The warship Vasa:
A Study of Lignin, Extractives and Acids in the oak wood
DINA DEDIĆ
Licentiate thesis
KTH Royal Institute of Technology Department of Fibre and Polymer Technology Division of Wood Chemistry and Pulp Technology
Stockholm 2013
Fibre and Polymer Technology KTH Royal Institute of Technology SE-100 44 Stockholm
Sweden
TRITA-CHE Report 2013:19 ISSN 1654-1081
ISBN 978-91-7501-678-8
©Dina Dedic Stockholm 2013
Tryck: US-AB, Stockholm 2013
3
Abstract
The oak wood timbers of the 17th century Swedish warship Vasa are weak. The weakening has been attributed to cellulose degradation, which is more extensive in the interior of the timbers compared to the surface regions. Further, the mechanism of cellulose degradation was attributed to oxidative reactions involving iron as a catalyst. In this work, the non-cellulosic wood components (lignin and extractives) in the wood of the Vasa have been studied in order to assess the level of degradation possibly caused by oxidative reactions. Because the interior of the wood timbers is more acidic and its cellulose more
depolymerized than at the near surface regions, a general study of organic acid- and iron concentrations with respect to different depths from the surface was also performed.
Characterization of lignin in the wood of the Vasa was done by means of wet chemical degradation (thioacidolysis) and subsequent GC-MS analysis of the degradation products, as well as CP/MAS 13C NMR spectroscopy of the wood.
Dry acetone- and water extracts of the wood were analyzed by 13C NMR spectroscopy and MALDI-TOF mass spectroscopy in order to study the presence of gallo- and ellagitannins. No severe lignin degradation and no discernible amounts of hydrolysable tannins in the oak wood of the Vasa were detected, indicating that oxidative reactions are not the primary route to cellulose depolymerization.
High amounts of oxalic acid (analyzed by HPIEC) and a low pH have been found in the interior of the wood timbers, supporting acid hydrolysis as the main mechanism of cellulose depolymerization. Analysis of the iron distribution using ICP AES shows that iron is most abundant in the surface of the timbers and decreases as the concentration of oxalic acid increases. Experimental work also shows that some iron species (rust) in the Vasa neutralize oxalic acid, thereby protecting the surface wood from acid hydrolysis.
Sammanfattning
Ekbalkarna i det svenska regalskeppet Vasa från 1600-talet är försvagade.
Försvagningen har tillskrivits nedbrytning av träets cellulosa, som är mer omfattande i balkarnas inre än i områden nära ytan. Mekanismen för
nedbrytningen tros vara oxidativa reaktioner med järn som katalysator. I detta arbete har lignin och extraktivämnen i Vasas ekträ studerats för att bedöma graden av nedbrytning som kan vara orsakad av oxidativa reaktioner. Eftersom balkarnas inre är surare och dess cellulosa mer nedbruten än vid ytområdena, studerades också koncentrationerna av organiska syror och järn med avseende på olika avstånd från ytan.
Karakterisering av lignin i Vasas trä gjordes med hjälp av kemisk nedbrytning (thioacidolys) och efterföljande GC-MS analys av nedbrytningsprodukterna samt CP/MAS 13C NMR-spektroskopi av trämjöl. Torkade aceton- och vattenextrakt av träet analyserades med 13C NMR-spektroskopi och MALDI-TOF
masspektroskopi för att studera förekomsten av gallo-och ellagtanniner. Ingen påtaglig nedbrytning av lignin och inga mätbara mängder tanniner kunde detekteras i Vasas trä, vilket pekar på att oxidativa reaktioner inte är den huvudsakliga orsaken till nedbrytning av cellulosan.
Höga mängder av oxalsyra (som analyserades med HPIEC) och ett lågt pH- värde har hittats i balkarnas inre, vilket stödjer sur hydrolys som den
huvudsakliga mekanismen för nedbrytningen. Analysen av järndistribution med ICP AES visar att järnkoncentrationen är högst i ytan av balkarna och minskar inåt i samband med att koncentrationen av oxalsyra ökar. Experimentellt arbete visar också att vissa järnföreningar (rost) som finns i Vasas trä neutraliserar oxalsyra och skyddar därmed träet i ytan från sur hydrolys.
5
List of Publications
1. Dedic, D., Iversen, T. and Ek, M. (2012) “Cellulose degradation in the Vasa:
The role of acids and rust”, Studies in Conservation. DOI 10.1179/2047058412Y.0000000069
2. Dedic, D., Sandberg, T., Iversen, T., Larsson, T. and Ek, M (2013) “Analysis of lignin and extractives in the oak wood of the 17th century warship Vasa”, Manuscript
Other publications not included in the thesis:
Dina Dedic, Tommy Iversen, Teresia Sandberg and Monica Ek, “Chemical Analysis of Wood Extractives and Lignin in the Oak wood of the 380 year old Swedish warship Vasa”, Proceedings of the 16th ISWFPC, Tjanjin, China, 2011 Dina Dedic, Tommy Iversen and Monica Ek “Degradation Reactions in the Vasa”, International conference Shipwrecks 2011, Stockholm, Sweden, 2011
Abbreviations
PEG – Polyethylene Glycol
Klason Lignin – Acid Insoluble Lignin S – Syringyl
G – Guaiacyl
GC-MS – Gas Chromatography Mass Spectroscopy
CP/MAS 13C NMR – Cross Polarization/Magic Angle Spinning Nuclear Magnetic Spectroscopy
13C NMR – Nuclear Magnetic Spectroscopy
MALDI-TOF MS – Matrix Assisted Laser Desorption Ionization - Time Of Flight Mass Spectroscopy
ICP AES – Inductively Coupled Plasma Atomic Emission Spectroscopy HPIEC – High Performance Ion Exchange Chromatography
7
Content
Chapter 1: Introduction ... 8
The chemical history of the Vasa ... 8
Degradation and weakening ... 9
Purpose of research ... 11
Objectives of research ... 11
Research approach ... 11
Chapter 2: Lignin ... 13
Chapter 3: Extractives ... 19
Chapter 4: Acids and Elements ... 23
Acid concentration ... 23
Elemental analysis ... 24
Neutralization of oxalic acid by rust ... 25
Chapter 5: Concluding remarks ... 27
Acknowledgements ... 28
References ... 29
Chapter 1: Introduction
Wood has served the human civilization for millennia, due to its availability and unique mechanical properties. Modern finds of archeological wooden artifacts tell a story about the historic culture that had created them and used them, about the climate and environment they have experienced since then and about the wood deterioration process itself. It is therefore of greatest importance that archeological wooden artifacts found in our time, also stay preserved for future generations.
Archeological wood differs from fresh wood in many ways, but the differences are equally substantial between archeological woods that have been preserved in different environments. Surfaces of dry, wooden artifacts may be exposed to weathering and insects on one hand and fungi and bacteria on the other hand depending on the surroundings the wood has been in contact with. A marine environment, in turn, presents a completely new set of issues conservators and wood scientists are faced with. For example, water causes wood cells to swell and become accessible to degrading microorganisms, in worst cases leading to moisture contents of 400% (Florian 1990), and enables diffusion of substances into and out of the wood material. The artifact is also exposed to a number of microenvironments within the marine environment, and is affected by salinity of the water, type of sediment, temperature, light and presence of oxygen.
Perhaps the mentioned concerns of the marine environment seemed not as vital to the crew of conservators and scientists who tended to the 17th century Swedish warship Vasa at the time of her salvage in 1961, as they would to the crew that has been studying the ship´s wood in an effort to understand the process of degradation that has taken place during the last 50 years. Since the humid summer of 2000 when alarming stains of yellow and acidic salts appeared on the hull of the Vasa, researchers in the fields of inorganic chemistry,
conservation, mechanics and wood chemistry have been involved in projects aiming to characterize the wood and to assist in designing suitable
conservation treatments.
The chemical history of the Vasa
The Vasa sank on her maiden journey in the Stockholm harbor in 1628 and lay submerged in the dirty, brackish water for 333 years. There, it was preserved in the dark, cold and anoxic environment. Because the brackish water of the Baltic Sea was not host to the wood degrading shipworm (Teredo navalis) normally
9 present in the oceans, the ship´s oak wood stayed physically intact apart from the outermost centimeters that were microbiologically degraded. Other wooden shipwrecks, such as the Mary Rose, a 16th century British warship that sank outside the coast of Isle of Wight, were less fortunate and were only preserved in the parts that were buried in the ocean sediment.
During the time the Vasa lay submerged, high amounts of iron and sulfur compounds had diffused into the wood along with other inorganic compounds and ions present in the seawater. The iron originated from cannon balls and bolts that held the ship´s timbers in place and the sulfur from the activity of sulfur reducing bacteria. Easily soluble compounds such as low molecular weight organic acids and extractives were, in turn, able to diffuse out of the wood. A part of the cellulosic material in the wood´s surface had disappeared due to erosion bacteria and physical wearing, leading to a porous wood surface and hence a buildup of the iron and sulfur compounds. Elemental analyses discussed in Chapter 4 and in other publications (Sandstrom et al. 2002; Almkvist 2008;
Fors 2008), respectively, generally show a gradient of iron and sulfur with high concentration in the surface and decreasing concentration the deeper into the timbers we measure.
After the salvage the Vasa was impregnated with polyethylene glycol (PEG) of various molecular weights to prevent crack formation and deformation during drying. The current distribution of PEG has the same pattern as the
aforementioned inorganic compounds, i.e. high abundance (both high and low molecular weight PEG) in the surface of the wood timbers and low to minimal in the interior (generally consisting of low molecular weight PEG). The conservation treatment with PEG continued for 17 years, during which
antifungal and antibacterial agents (sodium pentachlorophenolate, boric acid and borax) were added at occasions (Håfors 2010).
Degradation and weakening
It was not until the yellow, acidic stains appeared on the hull of the Vasa that the question of the wood´s condition was raised. Initially, focus was directed to the surfaces of the oak timbers and the distribution of oxidized and reduced sulfur compounds. It was shown that reduced sulfur compounds such as elemental sulfur were abundant in the surface areas where they were suspected to be a source of sulfuric acid that was formed through an oxidative reaction in which the iron in the wood acts as a catalyst (Sandstrom et al. 2002; Fors 2008).
Molecular weight analysis of the cellulose in the ship´s wood showed that the cellulose was indeed depolymerized, but a surprising outcome of the analysis indicated that the cellulose was, in fact, more degraded far from the surface of
the timbers, at a depth of 6-10 cm (Lindfors et al. 2008; Bjurhager 2011). It was also shown that the tensile strength of the wood followed the same pattern as the degraded holocellulose (cellulose and hemicellulose), i.e. the wood was weaker in the core of the timbers than in the near surface areas (Bjurhager et al.
2012). The overall loss of tensile strength was estimated to ca. 50% with as much as 80% in the most affected areas. For an essentially self-supporting ship that weighs more than 1000 tons, these results were worrying.
The production of sulfuric acid in the surfaces of the hull was thus no longer an immediate threat, since the near surface areas were shown to be in better condition (in terms of cellulose degradation and tensile strength) than the interior parts of the timbers. Focus was instead shifted to understanding why the near surface areas were better preserved and how they differed from the weaker interior wood. Researchers suspected that oxidative reactions involving iron and acid hydrolysis by oxalic acid, the latter being found in high amounts in the interior wood, might resolve some of the questions. It was also believed that the reduced sulfur compounds found in the surface of the timbers were, in fact, suppressing oxidative reactions through their antioxidant properties and that this was one of the reasons for the lower cellulose depolymerization in these areas (Almkvist and Persson 2008b; Almkvist and Persson 2008a; Fors 2008; Lindfors et al. 2008).
Because iron was known to be a catalyst in oxidative cellulose depolymerization (Emery et al. 1974; Rouchon et al. 2011), model studies using iron impregnated fresh oak were recently performed and showed similar depolymerization patterns of holocellulose as those observed in the Vasa (Almkvist and Persson 2008;
Norbakhsh et al. 2013). Preliminary treatment trials in which iron was extracted from small objects from the Vasa were thus initiated in an effort to slow down the reactions (Almkvist and Persson 2006; Fors 2008). The need to study the non-cellulosic components of the Vasa wood (lignin and extractives) was therefore evident, as they too were likely to be affected by oxidative reactions (Vivas et al. 1996; Bentivenga et al. 2003). Reactive functional groups (e.g. easily oxidized phenol groups) found in the non-cellulosic wood components (mainly lignin and extractives) could undergo an oxidation that initiated a catalytic cycle of Fenton type reactions in presence of iron (Elsander et al. 2000). These reactions could, in turn, lead to the formation of oxalic acid and other low molecular weight organic acids that could further depolymerize cellulose through acid hydrolysis. It was, therefore, necessary to study the pH and distribution and concentration of low molecular weight organic acids with respect to different depths from the surface of the timbers.
11
Purpose of research
The research presented in this thesis aims to support the preservation of the warship Vasa and to assist conservators in designing suitable conservation treatments by characterizing the wood of the ship and by understanding the chemical processes that are causing degradation.
Objectives of research
To study the non-cellulosic wood components (lignin and extractives) in the oak wood timbers of the Vasa in order to assess the level of degradation possibly caused by oxidative reactions.
To study the pH and distribution of low molecular weight organic acids with respect to different depths from the surface of the timbers in order to asses in which locations acid hydrolysis of cellulose could be threatening.
Research approach
In the initial phase of the project, cellulose depolymerization in the oak wood of the Vasa was attributed to oxidation reactions of Fenton type in which iron acts as a catalyst. A consequence of such oxidation reactions would be a degradation of lignin in the wood that, in turn, can be assessed through e.g. the quantification lignin subunits syringyl and guaiacyl (Sultanov et al. 1991). For direct chemical analysis of lignin, thioacidolysis (acid hydrolysis in the presence of ethanethiol) coupled with GC-MS was used. During the wet chemical procedure, lignin was degraded to produce mainly monomeric subunits syringyl and guaiacyl that were subsequently identified and quantified by GC-MS. A non-destructive method was also chosen to substantiate the results of the wet chemical procedure, namely solid state CP/MAS 13C NMR. Both methods required pre-treatment of the wood meal involving removal of hydrolysable and insoluble extractives, respectively, as well as iron. Lignin analysis and the results are discussed in detail in Chapter 2.
Some extractives in wood (tannins) may form highly reactive complexes with iron that have been shown to degrade cellulosic materials in so-called ink corrosion (Arpino et al. 1977; Rouchon-Quillet et al. 2004; Rouchon et al. 2011).
Presence of tannins in the Vasa wood could, therefore, support the hypothesis that oxidative reactions are the main mechanism of cellulose depolymerization.
In order to examine the presence of tannins, wood meal was extracted with an acetone:water mixture and the extracts were later analyzed by 13C NMR and MALDI-TOF MS (for low and high molecular weight tannins, respectively), results of which are presented in Chapter 3.
Previous studies have reported the high levels of low molecular weight organic acids, such as oxalic, glycolic, formic and acetic acids, in the wood of the Vasa.
Since oxalic acid is relatively acidic (pKa 1.25), it could participate in the hydrolysis of cellulose (Green et al. 1991). A detailed measurement of pH and quantification of low molecular weight organic acids was therefore performed with respect to increasing distance from the timber surface and is discussed in Chapter 4. In the final stage of the project, experiments in which oxalic acid was added to wood:water suspensions were performed in order to understand why a difference in pH between the surface and the interior of the wood timbers was observed.
Fig. 1 Project outline.
13
Chapter 2: Lignin
Lignin is a phenolic polymer that has the function of a fixating agent in and between wood cell walls by binding to the polysaccharides. It provides stiffness to the cell wall and thus the plant as a whole. Lignin also contributes in
protecting the plant against microbial attack and in making the plant more hydrophobic, thereby assisting in the plant´s water transport. Archeological wood, especially waterlogged, is generally richer in lignin because the polymer is less prone to degradation and dissolution than the polysaccharides. Apart from the porous outer centimeters of the timbers, the oak wood of the Vasa is, however, similar to fresh oak in terms of acid insoluble lignin (Klason lignin) content (Lindfors et al. 2008; Sandberg 2011). In this chapter, possible lignin degradation in the wood of the Vasa is studied in order to understand whether oxidative reactions have occurred since the time of the salvage.
Lignin in oak wood is comprised of the building units guaiacyl (G) and syringyl (S) linked by either condensed (carbon-carbon) bonds or non-condensed ether bonds. The most abundant linkage in hardwood is the β-O-4 bond shown in Fig.
2. A degradation of lignin in the Vasa would imply cleavage of β-O-4 ether bonds and a decrease of the S/G ratio, since syringyl (S) is more susceptible to oxidation than guaiacyl (G) (Sultanov et al. 1991).
Because the cellulose in the Vasa wood is more depolymerized in the interior of the oak timbers compared to the near surface layers, lignin from both locations was analyzed (1-3 cm and 5-8 cm in depth from the surface, respectively). As references, wood from fresh oak, still-waterlogged Vasa wood and two still- submerged Swedish shipwrecks of approximately the same age as the Vasa (Gröne Jägaren and Riksäpplet) were used. In this work, we have used both destructive and non-destructive analytical methods for the characterization of lignin.
Fig. 2 Lignin monomers syringyl (S), guaiacyl (S) and the β-0-4 bond in red.
Wet chemical lignin degradation (thioacidolysis) coupled with gas chromatography (GC-MS) is a relatively sensitive analytical method, well
established for characterization of typical and prominent lignin structures (Lin et al. 1992; Lapierre et al. 1995). It has been used to obtain the number of β-O-4 bonds and S and G building units that are the major constituents of native lignin in the oak wood of the Vasa and the reference samples. In the thioacidolysis procedure, the extractive free wood sample is subjected to a reagent mixture of boron trifluoride etherate, ethanethiol and dioxane. The boron trifluoride acts as a strong Lewis acid enabling side chain oxygen functions, such as the β-O-4 bonds, to be successively replaced by thioethyl groups (Fig. 3). The lignin macromolecule is therefore depolymerized into mono-, di-, and trimeric phenylpropane units with the major degradation products being the thioethyl derivatives of syringyl and guaiacyl. The erythro- and threo isomeric forms of the two latter can be identified and quantified by GC-MS. As seen in Fig. 4, peak pairs 2 and 4 are assigned to G and S units, respectively (Sandberg 2011).
Guaiacyl
Syringyl β-O-4’ linkage
6 6
5 5
4 4
3 3 2 2 1 1 Cα Cβ Cγ
OMe O
OMe
6 6
5 5
4 4
3 3 2 2 1 1 Cα Cβ Cγ
OMe O
6 6 5 5 4 4
3 3
2 2
OMe R
6 6
5 5
4 4
3 3 2 1 2 1
OMe O R
α α
β β γγ
O
1 1
15 Fig. 3 Reaction mechanism of lignin thioacidolysis, where A) shows the substitution at Cα and
B) shows the substitution at Cβ and Cγ, respectively (Rolando et al. 1992).
Fig. 4 Gas chromatogram of the trimethylsilylated (TMS) derivatives of guaiacyl and syringyl (peak pairs 2 and 4, respectively) originating from lignin in the oakwood of the Vasa. The
enlargement shows the area of interest between 20 and 25 minutes.
Results presented in Table 1 show no significant differences in the number of β- O-4 bonds or the S/G ratio between fresh oak and the dry, PEG impregnated Vasa wood. Also, the lignin from the still-waterlogged shipwrecks appears to have remained relatively unchanged.
An exception in Table 1 is the number of β-O-4 bonds in the near surface layers of the dry and PEG impregnated Vasa wood (3900 µmol/g Klason lignin). In these layers, PEG contributes significantly to the weight of the dry samples.
When the amount of acid insoluble lignin (Klason lignin) is determined, the PEG content in the wood is not taken into consideration and, consequently, a low percentage of Klason lignin is derived. Other reasons for a low percentage of Klason lignin could be the extraction of lignin during the time on the seabed and physical degradation of wood by erosion bacteria. Therefore, the absolute values of the S and G units as well as the number of β-O-4 bonds are affected but not the S/G ratio, which is comparable to the other samples.
Sample Klason lignin [%
dry wood]
Guaiacyl [µmol/g Klason lignin]
Syringyl [µmol/g Klason lignin]
β-O-4’
[µmol/g Klason lignin]
S/G (from GC-MS)
Peak intensity ratio (from CP/MAS 13C-
NMR)
Fresh oak 23 900 1600 2500 1.8 3
Vasa surface layers 17 1300 2600 3900 2.0 3
Vasa interior wood 26 900 1900 2800 2.0 2
Still-waterlogged
Vasa wood 28 700 2100 2800 3.0 3
Gröne Jägaren
(waterlogged) 30 800 1700 2500 2.0 3
Riksäpplet
(waterlogged) 28 700 1500 2200 2.0 3
Table 1: From the GC-MS analysis: Amounts of lignin monomeric units S and G; the connecting ether linkages β-O-4; S/G ratios. From the CP/MAS 13C-NMR analysis: peak
intensity ratios of the 153 and 148 ppm signals (etherified/non-etherified S units). Klason lignin is expressed as weight percentage of dry wood.
17 Solid state cross polarization magic angle spinning (CP/MAS) nuclear magnetic resonance (13C-NMR) spectroscopy is a non-destructive analytical technique that can provide supplementary structural information about the lignin in the native state (Bartuska et al. 1980; Kolodziejski et al. 1982; Haw et al. 1984; Haw et al.
1985). The signal at 153 ppm seen in Fig 5 is assigned to C-3/5 carbons of syringyl units involved in ether linkages at C4 (principally β-O-4 bonds) and the less intense signal at 148 ppm is assigned to C-3/4 carbons in guaiacyl and C-3/5 carbons in syringyl in which there instead is a hydroxyl at C4 (Haw et al. 1984;
Haw et al. 1985; Wikberg et al. 2004; Koenig et al. 2010). When β-O-4 bonds are cleaved in degraded lignin, the intensity of the 148 ppm signal increases and the intensity of the signal at 153 ppm decreases, making the relative intensities of the peaks at 153 (etherified S-units) and 148 (non-etherified S units) ppm an
indicator of the cleavage of β-O-4 bonds. Peak intensity ratios of the two signals at 153 and 148 ppm (etherified/non-etherified S units) were calculated and are presented in Table 1 for all samples. As can be seen, the peak intensity ratios of the analyzed samples are similar.
Fig. 5 Solid-state CP/MAS 13C NMR spectra of the analyzed wood samples showing the signals at 148 ppm and 153 ppm. In degraded lignin, the intensity of the 148 ppm signal
increases and the intensity of the signal at 153 ppm decreases.
153 ppm 148 ppm
In conclusion, differences between the dry, PEG impregnated Vasa and the reference oak wood samples (fresh oak and the still-waterlogged shipwrecks) that can be attributed to significant lignin degradation are not detected in our analysis.
An aggressive oxidative degradation mechanism is therefore not supported as the main contributor to cellulose depolymerization in the Vasa.
19
Chapter 3: Extractives
Wood extractives are non-structural compounds of low molecular weight that can relatively easily be separated from the wood using different solvents. The structure, composition and content of extractives vary greatly between species and even within a single tree, where they are generally more abundant in the bark, branches and knots. The extractives play a role in protecting the plant from disease and pests, as well as in growth regulation. There are a number of
different types of lipophilic extractives such as fatty acids, waxes and terpenoids.
This work focuses, however, on hydrolysable extractives (tannins) that are known to form complexes with metal ions and be involved in oxidative reactions that can cause degradation of cellulosic materials, known as ink corrosion (Arpino et al. 1977; Rouchon-Quillet et al. 2004; Rouchon et al. 2011). Although tannins are fairly water-soluble and easily oxidized, their presence in the wood of the Vasa could be an indication that an oxidation is still occurring.
In oak wood, hydrolysable tannins consist mainly of gallotannins and
ellagitannins where the galloyl- and hexahydroxydiphenic acid (HHDP) moieties, respectively, are esterified to a core molecule, generally glucose (Scalbert et al.
1988; Du Penhoat et al. 1991; Mämmelä et al. 2000). Common oak wood tannins consist of: gallic acid, ellagic acid, pentagalloylglucose (other polygalloylglucose esters are also common) and castalagin/vescalagin (isomers), grandinin/roburin E, roburin A/D and roburin B /C, some of which are shown in Fig. 6.
Fig. 6 Chemical structures of common oak wood tannins: a) gallic acid; b) ellagic acid; c) vescalagin/castalagin (1 and 2, respectively) and d) pentagalloylglucose (Vivas et al. 1995).
a) b)
c )
d)
The presence of tannins in the oak wood of the Vasa was studied by solution 13C NMR (for tannin Mw <600) and MALDI-TOF MS (for tannin Mw >600) of acetone:water extracts. Because the wood samples recently obtained from the ship contain high amounts of PEG, samples from still-waterlogged Vasa wood were also used. Analysis of still-waterlogged wood gives an indication of the tannin content before the salvage.
Castalagin/vescalagin was found in the mass spectra of still-waterlogged Vasa wood at m/z 957 (Fig. 7), although some overlap of pentagalloylglucose might occur because of the similar molecular weights of the three compounds. Also grandinin/Roburin E was detected at 1089, but only traces of the high molecular weight ellagitannins could be observed (but not assigned). Another interesting peak found in the extracts of still-waterlogged Vasa wood is detected at m/z 788.
The peak could not be assigned from literature review, but is suspected to be a disodium adduct ion of tetragalloylglucose originating from pentagalloylglucose (Nishizawa et al. 1983; Vivas et al. 1995). The 13C NMR spectrum of the extracts of still-waterlogged Vasa wood shows presence of ellagic acid (Li et al. 1999) but failed however, to confirm the glucose-carbon shifts of all of the above-
mentioned tannins (Fig. 8), indicating that these tannins are most likely present in amounts below the detection limit of the 13C NMR technique. In conclusion, extracts of still-waterlogged Vasa wood contain lower amounts of ellagitannins compared to fresh oak, suggesting that most of the gallo- and ellagitannins have been washed out during the time on the seabed.
In the corresponding MALDI-TOF MS and 13C NMR spectra of Vasa wood extractives, only PEG could be observed. Since PEG is present in high concentrations in the Vasa wood, it might have caused suppression of the ellagitannin peaks in both analyses. However, since only trace amounts of tannins are observed in the extracts of still-waterlogged Vasa wood, there are likely even less tannins in the dry and PEG impregnated Vasa wood. It is possible that most of the tannins present in the Vasa immediately after the salvage have been oxidized or washed out during the 17 year long wet
conservation treatment. These results imply, therefore, that oxidative reactions involving iron and ellagitannins (e.g. ink corrosion type of reactions) are less likely in the Vasa wood at present.
21 Fig. 7 MALDI-TOF MS spectrum of extractives from a) fresh oak and b) still-waterlogged Vasa oak wood, respectively. The tannins found in the extracts of fresh oak constitute mainly of ellagitannins and pentagalloylglucose, based on molecular weights of their ion adducts (m/z)
(Du Penhoat et al. 1991; Puech et al. 1999; Mämmelä et al. 2000).
Peak m/z [M+Na]+ Substance
A 788 n/a
B 957 Castalagin/Vescalagin C 1089 Grandinin/Roburin E
D 1874 Roburin A/D
E 2022 Roburin B/C
Table 2 The peaks represent the cations [M+Na]+ found in fresh oak extractives.
a) b)
Fig. 8 The 13C NMR spectrum of extractives from a) fresh oak, b) still-waterlogged Vasa wood and c) Dry and PEG impregnated Vasa wood. In a) glucose-carbon shifts of vescalagin/castalagin and pentagalloylglucose, and ellagic acid are observed (Nishizawa et al.
1983; Ishimaru Kanji 1988; Du Penhoat et al. 1991; Li et al. 1999). In b), only ellagic acid is detected and in c) PEG at 70 ppm and its end groups at 60 and 72 ppm,
respectively, are clearly visible (Spěváček et al. 2008).
23
Chapter 4: Acids and Elements
A number of different substances and elements have diffused into the wood of the Vasa ship during the time on the seabed and thereafter, during the 17 year long wet conservation treatment. As mentioned in Chapter 1, these elements are in general unevenly distributed with high concentrations near the surface of the timbers and low concentrations in the interior. In contrast, the degradation of cellulose is most extensive in the interior of the wood timbers where a high concentration of low molecular weight organic acids and a low pH is observed.
In this chapter, the differences between surface wood and the interior wood are discussed in terms of pH, acid concentration and distribution of elements (sulfur, S; iron, Fe; calcium, Ca; boron, B).
In order to analyze the Vasa wood from surface to interior, a cylindrical drilling core from a thick timber was cut into five slices. The core was approximately 9 cm in length, where the first two and the last 4-9 cm were considered as surface wood samples and interior wood samples, respectively. To avoid problems with microheterogeneity, a relatively large (approximately 2 g each) sample amount was used. Rusty salt deposits were taken at three locations in the ship and scraped from the timbers using a scalpel. As reference material, fresh oak and still waterlogged Vasa wood were used.
Acid concentration
The amount of low molecular weight acids in the wood samples was determined by cold-water extraction of wood meal and subsequent analysis by High
Performance Ion Exchange Chromatography (HPIEC), a well-established analytical technique for quantification of organic acids. The instrument was calibrated using standards of the analyzed acids (formic, glycolic, oxalic and acetic acids) that were distinguished based on elution times. Since HPIEC measures the acid anion concentration, it is suspected that some of the acid is present in the wood as salt. A pH measurement was performed on the cold- water extracted samples. As the measured pH regards a wood-water suspension in which the acids are diluted, the actual pH in the wood could be lower (Sithole 2005). Elemental analysis on untreated wood meal was performed using
Inductively Coupled Plasma – Atomic Emission Spectroscopy (ICP-AES).
The concentration of formic- and glycolic acids was found to be relatively constant throughout the oak timbers (Fig. 9), whereas the concentration of oxalic- and acetic acids increases with increasing distance from the surface of the
timbers, as previously reported (Almkvist 2008; Bjurhager 2011). In other words, high amounts of oxalic and acetic acids are present in the interior of the wood timbers where the cellulose is more depolymerized.
Because oxalic acid has an acid dissociation constant (pKa 1.25) much lower that any of the other analyzed acids, it is assumed to be the main contributor to the acidity of the wood. It is also suspected that oxalic acid is main cause of cellulose depolymerization. A plausible source of the acetic acid in the wood is acid catalyzed deacetylation of acetyl substituents on xylan units in hemicelluloses (Maloney et al. 1985), hydrolysis that could again be caused by oxalic acid.
Elemental analysis
Elemental analysis shows that the concentrations of iron and sulfur are high near the surface of the timbers and decrease with increasing depth (Fig. 9). This trend is less pronounced for calcium and boron (originating from anti-fungal treatment during the wet conservation (Håfors 2010), both of which are relatively evenly distributed in the timbers. The concentrations of sulfur and calcium in the interior, approximately 6-9 cm deep, are similar to those of fresh oak. A few centimeters from the surface of the timbers, the concentration of the analyzed elements is relatively constant.
Fig. 9 Concentration profile of elements and low molecular weight organic acids from the surface of the timbers in the Vasa to the interior.
0 1 2 3 4 5 6 7 8
1 3 5 7 9
Concentration [mg/g wood]
Distance from surface [cm]
Acetic acid Oxalic acid Formic acid Glycolic acid
0 2 4 6 8 10 12 14 16 18 20
1 3 5 7 9
Concentration [mg/g wood]
Distance from surface [cm]
Fe S B Ca
25
Neutralization of oxalic acid by rust
Iron is present in the Vasa in the form of iron and sulfur salts as well as various oxides and hydroxides (Fors 2008). Incidentally, a decrease in the total iron concentration coincides with an increase in oxalic acid concentration and a drop in pH as seen in Fig. 10.
Fig. 10 Concentration profiles of iron and oxalic acid as well as pH relative of distance from the surface of the oak wood timbers.
Historically, oxalic acid has been used to remove rust stains (Cutbush 1813). The chemistry of this reaction is based on the breaking of iron(III) hydroxide bridges to form iron oxalato complexes, thereby raising the pH (Baumgartner et al. 1983;
Cornell et al. 1987; Panias et al. 1996). As rust (iron(III) oxides and iron(III) hydroxides) is abundant on the surfaces of the Vasa timbers, it could neutralize oxalic acid and thus be a reason for the higher pH in this area. To investigate this, rust deposits from the surface and ground wood samples from the surface and interior of the timbers, respectively, were added to an aqueous solution of oxalic acid. The pH of the suspension was measured instantaneously and continuously during 48h. Suspensions containing rust rich surface wood and rust deposits evidently showed an increase of pH over the time of measurement (Fig. 11), whereas the pH in the suspension of interior wood (poor in rust and rich in oxalic acid) and a fresh oak reference, respectively, remained essentially unchanged. This suggests that the oxalic acid in the interior is not being neutralized, due to the fact that the neutralizing iron (rust) has been consumed.
In contrast, wood from surfaces areas rich in rust appears to be protected from acid hydrolysis by oxalic acid.
0 1 2 3 4 5 6
0 5 10 15 20 25 30 35 40
1 3 5 7 9
pH
Concentration
Distance from surface [cm]
Oxalic acid [mg/g water]
Fe[mg/g wood]
pH (25 °C)
Fig. 11 Change in pH of Vasa interior wood, surface wood, rusty salt deposit, waterlogged wood and fresh oak upon the addition of oxalic acid.
27
Chapter 5: Concluding remarks
In the beginning of the research project “A Future For Vasa”, cellulose degradation in the Swedish warship Vasa was suspected to be caused mainly by oxidative reactions in which iron acts as a catalyst. The work presented in this thesis shows, however, that an extensive oxidation of the non-cellulosic wood polymers has not occurred and suggests that oxidative reactions are not likely to be a main route to cellulose depolymerization in the Vasa. Firstly, the analytical work shows that no significant lignin degradation in the oak wood of the Vasa has occurred. Secondly, no discernible amount of tannins (that could be involved in corrosive reaction with iron) was found.
Furthermore, an explanation is given as to why the cellulose in surface of the timbers is less degraded. Our analysis shows that high amounts of oxalic acid is present in the interior parts of the timbers, which are characterized by low pH and depolymerized cellulose. Surface areas that are rich in rust, however, appear to be protected from acid hydrolysis, since the iron oxides and -hydroxides in rust increase the pH of the wood by binding oxalic acid. It should therefore be considered whether the ongoing iron extraction from small objects is necessary or in worst case, damaging.
The origin of oxalic acid remains a puzzle. Our suspicion is that the oxalic acid originates from microbes that may have been present during the wet
conservation period. Some of the most common wood degrading fungi and molds (e.g. Aspergillus niger) are known to produce oxalic acid (Espejo et al. 1991;
Gadd 1999). However, the low water content and the high acidity in the Vasa wood at present, as well as the previous treatments with boric acid, borax and PCP (pentachlorophenol) are reasons for the unlikelihood of microbial activity today. Since oxalic acid is not volatile, it has not been evaporating as the wood has been drying over the last decades. The increased concentration of the acid due to drying could be a reason for an increased hydrolysis of the
polysaccharides in the Vasa wood.
In conclusion, our research shows that lignin in the Vasa wood is stable but that the high amount of oxalic acid in the interior of the timbers is alarming. Further research regarding the rate of acid hydrolysis by oxalic acid and possible
neutralization methods is advised.
Acknowledgements
This work was performed at KTH, Royal Institute of Technology, as part of the Swedish National Maritime Museums’ research program A Future for Vasa. The financial support by The Swedish Research Council (VR), The Swedish
Foundation for Strategic Research (SSF), The Swedish Research Council for Environment, Agricultural Sciences and Spatial Planning (FORMAS), The Swedish Agency for Innovation Systems (VINNOVA) and KTH, Royal Institute of Technology is gratefully acknowledged.
I would like to express my sincerest gratitude to my supervisor Prof. Monica Ek at the Department of Fiber- and Polymer Technology for giving me the
opportunity to work with this exciting research and for always encouraging me to pursue my many ideas. I also owe many thanks to my co-supervisor Assoc.
Prof. Tommy Iversen at Innventia AB, this work could not have been done without his guidance and help.
My sincerest thanks to the staff at the Department of Fibre- and Polymer Technology at KTH, the staff at Innventia AB and the staff at the Vasa museum for brilliant collaborations, inspiring talks and mischievous laughs.
29
References
Almkvist, G. (2008). The Chemistry of the Vasa - Iron, Acids and Degradation.
Department of Chemistry - Faculty of Natural Resources and Agricultural Science. Uppsala, Swedish University of Agricultural Sciences. PhD Thesis (ISBN 978-91-85913-90-9).
Almkvist, G. and Persson, I. (2006). Extraction of iron compounds from wood from the Vasa. Holzforschung. 60: 678.
Almkvist, G. and Persson, I. (2008a). Analysis of acids and degradation products related to iron and sulfur in the Swedish warship Vasa. Holzforschung. 62: 694.
Almkvist, G. and Persson, I. (2008b) Fenton-induced degradation of
polyethylene glycol and oak holocellulose. A model experiment in comparison to changes observed in conserved waterlogged wood. Holzforschung 62: 704-708.
Arpino, P., Moreau, J. P., Oruezabal, C. and Flieder, F. (1977) Gas
chromatographic-mass spectrometric analysis of tannin hydrolysates from the ink of ancient manuscripts (XIth to XVIth century). Journal of Chromatography, A 134: 433-439.
Bartuska, V. J., Maciel, G. E., Bolker, H. I. and Fleming, B. I. (1980). Structural Studies of Lignin Isolation Procedures by 13C NMR. Holzforschung. 34: 214.
Baumgartner, E., Blesa, M. A., Marinovich, H. and Maroto, A. J. G. (1983) Heterogeneous electron transfer as a pathway in the dissolution of magnetite in oxalic acid solutions. Inorganic Chemistry. 22: 2224-2226.
Bentivenga, G., Bonini, C., D´Auria, M. and De Bona, A. (2003) Degradation of steam-exploded lignin from beech by using Fenton's reagent. Biomass and Bioenergy 24: 233-238.
Bjurhager, I. (2011). Effects of Cell Wall Structure on Tensile Properties of Hardwood: Effect of down-regulation of lignin on mechanical performance of transgenic hybrid aspen. Effect of chemical degradation on mechanical performance of archaeological oak from the Vasa ship. KTH, School of Chemical Science and Engineering (CHE), Fibre and Polymer Technology.
Stockholm, KTH Royal Institute of Technology. PhD Thesis (ISBN 978-91- 7415-914-1).
Bjurhager, I., Halonen, H., Lindfors, E.-L., Iversen, T., Almkvist, G., Gamstedt, K. and Berglund, L. A. (2012) State of degradation in archaeological oak from the 17th century Vasa ship - substantial strength loss correlates with reduction in (holo)cellulose molecular weight. Biomacromolecules. 13: 2521-2527.
Cornell, R. and Schindler, P. (1987) Photochemical dissolution of goethite in acid/oxalate solution. Clays and Clay Minerals. 35: 347-352.
Cutbush, J. (1813). The philosophy of experimental chemistry, Isaac Peirce.
Du Penhoat, C. L. M. H., Michon, V. M. F., Peng, S., Viriot, C., Scalbert, A. and Gage, D. (1991) Structural elucidation of new dimeric ellagitannins from
Quercus robur L. roburins A-E. Journal of the Chemical Society, Perkin Trans. 1 1653-1660.
Elsander, A., Ek, M. and Gellerstedt, G. (2000) Oxalic acid formation during ECF and TCF bleaching of kraft pulp. Tappi Journal. 83: 73-77.
Emery, J. A. and Schroeder, H. A. (1974) Iron-catalyzed oxidation of wood carbohydrates. Wood Science and Technology. 8: 123-137.
Espejo, E. and Agosin, E. (1991) Production and Degradation of Oxalic Acid by Brown Rot Fungi. Applied and Environmental Microbiology. 57: 1980-1986.
Florian, M.-L. E. (1990). Scope and History of Archeological Wood.
Archaeological wood: properties, chemistry, and preservation. Rowell, R. M. and Barbour, R. Journal of the Chemical Society: 3-32.
Fors, Y. (2008). Sulfur-Related Conservation Concerns in Marine Archaeological Wood : The Origin, Speciation and Distribution of Accumulated Sulfur with Some Remedies for the Vasa, Stockholm University, Department of Physical, Inorganic and Structural Chemistry, Stockholm. PhD Thesis (ISBN 978-91- 7155-624-0).
Gadd, G. M. (1999). Fungal Production of Citric and Oxalic Acid: Importance in Metal Speciation, Physiology and Biogeochemical Processes. Advances in Microbial Physiology. Poole, R. K., Academic Press. Volume 41: 47-92.
Green, F., Larsen, M. J., Winandy, J. and Highley, T. (1991) Role of oxalic acid in incipient brown-rot decay. Material und Organism. 26: 191-213.
31 Haw, J., F., Maciel, G., E. and Biermann, C., J. (1984). Carbon-13 Nuclear Magnetic Resonance Study of the Rapid Steam Hydrolysis of Red Oak.
Holzforschung. 38: 327.
Haw, J., F. and Schultz, T., P. (1985). Carbon-13 CP/MAS NMR and FT-IR Study of Low-Temperature Lignin Pyrolysis. Holzforschung. 39: 289.
Håfors, B. (2010). Conservation of the wood of the Swedish warship Vasa of A.D. 1628. Evaluation of polyethylene glycol conservation programmes.
Department of Conservation. Gothenburg, Gothenburg University. PhD Thesis (ISBN 978-91-7346-687-5).
Ishimaru Kanji, I. M., Nonaka Gen-Ichiro, Mihashi Kunihide, Iwase Yukiko and Nishioka Itsuo (1988) Tannins and Related Compounds. LXXII. : Isolation and Characterization of Mongolicanin (Procyanidino-ellagitannin), Mongolinin A, Acutissimin C and Vescalagin Carboxylic Acid, Novel Tannins from Quercus mongolica var. grosseserrata. Chemical and Pharmaceutical Bulletin. 36: 3319- 3327.
Koenig, A. B., Sleighter, R. L., Salmon, E. and Hatcher, P. G. (2010) NMR Structural Characterization of Quercus alba (White Oak) Degraded by the Brown Rot Fungus, Laetiporus sulphureus. Journal of Wood Chemistry and Technology. 30: 61-85.
Kolodziejski, W., Frye, J. S. and Maciel, G. E. (1982) Carbon-13 nuclear magnetic resonance spectrometry with cross polarization and magic-angle spinning for analysis of lodgepole pine wood. Analytical Chemistry. 54: 1419- 1424.
Lapierre, C., Pollet, B. and Rolando, C. (1995) New insights into the molecular architecture of hardwood lignins by chemical degradative methods. Research on Chemical Intermediates. 21: 397-412.
Li, X.-C., Elsohly, H. N., Hufford, C. D. and Clark, A. M. (1999) NMR assignments of ellagic acid derivatives. Magnetic Resonance in Chemistry. 37:
856-859.
Lin, S. Y. and Dence, C. W. (1992). Methods in lignin chemistry. Berlin, Springer Berlin Heidelberg.
Lindfors, E.-L., Lindström, M. and Iversen, T. (2008) Polysaccharide degradation in waterlogged oak wood from the ancient warship Vasa.
Holzforschung 62: 57-63.
Maloney, M. T., Chapman, T. W. and Baker, A. J. (1985) Dilute acid hydrolysis of paper birch: Kinetics studies of xylan and acetyl-group hydrolysis.
Biotechnology and Bioengineering. 27: 355-361.
Mämmelä, P., Savolainen, H., Lindroos, L., Kangas, J. and Vartiainen, T. (2000) Analysis of oak tannins by liquid chromatography-electrospray ionisation mass spectrometry. Journal of Chromatography, A 891: 75-83.
Nishizawa, M., Yamagishi, T., Nonaka, G.-i. and Nishioka, I. (1983) Tannins and related compounds. Part 9. Isolation and characterization of polygalloylglucoses from Turkish galls (Quercus infectoria). Journal of the Chemical Society, Perkin Trans. 1 0: 961-965.
Norbakhsh, S., Bjurhager, I. and Almkvist, G. (2013) Mimicking of the strength loss in the Vasa: model experiments with iron-impregnated recent oak.
Holzforschung (in print), DOI: 10.1515/hf-2012-0150.
Panias, D., Taxiarchou, M., Paspaliaris, I. and Kontopoulos, A. (1996) Mechanisms of dissolution of iron oxides in aqueous oxalic acid solutions.
Hydrometallurgy 42: 257-265.
Puech, J.-L., Mertz, C., Michon, V., Le Guernevé, C., Doco, T. and Hervé du Penhoat, C. (1999) Evolution of Castalagin and Vescalagin in Ethanol Solutions.
Identification of New Derivatives. Journal of Agricultural and Food Chemistry.
47: 2060-2066.
Rolando, C., Monties, B. and Lapierre, C. (1992). Thioacidolysis. Methods in Lignin Chemistry. Lin, S. and Dence, C., Springer Berlin Heidelberg: 334-349.
Rouchon, V., Duranton, M., Burgaud, C., Pellizzi, E., Lavedrine, B., Janssens, K., de Nolf, W., Nuyts, G., Vanmeert, F. and Hellemans, K. (2011) Room-
Temperature Study of Iron Gall Ink Impregnated Paper Degradation under Various Oxygen and Humidity Conditions: Time-Dependent Monitoring by Viscosity and X-ray Absorption Near-Edge Spectrometry Measurements.
Analytical Chemistry. 83: 2589-2597.
33 Rouchon-Quillet, V., Remazeilles, C., Bernard, J., Wattiaux, A. and Fournes, L.
(2004) The impact of gallic acid on iron gall ink corrosion. Applied Physics A:
Materials Science & Processing. 79: 389-392.
Sandberg, T. (2011). The Vasa Warship: Lignin Characterization by means of Thioacidolysis and CP/MAS 13C NMR. School of Chemical Science and Engineering (CHE). Stockholm, KTH, Royal Institute of Technology. Masters Thesis (ISSN 1654-1081).
Sandstrom, M., Jalilehvand, F., Persson, I., Gelius, U., Frank, P. and Hall-Roth, I.
(2002) Deterioration of the seventeenth-century warship Vasa by internal formation of sulphuric acid. Nature 415: 893-897.
Scalbert, A., Monties, B. and Favre, J.-M. (1988) Polyphenols of Quercus robur:
Adult tree and in vitro grown calli and shoots. Phytochemistry 27: 3483-3488.
Sithole, B. (2005). New Method of Measuring the pH of Wood Chips.
Proceedings of the 59th Appita Conference, Auckland, New Zealand, Technical Association of the Pulp and Paper Industry of South Africa.
Spěváček, J. and Baldrian, J. (2008) Solid-state 13C NMR and SAXS
characterization of the amorphous phase in low-molecular weight poly(ethylene oxide)s. European Polymer Journal. 44: 4146-4150.
Sultanov, V. S. and Wallis, A. F. A. (1991) Reactivities of Guaiacyl and Syringyl Lignin Model Phenols Towards Oxidation with Oxygen-Alkali. Journal of Wood Chemistry and Technology. 11: 291-305.
Vivas, N. and Glories, Y. (1996) Role of Oak Wood Ellagitannins in the Oxidation Process of Red Wines During Aging. American Journal of Enology and Viticulture 47: 103-107.
Vivas, N., Laguerre, M., Glories, Y., Bourgeois, G. and Vitry, C. (1995) Structure simulation of two ellagitannins from Quercus robur L. Phytochemistry 39: 1193- 1199.
Wikberg, H. and Liisa Maunu, S. (2004) Characterisation of thermally modified hard- and softwoods by 13C CPMAS NMR. Carbohydrate Polymers. 58: 461- 466.
