Archaeological

NIR spectroscopic monitoring of water adsorption/desorption process in modern and archaeological wood Tetsuya Inagaki1, Hitoshi Yonenobu2 and Satoru Tsuchikawa1 1 Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya 464-8601, Japan 2 Naruto University of Education, Naruto 772-8502, Japan Abstract We investigated the adsorption/desorption mechanism of water adsorption for modern and archaeological wood using near-infrared (NIR) spectroscopy. A mixture model of water was used to decompose the nearinfrared difference spectra into three components (free water molecules (S0), those with one OH group engaged in hydrogen bonding (S1) and those with two OH groups engaged (S2)). The variations of each water component with relative humidity could be explained by proposing a model that describes water absorption in three stages. It was concluded that the ageing phenomenon in wood is due to the decrease of adsorption sites on hemicellulose and amorphous cellulose.1 Introduction The physicochemical condition of the hydroxyl groups plays a key role in water adsorption/desorption process in wood. However, lacking an effective analytical technique, the behavior of water in wood is not well understood at a molecular level. Near-infrared (NIR) spectroscopy is useful in all facets of material assessment as a means of nondestructive measurement for biological materials (e.g. forest and agricultural products, textiles and so on).2 NIR spectroscopic information on biological materials is mostly related to water, which has specific absorption bands at 5200cm-1 (combination of stretching and deformation vibrations for OH) and 6900cm-1 (first overtone of the OH stretching vibration). 3 In this study, we report on the changes of water condition of modern and archaeological wood as inferred from NIR spectra. The objective was to investigate the water adsorption mechanism in wood using a structure model of water. NIR spectra were decomposed to three components of water. The temporal change in water adsorption mechanism was examined by comparing the analytical results between modern and archaeological wood. Experimental Section We used modern and archaeological wood samples of Hinoki cypress (Chamaecyparis obtusa). The modern wood sample was taken from a living tree. The archaeological sample was collected from an upright pillar of a historical building in Japan of which construction date was estimated to be the early 7th century by the documentary analysis. Archaeological wood and modern wood plates were 10×20×0.5 mm and 50×50×2 mm in tangential, radial and longitudinal directions of the samples, respectively. The wood plates were humidified gradually from an oven-dried to the fiber saturation state in a sealed desiccator, in which the internal temperature was maintained at 20 °C. Subsequently, the plates were dehumidified to an oven-dried state. Diffuse reflection NIR spectra were measured on a FT-NIR spectrophotometer (Bruker MATRIXF) with a fiber optic probe. 128 scans were signal-averaged at a spectral resolution of 8 cm-1. Result and Discussion Decomposition of NIR spectra Variation of the NIR spectra with water adsorption onto wood was analyzed based on the structural models for water molecules, which were classified into three components, free water molecules (S0), molecules with one OH group engaged in hydrogen bonding (S1) and molecules with two OH groups engaged in hydrogen bonding (S2).4 Curve fitting was undertaken to separate the NIR difference spectra using OPUS (ver. 4.0, Bruker Optik GmbH). Figure 1 shows the difference and the decomposed spectra of water in the modern sample. An areal integral for each component, AI(Sn), was calculated by (1), where in Sn is one of the structural models of water molecule (n=0,1,2), and DA(ν) is the difference absorbance at a wavenumber of ν. (1) AI ( S n ) DA( ) d Adsorption/desorption isotherm for the modern and archaeological wood Figure 2 shows the adsorption/desorption isotherms of the modern and archaeological wood samples. Both of the samples show hysteresis loops. The equilibrium moisture content of the archaeological samples reduced compared to modern samples at each RH level. This is due to the temporal decrease of holocelluloce of which OH groups form hydrogen bonding with ambient water molecules. Spectroscopic interpretation of the mechanism of water adsorption to wood Figure 3 plots the variation of the areal integral (AI(Sn)) (a) and the peak wavenumber (νp(Sn)) (b) for three water components in the NIR difference spectra of the modern sample versus RH. The archaeological sample shows the same tendencies. For the sake of simplicity, we assumed that an areal integral derived from the spectral decomposition is proportional to the amount of the adsorbed water molecules. S1 and S2 components show the hysteresis loop, whereas the S0 component does not clearly. The variation of AI(Sn) and νp(Sn) can be explained by classifying the RH range into following three stages: Stage I (RH=0-40 %), Stage II (RH=40-90 %) and Stage III (RH=90-100 %). Table 1 summarizes the spectroscopic characteristics. In Stage I, the water molecules interact with wood substance more strongly than the other stages since a monomolecular layer of water is formed. It is therefore suggested that the most of monomolecular layer is composed of the S2 component. The wavenumber, νp(S0) and νp(S1) showed almost the same value both in the adsorption and desorption processes at Stage I. This suggests that the water molecules adsorbed in wood substance consist predominantly of two of the structural models, namely, S2 and S0 (or S1) component. The S0 component is very likely to exist when adjacent water molecules are sparse so that the S0 component increased with an increase of RH at this stage. On the other hand, the S0 component decreased gradually at the RH more than 40 % possibly because of an expansion to the upper layers. In Stage II, the water molecules interact with adjacent water molecules, because two or more layers (multilayers) are formed on wood surface. The areal integral, AI(S0) decreased with an increase of RH. This might be due to the three-dimensional (inter-layer) expansion of water molecules and the increase of bonding force within water molecules. The shift of νp(S0) and νp(S2) to higher and lower wavenumbers, respectively, may also be caused by the inter-layer expansion of water molecules. Bulk water (or capillary condensed water) is observed at Stage III, where the water molecules exist in not only adsorbed but also bulk states. The increase of the S1 and S2 components in this RH range is directly associated with the increase of bulk water. Temporal changes of water adsorption mechanism to wood The areal integral ratio of each water component to the total integral (PI(Sn)) was calculated to compensate for the difference of absorbance due to the sample differences (2). AI ( S n ) (2) PI ( S n ) ( AI ( S 0 ) AI ( S1 ) AI ( S 2 )) 100% Figure 4 shows the variations of PI(Sn) and νp(Sn) for the modern and archaeological hinoki wood samples with RH. In Stage III (almost close to the fiber saturation point), PI(Sn) and νp(Sn) depend on the porous structure in wood, converging on the respective values for both of the modern and archaeological samples. In the case of Stage I and Stage II, relatively large differences in PI(Sn) was found between the modern and the archaeological samples. It can therefore be concluded that modern wood adsorbs water molecules strongly than archaeological (i.e. temporally degraded) wood, since the S2 component interacts more strongly to wood than the other components. It was suggested that the above-mentioned difference of water condition between the modern and archaeological wood occurs due to the decrease of adsorption site in hemicellulose or amorphous cellulose. Conclusion We investigated the adsorption/desorption mechanism of water in wood and its temporal change in archaeological wood, using NIR spectroscopy. The mixture model of water was adapted to explain the wood-water interaction, where the difference spectra were decomposed into three components. The variation of the areal integral and the peak wavenumber for each component with relative humidity could be explained by proposing a three stages model for water adsorption. Table 1: Adsorbed water condition in wood derived from NIR spectra. Areal intgral Water condition in wood 0-40 % AI(S 0) = AI(S 1) < AI(S 2) Monomolecular layer Stage II 40-90 % AI(S0) <AI(S1) < AI(S2) Multimolecular layer + Minor bulk water Stage III 90-100 % AI(S 0) << AI(S1) = AI(S 2) Multimolecular layer +Major bulk water Difference absorbance Stage I Difference spectrum of water in modern wood 0.6 S1 0.4 S2 0.2 S0 0 5400 5200 5000 Wavenumber (cm-1) 4800 Figure 1- Difference and decomposed spectra of water in the modern hinoki wood sample. Moisture content (%) Modern wood Archaeological wood 20 Desorption 10 Adsorption 0 20 40 60 80 100 RH (%) Figure 2- Adsorption/desorption isotherm constructed from the modern and archaeological hinoki wood samples. StageⅠ StageⅡ StageⅢ 200 150 AI (Sn) (a) S0 S1 S2 desorption 100 50 adsorption 0 (b) 5250 νp (Sn) (cm -1) 5200 5150 5100 S0 S1 S2 5050 5000 0 20 40 60 80 100 RH (%) Figure 3- Variations of the S0, S1 and S2 components with RH for the modern hinoki wood sample. StageⅠ StageⅡ StageⅢ StageⅠ StageⅡ StageⅢ 80 (a) S0 Modern wood Archaeological wood 70 5200 50 40 30 20 5150 5100 5050 10 Modern wood Archaeological wood 5000 0 80 (b) S1 Modern wood Archaeological wood 70 5250 (e) S1 νp (S1) (cm-1) 60 PI (S1) (%) (d) S0 5250 νp (S0) (cm-1) PI (S0) (%) 60 50 40 30 5200 5150 5100 20 5050 10 (c) S2 70 5200 50 νp (S 2) (cm-1) PI (S2) (%) Modern wood Archaeological wood (f) S2 5250 60 40 30 20 Modern wood Archaeological wood 10 0 Modern wood Arcaheological wood 5000 0 80 20 40 60 RH (%) 80 5150 5100 5050 100 5000 0 20 40 60 80 100 RH (%) Figure 4- Variations of the proportional areal integral PI(Sn) and the peak position νp(Sn) of the three water components (S0,S1 and S2) with RH. Acknowledgement This study was partly supported by Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (grant numbers 19380099 to ST and 19540493 to HY) and TOSTEM Foundation for Construction Materials Industry Promotion. References 1. T. Inagaki, H. Yonenobu and S. Tsuchikawa (2008): Near-Infrared Spectroscopic Monitoring of the Water Adsorption/Desorption Process in Modern and Archaeological Wood. Appl. Spectrosc. 62: 860-865. 2. S. Tsuchikawa (2007): A Review of Recent Near Infrared Research for Wood and Paper. Appl. Spectrosc. Rev. 42: 43-71. 3. S. Tsuchikawa and S. Tsutsumi (1998): Adsorptive and cappillary condensed water in biological material. J. Material Science Lletter 17, 661-663. 4. K. Buijs and G.R. Choppin (1963): Near-Infrared Studies of the Structure of Water. I. Pure Water. J.Chem. Phys. 39, 2035.