**3.5. Lignocellulosic materials**

Wood or lignocellulosic materials in general are ubiquitous in the environment and are therefore involved in many degradation processes of material mixtures such as municipal solid waste, landfill materials or soils. In the previous sections lignocellulosic material was often part of waste organic matter and evaluated together with other components. Although the degradation behaviour of the mixture can be or is different from those of single constituents, the knowledge of their composition, of differences and physico-chemical properties are useful for a better understanding of the behaviour of lignocellulosic material. The main structural wood polymers - cellulose, hemicelluloses, and lignin - are the most abundant biopolymers of the Earth's carbon cycle. These polymers form the lignocellulose complex in all woody tissues. The highly ordered structure of cellulose microfibril aggregates embedded in a matrix of hemicelluloses and lignin provides the basis for its mechanical strength [46] and for the resistance to microbial attack [47], to which also low molecular mass extractives contribute [48].

#### *3.5.1. Wood types / species and their composition*

The two wood types - hardwood and softwood – can be identified by FT-IR due to their different chemical composition. A PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of ATR-FT-IR spectra of different wood species belonging to hardwood (Poplar - *Populus × canadensis*, Beech - *Fagus sylvatica*, Birch - *Betula pendula*) or softwood (Pine - *Pinus sylvestris*, Spruce - *Picea abies*) shows their separation in the scores plot along PC1 (Figure 7A), which accounts for 77% of the spectral variation. The loadings plot of the first principal component (PC1) (Figure 7C) shows the variables (wavenumbers) that are responsible for the separation, describing the differences due to different contents or numbers of functional groups representing chemical compounds. A positive loading means that the samples with positive values in the scores plot have a higher number of the functional group represented by this wavenumber, e.g. hardwoods have a higher number of acetyl groups (C=O stretching vibration of the acetyl group at 1735 cm-1) than softwoods (Figure 7C) or vice versa softwoods have a lower number of acetyl groups. The acetyl groups derive from the acetic acid esters attached to the hemicelluloses. Comprehensive lists for the assignment of bands found in the infrared spectra of wood and acetylated wood can be found elsewhere [49-51]. The band at 1235 cm-1 corresponds to the C-O vibration of this acetyl ester. Three further remarkable bands at 1593 cm-1, 1510 cm-1, and 1268 cm-1 represent lignin. The lignin content and composition of hardwoods and softwoods are different. Softwood has higher lignin contents than hardwood which is represented in PC1 by the negative loading of the band at 1510 cm-1. Softwood lignin consists mainly of G-lignin (guaiacyl units) and hardwood lignin mainly of S-lignin (syringyl units). Besides small differences of the wavenumber at which the maxima of the bands appear the latter has a stronger band at 1593 cm-1. Therefore the loading of this band is positive (Figure 7C, PC1).

148 Multivariate Analysis in Management, Engineering and the Sciences

which also low molecular mass extractives contribute [48].

*3.5.1. Wood types / species and their composition* 

**3.5. Lignocellulosic materials** 

the particle was removed and the core was used for sample preparation. However, mineral particles were also found in the capillaries of historical charcoals. Their presence was confirmed by the higher percentage of inorganic matter. It can be assumed that other unknown chemical compounds were adsorbed at the surfaces of charcoal capillaries. The inclusion of non-charcoal compounds was considered as an intrinsic factor of the ageing process and therefore the whole sample was taken as "historical" charcoal. Figure 6B displays a PCA result based on DSC profiles of historical and recent charcoals. Historical

and recent charcoals are clearly segregated. The first PC explains 70% of the variance.

Wood or lignocellulosic materials in general are ubiquitous in the environment and are therefore involved in many degradation processes of material mixtures such as municipal solid waste, landfill materials or soils. In the previous sections lignocellulosic material was often part of waste organic matter and evaluated together with other components. Although the degradation behaviour of the mixture can be or is different from those of single constituents, the knowledge of their composition, of differences and physico-chemical properties are useful for a better understanding of the behaviour of lignocellulosic material. The main structural wood polymers - cellulose, hemicelluloses, and lignin - are the most abundant biopolymers of the Earth's carbon cycle. These polymers form the lignocellulose complex in all woody tissues. The highly ordered structure of cellulose microfibril aggregates embedded in a matrix of hemicelluloses and lignin provides the basis for its mechanical strength [46] and for the resistance to microbial attack [47], to

The two wood types - hardwood and softwood – can be identified by FT-IR due to their different chemical composition. A PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of ATR-FT-IR spectra of different wood species belonging to hardwood (Poplar - *Populus × canadensis*, Beech - *Fagus sylvatica*, Birch - *Betula pendula*) or softwood (Pine - *Pinus sylvestris*, Spruce - *Picea abies*) shows their separation in the scores plot along PC1 (Figure 7A), which accounts for 77% of the spectral variation. The loadings plot of the first principal component (PC1) (Figure 7C) shows the variables (wavenumbers) that are responsible for the separation, describing the differences due to different contents or numbers of functional groups representing chemical compounds. A positive loading means that the samples with positive values in the scores plot have a higher number of the functional group represented by this wavenumber, e.g. hardwoods have a higher number of acetyl groups (C=O stretching vibration of the acetyl group at 1735 cm-1) than softwoods (Figure 7C) or vice versa softwoods have a lower number of acetyl groups. The acetyl groups derive from the acetic acid esters attached to the hemicelluloses. Comprehensive lists for the assignment of bands found in the infrared spectra of wood and acetylated wood can be found elsewhere [49-51]. The band at 1235 cm-1 corresponds to the C-O vibration of this acetyl ester. Three further remarkable bands at 1593 cm-1, 1510 cm-1, and 1268 cm-1 represent lignin. The lignin

**Figure 7.** A-C: PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of the baseline-corrected and minimum-maximum normalised ATR-FT-IR spectra of different wood species (Pine - *Pinus sylvestris*, Spruce - *Picea abies*, Poplar - *Populus × canadensis*, Beech - *Fagus sylvatica*, Birch - *Betula pendula*); Scores plots of the first two principal components labelled according to wood type (A) and species (B) and their loading spectra (C)

The band at 1268 cm-1 typical for G-lignin shows negative loadings and scores, because more G-lignin is found in softwoods. The separation along PC2 (13%) is mainly due to different lignin contents of the different species. For further interpretation of the differences the reader is pointed to additional literature [49-52].

Lignin content and lignin composition are important wood parameters. Their fast and reliable determination is therefore of interest and was studied with infrared spectroscopic methods using both ranges the near infrared (NIR) [53-59] and the mid infrared (MIR) [56, 60] leading to many PLS-R models. Simple linear regressions between the ratio of bands heights (H1510 / H897) [60] and the lignin content (Figure 8A) result in prediction models with similar precision as those obtained form PLS-R models for MIR (Figure 8B) and NIR [58, 59].

Ageing and Deterioration of Materials in the Environment – Application of Multivariate Data Analysis 151

NIR [60, 66-72] and MIR [73, 74] spectroscopy have been used since about three decades to follow the chemical changes due to fungal decay. How less invasive spectroscopic and microspectroscopic methods contribute to understand fungal wood decay has been

The result of fungal decay of wood which has been exposed for a longer period can normally be seen at once. From the practical point of view when e.g. construction wood in service has to be evaluated the early degradation stages which cannot be seen are of special interest because the mechanical properties of wood are strongly influenced. Therefore spruce wood (*Picea abies* L. (Karst.)) was incubated with three strains of the selective whiterot fungi *Ceriporiopsis subvermispora* (namely FPL 90.031, FPL 105.752 and CBS 347.63) for 14 days (details in [60]). The PC1 – PC2 scores plot (Figure 9A) of a PCA, based on the second derivatives of the MIR spectra from 1800 cm-1 to 1490 cm-1, shows that the time course can be followed along PC1. From the loadings spectra (not shown here but in [60]) it is known that the lignin content decreases with increasing incubation time. Along PC3 (Figure 9B) a kind of clustering of the three strains was obtained. This means that the PCA also allowed separation of the three strains and as a consequence from the loading spectra (cp. Figure 2e in [60]) information about the different behaviour of the three strains can be gained. Besides small differences in oxidation products and water adsorption properties the number of acetyl ester groups is different, which points to slightly different decay pattern of hemicelluloses. A better separation of the three strains was obtained using NIR spectra [60].

**Figure 9.** A-B: PCA results of the second derivatives of the MIR spectra from 1800 cm-1 to 1490 cm-1: (A) scores plot PC1 – PC2; (B) scores plot PC1 – PC3; Adapted from [60] with the permission from IM

The degradation of wood exposed for a longer period, was investigated using three types of pine wood (*Pinus sylvestris*) samples: a) recent, b) strongly degraded in a forest, and c) subfossil wood. The latter one, which was found in the sediment of a lake in Finland, was

dated dendrochronologically to be 4000 – 5000 years-old [76].

reviewed recently [75].

Publications

**Figure 8.** A-B: Calibration with a good correlation (r = 0.965) between band-heights ratios H1510 / H897 from MIR spectra and lignin contents (A); Cross-validation result of the PLS-R model for the total lignin content determination (B) using the minimum - maximum normalised MIR spectra in the wavenumber range from 1745 cm-1 to 790 cm-1 with 4 PLS components, R2 = 0.89, and a RMSECV of 0.43%

#### *3.5.2. Degradation of wood*

Ageing and deterioration of wood is caused by light [61], temperature [62], moisture, microorganism [10], fungi [47, 63], and others, which influence physical and chemical properties.

Wood is a remarkably durable material. In nature, only higher fungi have developed biochemical systems to degrade the lignocellulose complex and perform the conversion and mineralisation of wood to carbon dioxide and water. Most of these fungi belong to basidiomycetes. Although they are phylogenetically closely related [64] their strategies of degrading wood are diverse: While brown-rot fungi degrade primarily the wood polysaccharides and leave behind a polymeric but highly modified lignin, simultaneous white-rot fungi degrade all polymeric wood constituents at similar rates. Selective white-rot fungi, which lack the ability to degrade cellulose efficiently, cause extensive delignification of wood. Ascomycetes and Deuteromycetes may cause soft-rot decay that leads to softening of wet wood. Cavity formations in wood cell walls are most characteristic for this decay type. Extensive reviews on decay pattern, chemistry, and biochemistry of microbial wood degradation are available [47, 63, 65].

NIR [60, 66-72] and MIR [73, 74] spectroscopy have been used since about three decades to follow the chemical changes due to fungal decay. How less invasive spectroscopic and microspectroscopic methods contribute to understand fungal wood decay has been reviewed recently [75].

150 Multivariate Analysis in Management, Engineering and the Sciences

Lignin content and lignin composition are important wood parameters. Their fast and reliable determination is therefore of interest and was studied with infrared spectroscopic methods using both ranges the near infrared (NIR) [53-59] and the mid infrared (MIR) [56, 60] leading to many PLS-R models. Simple linear regressions between the ratio of bands heights (H1510 / H897) [60] and the lignin content (Figure 8A) result in prediction models with similar precision as those obtained form PLS-R models for MIR (Figure 8B) and NIR [58, 59].

**Figure 8.** A-B: Calibration with a good correlation (r = 0.965) between band-heights ratios H1510 / H897 from MIR spectra and lignin contents (A); Cross-validation result of the PLS-R model for the total lignin content determination (B) using the minimum - maximum normalised MIR spectra in the wavenumber

Ageing and deterioration of wood is caused by light [61], temperature [62], moisture, microorganism [10], fungi [47, 63], and others, which influence physical and chemical

Wood is a remarkably durable material. In nature, only higher fungi have developed biochemical systems to degrade the lignocellulose complex and perform the conversion and mineralisation of wood to carbon dioxide and water. Most of these fungi belong to basidiomycetes. Although they are phylogenetically closely related [64] their strategies of degrading wood are diverse: While brown-rot fungi degrade primarily the wood polysaccharides and leave behind a polymeric but highly modified lignin, simultaneous white-rot fungi degrade all polymeric wood constituents at similar rates. Selective white-rot fungi, which lack the ability to degrade cellulose efficiently, cause extensive delignification of wood. Ascomycetes and Deuteromycetes may cause soft-rot decay that leads to softening of wet wood. Cavity formations in wood cell walls are most characteristic for this decay type. Extensive reviews on decay pattern, chemistry, and biochemistry of microbial wood

range from 1745 cm-1 to 790 cm-1 with 4 PLS components, R2 = 0.89, and a RMSECV of 0.43%

*3.5.2. Degradation of wood* 

degradation are available [47, 63, 65].

properties.

The result of fungal decay of wood which has been exposed for a longer period can normally be seen at once. From the practical point of view when e.g. construction wood in service has to be evaluated the early degradation stages which cannot be seen are of special interest because the mechanical properties of wood are strongly influenced. Therefore spruce wood (*Picea abies* L. (Karst.)) was incubated with three strains of the selective whiterot fungi *Ceriporiopsis subvermispora* (namely FPL 90.031, FPL 105.752 and CBS 347.63) for 14 days (details in [60]). The PC1 – PC2 scores plot (Figure 9A) of a PCA, based on the second derivatives of the MIR spectra from 1800 cm-1 to 1490 cm-1, shows that the time course can be followed along PC1. From the loadings spectra (not shown here but in [60]) it is known that the lignin content decreases with increasing incubation time. Along PC3 (Figure 9B) a kind of clustering of the three strains was obtained. This means that the PCA also allowed separation of the three strains and as a consequence from the loading spectra (cp. Figure 2e in [60]) information about the different behaviour of the three strains can be gained. Besides small differences in oxidation products and water adsorption properties the number of acetyl ester groups is different, which points to slightly different decay pattern of hemicelluloses. A better separation of the three strains was obtained using NIR spectra [60].

**Figure 9.** A-B: PCA results of the second derivatives of the MIR spectra from 1800 cm-1 to 1490 cm-1: (A) scores plot PC1 – PC2; (B) scores plot PC1 – PC3; Adapted from [60] with the permission from IM Publications

The degradation of wood exposed for a longer period, was investigated using three types of pine wood (*Pinus sylvestris*) samples: a) recent, b) strongly degraded in a forest, and c) subfossil wood. The latter one, which was found in the sediment of a lake in Finland, was dated dendrochronologically to be 4000 – 5000 years-old [76].

The results of a PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of the ATR-FT-IR spectra of pine wood (*Pinus sylvestris*) of varying degradation stages are shown in Figure 10. The scores plot of the first two principal components (Figure 10A) reveals that the two strongly degraded samples are far from the other ones along both axes. The loadings plot of PC2 (Figure 10C – PC2 A) shows that mainly polysaccharides, preferably hemicelluloses have been degraded. This conclusion is confirmed by the loss of the acetyl-ester band at 1735 cm-1.

Ageing and Deterioration of Materials in the Environment – Application of Multivariate Data Analysis 153

sample was degraded by brown-rot fungi. The loading spectrum of PC1 (PC1 B) of the PCA shown in Figure 10B shows similarity with the previous one. This means that the difference between the recent samples and the subfossil samples is an increase in the lignin content. In his thesis Stich [76] reviewed possible degradation types and mechanisms including the preservation of organic matter in subfossil and fossil samples [77]. Based on his spectroscopic and chemical results he concluded that a slow *in situ* hydrolysis had taken

The service life of wood for outdoor use such as for windows, doors, balcony, roofs, bridges, and other applications should be as high as possible. Wood is a remarkably naturally durable material. This natural durability in terms of decay resistance against fungi, varies in a wide range between species and even within species [78-80] and can also be predicted by infrared spectroscopy [78, 81, 82]. The natural durability of wood mainly depends on the extractives composition [79, 83] but also on the extractives content [79] and is in general higher in hardwoods than in softwoods. Besides the natural durability and biological control of wood decay against fungal infection [84] thermal treatment of wood has shown to improve the service life [62, 85, 86] as well as chemical modification of wood such as acetylation [86-91], butyrylation [92] or furfurylation [93]. Traditional wood protection methods employ chemicals [94-99] that are considered toxic and can adversely affect human health and the environment [100]. Serious efforts are made globally to develop alternative protection methods based on natural products with little or no toxicity. The implementation of these technologies progresses slowly because of certain limitations, including discrepancies between laboratory and field performance of natural products, variability in their efficacy related to exposure/environmental conditions, and legislation difficulties due to disagreements on setting standards defining the quality of their performance and use. However, information on those natural compounds that have shown promise for wood

Ageing and deterioration of different complex materials in the environment were characterised using FT-IR spectroscopy and thermal analysis. Due to large data pools generated by these analyses multivariate statistical methods were applied for data evaluation. Several examples of practical application and basic research were selected. Several samples originate from current investigations (section 3.3 and 3.4). The sample sets will be extended to elucidate the processes of ageing and deterioration. For compost application on soils the question of remaining organic matter in the long-term under the given climatic and soil conditions will be relevant. The contribution of different environmental factors should be revealed by multivariate data analysis. With regard to charcoals more historical samples from different regions are necessary to find out the contribution of pyrolysis temperature and time, wood species, applied technology and environmental conditions. Based on the collected data sets evaluation can be performed

*3.5.3. Natural durability of wood - Preventing wood from degradation* 

protection is available under defined interactive categories [100].

under diverse aspects by means of adequate multivariate methods.

place in these subfossil samples.

**4. Conclusions** 

The loading spectrum of PC1 (Figure 10C – PC1 A) shows that the strongly degraded sample consists almost exclusively of lignin. A comparison of this spectrum with a milled wood lignin spectrum confirmed this (not shown). Therefore it can be concluded that this

**Figure 10.** A-C: PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of the baseline-corrected and minimum - maximum normalised ATR-FT-IR spectra of pine wood (*Pinus sylvestris* L.) of varying degradation stages; Scores plots of the first two principal components labelled according to degradation stage with (A) and without the samples labelled "degraded" (B) and their loading spectra (C), whereas A refers to the PCA in (A) and B to the PCA shown in (B)

sample was degraded by brown-rot fungi. The loading spectrum of PC1 (PC1 B) of the PCA shown in Figure 10B shows similarity with the previous one. This means that the difference between the recent samples and the subfossil samples is an increase in the lignin content. In his thesis Stich [76] reviewed possible degradation types and mechanisms including the preservation of organic matter in subfossil and fossil samples [77]. Based on his spectroscopic and chemical results he concluded that a slow *in situ* hydrolysis had taken place in these subfossil samples.

#### *3.5.3. Natural durability of wood - Preventing wood from degradation*

The service life of wood for outdoor use such as for windows, doors, balcony, roofs, bridges, and other applications should be as high as possible. Wood is a remarkably naturally durable material. This natural durability in terms of decay resistance against fungi, varies in a wide range between species and even within species [78-80] and can also be predicted by infrared spectroscopy [78, 81, 82]. The natural durability of wood mainly depends on the extractives composition [79, 83] but also on the extractives content [79] and is in general higher in hardwoods than in softwoods. Besides the natural durability and biological control of wood decay against fungal infection [84] thermal treatment of wood has shown to improve the service life [62, 85, 86] as well as chemical modification of wood such as acetylation [86-91], butyrylation [92] or furfurylation [93]. Traditional wood protection methods employ chemicals [94-99] that are considered toxic and can adversely affect human health and the environment [100]. Serious efforts are made globally to develop alternative protection methods based on natural products with little or no toxicity. The implementation of these technologies progresses slowly because of certain limitations, including discrepancies between laboratory and field performance of natural products, variability in their efficacy related to exposure/environmental conditions, and legislation difficulties due to disagreements on setting standards defining the quality of their performance and use. However, information on those natural compounds that have shown promise for wood protection is available under defined interactive categories [100].

### **4. Conclusions**

152 Multivariate Analysis in Management, Engineering and the Sciences

1735 cm-1.

The results of a PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of the ATR-FT-IR spectra of pine wood (*Pinus sylvestris*) of varying degradation stages are shown in Figure 10. The scores plot of the first two principal components (Figure 10A) reveals that the two strongly degraded samples are far from the other ones along both axes. The loadings plot of PC2 (Figure 10C – PC2 A) shows that mainly polysaccharides, preferably hemicelluloses have been degraded. This conclusion is confirmed by the loss of the acetyl-ester band at

The loading spectrum of PC1 (Figure 10C – PC1 A) shows that the strongly degraded sample consists almost exclusively of lignin. A comparison of this spectrum with a milled wood lignin spectrum confirmed this (not shown). Therefore it can be concluded that this

**Figure 10.** A-C: PCA based on the fingerprint region (1800 cm-1 to 700 cm-1) of the baseline-corrected and minimum - maximum normalised ATR-FT-IR spectra of pine wood (*Pinus sylvestris* L.) of varying degradation stages; Scores plots of the first two principal components labelled according to degradation stage with (A) and without the samples labelled "degraded" (B) and their loading spectra (C), whereas

A refers to the PCA in (A) and B to the PCA shown in (B)

Ageing and deterioration of different complex materials in the environment were characterised using FT-IR spectroscopy and thermal analysis. Due to large data pools generated by these analyses multivariate statistical methods were applied for data evaluation. Several examples of practical application and basic research were selected. Several samples originate from current investigations (section 3.3 and 3.4). The sample sets will be extended to elucidate the processes of ageing and deterioration. For compost application on soils the question of remaining organic matter in the long-term under the given climatic and soil conditions will be relevant. The contribution of different environmental factors should be revealed by multivariate data analysis. With regard to charcoals more historical samples from different regions are necessary to find out the contribution of pyrolysis temperature and time, wood species, applied technology and environmental conditions. Based on the collected data sets evaluation can be performed under diverse aspects by means of adequate multivariate methods.
