The Physical Properties of Surface Layer Thermally Modified Wood and Energy Consumption in the Preparation Process

Abstract

Surface layer thermal modification (SLTM) not only improves the dimensional stability of wood but also effectively shortens production time. However, limited research has been conducted on how treatment conditions influence the properties of SLTM wood and energy consumption during the production process. This study compared the physical properties of SLTM wood with conventional thermally modified (CoTM) wood treated at 185 °C, 200 °C, and 215 °C, as well as the energy consumption during processing. Samples were subjected to SLTM at three temperatures (185 °C, 200 °C, and 215 °C) two times (2 h and 3 h) and two target surface layer thicknesses (6 mm and 12 mm). The results showed SLTM improved dimensional stability, with the anti-swelling efficiency (ASE) after water absorption increasing by a maximum of 2.4 times compared to CoTM185-2h. Increases in treatment temperature, time, and target surface layer thickness all contributed to enhanced ASE. At 96% relative humidity, SLTM wood samples exhibited lower equilibrium moisture content (16.6% to 19.2%) than CoTM185-2h (19.5%). SLTM also reduced the total treatment time by 1.9 h to 10.8 h compared to CoTM treatments. SLTM demonstrated energy savings ranging from 215 kW∙h to 1567 kW∙h, resulting in a reduction in CO₂ emissions by 73 kg to 729 kg per 1 m³ of processed wood. These findings suggest that SLTM provides a promising strategy for the energy-efficient utilization of plantation wood.

1. Introduction

With the continued development of the global economy and population growth, the demand for wood has steadily increased [1,2]. However, the reduction in global forest resources, coupled with intensified efforts to protect the ecological environment, has posed significant challenges regarding the raw material supply for the wood processing industry. Plantation wood, known for its rapid growth, has increasingly been used as a substitute for natural forest wood [3]. However, plantation wood has some serious drawbacks, including poor mechanical strength, dimensional instability, and susceptibility to biodegrading, which limit its widespread application in both residential and industrial sectors [4,5].

To enhance the dimensional stability, mechanical properties, and biodegradability of plantation wood, various modification methods have been employed, such as chemical impregnation [6,7], thermal treatment [8,9], and densification [10]. As a physical modification method, thermal treatment changes the chemical composition and structure of wood cell walls through high-temperature heating, thereby improving the dimensional stability and durability of the wood [11]. Because no chemical additives are required, thermal modification has been widely regarded as one of the most successful and effective methods for enhancing the dimensional stability of wood [12].

Despite its advantages, conventional thermal modification (CoTM) presents several challenges, including long processing times, high energy consumption, and a reduction in mechanical strength [13]. Surface layer thermal modification (SLTM) modifies only the surface layer of wood while maintaining the cell structure and properties of the core layer. This method not only reduces the treatment time but also minimizes the loss of mechanical strength in the wood. As an emerging thermal modification method, SLTM has attracted significant attention and research in recent years.

Ebner et al. [14,15] and Sorčan et al. [16]. used the “Yakisugi” method to prepare SLTM wood and confirmed that the treated wood exhibited enhanced weathering resistance, decay resistance, and flame retardancy. This method is a specific type of surface layer thermal modification, with a treatment temperature (up to over 450 °C) significantly higher than that of wood thermal treatment (160–240 °C), and is typically conducted in an oxygen-containing environment. As a result, hemicellulose, cellulose, and lignin undergo degradation. The improved properties of the samples are attributed to the formation of a charred layer [17]. The charred surface layer reduces water and moisture adsorption, and the decrease in surface pH improves decay resistance. Moreover, the char provides a passive insulating layer that inhibits further degradation of the wood beneath the char, which also contributes to the enhanced fire resistance of the treated wood [15]. However, the effectiveness of this layer often varies due to factors such as species or type of samples [18], and cracking in the charred layer can increase the water absorption of the samples [15]. Another common method for preparing SLTM wood is the “hot plate method”. During processing, the temperature of the hot plate can be controlled below 240 °C. Additionally, the surface of the sample is tightly adhered to the metal plate, and the steam formed by the vaporization of moisture in the sample creates a low oxygen processing atmosphere. Kymäläinen et al. [19]. treated spruce and pinus wood samples using this method, finding that the surface wettability decreased and water absorption varied depending on the species. While water absorption resistance increased in spruce samples, no improvement was observed in pine samples (Kymäläinen et al.) [20]. Later studies on the effects of treatment temperature and time showed that higher temperature and shorter treatment time reduced the moisture sorption and water absorption but also reduced the surface layer thickness [21]. It was confirmed that the hygroscopicity of SLTM samples was lower in the surface layer compared to the core layer, as discussed by Čermák et al. [22] and Ding et al. [23], who attributed this difference to variations in chemical composition and structure. Despite these advances, limitations remain in controlling surface layer thickness in SLTM wood prepared by the hot plate method.

Cyclic-gradient thermal modification has shown great potential in controlling surface layer thickness [24,25], with the prepared surface layer thickness reaching up to 12 mm. Samples treated at 215 °C exhibited excellent overall performance, with ASE increasing by more than 20%, and the transverse compressive strength being more than 40% higher than that of CoTM samples at the same temperature. However, research on this method remains limited. Furthermore, with the increasing emphasis on energy conservation and emission reduction, the energy consumption associated with wood thermal modification technologies has garnered significant attention. However, there is still a blank in the research on the energy consumption involved in SLTM. Therefore, this study aimed to prepare SLTM wood using cyclic-gradient thermal modification, establish the relationship between treatment conditions and wood properties, and reduce energy consumption in the thermal modification process. This study provides valuable insights and references for the energy-efficient utilization of plantation wood.

2. Materials and Methods

2.1. Materials

A typical fast-growing hardwood species, Poplar wood (Populus euramericana cv. ‘San Martina I-72/58’), was chosen as the experimental material. A total of 57 defect-free samples, each with dimensions of 50 mm × 50 mm × 360 mm (radial × tangential × longitudinal), were obtained from two logs sourced from the 12th to 19th growth rings. Of these, 21 samples were subjected to CoTM, while 36 samples were used for SLTM. Detailed information regarding the logs and the distribution of samples for each treatment condition are provided in

Table 1. Sample information for thermal modification.

Logs ID	Height Above Ground (m)	Growth Ring Region	Oven-Dry Density (g∙cm−3)	Dimensions of Samples (R × T × L, mm3)	Number of Samples for CoTM	Number of Samples for SLTM
P1	1.3–4	12–17	0.42 ± 0.02	50 × 50 × 360	14	24
P2	1.3–4	13–19	0.45 ± 0.04	50 × 50 × 360	7	12

.

2.2. SLTM and CoTM Processes

The preparation of SLTM and CoTM followed the methodology outlined in our previous research [24], as illustrated in Figure 1. CoTM was based on the Thermwood process, which comprised five stages. Stages I and II were focused on the drying of the samples, while Stages III, IV, and V corresponded to the heating, effective thermal treatment, and cooling stages, respectively.

Prior to SLTM, the samples were oven-dried. The drying procedure involved an initial phase at 60 °C for 4 h, followed by 80 °C for 4 h, and was completed at 103 ± 2 °C for 12 h. After drying, the samples were directly treated at the target temperature. SLTM was performed in three distinct stages: Stage I involved the heating phase, during which a temperature gradient was established between the surface and the target surface layer point of the sample. Stage II commenced when the surface temperature of the sample reached 180 °C and continued until the temperature at the target surface layer point also reached 180 °C, signifying the effective thermal treatment of the target surface layer. Stage III represented the cooling phase, during which the samples were removed from the treatment chamber and placed in a cooling box (−20–0 °C) until the temperature of the entire sample dropped below 100 °C. The heating, thermal treatment, and cooling cycles were repeated until the cumulative time in Stage II satisfied the required treatment conditions.

For SLTM, we established 12 groups by varying three treatment temperatures (185 °C, 200 °C, and 215 °C), two target surface layer thicknesses (6 mm and 12 mm), and two treatment times (2 h and 3 h). Similarly, for CoTM, 6 groups were prepared using the same temperatures and time. An additional group of oven-dried samples served as a control, and each group consisted of three replicates.

2.3. Mass Loss Analyses

The profile densities of the SLTM samples before and after treatment were measured using an X-ray contour densitometer (EWS, Mönchengladbach, Germany). The profile density after treatment was corrected and the mass loss (ML) distribution of the samples was calculated based on the vertical density profile (VDP) results.

Portions of the wood where ML exceeded the overall average were designated as the surface layer, while the remaining sections were classified as the core layer. The distance from the sample surface to the point where the ML equals the overall average ML was defined as the actual surface layer thickness (SL_A). The average surface layer mass loss (ML_SL) was calculated as the ratio of the integral area of ML within the actual surface layer region to the actual surface layer thickness. The distribution of ML in all samples was categorized into four gradients: 0% < ML ≤ 1%, 1% < ML ≤ 2%, 2% < ML ≤ 3%, and ML > 3%. The thickness proportions of these gradients were compared to evaluate differences in treatment intensity.

2.4. Color Distribution Analyses

The sampling for color testing of the samples is presented in Figure 1. For CoTM, the surface color of the samples was chosen to represent the overall color of the samples. For SLTM, the samples were sawed to expose three different thicknesses of the surfaces: the surface of the samples, the thickness of 6 mm or 12 mm (representing the boundary layer between the surface and core layers), and the thickness of 18 mm (representing the core layer). Color parameters were measured at five random positions on each surface for each of the three different thicknesses.

The color of the wood samples was determined using a spectrophotometer (3nh Technology Co., Ltd., Shenzhen, China) based on the CIE L*a*b* system. Measurements were conducted within a three-dimensional colorimetric framework, utilizing the L*, a*, and b* color coordinates, as described in previous studies [26]. Specifically, for the SLTM samples, the color was measured on the wood surface as well as at thicknesses of 6 mm, 12 mm, and 18 mm. Each test group, including SLTM, CoTM, and untreated samples, was evaluated using five replicate specimens. The total color change ΔE* is used to evaluate the color difference between various samples, according to ISO/CIE 11664–6 (2022) [27].

2.5. Water Absorption Test

Before the water absorption test, all samples were oven-dried at 103 ± 2 °C. They were then fully immersed in deionized water at 20 ± 2 °C. The samples for water absorption tests were cut as shown in Figure 1, with dimensions of 50 mm × 50 mm × 20 mm (radial × tangential × longitudinal). Each test group consisted of three replicate samples, for a total of 57 samples. The samples were periodically weighed, with surface water carefully removed before each weighing. The water absorption amount was calculated by Equation (1):

$$A_{water}=\left({\displaystyle \frac{m_i-m_0}{m_0}}\right)\times 100\%$$

(1)

where m_i is the mass of the samples after water soaking (g), and m₀ is the initial mass of oven-dry samples before water soaking (g). The maximum water absorption amount (Amax) was considered to be reached when the change in water absorption amount between the last two measurements was less than 5% [28]. Then, the anti-swelling efficiency (ASE) of the samples was calculated by Equation (2):

$$ASE=\left({\displaystyle \frac{{\alpha }_0-{\alpha }_t}{{\alpha }_0}}\right)\times 100\%$$

(2)

where α₀ is the volumetric swelling coefficient of untreated samples at Amax state, and α_t is the volumetric swelling coefficient of thermal modification samples at Amax state. The water absorption tests were conducted twice. In the first test, the samples had unsealed end faces, and the dimensional changes in the longitudinal, radial, and tangential directions were measured. After reaching the water saturation state, the samples were conditioned for 8 weeks at 20 °C and 65% relative humidity (RH). Subsequently, the end faces (cross-sections) of the samples were sealed with epoxy resin, and the samples were oven-dried at 103 ± 2 °C. In the second test, the water absorption amount was measured again.

2.6. Moisture Sorption Test

The moisture absorption and desorption of the samples were conducted progressively at 40 °C and a series of relative humidities (0%, 11%, 32%, 53%, 75%, and 96% RH), with the corresponding RH levels regulated using saturated salt solutions of P₂O₅, LiCl, MgCl₂, NaBr, NaCl, and K₂SO₄. An electronic balance with a precision of 0.001 g was used to record the mass changes in the samples. When the mass changes in the sample did not exceed 0.1%, it was considered to have reached equilibrium at the corresponding RH, and the samples were then moved to the next stage of moisture absorption or desorption. The variation in equilibrium moisture content (VEMC) was calculated using Equation (3):

$${\mathrm{V}}_{\mathrm{E}\mathrm{M}\mathrm{C}}={|\mathrm{E}\mathrm{M}\mathrm{C}}_1-{\mathrm{E}\mathrm{M}\mathrm{C}}_2|$$

(3)

where EMC₁ and EMC₂ represent the EMC of the samples under two consecutive RH conditions, and V_EMC is the absolute difference between EMC₁ and EMC₂. The rate of hygroscopic hysteresis (R_HR) was calculated using Equation (4):

$${\mathrm{R}}_{\mathrm{H}\mathrm{R}}={\displaystyle \frac{{\mathrm{R}}_{{\mathrm{E}\mathrm{M}\mathrm{C}}_{\mathrm{d}\mathrm{e}}}-{\mathrm{R}}_{{\mathrm{E}\mathrm{M}\mathrm{C}}_{\mathrm{a}\mathrm{b}}}}{{\mathrm{R}}_{{\mathrm{E}\mathrm{M}\mathrm{C}}_{\mathrm{a}\mathrm{b}}}}}\times 100\%$$

(4)

where ${\mathrm{R}}_{{\mathrm{E}\mathrm{M}\mathrm{C}}_{\mathrm{d}\mathrm{e}}}$ and ${\mathrm{R}}_{{\mathrm{E}\mathrm{M}\mathrm{C}}_{\mathrm{a}\mathrm{b}}}$ denote the EMC during desorption and absorption, respectively, under the same temperature and humidity conditions (%). The moisture sorption samples were cut to dimensions of 50 mm × 50 mm × 4 mm (radial × tangential × longitudinal, Figure 1). A total of 114 samples were prepared, with each experimental group consisting of six replicates.

2.7. Energy Consumption

Taking the calculation of energy consumption for the treatment of a 1 m³ wood sample with an initial moisture content of 12% as an example, and the size of the heat treatment chamber is 2 m × 2 m × 2 m, the specific design parameters refer to Wang et al. [29]. According to the research of Wang et al. [29] and Park et al. [30], the energy consumption during the treatment can be classified into the following seven categories: increase the temperature of wood substance (E₁), overcome the absorptive power of bound water (E₂), evaporate the water that is eliminated from wood (E₃), heat the air and produce the superheated steam in the chamber (E₄), descend the temperature of wood samples (E₅), heat loss of chamber structure (E₆) and fan use (E₇). E1 was calculated by Equations (5) and (6):

$$E_1=W_0·C_{wood}·(T_1-T_0)$$

(5)

$$C_{wood}=0.268+0.00055(T_1+T_0)$$

(6)

where W₀ is the oven dry weight of wood (kg), C_wood is the specific heat of wood substance (kJ∙kg⁻¹∙°C⁻¹), T₀ is the initial temperature of wood before heat treatment (°C), and T₁ is the final temperature of wood after heat treatment (°C).

E₂ was calculated by Equations (7) and (8):

$$E_2=W_0·h_{de}$$

(7)

$$h_{de}={10}^{1.2335-5.408{mc}_0}$$

(8)

where h_de is the heat of desorption per unit weight of wood (kJ/kg), and mc₀ is the bound water content in the wood.

E₃ was calculated by Equations (9) and (10):

$$E_3=W_0·∆mc·\left(\left(T_1-T_a\right)·c_{water}+h_{water}\right)$$

(9)

$$h_{water}=598.25-0.61T_1$$

(10)

where Δmc is the difference between initial and final moisture contents, T_a is the temperature of the air surrounding the chamber (°C); the environmental conditions before treatment are 20 °C and 65% RH), and h_water is the latent heat of vaporization (kJ∙kg⁻¹).

E₄ was calculated by Equations (11)–(15):

$$E_4=E_{4a}+E_{4w}+E_{4v}+E_{4d}$$

(11)

$$E_{4a}=c_{air}·m_{air}·\left(T_{a1}-T_{a0}\right)$$

(12)

$$E_{4w}=W_{water}·\left(100-T_{w0}\right)·c_{water}$$

(13)

$$E_{4v}=W_{water}·h_{water}$$

(14)

$$E_{4d}=W_{water}·\left(e_1-e_0\right)$$

(15)

where E_4a is the heat energy required to heat the air in the chamber, c_air is the specific heat of the air in the chamber (taken as 1.00 kJ∙kg⁻¹∙°C⁻¹), m_air is the mass of the air in the chamber (kg), Ta0 is the initial temperature of the air before heat treatment (°C), and T_a1 is the final temperature of the air after heat treatment (°C). Where E_4w is the heat energy required to heat the water from the current temperature to 100 °C, W_water is the amount of distilled water that produces superheated steam in the chamber (kg, according to the product of the specific heat capacity of superheated steam under treatment conditions and the volume of the chamber, it can be calculated that at least 6.57–8.81 kg of distilled water is required for CoTM and SLTM, here taken as 20 kg), T_w0 is the initial temperature of water before heating (°C), and c_water is the specific heat of the water (taken as 4.19 kJ∙kg⁻¹∙°C⁻¹). Where E_4v is the heat energy required to convert water at 100 °C to steam at 100 °C. Where E_4d is the heat energy required to convert the steam at 100 °C into superheated steam at treatment conditions, e₀ is the enthalpy of steam at 0.1 MPa and 100 °C (kJ∙kg⁻¹), and e₁ is the enthalpy of superheated steam at treatment conditions (kJ∙kg⁻¹). For CoTM, during the cooling process, it is necessary to control the cooling rate while the heating device is still running. E₅ was calculated using Equations (4) and (5). For SLTM, heating and cooling can be implemented through continuous production methods, and E₅ also includes infiltration cooling load (Einf), which was calculated by Equations (16) and (17):

$$E_{inf}=\eta G·\left({\rho }_{out}·c_{out}·\left(T_{out}-T_{in}\right)\right)+G·{\rho }_{out}·(d_{out}-d_{in})·(r_g+r_s)$$

(16)

$$G=0.221A·{\left(g{·H}_{door}\right)}^{{\displaystyle \frac{1}{2}}}·{\left({\displaystyle \frac{{\rho }_{in}-{\rho }_{out}}{{\rho }_{in}}}\right)}^{{\displaystyle \frac{1}{2}}}·{\left({\displaystyle \frac{2}{1+{\left(\frac{{\rho }_{in}}{{\rho }_{out}}\right)}^{\frac{1}{3}}}}\right)}^{{\displaystyle \frac{2}{3}}}$$

(17)

where η is the air curtain coefficient (taken as 0.8 [31]), G is the infiltration flow rate, ρ_in is the air density outside the cooling chamber, ρ_in is the air density in the cooling chamber, d_out is the moisture content of the air outside the cooling chamber (g∙kg⁻¹), din is the moisture content of the air in the cooling chamber (g∙kg⁻¹), and r_g and r_s is the latent heat of vaporization and fusion (kJ), respectively. Where A is the door area (taken as 4 m², the size of the cooling chamber is the same as the heating), H is the height of the door (taken as 2 m), and g is the acceleration due to gravity (assumed to be 9.8 m∙kg⁻¹). The heat treatment chamber structure refers to Wang et al. [29] and the cooling chamber structure refer to Niu et al. [31] and Tian et al. [32]

E₆ was calculated by Equation (18):

$$E_6=E_{6h}+E_{6c}=7.92·S_h{·R}_h·(T_2-T_a)·t_h+S_c{·R}_c·(T_{out}-T_{in})·t_c$$

(18)

where S_h is the surface area of the heat treatment chamber (m²), S_c is the surface area of the cooling chamber (m²), R_h is the thermal resistance of the heat treatment chamber (W∙(m²∙°C)⁻¹), R_c is the thermal resistance of cooling chamber (W∙(m²∙°C)⁻¹), T₂ is the temperature of heat treatment chamber after cooling, t_h is the heat treatment time and t_c is the cooling time (h).

E₇ was calculated by Equation (19):

$$E_7=P_f·n·t$$

(19)

where P_f is the power of one fan, n is the number of fans, and t is the work time of the fan.

The CO₂ emission intensity of coal-fired thermal power is 0.9–1.0 kg∙(kW∙h)⁻¹ [33,34]. In this study, this value was taken as 0.95.

2.8. Statistical Analysis

Statistical analyses were performed using SPSS 27.0 (IBM Corp., Armonk, NY, USA), and p values 0.05 were used to indicate statistical significance. The color parameters and dimensional stability of the samples under different thermal modification conditions were subjected to univariate post hoc multiple comparison analysis.

3. Results and Discussion

3.1. Mass Change

Figure 2 illustrates the VDP changes in samples before and after thermal treatment, revealing a reduction in density after treatment. At temperatures below 160 °C, chemical transformations in wood components were initiated, with hemicellulose undergoing deacetylation reactions [35]. The acetic acid released during this process not only accelerated the degradation of hemicellulose but also catalyzed the hydrolysis of cellulose [36]. Furthermore, the volatilization of extractives contributed to an increase in ML [37], leading to a consequent reduction in density. By comparing the oven-dry density of the samples before and after treatment, the radial distribution of ML was determined, as depicted by the purple solid line in Figure 2. The results indicated that under various treatment parameters, ML in the radial direction (corresponding to the thickness direction) followed a gradient, with greater ML observed in the surface layer compared to the core layer. These findings confirmed that the cyclic-gradient thermal treatment effectively produced SLTM wood under different treatment conditions.

Table 2. Treatment effects of various SLTM and CoTM samples.

Treatment Type	Temperature (°C)	SLT (mm)	Time (h)	SLA (mm)	MLSL (%)	ML (%)	PT with 0% < ML ≤ 1% (%)	PT with 1% < ML ≤ 2% (%)	PT with 2% < ML ≤ 3% (%)	PT with 3% ≤ ML (%)
SLTM	185	6	2	8.28	1.69	1.36	22.01	73.95	4.04	0
			3	6.50	2.34	1.67	21.18	79.51	1.01	0
		12	2	10.03	1.87	1.57	0	88.65	11.35	0
			3	9.23	2.66	2.21	0	57.72	35.05	7.23
	200	6	2	6.44	2.08	1.55	0	87.22	12.78	0
			3	5.84	2.74	1.96	0	76.58	16.07	7.35
		12	2	10.54	2.38	1.78	0	63.79	31.80	4.41
			3	11.64	3.08	2.38	0	43.31	17.30	39.39
	215	6	2	6.35	2.95	2.00	0	74.00	12.74	13.26
			3	6.49	3.74	2.40	0	71.98	7.07	20.95
		12	2	12.75	2.79	2.26	0	38.79	49.17	12.04
			3	12.61	3.54	2.96	0	32.56	26.71	40.73
CoTM	185	-	2	-	-	1.38	-	100	-	-
		-	3	-	-	1.84	-	100	-	-
	200	-	2	-	-	4.24	-	-	-	100
		-	3	-	-	5.68	-	-	-	100
	215	-	2	-	-	7.34	-	-	-	100
		-	3	-	-	9.69	-	-	-	100

Note: Target surface layer thickness (SL_T); actual surface layer thickness (SL_A); average surface layer mass loss (ML_SL); and proportion of thickness (PT).

summarizes the effects of SLTM under varying parameters. SL_A closely corresponded to SL_T, demonstrating the controllability of surface layer thickness preparation at different treatment conditions. As the treatment temperature, time, and target surface layer thickness increased, the proportion of samples (PT) with ML ≥ 3% rose, while the PT with 1% < ML ≤ 2% declined. This trend suggested that the intensity of treatment in the surface layer increased with the severity of the treatment conditions [21]. However, for samples treated at 185 °C with a target surface layer thickness of 6 mm, the PT with ML ≥ 2% did not exhibit a significant increase, and the reduction in the PT with 0% < ML ≤ 1% was negligible. This outcome was attributed to the relatively low treatment temperature, which limited the temperature gradient between the surface and core layers [20], resulting in comparable ML across both regions. Additionally, increasing the target surface layer thickness led to a higher PT with middle ML (1% < ML ≤ 2%), indicating a reduction in the difference in treatment intensity between the surface and core layers.

For CoTM, both the heating and cooling rates during the treatment process were controlled, resulting in a uniform temperature distribution throughout the samples, which in turn led to an even ML. The treatment intensity also increased with the severity of the treatment conditions. The ML of the SLTM samples was between that of CoTM185-2h and CoTM200-2h. Furthermore, the treatment intensity of the target surface layers of most SLTM samples was lower than that of CoTM200-2h (4.24%). Only when the temperature reached 215 °C did the ML_SL of some SLTM samples exceed this value. This was mainly influenced by the temperature distribution during the treatment process, as the target surface layer temperature of most SLTM samples did not reach 200 °C, or the exposure time was too short. In the study by Šeda et al. [21], the hot plate method achieved significant surface layer mass loss (4.5%–8.2%) in a short time (4 min–6 min). This was due to the metal plate’s ability to transfer heat rapidly to the wood surface, although the heat transfer depth was relatively shallow (approximately 2 mm). However, the treatment time would need to be extended accordingly to increase the treatment depth.

3.2. Color Changes

The color results of treated samples are presented in Figure 3. Figure 3a–d shows the variations in L*, a*, b*, and ΔE* values for untreated and CoTM samples. As the temperature and time increased, the L* values consistently decreased, while the a* values exhibited no discernible trend. The b* values initially increased and then decreased, and the ΔE* values consistently increased. These results were consistent with previous studies [26]. Wood contained chromophoric groups (carbonyl, carboxyl, unsaturated double bonds, conjugated structures, etc.) and auxochromic groups (hydroxyl groups, etc.), which were primarily found in the lignin structure [38]. As the intensity of the treatment increased, the cellulose and hemicellulose polysaccharides were reduced, leading to the generation of more carbonyl and carboxyl groups [39,40]. In addition, the increase in lignin content and the oxidation reactions in lignin caused changes in the color of the wood [39,41]. The extent of changes in L* and ΔE* values was positively correlated with the treatment intensity [42,43]. Furthermore, the effect of treatment temperature on the color change was found to be more significant than the time [44].

Figure 3e–h displayed the changes in L* and ΔE* values at the surface (Sf), the boundary between the surface and the core layers (Bd), and the core (Co) of SLTM samples. The L* and ΔE* values of SLTM samples exhibited an increasing trend from the surface toward the core. For all SLTM samples, the Sf values of L* and ΔE* were found to fall between those of CoTM200-2h (CoTM at 200 °C for 2 h) and CoTM215-2h (CoTM at 215 °C for 2 h). The L* and ΔE* values at Bd and Co were between those of the untreated samples and CoTM185-2h (CoTM at 185 °C for 2 h). These color changes also reflected differences in the treatment intensity at various positions within the SLTM samples [45,46].

When the target surface layer thickness was 6 mm, the difference in L* values between Bd and Co, as well as the difference in ΔE* value, increased as the treatment temperature and time were raised, indicating that increasing these two treatment parameters enhanced the differences in treatment intensity between the surface and core layers of SLTM wood. In contrast, when the target surface layer thickness was 12 mm, the L* and ΔE* values at Bd and Co showed no significant difference, suggesting that increasing the surface layer thickness reduced the differences in treatment intensity between the surface and the core layers.

3.3. Water Absorption and Dimensional Stability

Figure 4a,d illustrate the variations in water absorption amount of CoTM and SLTM samples over the soaking time, respectively. The water absorption amount of all samples gradually decreased with increasing soaking time and stabilized after 30 days. The results in

Table 3. Multiple comparisons for the water absorption of CoTM samples at different temperatures and times.

Temperature (°C)	Time (h)	Amax (%)	Duncan Grouping	ASE (%)	Duncan Grouping	Differential Swelling	Duncan Grouping
Untreated		86.6	A	-	-	3.3	A
185	2	81.6	AB	18.0	A	3.0	A
	3	80.1	ABC	21.4	A	3.0	A
200	2	74.3	ABC	37.9	B	2.9	A
	3	74.4	ABC	43.9	BC	2.8	A
215	2	65.2	BC	49.2	BC	2.8	A
	3	62.9	C	54.8	C	2.6	A

Note: The same letter in the same column indicates no significant difference between the two groups, while different letters indicate a significant difference, significance level of p < 0.05.

and

Table 4. Multiple comparisons for the water absorption of SLTM samples at different temperatures and times.

Temperature (°C)	SLT (mm)	Time (h)	Amax (%)	Duncan Grouping	ASE (%)	Duncan Grouping	Differential Swelling	Duncan Grouping
185	6	2	84.5	A	17.2	A	3.3	A
		3	83.8	A	23.5	ABC	3.2	A
	12	2	80.7	AB	21.0	AB	3.1	A
		3	80.8	AB	25.0	BC	3.1	A
200	6	2	79.7	AB	22.6	ABC	3.1	A
		3	78.9	AB	25.9	BC	3.1	A
	12	2	78.8	AB	25.0	BC	3.0	A
		3	77.6	AB	28.7	CD	3.0	A
215	6	2	76.0	AB	26.7	BC	2.9	A
		3	74.4	AB	33.3	BE	3.0	A
	12	2	74.7	AB	38.7	EF	2.9	A
		3	72.5	B	43.3	F	2.9	A

Note: The same letter in the same column indicates no significant difference between the two groups, while different letters indicate a significant difference, significance level of p < 0.05.

showed that for the CoTM samples, a significant reduction in water absorption occurred when the temperature exceeded 215 °C, while the effect of treatment time on water absorption was negligible [47]. For the SLTM samples, significant differences in water absorption were observed only under extreme conditions (between SLTM215-12-3h with SLTM185-6-2h or SLTM185-6-3h). The target surface layer thickness exhibited no significant influence on water absorption. The impact of thermal treatment on the water absorption properties of wood remains debated. On the one hand, the degradation of the amorphous regions of hemicellulose and cellulose [48], the polymerization and redistribution of lignin within the cell lumens and cell walls, and the closing of pits [49] reduced the water absorption capacity of the treated wood. On the other hand, the increased presence of micropores following thermal treatment generated higher capillary tension, resulting in greater water absorption amount [50,51].

The ASE values of CoTM and SLTM samples, as presented in Figure 4b,e, demonstrated that thermal modification significantly enhanced the dimensional stability of wood. The ASE values of CoTM samples ranged from 18.0% to 54.8%, which were similar to the findings of previous studies [52,53]. For SLTM samples, the ASE values ranged from 17.2% to 43.3%. Notably, the statistical analysis revealed that when the temperature of SLTM was below 215 °C, the effects of temperature, treatment time, and target surface layer thickness on ASE were not significant (Table 4). The ASE of the samples ranged between CoTM185-2h (18.0%) and CoTM200-2h (43.9%). However, when the temperature exceeded 215 °C (including 215 °C), the ASE of the sample increased significantly with the extension of processing time and the increase in target surface layer thickness. Therefore, to further improve the ASE of SLTM samples, increasing the treatment temperature, extending the treatment time at 215 °C, and increasing the target surface layer thickness (≤12 mm) proved to be effective approaches.

During the SLTM process at 215 °C, when the treatment time was 2 h or 3 h, and the target surface thickness increased from 6 mm to 12 mm, the ASE of the samples increased from 26.7% (SLTM215-6-2h) to 38.7% (SLTM215-12-2h), and from 33.3% (SLTM215-6-3h) to 43.3% (SLTM215-12-3h), respectively. Similarly, when the target thicknesses were 6 mm or 12 mm, and the treatment time was increased from 2 h to 3 h, the ASE of the samples increased from 26.7% (SLTM215-6-2h) to 33.3% (SLTM215-6-3h), and from 38.7% (SLTM215-12-2h) to 43.3% (SLTM215-12-3h), respectively. The influence of target surface layer thickness on ASE was more significant than the effect of treatment time.

Previous studies have highlighted the differing treatment intensities between the surface and core layers of SLTM wood, which influence overall dimensional stability [24]. Figure 4c,f depicted the differences in tangential and radial swelling of the treated samples. Unlike previous reports [54,55], the values of differential swelling after thermal modification were higher and did not show significant changes with varying treatment conditions. Nevertheless, the differential shrinkage results indicated that SLTM did not increase the risk of deformation or cracking of the samples. While differences in swelling rates were observed between the surface and core layers of SLTM samples, the surface layer, which received higher treatment intensity, effectively suppressed excessive deformation in the core layer. For CoTM, the value of differential swelling decreased with increasing temperature or time [56,57].

3.4. Moisture Sorption

Figure 5 presents the effects of CoTM and SLTM on the moisture sorption behavior of wood. As shown in Figure 5a₁, the absorption and desorption isotherms of CoTM wood followed a typical S-shaped curve, classified as Type II adsorption isotherms according to the IUPAC classification [58]. At the same RH levels, the EMC of thermally modified wood was lower than that of untreated wood, indicating that the hygroscopicity of wood decreased after treatment. Both increased temperature and prolonged time further reduced the EMC of the wood [59,60].

Figure 5a₂ shows the V_EMC of CoTM and untreated wood during the absorption and desorption. When the RH exceeded 60%, the V_EMC increased markedly, and the same phenomenon was observed in the studies by Hill et al. [61] and Čermák et al. [62]. This behavior was attributed to the glass transition of amorphous substances in the cell wall under high RH conditions [63], which weakened the rigidity of the cell wall’s network structure [61]. The increased relative slippage of macromolecular chains exposed more sorption sites, leading to greater absorption of multilayer moisture and a more rapid increase in EMC [59]. As temperature and time increased, the rate of EMC growth slowed, which was most often explained by the decrease in available adsorption sites due to the degradation of hemicellulose and the increase in cellulose crystallinity [64], and there was a negative effect of water molecules adsorbed in the monolayer on the multilayer adsorption when a medium or high RH was applied [65].

After thermal treatment, the R_HR of the CoTM samples increased (Figure 5a₃). Hysteresis was also due to the number of sorption sites [66]. Moisture adsorption accompanied swelling, creating volume in the cell wall matrix for the accommodation of moisture molecules. Conversely, the cell wall shrinks during desorption, removing accessible sorption sites. Thermal modification further reduced the number of these sites, thereby increasing the hygroscopic hysteresis ratio [11]. Typically, the extent of sorption hysteresis depends on the treatment temperature [59,64]. In this study, When the RH exceeded 60%, the sorption hysteresis of CoTM215 and CoTM200 samples were almost the same, which was consistent with a previous report [56]. Changes in the chemical composition of wood at different temperatures, especially the cross-linking of lignin, significantly impacted the hysteresis behavior of the treated wood [67].

Figure 5b₁–d₃ shows the moisture sorption behavior of SLTM samples. It can be seen that the moisture sorption of the SLTM sample exhibited the same trend as that of the CoTM sample. At 40 °C and 96% RH, the EMC of all SLTM samples ranged from COTM185-2h (19.5%) to COTM200-2h (14.4%), which was superior to the SLTM samples prepared using the “hot plate method” at 200 °C [23]. For SLTM samples treated at 185 °C and 200 °C, the effect of different treatment conditions on moisture sorption was minimal. Notably, at 215 °C, the sorption isotherms of the SLTM215-12-3h samples, as well as V_EMC, decreased significantly, especially under high RH conditions. The hygroscopic of SLTM samples was influenced by both the surface and core layers. When the temperature was high enough (215 °C), with the increase in surface layer thickness and time, the hygroscopic of the core layer also improved significantly, thereby significantly reducing the overall moisture absorption of SLTM samples.

The R_HR of SLTM samples showed a decreasing trend with the increase in RH. However, there was no significant change in the effect of treatment conditions on the R_HR, which differed from CoTM samples [59]. As mentioned earlier, the R_HR of wood was influenced by both the number of sorption sites and the interactions between chemical components, which requires further research.

3.5. Treatment Time and Energy Consumption

Figure 6a–c presents the average time of Stages I (heating), II (effective thermal treatment), and III (cooling) within a single cycle under different SLTM processes. Higher treatment temperatures significantly shortened the time of a single cycle, primarily due to an increased temperature gradient between the specimen’s interior and exterior, which accelerated the heating rate and markedly reduced the time required for Stage I. Conversely, an increase in the target surface layer thickness necessitated a longer Stage II to ensure that the entire surface layer reached the effective treatment temperature, thereby extending the total time for a single cycle. The time of a single cycle remained unaffected by either the effective treatment time or the time of cycles. Additionally, no significant differences were observed in the cooling times across the different treatment processes.

The total treatment time for SLTM and CoTM is shown in Figure 6d. The heating and cooling rates for CoTM followed the parameters reported in previous studies [68,69]. With the exception of the treatment process at 185 °C for 3 h, SLTM reduced the total treatment time by 1.9 h (SLTM185-6-2h) to 10.8 h (SLTM215-12-2h) compared to CoTM conducted under the same temperature and time conditions. In CoTM, lower heating and cooling rates resulted in a longer time for uniformly processing the wood [70,71]. The reason for the longer total treatment time in SLTM at 185 °C for 3 h was that the lower temperature resulted in a reduced heating rate, thereby extending Stage I time. The increased time of Stage II requires more cycle times. Additionally, the target surface layer thickness affected the total treatment time of SLTM. Although increasing the thickness of the target surface layer prolonged the time of a single cycle, it also extended the time of Stage II, reducing the number of cycles and thus shortening the total treatment time.

Table 5. The energy consumption and CO₂ emission for SLTM and CoTM.

Treatment Type	Temperature (°C)	Target Surface Layer Thickness (mm)	Time (h)	E1 (kJ)	E2 (kJ)	E3 (kJ)	E4 (kJ)	E5 (kJ)	E6 (kJ)	E7 (kJ)	Etotal (kJ)	Consumed Electrical Energy (kW∙h)	CO2 Production (kg)
SLTM	185	6	2	144,672	1614	44,565	63,385	1,117,074	1,628,698	973,901	3,973,908	1104	1049
			3	212,938	1614	44,565	63,385	1,755,401	1,737,032	1,011,917	4,826,851	1341	1274
		12	2	112,887	1614	44,565	63,385	797,910	1,509,109	931,738	3,461,207	961	913
			3	182,719	1614	44,565	63,385	1,436,237	1,964,683	1,092,442	4,785,644	1329	1263
	200	6	2	174,909	1614	44,565	65,248	1,463,439	1,510,883	912,038	4,172,696	1159	1101
			3	241,616	1614	44,565	65,248	2,113,857	1,808,361	1,008,115	5,283,376	1468	1394
		12	2	115,247	1614	44,565	65,248	813,022	1,351,204	860,544	3,251,443	903	858
			3	150,950	1614	44,565	65,248	1,138,231	1,569,320	931,046	3,900,972	1084	1029
	215	6	2	208,263	1614	44,565	67,109	1,823,038	1,498,720	891,302	4,534,610	1260	1197
			3	274,970	1614	44,565	67,109	2,485,961	1,836,157	991,872	5,702,247	1584	1505
		12	2	117,618	1614	44,565	67,109	828,654	1,288,486	828,749	3,176,794	882	838
			3	190,603	1614	44,565	67,109	1,491,577	1,650,179	936,576	4,382,222	1217	1156
CoTM	185	-	2	26,386	1614	49,450	55,467	11,896	2,999,016	2,413,440	5,557,269	1544	1467
			3	26,386	1614	49,450	55,467	11,896	3,096,794	2,448,000	5,689,606	1580	1501
	200	-	2	29,408	1614	49,450	56,072	14,918	3,284,943	2,499,840	5,936,245	1649	1567
			3	29,408	1614	49,450	56,072	14,918	3,391,610	2,534,400	6,077,472	1688	1604
	215	-	2	32,535	1614	49,450	56,676	18,045	3,597,536	2,586,240	6,342,095	1762	1674
			3	32,535	1614	49,450	56,676	18,045	3,713,092	2,620,800	6,492,211	1803	1713

summarizes the energy consumption and CO₂ emissions associated with SLTM and CoTM. SLTM not only reduced the total treatment time but also achieved a 12.1% to 49.9% reduction in energy consumption compared to CoTM conducted at the same temperature and time conditions. For every 1 m³ of wood processed, SLTM resulted in a reduction in CO₂ emissions by 73 kg to 729 kg compared to CoTM under the same conditions. As shown in Table 5, the cyclic heating-cooling process of SLTM directed most of the energy toward raising the temperature of the wood substance (E₁) and lowering the temperature of the specimens (E₃). In contrast, CoTM, with its longer heating and cooling phases, consumed a larger proportion of energy in heat loss through the chamber shell (E₆) and fan use (E₇).

3.6. Evaluation of SLTM and CoTM

Figure 7 provided a comprehensive analysis of the CoTM and SLTM samples, evaluating ML, ΔE* of the sample’s surface, ASE, EMC at 96% RH, and energy consumption, with level 5 representing optimal efficiency or performance. As shown in Figure 7a, the CoTM samples demonstrated excellent performance in individual properties. For example, achieving superior dimensional stability required more energy, resulting in a higher ML. In contrast, maintaining sample density and reducing energy consumption during production resulted in lower dimensional stability. SLTM wood exhibited a more balanced overall performance (Figure 7b–d), making it suitable for applications such as building structures. Moreover, the continuous production mode (which adhered to the cyclic thermal gradient in SLTM) and energy-efficient production processes endowed SLTM with significant potential for further development. Nevertheless, when the SLTM wood was processed, and the core layer was directly exposed to the external environment, other waterproofing methods, including coating, were effective in ensuring the proper use of SLTM wood. SLTM wood was expected to find applications in structural engineering, landscape design, garden construction, dock building, and other fields requiring materials with excellent mechanical strength and dimensional stability.

4. Conclusions

This study prepared SLTM samples using cyclic-gradient thermal treatment under three temperature conditions (185 °C, 200 °C, 215 °C), two time (2 h, 3 h), and two target surface layer thicknesses (6 mm, 12 mm). The differences in ML, color change, water absorption, moisture sorption, and energy consumption between SLTM and CoTM samples were analyzed under various treatment conditions.

The results demonstrated that controlling the thermal gradient between the surface and core layers enabled the controlled preparation of SLTM wood with specific surface layer thicknesses. After treatment, the ML of SLTM samples showed a distribution of large surface layer loss and small core layer loss, and the surface color of the SLTM wood was darkened, enhancing its decorative appearance. The dimensional stability of SLTM wood improved, with the ASE for water absorption increasing by a maximum of 2.4 times compared to CoTM185-2h. When the temperature exceeded 215 °C (including 215 °C), the water absorption ASE of samples significantly increased with the extension of treatment time and the increase in target surface layer thickness. The target surface layer thickness had a more significant influence on water absorption ASE than treatment time. At 96% RH, the EMC of SLTM wood ranged from 16.6% to 19.2%, and the hygroscopicity of SLTM samples decreased to varying degrees compared to CoTM185-2h (19.5%). The hygroscopic of SLTM samples was influenced by both the surface and core layers. When the temperature reached 215 °C, increasing the surface layer thickness and treatment time significantly reduced the hygroscopic of SLTM samples.

Furthermore, SLTM reduced the total treatment time by 1.9 h to 10.8 h and reduced energy consumption by 12.1% to 49.9% compared to CoTM at the same treatment conditions. Considering the water absorption and moisture sorption of SLTM samples were basically between those of CoTM185-2h and CoTM200-2h samples. Except for SLTM215-6-3h, SLTM saved 215 kW∙h to 215 kW∙h of energy, leading to a reduction in CO₂ emissions by 73 kg to 729 kg per 1 m³ of wood processed.

Further studies are needed to evaluate the mechanical properties of SLTM wood under different treatment conditions and optimize its performance in all aspects. However, this study provides valuable insights and references for the efficient and energy-saving utilization of plantation wood.