Energetic Films Realized by Encapsulating Copper Azide in Silicon-Based Carbon Nanotube Arrays with Higher Electrostatic Safety

Since copper azide (Cu(N3)2) has high electrostatic sensitivity and is difficult to be practically applied, silicon-based Cu(N3)2@carbon nanotubes (CNTs) composite energetic films with higher electrostatic safety were fabricated, which can be compatible with micro-electro mechanical systems (MEMS). First, a silicon-based porous alumina film was prepared by a modified two-step anodic oxidation method. Next, CNTs were grown in pores of the silicon-based porous alumina film by chemical vapor deposition. Then, copper nanoparticles were deposited in CNTs by electrochemical deposition and oxidized to Cu(N3)2 by gaseous hydrogen azide. The morphology and composition of the prepared silicon-based Cu(N3)2@CNTs energetic films were characterized by field emission scanning electron microscopy (FESEM), transmission electron microscopy (TEM) and X-ray diffraction (XRD), respectively. The electrostatic sensitivity of the composite energetic film was tested by the Bruceton method. The thermal decomposition kinetics of the composite energetic films were studied by differential scanning calorimetry (DSC). The results show that the exothermic peak of the silicon-based Cu(N3)2@CNTs composite energetic film is at the temperature of 210.95 °C, its electrostatic sensitivity is significantly less than that of Cu(N3)2 and its 50% ignition energy is about 4.0 mJ. The energetic film shows good electric explosion characteristics and is successfully ignited by laser.


Introduction
In recent years, micro-electro mechanical systems (MEMS)-related technologies have been widely used to fabricate miniature safe systems and micro-initiators [1][2][3][4]. The currently used lead-based primary explosives can hardly meet the requirements of the micro-initiator for energy due to the miniaturization of the ignition device. There is an urgent need to find an environmentally friendly substitute with higher safety and energy density [5][6][7][8][9][10]. Energetic films can be integrated directly on the ignitor without manipulating dangerous reactants [11,12], which makes the process safe and attractive for commercial products. Copper azide (Cu(N 3 ) 2 ) is a promising alternative to lead-containing primary explosives. Compared with lead azide and lead styphnate, the reaction product of copper azide has little harm to the environment. Hence, it is a green and environmentally friendly primary explosive [13][14][15]. In addition, copper azide has excellent initiation performance, and its limit charge of initiating pentaerythritol tetranitrate (PETN) is just 1/6 of that of lead azide [16,17]. In MEMS beam hit the evaporation material and the film began to deposit. The Ti layer served as a transition layer to increase the strength of the adhesion between the Al layer and the silicon substrate.
An improved two-step anodic oxidation method [28] was used to prepare the silicon-based porous alumina film, the device is shown in Figure 1. The preparation process can mainly be divided into three steps: porous alumina film, the device is shown in Figure 1. The preparation process can mainly be divided into three steps: First, 0.3 mol/L oxalic acid was used as the electrolyte, and ethanol/water (v/v = 1:1) as the electrolyte solvent; we kept the temperature of the electrolyte at 1-3 °C, and increased the voltage from 0 to 100 V; the speed of magnetic stirring was 430 r·min −1 , and the time of the first step of anodizing the silicon-based aluminum layer was 40 min; Next, the sample was immersed in a mixed solution of 1.8% CrO3 and 6% phosphoric acid at 65 °C for 3 h to remove the alumina film formed, and then the second step of anodizing was performed, the electrolyte was 0.5% phosphoric acid solution; Finally, the sample was immersed in a 5% phosphoric acid solution for 40 min at 30 °C. The purpose of this step was to enlarge the diameter of the pores. An aligned carbon nanotube array was prepared in a silicon-based porous alumina film channel using a chemical vapor deposition method. The silicon-based porous alumina film was placed in a tube furnace, and the temperature was raised to 700 °C under vacuum conditions, while a mixed gas of acetylene, hydrogen and argon was passed in. The vacuum degree was kept at 0.05 MPa, and the deposition time was 90 min. Then, the copper nanoparticles were deposited in the aligned carbon nanotubes by an electrochemical deposition method. The composite film was the working electrode and the platinum electrode was the counter electrode. The electrolytic solution was 0.08 mol/L copper sulfate pentahydrate and 0.4 mol/L boric acid, then thiourea or polyvinylpyrrolidone (PVP) were added as an additive and the current density was 0.1 mA·cm −2 . The effects of different additives on the deposition of copper nanoparticles in the carbon nanotubes were explored. Finally, the sample was immersed in a 0.5 mol/L sodium hydroxide solution for 1 h to remove the aluminum oxide film. We left it in the air for 24 h, and then the oxidation of the sample was characterized.
At normal temperature, nitric acid reacts with sodium azide to produce hydrogen azide gas. The gas-solid reaction of copper with hydrogen azide can produce copper azide. The reaction equation involved in the above process is as follows: HNO3 + NaN3→HN3 + NaNO3 (1) Cu + 3HN3→Cu(N3)2 + NH3 + N2 (2) First, 0.3 mol/L oxalic acid was used as the electrolyte, and ethanol/water (v/v = 1:1) as the electrolyte solvent; we kept the temperature of the electrolyte at 1-3 • C, and increased the voltage from 0 to 100 V; the speed of magnetic stirring was 430 r·min −1 , and the time of the first step of anodizing the silicon-based aluminum layer was 40 min; Next, the sample was immersed in a mixed solution of 1.8% CrO 3 and 6% phosphoric acid at 65 • C for 3 h to remove the alumina film formed, and then the second step of anodizing was performed, the electrolyte was 0.5% phosphoric acid solution; Finally, the sample was immersed in a 5% phosphoric acid solution for 40 min at 30 • C. The purpose of this step was to enlarge the diameter of the pores.
An aligned carbon nanotube array was prepared in a silicon-based porous alumina film channel using a chemical vapor deposition method. The silicon-based porous alumina film was placed in a tube furnace, and the temperature was raised to 700 • C under vacuum conditions, while a mixed gas of acetylene, hydrogen and argon was passed in. The vacuum degree was kept at 0.05 MPa, and the deposition time was 90 min. Then, the copper nanoparticles were deposited in the aligned carbon nanotubes by an electrochemical deposition method. The composite film was the working electrode and the platinum electrode was the counter electrode. The electrolytic solution was 0.08 mol/L copper sulfate pentahydrate and 0.4 mol/L boric acid, then thiourea or polyvinylpyrrolidone (PVP) were added as an additive and the current density was 0.1 mA·cm −2 . The effects of different additives on the deposition of copper nanoparticles in the carbon nanotubes were explored. Finally, the sample was immersed in a 0.5 mol/L sodium hydroxide solution for 1 h to remove the aluminum oxide film. We left it in the air for 24 h, and then the oxidation of the sample was characterized.
At normal temperature, nitric acid reacts with sodium azide to produce hydrogen azide gas. The gas-solid reaction of copper with hydrogen azide can produce copper azide. The reaction equation involved in the above process is as follows: The silicon-based Cu 2 O@CNTs composite film and Cu@CNTs composite film were placed in a closed container. We passed N 2 into it to remove the air and added 4 g·mL −1 of nitric acid and 2 Micromachines 2020, 11, 575 4 of 17 g·mL −1 of sodium azide solution. After 72 h of reaction, a silicon-based Cu(N 3 ) 2 @CNTs composite energetic film was obtained.
Field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) were used to characterize the micro-morphology of the samples and X-ray diffraction (XRD), high-resolution transmission electron microscopy (HRTEM), selective electron diffraction (SAED) and energy dispersive X-ray spectroscopy (EDS) were used to characterize the crystal structure and composition. The thermal decomposition kinetics of the Cu(N 3 ) 2 @CNTs composite energetic film were studied by differential scanning calorimetry (DSC).
The performance of the energetic films was tested by laser ignition experiments, and high-speed photography was used to record the ignition process. By changing the output voltage of the laser, the output energy and the optimal detonation energy can be obtained (84.1 mJ). The model of the high-speed camera used in this experiment was APHJ-100, the frequency of high-speed camera was 30,000 fps and the laser is a Nd:YAG Dawa-200 single pulse laser with a wavelength of 1064 nm and pulse width of 6 ns.
An electrical explosion performance test was performed on the energetic film. The schematic diagram of the electrical explosion experiment device is shown in Figure 2. A high-voltage ceramic capacitor was used as the energy storage capacitor. The capacitance was 0.22 µF and the initial explosion voltage was 1500 V. The high-voltage probe and the Rogowski current loop were used to test the voltage and current changes in the energetic film during the ignition process and recorded with an oscilloscope. The electric explosion curve was analyzed.
Micromachines 2020, 11, 575 4 of 17 The silicon-based Cu2O@CNTs composite film and Cu@CNTs composite film were placed in a closed container. We passed N2 into it to remove the air and added 4g·mL −1 of nitric acid and 2g·mL −1 of sodium azide solution. After 72 h of reaction, a silicon-based Cu(N3)2@CNTs composite energetic film was obtained.
Field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) were used to characterize the micro-morphology of the samples and X-ray diffraction (XRD), high-resolution transmission electron microscopy (HRTEM), selective electron diffraction (SAED) and energy dispersive X-ray spectroscopy (EDS) were used to characterize the crystal structure and composition. The thermal decomposition kinetics of the Cu(N3)2@CNTs composite energetic film were studied by differential scanning calorimetry (DSC).
The performance of the energetic films was tested by laser ignition experiments, and high-speed photography was used to record the ignition process. By changing the output voltage of the laser, the output energy and the optimal detonation energy can be obtained (84.1 mJ). The model of the highspeed camera used in this experiment was APHJ-100, the frequency of high-speed camera was 30,000 fps and the laser is a Nd:YAG Dawa-200 single pulse laser with a wavelength of 1064 nm and pulse width of 6 ns.
An electrical explosion performance test was performed on the energetic film. The schematic diagram of the electrical explosion experiment device is shown in Figure 2. A high-voltage ceramic capacitor was used as the energy storage capacitor. The capacitance was 0.22 μF and the initial explosion voltage was 1500 V. The high-voltage probe and the Rogowski current loop were used to test the voltage and current changes in the energetic film during the ignition process and recorded with an oscilloscope. The electric explosion curve was analyzed. The circuit was connected first, then the sample was soldered to the positive and negative terminals of the high-voltage output terminal of the printed circuit board (PCB), and the resistance was measured to be 0.4 Ω. The high voltage (1500 V) of the pulse power source acted on both ends of the high-pressure gas switch, and the trigger signal (3000 V) was output by the pulse power source, thereby turning on the gas switch, completing the high-voltage output and acting on both ends of the sample.
The electrostatic sensitivity test of the silicon-based Cu(N3)2@ CNTs composite energetic film was carried out by the Bruceton method. We tested the sample at a static ignition energy and ignition voltage to evaluate the electrostatic sensitivity of the sample prepared. The 50% ignition energy (E50%) of the sample was calculated by the critical ignition voltage (U50%). The circuit was connected first, then the sample was soldered to the positive and negative terminals of the high-voltage output terminal of the printed circuit board (PCB), and the resistance was measured to be 0.4 Ω. The high voltage (1500 V) of the pulse power source acted on both ends of the high-pressure gas switch, and the trigger signal (3000 V) was output by the pulse power source, thereby turning on the gas switch, completing the high-voltage output and acting on both ends of the sample.
The electrostatic sensitivity test of the silicon-based Cu(N 3 ) 2 @ CNTs composite energetic film was carried out by the Bruceton method. We tested the sample at a static ignition energy and ignition voltage to evaluate the electrostatic sensitivity of the sample prepared. The 50% ignition energy (E 50% ) of the sample was calculated by the critical ignition voltage (U 50% ). Figure 3a shows the XRD pattern of the silicon-based Al/Ti film. Compared with the standard pattern (PDF#01-1176), it is obvious that when 2θ is 38.53 • , the corresponding characteristic peak is the Al (111) plane. The aluminum film grows preferentially on the (111) crystal plane, which indicates that it has the lowest surface free energy in the direction of the (111) crystal plane. As in the atomic force microscope (AFM) diagram shown in Figure 3b, the aluminum film grows in an island-like structure, and the average surface roughness is about 30 nm. Figure 3c shows the EDS spectrum of the sample. The Si peaks, Al peaks and Ti peaks present in the spectrum are from the silicon substrate and the Al layer as well as the Ti layer. The O peak is from the oxidized alumina of the Al layer. The Au peak comes from the gold-spraying operation performed during the sample preparation to increase conductivity.

The Silicon-Based Al/Ti Film
Micromachines 2020, 11, 575 5 of 17 Figure 3a shows the XRD pattern of the silicon-based Al/Ti film. Compared with the standard pattern (PDF#01-1176), it is obvious that when 2θ is 38.53°, the corresponding characteristic peak is the Al (111) plane. The aluminum film grows preferentially on the (111) crystal plane, which indicates that it has the lowest surface free energy in the direction of the (111) crystal plane. As in the atomic force microscope (AFM) diagram shown in Figure 3b, the aluminum film grows in an island-like structure, and the average surface roughness is about 30 nm. Figure 3c shows the EDS spectrum of the sample. The Si peaks, Al peaks and Ti peaks present in the spectrum are from the silicon substrate and the Al layer as well as the Ti layer. The O peak is from the oxidized alumina of the Al layer. The Au peak comes from the gold-spraying operation performed during the sample preparation to increase conductivity.  Figure 4 shows a schematic diagram of the anodizing process. The surface of the Al film prepared by electron beam evaporation deposition is an island-like structure. The first step of anodizing caused some uniformly distributed dents on the surface of the aluminum film. During this process, an alumina film will also be produced. After the alumina film is removed, these dents become the initial sites for the growth of the pores during the second step of anodizing, allowing highly ordered pores to grow. After undergoing the pore enlarging process, the diameter of these channels becomes larger, so it is more suitable as a template for depositing carbon nanotubes.  Figure 4 shows a schematic diagram of the anodizing process. The surface of the Al film prepared by electron beam evaporation deposition is an island-like structure. The first step of anodizing caused some uniformly distributed dents on the surface of the aluminum film. During this process, an alumina film will also be produced. After the alumina film is removed, these dents become the initial sites for the growth of the pores during the second step of anodizing, allowing highly ordered pores to grow. After undergoing the pore enlarging process, the diameter of these channels becomes larger, so it is more suitable as a template for depositing carbon nanotubes. As depicted in Figure 5, the cross-section and surface FESEM images of the porous alumina film indicate the formation of a porous structure. Figure 5a,b shows the FESEM of the film before pore enlarging. As shown in Figure 5a, the thickness of the well-organized porous alumina membranes (PAM) is about 1.2 μm. Due to the increase in the oxidation time in the first step, the surface of the PAM is relatively flat (Figure 5b). Both the pore diameter and the pore spacing are 100 nm approximately. Figure 5c,d shows a cross-section and surface FESEM image of the PAM in a 5 wt% phosphoric acid solution at 30 °C for 40 min. After the pore treatment, the barrier layer of the alumina film was basically removed, indicating that the barrier layer was effectively removed during the pore enlarging process. The pore diameter of the PAM becomes larger and the diameter is 150 nm approximately and relatively regular and uniform (Figure 5d).

The Silicon-Based Cu@CNTs Composite Film
Carbon nanotube arrays can be prepared by chemical vapor deposition through the template method. In this process, acetylene served as a source of carbon. The carbon produced by the cracking of acetylene was deposited on the inner surface of the channel. Acetylene was continuously introduced into the tube furnace to ensure the integrity of the deposition process. Moreover, the alumina template as a catalyst improved the efficiency of the carbon nanotube growth process. As depicted in Figure 5, the cross-section and surface FESEM images of the porous alumina film indicate the formation of a porous structure. Figure 5a,b shows the FESEM of the film before pore enlarging. As shown in Figure 5a, the thickness of the well-organized porous alumina membranes (PAM) is about 1.2 µm. Due to the increase in the oxidation time in the first step, the surface of the PAM is relatively flat (Figure 5b). Both the pore diameter and the pore spacing are 100 nm approximately. Figure 5c,d shows a cross-section and surface FESEM image of the PAM in a 5 wt% phosphoric acid solution at 30 • C for 40 min. After the pore treatment, the barrier layer of the alumina film was basically removed, indicating that the barrier layer was effectively removed during the pore enlarging process. The pore diameter of the PAM becomes larger and the diameter is 150 nm approximately and relatively regular and uniform ( Figure 5d).  As depicted in Figure 5, the cross-section and surface FESEM images of the porous alumina film indicate the formation of a porous structure. Figure 5a,b shows the FESEM of the film before pore enlarging. As shown in Figure 5a, the thickness of the well-organized porous alumina membranes (PAM) is about 1.2 μm. Due to the increase in the oxidation time in the first step, the surface of the PAM is relatively flat (Figure 5b). Both the pore diameter and the pore spacing are 100 nm approximately. Figure 5c,d shows a cross-section and surface FESEM image of the PAM in a 5 wt% phosphoric acid solution at 30 °C for 40 min. After the pore treatment, the barrier layer of the alumina film was basically removed, indicating that the barrier layer was effectively removed during the pore enlarging process. The pore diameter of the PAM becomes larger and the diameter is 150 nm approximately and relatively regular and uniform ( Figure 5d).

The Silicon-Based Cu@CNTs Composite Film
Carbon nanotube arrays can be prepared by chemical vapor deposition through the template method. In this process, acetylene served as a source of carbon. The carbon produced by the cracking of acetylene was deposited on the inner surface of the channel. Acetylene was continuously introduced into the tube furnace to ensure the integrity of the deposition process. Moreover, the alumina template as a catalyst improved the efficiency of the carbon nanotube growth process.

The Silicon-Based Cu@CNTs Composite Film
Carbon nanotube arrays can be prepared by chemical vapor deposition through the template method. In this process, acetylene served as a source of carbon. The carbon produced by the cracking of acetylene was deposited on the inner surface of the channel. Acetylene was continuously introduced into the tube furnace to ensure the integrity of the deposition process. Moreover, the alumina template as a catalyst improved the efficiency of the carbon nanotube growth process. Figure 6 shows TEM images of the carbon nanotubes prepared at different deposition times. Figure 6a,c shows the TEM of the carbon nanotubes with a deposition time of 1.5 h, and Figure 6b,d shows the carbon nanotubes with a deposition time of 2 h. Carbon nanotubes replicate the pores of the PAM well and their outer diameter is basically the same as that of the alumina film, about 150 nm (Figure 6a,b). Its length is basically the same as the thickness of the alumina film, which is about 1.2 µm. The wall thickness of the carbon nanotubes in Figure 6c is 9-10 nm, and in Figure 6d is 13-14 nm. As the reaction time increases, the wall thickness increases accordingly. Since the crystallinity of the wall of the carbon nanotubes is poor, it is difficult to identify the number of layers of the carbon nanotubes.
Micromachines 2020, 11,575 7 of 17 Figure 6 shows TEM images of the carbon nanotubes prepared at different deposition times. Figure 6a,c shows the TEM of the carbon nanotubes with a deposition time of 1.5 h, and Figure 6b,d shows the carbon nanotubes with a deposition time of 2 h. Carbon nanotubes replicate the pores of the PAM well and their outer diameter is basically the same as that of the alumina film, about 150 nm (Figure 6a,b). Its length is basically the same as the thickness of the alumina film, which is about 1.2 μm. The wall thickness of the carbon nanotubes in Figure 6c is 9-10 nm, and in Figure 6d is 13-14 nm. As the reaction time increases, the wall thickness increases accordingly. Since the crystallinity of the wall of the carbon nanotubes is poor, it is difficult to identify the number of layers of the carbon nanotubes.   The addition of organic additives has a great impact on the size, morphology and crystal orientation of the electrodeposited particles and they can be used alone or in combination. In this  Figure 7a,b shows the Raman spectra of the carbon nanotubes when the deposition time is 1.5 h and 2 h, respectively. The I D /I G were 0.79 and 0.74, indicating that the crystallinity of the carbon nanotubes was poor. It shows that the crystallinity of the carbon nanotubes does not change significantly with the increase in deposition time.
Micromachines 2020, 11,575 7 of 17 Figure 6 shows TEM images of the carbon nanotubes prepared at different deposition times. Figure 6a,c shows the TEM of the carbon nanotubes with a deposition time of 1.5 h, and Figure 6b,d shows the carbon nanotubes with a deposition time of 2 h. Carbon nanotubes replicate the pores of the PAM well and their outer diameter is basically the same as that of the alumina film, about 150 nm (Figure 6a,b). Its length is basically the same as the thickness of the alumina film, which is about 1.2 μm. The wall thickness of the carbon nanotubes in Figure 6c is 9-10 nm, and in Figure 6d is 13-14 nm. As the reaction time increases, the wall thickness increases accordingly. Since the crystallinity of the wall of the carbon nanotubes is poor, it is difficult to identify the number of layers of the carbon nanotubes.   The addition of organic additives has a great impact on the size, morphology and crystal orientation of the electrodeposited particles and they can be used alone or in combination. In this  The addition of organic additives has a great impact on the size, morphology and crystal orientation of the electrodeposited particles and they can be used alone or in combination. In this work, 0.08 mol/L copper sulfate pentahydrate and 0.4 mol/L boric acid were used as the basic electrolytes. The following cases were compared: (a) no additives, (b) thiourea (TU) with a concentration of 0.01 g/L and (c) 0.01 g/L of TU and 0.1 g/L of polyvinylpyrrolidone (PVP). The electrochemical deposition time was 2 h. After the experiment, the products were characterized by TEM. Figure 8 shows the TEM images of these products. Figure 8a is the morphology of the sample without the additives. Obviously, the copper nanoparticles have a larger particle size and are distributed at the mouth of the carbon nanotubes, blocking the mouth. Figure 8b is the morphology of copper after adding TU at a concentration of 0.01 g/L. It can be found that the density of the copper particles in the CNTs is low. Figure 8c shows the morphology of copper deposited in the CNTs after adding both TU and PVP. Compared with Figure 8b, the deposition density of the copper nanoparticles in the CNTs has increased.
Micromachines 2020, 11, 575 8 of 17 work, 0.08 mol/L copper sulfate pentahydrate and 0.4 mol/L boric acid were used as the basic electrolytes. The following cases were compared: (a) no additives, (b) thiourea (TU) with a concentration of 0.01 g/L and (c) 0.01 g/L of TU and 0.1 g/L of polyvinylpyrrolidone (PVP). The electrochemical deposition time was 2 h. After the experiment, the products were characterized by TEM. Figure 8 shows the TEM images of these products. Figure 8a is the morphology of the sample without the additives. Obviously, the copper nanoparticles have a larger particle size and are distributed at the mouth of the carbon nanotubes, blocking the mouth. Figure 8b is the morphology of copper after adding TU at a concentration of 0.01 g/L. It can be found that the density of the copper particles in the CNTs is low. Figure 8c shows the morphology of copper deposited in the CNTs after adding both TU and PVP. Compared with Figure 8b, the deposition density of the copper nanoparticles in the CNTs has increased. The sample was placed in the air for 24 h, and its morphology is shown in Figure 9. Due to the rapid oxidation rate of the deposited copper nanoparticles in the air, which turns the copper into cuprous oxide, the morphology of the copper nanoparticles deposited in the CNTs has changed. In order to demonstrate the composition of the products, the products before and after the change were characterized.   The sample was placed in the air for 24 h, and its morphology is shown in Figure 9. Due to the rapid oxidation rate of the deposited copper nanoparticles in the air, which turns the copper into cuprous oxide, the morphology of the copper nanoparticles deposited in the CNTs has changed. In order to demonstrate the composition of the products, the products before and after the change were characterized.
Micromachines 2020, 11, 575 8 of 17 work, 0.08 mol/L copper sulfate pentahydrate and 0.4 mol/L boric acid were used as the basic electrolytes. The following cases were compared: (a) no additives, (b) thiourea (TU) with a concentration of 0.01 g/L and (c) 0.01 g/L of TU and 0.1 g/L of polyvinylpyrrolidone (PVP). The electrochemical deposition time was 2 h. After the experiment, the products were characterized by TEM. Figure 8 shows the TEM images of these products. Figure 8a is the morphology of the sample without the additives. Obviously, the copper nanoparticles have a larger particle size and are distributed at the mouth of the carbon nanotubes, blocking the mouth. Figure 8b is the morphology of copper after adding TU at a concentration of 0.01 g/L. It can be found that the density of the copper particles in the CNTs is low. Figure 8c shows the morphology of copper deposited in the CNTs after adding both TU and PVP. Compared with Figure 8b, the deposition density of the copper nanoparticles in the CNTs has increased. The sample was placed in the air for 24 h, and its morphology is shown in Figure 9. Due to the rapid oxidation rate of the deposited copper nanoparticles in the air, which turns the copper into cuprous oxide, the morphology of the copper nanoparticles deposited in the CNTs has changed. In order to demonstrate the composition of the products, the products before and after the change were characterized.      The Al and O peaks are very weak, they come from the PAM which has not been completely removed. The C peak comes from carbon nanotubes, while the Cu peak is nano-copper particles. The existence of the S peak indicates that the TU and copper ions have formed sulfides. Compared with the XRD pattern, it can be concluded that the S element is derived from CuS, and it is confirmed that the sulfur ions in thiourea and copper ions in the electrolyte form copper sulfide.    The Al and O peaks are very weak, they come from the PAM which has not been completely removed. The C peak comes from carbon nanotubes, while the Cu peak is nano-copper particles. The existence of the S peak indicates that the TU and copper ions have formed sulfides. Compared with the XRD pattern, it can be concluded that the S element is derived from CuS, and it is confirmed that the sulfur ions in thiourea and copper ions in the electrolyte form copper sulfide.   The Al and O peaks are very weak, they come from the PAM which has not been completely removed. The C peak comes from carbon nanotubes, while the Cu peak is nano-copper particles. The existence of the S peak indicates that the TU and copper ions have formed sulfides. Compared with the XRD pattern, it can be concluded that the S element is derived from CuS, and it is confirmed that the sulfur ions in thiourea and copper ions in the electrolyte form copper sulfide.     411) and (213) crystal planes of cuprous azide, respectively. Compared with the standard spectra (PDF#65-3288) and (PDF#21-0281), it can be found that the diffraction peaks at 36.523 • and 11.767 • correspond to the (111) crystal plane of cuprous oxide and (110) crystal plane of copper azide. The copper azide diffraction peak is very weak, which shows that cuprous oxide mainly formed cuprous azide after the azide reaction. The diffraction peak of copper azide is obvious in Figure 12b.  301) and (251) crystal planes of copper azide, respectively. It shows that copper was mainly converted into copper azide after the azide reaction, and no cuprous azide was formed.

The Silicon-Based Cu(N 3 ) 2 @CNTs Composite Energetic Film
Micromachines 2020, 11, 575 10 of 17 spectra (PDF#65-3288) and (PDF#21-0281), it can be found that the diffraction peaks at 36.523° and 11.767° correspond to the (111) crystal plane of cuprous oxide and (110) crystal plane of copper azide. The copper azide diffraction peak is very weak, which shows that cuprous oxide mainly formed cuprous azide after the azide reaction. The diffraction peak of copper azide is obvious in Figure 12b.
and (251) crystal planes of copper azide, respectively. It shows that copper was mainly converted into copper azide after the azide reaction, and no cuprous azide was formed.  Furthermore, the azide products were analyzed by HRTEM and SAED. Figure 13a The copper azide diffraction peak is very weak, which shows that cuprous oxide mainly formed cuprous azide after the azide reaction. The diffraction peak of copper azide is obvious in Figure 12b.

Properties of Composite Energetic Films
DSC was used to analyze the thermal behavior of the Cu(N3)2@CNTs composite film. As depicted in Figure 14, the exothermic temperature of the composite energetic material is 168.79 to 223.75 °C and a sharp exothermic peak was observed at 210.95 °C, indicating a rapid explosion reaction of the energetic film. The electrostatic sensitivity of the Cu(N3)2@CNTs energetic film was tested by the Bruceton method. The results are shown in Table 1. The U50% is 4.0 kV and E50% is 4.0 mJ. The energetic film was ignited by laser and the ignition process was recorded by high-speed photography ( Figure 15). The area of the flame reached its maximum at 33.3 μs, and the flame lasted about 166.7 μs.

Properties of Composite Energetic Films
DSC was used to analyze the thermal behavior of the Cu(N 3 ) 2 @CNTs composite film. As depicted in Figure 14, the exothermic temperature of the composite energetic material is 168.79 to 223.75 • C and a sharp exothermic peak was observed at 210.95 • C, indicating a rapid explosion reaction of the energetic film.

Properties of Composite Energetic Films
DSC was used to analyze the thermal behavior of the Cu(N3)2@CNTs composite film. As depicted in Figure 14, the exothermic temperature of the composite energetic material is 168.79 to 223.75 °C and a sharp exothermic peak was observed at 210.95 °C, indicating a rapid explosion reaction of the energetic film. The electrostatic sensitivity of the Cu(N3)2@CNTs energetic film was tested by the Bruceton method. The results are shown in Table 1. The U50% is 4.0 kV and E50% is 4.0 mJ. The energetic film was ignited by laser and the ignition process was recorded by high-speed photography ( Figure 15). The area of the flame reached its maximum at 33.3 μs, and the flame lasted about 166.7 μs. The electrostatic sensitivity of the Cu(N 3 ) 2 @CNTs energetic film was tested by the Bruceton method. The results are shown in Table 1. The U 50% is 4.0 kV and E 50% is 4.0 mJ. The energetic film was ignited by laser and the ignition process was recorded by high-speed photography ( Figure 15). The area of the flame reached its maximum at 33.3 µs, and the flame lasted about 166.7 µs.  Figure 16 shows the appearance of the film after laser ignition. The diameter of the firing area is about 4.274 mm. The firing position is concave. Spattering can be observed around it and black spots appear, which indicates that the film was melted during laser ignition.   Table 2. One of the envisaged applications of the composite energetic film designed in this article is as an initial energy-converting component in micro-detonation devices. The high-energy pulse capacitor gave a certain amount of energy to the composite energetic film, and the composite energetic film exploded to generate energy. The electric explosion characteristic mainly refers to the energy conversion efficiency of the energetic film in this process. The energy utilization ratio η is the ratio of ∫ • and 1/2C . ∫ • is the integral of • , and the integral interval is from 0 to t (the time when the current reaches the peak value). It is the energy obtained by the energetic film from the capacitor during the explosion. 1/2C is the energy stored in the highvoltage capacitor. The value of η shows that the energetic film has a good energy conversion ability during the electrical explosion process. Moreover, the time for the current to rise to the peak value is short, and the peak value of the current is large. These show that the energetic film has good electrical explosion properties. Therefore, it is expected to be used as a kind of energy conversion component in the micro-detonation device. Figure 15. High-speed photography of energetic films during laser ignition. Figure 16 shows the appearance of the film after laser ignition. The diameter of the firing area is about 4.274 mm. The firing position is concave. Spattering can be observed around it and black spots appear, which indicates that the film was melted during laser ignition.
Micromachines 2020, 11, 575 12 of 17 Figure 15. High-speed photography of energetic films during laser ignition. Figure 16 shows the appearance of the film after laser ignition. The diameter of the firing area is about 4.274 mm. The firing position is concave. Spattering can be observed around it and black spots appear, which indicates that the film was melted during laser ignition.   Table 2. One of the envisaged applications of the composite energetic film designed in this article is as an initial energy-converting component in micro-detonation devices. The high-energy pulse capacitor gave a certain amount of energy to the composite energetic film, and the composite energetic film exploded to generate energy. The electric explosion characteristic mainly refers to the energy conversion efficiency of the energetic film in this process. The energy utilization ratio η is the ratio of ∫ • and 1/2C . ∫ • is the integral of • , and the integral interval is from 0 to t (the time when the current reaches the peak value). It is the energy obtained by the energetic film from the capacitor during the explosion. 1/2C is the energy stored in the highvoltage capacitor. The value of η shows that the energetic film has a good energy conversion ability during the electrical explosion process. Moreover, the time for the current to rise to the peak value is short, and the peak value of the current is large. These show that the energetic film has good electrical explosion properties. Therefore, it is expected to be used as a kind of energy conversion component in the micro-detonation device.  Figure 17 shows the electric explosion characteristic curve of the energetic film. The energy utilization ratio η is obtained through Equation (3). The parameters such as current and voltage in the curve are listed in Table 2. One of the envisaged applications of the composite energetic film designed in this article is as an initial energy-converting component in micro-detonation devices. The high-energy pulse capacitor gave a certain amount of energy to the composite energetic film, and the composite energetic film exploded to generate energy. The electric explosion characteristic mainly refers to the energy conversion efficiency of the energetic film in this process. The energy utilization ratio η is the ratio of t 0 U·Idt and 1/2CU 2 0 . t 0 U·Idt is the integral of U·I, and the integral interval is from 0 to t (the time when the current reaches the peak value). It is the energy obtained by the energetic film from the capacitor during the explosion. 1/2CU 2 0 is the energy stored in the high-voltage capacitor. The value of η shows that the energetic film has a good energy conversion ability during the electrical explosion process. Moreover, the time for the current to rise to the peak value is short, and the peak value of the current is large. These show that the energetic film has good electrical explosion properties. Therefore, it is expected to be used as a kind of energy conversion component in the micro-detonation device.

Discussion
In this work, carbon nanotube arrays were prepared by chemical vapor deposition and template methods. Levin et al. [29] analyzed the existence form of an anodic aluminum oxide (AAO) template at different temperature stages during heating. When the temperature is above 1200 °C, the existence form of AAO is stable α-Al2O3. When the temperature reaches 230-500 °C, it changes to η-Al2O3, and the existence form is mainly γ-Al2O3 when the temperature is 450-750 °C. The η-Al2O3 and γ-Al2O3 have higher porosity and a large specific surface area, so they have better catalytic activity. The commonly used carbon source gas can only be thermally decomposed at high temperatures (800 °C). When acetylene was used as the carbon source, the aligned carbon nanotube array can be prepared in the AAO template channel at about 600 °C. It shows that the AAO template plays a catalytic role in the thermal decomposition of acetylene, so that it can be decomposed at lower temperatures. Liu et al. [30] studied that the catalytic process of the AAO template has two main steps. The first is the catalytic effect when the carbon source is decomposed, and the second is the catalytic effect during the growth of carbon nanotubes.
Yang et al. [31] studied the growth mechanism of carbon nanotubes prepared by the AAO templates and found that carbon atoms were deposited on the walls of the alumina template and the channels of the AAO template were copied to obtain the carbon nanotube structure. The carbon nanotubes are named a-CNTs (amorphous carbon nanotubes). By analyzing the Raman spectrum and TEM, it can be known that the carbon nanotubes obtained in this work are also a-CNTs. With reference to Yang's conclusion, we consider the mechanism of preparing carbon nanotube arrays through PAM, as shown in Figure 18.

Discussion
In this work, carbon nanotube arrays were prepared by chemical vapor deposition and template methods. Levin et al. [29] analyzed the existence form of an anodic aluminum oxide (AAO) template at different temperature stages during heating. When the temperature is above 1200 • C, the existence form of AAO is stable α-Al 2 O 3 . When the temperature reaches 230-500 • C, it changes to η-Al 2 O 3 , and the existence form is mainly γ-Al 2 O 3 when the temperature is 450-750 • C. The η-Al 2 O 3 and γ-Al 2 O 3 have higher porosity and a large specific surface area, so they have better catalytic activity. The commonly used carbon source gas can only be thermally decomposed at high temperatures (800 • C). When acetylene was used as the carbon source, the aligned carbon nanotube array can be prepared in the AAO template channel at about 600 • C. It shows that the AAO template plays a catalytic role in the thermal decomposition of acetylene, so that it can be decomposed at lower temperatures. Liu et al. [30] studied that the catalytic process of the AAO template has two main steps. The first is the catalytic effect when the carbon source is decomposed, and the second is the catalytic effect during the growth of carbon nanotubes.
Yang et al. [31] studied the growth mechanism of carbon nanotubes prepared by the AAO templates and found that carbon atoms were deposited on the walls of the alumina template and the channels of the AAO template were copied to obtain the carbon nanotube structure. The carbon nanotubes are named a-CNTs (amorphous carbon nanotubes). By analyzing the Raman spectrum and TEM, it can be known that the carbon nanotubes obtained in this work are also a-CNTs. With reference to Yang's conclusion, we consider the mechanism of preparing carbon nanotube arrays through PAM, as shown in Figure 18. First, the carbon source is cracked to generate carbon atoms under the catalysis of the PAM. Carbon atoms are deposited in the pores of the PAM. Under the autocatalysis of the PAM, the carbon atoms replicate the pore structure of the PAM and grow into carbon nanotubes at a relatively lower temperature. The carbon atoms generated by the continuous decomposition of the carbon source continue to enter the hollow lumen of the carbon nanotubes, deposit on the inner wall of the formed carbon nanotubes and repeat the growth of the carbon nanotubes. With the increase in the reaction time, the thickness of the carbon nanotube wall also increases and even solid carbon nanorods may be obtained. Therefore, carbon nanotubes of different thicknesses can be obtained by controlling the reaction time.
During the stage of electrochemical deposition, the type of additive has a great influence on the filling effect of the copper nanoparticles. Additives can be generally classified into three categories: halide ions, supporting electrolytes and organic additives. Xu [32] found that the addition of Cl − in the electrolyte would form a "chlorine bridge" under enhanced polarization conditions, changing the charge transfer mechanism of the charged particles in the solution. The role of the supporting electrolyte is generally considered to increase the charge transfer rate and the conductivity of the electrolyte.
By comparing the different morphologies of copper deposited in CNTs, it is obvious that when no additives were added, the crystal size of copper grew too large, resulting in the inability of copper nanoparticles to be deposited in CNTs. When TU was added, the particle size of the copper nanoparticles decreased significantly. These results confirm that TU as an additive has the effect of inhibiting the growth of copper particles. First, TU adsorbs on the surface of the copper particles and reacts with Cu 2+ to form compounds. These compounds become adsorption layers and cover the active sites of copper, preventing further reduction of Cu 2+ . Then TU and CuS generated by Cu 2+ are also adsorbed on the surface of the copper particles, making the copper atoms unable to reach the active site and inhibiting the growth of grains. When TU and PVP were added together, the particle size of the copper nanoparticles became smaller, indicating that the inhibitory effect is more pronounced when PVP and TU work together. Figure 19 shows the SEM images of the cross-section of CNTs grown on silicon-based porous alumina films, silicon-based Cu@CNTs film and silicon-based Cu(N3)2@CNTs film. It can be seen from Figure 19a that the carbon nanotubes were tightly attached to the inner walls of the pores of the porous alumina template, and the shape of the pores was well reproduced. Its upper end was open, providing an entrance for the deposition of the copper nanoparticles. Figure 19b shows the crosssection of the film after the copper nanoparticles were filled. The copper nanoparticles were relatively evenly distributed in the carbon nanotubes, without obvious agglomeration. After the azide reaction, the copper nanoparticles were converted to energetic copper azide (Figure 19c). It can be seen that its particles' size become slightly larger, and the ordered directional carbon nanotube arrays have not been destroyed during the azide process. First, the carbon source is cracked to generate carbon atoms under the catalysis of the PAM. Carbon atoms are deposited in the pores of the PAM. Under the autocatalysis of the PAM, the carbon atoms replicate the pore structure of the PAM and grow into carbon nanotubes at a relatively lower temperature. The carbon atoms generated by the continuous decomposition of the carbon source continue to enter the hollow lumen of the carbon nanotubes, deposit on the inner wall of the formed carbon nanotubes and repeat the growth of the carbon nanotubes. With the increase in the reaction time, the thickness of the carbon nanotube wall also increases and even solid carbon nanorods may be obtained. Therefore, carbon nanotubes of different thicknesses can be obtained by controlling the reaction time.
During the stage of electrochemical deposition, the type of additive has a great influence on the filling effect of the copper nanoparticles. Additives can be generally classified into three categories: halide ions, supporting electrolytes and organic additives. Xu [32] found that the addition of Cl − in the electrolyte would form a "chlorine bridge" under enhanced polarization conditions, changing the charge transfer mechanism of the charged particles in the solution. The role of the supporting electrolyte is generally considered to increase the charge transfer rate and the conductivity of the electrolyte.
By comparing the different morphologies of copper deposited in CNTs, it is obvious that when no additives were added, the crystal size of copper grew too large, resulting in the inability of copper nanoparticles to be deposited in CNTs. When TU was added, the particle size of the copper nanoparticles decreased significantly. These results confirm that TU as an additive has the effect of inhibiting the growth of copper particles. First, TU adsorbs on the surface of the copper particles and reacts with Cu 2+ to form compounds. These compounds become adsorption layers and cover the active sites of copper, preventing further reduction of Cu 2+ . Then TU and CuS generated by Cu 2+ are also adsorbed on the surface of the copper particles, making the copper atoms unable to reach the active site and inhibiting the growth of grains. When TU and PVP were added together, the particle size of the copper nanoparticles became smaller, indicating that the inhibitory effect is more pronounced when PVP and TU work together. Figure 19 shows the SEM images of the cross-section of CNTs grown on silicon-based porous alumina films, silicon-based Cu@CNTs film and silicon-based Cu(N 3 ) 2 @CNTs film. It can be seen from Figure 19a that the carbon nanotubes were tightly attached to the inner walls of the pores of the porous alumina template, and the shape of the pores was well reproduced. Its upper end was open, providing an entrance for the deposition of the copper nanoparticles. Figure 19b shows the cross-section of the film after the copper nanoparticles were filled. The copper nanoparticles were relatively evenly distributed in the carbon nanotubes, without obvious agglomeration. After the azide reaction, the copper nanoparticles were converted to energetic copper azide (Figure 19c). It can be seen that its particles' size become slightly larger, and the ordered directional carbon nanotube arrays have not been destroyed during the azide process. Carbon nanotubes have excellent electrical conductivity. When the copper azide is filled into the hollow cavity of the oriented carbon nanotube, the electrostatic charge will be dispersed on the surface of the carbon nanotube instead of accumulating on the surface of the copper azide. Therefore, the explosion of copper azide caused by the electrostatic charge can be avoided, and the electrostatic sensitivity of the composite energetic film can be effectively reduced. Compared with the E50% of pure nano-copper azide (0.05 mJ) reported before [33], the electrostatic sensitivity of the composite film is significantly reduced.
In future research, it will be an interesting direction to combine carbon materials with different structures and copper azide to obtain energetic materials with good performance. Copper azide can take advantage of its high energy density in composite materials, and carbon materials with different structures can isolate copper azide particles in various ways, thereby reducing the mechanical sensitivity of composite energetic materials. Meanwhile, the excellent conductivity of carbon materials can also reduce the electrostatic sensitivity of composite energetic materials. On the one hand, the advantages of copper azide are used, and on the other hand, carbon materials are used to make up for the shortcomings of copper azide as a primary explosive. The composite energetic films designed and prepared in this paper will also be further optimized to be applied to MEMS pyrotechnic devices. Furthermore, an in-depth study of the mechanism of azide reaction will also be a valuable work.

Conclusions
Energetic films realized by encapsulating copper azide in silicon-based carbon nanotube arrays were presented in this study, which opens a route to integrate primary explosives realized on Ticoated silicon wafers. A silicon-based alumina template was prepared using a modified anodization process. We prepared CNTs by chemical vapor deposition using the prepared template. Copper nanoparticles were embedded into the carbon nanotubes by means of electrochemical deposition. The prepared Cu@CNTs composite film reacted with hydrogen azide gas to synthesize the copper azide@CNTs composite energetic film with low electrostatic sensitivity. Moreover, this composite film stores a great quantity of energy and can be released rapidly. Therefore, it has great potential to replace the commonly used lead-containing detonator and achieve the miniaturization of the ignition device.   Carbon nanotubes have excellent electrical conductivity. When the copper azide is filled into the hollow cavity of the oriented carbon nanotube, the electrostatic charge will be dispersed on the surface of the carbon nanotube instead of accumulating on the surface of the copper azide. Therefore, the explosion of copper azide caused by the electrostatic charge can be avoided, and the electrostatic sensitivity of the composite energetic film can be effectively reduced. Compared with the E 50% of pure nano-copper azide (0.05 mJ) reported before [33], the electrostatic sensitivity of the composite film is significantly reduced.
In future research, it will be an interesting direction to combine carbon materials with different structures and copper azide to obtain energetic materials with good performance. Copper azide can take advantage of its high energy density in composite materials, and carbon materials with different structures can isolate copper azide particles in various ways, thereby reducing the mechanical sensitivity of composite energetic materials. Meanwhile, the excellent conductivity of carbon materials can also reduce the electrostatic sensitivity of composite energetic materials. On the one hand, the advantages of copper azide are used, and on the other hand, carbon materials are used to make up for the shortcomings of copper azide as a primary explosive. The composite energetic films designed and prepared in this paper will also be further optimized to be applied to MEMS pyrotechnic devices. Furthermore, an in-depth study of the mechanism of azide reaction will also be a valuable work.

Conclusions
Energetic films realized by encapsulating copper azide in silicon-based carbon nanotube arrays were presented in this study, which opens a route to integrate primary explosives realized on Ti-coated silicon wafers. A silicon-based alumina template was prepared using a modified anodization process. We prepared CNTs by chemical vapor deposition using the prepared template. Copper nanoparticles were embedded into the carbon nanotubes by means of electrochemical deposition. The prepared Cu@CNTs composite film reacted with hydrogen azide gas to synthesize the copper azide@CNTs composite energetic film with low electrostatic sensitivity. Moreover, this composite film stores a great quantity of energy and can be released rapidly. Therefore, it has great potential to replace the commonly used lead-containing detonator and achieve the miniaturization of the ignition device.