Optimization of Hydrothermal and Oleothermal Treatments for the Resistance of Dabema (Piptadeniastrum africanum (Hook.f.) Brenan) Wood

Abstract

This study evaluates the effects of hydrothermal and oleothermal treatments on the physical, colorimetric, and mechanical properties of Dabema wood. Samples were heated at 100, 160, and 220 °C for 2, 3.5, and 5 h. Equilibrium moisture content decreased from 13.16% in untreated wood to approximately 43% lower after hydrothermal treatment at 160 °C for 5 h and to 64% lower after oleothermal treatment at 220 °C for 5 h. Water absorption decreased from 78% in untreated samples to 25%–64% following hydrothermal treatment and to 17%–44% after oleothermal treatment. Hydrothermal treatment caused significant darkening, whereas oleothermal treatment maintained a lighter, more stable color. Mechanical properties improved substantially: in compression, MOE increased by 113% after oleothermal treatment at 220 °C for 5 h. In bending, MOR and MOE rose by 25%–35% under optimal oil-heat conditions. In tensile, MOE increased by 30%, and maximum tensile stress improved by up to 130%. Oleothermal treatments yielded the most stable enhancements, whereas severe hydrothermal treatments sometimes reduced mechanical performance despite improving moisture resistance. Multivariate analysis (PCA) and response surface methodology (RSM) indicate that oleothermal treatment at 160 °C for 3.5–5 h provides the best compromise between stiffness and color stability. Thermogravimetric analyses (TG/DTG) show hydrothermal treatment promotes hemicelluloses degradation, whereas oleothermal treatment stabilizes the cellulose–lignin network. Overall, hydrothermal treatment enhances dimensional stability, while oleothermal treatment achieves an optimal balance of stiffness, mechanical performance, and color retention. Deep color changes from furanic resin formation under hydrothermal conditions are strongly suppressed by oil during oleothermal processing, yielding lighter and more durable wood. For commercial applications such as furniture and structural components, oleothermal treatment is recommended, whereas hydrothermal treatment is more suitable when dimensional stability is prioritized over mechanical performance.

1. Introduction

Natural wood is valued for its high strength-to-weight ratio, attractive grain, and good workability [1,2]. Moreover, its biodegradability and renewability make it a widely used material in furniture and construction [3]. However, as a raw material, wood is sensitive to environmental conditions, particularly humidity, which can cause warping, cracking, and dimensional instability, thereby limiting its practical applications [4,5,6,7].

These limitations are particularly critical for tropical hardwoods used in structural and outdoor applications, where durability and dimensional stability are essential performance criteria [5,6]. Dabema (Piptadeniastrum africanum (Hook.f.) Brenan) is particularly valued in the timber industry for furniture, flooring, joinery, and structural applications, due to its combination of strength, workability, and inherent durability [1]. Improving Dabema’s performance under environmental stress thus justifies investigating treatments capable of enhancing its physical, colorimetric, and mechanical properties.

Tropical species such as Dabema (Piptadeniastrum africanum (Hook.f.) Brenan) are particularly sensitive to environmental variations due to the hygroscopic nature of wood, which strongly influences moisture uptake, dimensional stability, and mechanical behavior under changing climatic conditions [1]. As a result, their physical, hygroscopic, colorimetric, and mechanical properties can be significantly altered by thermal and oil-heat treatments, as widely reported for tropical hardwoods [8,9,10]. Specifically, hydrothermal treatments conducted in water or steam are known to modify wood–water interactions and the cell wall structure, while oil-heat treatments using vegetable oils primarily affect internal cohesion and mechanical performance through temperature-dependent thermo-oxidative processes [9,10,11]. These findings provide a framework to understand the vulnerability of Dabema and other tropical species and justify the investigation of thermal and oil-heat treatments to improve their physical and mechanical properties.

Given this sensitivity, several treatments have been developed to mitigate the resulting dimensional and mechanical variations. Hydrothermal, or wet-heat, treatments enhance dimensional stability by reducing hygroscopicity through partial hemicelluloses degradation and the reduction in hydroxyl groups [12,13]. These effects are generally observed for treatments performed at temperatures between approximately 160 and 250 °C, with longer exposure times increasing chemical degradation [14,15]. However, prolonged exposure or high temperatures can reduce mechanical strength, particularly the modulus of rupture (MOR) and modulus of elasticity (MOE), as observed by Mandraveli et al. [9] for tropical woods treated with palm oil at temperatures above 180 °C, and by Adewopo & Patterson [16] for temperate species treated under similar wet-heat conditions, confirming that thermal degradation affects flexibility and ductility. These results highlight the need to simultaneously assess dimensional and mechanical responses under controlled treatment conditions.

In addition to hydrothermal treatments, oil-heat treatments combine heat and lipid impregnation, partially filling microvoids and stabilizing the cellulose–lignin network. In this method, wood is heated while immersed in vegetable oil, typically in the range of 160–220 °C [9,15]. Hao et al. [17] demonstrated that this technique increases rigidity and limits compression deformation in bamboo treated at 180 °C, while Suri et al. [18] reported warmer, more saturated wood hues with improved color retention under light exposure for oleothermally treated wood samples. These effects are attributed to the protection of cellulose microfibrils by oil and reduced polysaccharide chain mobility, stabilizing the wood mechanically and dimensionally [19,20]. Moreover, as Piao et al. [21] highlighted, the presence of oil during heating limits the hydrolysis of hemicelluloses acetyl groups, thereby strongly restricting the formation of dark, water-repellent furanic resins, which are more pronounced during hydrothermal treatment and are otherwise generated under hydrothermal conditions and contribute to intense darkening and microstructural changes. These contrasting mechanisms suggest that hydrothermal and oleothermal treatments may lead to markedly different property evolutions under identical thermal conditions.

Despite these advances, the combined effects of temperature and duration on the physical, hygroscopic, colorimetric, and mechanical properties of Dabema remain insufficiently elucidated in a direct comparative study. Previous work often focused on a single treatment type or a limited number of properties, precluding a comprehensive comparison. Bessala et al. [22] reported that hydrothermal treatment greatly reduces water absorption while increasing brittleness in Afrormosia and Newtonia, whereas Mandraveli et al. [9] and Taghiyari et al. [11] observed that oleothermal treatment enhances internal cohesion and stabilizes color in tropical species. Therefore, a systematic comparison under identical experimental parameters is required to relate processing conditions to property changes in Dabema. Systematic comparisons under identical experimental parameters remain lacking, limiting objective evaluation of the advantages and drawbacks of each treatment for this species. Based on the literature, moderate oleothermal treatments are expected to simultaneously improve mechanical strength and dimensional stability, whereas hydrothermal treatments primarily increase rigidity and darken the wood, potentially reducing flexibility at high temperatures [23,24,25,26]. These expectations constitute hypotheses that are experimentally tested in this study.

To address existing knowledge gaps, this study aims to systematically characterize and compare the effects of hydrothermal and oleothermal treatments applied across a controlled range of temperatures and durations on the physical, colorimetric, and mechanical properties of Dabema wood. Previous studies indicate that thermal and oleothermal treatments can have contrasting effects on the physical and mechanical properties of temperate and tropical woods, justifying a systematic approach to assess their impact [9,10,15,16,17,18,19,20,21,27].

Thermogravimetric analyses (TG/DTG) will be employed to link macroscopic property changes to the underlying chemical transformations [28,29]. Physical, colorimetric, and mechanical properties will be evaluated simultaneously to provide an integrated understanding of the material’s behavior [30,31]. Multivariate statistical tools, including principal component analysis (PCA) and response surface methodology (RSM), will be used to systematically assess the influence of treatment parameters, cluster samples based on mechanical performance, and optimize treatment conditions. These analyses will enable the interpretation of experimental results in relation to the chemical mechanisms involved in thermochemical wood modification [32,33].

2. Materials and Methods

2.1. Raw Materials

Wood used in this study was heartwood of Dabema (Piptadeniastrum africanum (Hook.f.) Brenan, family Fabaceae), a tropical tree native to Central and West Africa, commonly employed in carpentry and local construction [1,10,34]. Sapwood was removed to minimize anatomical variability and ensure a uniform response to thermal treatments. At 12% moisture content, the heartwood exhibited a density of 0.70–0.85 g/cm³, a total tangential shrinkage of 8.5%, a radial shrinkage of 3.8%, and moderate natural durability against lignicolous fungi causing brown and white rot [10,34]. Wood is characterized by a straight to slightly interlocked grain, a fine to medium texture, and a light to reddish-brown color, with a clearly distinguishable, lighter-colored sapwood. Its mechanical properties support structural applications, with a modulus of rupture (MOR) of 60–100 MPa, a modulus of elasticity (MOE) of 8–12 GPa, and axial compression strength of 40–60 MPa [1]. Distilled water, neutral (pH ≈ 7), particle-free, and with negligible mineral content, was used for hydrothermal treatment to ensure reproducibility and uniform thermal modifications. Mayor peanut oil (Cameroon) was employed for oleothermal treatment; its typical properties are a density of 0.88–0.91 g·cm³ at 20 °C, kinematic viscosity of 30–60 mm²/s at 40 °C, free fatty acid content <0.5% FFA, peroxide value < 10 meq O₂/kg, smoke point of 180–230 °C, and flash point > 200 °C. These characteristics ensure thermal stability within the studied temperature range and limit oxidative degradation during treatment. Oil was verified prior to use for homogeneity and absence of oxidation, ensuring stable and controlled experimental conditions.

2.2. Hydrothermal and Oleothermal Modification of Dabema

Heartwood specimens of Dabema, conditioned at 12% moisture content, were subjected to hydrothermal and oleothermal treatments in an autoclave installed in an electric oven (HERAEUS D-6450 Hanau) following the same thermal cycle. A total of 285 specimens were processed for each treatment method, corresponding to 15 specimens for each combination of temperature and duration. For hydrothermal treatment, the specimens were placed in the autoclave containing water to generate a saturated steam atmosphere, whereas for oleothermal treatment, they were immersed in vegetable oil preheated to 50 °C, with an initial vacuum application (−0.08 MPa, 30 min), gentle agitation (≈60 rpm, 20 min), and an additional one-hour immersion at 50 °C to enhance oil penetration (Figure S1). In both cases, heating was performed from 50 °C to the target temperature (100–220 °C) at a rate of 5 °C/min, followed by an isothermal holding period of 2 h, 3.5 h, or 5 h, and then a controlled cooling at 10 °C/h down to 70 °C, before natural cooling to 50 °C. The autoclave provided a controlled, low-oxygen atmosphere, saturated with steam for hydrothermal treatment and filled with oil for oleothermal treatment, thereby preventing oxidation. The selected temperature and duration ranges align with recommendations for studying thermal modifications in tropical woods and for observing the progressive degradation of hemicelluloses, as well as the resulting physical and mechanical changes in the wood. Temperature and duration ranges were established based on the studies of [9,14,15,16,22].

2.3. Colorimetric and Physical Properties of Modified Wood

Wood specimens measuring 80 × 20 × 10 mm³ (Figure S2) for colorimetric tests and 20 × 20 × 20 mm³ (Figure S1) for the evaluation of physical properties were prepared to conduct a controlled study on the effects of hydrothermal and oleothermal treatments on the chromatic and physical properties of Dabema wood. The influence of treatment type, temperature (100, 160, and 220 °C), and treatment duration (2, 3.5, and 5 h) was specifically examined to quantify the modifications induced by these experimental conditions. Following treatment, the samples were conditioned at 28 ± 2 °C and 80 ± 5% relative humidity until constant mass was achieved, ensuring a comparable hygroscopic equilibrium state for all specimens prior to measurement. Colorimetric measurements (L*, a*, b*) were performed using an SC30 colorimeter (Figure S2) on each tangential face, with four (04) repetitions per face to enhance measurement accuracy and account for the surface heterogeneity of the material. These parameters, including color changes (ΔE*), chroma (C*), and hue angle (h*), were measured using the CIELab system to quantitatively characterize color variations induced by the thermal treatments and interpreted according to established literature methods [35,36,37].

$${∆E}^{\mathrm{*}}\mathrm{ab}=\sqrt{{(∆L^{*})}^2+{(∆a^{*})}^2+({∆b^{*})}^2}$$

(1)

$$C^{\mathrm{*}}=\sqrt{{\left(a^{\mathrm{*}}\right)}^2+{\left(b^{\mathrm{*}}\right)}^2}$$

(2)

$$h^{\mathrm{*}}=arctangent\left({\displaystyle \frac{a^{*}}{b^{\mathrm{*}}}}\right)$$

(3)

Physical properties, including hygroscopic equilibrium density (HED), equilibrium moisture content (EMC), volumetric shrinkage and swelling (VSH, VSW), and water absorption (WA), were determined after oven-drying the samples at 103 ± 2 °C, based on dry and saturated masses and volumes using the Archimedes method and in accordance with standard NF B 51-005 [38]. Each test was conducted on five (05) samples to ensure the reliability of the results.

$$\mathrm{H}\mathrm{E}\mathrm{D}={\displaystyle \frac{M}{V}}$$

(4)

where M (g) represents the mass of the specimen, while V (cm³) represents its volume.

$$\mathrm{E}\mathrm{M}\mathrm{C}\left(\%\right)={\displaystyle \frac{W_2-W_1}{W_1}}\times 100$$

(5)

where W₁ represents the oven-dry weight (g), while W₂ represents the constant weight after reconditioning (g).

$$\mathrm{V}\mathrm{S}\mathrm{H}\left(\%\right)={\displaystyle \frac{V_s-V_0}{V_s}}\times 100$$

(6)

where V_S is the volume in cm³ of the saturated specimen, and V₀ is the volume in cm³ of the oven-dry specimen.

$$\mathrm{V}\mathrm{S}\mathrm{W}\left(\%\right)={\displaystyle \frac{V_s-V_0}{V_0}}\times 100$$

(7)

where V_S is the volume in cm³ of the saturated sample, and V₀ is the volume in cm³ of the oven-dry specimen.

$$\mathrm{W}\mathrm{A}\left(\%\right)={\displaystyle \frac{M_s-M_0}{M_0}}\times 100$$

(8)

where M_s (g) represents the saturated mass of the test piece, while M₀ (g) represents its anhydrous mass.

2.4. Morphology of Wood Before and After Modification

Surfaces of the reference (Ref) samples and those subjected to hydrothermal and oleothermal treatments of Dabema were examined using a binocular stereo microscope (S/Dragon D75T/0750HT). Samples from the nineteen (19) experimental groups (G1 to G19) were sectioned with a microtome to obtain smooth, comparable surfaces without prior embedding, preserving the native cell wall structure [39]. Observations were conducted at a magnification corresponding to a 100 μm field of view, with three (03) independent observations per section to qualitatively assess the effects of the treatments on wood microstructure compared to the reference samples.

2.5. Mechanical Properties of the Modified Wood

Mechanical tests were carried out in accordance with the Chinese standards GB/T 1935 [40], GB/T 1936.1 [41] and GB/T 1938 [42] (Figure S3), corresponding to the evaluation of compression, bending, and tensile properties, respectively. A three-point bending configuration was adopted for the flexural tests. Prior to testing, all specimens were conditioned at 25 ± 2 °C and 82 ± 5% RH for 24 h. A total of (285) specimens were prepared for each mechanical test (compression, three-point bending, and tensile), corresponding to (15) replicates for each combination of treatment temperature and duration. Mechanical characterization was conducted in Bamenda (5°57′34.92″ N; 10°08′45.49″ E) using a LARYEE universal testing machine (model UE3450-300 kN), featuring a precision of ±0.5% and a crosshead speed of 2 mm·min⁻¹, ensuring stable and reliable loading conditions.

2.6. Multivariate Analysis (PCA, HCA) and RSM Optimization of Dabema Wood Treatments (Piptadeniastrum africanum (Hook.f.) Brenan)

The study focused on Dabema wood, including reference samples, as well as samples treated hydrothermally or oleothermally at three temperatures (100, 160, and 220 °C) and three treatment durations (2, 3.5, and 5 h). Data from mechanical tests (compression, bending, and hardness) were centered and scaled to eliminate the influence of differences in variable scales Huang et al. [43], using the transformation:

$$z_i= {\displaystyle \frac{x_i-\overline{x}}{S_x}}$$

(9)

z_i, x_i, S_x, and $\overline{x}$ represent, respectively, the standardized value, the original value, the standard deviation, and the mean.

2.6.1. Principal Component Analysis (PCA)

PCA was applied to group and visualize the samples based on their mechanical properties. The correlation matrix R was used, with coefficients calculated as follows:

$$r_{ij}= {\displaystyle \frac{cov\left(x_{i,}{x}_j\right)}{S_{xi}{S}_{xj}}}$$

(10)

The eigenvalues λ_k and their associated eigenvectors were used to extract the components that maximize the explained variance [33]. The selected axes followed the Kaiser criterion (λ_k) and a cumulative inertia threshold to ensure a reliable representation. Contributions, variable–axis correlations, and cos² values were used to interpret the distribution of samples, whether reference or treated, on the factorial planes.

2.6.2. Hierarchical Cluster Analysis (HCA)

To complement the analysis, an HCA was performed on the factor coordinates derived from the PCA, which enhances the robustness of the classification by eliminating redundancy among variables [32]. The similarity between two samples, j and k, was evaluated using the Euclidean distance:

$$d_{(j,k)}= \sqrt{\sum _{i=1}^n{(z_{ji}-z_{ki})}^2}$$

(11)

The resulting dendrogram allowed for the characterization of natural groupings among the samples and distinguished the effects related to hydrothermal and oil-heat treatments, as well as the reference state, according to the treatment temperature and duration.

2.6.3. Statistical Analysis and RSM

Data were processed using Microsoft Excel 10, SPSS 2.0, and Design-Expert (2021, version 13.05.0*64). A two-way ANOVA was performed to evaluate the effects of temperature and treatment duration on mechanical properties. Dunnett’s test was used to compare treated samples with reference samples. Normality and homogeneity of variances were verified using the Shapiro–Wilk and Bartlett tests at a 5% significance level, with α = 0.05 applied [3]. Following this analysis, the Response Surface Methodology (RSM) was implemented using a Central Composite Design (CCD) to optimize the combined effects of treatment temperature and duration on the mechanical properties of Dabema wood. The CCD included two factors (temperature and time) at five levels each, coded as −α, −1, 0, +1, and +α, to capture both linear, quadratic, and interaction effects. The design consisted of four factorial points, four axial points, and five center points, providing a total of 13 experimental runs per mechanical property. The selection of axial and center points ensured adequate estimation of curvature and reproducibility of the measurements. The following polynomial model was used:

Y = β₀ + β₁T + β₂t + β₁₁T² + β₂₂t² + β₁₂Tt + ε

(12)

where β₀ is the model constant, β₁ and β₂ represent the linear effects of temperature and duration, β₁₁ and β₂₂ capture the quadratic (nonlinear) effects of each factor, β₁₂ corresponds to the interaction effect between temperature and duration, and ε is the model residual error.

2.7. TG/DTG Thermogravimetric Analyses

Thermogravimetric analysis (TGA) coupled with differential thermogravimetry (DTG) was performed on Dabema (Piptadeniastrum africanum (Hook.f.) Brenan) powder samples to evaluate thermal stability and decomposition behavior. Approximately 10 mg of oven-dried, sieved powder (<100 µm) were heated from 30 °C to 900 °C at a rate of 10 °C/min under a nitrogen atmosphere (50 mL/min) using a NETZSCH STA 449 F3 analyzer. Mass loss curves and their derivatives were analyzed with Proteus software to identify characteristic degradation stages, including moisture loss, hemicelluloses depolymerization, cellulose pyrolysis, and lignin decomposition, consistent with patterns observed for lignocellulosic materials [13,28,29]. Triplicate measurements and the use of an empty crucible ensured reproducibility and baseline correction [13].

3. Results and Discussion

3.1. Colorimetric and Physical Properties of Modified Wood

Hydrothermal and oil-heat treatments induce significant colorimetric changes in Dabema wood, as evidenced by variations in L*, a*, b*, ΔE*, C*, and h* parameters (

Table 1. Characterization of physical, hygroscopic, and chromatic transformations of Dabema wood after hydrothermal and oleothermal treatments.

Code_Ech	Density g/cm3	EMC_% ± SD	VSH_% ± SD	VSW_% ±SD	WA_% ±SD	L*	a*	b*	∆E*	C*	h* (°)
Ref	0.963 ± 0.078	13.16 ± 0.43	15.80 ± 0.60	20.50 ± 0.75	78.0 ± 5.0	62.14 ± 1.24	4.96 ± 0.15	11.94 ± 0.36	-	12.929 ± 0.39	22.56 ± 0.34
Hyd100°2 h	0.928 ± 0.047	11.50 ± 0.32	14.20 ± 0.55	18.20 ± 0.70	64.0 ± 4.5	34.30 ± 0.69	−0.16 ± 0.00	1.98 ± 0.06	30	1.986 ± 0.06	−4.62 ± 0.07
Hyd100°3.5 h	0.980 ± 0.098	9.80 ± 0.30	13.50 ± 0.52	17.30 ± 0.68	59.0 ± 4.5	32.56 ± 0.65	−0.62 ± 0.02	1.08 ± 0.03	32	1.245 ± 0.04	−29.86 ± 0.45
Hyd100°5 h	1.010 ± 0.085	9.20 ± 0.28	12.80 ± 0.50	16.50 ± 0.65	54.0 ± 4.0	32.58 ± 0.65	−0.42 ± 0.01	1.72 ± 0.05	31.73	1.771 ± 0.05	−13.72 ± 0.21
Hyd160°2 h	1.020 ± 0.096	8.50 ± 0.26	11.20 ± 0.48	14.50 ± 0.60	48.0 ± 4.0	33.20 ± 0.66	−0.84 ± 0.03	1.32 ± 0.04	31.38	1.565 ± 0.05	−32.47 ± 0.49
Hyd160°3.5 h	1.060 ± 0.086	7.50 ± 0.27	10.50 ± 0.45	13.40 ± 0.55	43.0 ± 3.5	33.84 ± 0.68	−0.18 ± 0.01	3.42 ± 0.10	29.99	3.425 ± 0.10	−3.01 ± 0.05
Hyd160°5 h	1.080 ± 0.089	7.00 ± 0.25	9.80 ± 0.42	12.50 ± 0.50	39.0 ± 3.5	32.52 ± 0.65	−0.32 ± 0.01	2.70 ± 0.08	31.48	2.719 ± 0.08	−6.76 ± 0.10
Hyd220°2 h	1.120 ± 0.093	6.20 ± 0.30	8.00 ± 0.40	10.50 ± 0.48	33.0 ± 3.0	39.40 ± 0.79	2.02 ± 0.06	7.36 ± 0.22	23.38	7.632 ± 0.23	15.35 ± 0.23
Hyd220°3.5 h	1.100 ± 0.049	5.80 ± 0.28	6.50 ± 0.35	8.20 ± 0.40	29.0 ± 2.8	31.28 ± 0.63	−0.40 ± 0.01	2.38 ± 0.07	32.75	2.413 ± 0.07	−9.54 ± 0.14
Hyd220°5 h	1.130 ± 0.097	5.20 ± 0.26	5.20 ± 0.30	6.70 ± 0.35	25.0 ± 2.5	33.10 ± 0.66	1.12 ± 0.03	3.94 ± 0.12	30.37	4.096 ± 0.12	15.87 ± 0.24
Oleo100°2 h	1.050 ± 0.069	7.20 ± 0.22	10.50 ± 0.42	13.20 ± 0.55	44.0 ± 3.5	50.36 ± 1.01	12.08 ± 0.36	19.40 ± 0.58	15.65	22.854 ± 0.69	31.91 ± 0.48
Oleo100°3.5 h	1.080 ± 0.027	6.80 ± 0.22	9.80 ± 0.40	12.20 ± 0.50	40.0 ± 3.2	45.26 ± 0.91	10.54 ± 0.32	16.00 ± 0.48	18.24	19.160 ± 0.57	33.37 ± 0.50
Oleo100°5 h	1.100 ± 0.019	6.20 ± 0.20	9.20 ± 0.38	11.50 ± 0.48	37.0 ± 3.0	42.30 ± 0.85	8.38 ± 0.25	12.38 ± 0.37	20.14	14.950 ± 0.45	34.09 ± 0.51
Oleo160°2 h	1.120 ± 0.068	6.50 ± 0.25	8.00 ± 0.36	10.20 ± 0.45	34.0 ± 3.0	42.10 ± 0.84	8.04 ± 0.24	11.92 ± 0.36	20.27	14.378 ± 0.43	34.00 ± 0.51
Oleo160°3.5 h	1.140 ± 0.046	6.00 ± 0.24	7.20 ± 0.33	9.30 ± 0.42	30.0 ± 2.8	40.14 ± 0.80	6.00 ± 0.18	7.00 ± 0.21	22.57	9.220 ± 0.28	40.60 ± 0.61
Oleo160°5 h	1.180 ± 0.083	5.50 ± 0.22	6.50 ± 0.30	8.20 ± 0.38	27.0 ± 2.6	38.86 ± 0.78	7.36 ± 0.22	10.40 ± 0.31	23.45	12.741 ± 0.38	35.29 ± 0.53
Oleo220°2 h	1.160 ± 0.096	5.20 ± 0.28	5.50 ± 0.28	7.00 ± 0.35	23.0 ± 2.4	46.66 ± 0.93	11.30 ± 0.34	17.78 ± 0.53	17.71	21.067 ± 0.63	32.44 ± 0.49
Oleo220°3.5 h	1.170 ± 0.059	5.00 ± 0.26	4.80 ± 0.26	6.20 ± 0.32	20.0 ± 2.2	37.02 ± 0.74	8.62 ± 0.26	7.58 ± 0.23	25.76	11.479 ± 0.34	48.67 ± 0.73
Oleo220°5 h	1.200 ± 0.096	4.80 ± 0.24	4.20 ± 0.24	5.40 ± 0.30	17.0 ± 2.0	33.44 ± 0.67	4.30 ± 0.13	2.56 ± 0.08	30.20	5.004 ± 0.15	59.23 ± 0.89

). Lightness (L*) decreases sharply compared to the control sample, which measured 62.14. After hydrothermal treatment, L* ranges from 32.56 to 39.40, and after oil-heat treatment, from 33.44 to 50.36. According to color perception studies, variations greater than 2–3 units in L* are considered clearly perceptible to the human eye CIE 15 [37] and Hrčková et al. [36], indicating that the observed decreases correspond to visually obvious changes rather than subtle variations. This decrease indicates pronounced darkening, typical of deep thermal transformations of the lignocellulosic structure. These observations are consistent with Bessala et al. [22] and Tomak [44], who reported that heating wood in air or oil significantly reduces L*, especially at high temperatures.

For the red–green axis (a*), the responses differ according to the treatment. Hydrothermal treatment results in negative values, from −0.84 to −0.16, indicating a shift toward cool, greenish to bluish tones, particularly between 160 and 220 °C. Oil-heat treatment enhances red tones, with values ranging from 6.00 to 12.08. This intensification of warm hues is related to the decomposition of cell wall components and oxidation of extractives, as noted by Mandraveli et al. [9]. On the yellow–blue axis (b*), hydrothermal treatment markedly reduces yellow tones, from 11.94 in the control to 1.08–7.36. Oil-heat treatment maintains or slightly enhances warm hues at low temperatures, before decreasing at high temperatures. These results align with Suri et al. [18], who observed a decrease in b* at high temperatures during oil treatments of tropical woods.

Overall color changes (ΔE*) confirm the magnitude of these transformations. According to the CIE color difference scale, a ΔE* value of 1–2 corresponds to a just noticeable difference, values above 3 are clearly perceptible to the human eye, and values exceeding 10 indicate very strong and visually drastic color changes [36,37]. Hydrothermal treatment produces very high ΔE* values, from 30.37 to 32.75, which largely exceed these perceptibility thresholds and therefore correspond to drastic and immediately visible color shifts, while oil-heat treatment causes significant but more moderate changes, from 15.65 to 25.76, which remain well within the range of clearly perceptible to highly pronounced visual differences, consistent with Suri et al. [18], who reported increasing ΔE* with treatment temperature and duration.

Hue angle (h*) illustrates the chromatic divergence of the two treatments (Table 1), with low or negative values, from −32.47 to −3.01°, observed under hydrothermal treatment corresponding to dark, cool tones, while high values, from 31.91 to 59.23°, in oil-heat treatment indicate warm, saturated hues. Such large shifts in hue angle are well above reported perceptibility thresholds for wood color and result in clearly distinguishable visual appearances Mandraveli et al. [9].

These pronounced chromatic differences are closely associated with underlying chemical reactions. The observed color changes are key indicators of the chemical transformations occurring in the wood, particularly the formation of furanic resins, which influence both esthetic appearance and material properties such as water repellency and microstructural integrity [3,21]. Under hydrothermal conditions, the acidic environment generated by the hydrolysis of acetyl groups in hemicelluloses catalyzes the formation of furanic compounds through rearrangement and degradation of hemicelluloses [21]. These furanic compounds subsequently condense into dark, water-repellent furanic resins, explaining the pronounced darkening of hydrothermally treated specimens and the sharp reduction in their equilibrium moisture content (EMC) [3,18]. The formation of these resins also contributes to the development of microcracks within the cell walls due to structural rearrangements (Figure 1). In contrast, during oil-heat treatment, the presence of oil limits acetyl group hydrolysis by preventing direct water–wood contact and reducing oxygen availability [9]. As a result, furanic resin formation is strongly suppressed, with only trace amounts forming, leading to lighter coloration and more moderate chemical changes. This protective effect of oil, also reported by Zhang et al. [21] and Ninikas et al. [27], demonstrates that furanic synthesis is favored in aqueous, oxygen-rich environments, whereas oleothermal conditions effectively inhibit these reactions.

This contrast is explained by hemicelluloses degradation, reduction in hydroxyl groups, and formation of oxidation products, as indicated by Maulana et al. [8] and Mandraveli et al. [9]. Hydrothermal treatment thus cools and darkens the wood, whereas oil-heat treatment warms and visually stabilizes its color; this is consistent with the findings of Biwôlé et al. [3] on Eyong wood, confirming the structuring effect of hydrothermal treatments on color and density.

Physical and hygroscopic properties are also affected by the treatments. Wood density remains relatively stable, ranging between 0.928 and 1.200 g/cm³, indicating the absence of major structural degradation (Table 1). A slight increase in density is observed for several oil-heat treatments, with values reaching approximately 1.180–1.200 g/cm³, which may be attributed to lipid impregnation and partial filling of anatomical voids, in agreement with the observations reported by Suri et al. [18]. Equilibrium moisture content (EMC) decreases markedly relative to the control sample (13.16%), indicating reduced hygroscopicity and improved dimensional stability, with values spanning 6.2%–9.8% at 100 °C, 5.5%–8.5% at 160 °C, and 4.8%–6.2% at 220 °C. These trends align with Hsieh et al. [4] and Gao et al. [5], who showed that EMC decreases with increasing temperature and treatment duration.

Hydrothermal and oil-heat treatments also improve water stability. Shrinkage indices VSH and VSW confirm these observations. VSH drops from 15.80% in the control to 6%–8% at 220 °C under hydrothermal treatment and to 4.2%–5.5% under oil-heat treatment. VSW decreases from 20.50% to 6.7%–5.4%, indicating a significant reduction in volumetric shrinkage. Notably, oleothermal treatment demonstrates superior performance, providing the highest dimensional stability and effectively increasing hydrophobicity, thereby contributing to the long-term durability of Dabema wood [6,22,45]. These results are consistent with Bessala et al. [22], who noted the hydrophobic effect of fatty acids in oil-treated wood. Water absorption (WA) is strongly limited, decreasing from 78% in the control to 25%–64% at 220 °C under hydrothermal treatment and 17%–44% under oil-heat treatment. The oil protects the wood by limiting water penetration, as confirmed by Hill et al. [6]. Overall, thermal treatments, particularly oil-heat treatment, confer remarkable dimensional and water stability to Dabema wood. They reduce hygroscopicity, limit water absorption, and enhance durability. These physicochemical modifications are associated with the reduction in hydroxyl groups in the cell wall, contributing to the wood’s resistance and longevity under demanding environmental conditions [45].

3.2. Morphology of Wood Before and After Modification

Hydrothermal and oil-heat treatments induce pronounced alterations in the anatomical structure and mechanical performance of Dabema (Piptadeniastrum africanum (Hook.f.) Brenan) wood. According to Jančíková et al. [14], thermal treatments within the 160–250 °C range cause significant chemical and structural modifications of lignocellulosic polymers, resulting in microstructural defects observable at the cell wall level. These intrinsic changes provide a mechanistic basis for understanding the structural and functional responses observed in Dabema wood [46,47]. Figure 1 shows transverse sections of untreated wood (Ref) and wood subjected to hydrothermal (Hyd160°5 h, Hyd220°5 h) and oil-heat treatments (Oleo160°5 h, Oleo220°5 h), illustrating the direct effects of thermal modification. In untreated samples, vessels are partially occluded and poorly defined, with residual lumen contents [47]. Lighter zones surrounding the vessels, corresponding to paratracheal parenchyma, as well as fine horizontal lines of banded parenchyma, characterize the native anatomical organization and serve as a reference for evaluating treatment-induced changes [22].

Hydrothermal treatment at 160 °C for 5 h causes substantial vessel emptying, increasing lumen openness and clarity, reflecting extensive leaching of extractives and removal of parenchyma and vessel wall contents [12]. At 220 °C for 5 h, microcracks develop within the cell walls, indicating increased hydromechanical stresses and thermal degradation of hemicelluloses and the lignocellulosic matrix [16,45,47]. Paratracheal parenchyma zones become less distinct, reflecting advanced structural and chemical modifications. Similar degradation patterns have been reported for tropical woods by Ali et al. [12], while Huang et al. [48] observed a 10%–25% reduction in bending modulus in teak (Tectona grandis), accompanied by comparable microcracks. Thermal treatments are known to increase defects between cellulose microfibrils, reflecting genuine alterations in cell wall architecture [21,46].

In contrast, oil-heat treatment at 220 °C for 5 h more effectively preserves wood structure. Vessel lumens remain clearly identifiable and partially filled with oil, producing smoother lumen boundaries and reduced wall breakage [19]. Density reaches approximately 1.070 g/cm³ for both Oleo160°-5 h and Oleo220°-5 h samples. Preservation of parenchyma structures, combined with lipid impregnation, enhances wall cohesion and mechanical performance, in accordance with [22].

Water-related properties are strongly influenced by both treatments. Equilibrium moisture content (EMC) of untreated wood (13.16 ± 0.43%) gradually decreases with treatment severity, ranging from 11.50%–9.20% at 100 °C, 8.50%–7.00% at 160 °C, and 6.20%–5.20% at 220 °C for hydrothermal treatment [5,6]. Oil-heat treatment further reduces EMC, reaching 7.20%–6.20% at 100 °C, 6.50%–5.50% at 160 °C, and 5.20%–4.80% at 220 °C (Table 1), indicating reduced hygroscopicity and improved dimensional stability [7,49].

Density also increases with treatment severity. Untreated wood exhibits 0.963 ± 0.078 g/cm³, whereas hydrothermally treated samples range from 0.928 ± 0.047 to 1.130 ± 0.097 g/cm³, with the highest densities observed at 220 °C. Oil-heat treatment induces more pronounced densification, increasing from 1.050 ± 0.069 g/cm³ at 100 °C to 1.180 ± 0.083 g/cm³ at 160 °C and reaching 1.200 ± 0.096 g/cm³ at 220 °C (Table 1). These increases result from lipid impregnation and partial filling of anatomical voids, enhancing wall cohesion and mechanical performance [9,11,15].

These structural modifications directly influence water behavior. Vessels emptying under hydrothermal treatment and partial filling under oil-heat treatment reduce accessible hydroxyl groups and capillary pathways, limiting water penetration and volumetric shrinkage. Swelling indices (VSH and VSW) decrease from 15.80% and 20.50% in untreated wood to 6%–8% and 6.7% after hydrothermal treatment, and further to 4.2%–5.5% and 5.4% after oil-heat treatment. Water absorption also decreases, from 78% in untreated wood to 25%–64% and 17%–44% after hydrothermal and oil-heat treatments, respectively, in accordance with the findings of Bessala et al. [22] for Afrormosia elata and Newtonia spp. Consequently, hydrothermal treatment induces vessel emptying, partial mechanical weakening, and microcrack formation, particularly at high temperatures. Oil-heat treatment, on the other hand, stabilizes the anatomical structure, strengthens wall cohesion, increases density and mechanical strength, and significantly reduces hygroscopicity [14,15,18,45,47].

3.3. Mechanical Properties of the Modified Wood

3.3.1. Multivariate Analysis (PCA, HCA) and RSM Optimization of Dabema Wood Treatments (Piptadeniastrum africanum (Hook.f.) Brenan)

PCA applied to samples subjected to hydrothermal and oil-heat treatments clearly distinguishes their overall mechanical performance. The F1–F2 axes exhibit very high representativity, with 97.97% for compression, 98.51% for bending, and 96.97% for tensile properties (Figure 2, Figure 3, Figure 4 and Figure 5). Axis F1 primarily reflects stiffness and strength (MOE and σ), while axis F2 represents variations in elongation at break (ε) and flexibility (

Table 2. Cluster classification based on PCA for mechanical performance and recommended applications across compression, three-point bending, and tensile tests.

Cluster	PCA Mechanical Features	Experimental Codes	Mechanical Response	Potential Applications	Loading Mode (s)
C1	Very low mechanical values (σ, MOR, MOE −40 à −70%); fragile; high deformation	Ref; Hyd220°2 h; Hyd220°3.5 h; Hyd220°5 h; Hyd160°5 h	Severe thermal degradation (≥160–220 °C) → hemicelluloses loss, cell collapse, early failure	Non-structural uses: lightweight panels, decorative coverings	Compression, Three-point bending, Tensile
C2	Low strength and stiffness (MOR/MOE −40 à −60%); extreme deformation	Hyd100°2 h; Hyd100°3.5 h; Hyd100°5 h; Hyd160°2 h; Hyd160°3.5 h; Oleo100°5 h; Oleo160°5 h; Oleo220°3.5 h; Oleo220°5 h	High plasticization; reduced stiffness; fibers unable to carry maximum loads	Light structures; low-stress components; shock-absorbing parts	Compression, Three-point bending, Tensile
C3	Intermediate performance (MOR/MOE −10 à −25%); moderate ductility	Oleo100°2 h; Oleo100°3.5 h; Oleo160°2 h; Oleo160°3.5 h; Oleo220°2 h; Hyd100°5 h; Hyd100°3.5 h	Moderate treatments → limited mechanical reduction; good ductility	Furniture; semi-structural panels; moderate bending components	Compression, Bending, Tensile
C4	High performance (MOE loss < 10%–20%); σ stable or maximal; cellular consolidation	Hyd100°2 h; Hyd160°3.5 h; Hyd220°2 h; Hyd220°5 h; Oleo100°3.5 h; Oleo160°2 h; Oleo160°3.5 h; Oleo220°2 h	Consolidated microstructure; lignin–cellulose crosslinking	Frameworks; beams; structural elements; rigid components	Compression, Three-point bending, Tensile
C5	Very high MOE (10%–20% above control); maximum stiffness	Oleo100°5 h; Oleo160°5 h; Oleo220°3.5 h; Oleo220°5 h	Strong lignin–cellulose crosslinking; hygroscopic stabilization	Highly stressed bending applications; rigid technical components	Compression, Three-point bending, Tensile

). Aït-Sahalia & Xiu [43] demonstrated that PCA effectively extracts dominant factors from complex datasets, explaining the high variance captured by F1. Nadjet et al. [33] working on groundwater mineralization and pollutants, showed that F1 aggregates the major contributions while F2 isolates secondary effects. Our results follow the same logic of separation between stiffness, strength, and ductility.

Cluster 1 groups moderate oil-heat treatments between 100 and 160 °C. In bending, F-MOR and F-MOE losses remain limited to 10%–25%, with ductility preserved and microstructure intact. In tensile and compression, performance ranges from intermediate to high (Figure 6d–f). Compression displacement varies from 1.95 to 4.96 mm, with an average of 3.03 mm ± 0.97 mm and a mean strength of 2292 MPa ± 130 MPa (Table 2 and

Table 3. Mechanical performance of wood samples under compression, bending, and tensile loading with PCA cluster classification.

Loading Mode (s)	Variable	Min	Max	Mean	±SD	Mechanical Response	Associated PCA Cluster (s)
Compression	Elong_mm	1.948	4.958	3.032	0.974	Moderate deformation; ductile behavior under thermal damage	C1-C2 (high ε), C3 (moderate)
	ε_mm	0.032	0.083	0.051	0.016	Ductility is sensitive to microstructural collapse	C1-C2 (extreme deformation)
	σ_MPa	2110.94	2681.24	2291.6	129.7	Stress capacity is significantly reduced	C1-C2 (−40%–70%), C3 (moderate loss)
	MOE_MPa	26,092.67	70,836.53	49,461.36	13,540.12	Heterogeneous stiffness depending on treatment	C3 (−10%–25%), C4 (<10%–20%), C5 (10%–20%)
Three-point bending	Elong_mm	1.264	2.776	2.095	0.418	Bending deformation is moderately sensitive	C2 (high), C3-C5 (stable)
	ε_mm	0.009	0.020	0.016	0.003	Small bending strain; degradation amplifies ductility	C1-C2
	σ_MPa	53.7	83.5	67.6	10.0	MOR markedly decreases	C1-C2, C3 (moderate), C4-C5 (high MOR)
	MOE_MPa	3903.5	5751.6	4697.0	473.1	Flexural stiffness relatively stable	C3, C4, C5
	MOR_MPa	56.9	88.9	74.9	10.3	Rupture stress follows the degradation pattern	C1-C2 < C3 < C4-C5
Tensile	Elong_mm	1.561	4.667	2.451	0.700	Tensile deformation highly sensitive to microstructural damage	C1-C2
	ε_mm	0.076	0.244	0.124	0.036	High tensile ductility under degraded microstructure	C1-C2
	σ_MPa	33.99	63.44	48.55	7.76	Strongly reduced tensile resistance	C1-C2
	MOE_MPa	177.1	502.1	402.8	69.6	Significant stiffness loss in tensile	C1-C2
	MOR_MPa	36.89	90.65	56.96	12.11	High sensitivity to thermal and hydrothermal degradation	C1-C2

). Clusters C1-C2 exhibit strong strength reduction (−40% to −70%), whereas C3-C5 maintain higher stiffness (mean MOE 49,461.36 MPa ± 13,540.12 MPa). Perçin et al. [26] on poplar and beech showed that moderate treatments cause minor MOR and MOE reductions while maintaining mechanical stability. Nganko et al. [32] demonstrated that moderate temperatures enhance density, cohesion, and stiffness in thermochemically modified tropical sawdust briquettes, explaining the favorable strength–ductility combination in this cluster. Ali et al. [12] showed that lignin–cellulose crosslinking and crack reduction strengthen internal cohesion.

Cluster 2 includes prolonged hydrothermal treatments or those at lower temperatures (100–160 °C). F-MOR and F-MOE losses reach 40%–60%, while F-ε increases by 50%–120%, reflecting notable plasticization and material weakening. Bending elongation ranges from 1.26 to 2.78 mm (mean 2.10 mm ± 0.42 mm), and strength averages 67.6 MPa ± 10 MPa. Clusters C1-C2 show strongly reduced MOR (Table 3), while C3-C5 retain good stiffness (mean MOE 4697 MPa ± 473 MPa) and stable MOR (74.9 MPa ± 10.3 MPa). Broda et al. [25] showed that hydrothermal treatment strongly degrades hemicelluloses, decreasing MOR and stiffness while increasing deformability. Arriaga et al. [50] highlighted that hemicelluloses are the most water- and heat-sensitive polymers, consistent with Cluster 2 behavior.

Cluster 3 shows intermediate performance, with F-MOR and F-MOE losses limited to 10%–25% and moderate ductility (Figure 6d–f). Some hydrothermal and oil-heat treatments produce properties comparable to the highest-performing clusters (Figure 6a–c). Nakagawa et al. [51] on Western hemlock demonstrated that precise temperature control achieves a solid compromise between stiffness, hardness, and fracture energy despite moderate bending losses, which aligns with Cluster 3 outcomes. This cluster also highlights treatment substitution possibilities; for example, a hydrothermal treatment at 100 °C for 2 h can be replaced by an oil-heat treatment at 160 °C for 3.5 h, achieving similar losses with significant energy savings.

High-performance clusters 4 and 5 include the most severe oil-heat treatments, such as Oleo200°3.5 h, Oleo220°5 h, and Oleo160°5 h. Here, F-MOE remains high (90%–110%), occasionally with gains of 10%–20%, and F-MOR losses are limited, often below 10%–20%. These results indicate maximum stiffness, excellent dimensional stability, and consolidated microstructure (Table 3). Tufan & Üner [52] on Taurus cedar showed that severe treatments induce pronounced hydrophobization, strong OH group reduction, and notable mechanical stability improvement. Perçin et al. [26] indicated that relative lignin densification at high temperature strongly increases stiffness. Observed performances align with these mechanisms.

Overall, moderate oil-heat treatments improve MOR, MOE, and ductility simultaneously, while high-temperature hydrothermal treatments primarily enhance stiffness, sometimes at the expense of strength and flexibility. Prolonged low-temperature treatments and reference samples show the lowest performance, with F-MOR and F-MOE losses of 40%–60% and F-ε increasing 50%–120%. Tensile elongation ranges from 1.56 to 4.67 mm (mean 2.45 mm ± 0.70 mm), and mean strength is 48.6 MPa ± 7.76 MPa (Table 2 and Table 3). Clusters C1-C2 are highly sensitive to treatments, losing considerable strength and stiffness (mean MOE 402.8 MPa ± 69.6 MPa). Zhu et al. [53] demonstrated that PCA effectively differentiates Chinese wood species by stiffness, density, and strength, confirming that thermal treatments strongly structure property distribution in the factor space. Similarly, Ndiapi et al. [54] showed that fiber cohesion and density directly control mechanical performance in twelve Congo Basin species, explaining the superiority of oil-heat clusters, which maintain more coherent and stable microstructures. These results suggest that combining treatment type, temperature, and duration enables targeted modulation of mechanical properties, offering opportunities for energy-efficient optimization and process substitution depending on application requirements.

3.3.2. Mechanical Behavior Modeling

Quadratic regression models developed for compression, three-point bending, and tensile tests on samples subjected to hydrothermal and oil-heat treatments demonstrate high coefficients of determination (R²), ranging from 0.90 to 0.93. This indicates a strong explanatory power of the independent variables, particularly temperature (T) and time (t), which account for over 90% of the observed variability (Figure 7, Figure 8 and Figure 9,

Table 4. Estimation of regression coefficients for hydrothermal and oleothermal treatments applied to the compression test.

Treatment	Test	Coefficient	β	Df	Ss	Ms	F-Value	p-Value	CI 95%
Hydrothermal	Compression	Intercept	2043.52	1	-	-	-	-	[1971.92; 2115.13]
Hydrothermal	Compression	T-temp.	−38.53	1	8906.59	8906.59	1.67	0.2367	[−108.93; 31.88]
Hydrothermal	Compression	t-time	71.55	1	30,717.85	30,717.85	5.78	4.72 × 10−2	[1.15; 141.96]
Hydrothermal	Compression	T2	199.32	1	1.097 × 105	1.097 × 105	20.63	6.4 × 10−3	[95.55; 303.09]
Hydrothermal	Compression	t2	117.67	1	38,242.97	38,242.97	7.19	2.7 × 10−3	[13.90; 221.44]
Hydrothermal	Compression	Tt	−139.77	1	78,145.41	78,145.41	14.69	3.15 × 10−2	[−226.00; −53.55]
Hydrothermal	Compression	Model	-	5	3.486 × 105	69,718.92	13.11	1.9 × 10−3	-
Hydrothermal	Compression	Lack of fit	-	3	34,890.19	11,630.06	19.87	7.3 × 10−3	-
Hydrothermal	Compression	Residuals	-	3	37,231.90	5318.84	-	-	-
Oleothermal	Compression	Intercept	2282.08	1	-	-	-	-	[2261.32; 2302]
Oleothermal	Compression	T-temp.	34.01	1	6942.12	6942.12	15.52	5.6 × 10−3	[13.60; 54.43]
Oleothermal	Compression	t-time	−1.1883	1	0.2128	0.2128	5 × 10−4	0.9832	[−20.60; 20.23]
Oleothermal	Compression	T2	−83.30	1	19,165.43	19,165.43	42.85	3 × 10−4	[−113.39; −53.21]
Oleothermal	Compression	t2	79.43	1	17,424.37	17,424.37	38.96	4 × 10−4	[49.34; 109.52]
Oleothermal	Compression	Tt	5.65	1	127.69	127.69	0.2855	0.6097	[−19.35; 30.65]
Oleothermal	Compression	Model	-	5	33,584.55	6716.91	15.02	1.3 × 10−3	-
Oleothermal	Compression	Lack of fit	-	3	2463.94	821.31	4.93	0.0788	-
Oleothermal	Compression	Residuals	-	7	3130.74	447.25	-	-	-

,

Table 5. Estimation of regression coefficients for hydrothermal and oleothermal treatments applied to the three-point bending test.

Treatment	Test	Coefficient	β	Df	Ss	Ms	F-Value	p-Value	CI 95%
Hydrothermal	Bending	Intercept	80.81	1	-	-	-	-	[77.51; 84.11]
Hydrothermal	Bending	T-temp.	2.07	1	25.67	25.67	2.27	0.1756	[−1.18; 5.31]
Hydrothermal	Bending	t-time	3.83	1	88.01	88.01	7.79	2.69 × 10−2	[0.5842; 7.08]
Hydrothermal	Bending	T2	−9.72	1	261.08	261.08	23.09	2 × 10−3	[−14.51; −4.94]
Hydrothermal	Bending	t2	−7.51	1	155.67	155.67	13.77	7.5 × 10−3	[−12.29; −2.72]
Hydrothermal	Bending	Tt	4.21	1	70.98	70.98	6.28	4.07 × 10−2	[0.2372; 8.19]
Hydrothermal	Bending	Model	-	5	851.84	170.37	15.07	1.3 × 10−3	-
Hydrothermal	Bending	Lack of fit	-	3	78.96	26.32	601.34	<1 × 10−4	-
Hydrothermal	Bending	Residuals	-	7	79.14	11.31	-	-	-
Oleothermal	Bending	Intercept	88.76	1	-	-	-	-	[83.72; 93.80]
Oleothermal	Bending	T-temp.	−6.97	1	291.62	291.62	11.09	1.26 × 10−2	[−11.92; −2.02]
Oleothermal	Bending	t-time	2.70	1	43.85	43.85	1.67	0.2377	[−2.25; 7.65]
Oleothermal	Bending	T2	−13.61	1	511.50	511.50	19.45	3.1 × 10−3	[−20.91; −6.31]
Oleothermal	Bending	t2	−9.07	1	227.40	227.40	8.64	0.0217	[−16.37; −1.78]
Oleothermal	Bending	Tt	−4.08	1	66.75	66.75	2.54	0.1552	[−10.15; 1.98]
Oleothermal	Bending	Model	-	5	1570.52	314.10	11.94	2.6 × 10−3	-
Oleothermal	Bending	Lack of fit	-	3	184.13	61.38	1.427 × 105	<1 × 10−4	-
Oleothermal	Bending	Residuals	-	7	184.13	26.30	-	-	-

and

Table 6. Estimation of regression coefficients for hydrothermal and oleothermal treatments applied to the tensile test.

Treatment	Test	Coefficient	β	Df	Ss	Ms	F-Value	p-Value	CI 95%
Hydrothermal	Tensile	Intercept	79.88	1	-	-	-	-	[74.64; 85.12]
Hydrothermal	Tensile	T-temp.	3.44	1	70.86	70.86	2.48	0.1590	[−1.72; 8.59]
Hydrothermal	Tensile	t-time	−1.48	1	13.11	13.11	0.4597	0.5195	[−6.63; 3.68]
Hydrothermal	Tensile	T2	−25.98	1	1863.63	1863.63	65.34	1 × 10−4	[−33.58; −18.38]
Hydrothermal	Tensile	t2	−2.73	1	20.60	20.60	1.03	0.4235	[−10.33; 4.87]
Hydrothermal	Tensile	Tt	−2.71	1	29.32	29.32	14.69	0.3444	[−9.02; 3.61]
Hydrothermal	Tensile	Model	-	5	2492.04	498.41	17.47	<8 × 10−4	-
Hydrothermal	Tensile	Lack of fit	-	3	199.58	66.53	3563.26	<1 × 10−4	-
Hydrothermal	Tensile	Residuals	-	7	199.65	28.52	-	-	-
Oleothermal	Tensile	Intercept	77.10	1	-	-	-	-	[73.15; 81.06]
Oleothermal	Tensile	T-temp.	2.04	1	25.09	25.09	1.55	0.2535	[−1.84; 5.93]
Oleothermal	Tensile	t-time	−1.26	1	9.45	9.45	0.5831	0.4701	[−5.14; 2.63]
Oleothermal	Tensile	T2	−8.16	1	183.94	183.94	11.35	0.0119	[−13.89; −2.43]
Oleothermal	Tensile	t2	−16.84	1	783.32	783.32	48.33	2 × 10−4	[−22.57; −11.11]
Oleothermal	Tensile	Tt	0.8425	1	2.84	2.84	0.1752	0.6881	[−3.92; 5.60]
Oleothermal	Tensile	Model	-	5	1507.15	301.43	18.60	6 × 10−4	-
Oleothermal	Tensile	Lack of fit	-	3	106.56	35.52	20.62	6.8 × 10−3	-
Oleothermal	Tensile	Residuals	-	7	113.46	16.21	-	-	-

). A two-way ANOVA applied to both treatments reveals contrasting effects of temperature and duration depending on the mechanical loading mode (Equations (13)–(18)). For hydrothermal treatments, processing time significantly affects compression and bending, with p-values of 4.7 × 10⁻² and 2.7 × 10⁻², respectively. The 95% confidence intervals for these parameters exclude zero ([1.15; 141.96] for compression and [0.58; 7.08] for bending), confirming a positive influence of prolonged treatment (Table 4, Table 5 and Table 6). In contrast, temperature does not have a significant effect on these properties, as indicated by p-values exceeding 0.15 and confidence intervals that include zero ([−108.93; 31.88] for compression and [−1.18; 5.31] for bending). This influence relationship is reversed for oil-heat treatments, where temperature emerges as the dominant factor, with p-values of 5.6 × 10⁻³ and 1.26 × 10⁻² for compression and bending, and 95% confidence intervals that exclude zero ([13.60; 54.43] and [−11.92; −2.02]). Conversely, duration shows no significant effect (CI includes zero). Regarding tensile strength, neither temperature nor duration produces measurable effects; p-values exceed 0.15, and the 95% confidence intervals include zero, indicating that this property remains stable under the applied treatments (Table 4, Table 5 and Table 6).

These results indicate that variations in mechanical properties are closely associated with both the type of treatment applied and the mode of loading. The effects of hydrothermal treatments are primarily governed by exposure duration, which promotes hydrolysis of the wood’s polysaccharide components, as demonstrated by Ali et al. [12], along with cell wall relaxation phenomena, leading to alterations in the material’s stiffness and flexibility [48]. Despite these modifications, tensile behavior remains relatively stable, reflecting the strength and cohesion of longitudinal fiber bonding [6]. In contrast, the effects of oil-heat (oleothermal) treatments are governed primarily by temperature, through thermo-oxidative reactions that enhance internal cohesion and alter the wood’s compressive strength and overall stiffness [9,15]. These findings demonstrate that the judicious selection of treatment type, combined with precise control of thermal and exposure parameters, is critical for tailoring the mechanical properties of Dabema wood while preserving its intrinsic structural integrity.

In compression, hydrothermal treatment (R² = 0.90) exhibits a significant influence of temperature (β = −6.97, p = 1.26 × 10⁻²) as well as quadratic terms T² (β = −13.61, p = 3.1 × 10⁻³) and t² (β = −9.07, p = 2.17 × 10⁻²), with strength decreasing as treatment conditions deviate from the optimum, as reported by [3]. In contrast, oil-heat compression treatment (R² = 0.91) is primarily affected by immersion duration (β = 71.55, p = 4.72 × 10⁻²), while the temperature × time (Tt) interaction (β = −139.77, p = 3.15 × 10⁻²) highlights a strong sensitivity to extreme combinations of temperature and duration, in agreement with [11,19]. These findings indicate that each treatment type has a precise optimum of temperature and duration to maximize compressive strength, and deviations from these conditions result in a significant reduction in mechanical performance.

In three-point bending, hydrothermal treatment (R² = 0.92) is dominated by temperature (β = 34.01, p = 5.6 × 10⁻³) with notable contributions from quadratic terms T² (β = −83.30, p = 3 × 10⁻⁴) and t² (β = 79.43, p = 4 × 10⁻⁴). Oil-heat treatment (R² = 0.90) shows control by quadratic effects T² and t² (β ≈ −8.16 and −16.84, p ≤ 1.2 × 10⁻²), governing stress variation [9]. For tensile tests, hydrothermally treated samples (R² = 0.92) display strong dependence on temperature and quadratic effects, whereas oil-heat samples (R² = 0.93) demonstrate markedly higher mechanical stability [3]. This confirms the predominance of quadratic effects, the existence of a narrow optimization window for thermal and temporal parameters, and superior model coherence for oil-heat treatments.

According to Ali et al. [12], hydrothermal treatment modifies the chemical structure of wood, enhancing dimensional stability but reducing strength beyond a critical threshold. Huang et al. [48] demonstrated that combining moist heat and microwaves optimizes teak bending properties by minimizing structural degradation. Oil-heat treatment, combining thermal transfer with lipid compounds, provides a protective effect comparable to tannins described by Tomasi et al. [2], stabilizing the lignocellulosic matrix. Biwôlé et al. [3] showed that controlling pH and thermal parameters optimizes color and physical properties without compromising strength. Therefore, the combination of heat and lipids is highly effective for improving durability and mechanical performance, relying on well-established chemical and physical mechanisms, notably the stabilization of the lignocellulosic matrix and the limitation of thermo-oxidative degradation processes [20,48].

In addition, the kinetic modeling work of Fawzy et al. [55] provides further support for this interpretation. Their results demonstrate that lignocellulosic materials exhibit thermal reaction rates that are highly sensitive to the heating regime (isothermal, non-isothermal, or stepwise). Even slight deviations in temperature or exposure time can shift degradation pathways, which is consistent with the strong quadratic responses and the narrow optimization window observed in our hydrothermal and oil-heat treatments.

Y _{(Hydro.compression)} = 2043.52 − 38.53T + 71.55t + 199.32T² + 117.67t² − 139.77Tt

(13)

Y _{(Oleo.compression)} = 2282.08 − 34.01T − 1.1883t − 83.30T² + 79.43t² + 5.65Tt

(14)

Y _{(Hydr;bending)} = 80.81 + 2.07T + 3.83t − 9.72T² − 7.51t² + 4.21Tt

(15)

Y _{(Oleo.bending)} = 88.76 − 6.97T + 2.70t − 13.61T² − 9.07t² − 4.08Tt

(16)

Y _{(Hydro.tensile)} = 79.88 + 3.44T − 1.48t − 25.98T² − 2.73t² − 2.71Tt

(17)

Y _{(Oleo.tensile)} = 77.10 + 2.04T − 1.26t − 8.16T² − 16.84t² + 0.8425Tt

(18)

3.3.3. Mechanostructural Effects of Hydrothermal and Oleothermal Treatment on Mechanical Performance

Compression, three-point bending, and tensile tests reveal that hydrothermal and oil-heat treatments induce distinct mechanical responses (Table S1), closely linked to modifications in the cell wall and lignocellulosic structure. In compression, reference samples exhibit ε = 0.07 mm, σ = 2392.79 MPa, and an MOE of 38,674.38 MPa. Hydrothermal treatment produces divergent behaviors: Hyd100°2 h reduces MOE to 26,092.67 MPa and σ to 2156.04 MPa, whereas Hyd100°5 h strongly increases MOE to 46,196.71 MPa and σ to 2681.24 MPa (Figure 8, Figure 9, Figure 10, Figure 11, Figure 12 and Figure 13). This duality reflects the combined effect of hemicelluloses hydrolysis and lignocellulosic reorganization, as shown by Mandraveli et al. [9], Taghiyari et al. [11] and Lee et al. [15]. Adewopo & Patterson [16] indicate that, depending on time and temperature, hydrothermal treatment can either weaken the amorphous matrix or strengthen the cell wall via lignin condensation.

In this context, oil-heat treatment confirms a stabilizing effect. By reducing maximum deformation (ε = 0.03–0.04 mm) and increasing stiffness, it aligns with Hao et al. [17], who observed enhanced compression strength after hot oil impregnation. Oleo220°5 h reaches the highest MOE (70,836.52 MPa). These findings agree with Boonstra et al. [23], attributing improvement to relative increases in lignin and cellulose crystallinity, and with stabilizing mechanisms described by Haseli et al. [19] and Piao et al. [20]. The oil limits hygroscopicity and reinforces lignin/cellulose cohesion, contributing to stiffening and reduced deformation. These observations demonstrate that hydrothermal and oil-heat treatments mobilize different but complementary structural processes, explaining the variability in mechanical responses.

In bending, reference samples present ε = 0.02 mm, σ = 65.39 MPa, MOE = 4271.69 MPa, and MOR = 71.56 MPa. Hydrothermal treatment induces only limited modifications of the mechanical properties under moderate conditions (Table S1). At 100 °C for 2 h (Hyd100°2 h), σ decreases to 57.39 MPa, whereas extending the duration to 5 h (Hyd100°5 h) increases MOR to 80.92 MPa, indicating that treatment severity remains below the critical threshold of structural degradation. This behavior is attributed to limited hemicelluloses depolymerization and the absence of significant damage to the cellulose and lignin network. Similar observations were reported by Boonstra et al. [23] and Tjeerdsma et al. [24], who showed that mild hydrothermal conditions initiate hemicelluloses hydrolysis while promoting lignocellulosic rearrangements and increased cellulose crystallinity, thereby preserving or slightly enhancing bending performance. Perçin et al. [26] confirmed that reductions in MOR and MOE become significant only when temperature and exposure time exceed moderate levels, while Broda et al. [25] reported mainly hygroscopic changes under low-severity treatments. Mehrabi et al. [56] further emphasized that mechanical responses strongly depend on the combined effect of temperature and duration, with minimal losses or transient improvements under moderate conditions. Overall, these findings support the present results, confirming that hydrothermal treatment at 100 °C induces controlled chemical rearrangements rather than severe polymer degradation, thus maintaining or slightly improving bending performance (Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15).

Conversely, oil-heat treatment consistently increases σ and MOR. For instance, Oleo100°3.5 h exhibits σ = 83.51 MPa and MOR = 88.94 MPa, whereas Oleo220°3.5 h presents σ = 76.77 MPa and MOR = 83.63 MPa. This confirms the stabilizing action of oil on microfibrils [56]. The combined effect of temperature and time produces a maximum MOE of 5751.59 MPa, consistent with Hill et al. [6], who show that oil-heat treatment reduces microcracking from dry thermal treatments and promotes cell-wall stiffening. These mechanisms are further supported by Hao et al. [17], demonstrating improved internal cohesion and reduced early failure under bending, and by Suri et al. [18], highlighting the role of thermal-lipid interactions in mechanical stabilization. Moreover, this stabilizing trend is strongly reinforced by the Cis (Figure 16), which remains significantly narrower under oil-heat treatment, indicating greater mechanical reliability and reduced structural variability compared with hydrothermal treatment.

In tensile, reference samples show ε = 0.24 mm, σ = 38.44 MPa, MOE = 10,000 MPa, and MOR = 90.65 MPa. Hydrothermal treatment temporarily improves resistance for Hyd220°2 h (σ = 63.44 MPa, MOE = 11,900 MPa), but prolonged treatment Hyd220°3.5 h reduces σ to 35.58 MPa and MOR to 36.89 MPa (Table S1). Mandraveli et al. [9] and Ali et al. [12] report that hydrothermal treatment initially restructures the wood chemical matrix, thereby enhancing certain mechanical properties through partial hemicelluloses hydrolysis and stress relaxation within amorphous regions. This process can lead to a transient increase in stiffness and tensile strength. However, when treatment duration or temperature exceeds a critical threshold, Ali et al. [12] and Nurazzi et al. [13] show that extensive hemicelluloses degradation and cleavage of cellulose polymer chains become dominant, resulting in a collapse of load-bearing capacity and a marked reduction in tensile strength and MOR, which is consistent with the degradation observed for the Hyd220°3.5 h sample.

Oil-heat treatment stabilizes stiffness and reduces maximum deformation (ε = 0.08–0.18 mm). Oleo160°3.5 h exhibits σ = 55.00 MPa and MOR = 53.22 MPa, while Oleo220°2 h reaches σ = 58.13 MPa and MOR = 69.80 MPa. These effects align with studies showing that oil impregnation combined with heat decreases hygroscopicity, partially fills pores, and limits polysaccharide chain mobility. Hao et al. [17] and Liang et al. [57] the beneficial impact of oil-heat treatment on mechanical stability in bamboo and wood, reporting increased stiffness, reduced tensile deformation, and delayed mechanical degradation compared to wet-heat treatments due to the protective role of oil against hydrolysis. Finally, partial cell wall filling and hygromechanical stabilization, as reported by Salca et al. [49], are associated with reduced water accessibility to the cell wall polymers and improved stress transfer within the lignocellulosic matrix, which correspond with the observed values. Impregnation reduces water/matrix exchange and limits hydrolytic weakening, explaining why oil-heat-treated samples retain higher stiffness and lower deformation than prolonged hydrothermally treated samples.

3.4. TG/DTG Thermogravimetric Analyses

Combined TG and DTG analyses of Dabema wood reveal a thermal degradation sequence characteristic of dense tropical hardwoods, while also highlighting features related to its chemical composition (Figure 17). An initial mass loss of 7.86% occurs around 68 °C, with a pronounced DTG peak confirming that this stage corresponds to the removal of free and bound water, indicating low initial moisture content. At 159 °C, the TG curve exhibits a moderate inflection accompanied by a weak DTG signal, marking the onset of hemicelluloses destabilization, a common behavior in highly acetylated woods. The release of acetic acid at this stage promotes secondary reactions. Between 268 and 287 °C, hemicelluloses degradation intensifies, leading to a mass loss of approximately 15.02% and a pronounced DTG peak that reflects active polymer decomposition. This stage is associated with the release of light volatile compounds and an increase in the initial porosity of the wood. These observations are consistent with the findings of Nurazzi et al. [13], who emphasized the high thermal sensitivity of hemicelluloses and their central role in early mass losses of lignocellulosic biomaterials.

The major degradation phase occurs around 350 °C. At this temperature, the TG curve records a mass loss of 43.75%, while the DTG curve displays a sharp and intense peak, indicating rapid cellulose pyrolysis and a high proportion of crystalline cellulose, in agreement with previous reports Ninikas et al. [27] and Zhang et al. [28]. Between 380 °C and 600 °C, a gradual transition is observed. The reduced slope of the TG curve and the broad DTG shoulder indicate progressive lignin degradation. As an aromatic polymer, lignin decomposes over a wide temperature range. Isoconversional analyses conducted by El-Sayed et al. [29] showed that the variable activation energy of lignin results in slower degradation kinetics and a higher final residue. Above 600 °C, the TG and DTG curves stabilize, corresponding to the formation of thermally stable aromatic char. This char reflects moderate ash content and a highly condensed lignin structure. Studies on lignin-rich matrices by Mandraveli et al. [9] and Haseli et al. [19] confirm that thermal stability persists well above 300 °C, leading to the formation of dense and homogeneous char. These observations are fully consistent with the degradation kinetics recorded for Dabema wood.

TG/DTG profiles reported by Zhang et al. [21] provide complementary insights into the effects of thermal treatments. Moderate treatments between 120 °C and 200 °C induce only limited hemicelluloses alteration, which explains the low or transient hydrophobicity observed at these temperatures. In contrast, above 200 °C, TG/DTG curves for beech and fir exhibit strong attenuation of the DTG peak associated with hemicelluloses, indicating extensive polysaccharide degradation and the release of furanic monomers (Figure 18, Figure 19 and Figure 20). These monomers can reorganize into thermally stable cross-linked structures. The formation of these furanic networks represents the mechanism responsible for the durable hydrophobicity and reduced equilibrium moisture content observed in Dabema wood treated at high temperatures. The shift in TG curves toward increased thermal stability, also reported by Zhang et al. [21] further confirms the formation of these structures. Their development depends on exceeding a critical hemicelluloses depolymerization threshold rather than on wood species. Consequently, Dabema wood exhibits mass losses of 7.86%, 15.02%, and 43.75%, with characteristic peaks at 68 °C, 159 °C, 268–287 °C, and around 350 °C, followed by stabilization above 600 °C. These thermal signatures are fully consistent with established lignocellulosic degradation models and reflect strong polysaccharide cohesion combined with a particularly resistant lignin fraction.

4. Conclusions

This study demonstrates that hydrothermal and oil-heat treatments significantly improve the dimensional stability and water resistance of Dabema wood (Piptadeniastrum africanum (Hook.f.) Brenan), while inducing markedly different chromatic and mechanical responses depending on treatment conditions. Multivariate analysis (PCA) clearly differentiates treatment groups according to their mechanical behavior, and response surface methodology (RSM) identifies optimal processing windows that maximize mechanical performance while controlling color variations.

Hydrothermal treatment strongly enhances hygroscopic stability, reducing equilibrium moisture content by up to 64%, water absorption by 78%, and volumetric shrinkage to 6%–8%. However, at high severity, it promotes microcrack formation and mechanical losses reaching 40%–60% in bending. In contrast, oil-heat treatment provides superior structural performance, increasing density up to 1.200 g/cm³, improving compressive modulus of elasticity by 113%, and enhancing bending and tensile properties by 25%–35% and up to 130%, respectively, while maintaining greater color stability.

These contrasted behaviors are governed by distinct structural mechanisms. Hydrothermal treatment accelerates hemicelluloses degradation and furanic condensation, improving moisture resistance at the expense of mechanical integrity. Oil-heat treatment, by limiting hydrolysis and oxygen access, stabilizes the cellulose–lignin network through lipid impregnation, thereby reinforcing internal cohesion and stiffness.

Overall, oil-heat treatment is the most suitable option for structural and load-bearing applications, whereas hydrothermal treatment is preferable when hygroscopic stability is prioritized over strength. The combined PCA-RSM framework provides a robust basis for designing targeted thermal treatments, extending this approach to other tropical hardwoods, and supporting the sustainable valorization of Dabema wood in construction and furniture applications.