Determination of weight percent gain in solid wood modified with in situ cured furfuryl alcohol by near-infrared reflectance spectroscopy

https://doi.org/10.1016/j.chemolab.2008.02.002Get rights and content

Abstract

Near-infrared spectroscopy was used to determine the degree of treatment in Scots pine solid wood modified with in situ cured furfuryl alcohol. The treatment range was approximately 0 to 30% weight gain relative to wood dry mass. Reflectance near-infrared spectra were recorded – using a conventional, large sample cup – directly on wood surfaces without sample preparation. Partial Least Squares regressions were used to construct a model for weight percent gain; initially for one data set but later extended to an external test set not treated fully in the same way as the original set. This gave rise to an increase in the prediction error which was sought neutralised by addition of a small fraction of the new samples in the original calibration set. Four spectral ranges and 13 pre-processing methods were tested and the final model with addition of 30% of the new samples in the original calibration set resulted in a model with a Root Mean Square Error of Prediction of 1.7 ± 0.1% weight gain and a R2 of 0.97 ± 0.01.

Introduction

Impregnation of solid wood with a furfuryl alcohol monomer that is subsequently cured using heat and a catalyst to yield an inert polymer inside the cell wall structure, has for a long time been known to provide positive changes in wood properties. Durability towards acids and alkali was one of the first fields of interest [1], but improved mechanical properties [2], dimensional stability [3] and thereby improved biological durability was also discovered some 40 to 50 years ago in several studies. The process was then more or less dormant until Ryu et al. [4] presented work on the biological resistance of furfuryl alcohol-treated wood in 1992. In the following years two research groups simultaneously developed new catalyst-systems for implementation in the furfurylation process which later on made it possible to commercialize furfurylated wood [5].

When introducing new wood products to the market, quality control is essential. Besides the parameters normally documented for wood like modulus of elasticity, bending strength and the like, polymer mass is important for furfurylated wood, since this has been shown to associate with the enhanced properties [1], [4], [5], [6], [7]. In general, these properties are positively correlated with increasing polymer mass. The polymer mass is evaluated by the weight percent gain (WPG):WPG(%)=mofmoimoi×100%,where mof is the final oven-dry mass of the modified sample and moi is the initial oven-dry mass of the untreated sample.

Several spectroscopic methods can be used to monitor the quality of solid wood products directly in a production flow at- or in-line. Near-infrared (NIR) spectroscopy is often preferred routinely, since this has shown to have excellent prediction abilities in biological samples [8], [9], [10], [11] and require little or no sample preparation. Raman and UV–VIS spectroscopy have also been utilised; but to a smaller extent and for more specialized studies [12], [13]. The Mid-IR wavelength spectrum is mainly applied for studies of degradation in wood and generally for investigation of chemical compounds in wood [14], [15], [16]. In the present work focus is put on NIR spectroscopy and chemometrics hereon, since both VIS and Attenuated Total Reflectance IR (ATR-IR) measurements obtained on the same samples gave less good prediction models. ATR-IR results will be presented in another context, since the chemical information in these spectra provides additional and useful knowledge on the furfurylation treatment.

All measurements were performed in laboratory environments on small samples of furfurylated Scots pine (Pinus sylvestris, L.). The challenges in a possible up-scaling of such a method are not part of the present work, but several parameters related to the inherent problems with obtaining good spectra on wood will be touched upon. Eikenes et al. [17] initiated work on a system for quality control of furfurylated wood. In that study small samples of the diffuse-porous hardwood species Birch (Betula pendula, Roth.) were analysed using NIR spectroscopy on inner cross-sectional cuts. A root-mean-square error of prediction (RMSEP) referring to the WPG of 1.23 was reached on samples in the treatment range of 16.7 to 35.1 WPG using a PLS-model with six components.

To test the application range further, we chose to work on a commercially important softwood species and in a WPG range from approximately 0 to 30%. Furthermore, NIR was performed directly on external surfaces from clear tangential to clear radial cuts. Internal cracks – in some cases extending towards the surface – were present in a number of samples (mainly untreated). To summarise, several parameters complicated the recording of good spectra:

  • Softwoods have a heterogeneous layer-structure with alternating early and latewood cells with marked differences in both physical and chemical properties.

  • Wood surfaces are quite rough and sometimes even cracked which gives rise to scatter effects.

  • A coat of furfuryl alcohol polymer can be present on the external surface of samples. It was not known beforehand whether the amount of this coat was correlated to the average WPG of the sample.

  • Given different treatment degrees the samples will contain a varying amount of moisture.

Section snippets

Materials and methods

Two batches of modified wood were prepared. The first set (also termed ‘old’) was a large set meant for constructing the basic model. A second set (also termed ‘new’) was a smaller external test set made to test the prediction ability of the model constructed on the basis of the old data set.

Finding optimal wavelength range and best pre-processing

For this part, only the old samples were used in the analysis. There were two clear outliers in the data set, one reference measurement error and one sample with a spectrum behaving differently from all the remaining samples. The data set was then divided into a calibration and a validation set, as described above.

The first step in the analysis was to find the best pre-processing technique and the best NIR range. The best measurement range was selected based on the mean RMSECV per factor of all

Conclusions

This work shows that the degree of furfurylation (WPG) in Scots pine solid wood can be predicted by the use of NIR. Further, the conditioning of wood prior to NIR-analysis can be modelled away by the addition of as little as 30% of samples without the conditioning to the data with conditioned wood samples. The effect of the inclusion of the new samples into the calibration model suggests that the conditioning of the wood samples is only seen as an interfering signal in the NIR spectra. The

Acknowledgement

Åsmund Rinnan wishes to thank Frans van den Berg for good ideas during the work.

References (21)

  • M. Blanco et al.

    Use of near-infrared spectrometry in control analyses of acrylic fibre manufacturing processes

    Analytica Chimica Acta

    (1999)
  • K.K. Pandey et al.

    FTIR studies of the changes in wood chemistry following decay by brown-rot and white-rot fungi

    International Biodeterioration & Biodegradation

    (2003)
  • I.S. Goldstein

    The impregnation of wood to impart resistance to alkali and acid

    Forest Products Journal

    (1955)
  • A.P. Dunlop et al.

    The nature of furfuryl alcohol

    Industrial and Engineering Chemistry

    (1942)
  • A.J. Stamm

    Dimensional stabilization

  • J.Y. Ryu et al.

    Biological resistance of furfuryl-alcohol treated wood

  • S. Lande et al.

    Eco-efficient wood protection. Furfurylated wood as alternative to traditional wood preservation

    Management of Environmental Quality: An International Journal

    (2004)
  • W.L.E. Magalhaes et al.

    Treatment of Caribbean Pine by in-situ polymerization of styrene and furfuryl alcohol

    Journal of Applied Polymer Science

    (2004)
  • A.J. Stamm

    Dimensional Stabilization of wood with furfuryl alcohol resin

  • J.G. Pedersen et al.

    Non-destructive Evaluation of Wood by Near Infrared Reflectance Spectroscopy

    (1993)
There are more references available in the full text version of this article.

Cited by (36)

  • Montmorillonite-catalyzed furfurylated wood for flame retardancy

    2021, Fire Safety Journal
    Citation Excerpt :

    Kong et al. [15] found that PFA can encapsulate fire retardant and fix it in wood, thus reducing the leaching of fire retardant from furfurylated wood. Considering that most of the common acid catalysts, for instance, mineral acids (e.g. sulfuric acid), organic acids (e.g. maleic anhydride and p-toluenesulfonic acid), Lewis acid (e.g. ZnCl2) and iodine, may cause degradation of wood composition, posing negative impact on mechanical properties of wood [16–22]. MMT is suggested as an eco-friendly alternative catalyst since it exhibits potential in improving wood properties including thermal stability and mechanical properties [23].

  • Topochemical and light microscopic investigations of non-enzymatic oxidative changes at the initial decay stage of furfuryl alcohol-modified radiata pine (Pinus radiata) degraded by the brown rot fungus Rhodonia placenta

    2020, International Biodeterioration and Biodegradation
    Citation Excerpt :

    Schneider et al. (2000) showed that the FA solution filled the cell lumen completely or partially and penetrated the cell walls to some extent. That FA polymerization takes place within the wood cell wall has been shown on furfurylated wood by infrared microscopy (Venås and Rinnan, 2008), fluorescence microscopy (Thygesen et al., 2010) and cellular UV microspectrophotometry (UMSP) (Ehmcke et al., 2017) studies. The investigations revealed concordantly higher (poly)FA concentrations in lignin-rich parts, mainly in the cell wall corner regions (CC) and the compound middle lamella (CML).

  • Lignin-based catalysts for Chinese fir furfurylation to improve dimensional stability and mechanical properties

    2017, Industrial Crops and Products
    Citation Excerpt :

    Furfural, a renewable chemical from bioresources, is obtained xylose resulting from hydrolysis of pentosan-rich biomass, such as wood, and agricultural residues of sugarcane, corn and wheat (Wang et al., 2006). Furfuryl alcohol (FA) is a derivative of furfural through hydrotreating (Venås and Rinnan, 2008; Wang et al., 2006). FA resin, a brown red viscous liquid formed by polymerization of FA under the actions of acidic catalysts, is a kind of thermosetting polymer (Bertarione et al., 2008; Choura et al., 1996; Kim et al., 2014; Lande et al., 2004a).

  • Selection of the most informative near infrared spectroscopy wavebands for continuous glucose monitoring in human serum

    2016, Talanta
    Citation Excerpt :

    Afterwards, the prediction ability of such model is validated by applying it to a set of samples, which are not involved in the calibration. Many Multivariate Calibration methods such as Multiple Linear Regression (MLR), Principal Component Regression (PCR), Partial Least Squares (PLS), Ridge Regression (RR), Artificial Neural Networks (ANN), Support Vector Machines (SVM), etc. have been used in NIR spectral analysis [6,7]. Although biased regression methods such as RR, PCR and PLS can be used with all measured wavelengths, many of these wavelength variables are not essential for predicting the analyte of interest.

View all citing articles on Scopus
View full text