Impact of Demineralization on Various Types of Biomass Pyrolysis: Behavior, Kinetics, and Thermodynamics

Abstract

This study systematically investigates the effects of demineralization on the pyrolysis characteristics, kinetics, and thermodynamics of three biomass types (eucalyptus, straw, and miscanthus) using thermogravimetric analysis (TGA) combined with multiple kinetic approaches. The Coats–Redfern integral model was employed to determine the reaction mechanisms. The results indicate that the primary weight-loss temperature ranges for eucalyptus, straw, and miscanthus were 222.02~500.23 °C, 205.43~500.13 °C, and 202.30~490.52 °C, respectively. Demineralization increased the initial pyrolysis temperature and significantly enhanced the reaction rates. Kinetics analysis revealed that the ash content significantly influences the activation energy of the pyrolysis reaction. The average activation energies follow the trend eucalyptus (193.48 kJ/mol) < miscanthus (245.66 kJ/mol) < straw (290.13 kJ/mol). After demineralization, the activation energies of both straw and miscanthus pyrolysis decreased, with the largest reduction observed in straw, which dropped by 77.53 kJ/mol. However, the activation energy for eucalyptus pyrolysis increased by 12.52 kJ/mol after demineralization. The Coats–Redfern model and thermodynamic analysis demonstrated that each type of biomass followed distinct reaction mechanisms at different stages, which were altered after demineralization. Additionally, demineralization leads to higher ΔH and Gibbs free energy ΔG for eucalyptus, but lower values for straw and miscanthus, which indicate that the ash content has a significant impact on the biomass pyrolysis reaction. These findings provide fundamental insights into the role of ash in biomass pyrolysis kinetics and offer theoretical support for the design of pyrolysis reactors.

1. Introduction

Biomass energy boasts advantages such as wide availability, abundant reserves, and low-carbon environmental friendliness. Its resource utilization can optimize China’s energy structure, offering broad development prospects under the current dual-carbon strategic goals [1,2]. Currently, biomass and waste account for approximately 10% of global energy supply, and the estimated annual total available potential of biomass is projected to reach 1.08 × 10¹¹ tons of oil equivalent (toe) [3,4]. The efficient conversion and utilization of biomass thus present a critical challenge to address. Common biomass utilization technologies include physical, chemical, and biological conversion [5]. While all three have been developed for years, physical and biological conversion suffer from low energy efficiency, high environmental pollution, and poor conversion rates, making them unsuitable for large-scale, industrial biomass utilization [6,7]. Chemical conversion has emerged as the mainstream technology, with pyrolysis widely adopted due to its efficiency, environmental friendliness, and controllability [8].

Biomass primarily consists of cellulose, hemicellulose, and lignin, with their content varying significantly across different biomass types, typically ranging around 40–60%, 15–30%, and 10–25%, respectively [9]. For instance, hardwood biomass like poplar contains about 49% cellulose and 20% lignin [10], whereas agricultural biomass such as rice straw has 37% cellulose and 13.6% lignin [11]. Consequently, differences in composition greatly influence pyrolysis behavior. Recent research has focused on analyzing the pyrolysis characteristics of these three major components, including thermogravimetric behavior, product distribution, and reaction mechanisms [12,13,14,15,16,17].

Additionally, biomass contains inorganic minerals (ash), primarily elements such as K, Ca, Na, and Mg, with contents ranging from 0.5% to 15% [18]. Studies have shown that ash affects biomass pyrolysis. For example, Zhang et al. found that potassium ions promote the conversion of levoglucosan to furans during pyrolysis, increasing furfural yield [19]. Patwardhan et al. observed that magnesium and calcium ions facilitate the formation of furan derivatives (e.g., 2-furaldehyde, 5-hydroxymethylfurfural) while suppressing guaiacol formation in lignin pyrolysis [20]. Liu et al. reported that magnesium loading significantly alters cellulose pyrolysis pathways, leading to higher char yields during fast pyrolysis [21]. Similar ash-induced effects on pyrolysis products and mechanisms have been documented in other studies [22,23,24]. In summary, while numerous studies have explored ash’s influence on biomass pyrolysis, most involve externally adding alkali/alkaline earth metals to biomass, which may not fully reflect the actual effects of intrinsic ash.

Pyrolysis kinetics serve as the foundation for the design, scaling, optimization, and industrial application of pyrolysis processes, and are crucial for understanding the underlying mechanisms of pyrolytic reactions. Many researchers have conducted studies on the biomass pyrolysis kinetics; El-Naggar et al. [25] investigated the pyrolysis process of bagasse using kinetic methods such as KAS, FWO, and C-R, and found that the pyrolysis process conforms to the D3 reaction model, with ΔS, ΔH, and ΔG values of 1.54 J/mol·K, 164.2 kJ/mol, and 163.2 kJ/mol, respectively. Nawaz et al. [26] discovered that the pyrolysis process of rapeseed straw follows the F5 reaction model via C-R method, with an average activation energy of approximately 202 kJ/mol calculated by the isoconversional method, and obtained the thermodynamic parameters of the pyrolysis process. Similarly, Acikalin et al. [27] used the C-R method and the isoconversional method to study the pyrolysis kinetics of peanut shells, determining that the most suitable reaction mechanism for the pyrolysis stage of peanut shells was D5-D3, indicating two different reaction mechanisms during the devolatilization stage, and successfully calculated the thermodynamic parameters. Regarding the impact of ash content on biomass pyrolysis kinetics, Xu et al. [28] applied the iso-conversion method and found that the overall pyrolysis efficiency of microalgae increased after the addition of alkali and alkaline earth metals, with a reduction in the pre-exponential factor. Research by Tran et al. [29] has shown that potassium chloride, calcium chloride, and magnesium chloride significantly affect the pyrolysis kinetics of lignin. When 2.0 wt% Mg was added, the average activation energy for lignin pyrolysis decreased from 181.67 kJ/mol to 156.55 kJ/mol. Zhang et al. [30] investigated the effect of alkali metals on the pyrolysis characteristics of rice straw and found that the presence of alkali metals suppressed the formation of aldehydes and ethers, while also lowering the activation energy of rice straw pyrolysis. The current research also focuses on the impact of alkali and alkaline earth metals on biomass pyrolysis. However, the changes in pyrolysis kinetics and reaction mechanisms before and after ash removal remain unclear.

This study selects eucalyptus (hardwood), straw (agricultural), and miscanthus (herbaceous) as representative lignocellulosic biomass samples. Demineralization biomass is obtained via hydrofluoric acid washing. Thermogravimetric analysis is employed to investigate the thermal decomposition characteristics of raw and demineralized biomass, while KAS, FWO, Friedman, and C-R kinetic methods are applied to study pyrolysis kinetics and reaction mechanisms, elucidating ash’s impact on biomass pyrolysis. The findings of this study will further elucidate the complex pyrolytic mechanism network of biomass, clarify the influence of inherent ash content on biomass pyrolysis, and provide kinetic and thermodynamic data to support the optimization of biomass pyrolysis processes.

2. Materials and Method

2.1. Materials

The experiment used eucalyptus, straw, and miscanthus as raw biomass materials, all purchased from a biomass processing plant. Prior to the experiment, the raw materials were crushed, ground, and sieved to obtain samples with particle sizes of 100–250 μm, and then dried in an oven at 105 °C until constant weight was achieved. The dried samples were washed with hydrofluoric acid (purchased from Sigma-Aldrich, Shanghai, China) to remove inorganic components. The specific operation was as follows: 10 g of dried sample was placed in a beaker, 300 mL of 3 wt.% hydrofluoric acid was added, and magnetic stirring was performed for 1 h. After stirring, the solution was filtered and repeatedly rinsed with deionized water, then dried to a constant weight to obtain demineralized samples, which were labeled as HF-eucalyptus, HF-straw, and HF-miscanthus, respectively.

2.2. Thermogravimetric Experiment

The thermal decomposition experiment was conducted using a thermogravimetric analyzer (NETZSCH STA 449 F5 Jupiter^®, Selb, Germany). The temperature program consisted of two stages: drying stage—heating from 50 °C to 105 °C at 10 °C/min, then holding at 105 °C for 20 min; experimental stage—heating from 105 °C to 800 °C at heating rates of 5 °C/min, 10 °C/min, and 20 °C/min, respectively. The drying stage was designed to ensure complete drying of samples and reduce experimental errors. Before each heating rate experiment, a blank test was performed to correct baseline data. High-purity argon (>99.999%) was used as both carrier gas and protective gas, with flow rates of 50 mL/min and 20 mL/min, respectively. The sample loading amount was 10 ± 1 mg for each test.

2.3. Kinetic Analysis

This study employed three kinetic methods—Friedman, FWO, and KAS—to calculate thermogravimetric (TG) data and perform kinetic analysis. The TG data represent the weight loss of the sample as a function of temperature (TG-T). From these data, the conversion rate (α-T) was derived, defined as

$$\mathrm{\alpha }={\displaystyle \frac{m_i-m_T}{m_i-m_{\mathrm{\infty }}}}$$

(1)

where α is the conversion rate, $m_i$ and $m_{\infty }$ are the initial and final weights of the sample before and after pyrolysis, respectively, and $m_T$ is the sample weight at any given temperature.

The three kinetic methods are derived from the classical non-isothermal kinetic equation (Equation (2)). The transformed equations for the Friedman, FWO, and KAS methods are given in Equations (3)–(5), respectively.

$${\displaystyle \frac{d\alpha }{dT}}={\displaystyle \frac{1}{\beta }}A\mathrm{e}\mathrm{x}\mathrm{p}(-{\displaystyle \frac{E}{RT}})f(\alpha )$$

(2)

$$\ln \left(\beta {\displaystyle \frac{d\alpha }{dt}}\right)=\mathrm{l}\mathrm{n}\left[f\left(\alpha \right)A\right]-{\displaystyle \frac{E}{RT}}$$

(3)

$$\ln \left(\mathrm{\beta }\right)=\mathrm{l}\mathrm{n}{\displaystyle \frac{AE}{RG\left(\alpha \right)}}-{\displaystyle \frac{E}{RT}}$$

(4)

$$\ln \left({\displaystyle \frac{\beta }{T^2}}\right)=\mathrm{l}\mathrm{n}{\displaystyle \frac{AE}{RG\left(\alpha \right)}}-{\displaystyle \frac{E}{RT}}$$

(5)

where β is the heating rate (°C/min), T is the pyrolysis temperature (K), A is the pre-exponential factor (min⁻¹), E is the activation energy (kJ/mol), R is the universal gas constant (8.314 J/mol·K), and f(α) and G(α) are the reaction mechanism functions.

To determine the most suitable reaction mechanism function for the pyrolysis process, the Coats–Redfern (C-R) method was applied. The C-R equation is expressed as

$$\mathit{ln}\left[{\displaystyle \frac{g\left(\alpha \right)}{T^2}}\right]=\mathit{ln}{\displaystyle \frac{AR}{\beta E_a}}-{\displaystyle \frac{E_a}{RT}}$$

(6)

where g(α) is the integral form of the mechanism function f(α). The kinetic mechanism functions considered in this study are listed in

Table 1. General mechanism functions for solid-state thermal decomposition reactions.

Reaction Mechanisms	Symbol	$\mathit{f}\left(\mathit{\alpha }\right)$	$\mathit{g}\left(\mathit{\alpha }\right)$
Interface Reactions	R1	1	$\alpha$
	R2	$2{(1-\alpha )}^{1/2}$	$1-{\left(1-\alpha \right)}^{1/2}$
	R3	$3{\left(1-\alpha \right)}^{2/3}$	$1-{\left(1-\alpha \right)}^{1/3}$
Diffusion-Controlled Reactions	D1	${(1/2\alpha )}^{-1}$	${\alpha }^2$
	D2	$\mathit{ln}{[-ln(1-\alpha \left)\right]}^{-1}$	$\left(1-\alpha \right)\mathit{ln}\left(1-\alpha \right)+\alpha$
	D3	$3/2{\left(1-\alpha \right)}^{2/3}{\left[1-{\left(1-\alpha \right)}^{1/3}\right]}^{-1}$	${\left[1-{\left(1-\alpha \right)}^{1/3}\right]}^2$
	G-B	$3/2{\left[{\left(1-\alpha \right)}^{-1/3}-1\right]}^{-1}$	$\left(1-2/3\alpha \right)-{\left(1-\alpha \right)}^{2/3}$
	ZH	$3/2{\left(1-\alpha \right)}^{4/3}{\left[{\left(1-\alpha \right)}^{-1/3}-1\right]}^{-1}$	${\left[{\left(1-\alpha \right)}^{-1/3}-1\right]}^2$
Nucleation and Growth	A2	$2\left(1-\alpha \right){[-ln(1-\alpha \left)\right]}^{1/2}$	${[-ln(1-\alpha \left)\right]}^{1/2}$
	A3	$3\left(1-\alpha \right){[-ln(1-\alpha \left)\right]}^{2/3}$	${[-ln(1-\alpha \left)\right]}^{1/3}$
	A4	$4\left(1-\alpha \right){[-ln(1-\alpha \left)\right]}^{3/4}$	${[-ln(1-\alpha \left)\right]}^{1/4}$
Chemical Reactions	F1	$1-\alpha$	$-\mathit{ln}\left(1-\alpha \right)$
	F3/2	${\left(1-\alpha \right)}^{3/2}$	$2\left[{\left(1-\alpha \right)}^{-1/2}-1\right]$
	F2	${\left(1-\alpha \right)}^2$	$2{\left(1-\alpha \right)}^{-1}-1$
	F3	${\left(1-\alpha \right)}^3$	$\left[{\left(1-\alpha \right)}^{-2}-1\right]/2$
Power Law	P2/3	$2/3{\alpha }^{-1/2}$	${\alpha }^{3/2}$
	P2	$2{\alpha }^{1/2}$	${\alpha }^{1/2}$
	P3	$3{\alpha }^{2/3}$	${\alpha }^{1/3}$

.

2.4. Calculation of Thermodynamic Parameters

Thermodynamic parameters such as enthalpy (ΔH), Gibbs free energy (ΔG), and entropy (ΔS) play crucial roles in analyzing the enthalpy changes during pyrolysis, the spontaneity of the process, and reaction products and mechanisms. These parameters can be evaluated based on the activation energy derived from pyrolysis reaction kinetics, and their calculation equations are as follows:

$$\Delta H=E-R{T}_{\alpha }$$

(7)

$$\Delta G=E+R{T}_m\left(ln{\displaystyle \frac{K_b{T}_m}{hA}}\right)$$

(8)

$$\Delta S={\displaystyle \frac{\Delta H-\Delta G}{T_m}}$$

(9)

where E is the activation energy (kJ/mol), $T_{\alpha }$ is the temperature at conversion rate α (K), $T_m$ is the peak temperature of maximum mass loss in DTG curve (K), K_b is the Boltzmann constant (1.381 × 10⁻²³ J/K), h is the Planck constant (6.626 × 10⁻³⁴ J·s), A is the pre-exponential factor (min⁻¹), and R is the universal gas constant (8.314 J/(mol·K)).

3. Results and Discussion

3.1. Thermogravimetric Characteristics Analysis

Pyrolysis experiments were conducted on eucalyptus, straw, and miscanthus samples, as well as their demineralized counterparts, at heating rates of 5 °C/min, 10 °C/min, and 20 °C/min. The TG-DTG curves and characteristic thermogravimetric parameters are presented in Figure S1, Figure 1 and

Table 2. Characteristic thermogravimetric points of lignocellulosic samples before and after demineralization.

Samples	Heating Rate	Ti °C	Tf °C	Tmax1 °C	Dmax1 %/min	Tmax2 °C	Dmax2 %/min	Mf %
Eucalyptus	5	222.02	500.23	273.03	−1.74	338.59	−6.14	21.79
	10	232.90	509.32	278.61	−3.64	351.93	−11.05	23.10
	20	242.97	510.33	292.62	−8.70	360.90	−24.14	23.84
HF-eucalyptus	5	225.69	503.26	277.93	−9.16	342.27	−27.26	20.23
	10	232.49	505.61	285.82	−9.29	354.37	−25.43	21.57
	20	243.78	508.42	302.28	−10.43	364.98	−22.88	21.89
Straw	5	205.43	500.13	283.91	−2.34	300.51	−4.02	34.93
	10	211.68	510.25	291.94	−2.43	312.21	−4.08	33.01
	20	216.99	515.33	299.01	−2.57	317.10	−3.56	36.91
HF-straw	5	217.39	510.23	279.42	−2.71	324.31	−6.31	20.63
	10	220.12	514.46	282.96	−2.56	334.79	−5.93	21.35
	20	231.13	520.33	291.94	−2.75	343.08	−5.12	22.54
Miscanthus	5	202.30	490.52	280.65	−2.79	314.52	−4.62	21.83
	10	211.95	493.25	283.50	−2.44	323.50	−4.23	27.72
	20	210.46	500.24	289.63	−2.32	331.79	−4.12	28.01
HF-miscanthus	5	220.93	500.46	291.80	−2.93	333.15	−5.76	18.86
	10	233.99	508.65	296.56	−3.15	342.68	−5.32	18.90
	20	244.33	519.36	315.20	−3.09	351.38	−4.91	20.63

T_i refers to the temperature at the start of decomposition; T_f refers to the temperature at the end of decomposition; T_max1, T_max2 refers to the temperature of the first stage and second stage maximum mass loss rate; D_max1, D_max2 refers to the temperature of the first stage and second stage maximum mass loss; M_f refers to the residual mass at the end of decomposition.

, respectively.

As shown in Figure S1, the thermal degradation temperature ranges of eucalyptus, straw, and miscanthus shifted to higher temperature intervals when the heating rate increased from 5 °C/min to 20 °C/min. This phenomenon can be attributed to the thermal hysteresis effect, where heat transfer from the sample surface to the interior requires time, and higher heating rates prolong this process. The demineralized samples exhibited similar behavior. All three sample types displayed two distinct weight loss peaks in their thermogravimetric curves, corresponding to two degradation stages.

The TG and DTG curves of all samples obtained at 5 °C/min are displayed in Figure 1, and the corresponding pyrolysis characteristic points are listed in Table 2. It can be seen that the primary weight loss ranges for eucalyptus, straw, and miscanthus were 222.02~500.23 °C, 205.43~500.13 °C, and 202.30~490.52 °C, respectively. These differences arise from the distinct thermal decomposition temperature ranges of the lignocellulosic components (cellulose, hemicellulose, and lignin). The first stage (200~310 °C) primarily involved hemicellulose decomposition, with varying weight loss magnitudes among samples due to differences in hemicellulose content. The second stage showed more significant weight loss, corresponding to the overlapping decomposition of cellulose and lignin, resulting in a single peak. The residual solid contents after pyrolysis were 21.79%, 34.93%, and 21.83% for eucalyptus, straw, and miscanthus, respectively, indicating higher ash content in straw.

Following HF treatment for ash removal, Figure 1 reveals significant changes in the thermogravimetric behavior of demineralized samples compared to raw samples. The initial pyrolysis temperatures of demineralized samples were consistently higher, suggesting increased activation energy requirements, possibly due to the catalytic role of ash in promoting initial hemicellulose decomposition. However, the presence of ash was found to inhibit the overall thermal degradation rate, as evidenced by the higher maximum weight loss rates observed in demineralized samples (Table 2).

3.2. Analysis of Pyrolysis Reaction Activation Energy Variation Patterns

This study estimated the activation energies of eucalyptus, straw, miscanthus, and their demineralized samples using three kinetic methods (KAS, FWO, and Friedman) based on thermogravimetric data obtained at heating rates of 5 °C/min, 10 °C/min, and 20 °C/min. Figure 2 shows the variation trends of activation energy and correlation coefficients (R²) at different conversion rates. All R² values ranged within 0.9402~1.0000, indicating the reliability of activation energy values obtained by these three methods, which showed consistent trends.

As shown in Figure 2, for the three raw biomass materials, the activation energy of eucalyptus and miscanthus exhibited a steady increasing trend before the conversion rate reached 0.8. This stage mainly involved depolymerization and cracking reactions of hemicellulose, cellulose, and lignin, including cleavage of unstable side chains and glycosidic bonds in saccharide structures, decomposition of unstable functional groups, and breakage of C-O and C-C bonds between monomers [16,17,31]. When the conversion rate exceeded 0.8 at higher temperatures, the reactions mainly involved the rearrangement and condensation of polycyclic compounds in pyrolysis products, which required higher energy input, leading to a rapid increase in activation energy [32,33]. The activation energy curve of straw showed an inflection point at a conversion rate of 0.2, where hemicellulose and cellulose depolymerization predominantly occurred, suggesting that higher ash content increased the energy barrier for these reactions. Between conversion rates of around 0.2 and 0.6, the gradual decrease in activation energy revealed that ash only significantly affected the partial pyrolysis reactions of the three biomass components.

Considering potential errors in activation energy estimation at extremely low and high conversion rates, this study calculated average activation energies within the conversion rate range of 0.2~0.8 (

Table 3. Average activation energies of all samples at conversion rates of 0.2~0.8.

Samples	Average Activation Energy (kJ/mol)
	KAS	FWO	Friedman
Eucalyptus	193.48	203.62	203.71
HF-eucalyptus	205.70	216.14	200.95
Straw	290.13	299.10	296.96
HF-straw	218.43	228.28	219.43
Miscanthus	245.66	255.48	254.19
HF-miscanthus	205.52	215.52	211.54

). Among raw samples, eucalyptus showed the lowest average activation energy (193.48 kJ/mol), while miscanthus and straw exhibited progressively higher values (245.66 kJ/mol and 290.13 kJ/mol, respectively). This difference correlates with ash content—agricultural biomass straw contained the highest ash content (~20%), while hardwood biomass contained only 0.2~1% ash. After HF demineralization, the activation energies changed significantly. For eucalyptus with originally low ash content, HF-eucalyptus showed only a slight increase (≤12.52 kJ/mol) in average activation energy. In contrast, HF-straw and HF-miscanthus exhibited considerably lower average activation energies than their raw counterparts, with differences reaching 77.53 kJ/mol and 42.65 kJ/mol, respectively. These results demonstrate that ash content significantly affects pyrolysis activation energy, particularly for straw and miscanthus, where higher ash content corresponds to greater increases in pyrolysis energy barriers.

3.3. Analysis of Pyrolysis Reaction Models

To further investigate the pyrolysis mechanisms of eucalyptus, straw, and miscanthus, as well as the influence of ash on the reaction mechanisms, this study employed the Coats–Redfern (C-R) method to fit the thermogravimetric processes of each sample. The appropriate pyrolysis reaction mechanism functions were determined, and the calculated kinetic characteristic parameters are listed in

Table 4. Kinetic characteristic points of eucalyptus and HF-eucalyptus calculated using the C-R method.

Models	Eucalyptus						HF-Eucalyptus
	0.1 < α < 0.3		0.3 < α < 0.8		0.8 < α < 0.9		0.1 < α < 0.3		0.3 < α < 0.8		0.8 < α < 0.9
	R2	E	R2	E	R2	E	R2	E	R2	E	R2	E
R1	0.9713	58.59	0.9938	55.25	0.8065	−4.70	0.9807	56.01	0.9933	61.35	0.8503	−4.96
R2	0.9818	71.57	0.9899	80.54	0.9156	11.29	0.9885	68.96	0.9932	88.37	0.9272	10.97
R3	0.9827	72.84	0.9871	86.26	0.9219	13.58	0.9892	70.19	0.9915	94.71	0.9329	13.19
D1	0.9755	126.46	0.9947	120.51	0.9090	1.50	0.9838	121.40	0.9942	132.80	0.9148	1.06
D2	0.9808	140.58	0.9926	149.24	0.9054	17.83	0.9877	135.45	0.9944	163.60	0.9178	17.31
D3	0.9827	145.68	0.9871	172.52	0.9219	27.16	0.9892	140.38	0.9915	189.42	0.9329	26.37
G-B	0.9814	142.30	0.9909	156.93	0.9120	20.80	0.9882	137.11	0.9937	172.14	0.9239	20.20
ZH	0.9859	156.15	0.9725	224.81	0.9403	52.16	0.9918	150.50	0.9813	247.49	0.9495	50.63
A2	0.9844	37.71	0.9804	49.34	0.9324	9.58	0.9906	36.34	0.9871	54.25	0.9424	9.30
A3	0.9844	25.14	0.9804	32.81	0.9324	6.38	0.9906	24.23	0.9871	36.16	0.9424	6.20
A4	0.9844	18.86	0.9804	24.67	0.9324	4.79	0.9906	18.17	0.9871	27.12	0.9424	4.65
F1	0.9844	75.42	0.9804	98.68	0.9324	19.15	0.9906	72.69	0.9871	108.49	0.9424	18.59
F3/2	0.9867	79.43	0.9683	119.76	0.9434	30.07	0.9923	76.55	0.9780	131.92	0.9523	29.16
F2	0.9968	26.18	0.9099	100.19	0.9542	40.36	0.9941	25.29	0.9283	111.01	0.9618	39.15
F3	0.9921	92.32	0.9311	199.22	0.9556	76.58	0.9962	89.02	0.9471	220.36	0.9631	74.30
P3/2	0.9789	101.80	0.9954	97.90	0.8920	9.29	0.9862	98.08	0.9950	107.18	0.9055	9.03
P2	0.9789	33.93	0.9954	32.63	0.8920	3.10	0.9862	32.69	0.9950	35.73	0.9055	3.01
P3	0.9713	22.62	0.9954	21.75	0.8920	2.07	0.9862	21.79	0.9950	23.81	0.9055	2.01

E-kJ/mol.

,

Table 5. Kinetic characteristic points of straw calculated using the C-R method.

Models	Straw
	0.1 < α < 0.2		0.2 < α < 0.5		0.5 < α < 0.8		0.8 < α < 0.9
	R2	E	R2	E	R2	E	R2	E
R1	0.9977	73.69	0.9981	61.50	0.9849	56.53	0.9966	−6.20
R2	0.9979	85.89	0.9996	78.77	0.9957	89.95	0.9987	8.93
R3	0.9979	87.04	0.9998	81.59	0.9970	99.07	0.9980	10.71
D1	0.9980	156.20	0.9984	132.24	0.9873	123.01	0.8486	−1.26
D2	0.9980	169.46	0.9994	151.77	0.9935	160.94	0.9991	14.16
D3	0.9979	174.08	0.9998	163.18	0.9970	198.13	0.9980	21.43
G-B	0.9980	170.59	0.9996	155.58	0.9950	173.17	0.9989	16.48
ZH	0.9975	183.49	0.9995	187.25	0.9985	284.56	0.9940	40.80
A2	0.9977	44.69	0.9998	43.73	0.9984	59.66	0.9961	7.52
A3	0.9977	29.79	0.9998	29.15	0.9984	39.77	0.9961	5.01
A4	0.9977	22.34	0.9998	21.86	0.9984	29.83	0.9961	3.76
F1	0.9977	89.37	0.9998	87.46	0.9984	119.31	0.9961	15.04
F3/2	0.9974	92.95	0.9993	96.82	0.9983	154.78	0.9929	23.46
F2	0.9808	24.56	0.9780	53.44	0.9904	155.58	0.9882	31.33
F3	0.9964	104.26	0.9949	128.86	0.9921	293.96	0.9874	59.41
P3/2	0.9982	123.76	0.9987	106.11	0.9892	99.73	0.9987	7.42
P2	0.9982	41.25	0.9987	35.37	0.9892	33.24	0.9987	2.47
P3	0.9982	27.50	0.9987	23.58	0.9892	22.16	0.9987	1.65

E-kJ/mol.

,

Table 6. Kinetic characteristic points of HF-straw calculated using the C-R method.

Models	HF-Straw
	0.1 < α < 0.2		0.2 < α < 0.5		0.5 < α < 0.8		0.8 < α < 0.9
	R2	E	R2	E	R2	E	R2	E
R1	0.9990	76.96	0.9954	51.77	0.9850	56.53	0.7092	−4.10
R2	0.9995	89.43	0.9985	68.16	0.9958	89.89	0.8994	11.81
R3	0.9995	90.63	0.9989	70.60	0.9971	98.98	0.9062	14.21
D1	0.9991	162.94	0.9963	112.98	0.9874	123.01	0.0274	2.36
D2	0.9994	176.46	0.9980	131.32	0.9937	160.88	0.8885	18.63
D3	0.9995	181.26	0.9989	141.20	0.9971	197.97	0.9062	28.42
G-B	0.9994	178.10	0.9984	134.62	0.9951	173.07	0.8956	21.75
ZH	0.9997	191.02	0.9994	162.02	0.9985	284.15	0.9264	54.69
A2	0.9997	46.52	0.9993	37.84	0.9984	59.59	0.9177	10.03
A3	0.9997	31.02	0.9993	25.22	0.9984	39.72	0.9177	6.69
A4	0.9996	23.26	0.9993	18.92	0.9984	29.79	0.9177	5.02
F1	0.9997	93.05	0.9993	75.68	0.9984	119.17	0.9177	20.06
F3/2	0.9998	96.76	0.9993	83.78	0.9982	154.54	0.9298	31.51
F2	0.9918	25.48	0.9823	46.24	0.9900	155.20	0.9418	42.38
F3	0.9999	108.49	0.9965	111.49	0.9917	293.28	0.9435	80.44
P3/2	0.9992	128.88	0.9969	91.82	0.9893	99.72	0.8994	11.81
P2	0.9992	42.96	0.9969	30.61	0.9893	33.24	0.8743	3.23
P3	0.9992	28.64	0.9969	20.40	0.9893	22.16	0.8743	2.16

E-kJ/mol.

and

Table 7. Kinetic characteristic points of miscanthus and HF-miscanthus calculated using the C-R method.

Models	Miscanthus						HF-Miscanthus
	0.1 < α < 0.4		0.4 < α < 0.8		0.8 < α < 0.9		0.1 < α < 0.4		0.4 < α < 0.8		0.8 < α < 0.9
	R2	E	R2	E	R2	E	R2	E	R2	E	R2	E
R1	0.9931	67.21	0.9420	39.08	0.9320	−4.57	0.9731	68.58	0.9939	51.06	0.8409	−4.75
R2	0.9969	81.55	0.9784	63.55	0.9764	10.84	0.9839	83.24	0.9959	78.72	0.9248	11.16
R3	0.9975	83.33	0.9823	69.12	0.9788	12.94	0.9852	85.09	0.9951	85.45	0.9301	13.50
D1	0.9940	143.58	0.9544	88.05	0.1881	1.54	0.9766	146.48	0.9949	112.12	0.9177	1.24
D2	0.9963	159.46	0.9729	115.53	0.9726	17.33	0.9824	162.74	0.9964	143.48	0.9161	17.45
D3	0.9975	166.66	0.9823	138.24	0.9788	25.88	0.9852	170.18	0.9951	170.89	0.9301	27.00
G-B	0.9967	161.89	0.9765	123.01	0.9751	20.06	0.9834	165.24	0.9962	152.52	0.9218	20.49
ZH	0.9992	181.57	0.9927	190.22	0.9854	48.42	0.9899	185.59	0.9884	233.44	0.9455	52.83
A2	0.9985	43.50	0.9884	40.68	0.9826	8.99	0.9877	44.44	0.9923	50.10	0.9389	9.62
A3	0.9985	29.00	0.9884	27.12	0.9826	6.00	0.9877	29.63	0.9923	33.40	0.9389	6.41
A4	0.9985	21.75	0.9884	20.34	0.9826	4.50	0.9877	22.22	0.9923	25.05	0.9389	4.81
F1	0.9985	87.00	0.9884	81.37	0.9826	17.99	0.9877	88.88	0.9923	100.20	0.9389	19.24
F3/2	0.9994	92.73	0.9942	102.54	0.9865	27.75	0.9909	94.81	0.9861	125.63	0.9481	30.53
F2	0.9716	35.82	0.9956	95.97	0.9903	36.61	0.9903	36.98	0.9570	116.05	0.9569	41.52
F3	0.9996	111.56	0.9973	184.22	0.9908	69.28	0.9972	114.30	0.9650	223.41	0.9582	78.93
P3/2	0.9948	114.55	0.9632	73.45	0.9673	9.16	0.9795	116.84	0.9958	91.59	0.9050	8.99
P2	0.9948	38.18	0.9632	24.48	0.9673	3.05	0.9795	38.95	0.9958	30.53	0.9050	3.00
P3	0.9948	25.45	0.9632	16.32	0.9673	2.04	0.9795	25.96	0.9958	20.35	0.9050	2.00

E-kJ/mol.

, respectively.

Since the C-R kinetic method treats the pyrolysis process as a single reaction, the thermogravimetric curves and kinetic analysis of various samples reveal that the activation energy of the pyrolysis reaction changes with increasing temperature and conversion rate. Therefore, this study divides the entire pyrolysis process into multiple stages based on the different trends in activation energy. For example, the pyrolysis process of eucalyptus is divided into three stages based on conversion rate: 0.1 < α < 0.3, 0.3 < α < 0.8, and 0.8 < α < 0.9. As shown in Table 4, for the first stage of eucalyptus pyrolysis, the F2 model shows the highest fitting degree, belonging to the chemical reaction control model. The reaction rate is controlled by elementary steps of chemical bond breaking or formation, involving the simultaneous decomposition of multiple components, consistent with the activation energy trend of straw shown in Figure 2. As the conversion rate increases, the reaction mechanism changes in the second stage of eucalyptus pyrolysis (0.3 < α < 0.8), following the power law mechanism. This stage mainly involves the decomposition of hemicellulose and cellulose components, with the reaction progressing gradually from the outer surface to the inner surface of the particles. The release of volatiles during pyrolysis is limited by heat or mass transfer within the particles. The fitting degrees of various mechanisms are similar, and the higher activation energy value of 97.90 kJ/mol is selected. When the pyrolysis reaction reaches higher temperatures (0.8 < α < 0.9), the eucalyptus pyrolysis reaction enters the char-forming stage. In this stage, the fitting degrees of various mechanism models are relatively low, with the F3 model being the most suitable reaction mechanism. After ash removal, the data show that the reaction mechanism in the first stage of HF-eucalyptus pyrolysis changes to the F3 model, with a corresponding higher activation energy (89.02 kJ/mol). The second stage follows the same reaction mechanism (P3/2 model), also requiring higher activation energy (107.18 kJ/mol). The activation energy in the third stage shows little difference from that of eucalyptus pyrolysis. By comparing the reaction models and activation energies at each stage, it can be found that the activation energies of HF-eucalyptus pyrolysis are all higher than those of eucalyptus pyrolysis, indicating that the presence of ash in eucalyptus can reduce the activation energy required for pyrolysis and promote the pyrolysis reaction, verifying the trend observed in the thermogravimetric curves.

Table 5 and Table 6 list the reaction mechanisms followed by straw and HF-straw pyrolysis in four stages, which are P3/2, D3, ZH and D2, representing power law mechanism, nucleation and growth mechanism, interface reaction and diffusion effect mechanism, and diffusion control mechanism, respectively. For HF-straw, the reaction mechanism changes after demineralization, following the F3 reaction mechanism in the initial pyrolysis stage. The power law mechanism is mainly applicable to autocatalytic processes, and the change in initial pyrolysis mechanism further illustrates the effect of ash on straw pyrolysis. HF-straw follows the ZH mechanism in both the second and third stages, and changes to the F3 mechanism in the fourth stage. For miscanthus and HF-miscanthus pyrolysis, Table 7 shows that miscanthus exhibits different reaction mechanisms compared to eucalyptus and straw pyrolysis. The miscanthus pyrolysis process is divided into three stages, all following the F3 mechanism, indicating competitive reactions of multiple components during pyrolysis. The activation energies for the three stages are 111.56, 184.22, and 69.28 kJ/mol, respectively. After ash removal, the second stage of miscanthus pyrolysis changes to the D2 diffusion control mechanism.

These results demonstrate that the multi-stage pyrolysis reactions of different biomass types follow different reaction mechanisms, and ash has varying effects on biomass pyrolysis mechanisms and different controlling effects on reaction rates.

3.4. Analysis of Thermodynamic Parameters

The thermodynamic properties, such as enthalpy change (ΔH), Gibbs free energy change (ΔG), and entropy change (ΔS), are determined by Equations (7)–(9), respectively. The activation energies and reaction mechanisms obtained from the C-R method were used to calculate ΔH, ΔG, and ΔS for each sample in the range of 0.1 < α < 0.9, with the results shown in Figure 3; the data of pre-exponential factor and thermodynamic parameters are listed in Tables S1–S3.

ΔH is used to analyze the energy requirements of the pyrolysis process. A positive ΔH indicates that the pyrolysis reactions of all samples are endothermic, meaning that the pyrolysis process requires significant energy input. The average values of ΔH for eucalyptus, HF-eucalyptus, straw, HF-straw, miscanthus, and HF-miscanthus are 72.43 kJ/mol, 96.97 kJ/mol, 204.78 kJ/mol, 201.87 kJ/mol, 140.17 kJ/mol, and 126.66 kJ/mol, respectively. ΔG represents the thermodynamic driving force for the pyrolysis reaction. ΔG > 0 indicates that the reaction is not spontaneous. The magnitude of ΔG reflects the difficulty of the pyrolysis reaction path, with average values of 118.79 kJ/mol, 120.45 kJ/mol, 123.25 kJ/mol, 121.18 kJ/mol, 117.90 kJ/mol, and 110.98 kJ/mol, respectively. It can be observed that the trends of ΔH and ΔG for the pyrolysis of three types of biomass are consistent with the activation energy trends obtained by the isoconversional method (Section 3.2). The corresponding values for eucalyptus are lower than those for HF-eucalyptus, while the values for straw and miscanthus are higher than those for HF-straw and HF-miscanthus. This indicates that the ash content has a significant impact on the biomass pyrolysis reaction. ΔS is used to measure the degree of disorder in the system. Due to the multi-stage nature of the pyrolysis process in each sample, the ΔS values for different reaction stages exhibit both positive and negative trends, as shown in Figure 3, which helps to elucidate the potential mechanisms and energy dynamics in the design of the pyrolysis process.

4. Conclusions

This study systematically examined the pyrolysis kinetics and thermodynamic properties of three biomass feedstocks (eucalyptus, miscanthus, and straw) before and after demineralization. Key findings demonstrate that demineralization treatment elevates the initial pyrolysis temperature while significantly improving reaction rates. Moreover, the ash content significantly influences the activation energy of the pyrolysis reaction. The average activation energies follow the trend eucalyptus (193.48 kJ/mol) < miscanthus (245.66 kJ/mol) < straw (290.13 kJ/mol). After demineralization, the activation energies of both straw and miscanthus pyrolysis decreased, with the largest reduction observed in straw, which dropped by 77.53 kJ/mol. However, the activation energy for eucalyptus pyrolysis increased by 12.52 kJ/mol after demineralization. The application of the Coats–Redfern method reveals distinct multi-stage reaction mechanisms for each biomass type. For example, eucalyptus pyrolysis sequentially follows F2, P3/2, and F3 mechanisms, while demineralization modifies these reaction pathways and generally increases activation energies. Thermodynamic analysis indicates that demineralization leads to higher ΔH and Gibbs free energy ΔG for eucalyptus, but lower values for straw and miscanthus, which indicates that the ash content has a significant impact on the biomass pyrolysis reaction. The comprehensive kinetic and thermodynamic parameters obtained in this work provide fundamental data to support the optimization and design of biomass pyrolysis reactors. The findings offer valuable insights into the role of mineral content in biomass pyrolysis processes.