Thermal Decomposition Behaviour of Foundry Sand for Cast Steel in Nitrogen and Air Atmospheres

Sand casting is the most widely used casting technique, known for ages, even since ancient times. 4e main goal of this study was to determine the thermal decomposition behaviour of foundry sand for cast steel. We first tested the basic properties of foundry sand, including its proximate analysis, chemical composition, and particle size characteristics; we next monitored the thermal decomposition behaviour of foundry sand for cast steel via simultaneous thermal analysis. We focused on the mass loss of foundry sand for cast steel at different heating rates in nitrogen and air atmospheres. We adopted a novel method to calculate the volatile release characteristic index of foundry sand. 4e volatile content of foundry sand for cast steel was very low, so the volatile release characteristic index of the sand could not be strictly calculated according to this concept. We calculated the thermal decomposition kinetics parameters of foundry sand, namely, the activation energy and preexponential factor, under kinetics theory. To thoroughly test the fitting effect, we conducted a single-factor analysis of variance on the source of error. 4e results showed that the independent variable has a significant influence on the dependent variable and that the fitting equation we selected is feasible and effective.


Introduction
Since the reform and opening-up policies of recent decades, Chinese industries have rapidly developed [1][2][3], and the national economy has grown increasingly strong. e foundry industry is the foundation of the manufacturing industry and represents a basic guarantee of innovative, advanced technological equipment; to this effect, it is also an indispensable part of the national economy. China is a good example of a country that is well supported by its foundry industry. Foundry manufacturing is a process for preparing castings by pouring a liquid alloy into a cavity and then solidifying it gradually by refrigeration [4][5][6]. Despite their contribution to socioeconomic development, foundries also result in significant environmental pollution [7][8][9] and even casualties [10][11][12]. Sand casting occupies up to 90% of the foundry industry [13]. Research regarding the thermal behaviour of sand casting is significant with respect to safe production and casting quality. e heat conduction that occurs when molten alloy is poured into the cavity causes difficulty in researching sand casting thermal behaviour [14]. e local temperature of foundry sand during alloy casting exceeds 1600°C [15], so it is difficult to measure changes in sand temperature during casting. Numerical simulation is now a common approach to such research [16]; Procast is especially popular for simulating flow, temperature, and stress fields during the casting process [17,18]. e use of Procast simulations can shorten the production cycle. Zhang, for example, established an inverse conduction model with Procast to determine the heat flux and heat transfer coefficient at the metalsand mould interface in lost foam casting [18]. Although Procast can simulate temperature changes during the casting process, it unfortunately is not applicable to the thermal decomposition behaviour of foundry sand.
Previous thermal decomposition characteristics research has mainly centered around coal, wood, rice straw, and rice husk materials [19][20][21]. Parameters that are significant to the temperature-rise period of the materials (e.g., mass loss, decalescence, and heat release) can be monitored with a thermal analyser, as can the influence of heating rates and thermal decomposition atmospheres on the thermal decomposition characteristics. In addition, dynamic heat and vapor transport in a typical three-dimensional (3D) building structure, involving different types of environment-friendly concrete mixtures, had been simulated by using finite elements [22]. Coronado studied the thermal behaviour of foundry sand dust and monitored the acid gas emissions [23]; however, no previous researcher has investigated the thermal decomposition characteristics of foundry sand for cast steel. We hope to fill this gap.
During the pouring process, the heat conduction phenomena between the molten alloy and casting mould cause the mould temperature to rise rapidly. Foundry sand at the sprue is mainly thermally decomposed in an air atmosphere, and foundry sand in the cavity is mainly thermally decomposed in a nitrogen atmosphere. erefore, thermal decomposition experiments of foundry sand for cast steel were carried out in air and nitrogen atmospheres, respectively. e volatile release characteristic index is adopted to represent the pyrolysis behaviour of foundry sand [18]. A larger value indicates better sand release characteristics and a more facile pyrolysis reaction in the sample. To calculate the volatile release characteristic index of foundry sand, it is necessary to determine the initial temperature of volatile emission and mass loss corresponding to a mass loss rate of 0.1 mg/min [19]. e sample mass used for thermal decomposition experiments is usually 5 mg [21]; the initial temperature corresponds to a mass loss rate of 2 wt. % (min). SiO 2 comprises 96.62 wt. % of foundry sand [24], so the volatile content of foundry sand is very low, and the volatile release characteristic index cannot be strictly calculated according to this concept. In this paper, we adopt a novel method to calculate the volatile release characteristic index of foundry sand for cast steel. e thermal decomposition experiment presented here investigates the physical and chemical characteristics of foundry sand under heating, and thermodynamic and kinetic principles are applied during the research process. e thermal decomposition kinetic parameters include the activation energy and preexponential factor [25]. e activation energy refers to the energy needed for molecules to move from the normal state to an active state that is prone to chemical reaction [26], and the activation energy effectively reflects the stability of the material. e preexponential factor is determined by the thermal decomposition experiment, regardless of the reaction temperature and the concentration of the substance. e activation energy under kinetics theory can be calculated according to thermal decomposition data [27]. In the present study, the thermal decomposition kinetic parameters of foundry sand were calculated according to kinetics theory. After determining the thermal decomposition kinetics parameters of foundry sand, we roughly determined the fitting effect according to Pearson's R [19,21]. Xu et al. calculated the thermal decomposition kinetic parameters of foundry sand for cast iron under kinetics theory [27], but they did not test the validity of the calculation results. To rigorously test the fitting effect, we conducted single-factor analysis of variance on the source of error [28].
e main purpose of this study was to investigate the thermal decomposition behaviour of foundry sand for cast steel in nitrogen and air atmospheres. Before the thermal decomposition experiment, some important parameters were tested, such as proximate analysis data and chemical composition.
en, the thermal decomposition curves of foundry sand at different heating rates and atmospheres were analysed in detail. A new method was adopted to calculate the volatile release characteristic index of foundry sand, and its change trends in different heating rates and atmospheres were investigated. e thermal decomposition kinetics parameters of foundry sand were calculated under kinetics theory, and single-factor analysis of variance was carried out to test the validity of the calculation results.

Foundry Sand for Cast Steel
e foundry sand for cast steel that we used in this experiment was obtained from a research institute in Shenyang, China. We performed proximate analysis of the sample to ensure comprehensive investigation of its thermal decomposition behaviour in a high-temperature environment (TRGF-8000 Automatic Proximate Analyzer, Electronic and Technology Co., Ltd., of Tianrun Hebi, China). e foundry sand we tested is mainly composed of ash (98.82 wt. %), volatiles (2.15 wt. %), and moisture (0.03 wt. %). e fixed carbon content of the foundry sand was not detectable. e ash fraction indicates incombustible substances, or the residual offscourings after a given substance are completely burned. e main composition of foundry sand is ash, indicating that foundry sand is stable at high temperatures.

Chemical Composition Analysis.
e chemical composition of the foundry sand mainly includes SiO 2 . We determined the chemical composition of the sand on a ZSX Primus X-ray fluorescence (XRF) device (Rigaku Co., Japan). e XRF results are shown in Table 1.
One of the most important properties of foundry sand is stability at high temperatures [29][30][31]. e chemical composition of the foundry sand is mainly SiO 2 , indicating that the foundry sand is stable at high temperatures. e chemical composition analysis results of the foundry sand were consistent with the proximate analysis.

2.2.
ermal Decomposition Experiments. We conducted thermogravimetric analysis of the foundry sand on an STA449 F3 sensitive thermal balance (NETZSCH Group, Germany). e thermogravimetric (TG) curve was measured to obtain the weight loss, and a differential scanning calorimeter (DSC) was used to obtain phase transitions. e first derivative of the TG profile, that is, derivative thermogravimetric (DTG) curve, was used to obtain the weight loss rate. We conducted thermal decomposition experiments in a nitrogen atmosphere at heating rates of 20, 30, and 40°C/min and in an air (oxygen : nitrogen � 1: 4) atmosphere at a heating rate of 20°C/min. e thermal decomposition characteristics of foundry sand at temperatures above 100°C were the focus of this experiment due to the rapid temperature increase in foundry sand during pouring of the liquid alloy into the cavity; the temperature range of the thermal decomposition experiments was from 35°C to 1100°C. One sample was measured in each iteration of the experiment. e statistical data supporting this paper can be found in the electronic supplementary materials.

Volatile Release Characteristic Index.
e volatile release characteristic index was adopted to represent the pyrolysis behaviour of foundry sand [18]. A larger value indicates better sand release characteristics and a more facile pyrolysis reaction in the sample. e calculation procedures of the volatile release characteristic index were as follows.
Without regard to moisture release in the initial stage and inorganic salt decomposition in the later period, the volatile release characteristic index D is calculated as follows: where T s indicates the initial temperature of volatile emission and mass loss corresponding to a mass loss rate of 0.1 mg/min [19] in°C. T max indicates the temperature corresponding to the maximum mass loss rate in°C.

Kinetic eory.
e thermal decomposition data obtained under different heating rates and atmospheres can be used for the determination of pyrolysis reaction kinetics. e integral Coats-Redfern method, which has been successfully employed for investigations of the pyrolysis kinetic parameters, was employed for mathematical analysis [25]. e degradation rate is defined according to the equation of an Arrhenius-type kinetic model, and first-order-reactionbased Arrhenius theory is commonly employed in the kinetic analysis of pyrolysis [26]. e kinetic theory of the thermal decomposition of foundry sand is based on the rate of conversion [32]: where dα/dt denotes the conversion rate, calculated from the rate constant based on temperature k(T) and the reaction function f(α). e parameter α denotes the conversion of the reaction process: where m 0 , m ∞ , and m t denote the initial mass of the sample, the final mass of the sample, and the mass of the sample at a specific time t, respectively. e rate constant of foundry sand based on the first-order reaction is described as follows under the Arrhenius law [33]: where A denotes the preexponential factor, min −1 ; E denotes the activation energy, KJ/mol; R denotes the universal gas constant, 8.314 J/mol/K; and T denotes the thermodynamic temperature, K. e reaction function is expressed as follows [34]: For a constant heating rate β � dT/dt, the heating rate and equations (4) and (5) can be inserted into equation (2), which can then be rewritten as follows: e following equation can be obtained based on the Coats-Redfern method [34]: e term 2RT/E is far less than 1 and can be neglected, so equation (7) can be simplified as follows: For a given heating rate, the term −ln[−ln(1 − α)/T 2 ] varies linearly as 1/T with a slope of E/R; the intercept is −ln(AR/βE). e pyrolysis kinetic parameters, namely, activation energy E and preexponential factor A, can be calculated by equation (8).

Single-Factor Analysis of Variance.
Single-factor analysis of variance is a commonly used statistical approach to analyse the sources of error in data. e goal of single-factor analysis of variance is to determine whether the independent variable has a significant influence on the dependent variable by comparing the average of multiple normal distribution samples for equality [28]. Let factor A have r levels, that is, A 1 , A 2 , . . . , A r ; x ij indicates the jth observed value of the ith level. e specific procedures of single-factor analysis of variance are as follows [28].
First, two assumptions are made: indicating that the factor has less of an effect on the experimental results than does random error H 1 : μ 1 , μ 2 , · · · , μ k are not completely the same, indicating that the factor has a greater effect on the experimental results than does random error en, calculate the average value of the factor at different levels as follows: where n i indicates the total quantity of experimental data points. e average value of the factor at all levels can be determined as follows: where n � n 1 + n 2 + · · · n k .
To construct a statistical test, it is also necessary to calculate the total error sum of squares SST, factor error sum of squares SSA, and random error sum of squares SSE: e relationship among SST, SSA, and SSE is as follows: ese three error sums of squares are affected by the number of observations. e error sum of squares is larger when there are more observations. It is necessary to divide the error sum of squares by the corresponding degrees of freedom to eliminate the effect of observation quantity on the error sum of squares to determine the mean squares. e degrees of freedom for SST, SSA, and SSE are n−1, k−1, and n−k, respectively. e mean square of SSA is the "model mean square," MSA, which can be calculated as follows: e mean square of SSE is the "error mean square," MSE, which can be calculated as follows: e ratio of MSA to MSE obeys the F-distribution and is written as follows: e calculation results are shown in Table 2. Next, the critical value F α (k − 1, n − k) is determined according to the significance level α and F-distribution tables. Statistical decisions are made according to the test statistics F and critical value F α (k − 1, n − k). H 0 is rejected if F > F α , which indicates that the factor has a greater effect on the experimental results than does random error. H 0 is accepted if F < F α , which indicates that the factor has less of an effect on the experimental results than does random error.

Proximate Analysis of Foundry Sand.
ere is no previously published literature regarding the proximate analysis of foundry sand. Previous experimental materials submitted to proximate analysis include mainly coal, wood, rice straw, and rice husk [19][20][21]. e range of fixed carbon in coal varies widely, mainly because the metamorphic grade of different coal types differs. e fixed carbon content of foundry sand in this study was not detectable, indicating that pulverized coal was not added during the sand mixing process.
In contrast to other materials, such as coal, wood, rice straw, and rice husk [19][20][21], the main composition of foundry sand is ash. When pouring alloy metal into the cavity of a foundry, the adhesives, curing agents, and coatings in the sand quickly vaporize and pyrolyze, and the remaining substance is mainly ash. e greater the ash content, the more stable the material.

Chemical Composition Analysis of Foundry Sand.
e XRF results indicate that there were 18 chemical components in the foundry sand, in which the content of SiO 2 (94.6021 wt. %) was the largest; the main component was SiO 2 according to the proximate analysis. e SiO 2 content of silica sand can be as high as 96.62 wt. % [24]. Other materials (e.g., resin) were added in the process of sand mixing, which decreased the SiO 2 content to some extent. A small amount of Fe 2 O 3 was added to improve the hightemperature strength of the foundry sand. e chemical composition of foundry sand dust was tested, and the content of SiO 2 was only 50 wt. % [23]. Many impurities were introduced into the foundry sand dust in the process of sand shakeout, which drastically reduced SiO 2 .
e chemical composition of foundry sand mainly includes SiO 2 , and SiO 2 is stable in high-temperature environments, which will benefit the maintenance of the shape of the casting mould.

ermal Decomposition Behaviour of Foundry Sand in a Nitrogen
Atmosphere.
e mass of foundry sand gradually decreases during thermal decomposition, accompanied by heat absorption or release. e thermal decomposition data for foundry sand in nitrogen atmosphere can be found in Supplementary Materials S1, S2, and S3. e thermogram of our foundry sand sample at a heating rate of 20°C/min in a nitrogen atmosphere is shown in Figure 1.
In Figure 1, "TG" denotes "thermogravimetric," "DTG" is "derivative thermal gravimetric," and "DSC" is "differential scanning calorimetry." A downward DSC curve indicates heat release, and an upward curve indicates the absorption of heat.
As shown in Figure 1, the thermal decomposition of foundry sand can be divided into three stages. e first mass loss stage mainly occurred below 220°C, where the sand mass slowly decreased. e DSC curve shows an endothermic state, but no obvious endothermic peak appears. e thermal decomposition results of foundry sand dust indicated an endothermic peak in the DSC curve at 127°C due to the release of absorbed water [23]. e moisture content of foundry sand in this study was as low as 0.03 wt. %; there was no obvious endothermic peak corresponding to the release of moisture during thermal decomposition. e second mass loss stage took place from 220°C to 800°C, where volatile devolatilization occurred. ere is an endothermic peak at 573°C due to the transformation of quartz from the β to the α phase. e binding resin was pyrolyzed to produce gas molecules such as CH 4 and CO in this stage. e mass loss in this pyrolysis phase substantially contributed to the total mass loss, approximately 81 wt. %. Foundry sand dust was in an endothermic state at temperatures above approximately 200°C in an air atmosphere in a previous report [23], but our results were dissimilar. is difference was mainly due to the oxidation or combustion of impurities in the foundry sand dust sample.
e thermal decomposition of our sand was mainly attributable to its binding agent, and there was no obvious oxidation or combustion reaction in the nitrogen atmosphere. e third thermal decomposition stage mainly took place above 800°C. e remaining binding agent continued to decompose slowly, and the mass loss rate of the foundry sand was significantly reduced.
Interestingly, no flat region can be identified in the DSC curve in Figure 2. In other words, the foundry sand did not reach steady state by the final temperature of this experiment. e DSC curves of corncob show similar rising trends at a final temperature of 1200°C at 5°C/min and 20°C/min heating rates [34]. e DSC curve of rice husk has a downtrend at the final temperature of 1200°C at a 20°C/min heating rate [21]. e magnitude of the peak also appears to be unrelated to the magnitude of the DSC curve at the final temperature. In corncob, the peak magnitude is greater than the DSC curve magnitude at the final temperature [34], but the peak magnitude of rice husk is less than the DSC curve magnitude at the final temperature [21]. Four different diameters of corncob particles were used to investigate the effects of particle size on thermal decomposition behaviour [34]; all four samples exhibited downtrends in the DSC curves at the final temperature [34].
CO released by pyrolysis of binding resin may result in poisoning and asphyxiation accidents [35]. CO easily binds with haemoglobin to form carbonyl haemoglobin, which causes the haemoglobin to lose its ability to carry oxygen, leading to suffocation [36,37]. CO has toxic effects on the whole body, especially on the cerebral cortex [38,39].
erefore, corresponding safety measures should be adopted to prevent such accidents, such as improving safety management practices, identifying dangerous and harmful factors, and eliminating accident potential.

ermal Decomposition of Foundry Sand in Nitrogen and Air Atmospheres.
e thermal decomposition data for foundry sand in nitrogen and air atmospheres can be found in Supplementary Materials S1, S2, S3, and S4. e thermal decomposition behaviours of foundry sand at a 20°C/min heating rate in nitrogen and air atmospheres are shown in Figure 2.
As shown in Figure 2, the mass of foundry sand gradually decreased as the temperature increased in both the nitrogen and air atmospheres. Above 600°C, the mass of the sand basically remained unchanged. e mass loss of foundry sand was greater in air than in nitrogen, mainly due to the binding agent in the foundry sand being prone to pyrolysis in an air atmosphere. e mass loss of plywood is also higher in air than in nitrogen [20]. e mass loss rate of foundry sand from 250°C to 600°C was much larger in the air atmosphere than in the nitrogen atmosphere, but beyond this range, the rates were almost the same. e mass loss rate of foundry sand was the highest at approximately 450°C, and the thermal decomposition was the most severe at this time. e sand binding agent used in this study was furan resin. Methylene or methine bridges connecting furan resin chains begin to break at temperatures  [40]. Furan resin undergoes significant ring-opening reactions between 300°C and 350°C [40]. Benzene may also be formed from the fragments derived from said ring-opening reactions; in our experiment, the amount of benzene significantly increased starting at 400°C [41]. From 460°C to 650°C, benzene formed from the furan resin-derived intermediates [42]. In a previous study, Xu et al. only investigated the thermal composition characteristics of foundry sand for cast iron in a nitrogen atmosphere [27]. Foundry sand at the  sprue is mainly thermally decomposed in an air atmosphere, and foundry sand in the cavity is mainly thermally decomposed in a nitrogen atmosphere. is study investigated the thermal composition characteristics of foundry sand in air and nitrogen atmospheres, which enriches the theoretical knowledge base related to foundry sand.

ermal Decomposition of Foundry Sand at Different
Heating Rates. Foundry sand binding agents have different thermal conditions and different pyrolysis behaviours at different heating rates. e thermal composition behaviours of foundry sand at heating rates of 20, 30, and 40°C/min in a nitrogen atmosphere are shown in Figure 3.
As shown in the TG curves in Figure 3, the mass loss of foundry sand increased as the heating rate increased. e mass loss of foundry sand was 0.83 wt. % at a heating rate of 20°C/min but reached 0.99 and 1.36 wt. % at heating rates of 30°C and 40°C/min, respectively. An elevated heating rate appears to enhance the thermal decomposition of foundry sand in a high-temperature environment. e mass loss of corncob and peanut shells also increases as the heating rate increases [34,43]. e main component of foundry sand, SiO 2 , has high-temperature stability. erefore, the mass loss of foundry sand is relatively small in a high-temperature environment, which benefits sand casting moulding.
As shown in the DTG curves in Figure 3, the mass loss rate of foundry sand increased as the heating rate increased. e maximum mass loss rate was −2.33 × 10 −3 % (°C) at a heating rate of 20°C/min but reached −3.27 × 10 −3 and −5.42 × 10 −3 % (°C) at heating rates of 30°C and 40°C/min, respectively. Again, increasing the heating rate apparently enhances the thermal decomposition of foundry sand in a high-temperature environment. e maximum mass loss rate of peanut shell increased twofold to threefold as the heating rate increased in a previously published study [43], which is consistent with our results. e peak temperature of the DTG curve also increased as the heating rate increased. e maximum peak temperatures were 470, 472, and 485°C at heating rates of 20, 30, and 40°C/min, respectively. is thermal hysteresis was mainly caused by the temperature gradient between the support and the sample as well as between the inside and outside of the sample [34]. e mass loss of foundry sand trended downward as the heating rate increased. Generally, a slower heating rate results in greater mass loss of a material (e.g., bituminous coal) [44]. e main reason for this phenomenon is that a slower heating rate means longer exposure to elevated temperature and a longer thermal decomposition time. For the resin in this study, increasing the heating rate not only truncated the initial decomposition reaction time but also promoted the pyrolysis reaction [45]. e binding agent we used was furan resin, as mentioned above. Chen and Yeh reported that the residual mass fraction of resin trends down as the heating rate increases [46], which is consistent with our observations. A faster heating rate results in a faster resin decomposition rate, a greater amount of decomposition, and less crosslinking charring.
Xu et al. only investigated the thermal composition characteristics of foundry sand for cast iron at heating rates of 30°C and 40°C/min [27]. Increasing the heating rate appears to enhance the thermal decomposition of foundry sand in a high-temperature environment [27], which was consistent with our study. However, this study investigated the thermal composition characteristics of foundry sand at heating rates of 20, 30, and 40°C/min, and our conclusion is more persuasive.

Volatile Release Characteristic Index of Foundry Sand.
e volatile release characteristic index D of foundry sand can be determined according to the thermal decomposition process. e initial temperature T s of volatile emission and mass loss is first established by reference to a mass loss rate of 0.1 mg/min [19]. e sample mass is usually 5 mg when carrying out a thermal decomposition experiment [21], so the initial temperature T s corresponds to a mass loss rate of 2 wt. % (min). e initial temperature T s is unrelated to the mass of the sample after it is rewritten as mass loss percent per minute. In this study, the maximum mass loss of foundry sand was 1.36 wt. % in the nitrogen atmosphere and 2.08 wt. % in the air atmosphere, and the maximum mass loss rate was 0.2 wt. % (min) in the air atmosphere. e volatile release characteristic index D thus cannot be calculated according to this concept, as discussed above. To solve this issue, Xu et al. [27] proposed that the term T stv /m loss be 2.86 wt. %/min, where T stv indicates the threshold value for determining T s . e volatile release characteristics of foundry sand for cast steel were determined, as shown in Table 3.
As shown in Table 3, the volatile release characteristic index of foundry sand increased as the heating rate increased, indicating that an elevated heating rate benefits the thermal decomposition of foundry sand in a high-temperature environment. e mass loss of corncob also increases as the heating rate increases [34]. e results were consistent at different heating rates in the nitrogen atmosphere. e volatile release characteristic index of foundry sand in the air atmosphere was one order of magnitude larger than that in the nitrogen atmosphere, mainly because the sand binding agent is more prone to pyrolysis in air than in nitrogen. e results between nitrogen and air atmospheres were consistent.
Compared to other materials, the volatile release characteristic index of foundry sand is approximately 1/ 100,000th that of rice straw [21], 1/1,000th that of corncob [34], one-millionth that of wastewater solids, and 1/1,000th that of lignite [33]. Foundry sand evaporates less quickly than other materials, mainly because its volatile content is only approximately 2.15 wt. %. e main component of foundry sand, SiO 2 , has high-temperature stability. ese results are consistent with our proximate analysis and TG curve analysis.
Mathematical Problems in Engineering e activation energy E and preexponential factor A can be calculated by equation (8), as shown in Figure 4 and Table 4.
In Table 4, the temperature range 390°C-530°C corresponds to the main volatile release phase. After obtaining the fitting equation for the independent and dependent variables, the fitting effect can be roughly determined by Pearson's R. To thoroughly test the fitting effect, singlefactor analysis of variance is needed to determine the source of error (Table 5).
F 0.05 (1, 27) � 4.21 < 1877 was determined from F-distribution tables and p � 0 < 0.05. We found that the independent variable has a significant influence on the dependent variable in nitrogen and air atmospheres. e single-factor analysis of variance results also showed that the selected fitting equation is feasible and effective.
As shown in Table 4, the calculated thermal decomposition activation energy was higher in the air atmosphere than in the nitrogen atmosphere due to the presence of oxygen. Previous researchers have made similar observations [20,47,48] regarding high activation energy in oxidizing atmospheres. Regarding the thermal decomposition analysis of powdered ammonium nitrate, Yang [49] concluded that the activation energy was lower in air than in nitrogen. Previous scholars have calculated thermodynamic and kinetic parameters for atmospheres with varying oxygen concentrations to determine the change trend of activation energy parameters [50,51]; the activation energy does not increase or decrease strictly with increases in oxygen concentration. e activation energy may increase steadily with concentration until reaching a maximum and then decrease [27,52].
Compared to other materials, the activation energy of foundry sand was approximately one-third that of peanut shells [43], one-fourth that of corncob [34], 1/19th that of coal, and 1/10th that of wastewater solids [33]. Foundry sand  has high-temperature stability, as confirmed by the volatile release characteristic index: a great deal of activation energy is needed to convert volatiles into an active state. e activation energy of foundry sand is relatively small, however, which seems to contradict the volatile release characteristic index. is result is mainly due to the wide temperature range of thermal decomposition of the sand. Weight loss occurring between 390°C and 530°C in our sample (i.e., within the activation energy zone we calculated) contributed approximately 38 wt. % of the total weight loss. e weight loss of corncob in the pyrolysis zone contributes approximately 80 to 90 wt. % of the total weight loss [34]. Severe heat conduction phenomena occur when molten alloy is poured into the cavity of the foundry [15], resulting in the temperature of the casting mould rapidly rising. For sand casting, as the temperature of the casting mould rises, materials such as adhesives, curing agents, and coatings in the sand quickly vaporize and pyrolyze to produce gas [4,27]. If the gas cannot be discharged out of the cavity in time, there may be a blowhole defect or even an explosion accident [15]. erefore, investigating the thermal decomposition behaviour of foundry sand is helpful for obtaining an in-depth understanding of the change trends in casting mould temperature and cavity gas during the pouring process.

Conclusion
We investigated the thermal decomposition behaviour of foundry sand for cast steel in air and nitrogen atmospheres for the first time in this study, and the heating rates were 20,    Proximate analysis and chemical composition analysis of foundry sand indicate that foundry sand has stability in high-temperature environments, which will benefit the maintenance of the shape of the casting mould. e thermal decomposition of foundry sand was mainly attributable to its binding agent, and there was no obvious oxidation or combustion reaction of the sand in the nitrogen atmosphere. e change trends for the thermal decomposition of foundry sand can be determined by changing the thermal decomposition environment. e mass loss of foundry sand was greater in air than in nitrogen, mainly due to the binding agent of foundry sand being prone to pyrolysis in the air atmosphere. In addition, an elevated heating rate appears to enhance the thermal decomposition of foundry sand in a high-temperature environment.
To determine the volatile release characteristic index of foundry sand, a new method was adopted. e volatile release characteristic index of foundry sand increased as the heating rate increased, indicating that an elevated heating rate benefits the thermal decomposition of foundry sand in a high-temperature environment. e thermal decomposition kinetics parameters of foundry sand were obtained according to kinetics theory. e calculated thermal decomposition activation energy was higher in the air atmosphere than in the nitrogen atmosphere due to the presence of oxygen. To thoroughly test the fitting effects, we performed a single-factor analysis of variance on the source of error and found that the independent variable significantly influences the dependent variable.
Although this study achieved several significant results, to simplify the investigation, the binding materials used for the foundry sand only included furan resin. e thermal decomposition process of foundry sand may be different with different binding materials. Future work should focus on the influence of binding materials on the thermal decomposition process of foundry sand.
Data Availability e thermal decomposition data supporting this paper have been uploaded as electronic supplementary materials.

Conflicts of Interest
e authors declare no conflicts of interest.

Authors' Contributions
Jinjia Zhang, Kaili Xu, and Qingwei Xu guided the writing process; Shoukun Chen, Jinjia Zhang, Kaili Xu, and Qingwei Xu conducted the experiments and data analysis; and Shoukun Chen wrote the paper. All the authors gave their final approval for publication. Supplementary Materials e supplementary materials have been divided into four parts, and the corresponding concise descriptions of each supplementary material file are shown as follows. S1: thermal decomposition data of foundry sand at heating rate of 20°C /min in nitrogen. S2: thermal decomposition data of foundry sand at heating rate of 30°C/min in nitrogen. S3: thermal decomposition data of foundry sand at heating rate of 40°C/min in nitrogen. S4: thermal decomposition data of foundry sand at heating rate of 20°C/min in air. (Supplementary Materials)