Water Absorption Properties in Transverse Direction of Heat-Treated Chinese Fir Wood Determined Using TD-NMR

Abstract

Heat treatment is an environmentally friendly method that improves the moisture-resistant properties and increases the service life of timber. In this work, Chinese fir (Cunninghamia lanceolata [Lamb.] Hook.) wood was heat-treated in a chamber with steam at temperatures of 160, 180, 200 and 220 °C for 2 h, and the absorption of water was studied by gravimetric analysis and time domain nuclear magnetic resonance (TD-NMR). The results show that both the amount of bound water and free water decreased with the increasing treatment temperature. The water absorption of wood in the radial direction was faster than that in the tangential direction due to the existence of rays, and this difference remained after the heat treatment. The heat treatment at 220 °C had a significant effect on water absorption in the tangential direction of wood, and the moisture content (MC) was approximately 20% lower than that of samples absorbing water in the radial direction. T₂ (spin–spin relaxation time) distributions showed two main components which were associated with bound water and free water, and for samples absorbing water in the radial or tangential direction, there was only a difference in the amount of free water. The amount of free water significantly decreased for the samples that were heat-treated at 220 °C and absorbed water in the tangential direction, indicating that the high-temperature heat-treated samples tended to close the pits in wood cells.

1. Introduction

Wood is a naturally hydrophilic material and its main components are cellulose, hemicellulose and lignin matrix. The existence of hygroscopic groups makes it sensitive to moist conditions and its mechanical properties as well as dimensional stability are deeply influenced by the presence of water [1]. Furthermore, if it is exposed to water, wood is more prone to be decayed or worn and the growth of mold and fungi might be enhanced. Various methods have been developed to make wood a biologically durable material as well as to decrease water absorption and increase the dimensional stability of wood [2,3]. Among these methods, heat treatment is an important approach which can change the relationship between wood and moisture without introducing chemicals, thereby effectively improving the properties of wood.

Heat treatment is a method of exposing wood to an almost oxygen-free environment under high temperature conditions and causing changes in the chemical structure of the wood substance [4,5]. Hemicellulose first degrades during the heat treatment process, including acid hydrolysis and decarboxylation reactions [6]. Lignin is generally considered to be the most thermally stable component; however, with the production of various phenol decomposition products, lignin undergoes thermal degradation at relatively low temperatures [7]. In addition, high temperature will increase the crystallinity of cellulose [8]. Therefore, the amount of hydroxyl sites might be reduced and the pore structure of wood would be changed after heat treatment. All these changes might affect the state and the amount of water adsorbed by wood when being stored in a humid environment or immersed in water; consequently, the water absorption performance of wood would be influenced.

A commonly used method for assessing the water absorption performance of wood is that of calculating the MC after water immersion for a specific time. The traditional method to determine MC is the oven drying method. This method, however, cannot accurately obtain the amounts of different water states (free water, FW; bound water, BW) separately. The NMR technique can be used to investigate the internal structure of wood as well as water state and water content, which makes up for the lack of test methods for characterizing the water absorption performance of timber. Thus, NMR is being more and more widely used in many fields of wood science research for quantitively exploring the characteristics of wood in a quick and noninvasive manner. The NMR signals obtained by the free induction decay (FID) experiments were proven to be a feasible means of calculating the MC of wood [9,10] and the NMR signals of wood substance and water in wood decay at different times could be easily distinguished [11,12]. Furthermore, the peak value of the Carr–Purcell–Meiboom–Gill (CPMG) decay curve was used for fitting the wood MC, and the results indicate that there is a good linear relationship between the signal and MC [13]. Since the mobility and local environment of molecules are connected to the NMR signal decay rate, the water state can be distinguished by the water relaxation characteristics for both hardwood and softwood [14,15]. In addition, the integral peak areas of T₂ curves acquired using the CPMG pulse sequence and the inverse algorithm also have a good linear relationship with the wood MC and could have a good application in calculating the amount of BW and FW.

For wood with large dimensions, water absorption mainly occurs in the transverse direction. In order to see the difference in water absorption in the radial direction and tangential direction and the effect of heat treatment on it, a comparison of the water absorption properties of Chinese fir wood was carried out by the gravimetric analysis and time domain nuclear magnetic resonance (TD-NMR) technique in this study. The wood samples under investigation were immersed in water and weighed as a function of time of water immersion to observe the water absorption progress. The statistical analysis of changes in the average MC of the wood obtained by gravimetric method was conducted and then parts of the samples were selected to perform the TD-NMR experiment. To verify the feasibility of the TD-NMR experiment, the integral peak area of the T₂ curves were firstly used to fit the MC. Thereafter, water states as well as the change of amounts for BW and FW were analyzed according to the T₂ curves.

2. Materials and Methods

2.1. Materials

In this study, the heartwood of Chinese fir (Cunninghamia lanceolata [Lamb.] Hook.) was studied. The dimension of the specimen used for heat treatment was 150 (longitudinal, L) × 15 (radial, R) × 15 mm³ (tangential, T). The heat treatments were processed at 160, 180, 200 and 220 °C for 2 h in an airtight chamber. Throughout the heat treatment process, the oxygen content in the thermal treatment chamber was controlled to be no more than 2%. Before the heat treatment, the sample was stored under the air-dry condition, and the MC of the specimens was 9.8%–12.5%. After the heat treatment process, the specimens were cooled to room temperature in the chamber, and the final moisture content was in the range of 3.7%–7.6%. The number of the untreated references was 30 and 50 for each heat treatment group. After heat treatment, the references and each group of the heat-treated specimens were randomly divided into two equal parts: sealed tangential sections and cross-sections or radial sections and cross-sections by epoxy resin, respectively, using the method similar to that described in the previous study [16] and then conduct the experiment of the water absorption in the radial and tangential directions.

2.2. TD-NMR Experiments

The samples used for the TD-NMR experiments were cut in the middle of the heat-treated specimens, as described in their references above, in dimensions of 10(L) × 15(R) × 15(T) mm³. The NMR instrument used in this study was MesoMR23-060H-I (Niumag Co., Ltd., Suzhou, China) with a 0.52 Tesla permanent magnet (23 MHz proton resonance frequency) operating at 32.00 °C. The spectra were acquired using the CPMG pulse sequence. The number of echoes was 15,000 with 64 scans and the echo time was 0.15 ms. More details were the same as described in the previous study [16]. The simultaneous iterative reconstruction technique (SIRT) algorithm was used to achieve the T₂ distribution curve and the amount of BW and FW was calculated based on Equation (1):

$${\mathrm{MC}}_i=\frac{\left({\mathrm{Area}}_i/{\displaystyle \sum \mathrm{Area}}\right)\times \mathrm{WM}}{\mathrm{OW}}\times 100(\%)$$

(1)

where MC_i is the amount of BW or FW; Area_i is the integral area of BW peak or FW peak; ΣArea is the total integral area of the T₂ curve; WM is the water mass of the sample at each water absorption stage which can be calculated on a dry basis; and OW is the oven dry weight of the sample.

In the following analysis, the heat-treated samples are referred to as HT/X, where X represents the heat treatment temperature in degree Celsius, and when necessary, to distinguish the samples of water absorption in the radial direction (R direction) or the tangential direction (T direction), the prefixes R- or T- will be added to indicate the R direction or T direction. For instance, T-HT/220 indicates that the sample was heat-treated at 220 °C and absorbed water in the T direction.

3. Results and Discussion

3.1. MC Change during Water Absorption Process

Measured as a function of water immersion time and analyzed using the oven-drying method, the average MCs (with standard deviations) of the untreated and heat-treated samples are shown in Figure 1a,b. Within the first 30 days, the water absorption of the heat-treated samples was slower than that of the references in both the R direction and T direction. The rate of water absorption was decreased with the increased heat-treated temperature. While compared to the untreated samples, the water absorption rate was gradually increased due to the milder modification of HT/160 and HT/180, both in the R direction and T direction. After immersion for 30 days, the MC change of HT/160 was similar to the untreated samples and the absorption rate was higher than the untreated ones after water immersion for 45 days. Since the hemicellulose in the wood had already begun to degrade when being treated at 160 degrees, more pores might be formed in the wood for holding moisture, which was a possible reason for the higher MC after being immersed in water for long enough, and similar results have been reported in previous research [17]. The absorption rate of HT/180 was also close to the untreated sample after immersion for 45 days. Due to the stronger treatment, the rate of water absorption for HT/200 was lower than their untreated reference throughout the testing process, and the gap was widened along with the immersion time. Heat treatment at 220 °C had the strongest effect on water absorption, especially for samples absorbed water in the T direction, and the MC of the T-HT/220 was approximately 70% of its untreated reference after water absorption for 180 days.

The decrease in water absorption due to heat treatment within the first 30 days was mainly considered the result of changes in wettability. An increase was observed in the contact angle while the wetting tension was decreased for the heat-treated wood, both indicating that the wood samples were becoming more hydrophobic [18]. Therefore, it was difficult for the water to spread on the wood surface and penetrate the wood sample. However, according to the previous studies, the permeability of wood might increase due to milder heat treatment [19], and owing to the degradation of the courts, cracks appeared along the middle lamella and intercellular cavities [20]. Thus, the internal resistance of water transport was weaker compared with the untreated samples. This might be one of the reasons for the increases in the water absorption rate of HT/160 and HT/180 after prolonged immersion time. However, the degradation of the wood constituents becomes more severe along with the increase in treated temperature, and the micropores would be filled with the degradation products [21]. In addition, many pits were closed because of the heat treatment above 200 °C [14]. These alterations should be the most important reason for the reduction in water absorption of HT/200 and HT/220.

Figure 2 shows the MCs’ comparison of water absorption in the R direction and T direction for the samples which were under the same treated temperature as well as the references. As shown in Figure 2a, the water uptakes of R-control and T-control were almost the same after 180 days of water immersion, but the former was slightly higher than the latter. However, the curve revealed that the difference in MC between the two groups of samples was relatively obvious before the 150th day, revealing that the difference gradually decreased. This might indicate that there was a difference in water absorption rates between the R direction and T direction, but it should not affect the water absorption when upon reaching saturation. Although the water absorption of heat-treated samples was decreased compared to the references, no difference of the effect was founded between the R direction and T direction at temperatures below 200 °C. As shown in Figure 2b,c, the decrease in water absorption after the heat treatment was basically the same, and the water absorption was still slightly higher in the R direction. During the heat treatment progress, both the extractives and the products of thermal degradation would move towards the surface of the wood or evaporate from the wood [22]. These substances might partially remain in the surface layer of the wood and in consequence, impeding moisture from penetrating the wood, thereby decreasing the rate of water absorption. After heat-treatment to 220 °C, the water absorption of wood in the T direction significantly decreased and the MC of T-HT/220 was approximately 20% lower than R-HT/220. The possible explanation was the thermoplastic flow of the torus could bring about a filling of the courts after the heat-treated temperature above 200 °C [19], which would block the migration of moisture inside the wood. However, compared with the T direction, rays which exist in the R direction might play an important role during the water uptake process and result in higher water absorption.

3.2. Determination of the Water States

To evaluate the feasibility of the MC analysis of the wood sample during water absorption process using TD-NMR, the peak area of the T₂ relaxation curves was calculated using the inverse algorithm after conducting each CPMG experiment and then fitted with the MC obtained by the oven-drying method. The response and linear fitting curve of the MC to the peak area are shown in Figure 3a,b. The regression equations and the correlation coefficients (R²) are listed in

Table 1. The regression equations obtained by linear fitting of peak area and MC for samples with water absorption in the R direction.

Samples	Regression Equations	Adj.R2
R-Control	Y = 583.72x + 2900.81	0.997
R-HT/160	Y = 563.53x + 2386.21	0.996
R-HT/180	Y = 542.72x + 2635.67	0.995
R-HT/200	Y = 518.51x + 1751.49	0.997
R-HT/220	Y = 520.24x + 1520.01	0.998

and

Table 2. The regression equations obtained by linear fitting of peak area and MC for samples with water absorption in the T direction.

Samples	Regression Equations	Adj.R2
T-Control	Y = 554.04x + 2263.00	0.998
T-HT/160	Y = 571.94x + 2493.09	0.998
T-HT/180	Y = 482.26x + 2466.21	0.997
T-HT/200	Y = 515.90x + 1849.79	0.997
T-HT/220	Y = 531.70x + 1363.97	0.999

. According to the fitting curves and the correlation coefficients, the peak area and the MC has a good linear relationship and the R² was higher than 0.99 for all the samples, indicating that the peak areas obtained from the T₂ curves are proportional to the water amounts.

Figure 4 shows the water states of the wood samples determined by TD-NMR which were acquired from a single representative sample. The wood–water system has been widely researched using TD-NMR, and the literature has provided a good summary of relationships between the T₂ value and the states of water in softwood [23]. The T₂ of the BW was a few milliseconds and it was tens of milliseconds or more for FW according to the research results. The anatomical structure of softwood was relatively uncomplicated and only two obvious peaks were found in the water relaxation distribution curves which correspond to different water components, as shown in Figure 4. For all samples, the T₂ of BW was approximately 1 ms, and there was no obvious difference between the heat-treated samples and the untreated references. However, compared with the untreated references, the relaxation time of FW for the heat-treated wood significantly changed, and the change became more obvious with the increase in the heat-treated temperature. According to the relaxation theory of free fluid, the relaxation behavior of water inside wood was mainly relevant to its pore diameter and surface properties [24]. Heat treatment enhanced the surface hydrophobicity of wood due to changes in its chemical composition [25] which might be an important reason of the T₂ of FW becoming greater for the heat-treated wood. Another possible explanation for this was that the wood porosity and average pore diameters might be increased after the heat treatment according to the previous literature [26].

In addition, as for HT/200 and HT/220, a peak with T₂ of approximately tens of milliseconds appeared on the water relaxation distribution curve, as shown in Figure 4, and which was not found in untreated samples and samples heat-treated below 200 °C. The peak was not obvious on the curve for HT/200; however, it became larger for HT/220 and could be easily observed on the curve. As previously mentioned, the heat treatment might affect the pore structure and therefore affect the relaxation behavior of water inside the wood. Consequently, the above phenomenon indicates that a heat treatment temperature over 200 °C might have a serious effect on the pore structure of wood. According to a previous study [27], a peak with T₂ greater than 100 ms corresponds to water in the lumen of wood cells; by inference, the peak with T₂ of approximately tens of milliseconds might correspond to water in the pores with a diameter less than the cell lumen. During the heat treatment, the cell wall was compressed, the cell lumens became narrow and intercellular cracks may have appeared in the middle lamella [28]. This might be an interpretation of the appearance of a little peak on the water relaxation curve for HT/200 and HT/220.

3.3. Amount of Bound Water and Free Water

Over the past decade, the FID signal according to the NMR experiment has been used to determine the MC of wood [29]. The CPMG experiment was employed to investigate the MC of a wood sample also based on the peak area of a T₂ relaxation curve [30]. In this study, both the amounts of BW and FW were calculated using MCs acquired by the oven-drying method and the peak area of the T₂ curves. The change in the amount of BW and FW along with the water immersion time are illustrated in Figure 5. For a more direct comparison, the values of the BW content and FW content after 60 days of water immersion are listed in

Table 3. The content of BW and FW of the wood samples after 60 days of water absorption in the R direction.

Samples	R-Control	R-HT/160	R-HT/180	R-HT/200	R-HT/220
BW Content	28.92%	27.84%	27.43%	24.64%	19.83%
FW Content	81.00%	78.10%	74.20%	67.56%	60.17%

and

Table 4. The content of BW and FW of the wood samples after 60 days of water absorption in the T direction.

Samples	T-Control	T-HT/160	T-HT/180	T-HT/200	T-HT/220
BW Content	28.93%	28.32%	27.23%	24.21%	19.33%
FW Content	77.20%	74.12%	69.25%	62.20%	43.23%

.

According to the data in Table 3 and Table 4, no significant difference of the BW content was found between HT/160 and HT/180, but the value was slightly lower compared with the references. After the treatment temperature reached 200 °C, the BW content obviously decreased, and when the temperature rose to 220 °C, the value dropped by more than 30% compared to the untreated samples. As we all know, changes in wood chemistry and mass loss will occur after heat treatment [31]. The opinion of most researchers is that the hygroscopicity change of heat-treated wood is related to the reduction in available hydroxyl groups in the cell wall because of the degradation of chemicals [32]. At lower temperatures, wood chemical composition, especially that of cellulose and lignin, is less affected. Thus, the BW content is only slightly reduced for HT/160 and HT/180. At higher temperatures, the rearrangement or reorientation of cellulose molecules inside quasicrystalline amorphous regions and the crosslinking condensation reaction of lignin occur, whilst the wood chemical composition as well as wood pore structure are changed. The increase in matrix stiffness and lignin cross-linking would decrease the capacity of the expansion of the cell wall and lead to the reduction in the polylayer sorption [33]; as a result, the BW content markedly decreases as the heat treatment temperature rises.

The amount of FW in the heat-treated samples and untreated reference were measured as a function of the time of water immersion and the results are shown in Figure 5b,d. The values increased with prolonging water immersion time, but did not reach a saturation value after immersion in water for 60 days. For untreated samples, the absorption of FW was slightly faster in the R direction than that in the T direction, which can be inferred from Table 3 and Table 4. This may be caused by rays, which are oriented in the R direction. After the heat treatment, the difference in water transport speed between the R direction and the T direction still existed, and the gap became more obvious after heat treatment at 220 °C. The amount of FW for T-HT/220 was approximately 28% lower than R-HT/220 after water immersion for 60 days. Based on the above analysis, after 60 days of immersion, water absorption in the R direction or the T direction almost did not influence the BW content for the untreated samples or samples treated at the same temperatures, and the difference in MC was mainly caused due to the different amounts of FW. This might indicate that the heat treatment changes the wood chemical composition and pore structure, which affects the migration of liquid water but has no influence on the diffusion of water vapor inside it.

4. Conclusions

In this study, the water absorption performance of the heat-treated wood at different temperatures, i.e., 160, 180, 200 and 220 °C, as well as that of untreated samples was investigated as a function of the water immersion time. The MC analysis by the oven-drying method showed that after the heat treatment below 200 °C, the rate of water absorption slightly decreased, while higher treatment temperatures (200 °C and 220 °C) had an obvious effect on the water absorption rate. The comparison of MCs revealed that the water absorption in the R direction was faster than that in the T direction, and the difference still existed after the heat treatment. It is worth mentioning that the 220 °C-heat treatment had a significant effect on water absorption in the T direction, and the MC was significantly lower than that in the R direction (R-HT/220) after the same time of water immersion. T₂ distributions showed two main components corresponding to FW and BW, both in the case of the heat-treated samples and the references. Whereas there was also a component with T₂ of tens of milliseconds which belonged to FW for HT/200 and HT/220, the water amount was small. This might be because of changes in the pore structure caused by the chemical degradation of wood. For the heat-treated samples, the analysis of MCs obtained by the oven-drying method and the integral peak area of the T₂ curves showed that both the content of FW and BW decreased, which became more obvious after heat treatment at higher temperature. For the samples which were under the same heat-treatment conditions but in different water absorbing directions, the difference in MC was mainly due to the different amount of FW. Moreover, for the T-HT/220 samples, the significant reduction in FW content might be because the closing of pits reduces the access of water to the interior of wood.