The Influence of Solvent Content in Liquefied Wood and of the Addition of Condensed Tannin on Bonding Quality

Liquefi ed wood (LW) is a promising natural material that can be used as a part of the adhesive formulation. However, adhesive bonds made of LW only, have low durability. The aim of this study was, therefore, to increase the durability of adhesive bonds containing LW. LW was obtained with liquefaction of black poplar wood in ethylene glycol (EG) as the solvent and sulphuric acid (SA) as the catalyst. An optimal time of 120 minutes and a wood/EG mass ratio of 1:3 was defi ned for liquefaction at 180 °C. After liquefaction, the EG was evaporated in order to achieve a low solvent content LW with a fi nal mass ratio of 1:1. A hydroxyl number for 1:3 and 1:1 LW was determined in order to examine the reduction of hydroxyl groups. Four different adhesive mixtures were prepared: LW with a mass ratio of 1:1 (LW1:1 ), LW with a mass ratio of 1:3 (LW1:3 ), LW with a mass ratio of 1:1 and added condensed tannin (CT) (LW1:1 /CT), and LW with a mass ratio of 1:3 and added CT (LW1:3 /CT). The solid beech wood lamellas, which were bonded with these adhesive mixtures, were tested directly after bonding, and later on, after 7, 30 and 50 days. The test results indicated greater bonding shear strength in the case of LW1:1 compared to LW1:3. The addition of CT did not contribute to essentially higher shear strength values. The adhesive mixtures LW1:1 and LW1:1 /CT (uncured and cured) were analyzed using FT-IR spectroscopy. No signifi cant differences were observed between the cured LW1:1 and the LW1:1 /CT samples.


INTRODUCTION 1. UVOD
Adhesives are an indispensable part of wood-based composites.They are mainly composed of oil-based derivatives, which make them dependent on an ever-increasing oil price.One of the basic constituents of a large number of synthetic wood adhesives is formaldehyde, which is, however, potentially carcinogenic (IARC, 2004).Due to the free formaldehyde emission, the increasing prices of oil-based derivatives, strict environmental requirements, and increasing ecological awareness, there have been many attempts to produce wood adhesives based on natural and renewable sources.
Liquefi ed wood is one of the naturally-based products that has been developed in recent years.Liquefi ed wood is a product of the thermochemical reaction between wood (wood residues), solvent, and added catalyst.Liquefi ed wood can be used for the production of various biocopolymers.These biocopolymers include coatings (Budija et al., 2009;Kurimoto et al., 2000), various polymers (Wang et al., 2008;Doh et al., 2005), carbon fi bres (Xiaojun and Guangjie, 2010), foams (Alma and Shiraishi, 1998; Lee and Ohkita, 2004), and adhesives.
Over the last 20 years there have been many attempts to use liquefi ed wood as a part of the adhesive formulation.In the earlier years, the development of liquefi ed wood adhesives was based on liquefi ed wood that was prepared with phenol and added formaldehyde (Alma andBastürk, 2001, 2006;Li et al., 2004;Fu et al., 2006;Zhang et al., 2007).Many studies were performed in connection with the application of liquefi ed wood to epoxy resin systems (Kobayashi et al., 2000(Kobayashi et al., , 2001; Asano et al., 2007; Wu and Lee, 2010), and there have been some other attempts to blend liquefi ed wood with synthetic resins such as diisocyanates (Juhaida et al., 2010), urea-formaldehyde (Antonović et al., 2010), melamineurea-formaldehyde (Kunaver et al., 2010) and phenolformaldehyde resin (Ugovšek et al., 2010).
The above mentioned problem -low durability of liquefi ed wood based adhesives -could be potentially alleviated with the incorporation of a chemical substance that would help to crosslink the components of the liquefi ed wood.Based on the aim of creating an environmentally-friendly adhesive, the authors tried to use natural substances that are used as a part of adhesive mixtures.Tannins have been one of the most useful natural sources for wood adhesives (Gornik et al., 2000;Vázquez et al., 2002;Moubarik et al., 2009).They can be divided in two different classes, based on their chemical structure: hydrolyzable and condensed tannins.Almost all tannin-based wood adhesives are made from condensed tannins, due to their widespread availability and higher reactivity.The most reactive part of condensed tannin is resorcinol or the phloroglucinol A-ring of the fl avonoid unit.The resorcinol A-rings of mimosa (Accacia sp.) and quebracho (Schninopsis sp.) tannins show reactivity toward formaldehyde that is comparable to that of resorcinol (Pizzi, 2008).In this study spruce condensed tannin was used, and it was found that it contained approximately 60 % procyanidin and 40 % prodelphinidin (Behrens et al., 2003).Condensed tannins can thus be added to liquefi ed wood due to their high reactivity at elevated temperatures, so that they can react with the free phenolic and alcoholic hydroxyl groups that are present in the liquefi ed wood.
The objective of this study was to optimize the liquefaction of black poplar (Populus nigra L.) using ethylene glycol (EG) as the solvent, and the preparation of liquefi ed wood with a low content of EG.Additionally, the properties of adhesive mixtures using liquefi ed wood with different solvent contents and added condensed tannin (hydroxyl number and infrared spectroscopy) were studied.Finally the shear strength of adhesive bonds that were produced with different adhesive mixtures was tested.

Preparation of liquefi ed wood 2.1. Priprema utekućenog drva
Sawdust (fractions of 0.237 mm or smaller) of the black poplar (Populus nigra L.) was used for the produ ction of liquefi ed wood.Prior to the liquefaction process, the sawdust was dried in a laboratory oven (103 °C, 24 h).Ethylene glycol (p.a.grade) was used as the solvent, and sulphuric acid (p.a.grade) was used as a catal yst. 3 % of SA based on the EG mass was added.Li que fa ction was carried out in a 1000 mL three-neck glass reactor equipped with mechanical stirrer.The reac tor was immersed in an oil bath that was preheated to 180 °C.

Temporary amount of residue and liquefaction yield determination 2.2. Određivanje količine ostatka i postignutog utekućenja
Different mass ratios between wood and EG (1:1, 1:2, 1:3, 1:4, 1:5) and different liquefaction times were investigated in order to achieve the optimum results.By calculating the temporary amount of the residue (TAR), (Eq (1)) the optimum liquefaction time was assessed.During liquefaction, the samples of liquefi ed wood were dispossessed, diluted with a mixture of 1.4dioxane/water, and fi ltered through fi lter disks (Sartorius fi lter disks 388 grade/84/mm 2 ) every 15 minutes.The insoluble parts were dried in a laboratory oven (103 °C, 24 h), and weighed in order to calculate the TAR according to Eq (1).

TAR (1)
W 1 represents the mass of the fi lter paper with the dry residue (g), W 2 is the mass of the fi lter paper (g), W 3 is the mass of the dispossessed sample (g), and TAR represents the temporary amount of the residue (%).
After the optimum liquefaction time, the reactor was immersed in cold water in order to quench the reaction.The liquefi ed product was then diluted with a mixture of 1.4-dioxane and water (4/1, v/v), and fi ltered through fi lter disks (Sartorius fi lter disks 388 grade/84/mm 2 ) in order to remove the insoluble parts of the liquefi ed wood, with the aim of calculating the liquefaction yield (LY).The LY (i.e. the percentage of the 1.4-dioxane soluble part) was calculated for liquefi ed wood with wood/EG ratios of 1:2, 1:3 and 1:4, using Eq. ( 2).In order to obtain the liquefi ed wood containing EG, the mixture of 1.4-dioxane and water was evaporated at 55 °C using a rotavapor instrument.Evaporation was performed under reduced pressure achieved by means of a vacuum pump.

LY
(2) W 1 represents the mass of the fi lter paper with the dry residue (g), W 2 is the mass of the fi lter paper (g), W 4 is the mass of wood (g) and LY represents the liquefaction yield (%).After evaporation of the 1.4-dioxane, the EG in the LW was additionally evaporated (120 °C, 10 mbar) to achieve a fi nal mass ratio wood/EG of approximately 1:1.The mass of the evaporated EG was determined gravimetrically.The evaporation of the EG was also important due to very low viscosity of LW 1:3, which was problematic when applying LW to the wood surface prior to bonding.

Determination of hydroxyl number of liquefi ed wood 2.3. Određivanje hidroksilnog broja utekućenog drva
The hydroxyl (OH) numbers of liquefi ed wood with wood/EG ratios of 1:3 (LW 1:3 ) and 1:1 (LW 1:1 ) were determined according to standard ASTM D 4274-05, test method C -refl ux phthalation.0.45 g of sample LW 1:3 and 0.85 g of sample LW 1:1 was dissolved in 25 mL of a phthalic anhydride-pyridine reagent (115 g phthalic anhydride and 700 mL pyridine) and heated at 115 ± 2 °C, for 1h, under refl ux.After esterifi cation, 50 mL of pyridine was added through a condenser, and a phenolphthalein solution in pyridine was added.The mixture was titrated with 0.5 M sodium hydroxide.Due to the dark colour of the solution and the severe diffi culty in perceiving the colour change to pink, the titration end point was determined with a pH meter (Mettler Toledo, SevenEasy, pH meter S20).The end titration point was determined when a signifi cant change in the mV value occurred.

Preparation of adhesive mixtures 2.4. Priprema mješavine ljepila
The adhesive mixtures were prepared according to Table 1.Condensed tannin (CT) (Tanin Sevnica, Slovenia) from the Norway spruce (Picea abies L.) was used to prepare mixtures with different proportions of CT and LW.

Bonding and testing of specimens 2.5. Lijepljenje i testiranje uzoraka
Solid beech wood lamellas were used as a substrate for the preparation of two-layered test specimens, which were bonded according to EN 12765 by using conventional hot-pressing.Prior to the bonding, all of the beech wood lamellas were planed in order to ensure Table 1 Adhesive mixtures made of liquefi ed wood with different wood/solvent ratios and the addition of condensed tannin Tablica 1. Mješavine ljepila napravljene od utekućenog drva s različitim omjerom drva i otapala te s dodatkom kondenziranog tanina

Weight portion of CT
Težinski udjel CT-a % Fourier transform infrared spectroscopy (FT-IR) was used to investigate and compare the samples of LW 1:1 and LW 1:1 /CT (85/15) in their uncured state.Both samples were later on placed on aluminium foil and cured in an oven for 600 seconds at 200 °C.The cured LW 1:1 and LW 1:1 /CT (85/15) specimens were also investigated using a Perkin-Elmer Instruments Spectrum One, FT-IR spectrometer.All the spectra were compared using Perkin-Elmer Spectrum 6.3.5 software.The spectra were recorded by the ATR technique, using a HATR ZeSn Trough Plate 45°, over the 4000-650 cm -1 wave number range.The spectral resolution of the spectrometer was 1 cm -1 .

Priprema utekućenih proizvoda
The temporary amounts of residue versus the reaction time at different wood/EG ratios are shown in Figure 1.By calculating the TAR, the optimal liquefac-tion time was assessed.The optimum liquefaction time was assessed when the TAR was the least.The fi gure shows that the optimum liquefaction time for the ratios 1:3, 1:4 and 1:5 is above 120 min.With the aim of achieving the lowest possible (and optimal) wood/EG ratio, the 1:3 ratio turned out to be optimal.In the case of the ratio 1:2 the lowest TAR was at 135 minutes.After 135 minutes the TAR increased and re-condensation of the liquefi ed product occurred.In the case of 1:1 ratio, wood liquefaction did not occur, and the stirring in the reactor was aggravated.The ratio 1:1 turned out to be inappropriate for wood liquefaction.The optimal time for liquefying black poplar wood at 180°C was therefore 120 minutes.
The LY was determined for LW with wood/EG mass ratios of 1:2, 1:3 and 1:4.The highest LY was attained at a mass ratio of 1:3 (91 %) and the lowest LY at a mass ratio of 1:2 (84 %).LY for LW with a mass ratio of 1:4 was similar to LW with a mass ratio of 1:3.Due to the above mentioned purpose of achieving the lowest wood/EG ratio, the LY for 1:5 ratio was not investigated.At 1:1 ratio, liquefaction did not occur.A wood/EG mass ratio of 1:3 was therefore the most appropriate for the liquefaction of black poplar with EG, and 120 minutes at 180°C was required for the highest liquefaction yield.
The liquefaction process at the optimal conditions and the preparation of low-solvent liquefi ed product is shown in Figure 2. As can be seen, a theoretical wood/ EG mass ratio of 1:1 was obtained.The exact mass ratio of the fi nal product was actually 1:1.05.The evaporation of EG also contributed to the suitable viscosity of the LW when applying it to the wood surface.

Determination of hydroxyl number of liquefi ed wood 3.2. Određivanje hidroksilnog broja za utekućeno drvo
The OH numbers of the liquefi ed wood with different wood/EG mass ratios (1:3 and 1:1) was investi- Figure 1 The temporary amount of residue (TAR) at different liquefaction reaction times and different wood/EG ratios Slika 1. Privremena količina ostatka (TAR) pri različitim vremenima reakcije utekućenja i različitim omjerima drva i otapala gated as this could hypothetically affect the bonding properties of the specimens, and the durability of the bond-line.Table 2 shows the calculated OH number of the EG and determined OH number of the samples LW 1:3 and LW 1:1 .Liquefaction of the wood signifi cantly reduced the OH number of the liquefi ed product containing EG, which occurred due to the dehydration and thermal oxidation of the glycols, as well as due to the condensation reactions between the glycol and wood components, such as cellulose, hemicelluloses and lignin (Kunaver et al., 2010).The reduction of the OH number is also the consequence of solvent evaporation (Budija et al., 2009).It can be seen that the additional evaporation of EG to obtain a mass ratio of 1:1 lowered the number of OH groups to almost one half of the LW 1:3 OH value.The reduction of free OH groups could potentially affect the durability of the bond-line.This is because free OH groups are potentially reactive locations, which interact with other functional groups, but are also possibly weak points for water deterioration if the adhesive bond is inadequately cured.

Shear strength of tested specimens 3.3. Smicajna čvrstoća ispitivanih uzoraka
The initial shear strength values of the specimens bonded with EG liquefi ed wood adhesive mixtures were low (Table 3), and also, they did not exceed the requirements of the standard for any of the durability classes.It is interesting that they did not decrease drastically over the following period of time, which is typically the main problem of bonding only with LW (Ugovšek et al., 2010).It can also be seen that the shear strength values of the adhesive mixtures with a wood/ EG ratio of 1:1 exhibited higher shear strength values than those with a ratio of 1:3.This was probably due to the more suitable viscosity of the sample LW 1:1 , which contributed to the appropriate fl ow and penetration of the adhesive and wetting of the wood surface (Marra, 1992).The durability of the LW 1:1 adhesive bonds was also better compared to those corresponding to LW 1:3 .The unreacted EG in the LW 1:3 infl uenced negatively the shear strength durability.The latter could be correlated to fewer free OH groups in the LW 1:1 than in the LW 1:3 , which has more OH groups from the excessive EG.The free OH groups could be potential locations for water deterioration.
The addition of condensed tannin to the EG liquefi ed wood adhesive mixtures did not contribute signifi cantly to higher shear strength values of the adhesive bond.This could be due to the high viscosity of the adhesive mixture containing tannin, and, consequently, poor penetration and anchoring.It could also be due to the low solubility of the tannins in the liquefi ed wood, which has a very low pH value (less than 1).
As the curing mechanism and the rate of crosslinking at the given curing conditions (200 °C, 900 s) of the liquefi ed wood adhesive mixtures is not known, it should be investigated in the future.
The wood surface was considerably deteriorated in the area where the adhesive had been applied, which was probably the consequence of either liquefaction of the wood surface during pressing or acidic damage of the wood tissue due to very low pH of the adhesive mixture.The wood surface had, therefore, much lower shear strength values than normal (unaffected) wood, BLACK POPLAR (100 g), ETHYLENE GLYCOL (300 g), SULPHURIC ACID (9 g) CRNA TOPOLA (100 g), ETILEN GLIKOL (300 g), SUMPORNA KISELINA (  and a higher percentage of wood failure (Figure 3).The latter is, apart from poor water resistance, the main problem that needs to be worked on.

FT-IR analysis 3.4. FT-IR analiza
Fourier transform infrared spectroscopy (FT-IR) was used to investigate and compare samples of LW 1:1 and LW 1:1 /CT (85/15).Both samples were also cured (cured LW 1:1 and cured LW 1:1 /CT (85/15)) and their spectra were investigated.A comparison of all four spectra is shown in Figure 4.It can be seen that spectra corresponding to LW 1:1 and LW 1:1 /CT (85/15) are almost indistinguishable, which is proof that there were no new functional groups in the LW 1:1 /CT (85/15) adhesive mixture.
After curing, the spectra of the LW 1:1 and LW 1:1 / CT (85/15) mixtures changed.If the focus is placed on changes in peaks, then the most evident changes was at 1120 cm -1 , which could correspond to the reduction of -OH groups after curing (evaporation and reaction of the solvent) (Budija et al., 2009) or C-O stretch in the cellulose (Pandey and Pitman, 2003).Reduction of the peak at 1050 cm -1 can be associated with C-O stretch in the cellulose (glucose) (Morohoshi, 1991;Ibrahim et al., 2006;Gierlinger et al., 2008), and reduction in the peak at 883 cm -1 can be linked to antisymmetric out-of-phase stretching of cellulose (Morohoshi, 1991).All three peak reductions correspond to changes in the cellulose, or a reduction in the number of hydroxyl groups, which is a reasonable explanation and the consequence of curing.New peaks around 1090 cm -1 and 1020 cm -1 correspond to C-O ether vibrations (Budija et al., 2009).This indicates the presence of new compounds after curing.
A comparison of the spectra of the cured LW 1:1 and LW 1:1 /CT (85/15) is shown in Figure 5.It explains the possible differences due to the addition of condensed tannin.Both spectra are rather similar, but with three differences in the fi ngerprint region.Reduction of the peak at 1242 cm -1 could either be associated with a syringyl ring and C-O stretch in the lignin and xylan (Pandey and Pitman, 2003), or C-H and -OH deformation and C-O-C stretching vibration of the cellulose (Morohoshi, 1991;Gierlinger et al., 2008).The peak around 1065 cm -1 is attributed to pyranose ring stretching (Bouchard, 1990).The absorption at 853 cm -1 is associated with C-H outof-planes in positions 2, 5 and 6 of the G lignin units (vanillin) (Faix, 1991).It can be clearly seen that the latter absorption peak and the peak at 1242 cm -1 disappeared during the LW 1:1 /CT (85/15) curing.This means that during the curing of the LW 1:1 /CT (85/15) sample, vanillin and syringyl units of lignin, xylan or parts of cellulose could interact with the other functional groups or molecules.Despite the mentioned changes in the FT-IR spectra of the cured LW 1:1 and LW 1:1 /CT (85/15) samples, the effect of the addition of tannin to LW cannot be confi rmed, which also coincides with the results of the shear strength test.

ZAKLJUČCI
The bonding properties of low solvent content liquefi ed wood, with the addition of condensed tannin, have been investigated.It was determined that the mass Figure 4 FT-IR spectra of liquefi ed wood with a wood/EG mass ratio of 1:1 (LW 1:1 ), liquefi ed wood with a wood/EG mass ratio of 1:1 and the addition of condensed tannin (LW 1:1 /CT (85/15)), cured liquefi ed wood with a wood/EG mass ratio of 1:1 (LW 1:1 (cured)), and liquefi ed wood with a wood/EG mass ratio of 1:1 and the addition of condensed tannin (LW 1:1 /CT (85/15) (cured)) Slika 4. FT-IR spektar utekućenog drva s masenim omjerom drvo:EG u iznosu 1:1 (LW 1:1 ), utekućenog drva s masenim omjerom drvo:EG u iznosu 1:1 i dodanim kondenziranim taninom (LW 1:1 /CT (85/15)), utekućenog drva s masenim omjerom drvo: EG u iznosu 1:1 i osušen (LW 1:1 (cured)), utekućenog drva s masenim omjerom drvo:EG u iznosu 1:1, dodanim kondenziranim taninom i osušen (LW  liquefi ed wood (LW 1:1 and LW 1:3 ), and the liquefi ed wood with added condensed tannin, were tested for shear strength.It was found that the evaporation of ethylene glycol contributed to better durability of the adhesive bonds.The addition of condensed tannin did not contribute essentially to better durability or higher shear strength.The specimens did not fulfi l the requirements of the standard EN 12765 for any of durability classes.FT-IR spectra of the low solvent content LW, with and without the addition of condensed tannin, were investigated.Some minor differences among spectra were observed.The presence of new functional groups was not confi rmed.The results of this study indicated that LW containing less EG (a mass ratio of 1:1) exhibited better bonding properties than LW with an initial mass ratio of 1:3, and that the addition of condensed tannin did not contribute to an improvement in the bonding quality.

Table 1 )
. Each of the adhesive mixtures was applied by means of a roller, using an application rate of 200 g/m 2 .The press temperature was 200 °C, and the press time was 900 seconds.The specifi c press pressure was 1.5 MPa.