Journal of Chromatography A, 1109 (2006) 267–272

Qualitative and quantitative determination of extractives in heartwood of Scots pine (Pinus sylvestris L.) by gas chromatography Dag Ekeberg a,∗ , Per-Otto Flæte b , Morten Eikenes b , Monica Fongen b , Carl Fredrik Naess-Andresen a a

˚ Norway University of Life Sciences, Department of Chemistry, Biotechnology and Food Science, P.O. Box 5003, N-1432 As, b Norwegian Forest Research Institute, Høgskoleveien 8, N-1432 As, ˚ Norway Received 9 November 2005; received in revised form 5 January 2006; accepted 10 January 2006 Available online 3 February 2006

Abstract A method for quantitative determination of extractives from heartwood of Scots pine (Pinus sylvestris L.) using gas chromatography (GC) with flame ionization detection (FID) was developed. The limit of detection (LOD) was 0.03 mg/g wood and the linear range (r = 0.9994) was up to 10 mg/g with accuracy within ±10% and precision of 18% relative standard deviation. The identification of the extractives was performed using gas chromatography combined with mass spectrometry (GC–MS). The yields of extraction by Soxhlet were tested for solid wood, small particles and fine powder. Small particles were chosen for further analysis. This treatment gave good yields of the most important extractives: pinosylvin, pinosylvin monomethyl ether, resin acids and free fatty acids. The method is used to demonstrate the variation of these extractives across stems and differences in north–south direction. © 2006 Elsevier B.V. All rights reserved. Keywords: Pinewood extractives; Pinosylvin; Resin acids; Free fatty acids; GC–FID; GC–MS

1. Introduction As the Norwegian Pollution Control Authority has recently proposed the prohibition of traditional wood preservation methods like CCA (copper, chromium, arsenic) or creosote impregnation, there has been increased interest in natural wood durability and other, less hazardous, protecting agents. This has brought interest in the use of heartwood and lightwood as a construction timber with increased natural durability [1]. The heartwood is the inner part of the stem and can often be detected visually by its dark colour (see Fig. 1), caused by increased amounts of extractives [2]. Wood extractives are chemical components like resin acids, free fatty acids and phenolic compounds which exist in the heartwood of Scots pine among other trees. Historically, carpenters have often used Scots pine timber, especially the heartwood, since it is renowned for its structural qualities and durability. In a study on the improvement of natural wood durability, a method for qualitative and quantitative determination

∗

Corresponding author. Fax: +47 64965901. E-mail address: [email protected] (D. Ekeberg).

0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.01.027

of extractives in heartwood of Scots pine (Pinus sylvestris L.) was developed using gas chromatography (GC) combined with flame ionization detection (FID) or mass spectrometry (MS). Some of these extractives increase the natural wood durability against destruction by microorganisms and insects. A phenolic compound called pinosylvin, which is classified as a stilbene, has shown a great influence on natural wood durability and resin acids have a restraining effect on fungi [2]. In 1999, Bergstrøm et al. [3] described the changes of concentration and distribution in pinosylvin across sapwood and heartwood. They found that there was no pinosylvin in the sapwood and that the amount of pinosylvin in the inner heartwood was lower than that in the outer. The technique used was Fourier transform near-infrared (FT-NIR) Raman spectroscopy. Scots pine heartwood and lightwood are water repellent because of large amount of hydrophobic acids which block the hydroxyl groups of the wood cell walls and encrust cell cavities [4,5]. In addition, pinosylvin and pinosylvin methyl ether are considered to be determining factors active against decay [6]. The aim of this work was to develop a method for qualitative and quantitative determination of wood extractives using

268

D. Ekeberg et al. / J. Chromatogr. A 1109 (2006) 267–272

maric acid (95%, CAS number 79-54-9) and neoabietic acid (>99%, CAS 471-77-2) all Helix Biotech, New Westminster, Canada). Heptadecanoic acid (99%, CAS number 506-12-7, Fluka, Steinheim, Germany) and diethylstilbestrol (99%, CAS number 56-53-1, Sigma, Steinheim, Germany) were used as internal standard and for evaluation of performance. Acetone (analytical-reagent grade, Merck, Darmstadt, Germany) was used to prepare stock solutions of the internal standard and for the extraction procedure. The extracts were redissolved in absolute dry pyridine (puriss, H2 O < 0.05%, Fluka) and silylated with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA, ≥99%, CAS number 25561-30-2, Sigma) which was kept at 4 ◦ C under argon atmosphere. Stock solutions of the heptadecanoic acid and diethylstilbestrol were prepared by dissolving 140 and 200 mg in 50 mL acetone, respectively. Both stock solutions were kept in the dark at 4 ◦ C. 2.2. Wood samples

Fig. 1. The selection of 10 samples from trunk discs A and B to measure the distribution of extractives across heartwood and sapwood.

GC–FID or GC–MS. This method will be used to describe the variations in natural durability of Scots pine heartwood. Methods using different solvents for extraction of pinosylvin and pinosylvine methyl ether have been published. In 2000, Ingram et al. [7] published a method where they used dichlomethane and 1,4-dichlorobenzene in a Soxhlet extraction. Earlier benzene has been used [8]. Since benzene has been shown to be carcinogenic and halogenated solvents have been shown to be mutagenic, there is a need to use alternative solvents. Several workers have found acetone to be a suitable solvent for the analysis of extractives in pulp and paper samples [9 and references therein]. Acetone in combination with other solvents as e.g. water has also been used. In 2004, Ven¨al¨ainen et al. [10] published a method where acetone/water 80/20 v/v was used. The work consisted of estimating the limit of detection, determining linear range, accuracy and precision and identifying the wood extractives. To obtain the highest yield of wood extractives from the samples, studies were performed on how to process the wood samples before extraction. The last part of this work was done to see how the wood extractives were distributed across the heartwood and sapwood, and in the north–south and east–west directions of the heartwood.

The samples were taken from five different trees, A–E. All samples were from Scots pine, P. sylvestris L. The stem of a Scots pine tree consists of sapwood and heartwood. All wood samples used in this study were prepared from cross sectional sample discs taken from mature Scots pine trees at a trunk height of about 5 m above the base. The average height of the trees was 21 m and the average stem diameter at ground level was 41.5 cm. 2.3. Extraction of wood samples Ground wood samples (200 mg) were placed in cellulose filters (10 mm × 50 mm Munktell Filters, Grycksbo, Sweden), which were then sealed with wads of cotton wool. Both the filters and the cotton wads were washed in acetone prior to use. Acetone (pro analysis, 25 mL) and stock solution of internal standard (100 ␮L) were added to the extraction bottles. The filters and the bottles were placed in a Soxhlet extraction unit (FexIKA 50, IKA Werke, Staufen, Germany). The extraction unit was heated to 78 ◦ C and kept there for 4 min. The solvent was then cooled to 40 ◦ C. The heating and cooling cycle were repeated 36 times with a total time of 9 h. The extractions were started in the afternoon and the solvent was removed by vacuum evaporation in a rotary evaporator (RV06-ML, IKA Werke) at 50 ◦ C the next day. The extractives were then redissolved in absolute dry pyridine (1 mL) and silylated with addition of N,O-bis(trimethylsilyl)trifluoroacetamide (500 ␮L) and left at ambient temperature for minimum 1 h before analysis by GC–FID or GC–MS.

2. Experimental 2.4. Evaluation of sampling 2.1. Chemicals and standards The standards used for calibration and identification where pinosylvin and pinosylvin monomethyl ether (CAS number 22139-77-1 and 35302-70-6, respectively, Apin chemicals, Abingdon, UK), palustric acid (90–95%, CAS number 194553-5), pimaric acid (75–80%, CAS number 127-27-5), levopi-

A piece of Scots pine heartwood was divided into se 3 × 3 equal sized sub-samples (10 mm × 5 mm × 30 mm) in the longitudinal and tangential directions. Three sub-sets of three samples were chosen so that none of the samples within a sub-set had an origin next to each other in the original sample. The three sub-sets were then extracted as (a) solid wood, (b) small

D. Ekeberg et al. / J. Chromatogr. A 1109 (2006) 267–272

269

dimethylpolysiloxane as stationary phase (Chrompack, Middelburg, The Netherlands). Helium (99.9999%, from Hydro Gas and Chemicals, Rjukan, Norway) was used as carrier gas, at a constant flow of 1.0 mL/min. Splitless injection (the splitless time was set to 2 min) of 1 ␮L extracts was performed with an autosampler into an injector at 250 ◦ C. The GC column temperature was programmed as follows: 10 ◦ C/min from 40 to 190 ◦ C, 4 ◦ C/min from 190 to 300 ◦ C and 300 ◦ C isothermal for 7.5 min. The FID system was operated at 300 ◦ C with a gas flow of hydrogen (99.9999%, from Hydro Gas and Chemicals) at 40 mL/min and of artificial air (99.9995%, from Hydro Gas and Chemicals) at 450 mL/min. 2.8. GC–MS analysis

Fig. 2. The selection of samples from the inner and outer parts of the heartwood in the north–south and east–west directions.

particles, typically 2 mm × 1 mm × 1 mm, ground by an IKA Basic grinder and (c) fine powder from a Retch mill (RM2000) using two wolframcarbide balls. 2.5. Distribution of extractives across heartwood and sapwood The distribution of extractives across heartwood and sapwood was investigated in two trunk discs from two trees A and B. Ten samples were taken from each of the two trunk discs, as shown in Fig. 1. 2.6. Determination of extracting efficacy from heartwood Heartwood samples from five trees were delivered for determination of extractives. The samples were taken from trunk discs. The samples were taken from the inner (2 cm from the centre) and the outer part of the heartwood in the north–south and east–west directions (Fig. 2). The samples were treated as described for extraction of wood samples. Samples of heartwood and sapwood (10 mm × 5 mm × 30 mm) were treated as described under Section 2.4. A mixture of wood powder from several samples was used for the determination of the linear range, accuracy and precision. 2.7. GC–FID analysis The separation and quantification of the analytes were achieved using a gas chromatograph (Agilent 6890 Series, Agilent Technology, Wilmington, DE, USA) with a FID system, an auto sampler (Agilent 7683 Series) and a split/splitless injector (Agilent 7683 Series). The gas chromatograph was equipped with a CP-SIL 8 CB capillary column of 30 m × 0.25 mm I.D. and 0.25 ␮m film thickness with 5% phenyl- and 95%

The separation and identification of the analytes was performed using a GC8000 Top gas chromatograph from CE instruments, coupled with a Voyager Finnigan quadrupole mass spectrometer with the same type of column and temperature program used on the GC–FID system and with helium as carrier gas. The injections were performed manually. The gas chromatograph was equipped with a CP-SIL8CB low bleed/MS capillary column of 30 m × 0.25 mm I.D. (0.25 ␮m film thickness). Helium carrier gas was at a constant flow of 1.0 mL/min. The GC column temperature was programmed as follows: 10 ◦ C/min from 40 to 190 ◦ C, 4 ◦ C/min from 190 to 300 ◦ C and 300 ◦ C for 7.5 min. Injection volume 1 ␮L, The mass spectrometer was operated in electron ionization (EI) mode at 70 eV and the mass range was from m/z 40 to 500 with a scan time of 2.3 s. The transfer line was held at 250 ◦ C. 2.9. Determination of linear range, accuracy and precision The linear range (0.03–10 mg heptadecanoic acid/g wood sample, 16 levels) was determined with standard addition of heptadecanoic acid to wood samples. The samples were extracted and analyzed by GC–FID. The accuracy and precision were determined with three levels of standard addition of diethylstilbestrol ranging from 0.5–5 mg/g wood (n = 6). The method was evaluated with both peak area and height for accuracy. 2.10. Data handling and statistics Statistical analysis was performed using the SAS statistical program v. 8.0 (SAS Institute, USA) or SigmaPLOT v. 8.0 (SPSS ASC, The Netherlands). Identification of compounds was performed using the US National Institute of Standards and Technology (NIST) library and structures, version 98 unless otherwise stated. Peak area was used for quantification unless otherwise stated. 3. Results and discussion 3.1. Chromatographic considerations The Soxhlet acetone extraction of the wood samples gave a large number of wood extractives, which calls for a high

270

D. Ekeberg et al. / J. Chromatogr. A 1109 (2006) 267–272

Table 1 Data for the accuracy of the method based on standard addition of diethylstilbestrol (n = 6) based on the area of the peaks Standard addition (mg/g wood)

Found amount (mg/g wood)

Relative recovery (%)

RSD (%)

0.508 1.52 5.08

0.529 1.44 5.56

104 95 109

10 1 7

resolution method such as capillary GC. The detector responses for all the wood extractives were assumed to be equal to the detector response for heptadecanoic acid, which was used as internal standard (I.S.). The results are based on peak area, because this method gave better accuracy than the results for peak height, ±9% and ±19%, respectively. The accuracy of the method was measured as recovery of diethylstilbestrol based on standard addition. The precision of the method was based on analysis of one wood sample and 17 replicates. The data gave accuracy within ±9%. Data for accuracy are presented in Table 1. The analytical data gave a precision of 18% relative standard deviation or better, dependent on the analytes. Data for precision are presented in Table 2. The linear range of the method was tested and linearity was found up to 10 mg/g extractives in wood samples (r = 0.9994). The limit of detection (LOD), based on three times signal to noise ratio, was estimated to be 0.03 mg/g extractives in the wood samples and the limit of quantification (LOQ), based on 20 times signal-to-noise ratio, was estimated to be 0.17 mg/g extractives in wood samples. The identification of silylated wood extractives was achieved using computer matching of the mass spectra from the NIST library and by retention indices. For those mass spectra of extractives that were not found in the NIST library, reference compounds were analyzed with GC–MS for comparison. The samples contained a large number of wood extractives, and we therefore sought to identify only the wood extractives with a concentration higher than the LOQ in at least one of the samTable 2 Data for the precision of the method for extractives with amount above LOQ (n = 18) based on the area of the peaks Compound

Mean (mg/g)

±SD

Linolenic acid Podocarp-8(14)-en-15al,13a-methyl-13-vinyl Linoleic acid Oleic acid Unknown compound Pinosylvin monomethyl ether Pimaric acid Sandaracopimaric acid Pinosylvin Isopimaric acid Palustric acid Levopimaric acid Dehydroabietic acid Abietic acid Neoabietic acid Retinoic acid 7-Oxodehydroabietic acid

0.75 0.22 3.12 2.93 0.18 1.72 2.08 0.36 1.57 1.38 4.12 1.47 4.05 5.16 2.76 0.44 0.18

0.05 0.02 0.19 0.18 0.01 0.14 0.18 0.03 0.12 0.11 0.50 0.27 0.29 0.67 0.42 0.06 0.03

Table 3 Identification of wood extractives in Scots pine (Pinus sylvestris L.) TMS derivatives

tR (min)

Mach factor (%)

Peak number in Fig. 3

Hexadecanoic acidb Rhodoviolascina Heptadecanoic acid (IS) Linolenic acidb Podocarp-8(14)-en-15al,13amethyl-13-vinyla Linoleic acid Oleic acid Pinosylvin monomethyl etherb Pimaric acidb Sandaracopimaric acid Pinosylvinb Isopimaric acidb Diethylstilbestrolb Palustric acidb Arachidonic acid Dehydroabietic acid 11,14-Eicosadenoic acid Abietic acid Bis(2-ethylhexyl) phthalatea Retinoic acid 15-Hydroxy-dehydroabietic acid 7-Oxodehydroabietic acid Behenic acid 15-Hydroxy-7-oxodehydroabietic acid ß-Sitosterol

23.5 24.5 25.1 26.0 26.2

84 54 84 60 57

1 2 IS 4 6

26.4 26.5 28.2 28.4 28.7 28.9 29.0 29.3 29.4 29.8 29.9 30.1 30.6 32.5 32.7 33.4 33.7 34.3 36.9

97 80 96 94 94 96 90 89 92 22 93 40 95 29 61 89 96 67 88

7 8 11 12 13 14 nd 15 16 nd 18 19 20 nd 22 23 nd 24 nd

46.2

81

25

Identification was achieved using computer matching of the mass spectra with the NIST library or with mass spectra obtained from pure compounds. nd: not detected. a Not a TMS derivative. b Is also identified with a reference compound.

ples. In the retention area from 9 to 20 min many carbohydrates were identified. Carbohydrates occur in all living organisms and are not of interest in this method. Identified wood extractives in the retention area from 20 to 50 min are presented in Table 3 with retention time and probability based on the NIST library or mass spectra of standards. Fig. 3 shows a chromatogram in the retention area from 20 to 50 min of sample E, east direction, outer part of heartwood.

Fig. 3. Chromatogram of sample C, east direction, outer part of heartwood. The peak identification is in Table 3. Peaks 5, 9, 10, 17 and 21 were not identified. Peak 17 may be levopimaric acid TMS and peak 21 may be neoabietic acid TMS.

D. Ekeberg et al. / J. Chromatogr. A 1109 (2006) 267–272

271

Table 4 Multiple comparison test based on LS means differences (Tukey HSD, α = 0.05) on three different pre-treatments for extraction yields of pinosylvin monomethyl ether, pinosylvin, total resin acids and total fatty acids (levels not connected by same letter are significantly different) Treatment

Pinosylvin mono-methyl ether (mg/g wood)

Pinosylvin (mg/g wood)

Mean Fine powder Small particles Solid wood

A A B

1.69 1.66 1.06

Total resin acids (mg/g wood) Mean

A A B

0.61 0.61 0.41

Total fatty acids (mg/g wood)

Mean A B B

1.30 2.03 2.10

Mean A B C

0.80 1.78 2.63

n = 3 ± 2 std.

3.2. Results for evaluation of grinding process, pre-treatment of wood samples The different pre-treatments, (a) solid wood, (b) small particles and (c) powder, gave different yields for different extractives. The yields of pinosylvin and pinosylvin monomethyl ether were significantly higher for small particles (2 mm × 1 mm × 1 mm) and wood powder than for solid wood (5 mm × 10 mm × 30 mm), while the yield of resin acids was significantly higher for solid wood and small particles than for wood powder. The yield for fatty acids was significantly higher for the solid wood samples than for small wood particles, for which the yield was significantly higher than that for wood powder. The multiple comparisons between the three treatments (Table 4) were based on the statistical method called Tukey’s honest significant difference (Tukey’s HSD) (α = 0.05) [11]. The method was based on comparing pairs of means in balanced analysis of variance problems. A compromise was made between the best results of different extractives, and small particles were chosen for further analysis, because this pre-treatment gave good yields of pinosylvin, pinosylvin monomethyl ether, resin acids and free fatty acids, the most important extractives. 3.3. Distribution of extractives across heartwood and sapwood In the experiment where the distribution of extractives across heartwood and sapwood was measured, the two trees A and B gave different results. The results for tree A were corresponding to earlier work from Bergstrøm et al. [3] (Fig. 4, showing the results for pinosylvin and pinosylvin monomethyl ether), while the results for tree B were odd. Sample 8 in particular, was expected to have higher (about 4–5 mg/g wood) concentration of pinosylvin and pinosylvin monomethyl ether than samples from more central origin in the stem [3]. Based on this observation there is reason to believe that sample 8 for tree B may be a mixture of heartwood and sapwood. Since the scope is to describe the variation of pinosylvin content in heartwood, samples with a mixture of sapwood and heartwood are not discussed in this study. Samples 2 and 10 for tree B gave higher results for the total amount of extractives than expected. The samples for tree A showed that the total amount of extractives was higher in heartwood than in sapwood. The results for tree B did not show this so clearly. No pinosylvin or pinosylvin monomethyl

Fig. 4. The distribution of pinosylvin and pinosylvin monomethyl ether across heartwood and sapwood in trunk discs from (a) stem A and (b) stem B.

272

D. Ekeberg et al. / J. Chromatogr. A 1109 (2006) 267–272

ether was measured in the sapwood. Furthermore, the amounts of pinosylvin and pinosylvin monomethyl ether in the heartwood were highest in the outer part and decreased to the center of the trunk disc. 3.4. Determination of extractives in north-south and east-west directions of the heartwood The determination of extractives in the north–south and east–west directions of the heartwood was performed in the three trees C, D and E. A statistical test was performed on the total amount of extractives, on the amount of pinosylvin, the amount of pinosylvin monomethyl ether, the total amount of resin acids and the total amount of fatty acids. The statistical test was performed to see if there were significant differences in amount of extractives between the north–south and east–west directions of the heartwood. Furthermore, the differences in amount of extractives between the inner and outer parts of the heartwood were tested. In the statistical model the between-tree variation and the interactions between tree and direction, tree and part of heartwood, and direction and part of heartwood were also tested. The statistical model gave no significant effect of interaction. The results for the determination of extractives in the north–south and east–west directions gave no significant differences, yet there was a trend that showed a higher amount of extractives in the south direction compared to the other directions. The number of observations in this experiment was too small for certain determination of these differences. The test for variances in amount of extractives between the inner and outer parts of the heartwood showed no significant differences for total amount of extractives. On the other hand, the results for the amounts of pinosylvin and pinosylvin monomethyl ether, the total amount of resin acids and the total amount of fatty acids, showed significant differences between the inner and outer parts of the heartwood. The amounts of

pinosylvin and pinosylvin monomethyl ether in the outer part of the heartwood were higher than in the inner. The total amounts of resin acids and fatty acids in the outer part of the heartwood were lower than in the inner. The between-tree variation was significant for the total amount of extractives and for the amounts of the specified extractives pinosylvin, pinosylvin monomethyl ether, resin acids and fatty acids. Acknowledgements This study was made possible by financial support form the Norwegian Forest Research Institute and The Research Council of Norway. We are also grateful to Sigrun Kolstad for the pretreatment of the samples, Øystein Johnsen for statistical support and Nicholas Clarke for valuable assistance. References [1] L.E. Lindberg, S.M. Willfor, B.R. Holmbom, J. Ind. Microbiol. Biotechnol. 31 (2004) 137. [2] P.O. Flæte, A. Øvrum, Wood Focus Norway 25 (2002) 1. [3] B. Bergstrøm, G. Gustafsson, R. Gref, A. Ericsson, Trees – Structure and Function, Springer, Berlin, 1999. [4] J.H. Hart, J.F. Wardell, R.W. Hemingway, Phytopathology 65 (1975) 412. [5] J.H. Hart, in: J.W. Rowe (Ed.), Natural Products of Woody Plants II, Springer, Berlin, 1989, p. 861. [6] H. Erdtman, E. Ernnerfelt, Sven. Papperstidn. 47 (1944) 45. [7] L.L. Ingram Jr., M.C. Templeton, G.W. McGraw, R.W. Hemingway, J. Wood Chem. Technol. 20 (2000) 415. [8] J.W. Rowe, C.L. Bower, E.R. Wagner, Phytochemistry 8 (1969) 235. [9] B.B. Sithole, Appita J. 45 (1992) 260. [10] M. Ven¨al¨ainen, A.M. Harju, P. Sartanp¨aa¨ , P. Kainulainen, M. Tiitta, P. Velling, Wood Sci. Technol. 38 (2004) 109. [11] R. Christensen (Ed.), Analysis of Variance, Design and Regression. Applied Statistical Methods, CRC Press, London, 1996, p. 156.