TANNIN-BASED RESINS MODIFIED TO REDUCE WOOD ADHESIVE BRITTLENESS.
In South Africa and Brazil, natural tannins extracted from black wattle (Acacia mearnsii) bark are commercially available as the phenolic adhesive component for bonding wood. This preliminary work deals with the reduction of the brittleness of the tannin-based adhesive joint in such wood materials through the chemical modification of the condensed tannins before reacting with formaldehyde. The brittleness reduction was achieved by the introduction of adipic ester segments between the hydroxyl groups of the compact tannin macromolecules. The exploratory modified tannin-formaldehyde adhesive was tested for plywood mechanical properties under dry, water, and boil test procedures.
The tannins from black wattle (Acacia mearnsii) bark are commercially viable substitutes for phenolic resins in wood adhesives. Natural wattle tannins are polymers composed of condensed flavonoid units, presenting a large number of phenolic groups. The hydrogen atoms at the ortho or para position of the phenolic nuclei are suitable for aromatic eletrophilic substitution [5,8,15,24]. The condensation of tannins with formaldehyde, by acid- or base-catalysis, results in products that are useful for adhesive formulations. In addition to cost considerations, many tannin adhesives have limited industrial application due to some disadvantages such as low strength, poor adhesion, too fast or low reactivity, and brittleness. These undesirable characteristics may arise from the difficulty of forming methylene cross links during the cure with formaldehyde due to steric hindrance of the reactive hydrogen atoms or structural rearrangements of the bulky flavonoid structure [16-19,21,22]. In this preliminary work, a chemica lly modified tannin adhesive was formulated for improved flexibility. The modification involved the exploratory introduction of flexible, long linear six carbon chains to "open" the tannin structure through partial esterification of the free hydroxyl groups with adipoyl chloride.
MATERIALS AND METHODS
The commercial wattle-based tannin in the study contained 72 [+ or -] 2 percent tannins and 21 [+ or -] 1 percent nontannins and had 2.4 [+ or -] 1 percent on ash content . The tannin esterification was carried out in a reaction flask by slow addition of 20 percent w/w adipoyl chloride solution in chloroform on a stirred 5 percent w/w suspension of 5.3 mmol of tannin in 1,4-dioxane, in the proportion of 6 to 10 parts of adipoyl chloride to 1 part tannin, containing enough pyridine to retain the evolved hydrogen chloride. The resulting powder was treated with 5 percent hydrochloric acid aqueous solution to remove residual reagents, filtered, washed with cold water, and dried . The crude tannin ([FA.sub.q]) was fractionated and purified by lead acetate techniques  and the same procedure for esterification was done with the tannin ethanolic (FAl) and pure fractions (FTAN). The polysaccharide fraction (FPs) included in the crude material was also submitted to an esterification process to verify if som e reaction has occurred in this material. In order to remove the residual carbohydrate, the powder was treated with 1:1 ethanol-water mixture and filtered; then the residue was washed with hot water (80[degrees]C) and precipitated into ethanol; then the precipitate was filtered, dried, and submitted to the esterification reactions. The modification process was carried out considering that the macromolecule presents an average of four hydroxyl groups per single unit of tannin. Thus, 5.3 mmol of each tannin fraction and 8, 16, 24, 32, and 51 mmol of adipoyl chloride were used. The reaction was also examined under several reacting times (1, 2, 3, and 4 hr.) and temperatures (10[degrees], 30[degrees], 75[degrees], and 100[degrees]C). The characterization of these materials was done by infrared (IR) spectroscopy with 2 percent potassium bromide, by proton nuclear magnetic resonance ([H.sup.1] NMR) with chloroform (60 to 70 mg/mL) as solvent, and by gel permeation chromatography with dimethylformamide (0.1 mg/mL) a s solvent in a microstyragel column (Toyo Soda HLC 803, with Refractive Index Detector Model RI-2). The column was calibrated with a range of molecular weight (MW) standards, including ethylene polyoxide standards (25 000; 40 000; 73 000; 280 000; 1 200 000 MW), and ethylene polyglycol (4000; 7500; 14000 MW), and by the low MW ester compounds [C.sub.18][H.sub.26][O.sub.4] (306 MW), [C.sub.18][H.sub.16][O.sub.4] (298 MW), and [C.sub.12][H.sub.16][O.sub.2] (192 MW), which were prepared as model compounds.
RESIN FORMULATION AND ADHESIVE MIXES
The resins were prepared by using the crude tannin (R-1 nonmodified resin) and the esterified tannin (R-2 modified resin). First, the crude powder was treated with 1:1 ethanol-water mixture to remove the carbohydrates; the pH was then adjusted to 8.0 [+ or -] 0.1 with 25 percent sodium hydroxide (NaOH) aqueous solution. The temperature of the resulting solution was enhanced to 90[degrees] to 95[degrees]C, and the system was stirred for about 3 hours, brought to pH 6.5 to 7.0 by addition of acetic acid, cooled, and stored at room temperature. The resultant mixtures had a viscosity of 15 to 40 Mpa*s (cP). To dissolve the esterified tannin, the addition of some drops of an NaOH aqueous solution to the ethanol-water solution was required [16,22].
The adhesive formulation was prepared by stirring for 1 hour a mixture of 100 parts of R-1 or R-2 resins with 8 parts of paraformaldehyde (powder form) and 10 parts of wood flour (75 [micro]m, 200 mesh), then heating at 50[degrees]C for 30 minutes. A longer stirring time (3 hr.) without heating did not improve the adhesive. Brookfield viscosity  and pH of the adhesives were measured just before assembling the wood panels. The gel time of the mixtures were determined at pH 6.8 and 85[degrees]C in filler-free compositions . In order to study the influence of pH on the condensation reaction within the adhesive joint, pot-life determinations were run on the adhesive mixtures at 30[degrees]C with a Brookfield RVF viscometer. For comparative purposes, a typical commercial Cascophen [R] phenolformaldehyde (PF) resin, a resol-type resin, manufactured by Alba Industrias Quimicas, Brazil, was obtained.
PLYWOOD MANUFACTURE AND TESTS
Freijo (Cordia goeldiana) wood veneers  were conditioned to an equilibrium moisture content (EMC) of 8 to 10 percent. Two series of 20 by 20 cm three-ply panels, each ply 1 mm thick, were assembled according to ASTM D 1184-69  and compressed in a Carver press (Table 1). Series I panels were made with a glue spread of 100 g/[m.sup.2] single glueline (s.g.l) basis; Series II panels had 180 g/[m.sup.2] s.g.l. spread. In total, 12 panels were made for each adhesive.
After press, the panels were conditioned for 48 hours to an EMC of 8 to 10 percent and then the specimens were cut for mechanical tests. For each adhesive, the tension shear tests  included 12 specimens with no accelerated-aging treatment, 12 specimens that received 24-hour cold water immersions, and 12 specimens that were immersed in a 4-hour boiling water cycle. For the flexural tests , 10 dry specimens per adhesive type were prepared.
RESULTS AND DISCUSSION
In terms of product yields, the best results for the esterification of the natural condensed tannin were achieved at 75[degrees]C, with adipoyl chloride:tannin molar ratio of 6-10:1 and reaction time around 2 hours. Under these conditions, higher yields (72% to 75%) of modified material were obtained. The IR spectra of all tannin fractions taken before and after the esterification reaction showed peaks of absorption at 1155 to 1157 [cm.sup.-1], which is a characteristic of resorcinol-based compounds. But the absorption peak at 1200 [cm.sup.-1], also used to identify phenolic compounds, was not significant in these spectra [9-11,20,21]. The presence of hydroxyl groups suggests only a partial esterification of the tannin structure. The IR spectra did not show a presence of the polysaccharide fractions in the ester compounds at 1740 [cm.sup.-1] and 1728 [cm.sup.-1], nor were the fractions reacted with hydroxylamine and ferric chloride . The [H.sup.1] NMR detected a broad band with low intensity in the 6.8-t o-6.0-ppm region that was attributed to the five protons of the A- and B-aromatic rings; the protons linked to the [C.sub.2] and [C.sub.3] carbon atoms of the C-alicyclic ring were detected in a 5.4-to-4.3-ppm region. The proton band related to the C-ring hydroxyl group was absent in the esterified material, but was detected in a 4.0-to-3.6-ppm region in the nomnodified tannin. Besides this characteristic, a slightly broader band was detected in a 2.4-to-1.9-ppm region in the esterified tannin spectrum due to the methylene protons linked to the aliphatic chain with -O=C=O- substitution. A band in a 1.9 to 1.0 ppm that is characteristic of methylene protons was also observed [6,10].
The gel permeation chromatography applied to an esterified product of the purified fraction of tannin (FTAN) showed a curve in a 43.2-mL eluted volume, which corresponds to a low 1074 MW when plotted in the calibration curve. The original tannin material showed a MW greater than 1 000 000, which suggests an aggregate. However, some literature  states that natural condensed tannin MW is 3000, exceeding the limits of the calibration curve. The literature also states that tannin dissolved in dimethylformamide in a mycrostyragel column elutes all the material very near or at the exclusion volume. In this study, the limit was about 300 000 MW, utilizing methylene polyoxide. This high value indicates molecular aggregates that would have formed in this solvent .
RESIN FORMULATION AND ADHESIVE MIXES
A caustic treatment applied to the esterified tannin and original tannin (to hydrolyze the carbohydrates to simple sugars) is recommended by the literature to prevent the polysaccharide content in the crude material from increasing the viscosity [16,17,22]. In this exploratory work, the treatment has not affected the tannin esterified molecules, as the structure yields an intense peak at 1728 [cm.sup.-1] in the IR region, attributed to -C=O ester linkages.
The cross-link reactions in the adhesive were evaluated through the determination of gel-time and pot-life. The gel-time results at pH 6.8 and 85[degrees]C showed that the product formulated with the modified tannin is slower to cure (50 min.) than that from nonmodified tannin (20 mm). At pH 6.8, the resin solutions showed the same behavior as observed for the gel-time tests. However, at pH 7.5, the nonmodified tannin adhesive did not gel, reaching only 500 MPa.s (cP). The cure reaction in the alkaline range is much favored by the phenoxide type structure which is present in both nonmodified and modified tannin molecules. When the rate of reaction is enhanced there is a marked difference in the cure and in the viscosity. These effects are clearly important for application of the tannin-based adhesives on the wood surfaces.
PLYWOOD MANUFACTURE AND TESTS
Series II dry plywood shear tests (180 g/[m.sup.2] s.g.l. spread) for the modified and nonmodified (Group II) tannin adhesives showed an average shear strength of 1740 and 1960 kPa and an average wood failure of 17 and 93 percent, respectively (Table 2). These results could be expected after the chemical modification made to enhance the flexural properties of the tannin adhesive. Experiments showed that the increase of pH, heating time, and the reduction in the amount of glue on the veneer surface did not improve the joint performance. For comparison, dry plywood panel tests (Group II) run with a conventional PF adhesive at 100 g/[m.sup.2] s.g.l. basis presented 2630 kPa for average shear strength and 85 percent wood failure (Table 2).
Flexural load (kg) plots showed an interesting difference in behavior for both products; typical curves were obtained for all the plywood panel specimens in the flexural tests with the best adhesive resulting from the Series II (180 g/[m.sup.2] s.g.l. spread) panels (Fig. 1). With nonmodified tannin (AAP II), there was an average flexural strength of 20.40 Mpa (Table 2), with an abrupt decrease followed by rupture at 60 mm of deformation (Fig. 1). The modified tannin adhesive(AAP IV) showed an average flexural strength of 24.03 MPa (Table 2) followed by a gradual decrease with progressive deformation and rupture at 80 mm (Fig. 1). These values were analyzed by nonparametric statistics (data not provided) that show the validation of the results at a 5 percent significance level .
Flexible chains in the rigid molecule of natural tannin from black wattle bark were obtained by partial esterification with adipoyl chloride. The reaction retards the condensation with formaldehyde as expected. The compact tannin structure does not favor further condensation through methylol groups. This physio-chemical characteristic may be the cause for the brittleness shown by the tannin-formaldehyde condensates. Chemical modification of the tannin by introducing six-carbon atom chains acts as an internal plasticization of the macromolecule. This internal plasticization results in better distribution of forces throughout the laminate glue-lines, which leads to improved adhesive performance.
The authors are, respectively, Technologist, Instituto Nacional de Pesquisas da Amazonia (INPA), Caixa Postal 478, CEP 69011-970, Manaus, Brazil; and Research Scientists, Instituto de Macromoleculas, Caixa Postal 68525, CEP 21945-970, Rio de Janeiro, Brazil. This work was supported by the National Institute for Amazon Research, the Federal Univ. of Rio de Janeiro, and the National Council for Scientific and Technological Development, Brazil. Ms. Ana Barbosa gratefully acknowledges the assistance of Prof. Terry Sellers, Jr., Forest Prod. Lab., Mississippi State Univ., for kindly reviewing the manuscript and the International Tropical Timber Organization (ITTO) for the fellowship support. This paper was received for publication in December 1998. Reprint No. 8919.
(+.) Forest Products Society Member.
(1.) American Society for Testing and Materials. 1981. Standard methods for gel time and peak exothermic temperature of reacting thermosetting resins. ASTM D 2471-77. ASTM, West Conshohocken, Pa.
(2.) _____. 1981. Standard methods for viscosity of adhesives. ASTM D 1084-63. ASTM, West Conshohocken, Pa.
(3.) _____. 1981. Standard methods for flexural strength of adhesive bonded laminated assemblies. ASTM D 1184-69. ASTM, West Conshohocken, Pa.
(4.) _____. 1981. Strength properties of adhesives in plywood type construction in shear by tension loading. ASTM D 906/64. ASTM, West Conshohocken, Pa.
(5.) Barbosa, A.P. 1990. Chemical modification of black wattle (Acacia mearnsii) tannins to use in adhesives. MSc thesis. Instituto de Macromoleculas, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil. 196 pp. (in Portuguese, an English abstract is available.)
(6.) Brandt, E.V., D.A. Young, D. Ferreira, and D.G. Roux. 1987. Synthesis of condensed tannins. Part 20. Cycloconformations and conformational stability among derivatives of 'angular' tetraflavonoid profisetinidins. J. of the Chemical Soc., Perkin Transactions I. pp. 2353-2360.
(7.) Campos, H. 1983. Nonparametric Experimental Statistics. 4th ed. E.S.A. Luiz de Queiroz, Universidade de Sao Paulo (USP), Piracicaba, Brazil. 349 pp. (in Portuguese.)
(8.) Foo, L.Y. and L.J. Porter. 1986. Some recent advances in the chemistry of condensed tannins (proanthocyanidin polymers) relevant for their use as industrial chemicals. Appita 39(6):477-480.
(9.) Hemingway, R.W. and G.W. McGraw. 1978. Formaldehyde condensation products of model phenols for conifer bark tannins. J. of Liquid Chromatography 1(2):163-179.
(10. _____, _____, and J.J. Karchesy. 1981. Condensation of ortho- and parahydroxybenzyl alcohols with catechin as a model for use of methylolphenols as crosslinking agents in conifer bark polyflavonoid formulations. In: 2nd. Weyerhaeuser Science Symposium-Phenolic Resins: Chemistry and Applications, Tacoma, Wash. p. 273.
(11.) Kishimoto, J., T. Fukuta, T. Sakuno, and I. Furukawa. 1981. Studies on the IR absorption spectra of tannins. Bull. of the Eac. Agri., Tottori Univ., Japan. 33:65-69.
(12.) Loureiro, A., M.F. Silva, and J.C. Alencar. 1979. Characteristics of wood in the Amazon. Vol. I. CNPq/INPA, Imprensa Oficial do Estado do Amazonas, Manaus, Brazil. p. 166. (in Portuguese.)
(13.) Mano, E.B. and A.P. Seabra. 1987. Practice of Organic Chemistry. 3rd ed. Editora Edgard Blucher Ltda, Rio de Janeiro, Brazil. p. 99. (in Portuguese.)
(14.) Nekrasov, V.V. 1978. Practical Organic Chemistry - A Basic Course. Mir Publishers, Moscow. p. 282.
(15.) Pizzi, A. 1983. Tannin-based wood adhesives. In: Wood Adhesives: Chemistry and Technology. A. Pizzi, ed. Marcel Dekker, Inc., New York. pp. 177-246.
(16.) _____. 1978. Wattle-base adhesives for exterior grade particleboard. Forest Prod. J. 28(12):42-47.
(17.) _____ and A. Stephanou. 1994. A C-13 NMR-study of polyflavonoid tannin adhesive intermediates. Part 2. Colloidal state reactions. J. of Appl. Polymer Sci. 51(13):2125-2130.
(18.) _____ and _____. 1994. A C-13 NMR-study of polyflavonoid tannin adhesive intermediates. 1. Noncolloidal performance determining rearrangements. J. Applied Polymer Sci. 51(13):2109-2124.
(19.) _____ and _____. 1992. Theory and practice of non-fortified tannin adhesives. Holzforschung und Holzverwertung 44(4):62-68.
(20.) _____ and G.M.E. Daling. 1980. Infrared spectrometrical determination of resorcinol content of tannin-based wood adhesives. Holzforschung und Holzverwertung 32(2):39-41.
(21.) _____ and H.O. Scharfetter. 1978. The chemistry and development of tannin-based adhesives for exterior plywood. J. of Appl. Polymer Sci. 22:1745-1761.
(22.) _____, _____, and E.W. Kes. 1981. Adhesives and techniques open new possibilities for the wood processing industry. Part I. Experience with tannin based adhesives. Holz als Roh-und Werkstoff 39: 85-89.
(23.) Roux, D.G. 1955. The fractionation of black wattle extract. Leather Industries Res. Inst., Grahamstown, South Africa. pp. 48-51.
(24.) _____, D. Ferreira, H.K.L. Hundt, and E. Malan. 1975. Structure, stereo-chemistry, and reactivity of natural condensed tannins as basis for their industrial application. Appl. Polymer Symp. 28:335-353.
(25.) Tanac S.A. 1987. Personal communication on tannin analysis. Montenegro, Brazil. 2pp.
Plywood manufacturing conditions and adhesive mix characteristics. Manufacture conditions Parameters Series I Series II Glue spread 100 g/[m.sup.2] s.g.l. 180 g/[m.sup.2] s.g.l. Open assembling time 2 min. 2 min. Press temperature 120[degrees]C 120[degrees]C Pressure 130 kPa 980 kPa Pressing time 2 min. 2 min. Mix viscosity 485 MPa.s(R-1) 225 MPa.s(R-1) 370 MPa.s(R-2) 125 Mpa.s(R-2) Mix pH 7.8 to 7.9 7.3 to 7.4 Parameters Phenol-formaldehyde resin Glue spread 100 g/[m.sup.2] s.g.l. Open assembling time 30 min. Press temperature 135[degrees]C Pressure 980 kPa Pressing time 4.5 min. Mix viscosity 590 MPa.s Mix pH 12 Average shear and flexural strength results of plywood bonded with a PF control, a modified tannin, and a nonmodified tannin adhesives. Dry test Adhesive formulation code and test grouping [a] Shear strength [b] Wood failure (kPa x [10.sup.3]) (%) AAP I, Series I Group I 1.12 (0.05) [c] Delaminated nonmodified Group II 1.35 (0.10) Delaminated AAP II, Series I Group I 1.11 (0.07) 56 modified Group II 1.45 (0.07) 60 AAP III, Series II Group I 1.41 (0.18) 94 nonmodified Group II 1.96 (0.19) 93 AAP IV, Series II Group I 1.33 (0.14) 88 modified Group II 1.74 (0.14) 17 AFF PF Group I 2.37 (0.06) -- control resin Group II 2.63 (0.14) 85 24-hr. water soak test Adhesive formulation code and test grouping [a] Flexural strength [b] Shear strength [a] (kPa x [10.sup.3]) (kPa x [10.sup.3]) AAP I, Series I 1.14 (0.14) nonmodified -- 1.56 (0.05) AAP II, Series I 0.74 (0.04) modified -- 0.85 (0.04) AAP III, Series II 20.40 1.60 (0.06) nonmodified (0.69) 1.82 (0.15) AAP IV, Series II 24.03 0.6 (0.13) modified (0.78) 0.80 (0.08) AFF PF -- control resin -- -- 4-hr. boil test Adhesive formulation code and test grouping [a] Wood failure Shear strength [a] Wood failure (%) (kPa x [10.sup.3]) (%) AAP I, Series I 14 1.35 (0.12) 75 nonmodified 9 1.59 (0.06) 76 AAP II, Series I Delaminated 0.60 (0.07) Delaminated modified Delaminated 0.84 (0.07) Delaminated AAP III, Series II -- 1.41 (0.14) -- nonmodified 85 1.68 (0.09) 66 AAP IV, Series II Delaminated 0.60 (0.06) Delaminated modified Delaminated 0.75 (0.06) Delaminated AFF PF -- -- -- control resin -- -- --
(a.)The AAP I and AAP III adhesives were formulated with nonmodified tannin while the AAP II and AAP IV adhesives were formulated with modified tannin, under the conditions in Table 1. Series I and II refer to adhesive spreads of 100 and 180 g/[m.sup.2] s.g.l., respectively. Groups I and II are strength value groupings per ASTM D 906/64.
(b.)Each value is an average (standard deviations in parentheses) of 3 panels, 6 specimens per panel (18 total specimens).
(c.)The flexural strength values are the average results (standard deviations in parentheses) of a panel per each adhesive, 10 specimens per panel.
|Printer friendly Cite/link Email Feedback|
|Author:||BARBOSA, A.P.; MANO, E.B.; ANDRADE, C.T.|
|Publication:||Forest Products Journal|
|Date:||Sep 1, 2000|
|Previous Article:||COMPATIBILITY OF SOME TROPICAL HARDWOODS SPECIES WITH PORTLAND CEMENT USING ISOTHERMAL CALORIMETRY.|
|Next Article:||Popular Lumber Design Tool Updated.|