1(a). The agglomerates consist of irregular crystal geometry, dominated by sharp edged diamond shape crystals. However, reprocessing of raw CL-20 in laboratory by solvent evaporation and precipitation methods show agglomeration free well defined crystals Figs.1(b), 1 (c), and 1(d) as compared with those of raw CL-20. The crystal morphology of prepared ε-CL-20 by solvent evaporation without ultrasonication method show dominantly bi-pyramidal shape as shown in Fig.1(b) with sharp edges.

However, as shown in Fig.1(c), the sharp edge of bi-pyramidal crystal is smoothened increasingly using in-situ ultrasonication. Further, in-situ ultrasonication batch also show few bigger size near spherical particles surrounded by fine size particles. It clearly attribute that in-situ ultasonication greatly helps to promote sharp edge free crystals during crystallization. The crystal morphology of prepared ε-CL-20 by precipitation method show dominantly lozenge shape as shown in Fig. 1(d). The observed crystal habits are well matched with ε-CL-20 crystal morphology as reported in the literature⁹. By conducting several batches, the reproducible mean particle size of prepared ε-CL-20 by solvent evaporation method with, without in situ ultrasonication and precipitation methods is in the range of about 30 µm - 40 µm, 150 µm - 200 µm and 150 µm - 300 µm, respectively as shown in Table 1. The typical particle size distribution of prepared ε-CL-20 using both solvent evaporation methods and precipitation methods along with raw CL-20 is shown in Appendix (Fig. 1S) in the supporting information.

To begin with characterization of prepared ε-CL-20 by solvent evaporation and precipitation methods, we analyzed starting material raw CL-20. X-ray powder diffraction pattern for raw CL-20 and prepared ε-CL-20 is shown in Fig. 2 along with α-CL-20 and ε-CL-20 library standard over the region of 10-45° ¹²,¹³. It can be seen from Figs 2 (a) and 2 (b) that X-ray powder diffraction profile of raw CL-20 is well comparable with the library standard of α-CL-20. The prominent features of raw CL-20 at 2° = 12.0°, 13.6°, 13.8°, 17.4°, 17.9°, 20.1°, 24.9°, 27.4°, 27.9° and 28.7° are essentially those of α-CL-20. Figure 3 show the infrared spectra of raw CL-20 and prepared ε-CL-20 over the region 400 cm^-1 - 4000 cm^-1. While authors have recorded the spectra over the region 400 cm^-1 - 4000 cm^-1, only the regions 1700 cm^-1 – 1500 cm^-1 and 900 cm^-1- 600 cm^-1 are shown, as these are the regions that carry information on the polymorphs of CL-20. Table 2 shows the infrared frequencies of raw CL-20 and prepared ε-CL-20 along with the literature value.

The regions of 1700 cm^-1 – 1500 cm^-1 and 900 cm^-1 - 600 cm^-1 correspond to the NO₂ stretching and mostly bending and deformation (NNO bending, ring deformation, NO out of plane bending and ONO bending) vibrations, respectively. The main infrared spectral features of raw CL-20 occur at 623 cm^-1, 655 cm^-1, 688 cm^-1, 717 cm^-1, 739 cm^-1, 750 cm^-1, 763 cm^-1, 824 cm^-1, 834 cm^-1, 879 cm^-1, 1556 cm^-1, 1606 cm^-1, and 1618 cm^-1 as shown inFig. 3(a), which is well comparable with α-CL-20 as reported in the literature^14,16. We then recorded raw CL-20 and prepared ε-CL-20 spectrum using dispersed Raman spectrometer in the region of 100 cm^-1 - 4000 cm^-1 as shown in Fig. (4). Table 3 shows the Raman frequencies of raw CL-20 and prepared ε-CL-20 along with the literature value. The main Raman spectral features of raw CL-20 occur at 268 cm^-1, 280 cm^-1, 793 cm^-1, 822 cm^-1, 839 cm^-1, and 858 cm^-1 over the regions of 260 cm^-1 – 300 cm^-1 and 740 cm^-1- 880 cm^-1 as shown inFig. 4(a). As observed in X-ray powder diffraction and infrared experimental results, Raman spectrum of raw CL-20 is also well comparable with literature reported α-CL-20 forms^16,17. It confirms that the starting material raw CL-20 corresponds to α-CL-20 polymorphic phase. X-ray powder diffraction profile of prepared ε CL-20 showed the characteristic peaks at about 2θ = 10.7°, 12.6°, 12.8°, 15.7°, 16.3°, 25.8°, and 30.4° as compared those of raw CL-20 as shown in Fig. 2(b). As can be seen from Figs. 2 (c) - 2(d), X-ray powder diffraction features of prepared ε-CL-20 are well matched with previously reported ε-CL-20 form^12,13. The main infrared features of prepared ε-CL-20 are shown in Fig. 3(b). The presence of quartet and equivalent doublet near 750 cm^-1 (738 cm^-1, 744 cm^-1, 750 cm^-1 and 758 cm^-1) and 825 cm^-1 (820 cm^-1and 831 cm^-1) cm^-1, respectively in Fig. 3(b) establish the purity of ε-CL-20.

Further, ε-CL-20 shows a quartet-like near 1600 cm^-1 (1568 cm^-1, 1589 cm^-1, 1606 cm^-1 and 1632 cm^-1) in the region of asymmetric NO₂ stretching vibration. ε-CL-20 shows a distinct single Raman feature at 264 cm^-1 over the region of 260 cm^-1 - 300 cm^-1, corresponds to ring deformation, where it is at 280 cm^-1 for α-CL-20 along with weak feature at 268 cm^-1 as shown in Fig. 4(b). We also note that the β-CL-20 and γ-CL-20 polymorphs are also occur at 284 cm^-1 and 287 cm^-1 and 270 cm^-1, respectively in the region of 260-300 cm^-1as reported in the literature^16,17. Hence, the absence of feature near at 280 cm^-1 in prepared ε-CL-20 clearly indicates the higher polymorphic purity of ε-CL-20, which was assigned to be > 95%. The polymorphic purity of different batch ε-CL-20 by solvent evaporation and precipitation methods shows > 95%. Other Raman spectral features of prepared ε-CL-20 in the region of 740^-1-880 cm^-1occur at 789^-1, 819^-1, 832^-1 and 855 cm^-1 as shown in Fig. 4(b). The thermal behavior of raw CL-20 and prepared ε-CL-20 are recorded in nitrogen atmosphere when heated in crimped Aluminum pan from 30 ºC to 300 ºC at a heating rate of 10 ºC/min. Sample sizes of about 0.2 mg were employed to study decomposition behavior. However, to see solid-solid phase transition peak unambiguously, sample size increased of about 2 mg. The DSC curve in Fig. 5(a) demonstrates an endothermic behavior of raw CL-20 with a peak temperature, T_p= 174 ± 1 ºC. This peak is assigned to solid-solid α → γ phase transition and the peak temperature of solid-solid ε → γ phase transition is observed at 160 ±1 ºC Fig. 5(b), which are well comparable with value in reported literature¹⁸. Later than the solid-solid phase transition, the DSC profile exhibits a single step exothermic thermal decomposition for both raw CL-20 and ε-CL-20. Figure 6(a) show the onset and peak temperature of single step exothermic thermal decomposition behavior of raw CL-20 at T_o= 235 ± 1 ºC and T_p= 253 ± 1 ºC, respectively. Similarly, the onset and peak temperature of exothermic thermal decomposition behavior of prepared ε-CL-20 is recorded at T_o= 234 ± 1 ºC and Tp= 252 ± 1 ºC, respectively as shown in Fig. 6(b).

Table 4. shows true density characteristics of raw CL-20 and prepared ε-CL-20 by precipitation method along with the reported values. The measured true density using Helium gas pycnometer of raw CL-20 with the total moisture content of about 1.1% (weight percent) was 1.987 g/cm³, which is close to the ideal crystal density value of α-CL-20 hydrate (2.001 g/cm³ - 1.981 g/cm³) as reported in the literature⁴. However, the true density of prepared ε-CL-20 with 100 g batch and 500 g batch size by precipitation method were determined to be 2.038 g/cm³ and 2.043 g/cm³, respectively, which is well consistent with the maximum true density of ε-CL-20 crystal suggested by Johnston and Wardle¹⁵. The reported ideal crystal density of ε-CL-20 is 2.044 g/cm³. The void percentage of prepared ε-CL-20 with 100 g batch and 500 g batch size was quantitatively estimated from density difference in compared to ideal crystal density and was calculated to be 0.29% and 0.05%, respectively.

The low void percentage of prepared ε-CL-20 signifies the good crystal quality as reported in the literature¹¹. It is well known that shock sensitivity is reduced with the density growth of material, which suggests our prepared ε-CL-20 may have reduced shock sensitive¹⁹ than those of raw CL-20, which will be subject of our future studies. Further, the total moisture percentage of prepared ε-CL-20 is about ≤ 0.1%, which an essential criteria for formulation of high explosives and propellant compositions.

Preparation of the ε-CL-20 polymorph is carried out from raw CL-20 by solvent evaporation with and without insitu ultrasonication and precipitation methods. Preparation of ε-CL-20 by precipitation method is scaled upto batch size of 500 g. Prepared ε-CL-20 were well characterized using HPLC, vibrational spectroscopic, X-ray powder diffraction, scanning electron microscopic and DSC techniques. Raw CL-20 was assigned to be α-CL-20. The chemical and polymorphic purity of prepared ε-CL-20 was found to be about 98% and >95%, respectively and their obtained crystal morphology is dominantly bi-pyramidal or lozenge crystal shapes. Using solvent evaporation coupled with insitu ultrasonication method, the sharp edge of bipyramidal crystal is increasingly smoothened and also shows few bigger size spherical particles surrounded by fine size particles. The mean particle size of prepared ε-CL-20 by solvent evaporation coupled with insitu ultrasonication method is about 30 µm - 40 µm. The measured true density of prepared ε-CL-20 by precipitation method with 100 g and 500 g batch scale using He gas pycnometer was 2.038 g/cm³ and 2.043 g/cm³, respectively. The prepared ε-CL-20 has high true density with less percentage of voids, less total moisture content and free from agglomeration as compared with the starting raw CL-20.

The authors express gratitude to Director, High Energy Materials Research Laboratory (HEMRL), Pune for publication of the article. Authors are also thankful to Mr S.K. Sahoo and Mr P. Paramasivan from HEMRL, for their help in powder XRD and density characterizations, respectively.

1. Nielsen, A.T. Polycyclic amine chemistry. In Chemistry of energetic materials, edited by G.A. Olah & D.R. Squire. Academic Press, San Diego, US, 1991, pp. 95-124

2. Simpson, R.L.; Urtiew, P.A.; Ornellas, D.L.; Moody, G.L.; Scribner, K.J. & Hoffman, D.M. CL-20 performance exceeds that of HMX and its sensitivity is moderate. Propellants, Explosives, Pyrotechnics, 1997, 22(5), 249-55.

3. Li, J. & Brill, T.M. Kinetics of solid polymorphic phase transition of CL-20. Propellants, Explosives, Pyrotechnics, 2007, 32(4), 326-330.

4. Nielsen, A.T.; Chafin, A.P.; Christian, S.L.; Moore, D.W.; Nadler, M.P.; Nissan, R.A. & Vanderah, D.J. Synthesis of polyazapolycyclic caged polynitramines. Tetrahedron, 1998, 54(39), 11793-11812.

5. Foltz, M.F.; Coon, C.L.; Garcia, F. & Nicholas III, A.L. Thermal stability of the polymorphs of hexanitrohexaazaisowurtzitane, Pt I. Propellants, Explosives, Pyrotechnics, 1994, 19(1), 19-25.

6. Johnson, N.C. CL-20 sensitivity round robin. Indian Head Division Naval Surface Warfare Center, User Report No. MD 20640-5035, September 2003.

7. Hamilton, R.S.; Mancini, V.; Nelson, C. & Dressen, S.Y. High temperature crystallization of 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12-hexaazatetracyclo [5.5.0. 05,903,11]—dodecane. US Patent No. 7288648, 30 October, 2007.

8. Johnston, H.E. & Wardle, R.B. Process of crystallizing 2,4,6,8,10,12-hexanitro-2,4,6,8,10,12,-hexaazatetracyclo [5.5.0.0.5,903,11]-dodecane. US Patent No. 5874574, 23 February, 1999.

9. Mersmann, A. Crystallization technology handbook. Ed. 2. Marcel Dekker, New York, 2001.

10. Hoffman, D.M. Void density distribution in 2,4,6,8,10,12-Hexanitro-2,4,6,8,10,12-hexaazaisowurtzitane (CL-20) prepared under various conditions. Propellants, Explosives, Pyrotechnics, 2003, 28(4), 194-200.

11. Caulder, S.M.; Buess, M.L. & Nock, L.A. An analytical study of the crystal quality of e-hexanitrohexaazaisowurtzitane (CL-20) synthesized using several different crystallization techniques and intermediate precursors. Sci. Tech. Energetic Materials, 2005, 66, 406-409.

12. Engel, W. ICDD Grant-in-Aid. Fraunhofer-Inst. of Chemische Technologie, Pfinztal-Berghausen, Germany, 1999.

13. Engel, W. ICDD Grant-in-Aid. Fraunhofer-Inst. of Chemische Technologie, Pfinztal-Berghausen, Germany, 2000.

14. Holtz, E.V.; Ornellas, D.; Foltz, M.F. & Clarkson, J.E. The solubility of e-CL-20 in selected materials. Propellants, Explosives, Pyrotechnics, 1994, 19(4), 206-212.

15. Wardle, R.B.; Johnston, G.; Hinshaw, J.C.; Braithwaite, P. Synthesis of the cased nitramine HNIW (CL-20). Paper presentated at 27th International Annual Conference of ICT, Place 1996.

16. Kholod,Y.; Okovytyy, S.; Kuramshina, G. & Qasim, M. An analysis of stable forms of CL-20: A DFT study of conformational transitions, infrared and Raman spectra. J. Molecular Structure, 2007, 843(1-3), 14-25.

17. Gorb, L.; Leszczynski, J.; Goede, P.; Latypov, N. & Ostmark, H. Fourier transform Raman spectroscopy of the four crystallographic phases of α, β, γ and ε 2,4,6,8,10,12-Hexanitro-2,4,6,8,10,12-hexaazatetracyclo[5.5.0.05,9.03,11] dodecane (HNIW, CL-20). Propellants, Explosives, Pyrotechnics, 2004, 29(4), 205-08.

18. Foltz, M.F.; Coon, C.L.; Garcia, F. & Nicholas III, A.L. Thermal stability of the polymorphs of hexanitrohexaazaisowurtzitane, Pt II. Propellants, Explosives, Pyrotechnics, 1994, 19(3), 133-44.

19. Antoine, E.D.M.; Heijden, Van der & Bouma, R.H.B. Crystallization and characterization of RDX, HMX and CL-20. Crystal Growth Design, 2004, 4(5), 999-1007.

	Shri Mrinal Ghosh obtained his MTech from Indian Institute of Technology, Delhi in 2005. Presently working as Scientist C at High Energy Materials Research Laboratory (HEMRL), Pune. His research area includes: Polymorphism in crystalline explosive materials their preparation and characterization, vibrational spectroscopy and electron microscopic analysis.
	Dr V. Venkatesan obtained PhD (Chemistry) from Indira Gandhi Centre for Atomic Research, Kalpakkam in 2003. Presently, he is working as Scientist ‘D’ at HEMRL, Pune. His research and work experiences include: condensed phase and gas phase spectroscopy, crystallization, polymorphism, molecular and its cluster structure determination and computational calculations.
	Dr A.K. Sikder presently working as an Associate Director at the HEMRL, Pune. He has published about 85 research papers in national/ international journals. Earlier he worked on highly toxic organophosphorus compounds used as chemical warfare agents and their antidotes. His research work includes: High energy materials in a wide spectrum of subjects of defence interest.
	Dr (Mrs) Nirmala Sikder obtained her PhD from Jiwaji University, Gwalior, in 1996. Presently she is working as a Scientist at the HEMRL, Pune. She has published more than 35 papers in national/international journals. Earlier she worked on synthesis, characterization and hydrolysis studies on toxic organophosphorus compounds. Her research includes: Synthesis of high energy materials and their instrumental analysis.