![[BioMatNet Database - FAIR Program]](../images/Fair.gif) |
FAIR-CT95-0137
Natural Tannin-Based Adhesives for Wood Composite Products of Low or No Formaldehyde Emission |
 |
Contract No: | FAIR-CT95-0137 |
| Date Prepared: | January 1999 | |
| Source: | Final Report Executive Summary |
Four of the European tree species examined (Douglas fir, Radiata pine, Maritime pine and Calabrian pine) have barks that are suitable for the production of tannin extracts in sufficient yield, with a high tannin content and more than adequate reactivity to produce commercially and technically viable tannin adhesives for wood products. Their geographical distribution is also good in relation to established industrial extraction facilities. The tannins are similar irrespective of source, only slightly differing in degree of polymerisation, relative proportion of procyanidin or prodelphinidin tannins and reactivity with formaldehyde. However, there are some differences in the ratio of tannin to non-tannin components that affect viscosity and have to be taken into account when formulating adhesives. In contrast, Pinus sylvestris and Eucalyptus rostrata proved unsuitable production of wood adhesives. The first because the yield of extractable material was low and the second due to a low reactivity with formaldehyde
While direct solvent extraction of the bark gave unacceptable results both in yield and in composition, solvent extraction of dried water extracts of tannin gave higher purity tannins and was found to be a very acceptable procedure. With this approach solvent extraction always produced tannin extracts of much higher polyphenolic content and much lower carbohydrate content than water extraction. In general it did not alter the composition of the tannin extract or its polyphenolic composition. However, it appears to remove not only most of the carbohydrates but also some of the polyphenolic fractions, possibly some higher molecular weight fractions, although this is not quite clear. While extraction with ethyl acetate/acetone generally yielded extracts of slightly better polyphenolic content and lower carbohydrates content it generally produces a much lower yield of usable extract. Conversely, extraction with acetone alone (or preferably with 90% acetone and 10% water) gave much higher yields and only marginally lower polyphenolic and marginally higher carbohydrate contents. Hence, this method was more attractive both from a technical and an economic point of view. Upgrading of the polyphenolic content of water extracts of tannins by addition of solvent extracts of tannin led to the preparation of plywood panels that satisfied the quality requirements of the relevant European Norm standards for exterior plywood.
The nature of the extracts were confirmed not only by traditional wet chemistry analysis, but also by infra red spectrophotometry, by solid state 13 C NMR and by high-pressure liquid chromatography. The gel times of solvent extracted tannins appeared to be only marginally faster than those of water extracted tannins. Results indicated that the proportion of carbohydrates present had in general only a diluting effect on the inherent reactivity of the polyphenolic components of procyanidin-type (pine) tannin extracts.
Electron spin resonance (ESR) and thermomechanical analysis (TMA) studies were carried out to investigate the presence or lack of interference by ionic hardeners on the autocondensation of polyflavonoid tannins. Results indicated that in certain cases hardening by ionic co-reactants could be coupled with the simultaneous hardening of the tannin by radical autocondensation. Some co-reactants tend to depress the tannin radical autocondensation, but still leave a small contribution of this reaction to the formation of the final cross-linked network. Other co-reactants appear to enhance formation of the final network by synergy between ionic and radical mechanisms, while still others do not show any interference between the two types of reaction. Mechanisms describing the interaction between the two reactions have been proposed and discussed.
The thermomechanical analysis (TMA) study focussed directly on bonded joints during the process of adhesive hardening, on the synergy and interference between polycondensation and autocondensation of procyanidin and profisetinidin/ prorobinetinidin-type polyflavonoid tannins leading to network formation and hardening. These studies confirmed that at the higher temperatures characteristic of the hardening of tannin-based wood adhesives polycondensation could be coupled with simultaneous hardening of tannins by autocondensation. Results again suggested that some co-reactants appear to depress tannin autocondensation while still leaving a small contribution of this reaction to the formation of the final cross-linked network. Other co-reactants instead appear to enhance formation of the final network by synergy between the two condensation mechanisms.
Curves obtained by plotting the increase of modulus as a function of temperature obtained by thermomechanical analysis (TMA) of joints bonded with different polyflavonoid tannins/hardener systems show a succession of different plateaus, that were converted to peaks by plotting the first derivative. These plots showed considerable variability in relative intensity and sometime in relative position on the thermogram. This variability appeared to be due to the superimposition of the series of polycondensation reactions of the different phenolic nuclei, which constitute the flavonoid-repeating unit of each tannin with different hardeners. This was reflected in the pattern of plateaus and peaks induced by the non-cross-linked entanglement networks formed by the linear increase of the polymer in the early stages of the polycondensation. This succession of stages in the thermograms is described for both a procyanidin type and for a profisetinidin/prorobinetinidin type tannin. It forms the basis of the TMA results for the polycondensation of simple phenolic model compounds of the tannin nuclei as well as of linear, non-cross-linkable model polymers in the case of entanglement networks. The effect of different polycondensation hardeners on the extent of tannin autocondensation indicates that the more marked the polycondensation the earlier the entanglement networks appear and the lower is the temperature at which they appear. Autocondensation was shown to always participate to the formation of the final, hardened tannin network, the extent depending on the hardener used. However, the proportion of the network due to polycondensation appears to be related to the extent of water resistance of the final network, while the contribution of tannin autocondensation appears to be limited to the dry strength of the network.
The conclusion of the whole study is that when dry strength increases are required from the tannin adhesives in the production of interior grade panels it is definitely possible to increase the strength of the panel by favouring the autocondensation reaction. This enables a decrease in use of formaldehyde-yielding hardeners and a consequent decrease of formaldehyde emission by the panel. When water resistance of the panel needs to be improved, that depends on the level of polycondensation between the tannin and formaldehyde yielding hardeners, other techniques have to be used to decrease formaldehyde emission without loosing strength. Such options include addition of urea, or use of hexamine as hardener. It should be noted that urea gives greater strength than when formaldehyde is used alone, as it also participates in cross-linking and hence urea does not only work as a formaldehyde emission reducer.
Peculiar behaviour was found in tannin-based and other wood adhesives when hexamethylene tetramine (hexamine) was used as a hardener. In tannin adhesives using hexamine as hardener a flow problem was identified during hot curing of this adhesive/hardener system. The effect on adhesive performance was evaluated. The reason suggested for the problems with tannin/hexamine, providing a clear theoretical justification to the applied findings is that under many application conditions hexamine is not a formaldehyde-yielding compound, providing extremely low formaldehyde emissions in bonded joints. Evidence from 13C NMR confirmed that the main decomposition (and recomposition) mechanism of hexamine is not directly to formaldehyde but proceeds mainly through the formation of reactive imines rather than methylene bases. A very small amount of iminomethylene bases may also be formed. Results also confirmed that any species (tannin, resorcinol or other highly reactive phenol, melamine or other highly reactive amine or amide, organic or inorganic anion) with a strong negative charge under alkaline conditions is able to react with the intermediate species formed by decomposition (or recomposition) of hexamine. This reaction occurs far more readily than formaldehyde production, explaining the capability of wood adhesives formulations based on hexamine to give bonded panels of extremely low formaldehyde emission.
If no highly reactive negatively charged species are present then decomposition of hexamine proceeds rapidly to formaldehyde as reported in previous literature. The elucidation of this hexamine decomposition mechanism, and a scanning electron microscopy (SEM) investigation enabled a reason to be suggested for the formation of ambient temperature stiff gels in tannin/hexamine glue mixes without curing and to propose chemical structures for linear polymers formed.
Kinetics of the reaction of gelling of tannin-resorcinol-formaldehyde (TRF) and tannin- formaldehyde (TF) adhesive resins were followed by low-resolution impulsed proton NMR. S-shaped gel curves based on the relative variation of molecular mobility of solid phase protons in relation to liquid phase and total protons were obtained. The method appears to work well for both resins.
Glue-lines in boards made of wood particles using high moisture tolerance tannin-based adhesives were observed to show the unusual behaviour of producing a melting effect in the wood cell walls directly in contact with the adhesive. The high moisture content of up to 29% used in the preparation of these exterior grade particleboards and OSB industrial panels at standard pressing times causes considerable flow of the amorphous lignin and hemicelluloses in the wood directly in contact with the adhesive. This led to a composite, compact interphase in which loosened wood fibres are drowned in a hardened adhesive matrix. In this interphase the anatomical characteristics of the wood cell walls appear to be lost, apparently due to the marked decrease of the glass transition temperature of hemicelluloses and lignin at the high moisture contents used. This effect was replicated in boards prepared with a synthetic MUF adhesive using a moisture content level much higher than what this resin is normally capable of tolerating. However, this required considerable, uneconomic lengthening of the pressing time. It became clear only tannin-based adhesives have the characteristic of tolerating very high moisture contents at relatively fast press times, an economically much sought-after characteristic.
Applied studies indicated that tannin adhesives hardened with hexamine gave better mechanical performance at pressing temperature between 160ºC and 170ºC. This observation supported fundamental studies suggesting that too high a temperature will not allow sufficient formation of the reactive iminomethylene bases due to the faster decomposition to formaldehyde. Thus, at higher press temperatures (190ºC -210ºC), as normally used today in particleboard manufacture, a considerable proportions of hexamine will decompose to formaldehyde which will then react with the tannin. This leaves considerable amounts of free aminated compounds in the hardened glue line, which decreases its water resistance and performance. This also leads to the need to use relatively long press times. At the optimum pressing temperatures of 160ºC -170ºC formation of the fast reacting iminomethylene bases is maximised and formaldehyde formation is minimised. This allows for faster press times at a lower temperature (a very valuable characteristic) without impairing water resistance and mechanical performance. Hence, it is evident that greater care in the control of the board manufacturing parameters needs to be taken when using industrially tannin/hexamine systems than when using tannin/ paraformaldehyde systems.
The industrial plant trial results (a 2-hour industrial plant run) obtained with a tannin/hexamine system were as good as with the tannin/paraformaidehyde systems, as long as a fast flavonoid tannin such as pine was used. This requires a lower press temperature (but at equal press time) to be used, with resin treated wood particles of higher moisture content and a pressure profile as a function of press time changed to favour lower maximum pressures and longer degassing times. The need for the latter is a consequence of the pine tannin/hexamine system needing (rather than just tolerating) a higher moisture content in the wood particles. The need for higher moisture contents is the consequence of two factors, namely:
Hardening systems for fast reacting tannin adhesives for particleboard yielding results similar to those obtainable with hexamine hardener, both for the board internal bond (I.B.) strength performance and low formaldehyde emission, can also be obtained through the attainment of stable benzylamine bridges. These arise from a different mechanism than that of hexamine decomposition. Ammonia and ammonia-yielding compounds such as ammonium salts of weak acids, ammonium carbonate and particularly ammonium acetate, can be used in conjunction with paraformaldehyde for this purpose. In these cases gelling, curing acceleration and formation of benzylamine bridges in the hardened network are caused by reaction of formaldehyde with the tannin and the ammonia without passing through base stage as in the mechanism of decomposition of hexamine, but rather through the alternative route. NMR evidence for this mechanisms has been reported and a European patent has been obtained for this process by one of the partners.
A further process for zero formaldehyde emission was also developed and a European patent applied for. This consists in using tri(hydroxymethyl) nitro methane as hardener of the tannin. This system has proved particularly useful for the production of hardboard. Although plant trials had not been carried out when the project ended these were planned for 1999.
Research work on oriented strandboard (OSB) led to the pine tannin + paraformaldehyde + urea system being optimised from laboratory to semi-industrial scale. The work concentrated on the application of new procyanidin tannin + paraformaldehyde + urea adhesive systems to structural exterior grade oriented strandboard (OSB) manufactured in the laboratory with cycles imitating very modern continuous press lines (Kontiroll). This is the first time that tannin adhesives have been used in this application. It is also the first time using laboratory press cycles to reproduce a continuous press line: an application in which synthetic phenolic adhesives are known to give considerable problems (and hence they are not used for this application being replaced by other synthetic adhesives).
Work was concentrated on developing new and suitable laboratory press cycles (which were achieved by the middle of 1997) and on obtaining a good internal bond (I.B.) strength and other properties of the panel under varying conditions of application. At the same time efforts were made to obtain the correct density profile vs. thickness of the panel, so that it had the required balance of final properties. The results obtained were good and the use of the new pine tannins coupled with a somewhat more traditional type of hardening gave acceptable results for standard specifications. In addition new press cycles, in which the maximum pressure used was about half of that normally required for OSB (namely about 25 kg/cm 2 against the traditional value of 45 kg/cm2), were developed. In these cycles the decrease of pressure as a function of time, after the maximum of pressure is achieved, follows a slope slightly different from traditional OSB press cycles. These new press cycles result in a better profile of board density as a function of board thickness and hence a higher relative density of the board core leading to better internal bond (I.B.) strengths. Some other properties are also improved.
It was clear that these new press cycles are usable not only for tannin adhesives but for any other board adhesive with resin treated particles of higher moisture content. In this context the ability to use adhesives (such as the tannin-based products for example) capable of curing under high moisture content is equivalent to achieving the advantages of steam injection OSB processes without the steam injection part of the process, which can be substituted for by the high moisture content of the resinated wood particles. The much lower maximum pressure used is of particular interest as it also allows the manufacture with much lighter OSB presses with consequent economical and technical advantages.
The pine tannin adhesive for OSB system, the application conditions of which have now been perfected at laboratory and pilot plant level, needs now to be translated to a full scale industrial plant trial. In this context a few semi-industrial plant trials aimed at producing OSB panels of dimensions 2.5x 1.5 metres were carried out: the panels produced in this manner gave excellent results and confirmed that the technology developed at laboratory level is perfectly reproducible at industrial level and yields the same excellent results. At this stage it appears that the tannin adhesives with this type of hardener might be the only type of phenolic adhesives to be usable for this type of application, as well as being the only adhesive which appears to produce OSB panels in class 4 (marine). None of the MUF, PF, PMUF and isocyanates being capable of this fat industrial level. Both of these findings constitute a considerable 'first'.
As regards MDF, the standard formulation which appeared to work well was based on a tannin + formurea (a formaldehyde concentrate stabilised with urea in water and of composition urea= 20%-24% and formaldehyde, 50% - 58% according to source) plus additional urea to decrease formaldehyde emission. This proved to give good results as a liquid hardener and gave acceptable laboratory exterior type MDF panels. An early industrial plant trial (2 hours) was held in the biggest MDF plant in France and the results obtained were very encouraging. The system did not give any problems with the MDF blow line and operation was smooth and without any blockages. The only problem was that due to an error of calculation the panels were produced at lower density than normal MDF. In spite of this, they showed equal strength to the interior type MDF that was routinely produced by the same factory as well as having good exterior properties.
It is remarkable that excellent performance was obtained at lower density, as this was a new trend in being developed by commercial MDF panels at the time, with a wish to lower density but maintain performance. In this context the results obtained were indeed very 'modern'.
Excellent results were also obtained both at the laboratory and at the industrial plant level in the application of tannin/furfuryl alcohol adhesives to produce tempered hardboard. In this context a resin based on tannin/furfuryl alcohol, hence having very low, and almost no formaldehyde emission (this only coming from the heating of wood) was used in two full scale production trials of respectively 8 and 9 hours, for a total of approximately 2000 industrial hardboard panels. The results obtained were comparable (and sometime slightly better), at parity of all the other conditions, to those obtained with synthetic phenolic resins. However, this process had the advantage of using a natural non-polluting material (the tannin).
These results were extended to the preparation of hardboards with tannin/waste wood carbohydrates in presence of formaldehyde to reduce even further the cost of the natural raw material. In this case a mix of tannin/waste lignocellulosic carbohydrates and formaldehyde was used in a industrial plant. These trials gave acceptable results, indicating that the cost of the tannin could also be decreased by adding, at least for hardboards, the mix of hydrolysable tannins+ carbohydrates (mainly hydrolysed hemicelluloses) + lignin coming as waste from the defribrator producing the hardboard fibre.
Plywood panels were also made using the reaction of tannin with PMDI (polymeric isocyanate), without any addition of formaldehyde: the laboratory panels gave results acceptable for interior panels but were not suitable for exterior use. For this reason this line of investigation was discontinued. What was continued was an investigation of pMDI/tannin/paraformaldehyde mix to use for gluing under difficult conditions. This type of formulation was optimised at laboratory level, but as mentioned above, has only interest when formulations of particularly good performance are needed to bond difficult to glue timbers. In other cases the price of the adhesive is slightly on the high side and the presence of a synthetic product such as MDI renders it less attractive. Plywood was not actively pursued as a target but was used rather as one of the testing vehicles for adhesives formulations. However a plant trial of curved plywood for chair seats bonded with tannin/paraformaldehyde/urea formulations was carried out in a furniture factory in Italy. The results were technically encouraging and comparable to what obtained by the factory with UF resins, but as the type of product dominating this niche field is the very low cost urea-formaldehyde resin (just for interior products) there was no real interest, economical or technical, to pursue this further.
Preparation of tannin-resorcinol-formaldehyde cold set adhesives for glulam and fingerjointing by the traditional rather than the 'honeymoon' route (honeymoon pine tannin cold set adhesives that have now been industrial outside Europe for at least 9 years) were investigated. While very successful for the slower flavonoid tannins such as quebracho and mimosa, they were not very successful with pine tannin. This is because the A-rings of pine tannins (procyanidins) are much more reactive than the resorcinol chemicals added and as a consequence tannin reaction with formaldehyde to gel (cutting out resorcinol participation) becomes the favoured reaction. As a consequence this route was not pursued further.
A further route investigated was to decrease formaldehyde in PRF/tannin honeymoon system by relying on maximising autocondensation through the use of silica catalyst in alkaline environment. Although encouraging results were obtained, the joint bonded without any formaldehyde hardener only achieved half the strength necessary to pass the relevant standard and was suitable only for interior use. At very low amounts of formaldehyde added this dry strength performance was improved to around 85% of what was necessary to pass the relevant standard, while exterior strength was about half of what was required by the standard. While these results were encouraging they could not be improved to reach the levels required by the standard.
© Copyright 2006 Policy Statements
Updated
by CPL Press:
03/07/2007
- biomatnet@biomatnet.org
 |
 |
 |
News |
 |
Events |
|
|