Wood Science and Technology

, Volume 47, Issue 3, pp 523–535

MALDI-TOF, HPLC-ESI-TOF and 13C-NMR characterization of chestnut (Castanea sativa) shell tannins for wood adhesives


  • Gonzalo Vázquez
    • Department of Chemical Engineering, School of EngineeringUniversity of Santiago de Compostela
  • Antonio Pizzi
    • ENSTIB-LERMABNancy Université
  • M. Sonia Freire
    • Department of Chemical Engineering, School of EngineeringUniversity of Santiago de Compostela
  • Jorge Santos
    • Department of Chemical Engineering, School of EngineeringUniversity of Santiago de Compostela
  • Gervasio Antorrena
    • Department of Chemical Engineering, School of EngineeringUniversity of Santiago de Compostela
    • Department of Chemical Engineering, School of EngineeringUniversity of Santiago de Compostela

DOI: 10.1007/s00226-012-0513-8

Cite this article as:
Vázquez, G., Pizzi, A., Freire, M.S. et al. Wood Sci Technol (2013) 47: 523. doi:10.1007/s00226-012-0513-8


MALDI-TOF, HPLC-ESI-TOF and 13C-NMR techniques were used to analyse the structure of non-purified aqueous chestnut shell tannin extracts. In addition, the influence of the extraction agent (water or aqueous solutions of Na2SO3 and/or NaOH) on tannin structure was analysed by MALDI-TOF in order to select the extract with the best properties for wood adhesives. Using HPLC-ESI-TOF, catechin/epicatechin, gallocatechin/epigallocatechin, dicatechin structures, dicatechin structures without a hydroxyl group, galloyl-glucoses and ellagic acid were identified as the main monomeric components in the aqueous extract. 13C-NMR and MALDI-TOF spectra revealed that extracts are mostly composed of procyanidin and prodelphinidin structures although prorobinetidins might be also present. MALDI-TOF spectrometry was used to identify the extract oligomeric components. Extractions with Na2SO3 and/or NaOH produced changes in the predominant structures in the extracts and on the amount of sugar units linked to the flavonoid structures, which decreased in the presence of Na2SO3. Extract obtained using low Na2SO3 and NaOH concentrations (1.5 and 0.75 %, respectively) was selected as more suitable for wood adhesive preparation.


At present, wood adhesives industry is based on chemicals obtained from oil, such as phenol, formaldehyde or urea. The urgent need to find alternatives to the fossil fuels together with environmental concerns related to the use of formaldehyde have promoted to look for new raw materials that serve as substitutes for the oil derivatives for diverse applications, such as wood adhesives. In this respect, the use of polyflavonoid tannins as components of wood adhesives is one of the proposed alternatives, not only to prevent the use of formaldehyde but also to reduce its emission from wood panels (Pizzi 2006). However, the selection of the most appropriate tannin adhesives technology requires evaluation of tannin characteristics, including tannin structure, composition or polymerization degree.

The chestnut (Castanea sativa) represents 5 % (in m3 with bark) of the total wood stock in Galicia (NW of Spain) where the food industry uses ~7,000 t/year of chestnuts in the production of marron glacé, chestnut purée, etc. The peeling process generates a waste product, the shell, which represents ~10 % of the weight of whole chestnut, at present used as fuel. In a previous work (Vázquez et al. 2010), the extraction of polyphenols from chestnut shell for their application in the formulation of wood adhesives, in leather tanning and as antioxidant compounds, was analysed and the results revealed the potential of chestnut shell extracts for all the applications proposed.

The composition of tannin extracts from chestnut wood (Pasch and Pizzi 2002), bark (Garro-Gálvez et al. 1997) and flesh (Hwang et al. 2001) was determined and all belong to the group of hydrolysable tannins. With respect to the shell, Vasconcelos et al. (2010) found low molecular weight phenolics, condensed tannins and also ellagitannins in chestnut fruit pericarp and integument extracted with organic solvents.

13C-NMR is a technique commonly used to identify the structural characteristics and properties of potential importance for the use of tannin extracts as wood adhesives (Newman and Porter 1992; Pizzi and Stephanou 1993; Vázquez et al. 2004). Matrix-assisted laser desorption/ionization time of flight (MALDI-TOF) mass spectrometry (MS), introduced by Karas et al. (1987), has revealed itself to be also a powerful method for the characterization of both synthetic and natural polymers (Pasch and Pizzi 2002; Pasch et al. 2001).

In this work, in order to plan the use of chestnut shell extracts as components of adhesives for wood panels, the influence of the extraction agent (water and aqueous solutions of Na2SO3 and/or NaOH) on the structure and significance of the oligomers present in the extracts and on the polymerization degree of the polyflavonoid tannins was investigated using MALDI-TOF mass spectrometry and 13C-NMR solid phase spectrometry, respectively. The monomers present in chestnut shell extracts were also identified using RP-HPLC-ESI-TOF mass spectrometry.

Materials and methods


Chestnut fruit (Castanea sativa) shell (a mixture of the outer brown peel and the inner pellicle) was supplied by a food factory, Marrón Glacé, S.L. (Ourense, Spain). It was air dried to equilibrium moisture content, ground in a hammer mill, sieved, and the fraction of particle size between 0.1 and 2 mm was selected. Chemical composition of chestnut shell was determined in a previous work (Vázquez et al. 2008).

Extraction and concentration

The extraction experiments were carried out in a 10-L Pyrex glass reactor with mechanical stirring and temperature control. Chestnut shell and water were mixed at room temperature, heated and, once the selected temperature was attained, the alkali was added and contact time begun to run. After 1 h, the suspension was vacuum filtered, the solid residue was washed with water until a nearly colourless filtrate was obtained, and the extract together with the first water washings was concentrated by spray drying.

Solid/liquid ratio was maintained constant at 1/10 (w/w). Water and aqueous solutions of different alkaline chemicals such as sodium hydroxide and sodium sulphite (alone or combined) were used as extraction agents. Table 1 shows the extraction conditions of each extract obtained and the corresponding extraction yield.
Table 1

Extraction conditions tested for chestnut shell


Na2SO3 (% on oven-dried chestnut shell)

NaOH (% on oven-dried chestnut shell)

Temperature (°C)

Extraction yield (%)


























RP-HPLC-ESI-TOF mass spectrometry

Water extract (CS1) was analysed using an Agilent Technologies 1,100 HPLC and a Bruker Microtof ESI-TOF instrument. The different components of the sample were separated using a Zorbax Eclipse XDB-C18, 5 μm, (4.6 × 150 mm) column and a binary gradient of 2 % acetic acid in water for the mobile phase A and 0.5 % acetic acid in water/acetonitrile (1:1, v/v) for the mobile phase B. The linear gradient was from 10 to 55 % B from 0 to 50 min, from 55 to 100 % B from 50 to 60 min and from 100 to 10 % B from 60 to 65 min. The mass spectrometry analysis was performed in negative ion mode.

To elucidate the structural composition of the chestnut shell tannins, HPLC-ESI-TOF analysis of the following standard compounds was also performed under the conditions mentioned above: (+)-catechin hydrate, gallocatechin, (−)-epicatechin, procyanidin B2, quercetin-3-β-d glucoside, quercetin-3-o-rhamnoside, ellagic acid, (−)-gallic acid, isorhamnetin, kaempferol and tannic acid. The sample and the standards were dissolved in water to an initial concentration of 100–200 ppm.

Solid state 13C-NMR

The solid state 1D CP–TOSS (Cross-polarization–Total Suppression of Spinning Sidebands) 13C-NMR spectrum of the chestnut shell tannin extract CS1 was obtained at room temperature with a Varian Inova spectrometer 750 of 17.6 T. The speed of MAS rotation was of 9 kHz. A linear ramp was used for the cross-polarization, with a time contact of 5 ms, and a TPPM heteronuclear decoupling was used with a field strength of 74 kHz. The number of scans was 5,000, and the relaxation time between scans (d1) was 2.5 s.

MALDI-TOF–MS analysis

The method of Pasch et al. (2001) was used to analyse the samples. The spectra were recorded on a Kratos Kompact MALDI 4 instrument. The irradiation source was a pulsed nitrogen laser with a wavelength of 337 nm, and the duration of the laser pulse was 3 ns. The measurements were carried out using the following conditions: positive polarity; linear flight path; high mass (20 kV acceleration voltage); 100–150 pulses per spectrum. The delayed extraction technique was used applying delay times of 200–800 ns.

The chestnut shell tannin extracts (CS1–CS5) were dissolved in acetone (4 mg/mL), and then the sample solutions were mixed with an acetone solution (10 mg/mL acetone) of 2,5-dihydroxy benzoic acid used as matrix. For the enhancement of ion formation, NaCl was added to the matrix. The solutions of the samples and the matrix were mixed in equal amounts, and from 0.5 to 1 μL of the resulting solution was placed on the MALDI target. After evaporation of the solvent, the MALDI target was introduced into the spectrometer.

Results and discussion

HPLC-ESI-TOF, MALDI-TOF and 13C-NMR techniques have been used to analyse aqueous chestnut shell tannin extract (CS1) in order to get information on structure, reactivity and polymerization degree.

The monomeric components of the extract were identified by HPLC-ESI-TOF. The resulting signal was processed by the base peak method, and the chromatogram obtained is presented in Fig. 1. The main compounds found in the extract, assigned as shown in Table 2, were catechin/epicatechin, gallocatechin/epigallocatechin, dicatechin and the structure D observed earlier in pine bark tannins by Navarrete et al. (2010). Hydrolysable tannins such as galloyl-glucoses and ellagic acid were also observed.
Fig. 1

RP-HPLC-ESI-TOF chromatogram of chestnut shell extract CS1. (2) Galloyl, digalloyl and trigalloyl-glucose (3) galloyl and digalloyl-glucose (4) isorhamnetin (5) prorobinetidin and prodelphinidin (6) gallocatechin (7) di-epicatechin (8) catechin and dicatechin (9) ellagic acid (11) disaccharide (13) D structure

Table 2

Phenolic compounds in chestnut shell extract CS1


RT (min)


Identified compound

MS data (fragments)










Galloyl, digalloyl and trigalloyl-glucose

337.12, 481.07





Galloyl, digalloyl-glucose












Prorobinetidin and prodelphinidin


















Catechin and dicatechin






Ellagic acid












328.03, 343.05






236.11, 293.18, 375.19




Structure Da

263.17, 293.18, 307.16






231.11, 291.13, 351.15






RT retention time, MS Mass spectrometry

aIdentified by Navarrete et al. (2010); Std peaks identified with standard compounds, MZ peaks identified by their molecular weight, UP unknown peak

The 13C-NMR spectrum of the CS1 extract is shown in Fig. 2 where some of the characteristic bands of a typical condensed tannin pattern were observed: the band at 154 ppm, which belongs to the flavonoid C5, C7, attached to phenolic –OH groups on the A-ring, and the C9 of the same ring; the bands typical of procyanidin units at 120 ppm for the C6′, at 116 ppm for C2′ and C5′ and at 144 ppm for C3′ and C4′ (Zhang and Lin 2008); the band at 131 ppm for the C1′; the band at 85 ppm for the C2; the band at 104 for the C10; the broad band centred on 105 ppm for C4–C8 linkages and a small contribution of C4-C6 linkages (Newman and Porter 1992); the band corresponding to the unreacted C6, C8 and C10 at 97 ppm (Waver et al. 2006) and the C3 band at 72 ppm which could correspond to C3 sites in the interior chain and upper chain-ending positions (Lorenz and Preston 2002).
Fig. 2

CP-MAS 13C-NMR spectrum of chestnut shell extract CS1

No peaks were observed in the range of 170–180 ppm, which indicates that gallic acid residues linked to the C3 of the heterocyclic ring of the flavonoid structure were not present. The low intensity of the free C4 band (at 27–28 ppm) confirms the predominance of interflavonoid links and the MALDI-TOF finding that higher molecular weight oligomers are present (Fig. 3a). This observation does not allow to know whether the units in aqueous chestnut shell extract are mainly linked 4,8 or 4,6. The apparent predominance of the higher molecular weight repeating units appears to indicate that the aqueous extract (CS1) is mostly composed of procyanidins and prodelphinidins, some fisetinidin units appearing to be present in other extracts (see peak at 1,371 Da for CS5, Fig. 3e). As regards the presence of robinetidin units or prorobinetidin segments in the oligomers, this cannot be confirmed nor totally excluded as robinetidin units have the same molecular weight (Da) of catechin units. However, the great abundance of the higher molecular weight repeating units (306, 442) militate for the 290 Da unit being catechin units, hence for this tannin to be predominantly a procyanidin. The extreme reactivity of the phloroglucinol A-ring limits the use of this type of tannins as wood adhesives; this makes the tannins with resorcinol A-rings better for this application (Hemingway and McGraw 1978; Hillis and Urbach 1959). The relative intensity of the free (unreacted) C6 and C8 sites on A-rings at 97–98 ppm (Table 3), which is a set of very sensitive bands related directly to tannin reactivity and, indirectly, to tannin polymerization degree, revealed that chestnut tannins had a low polymerization degree.
Fig. 3

MALDI-TOF spectrum of chestnut shell extracts (a) CS1, (b) CS2, (c) CS3, (d) CS4 and (e) CS5 at 450–4,000 Da

Table 3

Relative band intensities (%) of the solid state CP-MAS 13C-NMR spectrum of the aqueous chestnut shell extract


C5, C7


C3′, C4′



C5′, C2′



C6, C8













Intensity, %










The bands in the region of 60–90 ppm from the carbons of carbohydrates, which can be added to the polyflavonoid structure, indicated the carbohydrate proportion in the extract. The relative intensity of these bands was lower than those of mimosa commercial extracts (Pizzi 1994).

Finally, the low intensity of the band at 120 ppm for the C6′, C5′ and C2′ of the B-ring revealed the presence of pyrogallol type B-rings together with catechol B-rings in the aqueous chestnut shell extract (Navarrete et al. 2010). The B-ring reaction with formaldehyde is faster with pyrogallol than with the catechol structures (McGraw et al. 1992).

The MALDI-TOF technique was used to complete the analysis of aqueous extracts and to compare chestnut shell tannins obtained under different extraction conditions, that is, changing the type and concentration of the extraction agent (water or aqueous solutions of Na2SO3 and/or NaOH). The MALDI-TOF spectra of chestnut shell extracts CS1-CS5 (Fig. 3a–e) showed that all consisted of varying proportions of these constituent monomers: catechin/epicatechin (A), epigallocatechin (B) and epicatechin gallate (C) (Fig. 4), with molecular weights (MW) of 290.3, 306.3 and 442.4 Da, respectively. However, the presence of two repeating structures with a lower molecular weight of 274.3 Da was also observed (Fig. 5): the fisetinidin (F) that has the same structure as catechin but without the –OH group in the C5 of the A-ring, and the structure E, that is the result of the loss of one gallic group of the epicatechin gallate structure. Additionally, the structure D with a MW of 528.0–529.8 Da, previously identified by Navarrete et al. (2010) in maritime pine tannin extract, was also found. The structure D is a dimer of the fisetinidin (or structure E) that has lost a –OH group in the C3′ of one of the two B-rings. Finally, monosaccharide structures (S) attached to the tannins, generally rhamnose, were also detected.
Fig. 4

Main flavonoid structures

Fig. 5

Structure of fisetinidin and E and D structures

The masses of the oligomer peaks in the spectra (Fig. 3a–e) have been calculated by combination of the masses of the catechin monomers according to the expression: M + Na = 23(Na) + 2(endgroup,H) + 288.3A + 304.3B + 440.4C + (262 or 263)(D/2) + 272.3F +180S. In the expression, A, B and C represent the three principal flavonoid monomers: catechin, epigallocatechin and epicatechin gallate, respectively, F is the fisetinidin, D is the number of the dimmer structure of fisetinidin that has lost one hydroxyl group, and S is the number of sugars added to the polyflavonoid structure. Online Resource 1 shows the majority of the dominant peaks observed by MALDI-TOF and the result of the combination of monomer units forming the series of different oligomers present in the samples. Analysing the series, the influence of the extraction agent on the structure of the chestnut shell extracts has been studied. Sulphitation of tannins generally affords tannins of lower viscosity and increases solubility due to the elimination of the heterocyclic ether group, the introduction of hydrophilic sulphonic and hydroxyl groups, the decrease of polymer rigidity, steric hindrance and intermolecular hydrogen bonding and the acid hydrolysis of the hydrocolloid gums and the interflavonoid bond. On the other hand, partial autocondensation can occur under strongly alkaline conditions (Pizzi 1983).

In the case of CS1 extract, the repeated unit of the flavonoids is the epigallocatechin (MW = 306) with a monosaccharide residue (MW = 180), generally rhamnose, which appears usually in the flavonoids extracted with water attached by an ether linkage to the O of the C3 site of the heterocycle of the tannin. Thus, the spectrum (Fig. 3a) showed a series of peaks with a difference of approximately 486 Da. The dominant series at 834.5, 1321.5, 1808.3 and 2295.7 Da is composed of fisetinidin structures linked together with B type structures (Fig. 4) and sugar units. The other series with other dominant peaks and separated by 486 Da indicated the occurrence of A type structures (catechin/epicatechin) (Fig. 4).

Epigallocatechin and the combination of catechin and epigallocatechin were the predominant structures in CS2, whereas in CS5 extracts, there were more D structures. With respect to CS3 and CS4 extracts, the most abundant structure was epigallocatechin, while epicatechin gallate was only present in CS4 extract.

The spectra of alkaline extracts (Fig. 3b–e) showed that there were no sugar residues when 1.5 % sodium sulphite was used alone (CS2), or very few when combined with a low sodium hydroxide concentration (CS5). However, the extracts obtained with 2.5 % sodium hydroxide alone (CS3) or mixed with 2.5 % sodium sulphite (CS4), showed a significant amount of monosaccharide residues in their structures. Thus, for both extracts, one of the dominant series at 1157.6, 1641.9, 2126.2 and 2615.5 Da was composed of catechin/epicatechin structures linked together with B type and sugar units as for CS1 extract. Sugars are harmful in the formulation of tannin adhesives as they reduce the strength and water resistance of the glued joints in proportion to the amount added (Pizzi 1983), therefore, their presence should be reduced as much as possible.

Taking into account the results obtained, the extraction with low sodium sulphite and sodium hydroxide concentrations (1.5 and 0.75 %, respectively) led to the tannin extracts CS5 with the best structural characteristics for wood adhesives formulation, including a low polymerization degree and low sugars content in combination with high extraction yield and good extract properties for wood adhesives such as Stiasny number, total phenols contents, etc., as found in a previous work (Vázquez et al. 2010).


HPLC-ESI-TOF, MALDI-TOF and 13C-NMR techniques have been use to analyse the structure of aqueous chestnut shell tannin extracts. The HPLC-ESI-TOF technique showed that the extract consisted of catechin/epicatechin, gallocatechin/epigallocatechin, dicatechin and the monomer of the dimer D structure. 13C-NMR and MALDI-TOF spectra revealed that the extract was mostly composed of procyanidins and prodelphinidins, although prorobinetidins might also be present. The MALDI-TOF technique was also used to compare chestnut shell tannins obtained under different extraction conditions. Changes in the predominant flavonoid structures (catechin/epicatechin, epigallocatechin and epicatechin gallate) and other minor structures (fisetinidin and structures E and D) and in the amount of sugars linked to them were observed. The extraction with low sodium sulphite and sodium hydroxide concentrations (1.5 and 0.75 %, respectively) led not only to a high extraction yield but also to the extracts with the lowest polymerization degree and a low sugar concentration in the flavonoid structure making them more suitable for wood adhesive preparation.


This work was funded by Ministerio de Ciencia e Innovación, Feder FUNDS and Plan E Fundy (CTQ2009-07539). Authors are also grateful to Ministerio de Ciencia e Innovación for a research grant awarded to J. Santos (BES-2006-13886).

Supplementary material

226_2012_513_MOESM1_ESM.pdf (66 kb)
Supplementary material 1 (PDF 65 kb)

Copyright information

© Springer-Verlag Berlin Heidelberg 2012