Impact-Induced Reaction Characteristic and the Enhanced Sensitivity of PTFE/Al/Bi2O3 Composites

In this paper, the reaction characteristic of a novel reactive material, which introduced bismuth trioxide (Bi2O3) into traditional polytetrafluoroethylene/aluminum (PTFE/Al), is studied. The effect of Bi2O3 with different content and particle size on the reaction behaviors of PTFE/Al/Bi2O3 are investigated by drop-weight test and X-ray diffractometer (XRD), including impact sensitivity, energy release performance under a certain impact, and reaction mechanism. The experimental results show that the content of Bi2O3 increased from 0% to 35.616%, the characteristic drop height of impact sensitivity (H50) of PTFE/Al/Bi2O3 reactive materials decreased first and then increased, and the minimum H50 of all types of materials in the experiment is 0.74 times that of PTFE/Al, and the particle size of Bi2O3 affects the rate of H50 change with Bi2O3 content. Besides, with the increase of Bi2O3 content, both the reaction intensity and duration first increase and then decrease, and there is optimum content of Bi2O3 maximizing the reaction degree of the PTFE/Al/Bi2O3. Furthermore, a prediction model for the impact sensitivity of PTFE-based reactive material is developed. The main reaction products include AlF3, xBi2O3·Al2O3, and Bi.


Introduction
Reactive materials, a special type of energetic material-including metal/polymer mixture, intermetallic, thermite, hydrides, and matrix material [1,2]-have been widely concerned. Metal/polytetrafluoroethylene (PTFE) as a representative, due to the high energy density and impact-induced reaction characteristics, has broad application prospects. PTFE, whose chemical formula is -(CF 2 -CF 2 )-, is a complicated semi-crystalline material with great stability and chemical inertness at room temperature [3]. It is noteworthy that PTFE has a special property: the as-polymerized melting temperature is about 341 • C, while the subsequent melting temperature is about 328 • C. PTFE has at least four known phases at different pressures and temperatures. Generally speaking, at the high temperature, PTFE mainly presents as pseudohexagonal or orthorhombic crystals, while at the low temperature, PTFE mainly presents as triclinic or hexagonal crystal [4]. At a higher temperature (more than 380 • C), because of PTFE decomposition, the volatile fluorinated gases are released, which have the ability to react with active metals releasing large amounts of heat [5].
In PTFE-based reactive material, PTFE acts not only as a reductant but also as a binder. It can closely wrap particles through high-temperature sintering, and its internal connection is relatively close. Therefore, PTFE-based energetic materials usually have excellent impact properties. Metals, as oxidants, participate in the reaction and release a large amount of heat. The properties of metals directly affect the energy release performance of reactive materials. A variety of metal/PTFE-based According to the above reaction equations, the stoichiometry ratio of PTFE/Al is 73.5:26.5, and the stoichiometry ratio of Al/Bi 2 O 3 is 10.96:89.04. Considering the reaction rate, this work prepared the material formulation with PTFE/Al and Al/Bi 2 O 3 as independent units. PTFE/Al and Al/Bi 2 O 3 were mixed in different proportions to form material with different formulations. A series of samples were prepared, whose information is listed in Table 2. The preparation process mainly includes mixing, cold isostatic pressing, and high-temperature sintering. Firstly, the PTFE, Al, and Bi 2 O 3 raw powders of a certain mass were added to the anhydrous ethanol solution and mixed by a blender for about 40 min, followed by a drying process at room temperature lasting 48 h. Then the mixed powder was placed in a mold with an inner diameter of 10 mm and cold uniaxial pressed at about 250 MPa. Finally, semi-finished samples were placed in a vacuum sintering oven. The oven temperature was raised to 370 • C at a rate of 60 • C/h, then holding at 370 • C for 4.5 h, and cooling to ambient temperature at a rate of 60 • C/h. In the high-temperature sintering process for a long time, PTFE melts and presents a viscoelastic state with a certain degree of fluidity. The polymer chain flows inside the sample to fill the pore formed in the preparation of the sample and improves its compactness. During the cooling process, PTFE recrystallization changes the state of the PTFE matrix in the sample from distributed distribution to integral distribution, which increases PTFE's bonding performance and improves the overall strength of the sample.
As shown in Table 2, with the increase of the content of bismuth oxide, the theoretical density of PTFE/Al/Bi 2 O 3 material theory increases and the energy contained in the unit mass theory decreases. Typically prepared samples (Φ 10 mm × 4 mm) with different formulations are shown in Figure 1. As shown in Figure 1, with the content of Bi 2 O 3 increasing, the color of the sample gradually changed from gray to gray-green. increases PTFE's bonding performance and improves the overall strength of the sample.
As shown in Table 2, with the increase of the content of bismuth oxide, the theoretical density of PTFE/Al/Bi2O3 material theory increases and the energy contained in the unit mass theory decreases. Typically prepared samples (Φ 10 mm × 4 mm) with different formulations are shown in Figure 1. As shown in Figure 1, with the content of Bi2O3 increasing, the color of the sample gradually changed from gray to gray-green.

Experimental Contents
The energy release performance of PTFE/Al/Bi 2 O 3 composites were investigated by a drop-weight device, as shown in Figure 2. In the drop-weight experiment, the test sample is placed in the center of the anvil and directly under the drop-weight. The drop-weight falls freely from a certain height H guided by two rods to ensure a planar impact. The test sample is compressed and reacts under the action of the compression load. The drop-weight (mass of 10 kg) carries the upper anvil drops from a height of up to 2 m. The impact sensitivities of each type are compared by the characteristic drop height (H 50 ), at which the material has a 50% probability of reaction. The test method adopts the well-known "up-and-down technique" [15]. The tests for each type of material were performed 28 times and at 5 cm intervals. In order to study the effect of Bi 2 O 3 on energy release performance, including reaction intensity, ignition delay time, and reaction duration, the drop-weight was dropped from the same height to impact the materials. The content of Al/Bi2O3 is 0, 5%, 10%, 20%, 30% and 40%, respectively.

Experimental Contents
The energy release performance of PTFE/Al/Bi2O3 composites were investigated by a dropweight device, as shown in Figure 2. In the drop-weight experiment, the test sample is placed in the center of the anvil and directly under the drop-weight. The drop-weight falls freely from a certain height H guided by two rods to ensure a planar impact. The test sample is compressed and reacts under the action of the compression load. The drop-weight (mass of 10 kg) carries the upper anvil drops from a height of up to 2 m. The impact sensitivities of each type are compared by the characteristic drop height (H50), at which the material has a 50% probability of reaction. The test method adopts the well-known "up-and-down technique" [15]. The tests for each type of material were performed 28 times and at 5 cm intervals. In order to study the effect of Bi2O3 on energy release performance, including reaction intensity, ignition delay time, and reaction duration, the dropweight was dropped from the same height to impact the materials. In order to record the sequences of the sample reaction behaviors in the drop-weight test, a Phantom V710 high-speed photography (Vision Research, Inc., Wayne, NJ, USA) was employed. The selected frame rate was 20,000 per second, so that a frame was taken every 50 μs. The resolution was 640 × 480 pixels and the exposure time was set to 10 μs. These settings were selected based on early testing and represent an optimal tradeoff between available lighting and the minimization of blur in the images.
The X-ray diffractometer (XRD, BRUKER D8 ADVANCE, Bruker, Karlsruhe Germany) was used to detect the reacted sample after the drop-weight test to analyze the reaction product of PTFE/Al/Bi2O3. The instrument parameters settings were as follows: the tube voltage was 40 Kv; the current was 40 Ma; Cu Kα radiation (λ = 0.15406 nm) was selected; the scanning range 2θ was 10°-90°; and the scanning speed was 5° min −1 .

Mesoscale Characteristics
In order to study the mesoscale characteristics of PTFE/Al/Bi2O3 composite material, scanning electron microscope (SEM, HITACHI S-4800, CamScan, Tokyo, Japan) was used. The mesoscale characteristics are shown in Figure 3. Figure 3a indicates that in the PTFE/Al/Bi2O3 composite, the In order to record the sequences of the sample reaction behaviors in the drop-weight test, a Phantom V710 high-speed photography (Vision Research, Inc., Wayne, NJ, USA) was employed. The selected frame rate was 20,000 per second, so that a frame was taken every 50 µs. The resolution was 640 × 480 pixels and the exposure time was set to 10 µs. These settings were selected based on early testing and represent an optimal tradeoff between available lighting and the minimization of blur in the images.
The X-ray diffractometer (XRD, BRUKER D8 ADVANCE, Bruker, Karlsruhe Germany) was used to detect the reacted sample after the drop-weight test to analyze the reaction product of PTFE/Al/Bi 2 O 3 . The instrument parameters settings were as follows: the tube voltage was 40 Kv; the current was 40 Ma; Cu K α radiation (λ = 0.15406 nm) was selected; the scanning range 2θ was 10 • -90 • ; and the scanning speed was 5 • min −1 .

Mesoscale Characteristics
In order to study the mesoscale characteristics of PTFE/Al/Bi 2 O 3 composite material, scanning electron microscope (SEM, HITACHI S-4800, CamScan, Tokyo, Japan) was used. The mesoscale characteristics are shown in Figure 3. Figure 3a indicates that in the PTFE/Al/Bi 2 O 3 composite, the PTFE matrix closely wrapped the Al and Bi 2 O 3 particles. Figure 3b-d show the distribution of elements in the sample. F, Al, and Bi elements represent the distribution of PTFE matrix, Al, and Bi 2 O 3 in the sample, respectively. The distribution of elements also shows that the Al and Bi 2 O 3 particles are uniformly distributed in the PTFE matrix and the particle and matrix are tightly bound.
Polymers 2019, 11, x FOR PEER REVIEW 5 of 17 elements in the sample. F, Al, and Bi elements represent the distribution of PTFE matrix, Al, and Bi2O3 in the sample, respectively. The distribution of elements also shows that the Al and Bi2O3 particles are uniformly distributed in the PTFE matrix and the particle and matrix are tightly bound.  Figure 4 presents the XRD pattern of PTFE/Al/Bi2O3 samples (Type D-1). As shown in Figure 4, it indicates that PTFE, Al, and Bi2O3 existed in the samples, and no new substances had formed. It means that no chemical reaction occurred in the sample during mixing and sintering, and only physical mixing occurred among the components. The components within the PTFE/Al/Bi2O3 composite still retained their original physical and chemical properties. Therefore, the physical and chemical properties of composite materials can be estimated from the properties of three components. Take the burning rate as an example, according to the superposition principle of burning rate of multi-component materials, the burning rate of Type F is about 253.6 m/s, about 127 times of that of Type A [13].   Figure 4, it indicates that PTFE, Al, and Bi 2 O 3 existed in the samples, and no new substances had formed. It means that no chemical reaction occurred in the sample during mixing and sintering, and only physical mixing occurred among the components. The components within the PTFE/Al/Bi 2 O 3 composite still retained their original physical and chemical properties. Therefore, the physical and chemical properties of composite materials can be estimated from the properties of three components. Take the burning rate as an example, according to the superposition principle of burning rate of multi-component materials, the burning rate of Type F is about 253.6 m/s, about 127 times of that of Type A [13].
physical mixing occurred among the components. The components within the PTFE/Al/Bi2O3 composite still retained their original physical and chemical properties. Therefore, the physical and chemical properties of composite materials can be estimated from the properties of three components. Take the burning rate as an example, according to the superposition principle of burning rate of multi-component materials, the burning rate of Type F is about 253.6 m/s, about 127 times of that of Type A [13].

Impact Sensitivity
The impact sensitivity of PTFE/Al/Bi 2 O 3 materials with different content was studied by the drop-weight test and compared by H 50 . The drop-weight experiments were recorded as shown in

Impact Sensitivity
The impact sensitivity of PTFE/Al/Bi2O3 materials with different content was studied by the drop-weight test and compared by H50. The drop-weight experiments were recorded as shown in     According to the above drop-weight test results, H50 of each type can be calculated as the following Equation: According to the above drop-weight test results, H 50 of each type can be calculated as the following Equation: where A is the lowest height in the test; B is the increment of height, B = 5 cm in this test; D is the number of reactions that occur in the test; i is the ordinal number that falls from low to high, starting from 0; and C i is the number of times that the samples react at the certain height. By Equation (3), the H 50 of each type is listed in Table 3. The variation trend of H 50 of PTFE/Al/Bi 2 O 3 with the content of Bi 2 O 3 is shown in Figure 7. where A is the lowest height in the test; B is the increment of height, B = 5 cm in this test; D is the number of reactions that occur in the test; i is the ordinal number that falls from low to high, starting from 0; and Ci is the number of times that the samples react at the certain height. By Equation (3), the H50 of each type is listed in Table 3. The variation trend of H50 of PTFE/Al/Bi2O3 with the content of Bi2O3 is shown in Figure 7.  The results show that the content of Bi2O3 has a significant influence on the impact sensitivity of PTFE/Al/Bi2O3 composites. Type A composite has the highest H50 reaching 100 cm, compared with composite adding the content of Bi2O3 from 4.452% to 35.616%. In detail, the H50 of PTFE/Al/Bi2O3 composite adding 75 μm-Bi2O3 decreases with the increase of the content of Bi2O3 ranging from 0% to 17.808%, while the H50 of the composite increases with the increase of the content of Bi2O3 ranging from 17.808% to 35.616%. The Type C-1 composite has the highest impact sensitivity, and the H50 of Type C-1 only 0.77 times of that of Type A. When the particle size of Bi2O3 is 150 μm, the H50 varies The results show that the content of Bi 2 O 3 has a significant influence on the impact sensitivity of PTFE/Al/Bi 2 O 3 composites. Type A composite has the highest H 50 reaching 100 cm, compared with . The Type C-1 composite has the highest impact sensitivity, and the H 50 of Type C-1 only 0.77 times of that of Type A. When the particle size of Bi 2 O 3 is 150 µm, the H 50 varies with the Bi 2 O 3 content, which is similar to that of Bi 2 O 3 particle size 75 µm.
The polynomial fitting result of H 50 with the change of content in Figure 7 shows that the particle size of Bi 2 O 3 has no fundamental influence on the general trend of H 50 with the change of content, but it has significant influence on the change rate of H 50 with the change of content, especially when the content of Bi 2 O 3 is higher than about 20%. It is worth noting that the minimum H 50 of PTFE/Al/ Bi 2 O 3 with large particle size Bi 2 O 3 is lower than that with small particle size Bi 2 O 3 . The reasons may be that Bi 2 O 3 particle size affects the failure behavior of PTFE/Al/Bi 2 O 3 composites under the compression, thus affecting the reaction performance of the composites.
The material is subjected to strong compression and failure, and the particles rub against each other, forming 'hot spots'. The addition of Bi 2 O 3 in the PTFE/Al system affects the impact sensitivity mainly from the following two aspects. As the content of Bi 2 O 3 increases, on the one hand, the number of hot spots in the material increases, making the reaction easier to be triggered, leading to the material impact sensitivity increase. On the other hand, in the PTFE/Al/Bi 2 O 3 , the contact area of active PTFE/Al may be reduced due to the separation caused by Bi 2 O 3 particles, and a portion of the impact energy is transferred to the Bi 2 O 3 particles, leading to the material impact sensitivity reduction.

Prediction Model for PTFE-Based Reactive Material Impact Sensitivity
In order to analyze the influence of Bi 2 O 3 content on impact sensitivity of PTFE-based reactive material, the model based on the 'hot spots' theory, is presented to predict the H 50 of materials. Two reactions, PTFE/Al and Al/Bi 2 O 3 , occurred in the PTFE/Al/Bi 2 O 3 composite material. In order to simplify the analysis, the following assumptions are made: (1) all particles in the material are distributed randomly and uniformly and tightly packed by the PTFE matrix without pores; (2) particles are spherical with average size.
As analyzed in Section 3.2, the content of Bi 2 O 3 affects the impact sensitivity of materials from two aspects: (1) as the number of Bi 2 O 3 particle increases, the number of hot spots formed by slide friction increases, which benefits to induce reaction; (2) as proportion of Al/Bi 2 O 3 reactants increases, the reaction threshold of PTFE/Al/Bi 2 O 3 composite increases, which impedes the reaction.
When the hot spots formed by interparticle slip is the dominant mechanism of induced reaction, the following three conditions are necessary: (1) the influence range of the friction hot spots includes the reactants; (2) the reaction temperature threshold is lower than the melting point of the friction particles; (3) the temperature of hot spot is higher than reaction threshold temperature. If all conditions are met, the material will be ignited. It is worth noting that due to the small size of the hot spot, the influence range of hot spot temperature is the particles involved in the friction and PTFE matrix. Therefore, the concept of effective hot spot is proposed, which refers to the hot spot that can induce the reaction. In PTFE/Al/Bi 2 O 3 composite material, effective friction hot spot refers to the hot spot formed by friction between Al and Al and Bi 2 O 3 , respectively. This is because the melting point of Al is only 933 K below the reaction threshold of Al and Bi 2 O 3 . Hot spots only could induce the reaction between Al and PTFE.
The temperature rise due to slide friction between particles is expressed as [16] ∆T = µLuπ 1/2 where ∆T is temperature rise caused by friction between particles, k 1 and k 2 are thermal conductivity of two particles, a is the radius of actual contact area, µ is the coefficient of friction, L is normal pressure, and u is relative velocity, respectively. It assumes that the average relative velocity between particles is equal to half of the drop mass velocity in the drop-weight test. Then, u when the drop height is H 50 can be expressed: where g is the gravitational acceleration which is 9.8 m·s -2 . The friction between two particles during mechanical impact belongs to a random independent event, and the friction probability of Al/Al and Al/Bi 2 O 3 fraction can be expressed as v 2 Al and 2v Al v Bi 2 O 3 by the volume fraction, respectively. Considering the uneven temperature and proportion of two-type hot spots, the weighted average temperature rises because of hot spots, and can be approximately expressed as where v Al and v Bi 2 O 3 are volume fraction of Al and Bi 2 O 3 , respectively. ∆T f is temperature rise. Assume the temperature rise caused by adiabatic shear and heating at creak tips mechanism is T other , ignoring variation because of Bi 2 O 3 content, the hot spot temperature of the material under the action of a drop mass can be expressed as where T, T 0 are ultimate and initial temperature of hot spot, respectively. k Al and k Bi 2 O 3 are thermal conductivity of Al and Bi 2 O 3 , respectively. Hot spot induced material reaction is not only related to hot spot temperature but also affected by hot spot size. The size of hot spots strongly depends on the chemical reaction rate. When the reaction rate k is higher than the critical reaction rate k c , the hot spot size satisfies the starting condition. According to Arrhenius formula, the rate of the chemical reaction can be expressed as where A is the pre-exponential factor; E a is the active energy; R is the gas content; and T is the temperature, respectively. When k > k c , that is to say, the hot spot size reaches the reaction condition, which is reflected as the chemical reaction temperature threshold at the macro level. The reaction temperature threshold can be expressed as Since constants (A, E a , and k c ) are affected by multiple factors, such as particle size and shape, it is difficult to get detailed data, so this paper directly introduces the reaction temperature threshold to consider the judgment of reaction parameters (A, E a , and k c ) on whether the reaction occurs or not. When PTFE-based granule composites are impacted, the particles within the materials are randomly rubbed against each other. Therefore, the reaction threshold temperature of PTFE/Al/Bi 2 O 3 composite with different Bi 2 O 3 content can be estimated as where T c is the reaction threshold temperature of composite material and ω Al/Bi 2 O 3 and ω PTFE/Al are the mass fraction of the reactant Al/Bi 2 O 3 , and PTFE/Al, respectively. T Al/PTFE and T Al/Bi 2 O 3 are the reaction threshold temperature of PTFE/Al and Al/ Bi 2 O 3 , respectively. When T ≥ T c , the material reaction is triggered. Combined with Equations (5), (7), and (10), the H 50 of PTFE/Al/Bi 2 O 3 composite material with the content of Bi 2 O 3 particles can be expressed as The parameters used in the model are listed in Table 4 [17]. Figure 8 depicts the calculated results of Equation (11). Table 4. The parameters in the model. The model analysis shows that when the content of Bi2O3 is low, the number and temperature of hot spots dominate H50 of the PTFE/Al/Bi2O3 composite materials, leading to the decline of the H50 with the increase of Bi2O3 content. When the content of Bi2O3 is high, the ignition threshold dominates H50, leading to the improvement of H50 with the increase of Bi2O3 content.
On the whole, the above model can predict the change law of H50 with Bi2O3 content, but compared with the experimental value, the model predicts the slow change rate of H50 with Bi2O3 content, which is because the model only considers the influence of friction hot spots and ignores the change of temperature rise under the influence of content induced by other mechanisms. When the Bi2O3 content was less than 5%, the model predict lower than the experimental results. The main reason is that friction is not the dominant mechanism of hot spot formation in the low particle content composite material. As PTFE matrix content increases, the material shows good toughness, and the temperature rise caused by adiabatic shear and heating at the crack tip is underestimated, so the prediction results of the model are lower than the experimental results. When the Bi2O3 content is in the range of 5%-30%, the model prediction is higher than the experimental results. According to the mechanical characteristic analysis of other PTFE/Al materials, the failure strain of the composite material in the range of particle content increases to the maximum and then declines, and the failure strain affects the exothermic change of the crack tip, leading to the prediction model error. When the Bi2O3 content is higher than about 30%, the model prediction is slightly better than the experimental composite, because friction contributes more to the formation of hot spots in the composite with high particle content. On the whole, the above model can predict the change law of H 50 with Bi 2 O 3 content, but compared with the experimental value, the model predicts the slow change rate of H 50 with Bi 2 O 3 content, which is because the model only considers the influence of friction hot spots and ignores the change of temperature rise under the influence of content induced by other mechanisms. When the Bi 2 O 3 content was less than 5%, the model predict lower than the experimental results. The main reason is that friction is not the dominant mechanism of hot spot formation in the low particle content composite material. As PTFE matrix content increases, the material shows good toughness, and the temperature rise caused by adiabatic shear and heating at the crack tip is underestimated, so the prediction results of the model are lower than the experimental results. When the Bi 2 O 3 content is in the range of 5%-30%, the model prediction is higher than the experimental results. According to the mechanical characteristic analysis of other PTFE/Al materials, the failure strain of the composite material in the range of particle content increases to the maximum and then declines, and the failure strain affects the exothermic change of the crack tip, leading to the prediction model error. When the Bi 2 O 3 content is higher than about 30%, the model prediction is slightly better than the experimental composite, because friction contributes more to the formation of hot spots in the composite with high particle content.   The content of Bi2O3 has a significant influence on the reaction duration. The reaction duration of the sample without Bi2O3 (Type A) is about 4.2 ms. In general, as the content of Bi2O3 increases, the reaction duration is prolonged; while, when the content of Bi2O3 exceeds a certain value, the reaction duration is shortened. When the particle size of Bi2O3 is 75 μm, there is no significant difference in reaction duration (about 5.2 ms) in range from 4.452% to 17.808%. When the content exceeds 17.808%, the reaction duration shortens. For the samples with 150 μm-Bi2O3, when the content of Bi2O3 is 8.904%, the reaction duration reaches about 6.3 ms at most. As the content of Bi2O3 continues to increase, the reaction duration of samples shortens. It can be illustrated from reactions, which occur in the PTFE/Al/Bi2O3 composite materials, PTFE/Al reacts at a lower temperature, but the burning rate of PTFE/Al is slow and the reaction has significant non-self-sustainability. Al/Bi2O3 reacts at a higher critical temperature, comparing with the PTFE/Al, and the burning rate of Al/Bi2O3 is faster. In the PTFE/Al/Bi2O3 composite, the PTFE/Al reaction occurs first, and releases a large amount of heat. A large amount of heat will induce the reaction between Al and Bi2O3. A large amount of heat released by the reaction Al/Bi2O3 and the high burning rate is conducive to the reaction, improving the self-sustainability of reaction. Thus, the reaction duration of the material is prolonged. However, the increase of Bi2O3 content means that the content of PTFE/Al decreases, and the heat released by the reaction of PTFE/Al decreases, which is not conducive to induce the secondary reaction of Al/Bi2O3, leading to more incomplete reaction and shorter reaction duration.

Energy Release Performance under a Certain Impact
The state of two samples, Type A (without Bi2O3) and Type D-1 (with 75 μm-Bi2O3), after impact, are shown in Figure 11. There is a large amount of carbon black remaining on the anvil after the reaction of Type A sample, while less carbon black after reaction of Type D-1 sample. The amorphous carbon generated by the reaction is mainly distributed on the edge of the residue sample. Besides, The content of Bi 2 O 3 has a significant influence on the reaction duration. The reaction duration of the sample without Bi 2 O 3 (Type A) is about 4.2 ms. In general, as the content of Bi 2 O 3 increases, the reaction duration is prolonged; while, when the content of Bi 2 O 3 exceeds a certain value, the reaction duration is shortened. When the particle size of Bi 2 O 3 is 75 µm, there is no significant difference in reaction duration (about 5.2 ms) in range from 4.452% to 17.808%. When the content exceeds 17.808%, the reaction duration shortens. For the samples with 150 µm-Bi 2 O 3 , when the content of Bi 2 O 3 is 8.904%, the reaction duration reaches about 6.3 ms at most. As the content of Bi 2 O 3 continues to increase, the reaction duration of samples shortens. It can be illustrated from reactions, which occur in the PTFE/Al/Bi 2 O 3 composite materials, PTFE/Al reacts at a lower temperature, but the burning rate of PTFE/Al is slow and the reaction has significant non-self-sustainability. Al/Bi 2 O 3 reacts at a higher critical temperature, comparing with the PTFE/Al, and the burning rate of Al/Bi 2 O 3 is faster. In the PTFE/Al/Bi 2 O 3 composite, the PTFE/Al reaction occurs first, and releases a large amount of heat. A large amount of heat will induce the reaction between Al and Bi 2 O 3 . A large amount of heat released by the reaction Al/Bi 2 O 3 and the high burning rate is conducive to the reaction, improving the self-sustainability of reaction. Thus, the reaction duration of the material is prolonged. However, the increase of Bi 2 O 3 content means that the content of PTFE/Al decreases, and the heat released by the reaction of PTFE/Al decreases, which is not conducive to induce the secondary reaction of Al/Bi 2 O 3 , leading to more incomplete reaction and shorter reaction duration.
The state of two samples, Type A (without Bi 2 O 3 ) and Type D-1 (with 75 µm-Bi 2 O 3 ), after impact, are shown in Figure 11. There is a large amount of carbon black remaining on the anvil after the reaction of Type A sample, while less carbon black after reaction of Type D-1 sample. The amorphous carbon generated by the reaction is mainly distributed on the edge of the residue sample. Besides, carbon also exists on the surface of the Type A sample. The phenomenon indicates that Type D-1 reacts more thoroughly than Type A under the drop mass height of 140 cm. That is probably because Bi 2 O 3 could react with amorphous carbon of PTFE/Al reaction product at high temperature [18], so the reaction degree is improved and the carbon black reduces. Compared with Type D-1 residue, Type A residue is more complete, which also indicates that the reaction rate of Type A is lower than that of Type D-1, and opportune Bi 2 O 3 content is conducive to the reaction.

Reaction Mechanism
In order to understand the chemical reaction mechanism of the PTFE/Al/Bi2O3 composites, the reacted residue of Type D-1 sample is analyzed by X-ray diffraction (XRD), and the result is shown in Figure 12. The results show that AlF3, Bi24Al2O39, and Bi48Al2O75 are produced during the reaction. Among them, AlF3 is the product of the reaction between Al and gaseous C2F4 generated by PTFE cracking. In addition, the reaction also generates amorphous carbon, which could not be detected by XRD. Bi24Al2O39 and Bi48Al2O75 are two kinds of mixed crystals formed by Bi2O3 and Al2O3 at high temperature. Bi2O3 reacts with Al to form Al2O3 and Bi, and unreacted melted Bi2O3 (melting point: 1098 K) wraps Al2O3 to form xBi2O3·Al2O3. Bi, the other product of the reaction between Al and Bi2O3, is also not detected in the product. It is because that, Bi, whose boiling point is only 1833 K, forms Bi vapor due to the high temperature during the reaction process. The formation of xBi2O3·Al2O3 crystal also indicates a small Al and Bi2O3 participation reaction. It implies that the materials react incompletely under the low impact load. The analysis in Section 3.4 also indicates that C and Bi2O3 may have subsequent reactions. Combining the analysis above, the possible chemical reaction process of the PTFE/Al/Bi2O3 composite can be described as:  Figure 11. Recovered residues and reaction vestiges.

Reaction Mechanism
In order to understand the chemical reaction mechanism of the PTFE/Al/Bi 2 O 3 composites, the reacted residue of Type D-1 sample is analyzed by X-ray diffraction (XRD), and the result is shown in Figure 12. The results show that AlF 3 , Bi 24 Al 2 O 39 , and Bi 48 Al 2 O 75 are produced during the reaction. Among them, AlF 3 is the product of the reaction between Al and gaseous C 2 F 4 generated by PTFE cracking. In addition, the reaction also generates amorphous carbon, which could not be detected by XRD 3 . Bi, the other product of the reaction between Al and Bi 2 O 3 , is also not detected in the product. It is because that, Bi, whose boiling point is only 1833 K, forms Bi vapor due to the high temperature during the reaction process. The formation of xBi 2 O 3 ·Al 2 O 3 crystal also indicates a small Al and Bi 2 O 3 participation reaction. It implies that the materials react incompletely under the low impact load. The analysis in Section 3.4 also indicates that C and Bi 2 O 3 may have subsequent reactions. Combining the analysis above, the possible chemical reaction process of the PTFE/Al/Bi 2 O 3 composite can be described as: Al + C 2 F 4 → AlF 3 + C  Figure 12. X-ray diffraction pattern of the reaction residue of Type D-1. Figure 12. X-ray diffraction pattern of the reaction residue of Type D-1.
The above analysis shows that when PTFE/Al/Bi 2 O 3 composite material is impacted, not only does it react with PTFE and Bi 2 O 3 , respectively, with Al as the oxidant, but also PTFE in the composite reacts with Bi as the oxidant; in addition, Bi 2 O 3 may further oxidize reaction product C. Therefore, Bi 2 O 3 can effectively improve the energy release characteristics of PTFE/Al-based energetic materials.
In general, the PTFE/Al/Bi 2 O 3 , a kind of granular composite reactive materials, would be initiated due to the 'hot spots', which are probably formed by the sliding friction, adiabatic shear, and heating at crack tips, during mechanical impact. Figure 13 presents the residue of Type D, including reacted and unreacted samples. There is a remarkable difference between reacted and unreacted sample residue.
There are several open cracks in the edge of the reacted residue, some of which have black marks and obvious melted marks (Figure 13a,c). While there are only open cracks at the unreacted residue edge (Figure 13b,d). These phenomena imply that the reaction is most likely to begin at the open crack in the material and that the formation of the crack is related to the material reaction. The black marks show that the sample reacts incompletely. The edge open cracks of Type D-2 residue are more obvious than that of Type D-1, indicating that the ductility of the material adding large particle size Bi 2 O 3 is weaker than that of adding small particle size Bi 2 O 3 material (Figure 13b,d). The above analysis shows that when PTFE/Al/Bi2O3 composite material is impacted, not only does it react with PTFE and Bi2O3, respectively, with Al as the oxidant, but also PTFE in the composite reacts with Bi as the oxidant; in addition, Bi2O3 may further oxidize reaction product C. Therefore, Bi2O3 can effectively improve the energy release characteristics of PTFE/Al-based energetic materials.
In general, the PTFE/Al/Bi2O3, a kind of granular composite reactive materials, would be initiated due to the 'hot spots', which are probably formed by the sliding friction, adiabatic shear, and heating at crack tips, during mechanical impact. Figure 13 presents the residue of Type D, including reacted and unreacted samples. There is a remarkable difference between reacted and unreacted sample residue. There are several open cracks in the edge of the reacted residue, some of which have black marks and obvious melted marks (Figure 13a,c). While there are only open cracks at the unreacted residue edge (Figure 13b,d). These phenomena imply that the reaction is most likely to begin at the open crack in the material and that the formation of the crack is related to the material reaction. The black marks show that the sample reacts incompletely. The edge open cracks of Type D-2 residue are more obvious than that of Type D-1, indicating that the ductility of the material adding large particle size Bi2O3 is weaker than that of adding small particle size Bi2O3 material ( Figure  13b,d).  Figure 14 shows the microscopic characteristics of the cracks in the reacted and unreacted samples. There are no reaction products, such as carbon and bismuth, in shear cracks of unreacted residues. There are fibers, whose direction is perpendicular to that of crack that can be observed clearly, from unreacted residue SEM images. The morphology of crack is similar to that of brittle fracture. Morphology features of unreacted residue are significantly different from those of reacted residue. The crack edge of the sample in which the reaction occurred was irregular and appeared coral-like, which is formed by melted and recrystallization of PTFE in the reaction zone during the reaction. It can be seen from Figure 14b that when the material reacts partially, the reaction zone is generally located at the crack, which also indicates that exothermic heat at the crack tip is an  Figure 14 shows the microscopic characteristics of the cracks in the reacted and unreacted samples. There are no reaction products, such as carbon and bismuth, in shear cracks of unreacted residues. There are fibers, whose direction is perpendicular to that of crack that can be observed clearly, from unreacted residue SEM images. The morphology of crack is similar to that of brittle fracture. Morphology features of unreacted residue are significantly different from those of reacted residue. The crack edge of the sample in which the reaction occurred was irregular and appeared coral-like, which is formed by melted and recrystallization of PTFE in the reaction zone during the reaction. It can be seen from Figure 14b that when the material reacts partially, the reaction zone is generally located at the crack, which also indicates that exothermic heat at the crack tip is an important factor for ignition. Combined with the results in Section 3.2, it indicates that the mechanism of PTFE/Al/Bi2O3 composite impact-induced reaction could be described as: (1) under mechanical impact, violent plastic deformation of the material results in the rise of the overall temperature of the material; (2) during the intense compression of granular composite material, violent friction occurs, resulting in hot spots and local temperature rise, which are randomly distributed with in the composite; (3) under the action of strong shear, cracks are formed at the edge of the material, and heat is released from the crack tip, making the crack tip to form hot spots. Considering the above three aspects, the hot spot at the crack tip of the material has the highest temperature and is the easiest to trigger the reaction.

Conclusions
In this paper, Bi2O3 was introduced into PTFE/Al reactive composite material. The reaction behaviors of this novel material was studied from energy release performance and reaction mechanism by the drop-weight test, XRD, and SEM. The main conclusions drawn are as follows: (1) The particles are uniformly distributed in PTFE/Al/Bi2O3 composite and bonded well with the PTFE matrix. The properties of PTFE/Al/Bi2O3 can be estimated from the properties of the components.
(2) Bi2O3 has a significant influence on PTFE/Al/Bi2O3 composite impact sensitivity. The content of Bi2O3 increased from 0% to 35.616%, the impact sensitivity of PTFE/Al/Bi2O3 composite increase first and then decrease, and the lowest H50 (Type C-2, 10%, 150 μm) was 74.04 cm, 0.74 times that of PTFE/Al. The model for predicting the PTFE-based reactive material impact sensitivity is presented, which is in good agreement with the experimental results. (3) Energy release performance of PTFE/Al/Bi2O3 strongly depend on the Bi2O3 content. At the drop height of 140 cm, as the content of Bi2O3 increases, the reaction intensity and duration first increases and then decreases. The maximum reaction duration (Type C-2, 10%, 150 μm) is about 6.3 ms, which is 1.5 times that of PTFE/Al. (4) The main reaction products include AlF3, xBi2O3·Al2O3, and Bi. In PTFE/Al/Bi2O3, Bi2O3, as a reductant, reacts with Al and C, which significantly improves the energy release characteristics of materials. The PTFE in the reaction zone is coral-like, which is significantly different from the morphology of the unreacted zone.  Combined with the results in Section 3.2, it indicates that the mechanism of PTFE/Al/Bi 2 O 3 composite impact-induced reaction could be described as: (1) under mechanical impact, violent plastic deformation of the material results in the rise of the overall temperature of the material; (2) during the intense compression of granular composite material, violent friction occurs, resulting in hot spots and local temperature rise, which are randomly distributed with in the composite; (3) under the action of strong shear, cracks are formed at the edge of the material, and heat is released from the crack tip, making the crack tip to form hot spots. Considering the above three aspects, the hot spot at the crack tip of the material has the highest temperature and is the easiest to trigger the reaction.

Conclusions
In this paper, Bi 2 O 3 was introduced into PTFE/Al reactive composite material. The reaction behaviors of this novel material was studied from energy release performance and reaction mechanism by the drop-weight test, XRD, and SEM. The main conclusions drawn are as follows: