l wC

ld

verAna

acce 8

The chemical characterisation of waterlogged archaeological wood is of fundamental importance to understand the degradation processes

structure which is likely to collapsewhen drying [14]. Due to thealmost complete loss of cellulosic components, the chemical

mass over 10,000), is insoluble in solvents and does not contain

for studying cellulose and lignin and also for investigating decayprocesses in archaeological wood [10,12].

During pyrolysis, the macromolecules are depolymerised byheat and the pyrolysis products can be identified by mass spec-

Available online at www.sciencedirect.com

(2characterisation of lignin is of primary importance in thediagnosis and conservation of waterlogged wood artefacts.The recovery, consolidation and preservation of waterloggedwooden artefacts, such as shipwrecks and archaeological objectsrecovered from underwater environments, is a particularly ar-duous conservation problem. Under favourable conditions of lowtemperature and low oxygen availability, wood artefacts cansurvive underwater in a surprisingly good state. Nevertheless, ithas been shown that some species of anaerobic bacteria canslowly degrade waterlogged wood even under near anoxicconditions, mainly by eroding the polysaccharides as a source ofnutrients. This leads to the formation of pores and cavities filledwith water, and transforms the wood into a soft and fragile

either chains of repeating units or easily hydrolysable bonds [5].These chemicalphysical properties make the chemical analysisof lignin extremely complex. Infrared spectroscopy [6,7], thermo-gravimetric methods [8], solid state 13C-NMR [9,10] andpyrolysis-gas chromatography based procedures [1014] havebeen adopted to deal with this problem in degraded or agedwood.

The use of pyrolysis gas chromatographic techniques in com-bination with mass spectrometry has proven to be of particularinterest in the study of macromolecules such as lignin, celluloseand hemicellulose. In particular, Py-GC/MS provides useful dataon the chemical structure of wood components, and has been usedpreliminary results, presented in this study, demonstrate the feasibility and the potential of DE-MS as a reproducible and rapid screening method forarchaeological waterlogged wood samples. 2007 Elsevier B.V. All rights reserved.

Keywords: Archaeological waterlogged wood; Lignin; Direct exposure electron ionisation mass spectrometry (DE-MS); Principal component analysis (PCA)

1. Introduction Lignin, a cross-linked poly-phenolic macromolecule (molecularundergone bywooden objects and consequently to develop suitable consolidation and conservation procedures. Lignin extracted from archaeologicalwaterlogged wood samples was characterized using direct exposure electron ionisation mass spectrometry (DE-MS). DE-MS achieves a massspectral fingerprint of the sample in a few minutes, avoiding any chemical pre-treatment and requiring only few micrograms of material.

Mass spectral data were put in relation to the chemical composition of lignin and evaluated by means of principal component analysis (PCA). TheAnalysis of lignin from archaeologicamass spectrometry (DE-MS) and P

F. Modugno a,, E. Ribechini a, M. Caa Dipartimento di Chimica e Chimica Industriale, Uni

b Soprintendenza ai Beni Archeologici per la Toscana, Laboratorio di

Received 3 August 2007;Available onlin

Abstract

Microchemical Journal 88 Corresponding author. Tel.: +39 050 2219303.E-mail address: frances@dcci.unipi.it (F. Modugno).

0026-265X/$ - see front matter 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.microc.2007.11.010aterlogged wood by direct exposureA evaluation of mass spectral data

erisi a, G. Giachi b, M.P. Colombini a

sit di Pisa, Via Risorgimento 35, I-56126 Pisa, Italylisi-Centro di Restauro, Largo del Boschetto 3, I-50135 Firenze, Italy

epted 21 November 2007January 2008

008) 186193www.elsevier.com/locate/microctrometry, after gas chromatographic separation. Especially in thecase of lignin, pyrolysis produces a variety of phenolic products,which consist of guaiacyl and syringyl structural units with an

aliphatic side chain in position 4 of the aromatic ring [15]. It isassumed that pyrolysis products represent, to a greater or lesserdegree, the structural units that make up the macromolecule.

The aim of this study was to use a direct in-source Py-massspectrometric technique, namely direct exposure electron ioni-sation mass spectrometry (DE-MS), to characterise archaeologi-cal waterlogged wood and lignin extracted from archaeologicalartefacts, in order to explore the potential of this technique toretrieve chemical information on archaeological wood materialsand to evaluate decay processes.

In DE-MS the sample undergoes a resistively controlledheating on a rhenium filament within the ion source of the mass-

spectrometer. The sample components are desorbed or, in the caseof macromolecules, pyrolysed over the heating range, and areionised and analysed as a function of time. DE-MS requires verysmall amounts of untreated sample (in the order ofmicrograms), itcan be performed rapidly (a few minutes for analysis and dataacquisition), and its sensitivity has been shown to be suitable foridentifying materials in the archaeometric field, with specificapplications in the analysis of wood-derived materials such asterpenic resins [16,17]. Direct and rapid analytical methods thatavoid sample pre-treatment are particularly welcome in thecultural heritage field, where the availability of sample is oftenlimited and contamination must be avoided.

187F. Modugno et al. / Microchemical Journal 88 (2008) 186193Fig. 1. DE mass spectra of a) lignin extracted from archaeological wood IS, identified as Pinaceae, and b) lignin extracted from spruce sound wood.

2.2. DE-MS

Samples were analysed by depositing a few particles ofpowder directly on the direct exposure probe filament by meansof a capillary. Each sample was analysed in duplicate.

The instrumentation (Thermo Electron Corporation, USA)was made up of a Direct Probe Controller and a Direct ExposureProbe (rhenium filament, current programmed mode 0 mA to1000 mA in 2 s then 1000 mA for 60 s), coupled with a Polaris Qion trap external ionisation mass spectrometer (electron impactionisation 70 eV). The source temperature was at 230 C. Themass spectrometer was scanned over an m/z range of 501000.

2.3. Data analysis

Unsupervised pattern recognition analysis of DE-MS massspectral data corresponding to the mass range 50500 m/z wasperformed by principal component analysis (PCA, Nipals algo-rithm) on the covariance matrix of centered data, after rownormalisation of the full 501000 spectra. The region 50500m/zwas selected because it contains all the pyrolysis fragmentscorresponding to lignin monomers and dimers. The software usedwas XLSTAT 6.0 (Addinsoft, Paris, France).

3. Results and discussion

ical Journal 88 (2008) 186193The few applications of in-source pyrolysis-mass spectrometry(direct temperature resolved mass spectrometry, DTMS) to woodand lignin described in the literature [18,19], have highlighted thatit is possible to relate mass spectral data to the chemical com-position of lignin and of wood in sediments [18]. Althoughdirectly combining pyrolysis with mass spectrometry does notoffer the detailed chemical information accomplished by Py-GC/MS, it nevertheless achieves a mass spectral fingerprint of thesamples within a few minutes. The application of pattern analysisbased on principal components (PCA) enables the mass profilesobtained for the various samples to be compared quantitatively inan easily readable manner. It also highlights similarities anddifferences between samples. Moreover, the examination ofloading plots permits the differences to be correlated to specificchemical features. This analytical approach is suitable for fastscreening even a large number of samples. It thus has greatpotential, when applied to wood artefacts of large dimensions,such as shipwrecks, where it can be used to monitor the state ofconservation of the wood in various regions of the timbers.

This first application of DE-MS to archaeological woodshows that it is a fast fingerprint tool that is able to discriminatebetween hardwood and softwood archaeological wood, and issensitive to differences in the chemical structure of lignin.

2. Materials and methods

2.1. Samples

The lignin was prepared and donated by ProfessorM. Orlandiof the Department of Environmental and Earth Sciences of theUniversity of Milano Bicocca (Milan, Italy). Lignin extractionand isolation was carried out using a recently published pro-cedure [9], based on a modification of the milled wood methoddeveloped by Holmbom et al. [20]. The lignin samples werederived from archaeological wood samples of different ages andorigins:

- two samples from bollards excavated in the Site of AncientPisa Ships (Pisa, Italy), provided by the ArchaeologicalSuperintendence of Tuscany: D (Ulmus sp), SR (hardwood,unknown species). The archaeological artefacts from the siteare dated to a period between the 4th century BC to the 2ndcentury AD;

- a wooden piece from a shipwreck found under shallowwaters(0.5 m deep) in Tantura Lagoon (Haifa, Israel) and dated 89th century AD, provided by the Institute ofMaritime Studiesof the University of Haifa: sample IS. Previous analysis hadidentified the wood as a conifer, probably from the Pinaceaefamily [10];

- a wooden piece from the shipwreck Epave du GrandConglou recovered near Marseille (France) and dated 2ndcentury BC, provided by Prof. I.D. Donato (Department ofChemistry of the University of Palermo, Italy): B44. Thewood was identified as oak (Quercus sp.).

188 F. Modugno et al. / MicrochemRecent lignin from spruce (soft wood) and birch (hard wood)extracted with the same procedure was used as references.All the samples examined showed the presence of a singlepeak in the total ion current (TIC) profile. The peaks show avariable degree of broadness and in some cases the presence of

Table 1Attribution of m/z peaks observed in archaeological lignin DE mass spectra

m/z Derivation Type of monomer G:guaiacyl S: siringyl

124 Guaiacol (M+) coniferylalcohol (M+-C3H4OH) G138 Methylguaiacol (M+) G150 Vinylguaiacol (M+) G137 Ethyguaiacol (M+ -CH3), Propylguaiacol

(M+ -CH2CH3), Coniferylalcohol (M+-C2H2OH)

G

164 Propenylguaiacol (M+) G151 Acetylguaiacol (M+-CH3), Vanillin (M

+-H),Propylguaiacol (M-CH3)

G

152 Vanillin (M+)154 Syringol (M+) S168 Methylsyringol (M+) S167 Ethylsyringol (M+-CH3), Propylsyringol

(M+-CH2CH3), Sinapylalcohol (M+-C2H2OH)

S

178 Propenyl-3-one-guaiacol (M+),prop-2-enalguaiacol (M+)

G

180 Vinylsyringol (M+) S194 Propenylsyringol (M+) S181 Acetylsyringol (M+-CH3) S208 Propenalsiryngol (M+) S210 Sinapylalcohol (M+) S272 Stilbene-type dimer (M+) GG dimer303 Stilbene-type dimer (M+) GS dimer332 Stilbene-type dimer (M+) SS dimer

+358 -resinol type dimer (M ) GG dimer388 -resinol type dimer (M+) GS dimer418 -resinol type dimer (M+) SS dimer

shoulders. This is probably due to the polymeric nature of ligninand the degree of the polymerization of lignin samples collectedfrom the various species of wood.

The mass spectrum is obtained by time-integration of themain peak in the total ion current (TIC) profile. The massspectra show high complexity, as expected in the direct massspectrometric analysis of the complex mixture of productsformed in the pyrolysis of lignin.

Themass spectra of lignin sample IS from the shipwreck fromIsrael (softwood) is shown in Fig. 1a and compared with themass spectra of the reference spruce lignin (Fig. 1b). Both thespectra are characterised by the occurrence of peaks indicative ofa guaiacyl lignin: m/z 124, corresponding to the molecular peak

of guaiacol (2-methoxy-phenol); m/z 137 ([guaiacol+CH2]+),

which could derive from several compounds formed in thepyrolysis of guaiacyl lignin including ethylguaiacol, propyl-guaiacol and coniferyl alcohol, by the loss of a methyl radical, anethyl radical and a C2H2OH

. radical respectively; m/z 151(guaiacol+CH2CH2

+); m/z 152, corresponding to the molecularion of vanillin. The peak at m/z 272 corresponds to the molecularpeak of a guaiacyl-guaiacyl dimer compound with a stilbenestructure (3,3'-dihydroxy-4,4'-dimethoxy-stilbene).

Table 1 lists the fragments identified and the most probableattribution, based on the mass spectra of guaiacyl and siringylcompounds determined by Py-GC/MS [10] and available fromthe literature [18].

189F. Modugno et al. / Microchemical Journal 88 (2008) 186193Fig. 2. DE mass spectra of a) lignin extracted from archaeological wood SR , unknown specie, and b) lignin extracted from sound birch wood.

arch

190 F. Modugno et al. / Microchemical Journal 88 (2008) 186193As can be seen in Figs. 2, 3 and 4, the mass spectra of birchlignin (hardwood) and of samples SR, B44 and D show thepresence of peaks characteristic of guaiacyl-syringyl lignin atm/z 124, 137, 151 deriving from guaiacyl monomers, and at m/z167, 181, 208, 210 deriving from syringyl monomers. Inparticular, the peaks at m/z 167 (syringol+CH2

+) and at m/z 181+

Fig. 3. DE mass spectra of lignin extracted from(syringol+CH2CH2) correspond to the main ion fragmentsderiving from sinapyl alcohol and alkylsyringol compounds,while the peak at m/z 210 corresponds to the molecular ion of

Fig. 4. DE mass spectra of lignin extracted from asinapyl alcohol. In the mass spectra of sample D (Fig. 4), theoccurrence of peaks at m/z 196 and at m/z 153 is significant.These peaks could derive from the demethylation reaction ofalkylsyringyl monomeric units. This reaction leads to the for-mation of corresponding catechol derivatives, whose fragmen-tation should form ion fragments at m/z 196 and at m/z 153. In

aeological wood B44, identified as Quercus sp.fact, the peaks at m/z 196 and at m/z 153 correspond to a loss of14 uma from peaks at m/z 210 and m/z 167, respectively. Due tothe fact that demethylation of alkylsyringyl units is a recognized

rchaeological wood D, identified as Ulmus sp.

icadecay index for lignin [12], the occurrence of peaks relative tolignin demethylation products can be considered as a way to

Fig. 5. PCA score plot of PC1 and PC2 of DE mass spectral data, accounting for78.5% of total variance. IS: conifer lignin from a 89th century AD shipwreckrecovered near Haifa (Israel); B44: oak lignin from a 2nd century BC shipwreckrecovered near Marseille (France); D: elm lignin from a bollard pertinent to theSite of Ancient Pisa Ships (Pisa, Italy); SR: D: hardwood lignin from a bollardpertinent to the Roman Harbour of san Rossore (Pisa, Italy); spruce: referencespruce lignin; birch: reference birch lignin.

F. Modugno et al. / Microchemestablish the degree of degradation.The ratio between the relative abundance of peak at m/z 167,

the most abundant from syringyl monomers, and that of peak atm/z 137, deriving from guaiacyl monomers, can be consideredas a rough indication of the S/G ratio for the lignin sample. Ascan be seen from the mass spectra, softwood samples do notcontain a peak at m/z 167, while hardwood samples showvariability in this ratio depending on the species of wood. In thelignin mass spectra, m/z fragments deriving from residualcellulose and hemicellulose are not evident, demonstrating theefficiency of the extraction and purification procedure adoptedto obtain lignin. In particular, if polysaccharides had been pre-sent in not negligible amounts, on the basis of the literature [19]and of the analysis of reference cellulose, the following m/zfragments would have been observed: 65, 73, 85, 79, 97, 127.

In all the mass spectra obtained for hardwood and softwoodlignin, in addition to the peaks deriving from the guaiacyl andsiringyl monomer structures, peaks attributable to ion fragmentsderiving from dimeric structures of the lignin macromoleculewere observed. Amongst these, the most abundant was at m/z272, which corresponds to the molecular ion of the GG dimerdihydroxydimethoxy-stilbene. In samples from hardwood spe-cies, peaks at m/z 302 and m/z 332 attributable to the homologueSG and SS dimers, were also observed. Moreover, the series ofpeaks atm/z 358, 388 and 418,which differ by 30Da (OCH3), canbe attributed to GG, GS and SS dimers with - resinol typestructures [18].

Due to the complexity of the mass spectra obtained in the directPy-MS analysis of lignin, principal component analysis (PCA)wasused as a pattern recognition technique to quantitatively comparethe mass spectra obtained and to highlight differences andsimilarities between samples and correlations between variables.

The DE mass spectra obtained from the archaeological ligninexamined and those from the two reference lignin extracted fromwood were compared by means of PCA. The data matrix wasconstituted by the mass spectra of the samples (in replicates) inthe range m/z 50500, where the intensities of each m/zfragment was expressed as percentage of the sum of theintensities of all the m/z fragments in the spectra (rownormalisation). The data matrix was column-centered and thecovariance matrix was used for PCA. Fig. 5 shows the scatterplot corresponding to the first two PCs, accounting for 78.5% ofthe total variance. The following couples of similar samples canbe highlighted: spruce lignin and IS, birch lignin and DD1, andSR and B44.

The loading plot of PC1 (Fig. 6) on the basis of the loadingsshows that this axis of variance concerns the ratio betweensyringyl and guaiacyl markers. The PC1+ is indicative of syringylunits and the PC1- of guaiacyl units. Actually, the two softwoodsamples, spruce and IS, scored lowest on PC1, because theycontained only guaiacyl lignin, and they were well differentiatedfrom the hardwood samples. The archaeological sample D1 wassimilar to the modern birch lignin, at an intermediate guaiacyl/syringyl ratio, while the archaeological samples B44 and SRformed a separate cluster with high syringyl content.

The scores of the samples on the PC3/PC1 map, shown inFig. 7, enhance the differences between archaeological (PC3+)and modern lignin (PC3).

It seems that PC2 and PC3 might be indicative of the degreeof conservation, in fact the reconstructed mass spectra of PC2and PC3 show that both the axes have significant loadings form/z 196 and m/z 153, which can be attributed to the frag-mentation catechol derivatives formed by the demethylation ofthe alkylsyringyl units. Peaks at m/z 153 and 196 are wellevident in the mass spectra of sample D. Moreover, PC3 showeda positive loading for m/z 110, corresponding to the molecularmass of catechol, which is also considered as a lignin de-gradation marker [12]. Neverthless, conclusions cannot bedrawn from PC3 values because this PC3 has a low significance(only 9% of total variance) and in fact shows a non negligiblevariability between replicated samples. The analysis of a highernumber of samples and the comparison of DE-MS data with Py-GC/MS data will permit to better evaluate the role of the for-mation of catechols in the degradation processes affecting ligninin waterlogged environment [21].

The results show that the differences in the mass spectra arerelated not only to the difference between guaiacyl lignin(spruce and IS) and guaiacyl/siringyl lignin (other samples), butalso to the ratio between guaiacyl and syringyl components andto the degree of demethylation undergone by lignin monomersin the course of degradation.

4. Conclusions

191l Journal 88 (2008) 186193This preliminary study has highlighted that DE-MS is a fastfingerprint tool which is able to discriminate between hardwood

ica192 F. Modugno et al. / Microchemand softwood archaeological wood and which is sensitive todifferences in the chemical structure of lignin. The mass spectraobtained can be quantitatively evaluated and compared bymeans of principal component analysis, and this method would

Fig. 6. Loading plots ofl Journal 88 (2008) 186193seem to be suitable for studying degradation processes in lignincontained in archaeological wood.

Further comparative investigations are in progress. They in-volve a larger number of archaeological and reference samples,

the first three PCs.

[4] L. Campanella, M. Tomassetti, R. Tomellini, Thermoanalysis of ancient,fresh and waterlogged woods, Journal of Thermal Analysis 37 (1991)19231932.

[5] Y. Lin, C.W. Dence (Eds.), Methods in Lignin Chemistry, Springer-Verlag,Berlin, 1992.

[6] K. Fackler, C. Gradinger, B. Hinterstoisser, Lignin degradation by white

193F. Modugno et al. / Microchemical Journal 88 (2008) 186193the aim being to establish the value of this technique in deter-mining the extent of degradation of lignin.

Fig. 7. PCA score plot of PC1 and PC3 of DE mass spectral data, accounting for72.2% of total variance. IS: conifer lignin from a 89th century AD shipwreckrecovered near Haifa (Israel); B44: oak lignin from a 2nd century BC shipwreckrecovered near Marseille (France); D: elm lignin from a bollard pertinent to theRoman Harbour of San Rossore (Pisa, Italy); SR: hardwood lignin from abollard pertinent to the Roman Harbour of San Rossore (Pisa, Italy); spruce:reference spruce lignin; birch: reference birch lignin.Acknowledgments

The authors wish to thank Dr. Y. Kahanov (University ofHaifa, Israel), Prof. I. Donato (University of Palermo) forhaving provided the archaeological samples, and Prof. M.Orlandi (University of Milano Bicocca) for having extractedlignin from archaeological woods. Economic support was fromthe Italian MIUR findings (PRIN Cofin05).

References

[1] P. Hoffmann, M.A. Jones, Structure and degradation processes forwaterlogged archaeological wood, in: R.M. Rawell, R.J. Barbour (Eds.),Archaeological Wood, Chemistry, Properties and Preservation, Advancesin Chemical Series, 225, American Chemical Society, Washington, 1990,pp. 3565.

[2] B. Hafors, Conservation of the Swedish warship Vasa from 1628, VasaStudies 18, Swedish National Maritime Museums, Stockholm, Sweden,2001.

[3] R.A. Blanchette, T. Nilsson, G.F. Daniel, A. Abad, Biological degradationof wood, in: R.M. Rawell, R.J. Barbour (Eds.), Archaeological Wood,Chemistry, Properties and Preservation, Advances in Chemical Series, 225,American Chemical Society, Washington, 1990, pp. 141174.rot fungi on spruce wood shavings during short-time solid-state fermen-tations monitored by near infrared spectroscopy, Enzyme and MicrobialTechnology 39 (2006) 14761483.

[7] B. Barker, N.L. Owen, Identifying softwoods and hardwoods by infraredspectroscopy, Journal of Chemical Education 76 (1999) 10761079.

[8] I. Crina, A. Sandua, M. Brebub, C. Lucac, I. Sandud, C. Vasileb,Thermogravimetric study on the ageing of lime wood supports of oldpaintings, Polymer Degradation and Stability 80 (2003) 8391.

[9] M. Bardet, M.F. Foray, Q.K. i Tra n, High-resolution solid-state CPMASNMR study of archaeological woods, Analytical Chemistry 74 (2002)43864390.

[10] M.P.Colombini,M.Orlandi, F.Modugno, E.L. Tolppa,M. Sardelli, L. Zoia,C.Crestini, Archaeologicalwood characterisation by PY/GC/MS,GC/MS,NMRand GPC techniques, Microchemical Journal 85 (2006) 164173.

[11] C. Saiz-Jimenez, J.W. de Leeuw, Pyrolysis-gas chromatography-massspectrometry of isolated, synthetic and degraded lignins, Organic Geochem-istry 6 (1984) 417422.

[12] P.F. van Bergen, I. Poole, T.M.A. Ogilvie, C. Caple, R.P. Evershed, Evidencefor demethylation of syringyl moieties in archaeological wood using pyro-lysis-gas chromatography/mass spectrometry, Rapid Communications inMass Spectrometry 14 (2000) 7179.

[13] J.C. Del Rio,M. Speranza,A. Gutierrez,M.J.Martinez, A.T.Martinez, Ligninattack during eucalypt wood decat by selected basidiomycetes: a Py-GC/MSstudy, Journal of Analytical and Applied Pyrolysis 64 (2002) 421431.

[14] V. Vinciguerra, A. Napoli, A. Bistoni, G. Petrucci, R. Sgherri, Wood decaycharacterization of a naturally infected London plane-tree in urban environ-ment using Py-/GC/MS, Journal of Analytical and Applied Pyrolysis 78(2007) 228231.

[15] C. Saiz-Jimenez,W.De Leeuw, Lignin pyrolysis products: their structures andtheir significance as biomarkers, Organic Geochemistry 10 (1985) 869876.

[16] M.P. Colombini, F. Modugno, E. Ribechini, Direct exposure electronionization mass spectrometry and gas chromatography/mass spectrometrytechniques to study organic coatings on archaeological amphorae, Journalof Mass Spectrometry, 40 (2005) 675687.

[17] F. Modugno, E. Ribechini, M.P. Colombini, Chemical study of triterpenoidresinous materials in archaeological findings by mean of direct exposureelectron ionization mass spectrometry (DE-MS) and gas chromatographymass spectrometry (GC/MS), Rapid Communications in Mass Spectro-metry 20 (2006) 17871800.

[18] E.R.E. Van der Hage, M.M. Mulder, J.J. Boon, Structural characterizationof lignin polymers by temperature-resolved in-source pyrolysis-massspectrometry and Curie-point pyrolysis-gas chromatography/mass spectro-metry, Journal of Analytical and Applied Pyrolysis 25 (1993) 149183.

[19] C. Saiz-Jimenez, J.J. Boon, J.I. Hedges, J.K.C. Hessels, J.W. De Leeuw,Chemical characterisation of recent and buried wood by analytical pyro-lysis. Comparison of pyrolysis data with 13C NMR and wet chemical data,Journal of Analytical and Applied Pyrolysis, 11 (1987) 437450.

[20] B. Holmbom, P. Stenius, Analytical methods, in: P. Stenius (Ed.), ForestProduct Chemistry, Papermaking Science and Technology, vol. 3, FapetOy, Helsinki, Finland, 2000, pp. 107172.

[21] M.P. Colombini, J. Lucejko, F.Modugno,M. Orlandi, E-L. Tolppa, L. Zoia,Study of degradation processes of lignin in archaeological waterloggedwood by pyrolytic techniques and NMR spectroscopy, in preparation.

Analysis of lignin from archaeological waterlogged wood by direct exposure mass spectrometry (D.....IntroductionMaterials and methodsSamplesDE-MSData analysis

Results and discussionConclusionsAcknowledgmentsReferences