Evolution of the Micropore Structure of Ammonium Perchlorate during Low-Temperature Decomposition and Its Combustion Characteristics

Ammonium perchlorate (AP) is a common oxidant in solid propellants, and its thermal decomposition characteristics at low temperatures (less than 240 °C) are key to the study of the thermal safety of propellants. Here, the low-temperature thermal decomposition characteristics of AP were investigated at 230 °C. The micromorphology of the low-temperature decomposition residues was characterized by scanning electron microscopy and 3D nano-computed tomography in order to analyse the evolution of microscopic pore structures, and the effect of the AP pore structure on combustion performance was then tested and analysed with a homemade closed bomb. The results demonstrate that the low-temperature decomposition of AP first occurs near the surface of the particles, simultaneously starting at multiple points and forming pores, and then gradually expands towards the interior until almost all of the pores connect with one other. Compared with ordinary AP, porous AP has a significantly improved combustion rate. When the ratio of porous AP to Al was 80:20, the peak pressure in the closed bomb was increased by 2.7 times; the rate of change in peak pressure increased 34 times, leading to a higher reaction speed and higher reaction intensity, and a typical explosion reaction occurred.


Introduction
Solid propellant charges are prone to burning or exploding under unexpected thermal stimulation, leading to catastrophic accidents [1,2]. Ammonium perchlorate (AP) is the most abundant ingredient in solid composite propellants and is widely used in solid propellant engines and aerospace boosters, where the AP content is over 70% [3,4]. AP has many advantages, including an ideal density, a favorable oxygen balance, and tailorable burn rates. While AP delivers excellent performance in propellants, it is an unstable compound when heated and has the potential to react violently and explode [5][6][7]. Its safety and thermal decomposition performance directly affect the safety and burning rate of solid propellants.
In general, the thermal decomposition of AP can be divided into two stages. Thermal degradation of AP occurring below the orthorhombic-to-cubic phase transition temperature of 240 °C is known as low-temperature decomposition [8][9][10][11]. There is a cessation of decomposition after an approximately 30 wt.% loss. By increasing the temperature, decomposition continues until complete decomposition [12][13][14]. During low-temperature decomposition, it is known that the AP particle surface becomes porous, which can be observed using microscopy. The remaining material contains pores up to a few microns in size; however, the chemical composition of the particle remains unchanged [15]. The reason for this has been described in the literature; pore growth is often attributed to crystalline defects, and the drastic drop in decomposition rate is often attributed to the exhaustion of defect nuclei [16][17][18][19].
Although there is an abundance of literature that examines AP decomposition, most of it has focused on bulk material decomposition rather than single particle behaviour [20,21]. However, due to the particularity of the low-temperature decomposition of AP, the formation and evolution of the aperture at low-temperature decomposition is necessary to understand the AP decomposition mechanism. Research that considered the structure of the residue of AP decomposition reported that only the structure of the surface or cross-section can be observed; the specific morphology inside the particles could not be observed [22]. However, only by understanding the evolution of pores during the lowtemperature decomposition of AP particles can we further analyse the effect of the lowtemperature decomposition of AP on the thermal safety of propellants.
The burning rate of the propellant can obviously be accelerated by using porous AP instead of ordinary AP in the propellant [23][24][25]. However, the pore structure may lead to convective combustion, cause combustion instability, and even increase the reaction intensity of the propellant when subjected to unexpected thermal stimulation, resulting in explosion. The above problems are closely related to the thermal decomposition characteristics of AP particles at low temperature.
Herein, to better explore the decomposition of AP particles at low temperature and its influence on the thermal safety of propellant, it is necessary to quantitatively study the microstructure evolution of AP particles and the effect of the pore structure on combustion properties during low-temperature thermal decomposition. First, the decomposition characteristics of AP at slow cook-off temperatures were quantitatively analysed. Then, the microstructure and its evolution were analysed by using scanning electron microscopy and 3D nano-computed tomography, which can identify and quantify size and number density of the pore location in the particle. Finally, the effect of porous AP on combustion performance was studied by using a closed bomb, and the response mechanism of the composite propellant under slow cook-off conditions was analysed.

Experiment and Measurement
In this study, AP was used with a particle size of 150 to 220 μm obtained from the Dalian potassium chlorate factory. A synchronous thermal analyser (TG-DSC, STA449F3 Jupiter ® , NETZSCH) by NETCH in Selb, Germany and a nanoVoxel-5000 series X-ray three-dimensional microscope by Sanying Precision Instruments in Tianjin, China with a minimum resolution of 300 nm were also used.
The mass of each AP sample was approximately 10 mg heated in an inert environment in the DSC-TG thermal analyser. First, the temperature was raised to 230 °C at a rate of 10 °C/min and then maintained at 230 °C constant. By changing the heating time, AP residues with different mass loss were prepared, as shown in Table 1. Figure 1 shows the mass loss process of AP under different heating conditions (where ★ represents the decomposition termination point at different heating times). The repeatability of the experiments was good.   Table 1. Figure 2d shows the initial AP where a smooth surface has no pores. When the mass loss was 29.9% (Figure 2a), numerous pores were spread across the surface of the entire particle. As the mass loss was reduced, the pore intensity was reduced, and when the loss rate was low, the pore distribution became uneven, although the pore size did not change, with a pore diameter between 2 and 5 μm. However, scanning electron microscopy could only observe the topography of the surface of the AP particles, and the internal distribution and decomposition of the particles could not be obtained. To further study the micromorphological evolution of AP under low-temperature decomposition, the "Digital Core" technique in rock physics was used to analyse porosity, specific surface area, and pore distribution uniformity by using high-precision 3D nano-computed tomography (CT). According to differences in the internal density of the sample, the attenuation degree of X-rays differed. When the density was high, the X-ray penetration was low, and the X-ray intensity received by the detector was weak, with the image reconstruction algorithm showing this as a lighter colour. The X-rays penetrated areas with low density, such as internal cracks and pores, and the Xray intensity received by the detector was strong, with the image reconstruction algorithm showing these areas in a darker colour. Based on this, areas with similar density can be distinguished by improving the image resolution. First, the decomposed AP particles were scanned using nano-CT, and the density distribution inside the AP particles was found to be uneven because of the pores. Second, the CT results were reconstructed using Avizo (v2019.1) software to establish a 3D structure model, and an individual AP particle was extracted from this model. Finally, the microstructure characteristics of porous AP particles were analysed by filtering, image enhancement and threshold segmentation.
Porous AP, with mass loss of 29.9%, 20.2% and 9.5%, was prepared by heating to 230 °C while varying the heating time. The microstructure of AP particles for different mass losses and how this changed with the degree of decomposition were studied. A sample with a mass loss of 29.9% is described as an example. 3D CT tomography was performed using a NanovoXel-5000 nano-CT (Sanying Precision Instruments in Tianjin, China) to obtain sliced images, with a scanning accuracy of 0.3 μm, as shown in Figure 3a. The threedimensional reconstruction of the CT image was carried out by Avizo software to establish the structure of the porous AP particles. A single AP particle was then extracted using image processing, as shown in Figure 3b. The surface became uneven, and the grey shaded areas show the location of the pores. By filtering, image enhancement and threshold segmentation, the pore spatial distribution characteristics of the particles could be obtained. Taking a 2D section, as shown in Figure 3c, the same colour demonstrates that the pores were connected to each other, with most of them connected and only a very small number of them isolated. The porosity for different slices and the overall porosity of the particles could be obtained by calculating the area ratio and volume ratio of the pore region. To facilitate the analysis of the pore size, a segment of a particle was considered, as shown in Figure 3d. The different colours represent different holes, and the local amplified 3D pores are shown on the right.   Figure 4 shows the typical state of the pore distribution at different mass losses. When the mass loss reached 9.5% (a), the pores were mainly concentrated near the surface of the particles, with no pores in the core of the particles. The pores connect to each other across a small range but are still in a separate state as a whole, which can be seen from the colour distribution. As the decomposition continued and the mass loss reached 20.2% (b), the pores had spread to the centre of the particles, with most pores being connected, although a few pores were still locally connected. When the mass loss reached 29.9% (c), the pores greatly increased, with almost all the pores being connected. As seen from the evolution of the pores, the low-temperature decomposition of AP particles began near the surface of the particles and from multiple points. During the heating process, they gradually expanded to the surrounding material and towards the centre. Then, local thermal decomposition occurred inside the particles.  Combined with the mass loss process, the low-temperature decomposition of AP can be divided into three stages, as shown in Figure 5. The first is the nucleation and phase interface reactions (I), where the AP has no mass loss. Second, a rapid reaction phase (II) starts from surface decomposition and produces gas, forms pores, and extends to the centre of the particles. This stage corresponds to the linear descending segment of the TG curve, with the mass loss increasing linearly over time. Finally, local decomposition occurs within the particles (Ⅲ) when the decomposition slows, and the mass loss decreases until decomposition stops, which corresponds to the last slow decline stage of the TG curve.

Porosity
To further study the generation and evolution of pores during AP decomposition, AP particles in the decomposition residue were randomly selected for microstructure analysis. According to the reconstructed 3D pore model, the porosity of different particle slices was analysed. Figure 6 shows the pore distribution of the CT slices, with the distance between each slice being 0.3 μm. The porosity was calculated layer by layer according to the CT scan image. The x axis is the relative distance from the first scanning plane, and the y axis is the level of porosity. When the mass loss was 9.5% (a), the porosity at both ends was relatively high, whereas it was relatively low in the middle, which is consistent with the above analysis results, giving an overall porosity of 9.56%. When the mass loss was 20.2% (b), the overall porosity was 19.49%, with little change between the different slices. This was slightly less than the mass loss, which may be due to the partial decomposition of the particle surface. When the mass loss was 29.9% (c), the pores were distributed between 20% and 30%, with the porosity being relatively evenly distributed in the different slices. The average porosity of the three particles was 28.25%, which was slightly less than the mass loss.

Particle Pore Characteristics
As shown in Figure 7, the AP particles were divided into several hollow spheres with the centre of mass of the particles as the centre of the sphere, spaced 10 μm in the radial direction. The porosity of the nonhollow spheres was calculated separately, and the uniformity of the pores and the internal distribution of the particles were analysed. For the mass loss of 9.5%, the calculated results are shown in Figure 8a, where there are no pores within 40 μm of the centre, and the number of pores gradually increases as distance increases. This shows that the initial decomposition of AP particles occurs near the surface of the particles and gradually extends to the interior. When the mass loss reaches 20.2%, as shown in Figure 8b, there are no pores near the centre of mass, with a number of pores beyond 20 μm, and porosity increases as distance increases. When the mass loss was 29.9%, as shown in Figure 8c, the maximum degree of low-temperature decomposition of AP was reached, and the pore distribution varied little with distance, resulting in a uniform pore distribution.  The equivalent diameter of the pores was calculated using the maximum sphere method, and the variation process of the different pore size ratios with mass loss was analysed. Figure 9 shows the distribution of the pore equivalent diameter of AP particles with mass loss of 9.5%, 20.2% and 29.9%. At 9.5%, there were more small pores, and 31.7% of all pores were below 2 μm. When the mass loss reached 20.2%, those below 2 μm decreased to 21.2%. When the mass loss reached 29.9%, the number of smaller pores was greatly reduced, and the pore distribution looks approximately Gaussian. The peak pore size was about 4 μm.
(a) 9.5% By analysing the characteristics of pore distribution and pore equivalent diameter in the AP particles, it can be seen that in the initial stage of decomposition the pores were mainly concentrated near the particle surface, with most of them being small pores, indicating that the decomposition of AP particles first occurred near the particle surface during low-temperature heating. AP decomposition generates gas, and the gas escapes the surface, leading to the formation of pores. As decomposition progresses, the pores increase in size and expand into the interior of the particles. Typically, when the pores reach approximately 5 μm in diameter they stop increasing; however, the process continues inside the particles until finally, at a mass loss of approximately 30%, the low-temperature decomposition of AP stops. At this point, the interior of the particles is full of irregular pores, with almost all pores connected and relatively evenly distributed.

Specific Surface Area
The porous structure formed by the decomposition of AP at low temperature will lead to an increase in specific surface area, with the ratio of surface area to volume being used as a parameter to evaluate the specific surface area. Based on the analysis of the established 3D pore model, the specific surface area of porous particles formed during decomposition was calculated using image processing and then compared with the AP specific surface area before decomposition. As shown in Table 2, the "Ratio" was the specific surface area of PAP to AP, where PAP is the porous AP formed after thermal decomposition. With an increase in the mass loss, more pores were produced inside the particles, and the specific surface area was greatly increased. When the mass loss reached 29.9%, the specific surface area increased 8.5 times.

Combustion Characteristics
To analyse the effect of the pore structure on combustion performance caused by lowtemperature decomposition of AP, a test was carried out using a homemade 13 mL closed bomb. The schematic diagram of the test system of the closed bomb is shown in Figure 10; the pressure sensor was a PCB113b22, with a range of 34.5 MPa and a frequency of 30 MHz. The test system adopts a pressure trigger, and the threshold is 0.5 MPa; the pressure data for 2.1 ms before trigger and 30 ms after trigger were recorded. The variation in combustion pressure over time in the binary system of porous AP (ordinary AP) and Al (median particle diameter of 10 μm) in a confined space was studied, as shown in Table 3. The effect of the pore structure on the burning rate performance was also analysed. Three tests were repeated for each proportion with a dose of 54 mg.  The experimentally derived pressure-time curve is shown in Figure 11, and after replacing the ordinary AP with porous AP, the P-t curve peak becomes increasingly steeper. When the ratio of porous AP to Al was 80:20, the pressure change was abrupt after ignition, forming a typical shock wave curve, which can be considered an explosion.   Figure 12 shows the relationship between the rate of pressure variation, dP/dt, and pressure, P, with Table 4 showing the peak pressure, peak time and rate of change in peak pressure under different conditions. dP/dt increases significantly when using porous AP instead of normal AP, demonstrating that porous AP can lead to a higher combustion rate, aggravating the reaction intensity and accelerating the combustion speed. The time to reach the peak was shortened, with the peak becoming higher during combustion. When the ratio of AP and Al was 80:20 the pressure peak had increased 2.7 times, with the maximum rate of change in pressure increasing 34 times. Therefore, the degree of reaction was significantly intensified. This is because the porous structure formed by partial decomposition of AP has a larger specific surface area. The porous structure of AP tends to lead to convective combustion to accelerate the combustion rate. When the proportion of porous AP reaches 80%, the combustion is transformed into an explosion.
(a) Ordinary AP (b) Porous AP Figure 12. dp/dt-p curve for ordinary and porous AP.

Conclusions
To summarize, the partial decomposition characteristics of AP at low temperature were studied, combined with nano-CT scanning and evaluation of the mass loss during decomposition. The low-temperature decomposition of AP can be divided into three stages: delay, rapid decomposition and slow reaction at 230 °C, which is the temperature for the slow cook-off of the propellant. The nano-CT results indicate that the decomposition of AP particles first occurs near the surface by forming pores, with the decomposition beginning at multiple points at the same time. The pores increase and expand gradually during the decomposition process. The pores do not expand indefinitely, and stop growing when their equivalent diameter reaches approximately 5 μm. When the mass loss reaches approximately 30%, the decomposition stops, and almost all of the pores become connected with one another. In the binary system of AP and Al, compared with ordinary AP, porous AP can significantly increase the burning rate, resulting in a higher reaction rate and higher reaction intensity.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.