Thermal decomposition mechanism of Co–ANPyO/CNTs nanocomposites and their application to the thermal decomposition of ammonium perchlorate

Abstract

A chemical precipitation method was used to prepare cobalt complexes of 2,6-diamino-3,5-dinitropyridine-1-oxide/carbon nanotube (Co–ANPyO/CNTs) nanocomposites. The structure and thermal analyses indicate that Co–ANPyO nanoparticles are well dispersed on the surface of CNTs with an average particle size of about 10 nm, the content of Co–ANPyO nanoparticles in nanocomposites is about 73.4 wt%. The thermal decomposition mechanism of Co–ANPyO and Co–ANPyO/CNTs nanocomposites were predicted based on thermogravimetry-differential scanning calorimetry (TG-DSC) and thermolysis in situ rapid-scan FTIR (RSFTIR) results. The thermal decomposition of Co–ANPyO and Co–ANPyO/CNTs nanocomposites contains two exothermic processes in the temperature range of 25–490 °C. The first exothermic process for Co–ANPyO/CNTs nanocomposites shifts towards lower temperatures compared to that of Co–ANPyO. The main products of the final residues for Co–ANPyO and Co–ANPyO/CNTs nanocomposites at 490 °C are Co₃O₄ and CoO, respectively. The catalytic performance of Co–ANPyO and Co–ANPyO/CNTs nanocomposites on thermal decomposition of ammonium perchlorate (AP) was investigated by TG-derivative thermogravimetry (DTG), DSC, non-isothermal kinetic and α–T kinetic curves analyses. The possible catalytic mechanism was also discussed and proposed. During the thermal decomposition process of AP with Co–ANPyO/CNTs nanocomposites, Co–ANPyO/CNTs nanocomposites might decompose and form Co₃O₄/CNTs and CoO/CNTs nanocomposites as high activity catalysts, which could accelerate the thermal decomposition of AP. Thus, Co–ANPyO/CNTs nanocomposites not only lower the decomposition temperature and activation energy, but also enhance the total heat of AP, which could not be achieved by the CNTs and Co–ANPyO alone. The way of preparing Co–ANPyO/CNTs nanocomposites presented in this work can be expanded to other energetic additives/CNTs nanocomposites used for AP and AP based propellants.

---

1 Introduction

As is well known, ammonium perchlorate (AP) is one of the most common oxidants in composite solid propellants, which has been used in various solid propellants. The thermal decomposition rate of AP-based propellants is related to the additives, which can influence the burning rate and pressure exponent of the propellant.^1–3 Considering its limitation in loading in composite solid propellants, it is important to improve decomposition efficiency of AP to satisfy the requirements of high energy generation at low burning temperatures. Recent work has shown that nano-sized metals, metal oxides and complex oxides (such as Fe, Co, Ni, TiO₂, Co₃O₄, Fe₂O₃, ZnO, CuO, NiO, LaCoO₃ and CuFe₂O₄) exhibit high catalytic properties for AP thermal decomposition.^4–16 However, the disadvantages of these non-energetic additives is obvious, an increase in their concentration may decrease the total energy of the solid propellant, which is one of the most important performance parameters for the solid propellant. In order to overcome these disadvantages, hunting for energetic additives with high catalytic activities is one of the future directions for the technology of AP thermal decomposition.^17–22 This will possibly be a hot research topic in both materials and chemistry fields.^23–32

Carbon nanotubes (CNTs) have attracted much attention from many researchers due to their superior mechanical, electrical and thermal properties.^33–37 In particular, their nano-scale size, low density, and high aspect ratio have put them into position as the promising candidate for reinforcement in composite materials. Recently, extensive attention has been paid on synthesis and characterization of nano-sized metals/CNTs and metal oxides/CNTs and their catalytic effects on thermal decomposition of 1,3,5-trinitrohexahydro-1,3,5-triazine (RDX) and AP.^38–43 These investigations show that the nano-size metals/CNTs and metal oxides/CNTs exhibit good catalytic effect on RDX and AP thermal decomposition, indicating metals/CNTs and metal oxides/CNTs nanocomposites can lead to a possible concerted effect or integration of the properties of the two components. However, energetic additives/CNTs nanocomposites have been investigated to a lesser extent than the other nanocomposites.

Inorganic 3d-transition cobalt compounds (Co, CoO, Co₃O₄, CuCo₂O₄, CoC₂O₄·2H₂O) have attracted tremendous attention due to their remarkable catalytic effects on the thermal decomposition of AP.^{7,8,13,24,44,45} The reasons for these cobalt compounds such as CoO and Co₃O₄ exhibit remarkable catalytic performance for thermal decomposition of AP due to their high and stable catalytic activity and selectivity toward the oxidation of ammonia, the intermediate product of AP decomposition.^8,13,44,46 Our early work reported the synthesis, crystal structure and properties of cobalt complex of 2,6-diamino-3,5-dinitropyridine-1-oxide (Co–ANPyO).³¹ Co–ANPyO is an insensitive, thermal stable energetic complex with a special coordination mode (Fig. 1), which has good compatibilities with RDX, cyclotetramethylene tetranitramine (HMX), nitrocellulose (NC) and aluminum (Al). Furthermore, differential scanning calorimetry (DSC) studies show that Co–ANPyO exhibits a good catalytic effect on the thermal decomposition of AP. However, the thermal decomposition of AP catalyzed by Co–ANPyO has been studied, while the underlying mechanism of Co–ANPyO additives in the thermal decomposition of AP is not clear. More importantly, the aggregation of the Co–ANPyO particles in nature may result in its insignificant catalytic performance compared to that of nano-sized cobalt compounds.^{7,8,13,24,44,45}

Fig. 1 Molecular structures of ANPyO (a) and Co–ANPyO (b).

The purpose of this work is to improve the catalytic ability of Co–ANPyO additive in the thermal decomposition of AP, and investigate the mechanism of AP thermal decomposition catalyzed by Co–ANPyO. Thus, we developed a new strategy for the synthesis and characterization of Co–ANPyO/CNTs nanocomposites, and characterized with FTIR spectroscopy, scanning electron microscopy (SEM), transmission electron microscope (TEM), high-resolution scanning transmission electron microscopes (STEM), Brunauer–Emmett–Teller (BET) and X-ray photoelectron spectroscopy (XPS). The thermal decomposition mechanism of Co–ANPyO and Co–ANPyO/CNTs nanocomposites was predicted base on thermogravimetry (TG)-DSC and thermolysis in situ rapid-scan FTIR (RSFTIR) results. The catalytic performance of Co–ANPyO and Co–ANPyO/CNTs nanocomposites on the thermal decomposition of AP was investigated by TG-derivative thermogravimetry (DTG), DSC, non-isothermal kinetic and α–T kinetic curves analyses. The possible catalytic mechanism was also discussed and proposed.

2 Experimental

2.1 Materials and instrumentation

All chemicals used were analytical grade, and purchased from commercial sources without further purification. CNTs were purchased from commercial sources.

The FTIR studies were conducted with use of a Bruker (55FT-IR) FTIR Spectrometer (500–4000 cm⁻¹). Elemental contents of carbon, hydrogen, and nitrogen were determined by a German Vario EL III analyzer. SEM images were obtained on JEM-2000CX scanning electron microscope. An H-8100 TEM operating at 200 kV accelerating voltage was used for TEM. XPS was performed with an American Thermo ESCALAB 250 electron spectrometer using Al K irradiation. DSC analyses were recorded on a TA-DSC-Q20 from 25 to 500 °C, TG-DTG analyses were conducted on a TGA/SDTA851eMETTLER TOLEDO from 25 to 500 °C. The detailed microscopic structure and the chemical composition of the Co–ANPyO/CNTs nanocomposites were investigated using high-resolution scanning transmission electron microscopes (Cs-corrected HR-STEM, JEM2010F and JEM2200FS operating at 200 kV, JEOL).

The conditions of TG-DTG and DSC were: sample mass, about 1.2–1.5 mg; N₂ flowing rate, 40 cm³ min⁻¹; heating rates (β), 2.5, 5, 10 and 15 °C min⁻¹, furnace pressures, 0.1 MPa; reference sample, α-A1₂O₃; type of crucible, aluminum pan with a pierced lid.

RSFTIR measurements were conducted using a Nicolet Model NEXUS 870 FT-IR Instrument and in situ thermolysis cell (Xiamen University, China) in the temperature range of 20–500 °C. Ar flowing rate, 10 cm³ min⁻¹. Heating rate: 10 °C min⁻¹. KBr pellet samples, well mixed by about 1.5 mg samples and 120 mg KBr were used. Infrared spectra in the range of 4000–500 cm⁻¹ were obtained by model DTGS detector at a rate of 15 files per min and 10 scans per file with 5 cm⁻¹ resolution.

2.2 Synthesis and preparation

2.2.1 Synthesis of ANPyO. ANPyO was prepared according to the literature.³¹ Anal. calcd (%): C, 27.89; H, 2.32; N, 32.54. Found: C, 27.87; H, 2.35; N, 32.52.

2.2.2 Synthesis of Co–ANPyO. Co–ANPyO was prepared according to the literature and recrystallized by N,N-dimethyl-formamide (DMF).³¹ Anal. calcd (%): C, 23.84; H, 2.38; N, 27.82. Found: C, 23.85; H, 2.39; N, 27.81. IR, (KBr, cm⁻¹): 3608, 3418, 3382, 3298, 3060, 1602, 1536, 1458, 1380, 1314, 1236, 1173, 1124, 1050, 951, 810, 744, 695, 620, 568.

2.2.3 Synthesis of Co–ANPyO/CNTs nanocomposites. Co–ANPyO/CNTs nanocomposites were prepared by chemical precipitation method, the synthesis process was described as follows:
Preparation of ANPyO solution. ANPyO (0.5 g, 2.3 mmol) was dissolved in DMF (60 ml) at 120 °C for 60 min, and kept the temperature of the solution at 60–70 °C for the next step.
Preparation of Co–ANPyO/CNTs nanocomposites. CNTs (0.5 g), tetrabutyl ammonium bromide (0.12 g) and Co(NO₃)₂·6H₂O (0.65 g, 2.3 mmol) were added in distilled water (300 ml) under ultrasound condition at 70 °C for 45 min, subsequently, the ANPyO solution was added into the above mixture stirring at 80 °C drop by drop, after that, the solution was stirred at 80 °C under ultrasound condition for another 2.5 h. Finally, the sample was collected by filtration, washed with distilled water several times, dried and heated at 50 °C for 7 h.

2.2.4 Preparation of Co–ANPyO/CNTs/AP and Co–ANPyO/AP mixtures. The Co–ANPyO/CNTs/AP and Co–ANPyO/AP mixtures were prepared by dry mixed. The content of the Co–ANPyO and Co–ANPyO/CNTs nanocomposites in AP was 5 wt%.

3 Results and discussion

3.1 Characterizations

Fig. 2(a) shows the SEM image of Co–ANPyO, in which the samples of Co–ANPyO were recrystallized by DMF. The shape of the samples is cuboid and smooth, with the average particle size and specific surface area of about 12 μm and 0.96 m² g⁻¹. The SEM image of CNTs (Fig. 2(b)) shows that most of the CNTs in the arrays appear to slightly tangle or curved, and these tube roots are isolated and have almost same diameter (about 50 nm), no particles are observed. Fig. 2(c) shows the TEM image of Co–ANPyO/CNTs nanocomposites, in which clearly discloses that there are no obvious changes in the morphology of CNTs, most of the CNTs in the TEM image almost keep the original state. However, it can be seen clearly that there are a lot of small Co–ANPyO nanoparticles uniformly deposited on the surface of CNTs, with the average particle size about 10 nm. The Co–ANPyO nanoparticles are anchored tightly and well dispersed on the surface of CNTs, forming compact coating, and the individual Co–ANPyO nanoparticles are well separated from each other. The growth of newborn Co–ANPyO nanoparticles is depressed due to the intense friction and collisions of the molecules created by ultrasound condition. With ultrasound condition of reactant in DMF and water, temperature and concentration gradients could be avoided, providing a uniform environment for the nucleation and growth of Co–ANPyO nanoparticles.⁴⁷ Furthermore, CNTs sheet could inhibit the aggregation of Co–ANPyO nanoparticles.^48,49 This helps to prevent Co–ANPyO nanoparticles from agglomerating and guarantee the efficient catalytic activities of the Co–ANPyO/CNTs nanocomposites.

Fig. 2 SEM images of Co–ANPyO (a), CNTs (b), TEM, STEM images of Co–ANPyO/CNTs nanocomposites (c and d) and corresponding elemental mapping images of C, Co, N, O (e and h).

STEM image and elemental mapping analysis of Co–ANPyO/CNTs nanocomposites (Fig. 2(d)–(h)) also suggest the presence of Co, C, N and O components in the Co–ANPyO/CNTs nanocomposites, which indicate that there are a lot of small Co–ANPyO nanoparticles uniformly deposited on the surface of CNTs.

The FTIR spectrum is used to further confirm the functional groups of CNTs, Co–ANPyO and Co–ANPyO/CNTs nanocomposites, respectively, as shown in Fig. 3. As can be seen in Fig. 3(a), the FTIR spectrum of CNTs discloses the presence of 1650(C=C) and 2090(C=C) cm⁻¹, which are in agreement with the functional groups of CNTs. As can be seen in Fig. 3(b), the FTIR spectrum of Co–ANPyO discloses the presence of 3608(H₂O), 3420(NH₂), 3337(NH), 3298(NH₂), 3064, 1124, 810, 744, 695(CH), 1583(NH, NH₂), 1524(C=C, C=N), 1458(NO₂), 1374(NO₂), 1308(N→O), 568(Co–N) and 620(Co–O) cm⁻¹, respectively, which are in agreement with the functional groups of Co–ANPyO.³¹ And the FTIR spectrum of Co–ANPyO/CNTs nanocomposites (Fig. 3(c)) shows that the characteristic peaks for CNTs and Co–ANPyO could be determined in the FTIR, while the intensity of the characteristic peaks for CNTs and Co–ANPyO turn to be weaken, maybe caused by the fine loading of Co–ANPyO nanoparticles on CNTs surface.

Fig. 3 FTIR spectrums of pristine CNTs, Co–ANPyO and Co–ANPyO/CNTs nanocomposites.

By XPS measurement of the CNTs, Co–ANPyO and Co–ANPyO/CNTs nanocomposites (Fig. 4), the chemical composition of Co–ANPyO/CNTs nanocomposites can be further confirmed. As shown in Fig. 4(a), in which discloses the presence of C and O for CNTs, respectively. C 1s XPS spectrum of CNTs in Fig. 4(b) shows four types of carbon with different chemical states are observed, in which appear at 284.3(C–C), 284.9(C–OH), 287.2(C=O) and 290.5 eV(COOH), respectively.⁵⁰ It is noted that the C/O ratio for CNTS in the composite is estimated to be 93 : 7, which indicates some degree of oxidation with three components that correspond to carbon atoms in different functional groups.

Fig. 4 XPS patterns of Co–ANPyO/CNTs nanocomposites, Co–ANPyO and CNTs(a), C 1s XPS spectrums of Co–ANPyO/CNTs nanocomposites, Co–ANPyO and CNTs (b), Co 2p XPS spectrums of Co–ANPyO/CNTs nanocomposites and Co–ANPyO (c) N 1s XPS spectrums of Co–ANPyO/CNTs nanocomposites and Co–ANPyO (d) and O 1s XPS spectrums of Co–ANPyO/CNTs nanocomposites and Co–ANPyO.

For the case of Co–ANPyO in Fig. 4(a), in which discloses the presence of C, N, O and Co, respectively. XPS spectrum of Co–ANPyO also exhibits two peaks at 781.1 and 796.0 eV, corresponding to Co2p_3/2 and Co2p_1/2 spin–orbit of Co–ANPyO, respectively, which confirm the formation of cobalt.⁸ The presence cobalt can be further confirmed by Co 2p, N 1s and O 1s XPS spectrums of Co–ANPyO in Fig. 4(c)–(e), respectively, in which the characteristic peaks close to 780.5(Co2p_3/2), 795.7(Co2p_1/2), 397.4 (Co–N) and 530.5(Co–O) eV. This confirms the formation of Co–N and Co–O bonds in the molecular structure of Co–ANPyO. As can be seen in Fig. 4(b), C 1s XPS spectrum of Co–ANPyO shows four types of carbon with different chemical states are observed, which appear at 283.1(C–H), 284.2(C=C), 285.0(C–N) and 286.6 eV(C=N), respectively. Fig. 4(d) and (e) also exhibit the N 1s and O 1s XPS spectrums of Co–ANPyO, six types of nitrogen and four types of oxygen with different chemical states are observed, in which appear at 397.4(N–Co), 398.7(N–H), 400.4(N–C), 401.1(N=C), 404.1(N–O, N→O), 405.9(N–O, NO₂), 530.5(O–Co), 531.2(O–H, H₂O), 532.0(O–N, N→O) and 533.7 eV(O–N, NO₂), respectively.

For the case of Co–ANPyO/CNTs nanocomposites in Fig. 4(a), in which discloses the presence of C, N, O and Co, respectively. The C 1s, N 1s, Co2p and O 1s XPS spectrums of Co–ANPyO/CNTs nanocomposites in Fig. 4 indicate that there are no obvious changes in the chemical states for N1s and Co2p, but some changes for C1s and O1s. The C 1s and O 1s XPS spectrums of Co–ANPyO/CNTs nanocomposites disclose that the content of C–C, C–OH, C=O and COOH groups significant increase compared to that of Co–ANPyO. The XPS results prove that the nanoparticles of Co–ANPyO are formed and load on the CNTs surface.

BET analyses are performed to further investigate the specific surface area of the nanoparticles, which is one of most important factors that guarantee the efficient catalytic activities of catalyst. As shown in Fig. 5, the Co–ANPyO/CNTs nanocomposites show a type H3 hysteresis loop in the range of P/P_o with the specific surface area of 43.1 m² g⁻¹, in which is higher than that of normal-sized Co–ANPyO (<1.0 m² g⁻¹).

Fig. 5 Nitrogen adsorption–desorption isotherms for Co–ANPyO/CNTs nanocomposites.

Base on SEM, TEM, STEM, FTIR, XPS and BET characterizations for CNTs, Co–ANPyO and Co–ANPyO/CNTs nanocomposites, we conclude that the Co–ANPyO nanoparticles are formed and well dispersed on the surface of CNTs in the chemical precipitation process. On the basis of our experimental results, a scheme has been presented to describe such a formation process of Co–ANPyO/CNTs nanocomposite, as illustrated in Fig. 6.

Fig. 6 A scheme shows a proposed formation route of Co–ANPyO nanoparticles onto the surfaces of CNTs sheets.

3.2 Thermal decomposition mechanism of Co–ANPyO and Co–ANPyO/CNTs

In order to realize the thermal decomposition mechanism of Co–ANPyO and Co–ANPyO/CNTs, thermal decomposition of CNTs, Co–ANPyO and Co–ANPyO/CNTs was investigated by TG-DSC and RSFTIR measurements at the heating rate of 10 °C min⁻¹, the results were shown in Fig. 7 and 8, respectively.

Fig. 7 TG-DSC curves of Co–ANPyO and Co–ANPyO/CNTs nanocomposites at the heating rate of 10 °C min⁻¹.

Fig. 8 The RSFTIR spectrums of Co–ANPyO at different temperatures.

3.2.1 Thermal decomposition of CNTs. As is well known, CNTs have good thermal and chemical stability, which are considered as good support for catalyst.³⁷ As shown in Fig. 7, there are no obvious changes in the TG-DSC curves for CNTs in the range of 25–500 °C, in which reveal that CNTs do not decompose during this process. Thus, we hypothesis that CNTs in Co–ANPyO/CNTs nanocomposites do not decompose during the heating process in the range of 25–500 °C.

3.2.2 Thermal decomposition mechanism of Co–ANPyO. The FTIR spectrum of Co–ANPyO at room temperature is shown in Fig. 3(b). The stretching vibration absorption of V_(C–NO2) peaks are at 1458 and 1374 cm⁻¹ for the NO₂, the V_(NH2) peaks are at 3420, 3298 and 1583 cm⁻¹ for the NH₂, the V_(CH) peaks are at 3064, 1124, 810, 744 and 695 cm⁻¹ for C–H, the V_(NH) peak is at 3337 cm⁻¹ for the NH, the V_(OH) peak is at 3608 cm⁻¹ for the H₂O, the V_(NO) peak is at 1308 cm⁻¹ for the N→O, the V_(Co–N) peak is at 568 cm⁻¹ for the Co–N and the V_(Co–O) peak is at 620 cm⁻¹ for the Co–O.

As can be seen in Fig. 8, the absorption peak of H₂O almost disappears at 200 °C, while the other characteristic groups do not change. Corresponding to the TG-DSC curves, there is a mass loss of 7.0% in this process, which corresponds well with the calculation value of 7.1%, while no obvious changes from the DSC curve. This process would be the loss of three H₂O molecules from the Co–ANPyO.

The first exothermic stage for the Co–ANPyO occurs in the range of 250.0–303.9 °C with the peak temperature at 287.8 °C. Corresponding to this process, there is a mass loss of 33.1% from the TG curve. It can be seen that the absorption peaks of NH₂ and NO₂ decrease, and the absorption peaks of C–H, N →O, Co–O and Co–N almost disappear. Furthermore, the new absorption peaks at 665, 548, 460 and 430 cm⁻¹ prove the existence of Co₃O₄ and CoO in the solid residue.^44,51 This process would be the Co–O, Co–N bonds breaking of the Co–ANPyO and the ring breaking of the ligands, which may be attributed to the partial decomposition of Co–ANPyO.

The second exothermic stage occurs in the range of 303.9–380.3 °C with the peak temperature at 315.0 °C. Corresponding to this process, there is a mass loss of 16.9% from the TG curve. The cleavage of the amino-groups and nitro-groups can be confirmed by the disappearance of the absorption bands of V_(NH2), V_(C–NO2) and V_(NH), respectively. The breaking of the pyridine ring can be confirmed by the disappearance of the absorption bands of V_(C–H), V_(N→O), V_(C=C) and V_(C=N), respectively. While the new absorption peaks at 665, 548, 460 and 430 cm⁻¹ for Co₃O₄ and CoO keep increasing. The other new absorption peak at 2165 cm⁻¹ prove the existence of Co(NCO)₂ in the solid residue.⁵² But the intensity of Co(NCO)₂ is weaker compared to that of Co₃O₄ and CoO, which implies that Co(NCO)₂ might be the by-product for solid residue. This process would be the Co–O, Co–N bonds breaking of the Co–ANPyO and the ring breaking of the ligands, which may be attributed to the completely decomposition of Co–ANPyO.

On the TG curve, there still is a slow mass loss of 3.1% from 380.3 to 490 °C. Corresponding to this process, there are no obvious changes from DSC and RSFTIR curves. Therefore, the decomposition pathway of Co–ANPyO may be described as follows:

3.2.3 Thermal decomposition mechanism of Co–ANPyO/CNTs. As can be seen in Fig. 7, there are no obvious difference in the thermal decomposition behaviors for Co–ANPyO and Co–ANPyO/CNTs nanocomposites, but some changes in decomposition pattern. The first exothermic stage of Co–ANPyO/CNTs nanocomposites occurs in the range of 138.7–278.6 °C with the peak temperature at 248.1 °C, which notably shifts towards lower temperatures compared to that of Co–ANPyO. However, there are no obvious changes in the second exothermic stage of Co–ANPyO/CNTs nanocomposites compared to that of Co–ANPyO. Nano-size particles exhibit an increase in the ratio of surface atoms to interior atoms compared to that of normal-size particles.⁵³ This may lead to a higher surface energy and result in the decrease of the temperature for thermal decomposition process. This is the main reasons for above phenomenon. Furthermore, the absorption peaks at 2165, 665, 548, 460 and 430 cm⁻¹ also prove the existence of Co₃O₄, CoO and Co(NCO)₂ in the solid residue at 490 °C, which are in agreement with that of Co–ANPyO. According to the mass loss of Co–ANPyO and Co–ANPyO/CNTs nanocomposites, about 73.4 wt% of Co–ANPyO deposited on the surface of CNTs.

Based on above results, we propose that the main solid products for thermal decomposition of Co–ANPyO and Co–ANPyO/CNTs nanocomposites are Co₃O₄ and CoO, respectively, which may be contributed to the catalytic effects of AP thermal decomposition.

3.3 Catalytic effects of Co–ANPyO and Co–ANPyO/CNTs nanocomposites on the thermal decomposition of AP

3.3.1 Thermal decomposition of pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP. Recently, the thermal decomposition of AP catalyzed by CNTs has been investigated by Liu.⁵⁴ This investigation shows that CNTs exhibits insignificant catalytic effects on the thermal decomposition of AP.

TG, DTG and DSC results of the decomposition of AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP were shown in Fig. 9 and Table 1, respectively. As shown in Fig. 9(a), the DSC curve for pure AP reveals that the thermal decomposition of AP takes place in three steps: the endothermic phase transition at 242.3 °C (ascribed to a phase transition of AP from orthorhombic to cubic), the low-temperature decomposition (LTD) at 332.2 °C and the high-temperature decomposition (HTD) at 432.5 °C (contributed to intermediate products such as NH₃ and HC1O₄ and a complete one to volatile products respectively).⁵⁵ Corresponding, the TG, DTG curves for pure AP reveal that the 20 wt% weight loss at LTD is attributed to the partial decomposition of AP. The 75 wt% weight loss at HTD is attributed to the complete decomposition of the intermediate to volatile products.⁵⁵

Fig. 9 TG, DTG and DSC curves for AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP at the heating rate of 10 °C min⁻¹.

Table 1 TG-DTG and DSC results for pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP in LTD and HTD processes^a

Sample	Endothermic peak/°C	LTD				HTD				ΔH/J g^−1
		To/°C	Tp/°C	Te/°C	Mass loss/%	To/°C	Tp/°C	Te/°C	Mass loss/%
a Note: To, onset temperature of decomposition for DSC curve. Te, end temperature of decomposition for DSC curve. Tp, peak temperature of decomposition for DSC curve.
Pure AP	242.3	304.3	332.2	353.9	20	363.1	432.5	443.6	75	655
Co–ANPyO/AP	246.8	256.5	288.1	321.0	11	321.0, 382.7	348.0, 389.4	382.7, 426.5	49, 40	1084
Co–ANPyO/CNTs/AP	245.1	282.8	337.8	383.4	100	—	—	—	—	1850

---

When Co–ANPyO is added, the DSC result for Co–ANPyO/AP reveals that the HTD process crakes into two steps, and the LTD and HTD processes shift towards lower temperatures compared to that of pure AP. Corresponding, the TG, DTG curves for Co–ANPyO/AP reveal that there are 11.0 wt% weight loss at LTD process and 89.0 wt% weight loss at HTD process. This shows that the LTD process for Co–ANPyO/AP is narrowed down, while the HTD process for Co–ANPyO/AP is enlarged. The overall heat for HTD and LTD processes (1084 J g⁻¹) is 429 J g⁻¹, which is higher than that of pure AP. The TG, DTG and DSC curves of Co–ANPyO/AP show that Co–ANPyO additive has no effects on the crystallographic transition temperature, but some changes in the decomposition pattern. AP decomposition is accelerated in the presence of the Co–ANPyO.

The TG, DTG and DSC results for Co–ANPyO/CNTs/AP show a difference thermal decomposition behavior compared to that of pure AP and Co–ANPyO/AP. Typically, the HTD process for Co–ANPyO/CNTs/AP disappeared, while the LTD for Co–ANPyO/CNTs/AP shows a peak temperature closer to that observed for pure AP, but higher than that of Co–ANPyO/AP. This indicates that Co–ANPyO/CNTs nanocomposites do not influence the primary dissociation of AP into ammonia and perchloric acid, but most likely catalyze the secondary process involved in AP decomposition, and accelerates the overall decomposition process. Similar effects have been found from nano-sized Co, CoO, Co₃O₄, CuCo₂O₄, CoC₂O₄·2H₂O, but have not been discussed in detail.^{7,8,13,24,45,56} This indicates a difference catalytic mechanism for Co–ANPyO and Co–ANPyO/CNTs nanocomposites. The LTD process for the mixture systems (337.8 °C) is 94.7 °C, which is lower than HTD of pure AP. And decomposition heat of 1850 J g⁻¹ is observed in the presence of Co–ANPyO/CNTs nanocomposites, nearly triple the decomposition heat for pure AP.

Based on above results, it show that both of CNTs and Co–ANPyO exhibit insignificant catalytic effects on the thermal decomposition of AP. However, because of the combination of the CNTs and Co–ANPyO for the catalysis of AP decomposition, Co–ANPyO/CNTs nanocomposites not only lower the decomposition temperature, but also enhance the total heat of AP, which could not be achieved by the CNTs and Co–ANPyO alone. Furthermore, a lot of research on the decomposition of AP catalyzed by cobalt compounds (such as Co, CoO, Co₃O₄, CuCo₂O₄, CoC₂O₄·2H₂O, Co–ANPyO) and nanocomposites (such as Co₃O₄/grapheme oxide, Co₃O₄/CNTs) has been carried out.^{7,8,13,24,31,45,56–59} And some of these cobalt catalysts can decrease exothermic peaks of LTD and HTD processes for AP to slightly lower temperature than that of Co–ANPyO/CNTs nanocomposites. While this work shows that AP with Co–ANPyO/CNTs nanocomposites shows notably higher decomposition heat than that of AP with all of these cobalt catalyst, which suggests better catalytic effect than these cobalt catalysts.

3.3.2 Nonisothermal reaction kinetics of pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP. DSC and kinetic parameters of the overall decomposition processes for pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP, estimated by the Kissinger's method, were given in Table 2.⁶⁰ A decrease in the apparent activation energy (E_a) of LTD and HTD processes for Co–ANPyO/AP are observed in the catalysed systems. Similar situation in the E_a of LTD process for Co–ANPyO/CNTs/AP is observed, but exhibits higher E_a than that of LTD process for Co–ANPyO/AP. The E_a of LTD and HTD processes for AP decomposition, associates with the primary dissociation step and completely decomposition step.⁶¹ The above results indicate that Co–ANPyO not only influences the primary dissociation of AP, but also influences the completely decomposition of AP. While Co–ANPyO/CNTs nanocomposites do not influence the primary dissociation of AP, which is in agreement with the TG-DTG and DSC results.

Table 2 Summary of DSC and Kinetic parameters results for pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP in LTD and HTD processes

Sample	β/(°C min^−1)	ΔH/J g^−1	LTD			HTD
			Tp/°C	Ea/(kJ mol^−1)	R^2	Tp/°C	Ea/(kJ mol^−1)	R^2
Pure AP	2.5	208	310.9	173.9	0.9975	405.5	185.6	0.9895
	5	244	321.5			421.2
	10	655	332.2			432.5
	15	689	340.2			443.1
Co–ANPyO/AP	2.5	855	263.5	139.7	0.9855	313.5, 342.7	158.5, 90.47	0.9781, 0.9901
	5	1039	273.7			328.6, 366.8
	10	1084	288.1			339.8, 386.5
	15	1221	292.5			345.2, 405.7
Co–ANPyO/CNTs/AP	2.5	1767	310.8	148.5	0.9765	—	—	—
	5	1833	320.6			—
	10	1850	337.6			—
	15	1897	342.6			—

---

It can be seen that both of Co–ANPyO/CNTs nanocomposites and Co–ANPyO increase the overall heat for HTD and LTD processes of AP. In general, Co–ANPyO/CNTs nanocomposites exhibit higher catalytic activity than that of Co–ANPyO.

3.3.3 α–T kinetic curves of pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP. To explore the thermal decomposition mechanism of pure AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP, the corresponding α–T (α is the extent of conversion, 0 < α < 100) kinetic curves (Fig. 10) were obtained by dealing the TG curves and compared. It has been shown that the thermal decomposition of Co–ANPyO/AP starts at lower temperatures than that of pure AP. This temperature difference points toward the AP decomposition catalyzed by Co–ANPyO.

Fig. 10 α–T kinetic curves for AP, Co–ANPyO/AP and Co–ANPyO/CNTs/AP at the heating rate of 10 °C min⁻¹.

However, It can be seen from the figure that, during the stages decomposition (α < 10), the thermal decomposition of Co–ANPyO/CNTs/AP starts at higher temperatures than that of pure AP and Co–ANPyO/AP for the same extent of conversion. While during the stages decomposition (10 < α < 95), the thermal decomposition of Co–ANPyO/CNTs/AP shows significantly lower temperature than that of pure AP and Co–ANPyO/AP for the same extent of conversion. This indicates a difference catalytic mechanism for AP catalyzed by Co–ANPyO and Co–ANPyO/CNTs nanocomposites, which is in agreement with the TG, DTG, DSC and nonisothermal reaction kinetics results.

By means of α–T kinetic curves investigation, Co–ANPyO/CNTs nanocomposites exhibit higher catalytic activity than that of Co–ANPyO.

3.4 Possible catalytic mechanism of Co–ANPyO and Co–ANPyO/CNTs on the thermal decomposition of AP

Based on above experimental results, we conclude that Co–ANPyO/CNTs nanocomposites play better catalytic role than that of Co–ANPyO on AP thermal decomposition. TG, DTG, DSC, non-isothermal kinetic and α–T kinetic curves analyses indicate that Co–ANPyO/CNTs nanocomposites do not influence the primary dissociation of AP into ammonia and perchloric acid, but most likely catalyze the secondary process involved in AP decomposition. While Co–ANPyO influences both the LTD and HTD processes for AP decomposition. What is reason for this phenomenon?

We have proposed that (3.2) the main solid products for thermal decomposition of Co–ANPyO and Co–ANPyO/CNTs nanocomposites are Co₃O₄ and CoO, respectively. AP decomposition involves two crucial steps: (1) ammonia oxidation and (2) dissociation of ClO₄⁻ species into ClO₃⁻ and O₂.⁴⁶ In the first step, Co₃O₄ and CoO exhibit high and stable catalytic activity and selectivity toward ammonia oxidation, thus promoting the partial decomposition of AP.⁴⁶ In the second step, Co₃O₄ and CoO belong to p-type semiconductor, which could release of O²⁻ ions, form easy melting eutectics or intermediate amine compounds with AP, thus promoting the completely decomposition of AP.⁶² According to proton transfer mechanism, it can be proposed that the Co–ANPyO decomposes and releases a large amount of heat itself. This enhances the total heat of the AP mixture, as well as the formation of Co₃O₄ and CoO in situ on the AP surface. Therefore, the Co₃O₄ and CoO used here can speed up above two controlling steps.^56,63

For the case of Co–ANPyO/CNTs nanocomposites, CNTs exhibits some catalytic effects on the thermal decomposition of AP.⁵⁴ And we have concluded that the thermal decomposition process for Co–ANPyO/CNTs nanocomposites shifts towards lower temperature compared to that of Co–ANPyO (3.2). Thus, this might result in formation Co₃O₄/CNTs and CoO/CNTs nanocomposites in situ at lower temperatures. More importantly, CNTs sheet could inhibit the aggregation of the Co₃O₄ and CoO nanoparticles. We hypothesize that CNTs is more likely to aggregate and expose active sites to absorb ammonia which covers the CNTs surface, then create a supersaturated atmosphere of NH₃.^64,65 This might result in depriving the catalytic activities of Co₃O₄/CNTs and CoO/CNTs nanocomposites. As a result, the LTD process Co–ANPyO/CNT/AP shifts towards relatively higher temperature. As the temperature rises, the catalytic activities of Co₃O₄/CNTs and CoO/CNTs nanocomposites recover, the ammonia would be oxidized with a higher speed catalyzed by Co₃O₄/CNTs and CoO/CNTs nanocomposites. And the completely decomposition of AP would be accelerated with a higher speed catalyzed by Co₃O₄/CNTs and CoO/CNTs nanocomposites.

A possible mechanism of AP thermal decomposition catalyzed by Co–ANPyO/CNTs nanocomposites is proposed, as shown in Fig. 11.

Fig. 11 Mechanism of AP thermal decomposition catalyzed by Co–ANPyO/CNTs nanocomposites.

4 Conclusion

Co–ANPyO nanoparticles were obtained by chemical precipitation method using CNTs, Co(NO₃)₂·6H₂O and ANPyO as raw materials. The Co–ANPyO nanoparticles are well dispersed on the surface of CNTs with the average particle size about 10 nm, the content of Co–ANPyO nanoparticles in nanocomposites is about 73.4 wt%. Thermal analyses show that the thermal decomposition of Co–ANPyO and Co–ANPyO/CNTs nanocomposites contains two exothermic processes in the temperature range of 25–490 °C. The main products of final residues for them at 490 °C are Co₃O₄ and CoO, respectively. TG, DTG, DSC, non-isothermal kinetic and α–T kinetic curves investigations show that Co–ANPyO/CNTs nanocomposites not only lower the decomposition temperature, but also enhance the total heat of AP, which could not be achieved by the CNTs and Co–ANPyO alone. The way of preparing Co–ANPyO/CNTs nanocomposites presented in this work can be expanded to other energetic additives/CNTs nanocomposites used for AP and AP based propellants.

Acknowledgements

We gratefully acknowledge the financial support from Nanjing University of Science and Technology and Xi'an Modern Chemistry Research Institute. This work is supported by Five-Year (2011–2015) Pre-research Project (no. 62201070102).

References

1. S. Vyazovkin and C. A. Wight, Chem. Mater., 1999, 11, 3386.
2. L. Bereczk, K. Marthi and P. Huszthy, J. Therm. Anal. Calorim., 2004, 78, 449.
3. H. Xu, X. Wang and L. Zhang, Powder Technol., 2008, 185, 176.
4. A. Dey, J. Athar and P. Varma, RSC Adv., 2015, 5, 1950.
5. X. F. Guan, L. P. Li and J. Zheng, RSC Adv., 2011, 1, 1808.
6. D. L. Reid, A. E. Russo and R. V. Carro, Nano Lett., 2007, 7, 2157.
7. H. Y. Zhao, L. Guo and S. F. Chen, RSC Adv., 2013, 3, 19929.
8. S. S. Liu, X. Y. Jing and J. Y. Liu, J. Solid State Chem., 2013, 197, 345.
9. Z. Ma, F. Li and H. Bai, Propellants, Explos., Pyrotech., 2006, 31, 447.
10. X. Sun, X. Qiu and L. Li, Inorg. Chem., 2008, 47, 4146.
11. J. Yin, Q. Lu and Z. Yu, Cryst. Growth Des., 2009, 10, 40.
12. J. Zhu, G. Zeng and F. D. Nie, Nanoscale., 2010, 2, 988.
13. J. Wang, Z. Q. Qiao and L. Zhang, CrystEngComm, 2014, 16, 8673.
14. L. J. Chen, G. S. Li and P. Qi, J. Therm. Anal. Calorim., 2008, 92, 765.
15. Y. P. Wang, X. J. Yang and L. D. Lu, Powder Technol., 2008, 185, 231.
16. T. Liu, L. S. Wang and P. Yang, Mater. Lett., 2008, 62, 4056.
17. G. Singh and S. P. Felix, Combust. Flame, 2003, 132, 422.
18. Z. Cheng, G. Zhang and X. Fan, Inorg. Chim. Acta, 2014, 421, 191.
19. Q. Li, Y. He and R. F. Peng, RSC Adv., 2015, 5, 24507.
20. Q. Yang, S. Chen and G. Xie, J. Hazard. Mater., 2011, 197, 199.
21. M. Friedrich, J. C. Gálvez-Ruiz and T. M. Klapötke, Inorg. Chem., 2005, 44, 8044.
22. D. E. Jones, K. Armstrong and T. Parekunnel, J. Therm. Anal. Calorim., 2006, 86, 641.
23. R. P. Singh, R. D. Verma and D. T. Meshri, Angew. Chem., Int. Ed., 2006, 45, 3584.
24. M. Zou, X. H. Jiang and L. D. Lu, J. Hazard. Mater., 2012, 225, 124.
25. N. Li, M. H. Cao and Q. Y. Wu, CrystEngComm, 2012, 14, 428.
26. K. Karaghiosoff, T. M. Klapotke and P. Mayer, Inorg. Chem., 2008, 47, 1007.
27. H. B. Liu, Q. Z. Jiao and Y. Zhao, J. Alloys Compd., 2010, 496, 317.
28. M. Kaiser, U. Ticmanis and A. Hammerl, Inorg. Chem., 2001, 40, 3570.
29. Y. Shvedenkov, M. Bushuev and G. Romanenko, Eur. J. Inorg. Chem., 2005, 167, 1678.
30. J. J. Liu, Z. L. Liu and J. Cheng, J. Solid State Chem., 2013, 197, 198.
31. J. J. Liu, Z. L. Liu and J. Cheng, RSC Adv., 2013, 3, 2917.
32. J. J. Liu, Z. L. Liu and J. Cheng, J. Solid State Chem., 2013, 200, 43.
33. P. M. Ajayan and J. M. Tour, Nature, 2007, 447, 1066.
34. R. H. Baughman, A. A. Zakhidov and W. A. Heer, Science, 2002, 297, 787.
35. T. W. Ebbesen, H. J. Lezec and H. Hiura, Nature, 1996, 382, 54.
36. P. Kim, L. Shi and A. Majumdar, Phys. Rev. Lett., 2001, 87, 2155021.
37. M. F. Yu, O. Lourie and M. J. Dyer, Science, 2000, 287, 636.
38. X. J. Zhang, W. Jiang and D. Song, Mater. Lett., 2008, 62, 2343.
39. H. Ren, Y. Liu and Q. Jiao, J. Phys. Chem. Solids, 2010, 71, 149.
40. P. Cui and A. Wang, J. Saudi Chem. Soc., 2014 DOI:10.1016/j.jscs.2014.09.010 , in press.
41. P. Cui, F. Li and Z. Jian, Propellants, Explos., Pyrotech., 2006, 31, 452.
42. X. Zhang, W. Jiang and D. Song, Propellants, Explos., Pyrotech., 2009, 34, 151.
43. M. Zheng, Z. Wang and J. Wu, J. Nanopart. Res., 2010, 12, 2211.
44. E. Santacesaria, A. Morini and S. Carra, Combust. Flame, 1978, 31, 17.
45. E. Alizadeh-Gheshlaghi, B. Shaabani and B. Shaabani, Powder Technol., 2012, 217, 330.
46. (a) Y. Matatov-Meytal and M. Sheintuch, Ind. Eng. Chem. Res., 1998, 37, 309; (b) K. Schmidt-Szalowski, K. Krawczyk and J. Petryk, Appl. Catal., A, 1998, 175, 147.
47. H. Zhu, C. Zhang and Y. Yin, J. Cryst. Growth, 2004, 270, 722.
48. W. X. Chen, J. Y. Lee and Z. Liu, Mater. Lett., 2004, 58, 3166.
49. R. Licheri, R. Orrù and C. Musa, Mater. Lett., 2008, 62, 432.
50. V. Datsyuk, M. Kalyva and K. Papagelis, Carbon, 2008, 46, 833.
51. N. N. Binitha, P. V. Suraja and Z. Yaakob, J. Sol-Gel Sci. Technol., 2010, 53, 466.
52. T. B. Brill and T. L. Zhang, Combust. Flame, 2000, 121, 662.
53. B. Gao, P. Wu and B. Huang, New J. Chem., 2014, 38, 2334.
54. J. Liu, W. Jiang and Z. Wang, Acta Chim. Sin. (Chin. Ed.), 2007, 65, 2725.
55. L. Song, S. Zhang and B. Chen, Colloids Surf., A, 2010, 360, 1.
56. L. P. Li, X. F. Sun and X. Q. Qiu, Inorg. Chem., 2008, 47, 8839.
57. C. Xu, X. Wang and J. Zhu, J. Mater. Chem., 2008, 18, 5625.
58. J. Zhao, Z. S. Liu and Y. L. Qin, CrystEngComm, 2014, 16, 2001.
59. J. X. Liu, F. S. Li and W. Jiang, J. Solid Rocket Technol., 2007, 30, 243.
60. H. E. Kissinger, Anal. Chem., 1957, 29, 1702.
61. R. A. Chandru, S. Patra and C. Oommen, J. Mater. Chem., 2012, 22, 6536.
62. S. A. Makhlouf, J. Magn. Magn. Mater., 2002, 246, 184.
63. G. Singh, I. P. S. Kapoor and S. H. Dubeyl, J. Sci. Conf. Proc., 2009, 1, 11.
64. P. W. M. Jacobs and A. Russell-Jones, AIAA J., 1967, 5, 829.
65. S. S. Joshi, P. R. Paul and V. I. V. Krishnamurthy, AIAA J., 2008, 58, 721.

---