Ten of the 12-year-old Populus (Zhonglin 46 Populus) were procured from Nongcui Garden of Anhui Agricultural University, Hefei, China. They were sawed into 200 mm × 150 mm × 30 mm (length × width × thickness) and air-dried. Then, the timbers were cut to the dimensions of 100 mm × 20 mm × 5 mm (length × width × thickness). All the samples were defect-free and flat-sawn. Calcium chloride (CaCl₂) and sodium carbonate (Na₂CO₃) were analytically pure purchased from Sinopharm Chemical Reagent Co., Ltd. (China).

Using the water vapor protection method, a self-made high-temperature test box was used to conduct high-temperature HT on the poplar wood, and the cistern is installed as a steam generating device [26]. The modification process and treatment levels are shown in Fig. 1 and Table 1.

The CaCl₂ solution and Na₂CO₃ solution were used to prepare the CaCO₃ modified wood samples. First, untreated and heat-treated samples were impregnated in the concentration of 0.4, 0.8, 1.2, 1.6, 2.0 mol/L CaCl₂ solution in a vacuum drying oven at a temperature of 60°C and vacuum decompression filtration for 4 times, then maintained −0.09 MPa for 4 h. After the impregnation process of CaCl₂ solution, the samples were oven-dried at a temperature of 60°C for 6 h. Next, the same impregnation process using Na₂CO₃ solution was conducted. The optimized solution concentration will be selected for further study based on the investigation of impregnation efficiency, here we found is 1.2 mol/L.

Twelve specimens in each groups were oven dried at 103°C to a constant weight and the weight gain (WPG, %) of wood specimens after in-situ synthesis by CaCO₃ were calculated as:(1) WPG=M1−M0M0×100

where M₀(g) is the absolute dry mass of the specimen before impregnation treatment, and M₁(g) is the absolute dry mass of the specimen after impregnation treatment.

Before conducting mechanical properties testing, fifteen specimens in each group were subjected to humidification balance treatment in a conditioning chamber with a temperature of 20°C and relative humidity of 65%. The surface Shore D hardness (HD) test was carried out using a TH210 Shore hardness tester. A high precision double column universal mechanical testing machine (AG-X plus, Shimadzu, Japan) with a load cell of 5 kN was employed to measure the modulus of rupture (MOR) and modulus of elasticity (MOE) of the samples. According to GB/T 17657–2013, the three-point bending test was applied, the span of the samples is 50 mm, and indenter drop speed is 5 mm/min [27].

A high-speed universal pulverizer (FW-100, Beijing Zhongxing Weiye Instrument Co., Ltd., China) was employed to prepare wood powder samples. Next, thermal stability tests were accomplished using a thermogravimetric analyzer (TG 209F3, NETZSCH, Germany). For this test, three replicates of 4–6 mg wood powder underwent 60 ml/min nitrogen flow rate test conditions from room temperature (23°C) to 700°C with a heating rate of 15 K/min. The TG and DTG curves were obtained by processing the calculated data using origin.

Hierarchical cluster analysis (HCA) was performed using IBM SPSS Statistics 22.0 by the between-group linkage method [28] The values of the surface hardness (HD), the modulus of rupture (MOR), and modulus of elasticity (MOE) were taken into consideration, respectively.

Fourier transform infrared (FTIR) spectroscopy was carried out on an FTIR spectrometer (TENSOR II, BRUKER, Germany) with a wavenumber range from 4000 to 400 cm⁻¹, resolution of 8 cm⁻¹, and a frequency of 32 times.

X-ray diffraction (XRD) analysis (XD6, PERSEE, China) was employed with Cu Ka radiation (0.15406 nm) and a measurement range of 10°–80°. Segal’s empirical method was used to calculate the crystallinity of the wood cellulose [29].

Scanning electron microscopy (SEM Sirion 200, FEI, USA) was used to analyze the grain morphology of the in-situ synthesized CaCO₃. All samples were Au coated prior to examination.

Differences between samples were tested for significance by the Tukey honest significance test using origin for the values of the surface hardness (HD), the modulus of rupture (MOR), and modulus of elasticity (MOE).

Fig. 2 shows the weight gain (WPG, %) of the modified samples at different treatment levels (different impregnating solution concentration levels). The WPG of CI, CI-HT-1.5, and CI-HT-3 elevated as the impregnating solution concentration increased.

Different impregnation solution concentrations significantly affected the synthesis of CaCO₃ in poplar wood [30]. When the impregnating solution concentration was 0.4 mol/L, the weight gain value was relatively low, the WPG of CI, CI-HT-1.5, and CI-HT-3 was 5.79%, 6.60%, and 4.76%, respectively. When the impregnating solution concentration is 0.4 mol/L, WPG showed a nonlinear increase trend with the increase of the impregnating solution concentration. The WPG of CI, CI-HT-1.5, and CI-HT-3 all reached the maximum value at the impregnating solution concentration of 2.0 mol/L with 15.94%, 14.80%, and 15.30%, respectively. When the concentration of CaCl₂ precursor solution increased from 0.8 to 1.2 mol/L, the WPG had the highest rate of change, but the WPG of CI, CI-HT-1.5, and CI-HT-3 decreased successively. The possible reason is that water absorption of wood decreases with increasing heat treatment time, making it difficult to provide stable attachment sites for water molecules and CaCO₃ crystallization [31–33].

When the impregnating solution concentration increased from 1.2 to 1.6 mol/L, the WPG of CI, CI-HT-1.5, and CI-HT-3 only increased by 1.05%, 1.00%, and 2.03%. The main reason for the lower increase in WPG is that the high concentration of the impregnating solution led to the formation of CaCO₃, which is prone to agglomeration. However, adherence to the wood surface can block some ducts or grain pores, preventing further calcium carbonate production inside the wood. Therefore, as described in the literature, WPG is related to the precursor solution concentration, and the high concentration is not conducive to the formation of CaCO₃ inside the wood [34]. Consequently, the impregnating solution concentration in subsequent experiments was preferentially 1.2 mol/L.

The equilibrium moisture content (EMC), the surface hardness (HD), the modulus of rupture (MOR), and the modulus of elasticity (MOE) , and the density of the untreated, heat-treated, and CaCO₃ modified samples are shown in Figs. 3, 4 and Table 2.

The EMC of the HT-1.5 and HT-3 were reduced, as shown in Fig. 3. This is due to the fact that poplar wood undergoes a HT process in which the hemicellulose is subjected to pyrolysis and hydrolysis reactions, reducing the number of hydrophilic groups such as free hydroxyl groups [35]. In addition, the crystallinity of the cellulose and the size of the crystals in the heat-treated wood increase due to the heat, which also reduces its hygroscopicity, resulting in a lower equilibrium moisture content [35]. Compared with HT-1.5 and HT-3, the EMC of CI-HT-1.5 and CI-HT-3 increased slightly. However, they were obviously lower than that of UT. The EMC of CI reduced duo to the mineralization process, CI had the highest increment of the CaCO₃. Previous study shown that the CaCO₃ deposited in the internal void structure of poplar wood would impede the movement of moisture, thus reducing the hygroscopicity [36].

Table 2 shows that the increasing rate of density for each group is consistent with the results of the WPG. The density of the heat-treated samples decreased due to the degradation of cell wall components during the HT process [37]. The mechanical properties of UT are similar to the literature, while the MOR and MOE of heat-treated wood is lower [35]. Compared to UT, the mechanical properties of the HT-1.5 and HT-3 decreased. The HD value of HT-1.5 and HT-3 were 41.95 and 43.30, which were 12.15% and 9.32% lower than that of UT. Besides, the MOR and MOE of the HT-3 were 29.32% and 6.43% lower than in the UT samples.

Fig. 4 indicates the HD and MOR of untreated and heat-treated samples were enhanced by in-situ filling with CaCO₃. Although the impregnation process may negatively affect wood mechanical properties, the optimization of the wood structure by the filling effect of CaCO₃ compensated for this loss, except for stiffness. In other words, the in-situ synthesis of CaCO₃ filled some of the voids in wood cell walls and cell lumina [38].

The HD values of CI, CI-HT-1.5, and CI-HT-3 were 51.23, 49.65, and 50.38, respectively, improved by 7.29%, 18.36%, and 16.35%, respectively, all with highly significant differences from the UT (P < 0.01). The in-situ synthesis of CaCO₃ significantly improved the bending strength of CI (P > 0.05). The bending strength of CI-HT-1.5, and CI-HT-3 improved to same extent. The MOR values were 59.11 and 61.12 MPa. The decrease in MOE of CI, CI-HT-1.5 and CI-HT-3 is found in Table 2, compared with UT, HT-1.5, HT-3, all with highly significant differences (P < 0.01). It was indicated that CaCO₃ crystals could effectively enhance the cell wall of heat-treated wood and improve its surface hardness [34], and improve the bending strength accordingly.

Cluster analysis was done based on values of the HD, MOR, and MOE values; the dendrogram of those clusters is shown in Fig. 5. The dendrogram of HCA for various samples regarding HD shows two main clusters; the CI and CI-HT-3 were classified into the first group, which also contains UT and CI-HT-1.5. The HT-1.5 and HT-3 were classified into the second group, further proving that the HD of the heat-treated wood was improved by the in-situ filling with CaCO₃.

For MOR, the dendrogram of HCA was grouped into two blocks, CI-HT-1.5, HT-3, CI-HT-1.5, and HT-1.5 were classified into the first group. Regarding the MOE dendrogram, the CI, CI-HT-1.5, and CI-HT-3 were classified into the same group, and UT, HT-200–1.5, and HT-200–3 belonged to another group. It also verified that CaCO₃ in-situ synthesized improves the deficiency of surface hardness and mechanical properties of the heat-treated wood.

Fig. 6 represents the TG/DTG curves of the CaCO₃ modified samples. The TG curves of HT-1.5 and HT-3 samples were assembled with that of UT, so as stated in the literature, high-temperature HT cannot significantly affect the pyrolytic properties of the wood [39]. It worth noting that the thermal stability of poplar wood was enhanced after the modification of in-situ filling with CaCO₃.

The thermal behavior of poplar wood was divided into four stages (Fig. 6) [25]. The initial stage is the drying stage, which is attributed to the evaporation and desorption of residual water in the sample with the temperature below 120°C–150°C. Afterwards the unstable components of the wood, such as hemicellulose, decompose in small amounts. The cellulose and hemicellulose pyrolysis mostly transpired in the third stage with the temperature from 237.60°C to 404.23°C. The weight loss of the wood at the third stage reached 73.24%. The temperature of 361.82°C was correspondent to the maximum rate of pyrolysis. In the last stage, the temperature exceeds 404.23°C, and the wood mass stabilizes, mainly due to the decomposition of char products [40].

Compared to UT, the char residue of HT-1.5 decreased by 1.01%, and the maximum weight loss rate increased by 11.33%; there was no distinct difference in the temperature corresponding to their maximum weight loss rate. For the HT-3, the char residue increased by 3.11%, and the temperature corresponding to the maximum weight loss decreased by 12.4°C, indicating that high-temperature and long duration HT may negatively influence wood thermal stability. The DTG curves of HT-3/CI-HT-3 in the second stage were plateau, and no shoulder peak was generated (Fig. 6c). This phenomenon is relevant to hemicellulose decomposition, of which the rate of degradation increases with the prolonged HT time [40,41].

Because of the in-situ filling with CaCO₃, the char residue of CI, CI-HT-1.5, and CI-HT-3 increased by 9.09%, 9.31%, and 9.69%, respectively. Furthermore, the maximum weight loss rate decreased by 11.48%, 19.80%, and 17.79%, respectively. It could be explained that the in-situ synthesis of CaCO₃ can effectively improve the thermal stability of heat-treated poplar wood, because of the filling of the CaCO₃.

Fig. 7 demonstrates the FTIR spectrum of the CaCO₃ modified wood. The HT changes the wood structure and degrades cell wall compounds and wood extractives [42]. Besides, there is an insignificant difference between the chemical composition of the HT-1.5 and HT-3 wood (Figs. 7b and 7c) [43]. Table 3 lists the absorbance ratios of the modified poplar wood samples.

Wood is mainly composed of cellulose, hemicellulose, and lignin [44]. The peak height ratio of the characteristic absorption peaks of lignin to cellulose, hemicellulose and carbohydrates can be used to determine the relative lignin content [45]. The characteristic peak at 1510 cm⁻¹ is the aromatic skeleton of C=C vibration, which, like 1427 cm⁻¹, is the characteristic peak representing lignin in the cell wall [46–48]. Moreover, a spectral peak at 1053 cm⁻¹ corresponds to C-O deformation in cellulose 0. I1053/I1427 reflects the relative content of lignin, and it can be seen from Table 3 that its relative value does not change with increasing heat treatment duration [49]. The peak at 1374 cm⁻¹ depicts the presence of C-H deformation, aliphatic C-H stretching in phenol and methyl group of cellulose and hemicellulose [50,51]. And the peak at 1159 cm⁻¹ represents C-O-C symmetric stretching in pyranose rings, C=O stretching in aliphatic groups of cellulose and hemicellulose [51,52]. I1510/I1374 andI1510/I1159 reflect the relative content of holocellulose, there were no changes between HT-1.5 and HT-3 [49]. The peak at 898 cm⁻¹ is attributed to C-H stretching out of the plane of the aromatic ring of glucose in cellulose 0.I898/I1427 and I898/I1510 reflect the relative content of cellulose, increased with the extension of HT duration [49]. It follows that the constant wood total cellulose and the gradual increase in cellulose indicate the degradation of hemicellulose, which is similar to the results reported in the literature [53,54]. The thermal degradation becomes more intense as the treatment time increases, and the hemicellulose is interspersed between the microfibril and lignin, which acts like a bond, thus causing a reduction in the mechanical strength of the thermally modified material [55].

Fig. 7 shows a strong peak at 1738 cm⁻¹, mainly caused by the vibration of -C=O- [46]. However, this vibration is weakened after the in-situ synthesis of the sample by CaCO₃, which is perhaps relevant to the acetic acid produced by the hydroxyl group in part of the sugar group in hemicellulose hydrolyzed with the acetyl group. During the modification process, the acetic acid reacts with the alkaline impregnating solution of the modified samples, causing partial hydrolysis of the acetyl group consisting of -CH₃ and C=O. Consequently, the intensity of the peak of the acetyl group in the infrared spectrogram of the modified wood samples is weakened [30,56]. Furthermore, the characteristic peaks of calcium carbonate at 1427 and 877 cm⁻¹ were not detected in the unmodified wood and thermally modified wood samples. The peak at 877 cm⁻¹ assigned to CO₃^2- [57] indicates that CaCO₃ was deposited uniformly in the wood cell lumen.

Table 4 shows the changes in relative crystallinity. Wood cellulose showed characteristic reflections at 2θ of 16.0° (101) and 22.0° (002). The crystallinity of the heat-treated wood increased significantly with increasing HT duration, which is probably because of the of amorphous cellulose degradation [57]. Besides, pseudocrystalline regions are formed in the amorphous regions due to the re-formation of crystalline degradation products of the hemicellulose and lignin during the steam and heat process [58]; In addition, heat-induced stress release within the wood may also cause the increase in pseudo crystallinity [59,60]. Wood properties are highly dependent on the chemical compositions, the orientation of cellulose microfibrils, the molecular interactions of the cell wall polymer assembly, and changes in crystallinity [59,61]. The crystallinity of the CI, CI-HT-1.5, and CI-HT-3 samples slightly decreased after the in-situ synthesized CaCO_3. The reasons are as follows: on the one hand, the alkaline Na₂CO₃ solution degrades some of the hemicellulose and lignin, this leads to the enlargement of the void size of the amorphous region in the cell wall. The increase in the area of the amorphous region leads to a decrease in crystallinity; on the other hand, the solution can cause cellulose to swell and enter the amorphous zone, which leads to greater spacing between cellulose molecular chains, making crystallinity of mineralized samples reducing [62].

The crystal structure and physical phase of the untreated and CaCO₃ modified wood samples were determined and analyzed by XRD, as shown in Fig. 8. Compared with UT, the intensity of these two peaks decreased significantly in the treated samples of HT-1.5, while there was no significant change in the ones of HT-3. There were no other characteristic diffraction peaks observed in Fig. 8, indicating the absence of other impurities in the test samples. In addition, the sufficient crystallinity of the CaCO₃ was synthesized in the modified samples.

The diffraction peaks at 2θ = 25.1°, 39.6°, 44.0°, and 50.0° corresponded to the (110), (205), (300), and (118) planes of CaCO₃, respectively, which could be indexed to the vaterite (JCPDS 33-0268); The peaks at 2θ = 27.3° and 33.0° corresponded to the (021) and (012) planes of CaCO₃, respectively, which could be indexed to the aragonite (JCPDS 41-1475); The peaks at 2θ = 29.6°, 36.1°, 47.7°, and 48.7° corresponded to the (104), (110), (018) and (116) planes of CaCO₃, respectively, which could be indexed to the well-crystallized calcite (JCPDS 05-0586). Furthermore, the CaCO₃ crystal types in the samples of CI, CI-HT-1.5, and CI-HT-3 were approximately the same. In another word, no matter if the samples were heat-treated or not, wherein the CaCO₃ crystal is mainly in the form of calcite.

The morphologies and microstructures of the CaCO₃ modified wood were investigated by SEM. Figs. 9a–9c show that a layer of CaCO₃ particles covered the surface of CI wood. While the pore size of the CaCO₃ distribution was narrower, the diameter particles were larger and rhombic polyhedral in the heat-treated wood.

Figs. 9d and 9e are SEM images of CI-HT-1.5, the CaCO₃ particles were evenly distributed in the cell cavity of the wood fiber, and the CaCO₃ was regularly arranged. Furthermore, CaCO₃ particles were distributed in patches on the walls of the vessels, and there was also a distribution of granular CaCO₃ in the ray cells consistent with cross-sectional observations. The evenly distributed lumps of CaCO₃ can be seen on the inner the duct wall and the grain pore area on the wood rays. On top of that, the grain pores were incompletely covered by the CaCO₃ particles. This finding indicates that the CaCl₂ and Na₂CO₃ solutions can flow in the wood vessel and ray cells during the impregnation process, resulting in a more efficient deposition modification.

The SEM images of CI-HT-3 were shown in Figs. 9g and 9i. There were well-proportioned CaCO₃ particles generated in inner wall of the vessels and the inner wall of the cell wall. Some particles embedded in the interior of the cell wall and covered by wood fibers, which was due to the reaction of the precursor solution penetrated inside the cell wall to produce CaCO₃. It can be found that CaCO₃ particles were distributed in patches on the inner wall of the vessels, interestingly, with some spreading outward with the nucleus as the center. In the meantime, the primary nuclei were anchored into the cell wall region where the larger particles would grow at the expense of the smaller ones by the so-called “Ostwald ripening process” [63]. As the result, the CaCO₃ particles are mostly spherical or small clusters in the radial section, as shown in Figs. 9h and 9i. In this growth process, some constant precipitation of very fine primary nuclei of CaCO₃ occurred on the inner wall of the duct and the cell wall where the Na₂CO₃ solution was just injected.