Abstract
A total of 120 Acacia melanoxylon R. Br. (Australian blackwood) stem discs, belonging to 20 trees from four sites in Portugal, were used in this study. The samples were kraft pulped under standard identical conditions targeted to a Kappa number of 15. A Near Infrared (NIR) partial least squares regression (PLSR) model was developed for the Kappa number prediction using 75 pulp samples with a narrow Kappa number variation range of 10 to 17. Very good correlations between NIR spectra of A. melanoxylon pulps and Kappa numbers were obtained. Besides the raw spectra, also pre-processed spectra with ten methods were used for PLS analysis (cross validation with 48 samples), and a test set validation was made with 27 samples. The first derivative spectra in the wavenumber range from 6110 to 5440 cm-1 yielded the best model with a root mean square error of prediction of 0.4 units of Kappa number, a coefficient of determination of 92.1%, and two PLS components, with the ratios of performance to deviation (RPD) of 3.6 and zero outliers. The obtained NIR-PLSR model for Kappa number determination is sufficiently accurate to be used in screening programs and in quality control.
Download PDF
Full Article
Kappa Number Prediction of Acacia melanoxylon Unbleached Kraft Pulps using NIR-PLSR Models with a Narrow Interval of Variation
António J. A. Santos,a,b,c Ofélia Anjos,a,d Rogério Simões,b José Rodrigues,c and Helena Pereira a*
A total of 120 Acacia melanoxylon R. Br. (Australian blackwood) stem discs, belonging to 20 trees from four sites in Portugal, were used in this study. The samples were kraft pulped under standard identical conditions targeted to a Kappa number of 15. A Near Infrared (NIR) partial least squares regression (PLSR) model was developed for the Kappa number prediction using 75 pulp samples with a narrow Kappa number variation range of 10 to 17. Very good correlations between NIR spectra of A. melanoxylon pulps and Kappa numbers were obtained. Besides the raw spectra, also pre-processed spectra with ten methods were used for PLS analysis (cross validation with 48 samples), and a test set validation was made with 27 samples. The first derivative spectra in the wavenumber range from 6110 to 5440 cm-1 yielded the best model with a root mean square error of prediction of 0.4 units of Kappa number, a coefficient of determination of 92.1%, and two PLS components, with the ratios of performance to deviation (RPD) of 3.6 and zero outliers. The obtained NIR-PLSR model for Kappa number determination is sufficiently accurate to be used in screening programs and in quality control.
Keywords: Acacia melanoxylon; Kappa number; NIR; RPD
Contact information: a: Centro de Estudos Florestais, Instituto Superior de Agronomia, Universidade de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, Portugal; b: Universidade da Beira Interior, Unidade de Materiais Têxteis e Papeleiros, 6201-001 Covilhã, Portugal; c: Tropical Research Institute of Portugal (IICT), Forest and Forest Products Centre, Tapada da Ajuda, 1349-017 Lisboa, Portugal; d: Instituto Politécnico de Castelo Branco, Escola Superior Agrária, Apartado 119, 6001-909 Castelo Branco, Portugal;
* Corresponding author: hpereira@isa.ulisboa.pt
INTRODUCTION
Acacia melanoxylon R. Br. (blackwood) is one of the main wattle species valued for its wood in its natural region of Australia. In Europe, where it was introduced as an ornamental tree and for stabilization of sand dunes (Humphries 1996), the species has not yet found its place in the wood processing industry. However recent studies researched the technological quality of European grown A. melanoxylon regarding e.g. heartwood variation (Knapic et al. 2006), wood basic density (Igartúa and Monteoliva 2009; Santos et al. 2012), wood macroscopic properties (Santos et al. 2013), and mechanical properties (Santos et al. 2007; Machado et al. 2014).
A. melanoxylon is also an interesting species for pulping, and pulp yields and properties were already evaluated (Santos et al. 2005, 2006, 2012; Anjos et al. 2011; Lourenço et al. 2008). One important pulping quality parameter is the delignification degree, measured by the residual lignin content in the pulp. Usually it is determined as the Kappa number, which is used in the pulp industry as a process control tool to maximize yield and quality while minimizing consumption of energy and chemicals.
The determination of Kappa number relies on the oxidation of the residual lignin in the pulp and is a standard method (TAPPI T 236 om-13 1999). This classical wet-chemistry method is, however, tedious and impractical when a large number of samples has to be evaluated, for instance in screening programs. This has led to the search for fast, non-destructive, reproducible, and accurate determination methods.
Near-infrared (NIR) spectroscopy is a method that complies with such requirements and has been applied to determination of various properties in numerous materials. For instance, it has been used to estimate wood chemical components including lignin (Schwanninger and Hinterstoisser 2001; Kelley et al. 2004; Alves et al. 2006; Yao et al. 2010; Zhang et al. 2011) and cellulose (Wright et al. 1990; Schimleck et al. 1997; Poke and Raymond 2006; Hein et al. 2009), as well as for the determination of wood properties (Hauksson et al.2001; Kelley et al. 2004; Sykes et al. 2005) and pulp yield (Michell 1995; Hodge and Woodbridge 2004; Schimleck et al. 2005; Downeset al. 2009). NIR spectroscopy was also applied for the determination of the Kappa number of hardwood pulps (Easty et al. 1990; Michell 1995; Malkavaara and Alen 1998; Fardim et al. 2002; Terdwongworakul et al. 2005) and of softwood pulps (Malkavaara and Alén 1998; Alves et al. 2007). However, in most cases, the models were developed using samples in which the range of Kappa numbers was controlled by varying the pulping conditions, e.g. the wood chips were delignified to different degrees targeted to achieve a suitable variation of Kappa numbers e.g. 58 to 100 in pine pulps (Alves et al. 2007), 30 to 90 in mixed pulps (Antti et al. 2000); 3 to 34 in hardwood softwood mixed pulps (Easty 1990), and 16 to 27 in eucalypt pulps (Birkett and Gambino 1989). This does not correspond to normal pulping situations where the pulping conditions are constant and the Kappa numbers may fall within a narrow range of 13 to 20.
This paper focuses on the development of a NIR-PLSR model for the prediction of the Kappa number of Kraft pulps obtained under such conditions, e.g. identical pulping conditions and a narrow variation range, using A. melanoxylon pulping as a case study. It was aimed at obtaining models performing better than those found in the literature, which are not precise enough for screening purposes, according to AACC Method 39-00 (1999), due to failing the ratios of performance to deviation (RPD) criteria.
EXPERIMENTAL
Sample Preparation
A total of 120 A. melanoxylon wood discs, belonging to 20 trees from four sites in Portugal, were used in this study. Detailed information on the samples, sites, and stands is available elsewhere (Santos et al.2013). Samples of 25 g oven-dry wood milled using a knife mill (Retsch) with a 1 mm output screen and the fraction coarser than 0.25 mm was selected to Kraft pulping using a multi-batch digester system under the following reaction conditions: active alkali charge 21.3% (as NaOH); sulfidity 30%; liquor/wood ratio 4/1; time to temperature of 160 ºC, 90 min; time at temperature of 160 ºC, 90 min. These conditions were set after a few preliminary experiments to obtain the target Kappa number of 15, as needed for papermaking applications. The pulp was washed according to Santos et al. (2012). Under these conditions the pulp yield ranged from 47.0 to 58.2%. The pulp yield range was wide corresponding to the large variation of heartwood proportion in the wood of the different discs, as shown in the same samples by Santos et al. (2013) e.g. 38 to 45% at 65% height level and 2 to 15 % at the tree top.
The Kappa number of the pulp samples was determined according to the standard method TAPPI 236 om-13 (1999). A selection of 75 samples was made from the total samples, based on those with Kappa numbers between 10 and 17.
Pulps discs were produced with the washed and screened pulps, in Glass Vacuum Filter Holder 16307 – Sartorius with 0.2 µm membrane filter, and a diameter matching the diameter of the spinning cup module of the near infrared spectrometer.
Spectra Collection
The pulp disc samples were conditioned in a climatic chamber at 60 ºC for a period of 48 hours before spectral acquisition. NIR spectra of pulp fibres in disc form were collected in the wavenumber range from 12000 to 3800 cm-1 with a near infrared spectrometer (BRUKER, model Vector 22/N) in the diffuse reflectance mode, using a spinning cup module. Each spectrum was obtained with 100 scans at a spectral resolution of 16 cm-1.
Data Processing
The samples were randomly divided into a calibration set containing 48 samples and a validation set (test set) containing 27 samples by means of a principal component analysis. The processing was done according to Alves et al. (2007) in two steps. First, the infrared data from the calibration samples were regressed against the Kappa number, and by means of full cross validation with one sample omitted a significant number of PLS components (rank) was obtained using OPUS Quant 2 software (BRUKER). Besides using the raw spectra, the following pre-processing methods were applied prior to calculation of the PLSR models: multiplicative scatter correction (MSC), first derivative (1stDer), second derivative (2ndDer), vector normalization (VecNor), straight line subtraction (SLS), minimum maximum normalization (MinMax), constant offset (ConOff), and combinations of them. In a second step, the validation of the PLSR models was performed using the independent test set.
The quality of the calibration models was assessed by means of cross validation and by using the test set validation results by determining their coefficient of determination (r2), root mean square error of cross validation (RMSECV), root mean square error of prediction (RMSEP), and the residual prediction deviation or ratio of performance to deviation (RPD). The selection of the final model was based on its predictive ability assessed by the number of samples classified as outsiders and/or outliers that should be as less as possible (Gierlinger et al. 2002; Alves et al. 2007).
RESULTS AND DISCUSSION
The Kappa number of the unbleached pulps of A. melanoxylonobtained under the same pulping conditions ranged from 10.0 to 16.8, with an average of 13.6 and a standard deviation of 1.3 (Table 1), corresponding to a sample with a narrow range of Kappa numbers, as it was the objective of this work. The two sets showed similar statistics.
The spectral range from 6110 to 5440 cm-1 was used for calibration. This range was found by automated optimization and has also already used by Alves et al. (2007). The NIR based data from the 48 samples of the calibration set were regressed against their experimentally determined Kappa number, and the results obtained with the various pre-processing of the raw spectral data are summarized in Table 2. The obtained rank ranged from one (2ndDer) to five (no spectral pre-processing), while the coefficients of determination (R2) ranged from 76.4% to 87.9%, and the root mean square error of cross validation (RMSECV) from 0.5 to 0.6. When using the test set validation, the coefficients of determination ranged from 85.7% to 92.1%.
Table 1. Number of Samples and Range of Kappa Number in the Calibration Set and Test Set (SD is standard deviation)
The best model using the first derivative (1stDer) of the spectral data was selected, and the NIR-PLSR predicted versus the laboratorial determined Kappa numbers are shown in Fig. 1. Both the cross-validation and the validation show high correlation between the predicted and the determined Kappa number values, with an RMSEP of 0.4, a rank of three, and no outliers.
Table 2. Results for Cross-validation and Test Set Validation*
Fig. 1. NIR PLSR predicted versus laboratory determined Kappa number for blackwood pulps: a) cross validation (47 samples), b) test set validation (27 samples).
This is the first time that the Kappa number of hardwood unbleached kraft pulps could be predicted with such high accuracy and precision (0.6 Kappa units). In fact, previous models for Kappa number determination using NIR spectral data showed higher prediction errors, as summarized in Table 3.
Zhang et al. (2011) developed a model for Acacia spp. with root mean square error of cross validation (RMSECV) of 1.0 units of Kappa number and predictive ability corresponding to a RMSEP of 0.8 units of Kappa number, while Monrroy et al. (2008) reported 3.6 and 2.9 units of Kappa, respectively for Eucalyptus globulus. Alves et al. (2007) reported a root mean square error of prediction of 2.3 units of Kappa number for Pinus pinaster kraft pulps. Antti et al. (2000) developed a model with a predictive ability corresponding to a RMSEP of 1.7. Easty et al. (1990) calibrated the Kappa number of hardwood-softwood blends using the second derivative of the NIR spectra at 5952 cm-1 with an R2 of 0.99, but did not validate the result.
Birkett and Gambino (1989) determined multiple linear regression calibration models for Kappa number with 95% confidence limits of ± 1.4 for Eucalyptus grandis pulp and of ± 2.1 for mixed pine pulps. The best model so far published presents a root mean square error of prediction of 0.8 units of Kappa number, a coefficient of determination of 80.0 % within a range of variation of 14 Kappa units. The model developed with the largest Kappa units range (71) has a root mean square error of prediction of cross validation of 3.2.
Table 3. Kappa Range of the Different Models Published in the Bibliography
The ratios of performance to deviation (RPD) may be used to evaluate whether the prediction models fulfill the requirements of AACC Method 39-00 for screening in breeding programs or for quality control that require a RPD ≥ 2.5 or RPD ≤ 5, respectively (AACC 1999). The RPD was introduced by Williams and Norris (2004) as the ratio between the standard deviation of the reference data of the validation set and the standard error of prediction of a cross-validation or of the test set validation. In this case, the RPD for the validation of the NIR PLS-R model was 3.6 (Table 2), thereby allowing the conclusion that the method is applicable for screening in breeding programs.
The use of this statistical tool to assess near infrared-based partial least squares regression (NIR PLS-R) models has been very infrequent for forest products. The only published use of RPDs was dedicated to wood density (Santos et al. 2012; Alves et al. 2012; Inagaki et al. 2012).
A reliable model will be obtained only if the model is able to predict the test samples in a proper way (Antti et al. 2000; 1996), as it is commonly agreed. But according to the results presented here, this is not sufficient, because complementary data of an external validation is needed, stressing the importance of an additional step (evaluation) before making the decision in model selection. Following that, the 1stDer spectra gave the best model, with the highest coefficient of determination (R2), lowest root mean square error of prediction (RMSEP), and highest residual prediction deviation or ratio of performance to deviation (RPD).
CONCLUSIONS
- The results demonstrated the potential of using NIR spectra for prediction of Kappa numbers in well delignified unbleached hardwood pulps with an accuracy that fulfils the requirements to be used in screening programs and in quality control processes.
- The developed NIR-PLSR models improved the prediction ability in relation to previously published models and showed the possibility of using a narrow range of Kappa number values to perform high quality calibration and validation processes.
ACKNOWLEDGMENTS
The first author acknowledges a scholarship (SFRH/BD/42073/2007) given by Fundação para a Ciência e a Tecnologia (FCT), Portugal. Centro de Estudos Florestais is a research unit funded by FCT (PEst-OE/AGR/UI0239-2014).
REFERENCES CITED
AACC-Standards (1999). “Near-infrared methods-guidelines for model development and maintenance,” American Association of Cereal Chemists (AACC). AACC Method 39-00:15.
Alves, A., Santos, A., da Silva Perez, D., Rodrigues, J., Pereira, H., Simões, R., and Schwanninger, M. (2007). “NIR PLSR model selection for Kappa number prediction of maritime pine kraft pulps,” Wood Sci. Technol. 41(6), 491-499.
Alves, A., Santos, A., Rozenberg, P., Pâques, L., Charpentier, J., Schwanninger, M., and Rodrigues, J. (2012). “A common near infrared (NIR) – based partial least squares regression (PLS-R) model for the prediction of wood density of Pinus pinaster and Larix × eurolepis,” Wood Sci. Technol. 46(1-3), 157-175.
Alves, A., Schwanninger, M., Pereira, H., and Rodrigues, J. (2006). “Calibration of NIR to assess lignin composition (H/G ratio) in maritime pine wood using analytical pyrolysis as the reference method,” Holzforschung 60(1), 29-31.
Anjos, O., Santos, A., and Simões, R. (2011). “Effect of Acacia melanoxylon fibre morphology on papermaking potential,” Appita J.64(2), 185-191.
Antti, H., Alexandersson, D., Sjöström, M., and Wallbäcks, L. (2000). “Detection of Kappa number distributions in kraft pulps using NIR spectroscopy and multivariate calibration,” TAPPI J. 83(3), 102-108.
Antti, H., Sjöström, M., and Wallbäcks, L. (1996). “Multivariate calibration models using NIR spectroscopy on pulp and paper industrial applications,” J. Chemometr. 10(5-6), 591-603.
Birkett, M., and Gambino, M. (1989). “Estimation of pulp Kappa number with near-infrared spectroscopy,” TAPPI J. 72(9), 193-197.
Downes, G., Meder, R., Hicks, C., and Ebdon, N. (2009). “Developing and evaluating a multisite and multispecies NIR calibration for the prediction of Kraft pulp yield in eucalypts,” Southern Forests. J. Forest. Sci. 71(2), 155-164.
Easty, D., Berben, S., DeThomas, F., and Brimmer, P. (1990). “Near-infrared spectroscopy for the analysis of wood pulp: Quantifying hardwood-softwood mixtures and estimating lignin content,” TAPPI J. 73(10), 257-261.
Fardim, P., Ferreira, M., and Duran, N. (2002). “Multivariate calibration for quantitative analysis of eucalypt kraft pulp by NIR spectrometry,” J. Wood Chem. Technol. 22(1), 67-81.
Gierlinger, N., Schwanninger, M., Hinterstoisser, B., and Wimmer, R. (2002). “Rapid determination of heartwood extractives in Larix sp. by means of Fourier transform near infrared spectroscopy,” J. Near Infrared Spec. 10(3), 203-214.
Hauksson, J., Bergqvist, G., Bergsten, U., Sjostrom, M., and Edlund, U. (2001). “Prediction of basic wood properties for Norway spruce. Interpretation of Near Infrared spectroscopy data using partial least squares regression,” Wood Sci. Technol. 35(6), 475-485.
Hein, P., Sá, V., Bufalino, L., and Mendes, L. (2009). “Calibrations based on near infrared spectroscopic data to estimate wood-cement panel properties,” BioResources 4(4), 616-623.
Hodge, G., and Woodbridge, W. (2004). “Use of near infrared spectroscopy to predict lignin content in tropical and sub-tropical pines,” J. Near Infrared Spec. 12(2), 381-390.
Humphries, J. (1996). “Guia Fapas. Árvores de Portugal e Europa,” JR, Porto, pp 212-216.
Igartúa, V., and Monteoliva, S. (2009). “Densidad básica de la madera de Acacia melanoxylon R. Br. en relación con la altura de muestreo, el árbol y el sitio,” Invest. Agrar.: Sist. Recur. For. 18(1), 101-110.
Inagaki, T., Schwanninger, M., Kato, R., Kurata, Y., Thanapase, W., Puthson, P., and Tsuchikawa, S. (2012). “Eucalyptus camaldulensisdensity and fiber length estimated by near-infrared spectroscopy,” Wood Sci. Technol. 46(1-3), 143-155.
Kelley, S., Rials, G., Groom, L., and So, C. (2004). “Use of near infrared spectroscopy to predict the mechanical properties of six softwoods,” Holzforschung 58(3), 252-260.
Kelley, S., Rials, G., Snell, R., Groom, L., and Sluiter, A. (2004). “Use of near infrared spectroscopy to measure the chemical and mechanical properties of solid wood,” Wood Sci. Technol. 38(4), 257-276.
Knapic, S., Tavares, F., and Pereira, H. (2006). “Heartwood and sapwood variation in Acacia melanoxylon R. Br. trees in Portugal,” Forestry 79(4), 371-380.
Lourenço, A., Baptista, I., Gominho, J., and Pereira, H. (2008). “The influence of heartwood on the pulping properties of Acacia melanoxylon wood,” J. Wood Sci. 54(6), 464-469.
Machado, J., Louzada, J., Santos, A., Nunes, L., Anjos, O., Rodrigues, J., Simões, R., and Pereira, H. (2014). “Variation of wood density and mechanical properties of blackwood (Acacia melanoxylon R. Br.),” Materials & Design 56(4), 975-980.
Malkavaara, P., and Alén, R. (1998). “A spectroscopic method for determining lignin content of softwood and hardwood kraft pulps,” Chemometr. Intel.l Lab. 44, 287-292.
Michell, A. (1995). “Pulpwood quality estimation by near-infrared spectroscopic measurements on eucalypt woods,” Appita J. 48(6), 425-428.
Monrroy, M., Mendonça, R., Baeza, J., Ferraz, A., and Freer, J. (2008). “Estimation of hexenuronic acids and Kappa number in kraft pulps of Eucalyptus globulus by Fourier transform near infrared spectroscopy and multivariate analysis,” J. Near Infrared Spec. 16(2), 121-128.
Poke, F., and Raymond, C. (2006). “Predicting extractives, lignin, and cellulose contents using near infrared spectroscopy on solid wood in Eucalyptus globulus,” J. Wood Chem. Technol. 26(2), 187-199.
Santos, A., Alves, A., Simões, R., Pereira, H., Rodrigues, J., and Schwanninger, M. (2012). “Estimation of wood basic density of Acacia melanoxylon (R. Br.) by near infrared spectroscopy,” J. Near Infrared Spec. 20(2), 267-274.
Santos, A., Amaral, M., Gil, N., Anjos, O., Pereira, H., and Simões, R. (2012). “Influence on pulping yield and pulp properties of wood density of Acacia melanoxylon,” J. Wood Sci. 58(6), 479-486.
Santos, A., Anjos, O., and Simões, R. (2005). “Avaliação da Qualidade do Papel Produzido com Fibra de Acacia spp.,” Silva Lus.13(2), 249-266.
Santos, A., Anjos, O., and Simões, R. (2006). “Papermaking potential of Acacia dealbata and Acacia melanoxylon,” Appita J. 59(1), 58-64.
Santos, A., Simões, R., and Tavares, M. (2013). “Variation of some wood macroscopic properties along the stem of Acacia melanoxylonR. Br. adult trees in Portugal,” Forest Syst. 22(3), 463-470.
Santos, A., Teixeira, A., Anjos, O., Simões, R., Nunes, L., Machado, J., and Tavares, M. (2007). “Utilização potencial do lenho de Acacia melanoxylon a crescer em povoamentos puros ou mistos com Pinus pinaster pela indústria florestal portuguesa,” Silva Lus. 15(1), 57-78.
Schimleck, L., Payne, P., and Wearne, R. (2005). “Determination of important pulp properties of hybrid poplar by near infrared spectroscopy,” Wood Fiber Sci. 37, 462-471.
Schimleck, L., Wright, P., Michell, A., and Wallis, A. (1997). “Near-infrared spectra and chemical compositions of E. globulus and E. nitens plantation woods,” Appita J. 50(1), 40-46.
Schwanninger, M., and Hinterstoisser, B. (2001). “Determination of the lignin content in wood by FT-NIR,” 11th ISWPC, International Symposium on Wood and Pulping Chemistry, Nice, France, Centre Technique Papeterie III: 641-644.
Sykes, R., Li, B., Hodge, G., Goldfarb, B., Kadla, J., and Chang, H. (2005). “Prediction of loblolly pine wood properties using transmittance near-infrared spectroscopy,” Can. J. Forest Res. 35(10), 2423-2431.
TAPPI (1999). “Kappa number of pulp, test method T 236 om-13,” TAPPI Standard Test Methods.
Terdwongworakul, A., Punsuwan, V., Thanapase, W., and Tsuchikawa, S. (2005). “Rapid assessment of wood chemical properties and pulp yield of Eucalyptus camaldulensis in Thailand tree plantations by near infrared spectroscopy for improving wood selection for high quality pulp,” J. Wood Sci. 51, 167-171.
Zhang, H., Shuping, S., Lang, Q., Zhang, J., and Pu, J. (2011). “Rapid prediction models for minimally destructive kappa number and pulp yield of Acacia spp. with near infrared reflectance (NIR) spectroscopy,” BioResources 7(1), 616-623.
Williams, P., and Norris, K. (2004). Near-Infrared Technology in the Agricultural and Food Industries, American Association of Cereal Chemists, Inc, St. Paul, 296.
Wright, J., Birkett, M., and Gambino, M. (1990). “Prediction of pulp yield and cellulose content from wood samples using near infrared reflectance spectroscopy,” TAPPI J. 73(8), 164-166.
Yao, S., Wu, G., Xing, M., Zhou, S., and Pu, J. (2010). “Determination of lignin content in Acacia spp. Using near-infrared reflectance spectroscopy,” BioResources 5(2), 556-562.
Article submitted: July 26, 2014; Peer review completed: August 20, 2014; Revised version received: September 18, 2014; Accepted: September 21, 2014; Published: September 23, 2014.