FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones

S. Raiskila, M. Pulkkinen, T. Laakso, K. Fagerstedt, Mia Löija, Riitta Mahlberg, Leena Paajanen, Anne-Christine Ritschkoff, Pekka Saranpää

Research output: Contribution to journalArticleScientificpeer-review

54 Citations (Scopus)

Abstract

Our purpose was to develop a FTIR spectroscopic method to be used to determine the lignin content in a large number of samples and to apply this method studying variation in sapwood and heartwood lignin content between three fast-growing cutting clones grown in three sites. Models were estimated with 18 samples and tested with 6 samples for which the Klason lignin + acid soluble lignin content had been determined. Altogether 272 candidate models were built with all-subset regressions from the principal components estimated from differently treated transmission spectra of the samples; the spectra were recorded on KBr pellets of sieved and unsieved unextracted wood powder and subjected to four different preprocessings and two different wavenumber selection schemes. The final model showed an adequate fit in the estimation data (R2 = 0.74) as well as a good prediction performance in the test data (R2P = 0.90). This model was based on the wavenumber range of 1850–500 cm–1 of the line-subtraction-normalised spectra recorded from sieved samples. The model was used to predict lignin content in 64 samples of the same material. One of the clones had a slightly lower sapwood lignin content than the two other clones. The fertile growing site with fast growing trees showed slightly higher sapwood lignin content compared with the other two sites. The model was also used to predict the lignin content in the earlywood of 45 individual annual rings. Variation between individual stems and between annual rings was found to be large. No correlation was found between the lignin content and density of earlywood.
Original languageEnglish
Pages (from-to)351–371
JournalSilva Fennica
Volume41
Issue number2
Publication statusPublished - 2007
MoE publication typeA1 Journal article-refereed

Fingerprint

Picea abies
lignin
clone
clones
prediction
acids
acid
sapwood
earlywood
sampling
cutting (process)
heartwood
individual variation
pellets
powders
stem
stems
methodology

Keywords

  • FTIR
  • lignin
  • Norway spruce
  • PCR
  • principal component regression

Cite this

Raiskila, S., Pulkkinen, M., Laakso, T., Fagerstedt, K., Löija, M., Mahlberg, R., ... Saranpää, P. (2007). FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones. Silva Fennica, 41(2), 351–371.
Raiskila, S. ; Pulkkinen, M. ; Laakso, T. ; Fagerstedt, K. ; Löija, Mia ; Mahlberg, Riitta ; Paajanen, Leena ; Ritschkoff, Anne-Christine ; Saranpää, Pekka. / FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones. In: Silva Fennica. 2007 ; Vol. 41, No. 2. pp. 351–371.
@article{c2c9fecb074d4fec8bf7898e8f56ba9e,
title = "FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones",
abstract = "Our purpose was to develop a FTIR spectroscopic method to be used to determine the lignin content in a large number of samples and to apply this method studying variation in sapwood and heartwood lignin content between three fast-growing cutting clones grown in three sites. Models were estimated with 18 samples and tested with 6 samples for which the Klason lignin + acid soluble lignin content had been determined. Altogether 272 candidate models were built with all-subset regressions from the principal components estimated from differently treated transmission spectra of the samples; the spectra were recorded on KBr pellets of sieved and unsieved unextracted wood powder and subjected to four different preprocessings and two different wavenumber selection schemes. The final model showed an adequate fit in the estimation data (R2 = 0.74) as well as a good prediction performance in the test data (R2P = 0.90). This model was based on the wavenumber range of 1850–500 cm–1 of the line-subtraction-normalised spectra recorded from sieved samples. The model was used to predict lignin content in 64 samples of the same material. One of the clones had a slightly lower sapwood lignin content than the two other clones. The fertile growing site with fast growing trees showed slightly higher sapwood lignin content compared with the other two sites. The model was also used to predict the lignin content in the earlywood of 45 individual annual rings. Variation between individual stems and between annual rings was found to be large. No correlation was found between the lignin content and density of earlywood.",
keywords = "FTIR, lignin, Norway spruce, PCR, principal component regression",
author = "S. Raiskila and M. Pulkkinen and T. Laakso and K. Fagerstedt and Mia L{\"o}ija and Riitta Mahlberg and Leena Paajanen and Anne-Christine Ritschkoff and Pekka Saranp{\"a}{\"a}",
year = "2007",
language = "English",
volume = "41",
pages = "351–371",
journal = "Silva Fennica",
issn = "0037-5330",
number = "2",

}

Raiskila, S, Pulkkinen, M, Laakso, T, Fagerstedt, K, Löija, M, Mahlberg, R, Paajanen, L, Ritschkoff, A-C & Saranpää, P 2007, 'FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones', Silva Fennica, vol. 41, no. 2, pp. 351–371.

FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones. / Raiskila, S.; Pulkkinen, M.; Laakso, T.; Fagerstedt, K.; Löija, Mia; Mahlberg, Riitta; Paajanen, Leena; Ritschkoff, Anne-Christine; Saranpää, Pekka.

In: Silva Fennica, Vol. 41, No. 2, 2007, p. 351–371.

Research output: Contribution to journalArticleScientificpeer-review

TY - JOUR

T1 - FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones

AU - Raiskila, S.

AU - Pulkkinen, M.

AU - Laakso, T.

AU - Fagerstedt, K.

AU - Löija, Mia

AU - Mahlberg, Riitta

AU - Paajanen, Leena

AU - Ritschkoff, Anne-Christine

AU - Saranpää, Pekka

PY - 2007

Y1 - 2007

N2 - Our purpose was to develop a FTIR spectroscopic method to be used to determine the lignin content in a large number of samples and to apply this method studying variation in sapwood and heartwood lignin content between three fast-growing cutting clones grown in three sites. Models were estimated with 18 samples and tested with 6 samples for which the Klason lignin + acid soluble lignin content had been determined. Altogether 272 candidate models were built with all-subset regressions from the principal components estimated from differently treated transmission spectra of the samples; the spectra were recorded on KBr pellets of sieved and unsieved unextracted wood powder and subjected to four different preprocessings and two different wavenumber selection schemes. The final model showed an adequate fit in the estimation data (R2 = 0.74) as well as a good prediction performance in the test data (R2P = 0.90). This model was based on the wavenumber range of 1850–500 cm–1 of the line-subtraction-normalised spectra recorded from sieved samples. The model was used to predict lignin content in 64 samples of the same material. One of the clones had a slightly lower sapwood lignin content than the two other clones. The fertile growing site with fast growing trees showed slightly higher sapwood lignin content compared with the other two sites. The model was also used to predict the lignin content in the earlywood of 45 individual annual rings. Variation between individual stems and between annual rings was found to be large. No correlation was found between the lignin content and density of earlywood.

AB - Our purpose was to develop a FTIR spectroscopic method to be used to determine the lignin content in a large number of samples and to apply this method studying variation in sapwood and heartwood lignin content between three fast-growing cutting clones grown in three sites. Models were estimated with 18 samples and tested with 6 samples for which the Klason lignin + acid soluble lignin content had been determined. Altogether 272 candidate models were built with all-subset regressions from the principal components estimated from differently treated transmission spectra of the samples; the spectra were recorded on KBr pellets of sieved and unsieved unextracted wood powder and subjected to four different preprocessings and two different wavenumber selection schemes. The final model showed an adequate fit in the estimation data (R2 = 0.74) as well as a good prediction performance in the test data (R2P = 0.90). This model was based on the wavenumber range of 1850–500 cm–1 of the line-subtraction-normalised spectra recorded from sieved samples. The model was used to predict lignin content in 64 samples of the same material. One of the clones had a slightly lower sapwood lignin content than the two other clones. The fertile growing site with fast growing trees showed slightly higher sapwood lignin content compared with the other two sites. The model was also used to predict the lignin content in the earlywood of 45 individual annual rings. Variation between individual stems and between annual rings was found to be large. No correlation was found between the lignin content and density of earlywood.

KW - FTIR

KW - lignin

KW - Norway spruce

KW - PCR

KW - principal component regression

M3 - Article

VL - 41

SP - 351

EP - 371

JO - Silva Fennica

JF - Silva Fennica

SN - 0037-5330

IS - 2

ER -

Raiskila S, Pulkkinen M, Laakso T, Fagerstedt K, Löija M, Mahlberg R et al. FTIR spectroscopic prediction of Klason and acid soluble lignin variation in Norway spruce cutting clones. Silva Fennica. 2007;41(2):351–371.