Development of a novolak-based hydroxymethylated resorcinol coupling agent for wood adhesives.
A hydroxymethylated resorcinol (HMR) coupling agent was previously shown to enhance the exterior durability of bonds between wood and epoxy or other thermosetting wood adhesives, as well as bonds between phenol-resorcinol-formaldehyde and southern pine treated with chromated copper arsenate. When originally developed, HMR had no shelf life, so the resorcinol and formaldehyde components had to be mixed and reacted for several hours before being ready for application. Now we have developed an HMR coupling agent in the form of a liquid novolak prepolymer that remains chemically stable in storage. The novolak HMR is deficient in formaldehyde in its prepolymer state, which makes it stable. With the addition of formaldehyde, the reaction resumes. In the most effective formulation, one-quarter of the formaldehyde was incorporated into the novolak, and three-quarters of the formaldehyde was added to the novolak when it was ready for use. In this study, the reactivated HMR was usable immediately and up to at least 7 h ours after the final addition of formaldehyde. The new form of HMR is as effective a coupling agent as the original HMR (based on ASTM D 2559), and capable of producing epoxy bonds to wood that easily pass the delamination requirements (5% maximum for softwoods) of ASTM D 2559 after exposure to the severe accelerated aging test. Differential scanning calorimetry and carbon- 13 nuclear magnetic resonance were used to determine the reactivity and changing chemical structure of the novolak HMR.
In 1995, researchers at the USDA Forest Service, Forest Products Laboratory, reported on a coupling agent that greatly enhanced the durability of the bond between adhesive and wood in exterior use (Vick et al. 1996, 1995). This coupling agent was hydroxymethylated resorcinol (HMR), which is a dilute, reactive mixture of resorcinol, formaldehyde, and alkaline catalyst in water. Epoxy adhesives do not develop bonds of exterior durability to wood, nor does phenol-resorcinol-formaldehyde (PRF) adhesive effectively bond to southern pine treated with chromated copper arsenate (CCA) preservative. However, if the wood is first primed with HMR, bonds of the epoxy to wood become durable enough to meet the requirements of the most demanding standards. As a result, it is possible to make highly durable epoxy bonds between fiber-reinforced plastics and wood because of the enhanced adhesion of epoxy to wood afforded by the HMR coupling agent (Vick et al. 1998, 1996, 1995; Vick and Okkonen 1997; Vick 1996). Also, PRF adhesi ve makes similarly durable bonds to HMR-primed CCA-treated southern pine (Vick 1996, 1995).
The original HMR had characteristics that inhibited its use in industrial processes. First, all ingredients had to be carefully weighed before each use, which was time consuming and subject to error. Second, HMR had no storage life, so each batch had to be mixed and reacted 4 hours before use, which could be a serious problem if production was interrupted unexpectedly. Production would then be delayed until a fresh batch could be prepared. And third, since a batch was only useful for about 4 hours, two batches would be needed for each 8-hour shift.
Research on HMR during its development provided information on reactions occurring at various times during the period when HMR effectively improved adhesion (Christiansen 2000; Vick et al. 1998). The formaldehyde, which is preferably added last, reacts during the first few minutes with resorcinol to predominantly form hydroxymethyl groups on the resorcinol molecules. The hydroxymethyl groups on resorcinol react with other resorcinol molecules to produce methylene bridges, forming dimers, trimers, and higher oligomers of resorcinol-formaldehyde resins. Methylene bridge development was detected by carbon-13 nuclear magnetic resonance ([C.sup.13]-NMR) after an initial 20 minutes of reaction, but development continued for at least 19 hours. Because hydroxymethylation reached its maximum after 1 hour, that reaction alone was insufficient to account for all the properties necessary for the effective coupling action of HMR. Maximum adhesion occurs after 4 hours of reaction. We surmised that an increase of molecular weight was also a factor in improving adhesion. A higher molecular weight component would ensure that some of the coupling agent remained on the surface to bond to the bulk adhesive, while much of the lower molecular weight component penetrated deeply into the cell wall structure to establish a firm hold on the wood. A broad molecular weight distribution with a low molecular weight component that could penetrate cell walls was advocated by Nearn (1974) as the way to produce the most durable bonds to wood.
Methods of stabilizing HMR were considered. Two trials of freeze-drying were performed but were not promising, because HMR continued to react during the freeze-drying process to yield an insoluble and cured resin mass.
Another way to chemically stabilize HMR would be to prepare it in a novolak form similar to that used in resorcinol-formaldehyde resin adhesives. This meant that the resorcinolic component would be reacted with less than the total amount of formaldehyde needed for complete curing. However, enough formaldehyde would be used to produce a product with an average molecular weight, at a selected level, beyond the monomeric stage. Because formaldehyde eventually would be consumed to produce methylene bridges between aromatic rings, the novolak could not react further and would remain stable during long-term storage. A novolak resorcinol-formaldehyde could be shipped to a wood-processing mill in this condition. Once a separate formaldehyde component was added to the novolak component to initiate the final curing reaction, little waiting time would be required before the adhesive could be used. This idea was pursued.
The major objective of this work was to determine whether a resorcinol-formaldehyde novolak in a dilute, alkaline, aqueous solution could be reacted with additional formaldehyde solution (formalin) to make a ready-to-use and effective HMR coupling agent. A secondary objective was to determine the useful application life of the reactivated HMR. Useful life is the time that elapses between the formaldehyde being mixed in and the HMR losing its effectiveness, which is the point when the resulting epoxy bonds would fail to meet the delamination requirements of ASTM D 2559 cyclic exposures (ASTM 1998). A third objective was to determine what formaldehyde-to-resorcinol molar ratio (F/R) of the novolak would produce the most effective coupling agent, as indicated by the ASTM D 2559 tests of resistance of adhesive bonds to delamination.
The chemicals for the novolak HMR (n-HMR) were crystalline resorcinol, 37.1 percent aqueous formaldehyde solution (formalin), 3 M sodium hydroxide, and deionized water. The sources for these chemicals were given in a previous publication (Vick et al. 1998). The wood was Douglas-fir lumber at approximately 9.5 percent equilibrium moisture content (EMC), knife-planed to 19-mm (3/4-in.) thickness the day before the primer was applied.
The adhesive was FPL 1A epoxy, which was formulated as shown in Table 1. These ingredients were mixed just before use. Diglycidyl ether of bisphenol A (DGEBA) epoxy resin was D.E.R. 331 from Dow Chemical Company (Midland, Michigan); benzyl alcohol (99%) was from Aldrich Chemical Company (Milwaukee, Wisconsin); the hydrophobic fumed silica was Cab-O-Sil N70-TS grade from Cabot Corp. (Tuscola, fllinois); and the triethylenetetramine hardener was D.E.H. 24 from Dow Chemical Company.
In these experiments, the final n-HMR was to have the same composition as the original HMR. Formaldehyde-deficient novolaks at several F/R molar ratios were formed by mixing the original HMR ingredients but leaving out predetermined portions of formaldehyde solution. Because of the previously measured low degree of reaction of HMR at its earliest effectiveness (about 25% of total) (Vick et al. 1998), we expected that the FIR molar ratio of the novolak should be considerably below the 0.7 to 0.8 molar ratios of typical resorcinol novolak adhesive resins. We decided to test novolak compositions near 25 percent of the 1.54 total F/R molar ratio of HMR, which is 0.39. Exploratory tests indicated that a molar ratio of 0.23 was too low (Christiansen et al. 2001). The percentages of ingredients in test novolaks at three molar ratios are given in Table 2.
The n-HMR was applied between 0 and 3 hours (up to 7 hr. for molar ratio 0.39) after the final formaldehyde was added. Dodecyl sulfate sodium salt (98% pure, from Aldrich Chemical Company) was added to the n-HMR (at 0.5% by weight), just prior to application, to aid the wetting of the resinous wood surfaces.
Free water can decrease the effectiveness of bonding of epoxy adhesives with wood, so water must be evaporated from the HMR-primed wood surfaces before the adhesive is spread. Because epoxy cure is much more sensitive to wood moisture content than most present wood adhesives, water was dissipated by conditioning the primed wood overnight at 50 percent relative humidity (RH) and 23[degrees]C. However, using the previous version of HMR to improve the bond between aqueous PRF adhesive and CCA-treated wood, we found that only 1 hour of drying was necessary.
A control set of epoxy-bonded assemblies without the coupling agent was also tested alongside those tested with the novolak-based coupling agents.
Residual reactivities of various n-HMR coupling agents were measured as a function of reaction time by differential scanning calorimetry (DSC). The novolaks were prepared weeks before addition of the final formaldehyde. All the novolak-forming reactions were monitored to ensure that all their reactivity had dissipated within 24 hours. Reaction times were measured from the time the final formaldehyde was added to the novolaks. The equipment and techniques used for DSC analysis of residual heat of reaction were the same as those previously reported (Vick et al. 1998), except that the starting temperature for each scan was 20[degrees]C.
Carbon-13 NMR was used to record changes of resonance signals with time for chemical groups formed after adding the final formaldehyde to the novolak. The equipment and techniques for [[blank].sup.13]C-NMR were described in a previous paper (Vick et al. 1998). Carbon-13 formaldehyde was used to make the novolak with F/R = 0.39 and, 17 days later, to activate the novolak to create the coupling agent with final F/R = 1.54. Only the second reaction, that of the novolak with the final formaldehyde, was followed by spectroscopy.
Preparation and testing of delamination specimens
Each laminated assembly was made up of six pieces of lumber, each 19 mm thick, 76 mm wide, and 305 mm long. Wood surfaces to be bonded were brushed with 0.15 kg/[m.sup.2] HMR. After the lumber pieces were dried overnight, epoxy adhesive was applied to the wood surfaces by roller to a total of 0.34 kg/[m.sup.2]. Laminae were quickly joined together to build an assembly. Closed assembly time was 60 minutes for the first bondline and 50 minutes for the last. After the assemblies were placed in a cold-press, the pressure was increased until a small amount of adhesive squeezed out along the full length of each bondline. In a previous study, the pressure was measured at about 69 kPa. The assemblies were kept under pressure overnight. Then, to ensure that all bondlines were cured to the same degree, the assemblies were heated to 71[degrees]C for 5 hours. The RH of the oven was increased to maintain the EMC of the wood so that bond lines would not be stressed by shrinkage of the wood while the resin cured.
Three delamination specimens were cut as 76-mm-long cross sections from each six-ply assembly. Each set of three specimens from one assembly was considered one replicate. The specimens were subjected to the severe cyclic delamination procedure of ASTM D 2559 (ASTM 1998). Industry standard ANSI/AITC A 190.1-1992 (AITC 1992) specifies that all wet-use adhesives intended for exterior service in structural lumber laminates must be qualified according to this ASTM specification. Each treatment was replicated four times for most tests, but two replicates were used for tests that extended beyond a 3-hour reaction time.
The ASTM D 2559 cyclic delamination test procedure is as follows (ASTM 1998):
* First cycle: step 1, vacuum soak in 180 to 21[degrees]C water at 84 kPa for 5 minutes; step 2, pressure soak in 18[degrees] to 21[degrees]C water at 517 kPa for 1 hour; step 3, repeat steps 1 and 2; step 4, dry at 66[degrees]C for 21 to 22 hours.
* Second cycle: step 1, steam at 100[degrees]C for 1 to 1.5 hours; step 2, add water at 180 to 21[degrees]C and pressure soak at 517 kPa for 40 minutes; step 3, repeat step 4 from the 1st cycle.
* Third cycle: repeat all steps from the first cycle.
Immediately after the final cycle, delamination was measured along all end-grain surfaces to the nearest 1.0 mm with a machinist's scale under a stereomicroscope. Delamination was measured from five bondlines on each end of the three cross sections from each assembly. Delamination was expressed as a percentage of total bondline length for each specimen and assembly. Statistical analysis was based on these delamination percentages.
Results and discussion
Reactivity and reaction products
Immediately after mixing, early reactions produced hydroxymethyl and hemiformal groups. Since part of those early reactions occurred before samples could be scanned, the exact heat of reaction could not be obtained by DSG. Also, measuring progressively smaller residual heats of reaction after 8 hours of reaction time became difficult, particularly as very dilute solutions were being analyzed.
The DSC measurements of heat of reaction as a function of time are shown in Figure 1 for three n-HMR coupling agents prepared from different novolak FIR molar ratios (0.31, 0.39, and 0.46). Nearly all of the potential reactivity of the novolak with formaldehyde was consumed within the first 24 hours.
When the heat of reaction data was transformed into natural logarithms and plotted against time, linear relations appeared to fit the data (Fig. 2), indicating an apparent first order reaction. Equations of the following form were fitted to the data:
LnQ = a - bt
where Q is the heat of reaction in J/g, and t is time in hours. The [r.sup.2] values for the goodness of fit ranged from 0.98 to 0.99, confirming the linearity of the logarithmic relation. Values from fitting the equations to the curves are given in Table 3. This information allows one to confidently extrapolate to a time at which the reactivity falls to inconsequential levels. For example, at about 24 hours, Ln(heat) --1.00, which means that the residual reaction heat value was 0.37 J/g, whereas values recorded after a few minutes of reaction were above 20 J/g.
The course of reaction of the 0.39 F/R novolak with additional formaldehyde is presented as integrated areas of the (13) C signals for the various formaldehyde-derived moieties (Table 4). Signals from structures were separated into three categories. As Figure 3 shows, formaldehyde was quickly consumed by the novolak to form hydroxymethyl groups. Within 23 minutes, about 20 percent of the total formaldehyde-derived material (27% of recently added formaldehyde) had formed hydroxymethyl groups. It took approximately 20 minutes to spread HMR onto laminates to be bonded in these tests. In comparison, for the original HMR, nearly 70 percent of the formaldehyde was converted to hydroxymethyl groups and only 2 percent to methylene bridges within the 23-minute period (Christiansen 2000). The n-HMR started with 23 percent of total formaldehyde as methylene bridges, whereas the original HMR required more than 4 hours to build methylene bridges to this level. The calculated average molecular weight for a resorcino 1-form aldehyde novolak of 0.4 F/R is 192, whereas resorcinol has a molecular weight of 110. This means that the average molecule in the solution contained only about two resorcinol units.
A calculation of the ratio of novolak F/R = 0.39 to total F/R = 1.54, at the same resorcinol level, shows that 25.3 percent of total formaldehyde was added to make the novolak. This initial formaldehyde should finally form methylene bridges, with only a small amount of formaldehyde forming side products from the Cannizzaro reaction (Walker 1975, p. 214). The NMR peak areas in Table 4 indicate that in the earliest reaction period, sampled 5 to 12 minutes after addition of the final formaldehyde, the number of carbons as methylene groups and as methanol constituted an average 24 percent of the peak area of species associated with formaldehyde additions. At this stage, most of these peak areas were contributed by the original novolak. Thus, there is good agreement between calculated and experimental values for the fate of formaldehyde in forming the novolak. The ratio of methylene bridges that linked only 4- (or 6-) positions on a ring to those that linked to either one or two 2-positions was 4.8 for the novolak . Thus, there was good opportunity for branching and crosslinking with the reactive 2- position in subsequent reactions.
As with the original HMR, the activated n-HMR, after 15 to 19 hours, still had substantial numbers of unreacted hydroxymethyl groups (16.3% of the added formaldehyde). Furthermore, 15.2 percent of the total formaldehyde in n-HMR was still unreacted formalin, whereas only 2 percent formalin remained unreacted in the original HMR during this period. This difference in residual formalin content indicates that less formaldehyde reacted with the resorcinolic structures in the novolak-based HMR than reacted in the original HMR during the same period. This suggests an opportunity to decrease the total formaldehyde that needs to be added. As the reaction of formaldehyde and novolak proceeded, the percentage of total methylene-bridge carbons that were linked to at least one resorcinol 2- position (Fig. 4) increased from 17.3 percent in the novolak to 37.8 percent in the activated n-HMR during the 15- to 19-hour period. Because the 2-position is less reactive than the 4- or 6- positions and thus will often react after both 4- and 6- positions have reacted, substantial branching and some crosslinking probably occurred in this period. The source of the signals at 38 to 49 ppm, however, is still unresolved. As in a previous study of HMR (Christiansen 2000), we could not conceive of any expected structure that would give such signals, although the chemical shift and time of appearance suggest that the structure may consist of a methylene carbon linked to at least one resorcinolic ring.
Resistance to delamination
Assemblies without n-HMR (controls) suffered severe delamination (Table 5). However, those assemblies bonded with n-HMR coupling agent reacted between 1 and 3 hours showed delamination well below the ASTM-specified 5 percent maximum. Furthermore, the n-HMR made from the novolak with F/R molar ratio of 0.39 passed the specification at 0 hour reaction time. This means that n-HMR could be applied to wood immediately after mixing the two components.
Statistical analysis of percentages of delamination indicated that the epoxy-bonded control assemblies (without coupling agent) were significantly different from any of those bonded with any of the coupling agents (Table 5). Delamination of the control assemblies was severe (49.5%), far surpassing the 5 percent maximum specified by the ASTM standard. A second statistically separate group contained all F/R molar ratios at 0-hour reaction time. A third significantly different group consisted of all the other molar ratios at 1-, 2-, and 3-hour reaction times, as well as those at the 0-hour reaction time with molar ratios of 0.31 and 0.39. This third group included all those assemblies that passed the specification plus one with average delamination of only 7.6 percent. Based on the results in Table 5, of the three F/R molar ratios (0.31, 0.39, and 0.46), the n-HMR coupling agent using F/R of 0.39 was found to be the most effective.
These encouraging results prompted an extension of the experiment with the 0.39 molar ratio novolak to see whether delamination would remain low at even longer reaction times. Consequently, epoxy-bonded assemblies were primed after 5- and 7-hour reaction times. Only two assemblies were laminated at each reaction time. For this small set, the averages of delamination were also below the 5 percent limits (Table 5). Thus, it appears that a novolak-based HMR coupling agent with a novolak F/R molar ratio of 0.39 could be applied immediately and also be expected to last a whole 8-hour shift.
An n-HMR coupling agent is as effective as the original one-step HMR in providing exterior durable bonds of epoxy adhesive to Douglas-fir wood laminates. When n-HMR with F/R of 0.39 is reactivated with the second formaldehyde component, it can be used immediately.
The F/R ratio will affect the resistance to delamination of the bonded assemblies when the coupling agent is applied at different times after mixing. HMRs based on novolaks with molar ratios from 0.31 to 0.46 and reacted for 1 to 3 hours before being applied to wood easily met the percentage delamination requirements of ASTM D 2559. The n-HMR coupling agent that delaminated the least had a novolak molar ratio of 0.39 and had acceptable delamination resistance after 0 to 7-hour reaction times.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
Table 1. - Ingredients of epoxy FPL 1A. Ingredients Parts by weight Diglycidyl ether of bisphenol A 100.0 (DGEBA) epoxy resin Benzyl alcohol 12.5 Hydrophobic fumed silica 2.5 Triethylenetetramine hardener 11.1 Table 2. - Ingredients of test novolaks and complete HMR coupling agents. Weight percentages for different novolak F/R molar ratios Ingredients 0.31 0.39 0.46 Resorcinol (solid) 3.34 3.34 3.34 Water (deionized) 90.43 90.43 90.43 Sodium hydroxide 2.44 2.44 2.44 (3 M solution) Formalin (37% solution) for novolak 0.76 0.95 1.14 for final activation 3.03 2.84 2.65 Table 3. - Parameters for statistically fitting straight lines to differential scanning calorimetry heat of reaction versus time for three n-HMR coupling agents based on novolaks with different F/R molar ratios. (a) Novolak F/R ratio a b [r.sup.2] 0.31 3.083 0.138 0.989 0.39 3.378 0.213 0.978 0.46 3.104 0.176 0.996 a Parameters for LnQ = a - bt, where Q = heat of reaction (J/g); t = reaction time (hr.); [r.sup.2] is the goodness-of-fit parameter. Table 4. - Integrated areas of NMR signals for various chemical shift regions (with attributed chemical species), as a function of reaction time of n-HMR based on novolak with 0.39 F/R. (a) Fractional areas of total formaldehyde- related NMR signals Methylene bridges Time Avg. 2,4-/2,2- 4,4-/4,6- Unassignrd period time 20 to 29 ppm 20 to 38 ppm 38 to 49 ppm (hr.) 0 min 0.00 0.17 0.81 0.000 5 to 12 min. 0.14 0.04 0.18 12 to 19 min. 0.26 0.05 0.17 19 to 26 min. 0.38 0.05 0.17 26 to 33 min. 0.49 0.04 0.17 33 to 40 min. 0.61 0.05 0.18 40 to 47 min. 0.73 0.05 0.18 47 to 54 min. 0.84 0.05 0.18 54 to 61 min. 0.96 0.05 0.17 61 to 75 min. 1.13 0.07 0.17 75 to 89 min. 1.37 0.07 0.18 89 to 103 min. 1.60 0.08 0.18 103 to 117 min. 1.83 0.08 0.18 117 to 131 min. 2.07 0.09 0.19 131 to 145 min. 2.30 0.09 0.20 145 to 159 min. 2.53 0.10 0.20 0.018 159 to 187 min. 2.88 0.11 0.20 0.045 187 to 215 min. 3.35 0.12 0.21 0.036 215 to 243 min. 3.82 0.12 0.22 0.043 243 to 271 min. 4.28 0.14 0.23 0.043 271 to 331 min. 5.02 0.15 0.24 0.047 331 to 391 min.. 6.02 0.17 0.25 0.050 7 to 9 hr. 8.00 0.19 0.27 0.057 9 to 11 hr. 10.00 0.21 0.30 0.064 11 to 13 hr. 12.00 0.22 0.31 0.069 13 to 15 hr. 14.00 0.23 0.31 0.075 15 to 19 hr. 17.00 0.24 0.31 0.083 Fractional areas of total formaldehyde- related NMR signals Hydroxymethyl groups Time Methanol 2- 4/6- period 49.5 to 53 ppm 54 to 59 ppm 60 to 65 ppm 0 min 0.03 0.000 0.000 5 to 12 min. 0.02 0.041 0.065 12 to 19 min. 0.02 0.051 0.105 19 to 26 min. 0.02 0.065 0.139 26 to 33 min. 0.02 0.076 0.168 33 to 40 min. 0.02 0.085 0.192 40 to 47 min. 0.02 0.091 0.210 47 to 54 min. 0.02 0.101 0.226 54 to 61 min. 0.02 0.110 0.242 61 to 75 min. 0.02 0.110 0.268 75 to 89 min. 0.02 0.119 0.276 89 to 103 min. 0.02 0.123 0.285 103 to 117 min. 0.02 0.125 0.296 117 to 131 min. 0.02 0.130 0.299 131 to 145 min. 0.02 0.136 0.302 145 to 159 min. 0.02 0.130 0.299 159 to 187 min. 0.02 0.126 0.281 187 to 215 min. 0.02 0.133 0.284 215 to 243 min. 0.02 0.131 0.276 243 to 271 min. 0.02 0.128 0.266 271 to 331 min. 0.02 0.125 0.250 331 to 391 min.. 0.02 0.125 0.228 7 to 9 hr. 0.02 0.112 0.196 9 to 11 hr. 0.02 0.103 0.155 11 to 13 hr. 0.02 0.098 0.126 13 to 15 hr. 0.03 0.091 0.106 15 to 19 hr. 0.03 0.086 0.088 Hemiformal, Formalin+ Time 4-alpha (b) beta-HF (c) period 66 to 70 ppm 81 to 96 ppm 0 min 0.000 0.000 5 to 12 min. 0.015 0.639 12 to 19 min. 0.016 0.600 19 to 26 min. 0.017 0.545 26 to 33 min. 0.020 0.499 33 to 40 min. 0.022 0.462 40 to 47 min. 0.021 0.427 47 to 54 min. 0.022 0.403 54 to 61 min. 0.020 0.380 61 to 75 min. 0.019 0.347 75 to 89 min. 0.016 0.323 89 to 103 min. 0.017 0.294 103 to 117 min. 0.018 0.285 117 to 131 min. 0.016 0.250 131 to 145 min. 0.014 0.234 145 to 159 min. 0.014 0.226 159 to 187 min. 0.013 0.202 187 to 215 min. 0.010 0.186 215 to 243 min. 0.009 0.175 243 to 271 min. 0.008 0.171 271 to 331 min. 0.010 0.157 331 to 391 min.. 0.007 0.149 7 to 9 hr. 0.009 0.142 9 to 11 hr. 0.010 0.142 11 to 13 hr. 0.010 0.144 13 to 15 hr. 0.009 0.149 15 to 19 hr. 0.011 0.152 (a)Entries in the first row are for novolak without additional formaldehyde; data after the first row are for novoak plus formaldehyde. (b)Hemiformal, 4-alpha, are signals assigned to hemiformal carbons attached to a 4- or 6- position on a resorcinolic ring. (c)Beta-HF are signals assigned to hemiformal carbons not attached directly to a resorcinolic ring. Table 5. - Aaveraged delamination and statistical comparison of epoxy-bonded laminates primed with n-HMR coupling agents by reaction time. (a) Delamination (c)% Novalak F/R molar ratio Reaction time (b) Control (d) 0.31 0.39 0.46 (hr) -- 49.5 A 0 7.6 BC 3.8 BC 11.8 B 1 1.5 C 0.8 C 1.1 C 2 2.7 C 0.9 C 2.1 C 3 2.5 C 1.1 C 0.6 C 5 3.0 (e) 7 2.6 (e) (a)Average values for sample group identified by the same capital letter are not significantly different at the 0.05 level of probability, whereas those with different capital letters are statistically different. Those with the BC designation statistically overlap both the B and C groups. (b)The time between when n-HMR was constituted and when it began to be applied to the wood. (c)The ASTM D 2559 specified maximum allowable delamination for softwoods is 5 percent. (d)Control specimens were not primed with n-HMR. (e)Values for 5 and 7 hours were generated from only two replicates, rather than the four replicates used for the other conditions.
American Institute of Timber Construction (AITC). 1992. American national standard for wood products: Structural glued laminated timber. ANSI/AITC A190.1-1992. AITC, Englewood, CO. 16 pp.
American Society for Testing and Materials (ASTM). 1998. Standard specification for adhesives for structural laminated wood products for use under exterior (wet-use) exposure conditions. ASTM D 2559-97a. In: Annual Book of ASTM Standards, Vol. 15.06 Adhesives. ASTM, West Conshohocken, PA. pp. 157-161.
Christiansen, A.W. 2000. Resorcinol-formaldehyde reactions in dilute solution observed by carbon-13 NMR spectroscopy. J. of Appl. Polymer Sci. 75:1760-1768.
_____, C.B. Vick, and E.A. Okkonen. 2001. A novolak-based hydroxymethylated resorcinol coupling agent for wood bonding.
In: Proc. of Wood Adhesion 2000. Forest Prod. Soc., Madison, WI.
Nearn, W.T. 1974. Application of the ultra-structure concept in industrial wood products research. Wood Sci. 6(3):285-293.
Vick, C.B. 1995. Coupling agent improves durability of PRF bonds to CCA-treated southern pine. Forest Prod. J. 45(3):78-84.
_____. 1996. Hydroxymethylated resorcinol coupling agent for enhanced adhesion of epoxy and other thermosetting adhesives to wood. In: Proc. of the Wood Adhesives 1995 Symp. Proc. 7296. Forest Prod. Soc., Madison, WI. pp. 47-55.
_____ and E.A. Okkonen. 1997. Structurally durable epoxy bonds to aircraft woods. Forest Prod. J. 47(3):71-77.
_____, A.W. Christiansen, and E.A. Okkonen. 1998. Reactivity of hydroxymethylated resorcinol coupling agent as it affects durability of epoxy bonds to Douglas-fir. Wood and Fiber Sci. 30(3):312-322.
_____, K. Richter, and B.H. River. 1996. Hydroxymethylated resorcinol coupling agent and method for bonding wood. U.S. patent 5,543,487. August 6. 17 pp.
_____, _____, _____, and A.R. Fried. 1995. Hydroxymethylated resorcinol coupling agent for enhanced durability of bisphenol-A epoxy bonds to Sitka spruce. Wood and Fiber Sci. 27(1):2-12.
Walker, J.F. 1975. Formaldehyde. 3rd ed. Robert E. Krieger Pub. Co., Huntington, New York.
Alfred W. Christiansen *
* Forest Products Society Member.
The authors are, respectively, Chemical Engineer, Research Scientist (retired), and Physical Science Technician (retired), USDA Forest Serv., Forest Prod. Lab., One Gifford Pinchot Dr., Madison, WI 53726-2398. The use of trade or firm names in this publication is for reader information and does not imply endorsement by the U.S. Dept. of Agriculture of any product or service. We thank Kolby C. Hirth for obtaining nuclear magnetic resonance spectra on samples and Cherilyn A. Hatfield for statistically analyzing data from the delamination tests. This paper was received for publication in September 2001. Article No. 9367.
|Printer friendly Cite/link Email Feedback|
|Author:||Christiansen, Alfred W.; Vick, Charles B.; Okkonen, E. Arnold|
|Publication:||Forest Products Journal|
|Date:||Feb 1, 2003|
|Previous Article:||Relationship between site index and several mechanical properties of Turkish calabrian pine (Pinus brutia Ten).|
|Next Article:||Several examinations on the static bending test methods of wood using todomatsu (Japanese fir).|