Temperature-Dependent Creep Behavior and Quasi-Static Mechanical Properties of Heat-Treated Wood

: In this paper, Berkovich depth-sensing indentation has been used to study the effects of the temperature-dependent quasi-static mechanical properties and creep deformation of heat-treated wood at temperatures from 20 ◦ C to 180 ◦ C. The characteristics of the load–depth curve, creep strain rate, creep compliance, and creep stress exponent of heat-treated wood are evaluated. The results showed that high temperature heat treatment improved the hardness of wood cell walls and reduced the creep rate of wood cell walls. This is mainly due to the improvement of the crystallinity of the cellulose, and the recondensation and crosslinking reaction of the lignocellulose structure. The Burgers model is well ﬁtted to study the creep behavior of heat-treated wood cell walls under different temperatures.


Introduction
As a kind of complex natural polymer material, wood exhibits obvious viscoelastic properties, such as creep and stress relaxation [1][2][3]. The viscoelasticity of wood is very sensitive to changes in service temperature, moisture content and wood species [4][5][6]. Wood heat treatment can effectively improve the defects of cracking, warping deformation and easy decay of wood material [7,8]. Therefore, heat-treated wood with elegant color and dimensional stability is widely used for furniture, flooring, cladding and outdoor utilities [9], while the mechanical properties of the heat-treated wood are critical for its modification and application [10]. Most previous studies were focused on the changes of micromechanical properties of cell walls with different heat treatments that use nanoindentation [11,12]. Nanoindentation is widely used to characterize the quasi-static mechanical properties of metals and polymers at the micro scale [13][14][15]. The quasi-static mechanical properties of samples are measured by the stress-strain curves under constant load [15]. Nanoindentation is also applicable to the study of plastic deformation and dislocations of materials [16][17][18]. The creep of materials such as polymers and soft metals has been studied using data collected during the pressing process [19,20]. Nanoindentation is also applied to characterize the viscoelastic properties of biological tissue structure by extending the load-holding time [21]. In previous studies, the creep model has provided an interesting explanation for the rheological behavior of chemical-modified materials [22][23][24]. Creep is an important form of viscoelastic behavior of wood, which includes stress relaxation and dynamic viscoelasticity [25]. As a complex organic polymer, wood exhibits viscoelastic behavior due to the movement of molecular chain segments, flow within molecules, slip between molecules, crystallization and molecular orientation [26,27].
There is no published study on the micro-mechanical creep deformation of heattreated wood under different service temperatures. This study of the viscoelasticity of heat-treated wood provides the fundamental theoretical basis for heat treatment. The Larch (Larix gmelinii (Rupr.) Kuzen) is a rapid planting species in artificial forests, which is widely distributed in Northern and eastern China due to its excellent adaptability. The dimensions of the sample are 10 mm, 8 mm and 8 mm (axial, radial and tangential), respectively. The bottom and top of the samples are parallel and perpendicular to the axial section of wood cell wall. Samples were precisely prepared from the same growth ring and the late wood location of the growth ring was marked. The samples were placed in an oven for 48 h conventional drying at 103 • C. All samples were then placed in moisture-proof plastic bags for high temperature heat treatment.

Heat Treatment
After a conventional drying process, the wood specimens were treated by 1 tube furnance (Thermo Scientific, Asheville, NC, USA) at 180 ± 1 • C and 210 ± 1 • C for 6 h. Nitrogen was used as the shielding and heat conduction gas for heat treatment and the gas flow rate was 20 mL/min. All specimens were then deposited in an oven at 103 • C to await nanoindentation testing. The specific parameters of the heat treatment process were consistent with previous studies [28]. H0 refers to the oven-dry sample, and H1 and H2 are the heat-treated samples at 180 • C and 210 • C.
Quasi-static nanoindentation: The wood samples were fixed onto a sample holder. The samples were well prepared into a pyramid shape. Then, the tips of samples were smoothed using an ultramicrotome equipped with glass and diamond knifes. Preparation of the samples followed the steps represented by Meng [5].
A TriboIndenter (Hysitron Inc., Minneapolis, MN, USA), integrated with a Berkovich diamond tip and operated in open-loop control mode, was applied to measure the wood S2 layer cell wall laminae.
On the basis of nanoindentation, the micro-mechanical properties (hardness and elastic modulus) were obtained from the load-displacement curves: Hardness: P max is the peak load force and A is acquired by the projected contact area. The sample's reduced elastic modulus (Er) was obtained as follows: where dP/dh (stiffness) is the slope of the line of the unloading process. The thermal stage is used to achieve a temperature control stage for different service temperatures by means of the nanoindentation test. The three-segment load ramp of the indentation experiment was performed, including: load time for 5 s, load-holding process for 30 s and unloading process for 5 s, as seen in Figure 1. The peak load force of nanoindentation was 400 µN for all tests. An indentation depth of 150~200 nm was achieved on the wood cell wall, as seen in Figure 2. Figure 2a shows the morphology of the cell wall under scanning electron microscope, and Figure 2b-d show the SPM scanning images before and after indentation to verify the validity of the indentation data. The samples were indentation tested at 20 • C, 60 • C, 100 • C, 140 • C and 180 • C for wood specimens. The nanoindentations were performed under ambient relative humidity, which was 45 ± 2% during the experimental procedure.
images before and after indentation to verify the validity of the indentation data. The ples were indentation tested at 20 °C, 60 °C, 100 °C, 140 °C and 180 °C for wood speci The nanoindentations were performed under ambient relative humidity, which wa 2% during the experimental procedure.

Viscoelastic Properties-Creep Behavior Nanoindentation Test
The short-term viscoelastic behavior of heat-treated wood cell walls was invest under constant load. Nanoindentation equipped with a Berkovich tip was used to the effect of the temperature-dependent hardness, modulus and short-term mech creep behavior from room temperature to 180 °C. For nanoindentation, the creep rate was obtained through the following equations. The stress can be calculated as fo where d /dt is the creep strain rate, h is the indent depth, t is the creep time, is the P is the constant peak loading force and A is the indent contact area. A is the pro contact area given in our present work.

Viscoelastic Properties-Creep Behavior Nanoindentation Test
The short-term viscoelastic behavior of heat-treated wood cell walls was investigated under constant load. Nanoindentation equipped with a Berkovich tip was used to study the effect of the temperature-dependent hardness, modulus and short-term mechanical creep behavior from room temperature to 180 • C. For nanoindentation, the creep strain rate was obtained through the following equations. The stress can be calculated as follows.
where dε/dt is the creep strain rate, h is the indent depth, t is the creep time, σ is the stress, P is the constant peak loading force and A is the indent contact area. A is the projected contact area given in our present work. Based on the curves of indent depth versus time, the creep strain rate can be obtained. The creep strain rates of samples are strongly dependent on the stress. The creep data of wood cell wall can be obtained by nanoindentation test. The strain and creep compliance of wood cell wall can be calculated as follows. Since and Forests 2021, 12, 968

of 14
Then, Equation (3) can be rewritten as Before nanoindentations, the wood samples were held at the testing temperature for 10 min to attain thermal equilibrium. The thermal stage was also added to measure the creep behavior of samples in different temperature conditions (ambient condition, 60 • C 100 • C, 140 • C and 180 • C). At least 10 measurements were repeated and the average value, that could reduce the noise of the creep curves, was calculated. Burgers model The Burgers model was selected to simulate the experimental data and to investigate the heat-treated and untreated wood at different temperatures, respectively. The schematic image of the Burgers model is represented in Figure 3.

21, 12, x FOR PEER REVIEW
100 °C, 140 °C and 180 °C). At least 10 measurements were repeated a that could reduce the noise of the creep curves, was calculated. Burgers model The Burgers model was selected to simulate the experimental d the heat-treated and untreated wood at different temperatures, re matic image of the Burgers model is represented in Figure 3.

Thermal Drift
The temperature of the hot stage was monitored by a thermo sample surface to detect the real time surface temperature. Figure drift rate during the nanoindentation test. The thermal drift rate of e increases with the increase in the test temperature. After proper te attained, all the drift rates are controlled and kept at less than 1.2 thermal drift has little effect on the test results [29]. By applying the Burgers model, the creep compliance can be rewritten as follows: E B e and E B d are the spring constants, while η B 1 and η B 2 are the viscosity parameters of pure viscous damper shown in Figure 3.

Thermal Drift
The temperature of the hot stage was monitored by a thermocouple placed on the sample surface to detect the real time surface temperature. Figure 4 shows the thermal drift rate during the nanoindentation test. The thermal drift rate of each sample gradually increases with the increase in the test temperature. After proper temperature balance is attained, all the drift rates are controlled and kept at less than 1.2 nm/s. Therefore, the thermal drift has little effect on the test results [29]. Figure 5a,b show the reduced elastic modulus and hardness of the wood cell wall of heat-treated and untreated wood at different temperatures, where the error line represents the standard deviation of the samples. The elastic modulus and hardness of untreated samples increased with the increase in ambient temperature from 20 • C to 100 • C. The increasing trend is mainly caused by the reduction in the moisture content of the samples during the temperature increase [28]. The reduction in moisture content of wood caused the decrease in the mobility of cellulose molecular chains and chain segments, which contributes to the increase in elastic modulus and hardness [18]. The moisture content of the heat-treated wood is lower than that of the untreated wood, so the hardness and elastic modulus of the heat-treated wood are less affected by the ambient temperature (20 • C to 100 • C). The hardness and elastic modulus of H0 decreased noticeably with the service temperature being raised above 140 • C, which included the softening process of wood and the mild pyrolysis of wood components (especially the hemicellulose degradation) [7]. The gas environment of the sample in the nanoindentation test process caused the oxidation reaction of the extract and the pyrolysis of the wood's main components [8]. Large deviation values of H0 hardness and elastic modulus at 140 • C and 180 • C also indicate that wood cell walls undergo complex chemical reactions at this time. In other words, the nanoindentation test at 180 • C can also be viewed as an air heat treatment, while the hardness and elastic modulus of H1 and H2 samples showed small fluctuations, indicating that the high temperature (140 • C~180 • C) environment has less effect on the heat-treated samples. The heat-treated wood has better high temperature resistance, which is caused by the decrease in the hemicellulose content, cellulose recondensation and cross-linking reactions of the lignin structure [5].

Thermal Drift
The temperature of the hot stage was monitored by a thermocouple p sample surface to detect the real time surface temperature. Figure 4 show drift rate during the nanoindentation test. The thermal drift rate of each sam increases with the increase in the test temperature. After proper temperatu attained, all the drift rates are controlled and kept at less than 1.2 nm/s. T thermal drift has little effect on the test results [29].

Temperature on Creep Behaviors
Generally, wood creep occurs under compressive stress, which is related to the variation of wood species, mechanical properties, stress level, temperature, moisture content and other factors [1,4,28]. The parameters of the creep behavior of the cell walls were obtained from the data recorded in the nanoindentation peak force-holding process. The displacement-time curve of untreated wood at different service temperatures is shown in Figure 6. The displacement increased rapidly at the beginning of loading and then slowed down gradually. This creep behavior consists of instantaneous elastic deformation, viscoelastic deformation and viscoelastic degeneration [19,20]. The maximum creep displacement of H0 occurs at 20 • C. For the H0 nanoindentation test at 180 • C, the creep displacement increased significantly, and the curve fluctuated noticeably; this is mainly caused by the degradation of hemicellulose and the softening of lignin [16]. For the high temperature nanoindentation test, the movement of the cellulose molecular chain was activated, the free volume gradually increased, and the intermolecular chain began to slip [30]. Hydrogen bonds between hemicellulose and cellulose in the amorphous region were broken, which also promoted the creep displacement [16,28]. while the hardness and elastic modulus of H1 and H2 samples showed sm indicating that the high temperature (140 °C~180 °C) environment has le heat-treated samples. The heat-treated wood has better high tempera which is caused by the decrease in the hemicellulose content, cellulose and cross-linking reactions of the lignin structure [5].  the degradation of hemicellulose and the softening of lignin [16]. For the h nanoindentation test, the movement of the cellulose molecular chain w free volume gradually increased, and the intermolecular chain began to gen bonds between hemicellulose and cellulose in the amorphous regi which also promoted the creep displacement [16,28].

Heat Treatment on Creep Behaviors
According to the microstructure of wood, the cell wall can be seen a cellulosic composites [13]. The nanoindentation is mainly applied to th cell wall, which is the thickest part of the cell wall [1,5]. Figure 7 shows t ance for different test temperature conditions (20 °C, 60 °C, 100 °C, 140 ° H0, H1 and H2. Comparing the H0, H1 and H2 curves at the same test t clear that the heat treatment completely changes the wood cell wall cr Figure 7, the creep compliance of the cell wall is decreased after heat tr temperature (20 °C). In other words, heat treatment can effectively redu havior of the wood cell walls at 20 °C. H2 exhibits minimal creep compl temperature to 100 °C, while it represents the maximum creep complian creep of H2 sample increases rapidly when ambient temperature exceed tion, the H0 and H1 curves overlap when the ambient temperature reach is most likely caused by hemicellulose Softening. As a complex physical m cess, heat treatment includes the degradation and modification of hemic degradation and crystallization of amorphous cellulose and the conden zation of lignin [31,32]. After heat treatment, the reinforced wood cell w creep, which was closely correlated with the tighter bond of the wood ce

Heat Treatment on Creep Behaviors
According to the microstructure of wood, the cell wall can be seen as complex lignocellulosic composites [13]. The nanoindentation is mainly applied to the S2 layer of the cell wall, which is the thickest part of the cell wall [1,5]. Figure 7 shows the creep compliance for different test temperature conditions (20 • C, 60 • C, 100 • C, 140 • C and 180 • C) of H0, H1 and H2. Comparing the H0, H1 and H2 curves at the same test temperature, it is clear that the heat treatment completely changes the wood cell wall creep behavior. In Figure 7, the creep compliance of the cell wall is decreased after heat treatment at room temperature (20 • C). In other words, heat treatment can effectively reduce the creep behavior of the wood cell walls at 20 • C. H2 exhibits minimal creep compliance from room temperature to 100 • C, while it represents the maximum creep compliance at 180 • C. The creep of H2 sample increases rapidly when ambient temperature exceeds 100 • C. In addition, the H0 and H1 curves overlap when the ambient temperature reaches 140 • C, which is most likely caused by hemicellulose Softening. As a complex physical modification process, heat treatment includes the degradation and modification of hemicellulose [30], the degradation and crystallization of amorphous cellulose and the condensation polymerization of lignin [31,32]. After heat treatment, the reinforced wood cell walls showed less creep, which was closely correlated with the tighter bond of the wood cell wall.

Burgers Model
Nanoindentation is widely applied in the microscopic creep deformation of materials [33,34]. This study shows that nanoindentation technology can effectively test the heat treatment of wood cell wall creep behavior. The experimental data are simulated using the four component Burgers model. Figures 8 and 9 show the experimental data for wood cell wall creep compliance and the red line is creep compliance simulated by the Burgers model. The creep simulated curves are in good agreement with the test data, with a goodness of fit of 0.99. The Burgers model is suitable for forecasting the creep behavior of wood cell walls. According to nanoindentation test data, the creep parameters of wood cell walls are obtained based on Burgers model fitting. The parameters of the Burgers model include E1, E2, η1, η2 and τ0: elastic modulus, viscoelastic modulus, plasticity coefficient, viscoelasticity and relaxation time. These parameters are listed in Table 1.

Burgers Model
Nanoindentation is widely applied in the microscopic creep deformation of materials [33,34]. This study shows that nanoindentation technology can effectively test the heat treatment of wood cell wall creep behavior. The experimental data are simulated using the four component Burgers model. Figures 8 and 9 show the experimental data for wood cell wall creep compliance and the red line is creep compliance simulated by the Burgers model. The creep simulated curves are in good agreement with the test data, with a good-   Forests 2021, 12, x FOR PEER REVIEW Figure 9. The test data and simulated curves of H2 heat-treated wood cell wall.   The creep deformation of wood is greatly affected by heat treatment, environmental temperature and humidity [5,35]. At room temperature, the Burgers models E1, E2, η1 and η2 of the wood cell walls increased gradually with the increase in heat treatment intensity, as seen in Table 1. This demonstrates that heat treatment can reduce wood cell wall creep behavior at room temperature, which is consistent with results obtained using the quasistatic nanoindentation test [18,28]. When the service temperature increased (60~140 • C), the elastic modulus, viscoelastic modulus, plastic coefficient and viscoelasticity of the untreated specimen showed a significant increase, which is related to the softening of the wood cell wall. The creep parameters of the Burgers model of the untreated sample were significantly reduced and there was a significant fluctuation in the creep curve at 180 • C, which was caused by real-time softening and pyrolysis of the wood components during the 180 • C nanoindentation test. Compared with the untreated sample, the creep curve of H2 is relatively stable at 180 • C. The results showed that the Burgers model was well fitted to the creep curves of untreated and heat-treated Larix cell wall. The difference between the fitting data of the Burgers model and the experimental data is accurately controlled.

Conclusions
After reasonable temperature balance treatment, the nanoindentation overcomes the limitations of the inherent room temperature test and can be used to test the microscopic creep behavior under high temperature conditions. The hardness of the wood cell walls significantly increased and the creep rate decreased noticeably after heat treatment. The improvement in the micromechanical properties of wood after heat treatment is mainly due to the recondensation and cross-linking reaction of the lignocellulosic structure and the increase in the crystallinity of cellulose.
It has been verified that the parameters of creep behavior fitted by the Burgers model are sensitive to changes in ambient temperature. The Burgers model can adequately fit the creep of wood cell walls under temperature response.
The nanoindentation test during the peak stress holding stage also demonstrated the softening of wood components at high temperature. It was also proven that the Burgers model is appropriate to study the viscoelastic creep behavior of wood cell walls.