Study on Thermal Decomposition Behavior, Gaseous Products, and Kinetic Analysis of Bis-(Dimethylglyoximato) Nickel(II) Complex Using TG-DSC-FTIR-MS Technique

: The ﬁber-like bis-(dimethylglyoximato) nickel(II) complex, Ni(DMG) 2 was successfully synthesized. The obtained samples were characterized by SEM-EDS, FT-IR, XRD, and XPS. The TG-DSC-FTIR-MS coupling technique was used to characterize the thermal decomposition behavior and evolved gas analysis of Ni(DMG) 2 . The non-isothermal decomposition reaction kinetic parameters were obtained by both combined kinetic analysis and isoconversional Vyazovkin methods. It was found that Ni(DMG) 2 begins to decompose at around 280 ◦ C, and a sharp exothermic peak is observed in the DSC curve at about 308.2 ◦ C at a heating rate of 10 ◦ C · min − 1 . The main gaseous products are H 2 O, NH 3 , N 2 O, CO, and HCN, and the content of H 2 O is signiﬁcantly higher than that of the others. The activation energy obtained by the combined kinetic analysis method is 170.61 ± 0.65 kJ · mol − 1 . The decomposition process can be described by the random nucleation and growth of the nuclei model. However, it was challenging to attempt to evaluate the reaction mechanism precisely by one ideal kinetic model.


Introduction
Combustion catalysts, as one of the main ingredients, play an essential role in the modification of the burning rate and combustion stability of solid propellants [1][2][3]. A variety of combustion catalysts, such as transition metal oxides (e.g., CuO, PbO, NiO) [4][5][6][7][8], compounds (e.g., lead salicylate, copper adipate, copper salicylate) [9][10][11], and nano-metal particles (e.g., Ni, Bi) [12][13][14], have been up to now studied extensively. In the combustion process of solid propellants, the combustion catalyst acts as a heterogeneous catalyst, and its catalytic activity primarily depends on its active sites. Generally, the selection principle and method of the combustion catalyst are their catalytic activity and energy.
Although transition metal oxides such as NiO are a class of high-efficiency combustion catalysts for solid propellants and have excellent catalytic effect [15], they have no contribution to the combustion except for their catalytic effect. With the requirement to further enhance the burning rate, it is usually required to add more catalyst. Then the energy property of the solid propellants decreases because of a large amount of transition metal oxides, which cannot release any energy during combustion. Therefore, developing new energetic burning catalysts to increase the burn rate and reduce energy losses is crucial for the solid propellant. This problem could be overcome by using energetic combustion catalysts, in which an exothermic effect can be generated during the decomposition and combustion process of the propellants [16].
Bis-(dimethylglyoximato) nickel(II) complex, Ni(DMG) 2 , is usually synthesized by the reaction of dimethylglyoxime and Ni 2+ cations. This reaction has been much used to detect and estimate the presence of Ni in solution [17,18]. For the Ni(DMG) 2 complex, the thermal decomposition reaction is a rapid exothermic reaction, which produces a lot of heat during the process. Additionally, the main thermolysis products of Ni(DMG) 2 complex are NiO and carbon [19][20][21]. NiO by this method of formation in-situ usually has a higher catalytic activity than the direct addition of NiO. The carbon framework formed by the thermal decomposition of Ni(DMG) 2 complex can also increase the dispersion of NiO species on the burning surface of solid propellants, and then the catalytic efficiency of NiO can be further improved. Thus, the Ni(DMG) 2 complex can be used as a new type of energetic burning catalysts that not only enhances the burning rate, but also results in a low energy loss to the solid propellant system.
As is known, the thermal decomposition behavior of combustion catalysts has a significant effect on the catalytic property of combustion. The studies of the structure, decomposition kinetics, and thermal behaviors are very important for the selection and application of new combustion catalysts. Some of the previous studies on thermal decomposition reaction of Ni(DMG) 2 complex were mainly focused on the different properties of NiO nanoparticles, which were obtained by the thermal pyrolysis of the Ni(DMG) 2 precursor [19,21]. However, the correlation study between the thermal decomposition mechanism, the formation of gas products, and the reaction kinetics of Ni(DMG) 2 complex has been scarcely reported.
The main objective of this paper is to investigate the thermal decomposition reaction behavior and non-isothermal decomposition reaction kinetics of the Ni(DMG) 2 complex. The reaction models and kinetic parameters obtained by combined kinetic analysis and isoconversional Vyazovkin methods are also reported. Figure 1a,b represents the SEM images of the as-prepared Ni(DMG) 2 complex. It can be seen from Figure 1 that the surface morphology of the Ni(DMG) 2 is a nanofiber-like structure with different lengths, the average diameter is about 200 nm, and the fibers are in close proximity to each other. EDS is used to analyze the element types and the relative contents of atoms. The EDS image of Ni(DMG) 2 complex (Figure 1d)  propellants decreases because of a large amount of transition metal oxides, which cannot release any energy during combustion. Therefore, developing new energetic burning catalysts to increase the burn rate and reduce energy losses is crucial for the solid propellant. This problem could be overcome by using energetic combustion catalysts, in which an exothermic effect can be generated during the decomposition and combustion process of the propellants [16]. Bis-(dimethylglyoximato) nickel(II) complex, Ni(DMG)2, is usually synthesized by the reaction of dimethylglyoxime and Ni 2+ cations. This reaction has been much used to detect and estimate the presence of Ni in solution [17,18]. For the Ni(DMG)2 complex, the thermal decomposition reaction is a rapid exothermic reaction, which produces a lot of heat during the process. Additionally, the main thermolysis products of Ni(DMG)2 complex are NiO and carbon [19][20][21]. NiO by this method of formation in-situ usually has a higher catalytic activity than the direct addition of NiO. The carbon framework formed by the thermal decomposition of Ni(DMG)2 complex can also increase the dispersion of NiO species on the burning surface of solid propellants, and then the catalytic efficiency of NiO can be further improved. Thus, the Ni(DMG)2 complex can be used as a new type of energetic burning catalysts that not only enhances the burning rate, but also results in a low energy loss to the solid propellant system.

Morphology and Structure Characterization
As is known, the thermal decomposition behavior of combustion catalysts has a significant effect on the catalytic property of combustion. The studies of the structure, decomposition kinetics, and thermal behaviors are very important for the selection and application of new combustion catalysts. Some of the previous studies on thermal decomposition reaction of Ni(DMG)2 complex were mainly focused on the different properties of NiO nanoparticles, which were obtained by the thermal pyrolysis of the Ni(DMG)2 precursor [19,21]. However, the correlation study between the thermal decomposition mechanism, the formation of gas products, and the reaction kinetics of Ni(DMG)2 complex has been scarcely reported.
The main objective of this paper is to investigate the thermal decomposition reaction behavior and non-isothermal decomposition reaction kinetics of the Ni(DMG)2 complex. The reaction models and kinetic parameters obtained by combined kinetic analysis and isoconversional Vyazovkin methods are also reported. Figure 1a,b represents the SEM images of the as-prepared Ni(DMG)2 complex. It can be seen from Figure 1 that the surface morphology of the Ni(DMG)2 is a nanofiber-like structure with different lengths, the average diameter is about 200 nm, and the fibers are in close proximity to each other. EDS is used to analyze the element types and the relative contents of atoms. The EDS image of Ni(DMG)2 complex (Figure 1d   Ni(DMG) 2 complex was dried under vacuum for 2 h, and its XRD pattern is shown in Figure 2a. The very narrow and sharp high-intensity diffraction peak of the diffraction pattern with a preferentially exposed (110) plane at 2θ of 10.0 • , indicates a highly crystalline structure. The average grain size calculated by the Debye-Scherrer formula based on the most intense (110) diffraction peak was about 125.0 nm. The remaining peaks of Ni(DMG) 2 (200) [22][23][24][25]. From the diffraction pattern in Figure 2a, Ni(DMG) 2 is attributed to the orthorhombic system, Ibam space group with a = 16.575 Å, b = 10.423 Å, c = 6.474 Å and α = β = γ = 90 • [26][27][28][29], which confirms the good chemical purity of the materials. Ni(DMG)2 complex was dried under vacuum for 2 h, and its XRD pattern is shown in Figure 2a. The very narrow and sharp high-intensity diffraction peak of the diffraction pattern with a preferentially exposed (110) plane at 2θ of 10.0°, indicates a highly crystalline structure. The average grain size calculated by the Debye-Scherrer formula based on the most intense (110) diffraction peak was about 125.0 nm. The remaining peaks of Ni(DMG)2 observed at the 2θ values of 10.63, 19.35, 26.13, 27.52, 28.17, 29.36, 32. The FT-IR spectrum of the Ni(DMG)2 complex is shown in Figure 2b. The characteristic highest intensity absorption bands at 1240 and 1101 cm -1 observed in the sample can be attributed to the asymmetrical and symmetrical stretching vibrational modes of N-O. The strong absorption band at 1572 cm −1 can be assigned to the C=N stretching vibration mode of the sample [30,31], while the bands at 520 and 429 cm −1 can be attributed to the asymmetrical and symmetrical stretching vibrational modes of Ni-N [32,33]. The adsorption bands at 1368 and 990 cm −1 are attributed to the N-OH bond bending vibrational and N-O symmetric stretching vibrational models, respectively [32]. While the appearance of a band at 752 cm −1 is due to the C=N-O deformation vibration of the oxime group [34,35]. The broad absorption band at 3441 cm −1 is related to the O-H bonding stretching vibrational modes [32]. Furthermore, the weak band at 1788 cm −1 indicates a strong intramolecular hydrogen bonding in the Ni(DMG)2 complex [36][37][38][39]. Also, the absorption bands related to asymmetrical and symmetrical stretching vibrational modes of C-H are observed around 3046 and 2925 cm −1 , respectively [40]. These results verified the formation of the Ni(DMG)2 complex. The FT-IR spectrum of the Ni(DMG) 2 complex is shown in Figure 2b. The characteristic highest intensity absorption bands at 1240 and 1101 cm -1 observed in the sample can be attributed to the asymmetrical and symmetrical stretching vibrational modes of N-O. The strong absorption band at 1572 cm −1 can be assigned to the C=N stretching vibration mode of the sample [30,31], while the bands at 520 and 429 cm −1 can be attributed to the asymmetrical and symmetrical stretching vibrational modes of Ni-N [32,33]. The adsorption bands at 1368 and 990 cm −1 are attributed to the N-OH bond bending vibrational and N-O symmetric stretching vibrational models, respectively [32]. While the appearance of a band at 752 cm −1 is due to the C=N-O deformation vibration of the oxime group [34,35]. The broad absorption band at 3441 cm −1 is related to the O-H bonding stretching vibrational modes [32]. Furthermore, the weak band at 1788 cm −1 indicates a strong intramolecular hydrogen bonding in the Ni(DMG) 2 complex [36][37][38][39]. Also, the absorption bands related to asymmetrical and symmetrical stretching vibrational modes of C-H are observed around 3046 and 2925 cm −1 , respectively [40]. These results verified the formation of the Ni(DMG) 2 complex.

Morphology and Structure Characterization
The surface elemental components and chemical state of the obtained Ni(DMG) 2 complex were characterized by XPS. The survey XPS spectra of the Ni(DMG) 2 complex are shown in Figure 2c. The calculation results show that the atomic percent rate of N, O, C, and Ni are about 18.21%, 20.00%, 58.26%, and 3.53% in the Ni(DMG) 2 complex, respectively. Thus, it can be confirmed that the Ni(DMG) 2 complex is of high purity. As shown in Figure 2d, the main characteristic peaks at 853.3 eV and 870.5 eV are assigned to Ni 2p3/2 and Ni 2p1/2, respectively. The peaks centered at 857.0 eV and 874.0 eV are the satellite peaks of Ni 2p3/2 and Ni 2p1/2, respectively. Meanwhile, the peak of Ni 2p1/2 at 870.5 eV also indicates the oxidation state of Ni in as-prepared samples is in the Ni 2+ state [22]. This also further confirmed the formation of the Ni(DMG) 2 .

Thermal Decomposition Behavior
To get a better understanding of thermal decomposition characteristics of the Ni(DMG) 2 complex, TG-DSC methods were adopted to study the thermal decomposition reaction behaviors of the Ni(DMG) 2 complex. TG and DSC curves for Ni(DMG) 2 complex at a heating rate of 10 • C·min −1 under a high purity argon atmosphere at a flow rate of 50 mL·min −1 are shown in Figure 3. One can see that the thermal decomposition reaction proceeds rapidly within the temperature range of 280-330 • C with a large exothermic effect. The sharp exothermic peak at 308.2 • C is observed in the DSC curve. The exothermic peak area is about 686.3 J·g −1 . The total mass loss is about 42.3%. The color of the solid residue after the thermal decomposition reaction is black. When the Ni(DMG) 2 completely decomposes, the theoretical value of mass loss for the formation of Ni or NiO is about 79.7% or 74.1%, respectively. This indicates that there are many other residual products, such as carbon formed in this decomposition process.
Catalysts 2020, 10, x FOR PEER REVIEW 4 of 13 The surface elemental components and chemical state of the obtained Ni(DMG)2 complex were characterized by XPS. The survey XPS spectra of the Ni(DMG)2 complex are shown in Figure 2c. The calculation results show that the atomic percent rate of N, O, C, and Ni are about 18.21%, 20.00%, 58.26%, and 3.53% in the Ni-DMG complex, respectively. Thus, it can be confirmed that the Ni-DMG complex is of high purity. As shown in Figure 2d, the main characteristic peaks at 853.3 eV and 870.5 eV are assigned to Ni 2p3/2 and Ni 2p1/2, respectively. The peaks centered at 857.0 eV and 874.0 eV are the satellite peaks of Ni 2p3/2 and Ni 2p1/2, respectively. Meanwhile, the peak of Ni 2p1/2 at 870.5 eV also indicates the oxidation state of Ni in as-prepared samples is in the Ni 2+ state [22]. This also further confirmed the formation of the Ni(DMG)2.

Thermal Decomposition Behavior
To get a better understanding of thermal decomposition characteristics of the Ni(DMG)2 complex, TG-DSC methods were adopted to study the thermal decomposition reaction behaviors of the Ni(DMG)2 complex. TG and DSC curves for Ni(DMG)2 complex at a heating rate of 10 °C·min −1 under a high purity argon atmosphere at a flow rate of 50 mL·min −1 are shown in Figure 3. One can see that the thermal decomposition reaction proceeds rapidly within the temperature range of 280-330 °C with a large exothermic effect. The sharp exothermic peak at 308.2 °C is observed in the DSC curve. The exothermic peak area is about 686.3 J·g −1 . The total mass loss is about 42.3%. The color of the solid residue after the thermal decomposition reaction is black. When the Ni(DMG)2 completely decomposes, the theoretical value of mass loss for the formation of Ni or NiO is about 79.7% or 74.1%, respectively. This indicates that there are many other residual products, such as carbon formed in this decomposition process. To investigate the thermal decomposition mechanism of Ni(DMG)2 complex, gas products released during the thermal decomposition of the Ni(DMG)2 complex were detected simultaneously by mass spectra and FT-IR spectroscopy. The mass spectra of the principal gaseous products of Ni(DMG)2 compound are shown in Figure 4a. It can be seen that the gas released from the Ni(DMG)2 complex begins to appear at about 280 °C. The ion current number (m/z) of the major process products are 18, 17, 44, and 28, among which the ion current intensity of m/z = 18 is the highest. It is believed to be H2O as the dimethylglyoxime can undergo a dehydration reaction to produce a heterocyclic molecule [41]. The N-OH is also the weakest bond in the dimethylglyoxime molecule, so the H2O forms very easily during the thermolysis of Ni(DMG)2. The gas product with m/z = 17 could be NH3. For the m/z of 44, two possible species are CO2 or N2O. The MS signals at m/z = 28, can be attributed to CO or N2. It can also be seen that there are some weak ion currents of m/z = To investigate the thermal decomposition mechanism of Ni(DMG) 2 complex, gas products released during the thermal decomposition of the Ni(DMG) 2 complex were detected simultaneously by mass spectra and FT-IR spectroscopy. The mass spectra of the principal gaseous products of Ni(DMG) 2 compound are shown in Figure 4a. It can be seen that the gas released from the Ni(DMG) 2 complex begins to appear at about 280 • C. The ion current number (m/z) of the major process products are 18, 17, 44, and 28, among which the ion current intensity of m/z = 18 is the highest. It is believed to be H 2 O as the dimethylglyoxime can undergo a dehydration reaction to produce a heterocyclic molecule [41]. The N-OH is also the weakest bond in the dimethylglyoxime molecule, so the H 2 O forms very easily during the thermolysis of Ni(DMG) 2 . The gas product with m/z = 17 could be NH 3 . For the m/z of 44, two possible species are CO 2 or N 2 O. The MS signals at m/z = 28, can be attributed to CO or N 2 . It can also be seen that there are some weak ion currents of m/z = 12, 15, and 27 during the decomposition process, which indicate the existence of C, CH 3 • , and HCN. The HCN may come from the reaction of NH 3 and C at high temperature.
Catalysts 2020, 10  To further confirm the gaseous products of Ni(DMG)2 complex, FT-IR spectroscopy was used. Figure 4b shows the FT-IR results of the main gaseous products at a different temperatures. Two significant changes in the FT-IR spectra can be found in the wavenumber region of 650-1000 cm −1 and 2000-2400 cm −1 . In the region between 650 and 1000 cm −1 , some new absorption peaks of gaseous products appear from 288.7 °C. An absorption peak at 668 cm −1 is attributed to the vibration of HCN, and the double peaks at 930 and 965 cm −1 are attributed to the vibrations of NH3. Combined with the MS signals of m/z = 28 in Figure 4a, we can observe that the absorbance peaks at 2107 and 2185 cm −1 should be CO. Two bands at 2291 and 2355 cm −1 are assigned to CO2 vibrations, which indicate that the MS signal of m/z = 44 in Figure 4a may be CO2 [42]. However, NH3 can be formed from the C=N bond only as a result of a reduction reaction, and the formation of CO2 is the result of deep oxidation of the ligand. These reactions cannot occur simultaneously. Additionally, there is an excess of reducing elements in the Ni(DMG)2 molecule and the formation of CO2 during the thermolysis of Ni(DMG)2 is unlikely. Most likely, this is atmospheric CO2 or the oxidation product of Ni(DMG)2 by air. So, the MS signal of m/z = 44 in Figure 4a must be N2O. Generally, the characteristic absorption peak of N2O is at about 2237 cm −1 , it can be affected by the absorption peak of CO2 at 2291 cm −1 . Thus, N2O is difficult to distinguish from the FT-IR results. According to Figure  4b, the peaks at 2939 and 2968 cm −1 are generally related to CH3 vibrations, corresponding to m/z = 15 (Figure 4a). Therefore, by combining the results of the MS and FT-IR spectra, we can confirm that the main gaseous products of Ni(DMG)2 complex are H2O, NH3, N2O, CO, and HCN.

Non-Isothermal Decomposition Reaction Kinetics
Usually, the solid-state reaction rate (dα/dt) can be described as: where T is the temperature, α is the extent of conversion, A is the pre-exponential factor, t is the time, R is the gas constant, E is the activation energy, and f( α) is the reaction model. Some of the solid-state reaction kinetic models are listed in the literature [43]. Under the non-isothermal conditions, the experiments were carried out at a constant reaction rate (β=dT/dt). Then Equation (1) can be rearranged as follows: To further confirm the gaseous products of Ni(DMG) 2 complex, FT-IR spectroscopy was used. Figure 4b shows the FT-IR results of the main gaseous products at a different temperatures. Two significant changes in the FT-IR spectra can be found in the wavenumber region of 650-1000 cm −1 and 2000-2400 cm −1 . In the region between 650 and 1000 cm −1 , some new absorption peaks of gaseous products appear from 288.7 • C. An absorption peak at 668 cm −1 is attributed to the vibration of HCN, and the double peaks at 930 and 965 cm −1 are attributed to the vibrations of NH 3 . Combined with the MS signals of m/z = 28 in Figure 4a, we can observe that the absorbance peaks at 2107 and 2185 cm −1 should be CO. Two bands at 2291 and 2355 cm −1 are assigned to CO 2 vibrations, which indicate that the MS signal of m/z = 44 in Figure 4a may be CO 2 [42]. However, NH 3 can be formed from the C=N bond only as a result of a reduction reaction, and the formation of CO 2 is the result of deep oxidation of the ligand. These reactions cannot occur simultaneously. Additionally, there is an excess of reducing elements in the Ni(DMG) 2 molecule and the formation of CO 2 during the thermolysis of Ni(DMG) 2 is unlikely. Most likely, this is atmospheric CO 2 or the oxidation product of Ni(DMG) 2 by air. So, the MS signal of m/z = 44 in Figure 4a must be N 2 O. Generally, the characteristic absorption peak of N 2 O is at about 2237 cm −1 , it can be affected by the absorption peak of CO 2 at 2291 cm −1 . Thus, N 2 O is difficult to distinguish from the FT-IR results. According to Figure 4b, the peaks at 2939 and 2968 cm −1 are generally related to CH 3 vibrations, corresponding to m/z = 15 (Figure 4a. Therefore, by combining the results of the MS and FT-IR spectra, we can confirm that the main gaseous products of Ni(DMG) 2 complex are H 2 O, NH 3 , N 2 O, CO, and HCN.

Non-Isothermal Decomposition Reaction Kinetics
Usually, the solid-state reaction rate (dα/dt) can be described as: where T is the temperature, α is the extent of conversion, A is the pre-exponential factor, t is the time, R is the gas constant, E is the activation energy, and f (α) is the reaction model. Some of the solid-state reaction kinetic models are listed in the literature [43]. Under the non-isothermal conditions, the experiments were carried out at a constant reaction rate (β = dT/dt). Then Equation (1) can be rearranged as follows: The kinetic triplet, E, A, and f (α), is very important for kinetic studies, while it can be used to explain the reaction mechanism and make a kinetic prediction [44][45][46]. It is also the main aim of kinetic studies to obtain a kinetic triplet. The combined kinetic analysis method has been proved to be a promising method for determining the three kinetic parameters. It can determine all the kinetic parameters from the simultaneous treatment of experimental data, which can be obtained under any heating conditions [47][48][49]. In the combined kinetic analysis method, the reaction kinetic model can be described in the following general form: This equation is commonly called the modified Šesták-Berggren empirical equation [50]. Equation (3) can acts as an umbrella. The most common physical models and their possible deviations from ideal conditions can be fitted by merely adjusting the parameters c, n, and m [47].
By substituting Equation (3) into Equation (2) and taking the logarithm, Equation (2) can be rearranged as the following equation: To obtain the parameters of Equation (4), one should simultaneously substitute the experiment data of α-T and (dα/dT)-T obtained at different heating rates. When the linear correlation coefficient (r) between the left-hand side of Equation (4) and the reciprocal temperature (1/T) reaches the maximum, the parameters c, n, and m can be obtained by the linear fitting method. Considering that the relative experimental errors of the kinetic data are larger at α < 0.1 and α > 0.9, only the data in the range of α = 0.1-0.9 were considered. Although the f (α) used in this method is an empirical kinetic equation, it does not have any previous assumption on the kinetic model of the reaction. It is very convenient to investigate the complex decomposition reactions, in which the f (α) estimated by experiment cannot match any ideal kinetic models. Figure 5a shows the DSC curves of the Ni(DMG) 2 complex in N 2 flow at different linear heating rates of 5, 8, 10, and 12 • C·min −1 . The DSC parameters of Ni(DMG) 2 complex at different heating rates are listed in Table 1. It can be observed that the exothermic peaks are well-formed. According to Table 1, the onset, peak, and end temperatures (T onset , T p, and T end ) moved to high temperature with the increase of heating rate, respectively. The heat release (∆Q) of the thermal decomposition process decreases with the increase of heating rate. The decomposition heat at a heating rate of 10 • C·min −1 is 1412 J·g −1 , and the decomposition peak temperature (309.2 • C) is very close to the TG-DSC result (peak temperature of 308.2 • C) in Figure 3 at the same heating rate.
heating rates of 5, 8, 10, and 12 °C·min −1 . The DSC parameters of Ni(DMG)2 complex at different heating rates are listed in Table 1. It can be observed that the exothermic peaks are well-formed. According to Table 1, the onset, peak, and end temperatures (Tonset, Tp, and Tend) moved to high temperature with the increase of heating rate, respectively. The heat release (ΔQ) of the thermal decomposition process decreases with the increase of heating rate. The decomposition heat at a heating rate of 10 °C·min −1 is 1412 J·g −1 , and the decomposition peak temperature (309.2 °C) is very close to the TG-DSC result (peak temperature of 308.2 °C) in Figure 3 at the same heating rate.  The corresponding non-isothermal conversion (α-T) curves for the thermal decomposition process of Ni(DMG)2 complex at the different heating rates are shown in Figure 5b. It can be seen that all α-T curves have a sigmoidal character, and they are moved to higher temperature with the heating rate increase. The α-T data, in the range of 0.1 < α < 0.9, were subjected to combined kinetic analysis studies for determining the kinetic parameters of the reaction. Figure 5c shows the calculation results of the reaction by the combined kinetic analysis method through the plot of the left-hand side of Equation (4) versus 1/T. The kinetic parameters obtained by the combined kinetic analysis methods are also summarized in Table 1. As it can seen, the experimental data in the range of α fit closely to a straight line, the linear correlation coefficient (r) can achieve 0.9977 for the fitting parameters n = 0.860 and m =1 .009. The activation energy and ln(cA) obtained from the slope and intercept of the fitting line are 170.61 ± 0.65 kJ·mol −1 and 36.46 ± 0.13, respectively.
To further confirm the possible reaction mechanism functions, the differential master plot method, which is defined by Equation (5), can be applied [51]. Figure 5d shows the theoretical f(α)/f(0.5) curves for different models (lines), and four experimental f(α)/f(0.5) results (symbols) obtained by using Equation (5). It can be seen from Figure 5d that the experimental f(α)/f(0.5) results  The corresponding non-isothermal conversion (α-T) curves for the thermal decomposition process of Ni(DMG) 2 complex at the different heating rates are shown in Figure 5b. It can be seen that all α-T curves have a sigmoidal character, and they are moved to higher temperature with the heating rate increase. The α-T data, in the range of 0.1 < α < 0.9, were subjected to combined kinetic analysis studies for determining the kinetic parameters of the reaction. Figure 5c shows the calculation results of the reaction by the combined kinetic analysis method through the plot of the left-hand side of Equation (4) versus 1/T. The kinetic parameters obtained by the combined kinetic analysis methods are also summarized in Table 1. As it can seen, the experimental data in the range of α fit closely to a straight line, the linear correlation coefficient (r) can achieve 0.9977 for the fitting parameters n = 0.860 and m =1.009. The activation energy and ln(cA) obtained from the slope and intercept of the fitting line are 170.61 ± 0.65 kJ·mol −1 and 36.46 ± 0.13, respectively.
To further confirm the possible reaction mechanism functions, the differential master plot method, which is defined by Equation (5), can be applied [51]. Figure 5d shows the theoretical f (α)/f (0.5) curves for different models (lines), and four experimental f (α)/f (0.5) results (symbols) obtained by using Equation (5). It can be seen from Figure 5d that the experimental f (α)/f (0.5) results at different heating rates are not in good agreement with any one of the theoretical f (α)/f (0.5) curves. However, it should be noted that the experimental curve is in partial agreement with random nucleation and 3D growth of nuclei (A3) model in the range of α < 0.5 and random nucleation and 2D growth of nuclei (A2) model in the range of α > 0.5. This may indicate that the decomposition process of Ni(DMG) 2 complex is controlled by two processes. In the initial decomposition process, it follows the random nucleation and 3D growth of the nuclei model. As the reaction proceeds, the decomposition process gradually shifts to the 2D growth of the nuclei model.
Generally, the dependence of apparent activation energy (E α ) on α, obtained by the isoconversional methods, is important for distinguishing and treating the multistep reaction kinetics. A significant variation of E α -α indicates that the reaction process is kinetically complex, and one cannot apply any one of the ideal kinetic models to describe a possible reaction mechanism of such a process throughout the whole range of experimental conversions and temperatures [43]. To obtain the dependence of E α -α, the integral isoconversional method developed by Vyazovkin [52][53][54] was used. For a series of runs performed at different heating rates, the E α value can be determined by minimizing the following function [52]: where the temperature integral is solved numerically using the Runge-Kutta method. Then the dependence of E α on α can be obtained by minimizing the Φ(E α ) function repeatedly for each value of α. Figure 6 shows the dependence of E α on α obtained by the Vyazovkin method for the thermal decomposition of Ni(DMG) 2 complex. It indicates that the activation energy of the decomposition process has significantly changed at high and low conversions. The reason is that the relative experimental errors of the kinetic data are larger at α < 0.1 and α > 0.9. The average activation energy is 167.80 ± 1.49 kJ mol −1 in the range of 0.1 ≤ α ≤ 0.9, which is almost identical to the result obtained by the combined kinetic analysis method. Also note that, the change of E α in the range of 0.1 < α < 0.6 was significantly greater than the 0.6 < α < 0.9. This indicates that the thermal decomposition process of Ni(DMG) 2 complex was a complex multistep reaction, so it is extremely difficult to properly evaluate the reaction mechanism by the ideal kinetic model. energy is 167.80 ± 1.49 kJ mol −1 in the range of 0.1 ≤ α ≤ 0.9, which is almost identical to the result obtained by the combined kinetic analysis method. Also note that, the change of Eα in the range of 0.1 < α < 0.6 was significantly greater than the 0.6 < α < 0.9. This indicates that the thermal decomposition process of Ni(DMG)2 complex was a complex multistep reaction, so it is extremely difficult to properly evaluate the reaction mechanism by the ideal kinetic model.  Figure 6. E α -α curve for the thermal decomposition of Ni(DMG) 2 complex.

Materials
Dimethylglyoxime (DMG), nickel chloride hexahydrate (NiCl 2 ·6H 2 O), and absolute ethanol were all obtained from the Aladdin Chemical Reagent Co., Ltd. (Shanghai, China). All of the chemical reagents were analytical grade, commercially available and used without purification further.

Synthesis of Ni(DMG) 2
An amount of 100 mL absolute ethanol was mixed with 1.0015 g dimethylglyoxime (DMG) and 1.0303 g nickel chloride hexahydrate (NiCl 2 ·6H 2 O), followed by stirring until completely dissolved. The mixed solution was poured into a round bottom flask and the pH adjusted of the solution to pH 8. Then the solution was heated and stirred at 85 • C for 30 min, and reacted for 2 h at 55 • C with no agitation. The obtained precipitate was washed with ethanol. The obtained red precipitate was Ni(DMG) 2 . Finally, the dried sample was heated at 400 • C for 2 h and kept in an argon atmosphere. Elemental analysis (%), calculated for Ni(DMG) 2

Apparatus and Measurements
Field-emission environment scanning electron microscope (SEM, FEI Quanta 600, Hillsboro, OR, USA) was used to characterize the morphology of the sample. Energy-dispersive X-ray spectroscopy (EDS, OXFORD INCA Penta FET × 3) was used to roughly examine the composition, which was attached with SEM.
The phase structure and purity of the Ni(DMG) 2 were confirmed by powder X-ray diffraction (XRD, PANalytical X'Pert PRO, Amsterdam, Netherlands). The radiation source of the X-ray powder Diffractometer is Cu K α (λ = 1.5418 Å) at 40 kV and 40 mA. The testing temperature was ambient temperature, and the range of 2θ was 5 • to 90 • with steps of 0.05 • .
A Fourier Transform Infrared (FT-IR, Bruker Tensor 27, Karlsruhe, Germany) spectrometer was used to record the infrared spectra. The samples were pressed into pellets. The mass ratio of KBr pellets with KBr to sample was approximately 30:1. The spectral range was 4000-400 cm −1 with a resolution of 4 cm −1 .
The information of element composition was obtained by X-ray photoelectron spectroscopy (XPS, NEXSA, Thermo scientific, Waltham, MA, USA) instrument and the radiation source was Al K α (1468.68 eV). The binding energy of the C1s peak at 284.6 eV was used to calibrate the obtained XPS peaks.
The thermal decomposition behavior and kinetic analysis of the sample were investigated by differential scanning calorimetry (DSC). A NETZSCH DSC204 instrument (Selb, Bavaria, Germany) equipped with a water-cooling system which was used to carry out the DSC measurement. The instrument was calibrated by an indium standard. In this work, the mass of each sample was about 1.0 mg. The test temperature range was room temperature to 500 • C. The heating rates were 5.0, 8.0, 10.0, and 12.0 • C·min −1 . The carrier gas was nitrogen at a flow rate of 50 mL·min −1 .
Thermal decomposition and evolved gas analysis of Ni(DMG) 2 were performed using thermogravimetric analysis, differential scanning calorimetry, Fourier transform infrared spectrometer and mass spectrometer (TG-DSC-FTIR-MS) coupling technique with a STA449C thermal analyzer (NETZSCH, Selb, Bavaria, Germany), a QMS-403C mass spectrometer (NETZSCH, Selb, Bavaria, Germany), and a 5700 infrared spectrometer (Nicolet, Madison, WI, USA). The thermal decomposition measurements were performed under a high purity argon atmosphere at a flow rate of 50 mL·min −1 . The test temperature range was 30 to 500 • C, and the heating rate was 10 • C·min −1 . The sample mass was about 2.0 mg. The transfer tubes were used to directly transfer the evolved gases during the thermal decomposition process, and finally detected by MS and FT-IR.

Conclusions
As a potential combustion catalyst, the fiber-like Ni(DMG) 2 complex was successfully synthesized and characterized by SEM-EDS, FT-IR, XRD, and XPS. The thermal decomposition behavior and gaseous products of the Ni(DMG) 2 complex were characterized by TG-DSC-FTIR-MS techniques. The decomposition reaction of Ni(DMG) 2 occurs in the temperature range 280-330 • C with a large exothermic effect. The experimental total mass loss during the thermal decomposition of Ni(DMG) 2 is lower than the theoretical value of complete decomposition. The gas products released during the thermal decomposition were detected simultaneously by MS and FT-IR, and the results show that the main gas products are H 2 O, NH 3 , N 2 O, CO, and HCN. The ion current intensity of H 2 O (m/z = 18) is the highest according to the MS results.
The non-isothermal decomposition reaction kinetic parameters were obtained by both isoconversional Vyazovkin methods and the combined kinetic analysis method. The activation energy obtained by the combined kinetic analysis method was 170.61 ± 0.65 kJ·mol −1 , which is almost the same as that obtained by the isoconversional Vyazovkin method. The initial decomposition process follows the random nucleation and three-dimensional growth of the nuclei model. As the reaction proceeds, the decomposition process gradually shifts to the two-dimensional growth of the nuclei model. The dependence of E α on α obtained by the Vyazovkin method also indicates that the thermal decomposition process is a complex multistep reaction.