Developing an industrial friendly process for hydroxymethyl resorcinol (HMR) priming of wood using a novolak-based HMR.
Creating exterior durable bonds to wood using epoxy adhesives or other nonaqueous adhesives often requires the application of a primer or coupling agent. For durable wood-to-wood and wood-to-fiber-reinforced-polymer (FRP) bonding, the use of a hydroxymethylated resorcinol (HMR) coupling agent is recommended as a primer. In this paper, an experiment aimed at reducing the processing time required for HMR priming using a novolak-based HMR primer and the application of infrared heat emitted from lamps is described. This experiment evaluated the sensitivity of the entire HMR-adhesive system by assessing important process variables including HMR drying time, spread rate, and solids content. The experiment utilized a quick, accurate, and widely used technique to evaluate both shear strength and percentage of wood failure: a compression shear block test. The best bonding performance was measured at a novolak-based HMR drying time between 15 and 20 minutes. Both test conditions showed a high shear strength combined with a high percentage of wood failure. Finally, this experiment contributed to the knowledge required for the development of an industrial friendly process using HMR as a wood primer.
Fiber-reinforced polymer (FRP) composites have offered new opportunities to repair or strengthen existing wood structures in buildings and bridges (Lopez-Anido et al. 2002). Many of the adhesives used for bonding FRPs to wood do not adhere properly to wood, whereas commonly used structural adhesives, such as melamine-formaldehyde, phenol-formaldehyde, and resorcinol-formaldehyde often cannot be used for bonding FRPs to wood. To improve the bonding strength of FRP to wood, a coupling agent is often used. For FRP-to-wood bonding, the most commonly used coupling agent is hydroxymethylated resorcinol (HMR). HMR improves the wood-to-wood bonding strength of epoxy resins (Vick et al. 1995, 1998), but also improves the bonding of vinyl ester resin (Lopez-Anido et al. 2000), and phenol-resorcinol-formaldehyde (PRF) resin on preservative-treated southern pine (Vick 1995). It is hypothesized that HMR acts as a link between the wood substrate and the resin matrix. The linkages are thought to range from covalent to ether and hydrogen bonding (Vick and Okkonen 1997) depending on the chemical functional groups comprising the involved reactants. Gardner et al. (2000) found that HMR-treated wood has an increased polar surface energy, which enhances the interaction between the HMR-treated wood surface and the adhesive. This may promote both strong secondary interactions and the possible formation of covalent bonds during adhesive curing.
The HMR coupling agent is applied to the wood surface as an aqueous solution with low solids content (5%). It contains four main components: resorcinol, formaldehyde, sodium hydroxide, and water (Table 1). At present, two types of HMR have been evaluated in laboratory studies. For wood-to-wood bonding using epoxy adhesives, the original version of HMR was used by Vick et al. (1995). This HMR solution was allowed to react for 4 hours at room temperature. After 4 hours, the solution was applied to a freshly planed wood surface. Because of the amount of water contained in the HMR solution (95%) the HMR-treated surface needs to dry out prior to adhesive application. During drying, water evaporates from the wood surface while the resorcinol and formaldehyde react. A drying time of 18 to 24 hours is used at ambient conditions to bond nonaqueous adhesives, such as epoxy resins, making HMR commercially cumbersome to use. Care of freshly HMR-treated wood is required because contamination with dust or other contaminants can decrease the effectiveness of the HMR.
To reduce the reaction time of the HMR solution, Christiansen et al. (2000) developed a novolak-based HMR (n-HMR) coupling agent that is able to react within 1 hour of mixture at ambient temperature. This development of a modified HMR makes the coupling agent more user-friendly. However, a 24-hour drying time is still required for the HMR-treated wood surface, and this step should be reduced to be more attractive to commercial manufacturers of FRP wood composites. A shorter drying time of a few minutes, possibly a few seconds, could improve the industrial use of HMR. Decreasing the drying time of HMR could make possible the commercially viable industrial production of reinforced structural parts.
The purpose of this experiment was to significantly reduce the processing time of an n-HMR coupling agent from 24 hours to less than 30 minutes by applying an infrared heating drying step. The results of this experiment were obtained on wood-to-wood adhesive-bonded samples using epoxy resin. Although only wood-to-wood bonding was studied, it is envisioned that the experimental results could be extended to FRP-to-wood bonding procedures without concerns.
Experimental materials and methods
n-HMR coupling agent
For this experiment, an n-HMR coupling agent was used. Table 1 shows the ingredients for the n-HMR solution at 5 and 10 percent solids content. The primer has two chemical states and the mixing procedure is divided into two steps. The first mixing step provided an n-HMR solution. In the second step, the coupling agent is modified from the novolak status to the activated status by the addition of formaldehyde (see Christiansen et al. 2000). The n-HMR from the first mixing step can be stored at ambient conditions for at least 3 days, but no longer than 6 days. Before the coupling agent was applied to the wood, the final amount of formaldehyde was added to the solution. At this stage, the formaldehyde/resorcinol ratio increases from 0.39 to 1.54 and the n-HMR solution becomes activated n-HMR. To ensure adequate catalysis of the HMR reaction, the pH was determined after the final addition of formaldehyde solution, and, if necessary, the pH was adjusted by adding 3-molar sodium hydroxide solution. The final pH was held in a range between 8.5 and 9.0. This solution was allowed to react for 1 hour. At the end of the reaction time, dodecyl sulfate sodium salt was added to the n-HMR solution to improve the wetting of the wood surface. Within the next 1 to 3 hours, the n-HMR solution was applied to wood.
The experiment was conducted using FPL-1 epoxy adhesive (Table 2). This adhesive is derived from the reaction of bisphenol-A with epichlorohydrin to form a diglycidether of bisphenol-A (DGEBA). The epoxy resin (D.E.R. 331) was obtained from Dow Chemical Company, Midland, Michigan. Benzyl alcohol (99%), triethylenetetramine (60%, technical grade), and hydrophobic fumed silica were obtained from Aldrich Chemical Company, Milwaukee, Wisconsin.
The experiment was conducted on flat-sawn hard maple (Acer saccharum). The lumber was "Select" grade. The dimensions of the lumber were 3000 mm by 125 mm by 19 mm (length by width by thickness). The boards were stored at standard conditions (24[degrees]C, 65% relative humidity) for at least 8 weeks and had an equilibrium moisture content between 10 and 12 percent at the end of conditioning.
n-HMR drying apparatus
The n-HMR-treated surfaces were dried in the apparatus depicted in Figure 1. The apparatus was constructed from plywood boards having a thickness of 12.5 mm. The dimensions of the drying apparatus were 650 mm by 650 mm by 250 mm (height by length by depth). The specimen (1) was placed under four 250-W infrared lamps IR40 (125-mm diameter) from Phillips Lighting Co., Somerset, New Jersey, at a vertical distance of 230 mm. The lamps (2) were mounted at a horizontal distance of 25 mm from one another. The total span covered by the lamps was 650 mm. The infrared light emitted from the lamps heated the surface of the specimen contributing both to evaporating the water in the n-HMR solution and also shortening the HMR-wood reaction time. A laser beam temperature measurement device Raynger[R] unit (3) measured the temperature of the n-HMR-treated surface. The distance from the measuring point at the sample surface to the Raynger[R] unit was about 400 mm. Both hot air and evaporated n-HMR solution escaped from the apparatus through vents on the top of the drying apparatus (4).
Maple wood-epoxy laminates were bonded in press clamps to produce compression shear block specimens (Fig. 2). Eight press clamps were used to produce 16 wood laminates per day. The press clamps were made from mild steel. The press plates had the dimensions of 300 mm by 180 mm by 12.5 mm (length by width by thickness). Four threaded rods were used as the clamping device. Each rod had a diameter of 12 mm. The rods were mounted in a longitudinal distance of 150 mm and in a cross-distance of 150 mm. For wood laminates having a clamping area of 250 by 115 mm, a clamping pressure of 350 kPa was applied.
[FIGURE 1 OMITTED]
[FIGURE 2 OMITTED]
The experimental was conducted as a completely random design in a factorial arrangement (Steel et al. 1997) and examined the following factors.
Factor A: n-HMR drying time
1 5 minutes equivalent to about 45[degrees]C
2 10 minutes equivalent to about 55[degrees]C
3 15 minutes equivalent to about 60[degrees]C
4 20 minutes equivalent to about 65[degrees]C
5 24 hours at standard conditions
Factor B: n-HMR spread rate
1 146 g/[m.sup.2]
2 220 g/[m.sup.2]
Factor C: n-HMR solids content
Each treatment combination was replicated 7 times, giving a total number of 140 laminates. A single laminate consisted of two layers of hard maple boards, each having the dimensions of 250 mm by 115 mm by 15 mm (length by width by thickness). The following linear additive model was used:
[Y.sub.ijkl] = [mu] + [[alpha].sub.i] + [[beta].sub.j] + [[gamma].sub.k] + ([alpha][beta])[.sub.ij] + ([alpha][gamma])[.sub.ik] + ([beta][gamma])[.sub.jk] + ([alpha][beta][gamma])[.sub.ijk] + [[epsilon].sub.ijkl]
where [Y.sub.ijkl] = observation (response variable); [mu] = mean of the sample population; [[alpha].sub.i] = average effect of the n-HMR drying time; [[beta].sub.j] = average effect of the n-HMR spread rate; [[gamma].sub.k] = average effect of the n-HMR solids content; ([alpha][beta])[.sub.ij] = interaction between n-HMR drying time and n-HMR spread rate; ([alpha][gamma])[.sub.ik] = interaction between n-HMR drying time and n-HMR solids content; ([beta][gamma])[.sub.jk] = interaction between n-HMR spread rate and n-HMR solids content; ([alpha][beta][gamma])[.sub.ijk] = interaction between HMR drying time and n-HMR spread rate and n-HMR solids content; [[epsilon].sub.ijkl] = experimental error, which is not explained by the model.
A single treatment combination provided four response variables described as follows.
** Bondline shear strength at standard conditions
** Bondline shear strength for the water-soaked condition
** Bondline wood failure percentage at standard conditions
** Bondline wood failure percentage for the water-soaked condition
The hard maple boards were planed with a feeding speed of 10 m/min. After being planed, the boards were stored at room conditions for no longer than 4 hours. All possible treatment combinations were assigned with random numbers provided by SAS[R] (SAS Institute Inc. 1985). Immediately after n-HMR treatment, the boards were placed beneath the infrared heat lamps. When the first laminate was almost finished drying, the second laminate was treated with n-HMR. During the drying of the second laminate, the epoxy resin was prepared. Immediately after drying, the adhesive was applied and the two laminates were bonded in the press-clamp. The spread rate of epoxy resin was 500 g/[m.sup.2] per single bondline. The clamping pressure was approximately 350 kPa. The laminates were allowed to cure overnight. The bonded laminates were stored under standard conditions (23 [+ or -] 2[degrees]C and 65% relative humidity) for at least 3 days. From each laminate, six compression shear blocks were prepared. The size of the compression shear block specimens complied with specifications in ASTM D 905 (1994). The compression shear blocks were stored 1 day at standard conditions. The shear strength and percentage of wood failure were determined for the two conditions. Three compression shear blocks were tested at standard conditions, and three compression shear blocks were tested under the water-soaked condition. The total number of compression shear blocks tested in the experimental matrix was 840, where 420 were tested at standard conditions and 420 were tested for the water-soaked condition.
The shear strength at standard conditions was determined at 23 [+ or -] 2[degrees]C and 65 percent relative humidity. The shear strength for the water-soaked conditions was determined after soaking in water under vacuum (635 mm mercury) for 20 minutes followed by soaking in water under 520 kPa pressure for another 20 minutes. The shear strength was determined using an INSTRON 8801 testing frame. The loading was in shear by compression with a loading speed of 1.27 mm/min. After determining the shear strength, the percentage of wood failure of the bondline area was estimated to the nearest 5 percent. The three shear block strength values and three wood failure values of each condition were averaged. In the analysis of variance (ANOVA), this mean was used to determine the treatment effect of the particular treatment combination. Furthermore, by treatment separation, the 4 degrees of freedom of factor A (HMR drying time) were separated to four single degrees of freedom F-tests. Following treatment effects, mean separation was analyzed using contrast coefficients.
[FIGURE 3 OMITTED]
** standard n-HMR drying time (24 hours) vs. infrared drying times (5 to 20 minutes)
** 5-minute n-HMR drying time vs. 10-to-20-minute n-HMR drying times
** 10-minute n-HMR drying time vs. 15-to-20-minute n-HMR drying times
** 15-minute n-HMR drying time vs. 20-minute n-HMR drying time
In addition to the treatment separation, the treatment means were analyzed using Duncan's Multiple Range Test (DMRT). An alpha level of 0.05 was used. SAS[R] software (SAS Institute Inc. 1985) was used for separately analyzing each of the four response variables.
Results and discussion
Impact of drying time on board temperature
For all response variables, the n-HMR drying temperature (Factor A) increased with increased n-HMR drying time. Figure 3 shows a typical temperature curve of an n-HMR-treated surface being processed in the drying apparatus. Here, the laminate had an assigned n-HMR treatment combination of 220 g/[m.sup.2] spread rate, solids content of 5 percent, and a drying time of 20 minutes. Before drying, the surface temperature of the board was 25[degrees]C. At the end of the 20-minute process time, the temperature reached 80[degrees]C. This temperature change contributed to increased reactivity of the HMR as well as to evaporation of water.
General description of the collected experimental data
Because this experiment was aimed at reducing HMR processing time, it pushed the limits of the epoxy bonding of wood using HMR treatment. The data set response variables were impacted by both nonnormality and inequality of variances. The two response variables (shear strength dry and wood failure dry) showed nonnormality of the residuals in the Shapiro-Wilk Test. However, the variables shear strength wet and wood failure wet were normally distributed. Box plots for all response variables showed outliers within a range of 1.5 to 3.0 times of the inner-quartile. The Levine's Test showed inequality of variances for all four response variables. The n-HMR drying time had the strongest impact on variance inequality.
To eliminate nonnormality and inequality, the datasets were transformed. The response variables shear strength dry and wet were transformed using a logarithm of the base 10. The response variables wood failure dry and wood failure wet were transformed using an arcsine transformation. Additionally, a few observations were eliminated from three out of the four datasets. The elimination of observations was based on the residuals and their distance from the center. The observations having the largest deviation to both sides of the center of the box plot were eliminated. Only two observations were dropped per dataset. No elimination of data was necessary on the variable percentage of wood failure under dry test conditions.
The ANOVA on the variable shear strength dry showed the following results. At [alpha] = 0.05, the ANOVA indicated no significant differences among the main factors of the factorial. However, the treatment separation using contrast coefficients showed a significant difference between the shortest n-HMR drying time of 5 minutes compared to all other drying times (10, 15, and 20 min). This was confirmed by DMRT. No significant differences were found in the other interactions. The DMRT showed no significant differences between n-HMR spread rates and between the n-HMR solids content.
The ANOVA on the variable wood failure dry provided the following results. At [alpha] = 0.05, the ANOVA indicated a significant difference between the levels of the HMR drying time. The treatment separation using contrast coefficients showed a significant difference between the shortest n-HMR drying time of 5 minutes compared to all other drying times (10, 15, and 20 min). This was confirmed by the DMRT. No significant differences were found in all other interactions. The DMRT showed no significant differences between n-HMR spread rates and between n-HMR solids contents.
The results of not meeting the assumptions of homogeneity and equality of variances did not allow analysis of the response variables shear strength and percentage of wood failure for the water-soaked conditions using the ANOVA test or other parametric statistics. For both variables, the Shapiro-Wilk Test showed nonnormality of the residuals and the Levine's test showed inequality of variances. Dropping of more than two observations per dataset would lead to a mis-interpretation of the entire experiment. Therefore, interaction charts were used to interpret the ANOVA results more precisely. All interaction charts were based on the nontransformed and complete datasets.
Impact of process variables on adhesive bonding properties
For the variable shear strength dry, the lowest shear strength was achieved at the shortest n-HMR drying time (5 min). A shear strength of about 18 MPa was significantly lower than all other levels of n-HMR drying time (Fig. 4).
A similar situation was observed for the variable wood failure dry (Fig. 5). The lowest wood failure was achieved at the shortest n-HMR drying time (5 min). Also, the values of wood failure do not follow a linear trend, but a general trend can be seen. By increasing the drying time, the wood failure increases in the same manner as the shear strength, as discussed previously.
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
For the response variable shear strength wet in Figure 6, it is obvious that there is a significant treatment effect. Under the wet test conditions, the shear strength was affected to a greater extent by the n-HMR drying time than under the dry test conditions. At the lowest level of n-HMR drying time (5 min), the shear strength was the lowest. A shear strength of 4 MPa is below the 6-MPa minimum level for structural adhesives. By increasing the n-HMR drying time to 15 minutes, the shear strength attained a maximum value of 9 MPa. Furthermore, laminates dried at standard conditions did not meet the minimum standard for structural adhesives. Here, the shear strength varied from 4.5 to 7 MPa.
[FIGURE 6 OMITTED]
The same performance pattern occurred for the response variable percentage of wood failure tested under wet conditions. Here, the wood failure also increased with increasing n-HMR drying time, followed by a decreased wood failure on laminates cured at ambient conditions (Fig. 7).
Practical considerations from the study
The n-HMR drying experiment showed that the entire n-HMR-drying process was only sensitive to drying time. It was not sensitive to minor changes in spread rate and solids content. This means minor changes do not impact the bond performance. This is very convenient for the industrial application of n-HMR. The set-up of a production line has a certain degree of freedom. This freedom allows small deviations in spread rate and solids content without causing a loss in the bonding quality. The process of bonding the laminates in a heated environment improved the bonding performance. The use of a heated treatment process for processing the HMR coupling agent may also reduce the need for the postcuring of the epoxy adhesive in the production process, as is often recommended to achieve structural, water proof bonds. The use of a similar industrial drying station could replace the expense of postcuring epoxy-bonded wood composites in autoclaves. In future work aimed at optimizing the bonding process, it might be possible to shorten the n-HMR drying time using either stronger infrared heat lamps or by placing the wood laminates at a closer distance to the heat source.
The entire process, from planing of the surface of the laminates to bonding in the press clamps, was carried out as an industrial process. The experiment provided successful results and simulated a proposed industrial process.
[FIGURE 7 OMITTED]
The n-HMR drying time affected the response variables shear strength and percentage of wood failure to the greatest extent. The drying apparatus showed the lowest bond performance at an n-HMR drying time of 5 minutes. Under the conditions evaluated in this experiment, the 5-minute drying time was too short and resulted in unacceptable adhesive bond strength. At an n-HMR drying time of 10 minutes, the bonding performance attained the same level as for laminates bonded under standard conditions. The best bonding performance was measured at an n-HMR drying time between 15 and 20 minutes. Both test conditions showed a high shear strength combined with a high percentage of wood failure. The n-HMR spread rate did not significantly impact the bond properties. Because of the higher amount of water within the n-HMR solution, we recommend using the lower spread rate of 146 g/[m.sup.2]. Less water applied to the wood surface requires less heat to dry the n-HMR solution. The n-HMR solids content also did not impact the bond performance. There was no significant interaction measured, and it is recommended to use an n-HMR solids content of 5 percent.
American Society for Testing and Materials (ASTM). 1994. Standard test method for strength properties of adhesive bonds in shear by compression loading. ASTM D 905-94. In: Annual Book of ASTM Standards. ASTM, West Conshohocken, PA.
Christiansen, A.W., C.B. Vick, and E.A. Okkonen, 2000. A novolak-based hydroxymethylated resorcinol coupling agent for wood bonding. Wood Adhesives 2000, abstract from an international symposium sponsored by the USDA Forest Serv., Forest Products Lab., Madison, WI. pp. 245-250.
Gardner, D.J., W.T. Tze, and S.Q. Shi. 2000. Adhesive wettability of hydroxymethyl resorcinol (HMR) treated wood. Wood Adhesives 2000, abstract from an international symposium sponsored by the USDA Forest Serv., Forest Products Lab., Madison, WI. pp. 321-327.
Lopez-Anido, R., D.J. Gardner, and J.L. Hensley. 2000. Adhesive bonding of eastern hemlock glulam panels with E-glass/vinyl ester reinforcement. Forest Prod. J. 50(1112):43-47.
________, L. Muszynski, D.J. Gardner, B. Goodell, Y. Hong, L. Eisenheld, and B. Herzog. 2002. Performance-based material evaluation of FRP composite reinforcement bonded to glulam members. In: Proc. of the 10th U.S.-Japan Conf. on Composite Materials. F.K. Chang, ed. DEStech Publications, Inc., Lancaster, PA. pp. 462-471.
SAS Institute Inc. 1985, SAS[R] user's guide: Statistics, 5th ed. SAS Inst. Inc., Cary, NC.
Steel, R.G.D., J.H. Torrie, and D.A. Dickey, 1997. Principles and Procedures of Statistics, 3rd ed. McGraw-Hill, New York.
Vick, C.B. 1995. Coupling agent improves durability of PRF bonds to CCA-treated southern pine. Forest Prod. J. 45(3):78-84.
________ 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, B.H. River, and A.R. Fried, Jr. 1995. Hydroxymethylated resorcinol coupling agent for enhanced durability of bisphenol-A epoxy bonds to Sitka spruce. Wood and Fiber Sci. 27(1):2-12.
Douglas J. Gardner*
The authors are, respectively, Former Graduate Research Assistant and Professor of Wood Sci. and Technology, Advanced Engineered Wood Composite Center, Univ. of Maine, Orono, ME; (email@example.com). The authors gratefully acknowledge the grant funding USDA National Research Initiative Competitive Grants Program Grant # 2001-35103-11191, and McIntire Stennis grant ME 09615. This is paper #2688 of the Maine Agricultural and Forestry Experiment Station. This paper was received for publication in March 2004. Article No. 9860.
*Forest Products Society Member.
Table 1. -- Composition and solids content of n-HMR. n-HMR solids content HMR ingredients 5% 10% (g) Crystalline resorcinol 3.34 6.68 Deionized water 90.43 80.86 3-molar sodium hydroxide solution 2.44 4.88 Stage 1: Formaldehyde solution (37% formalin) 0.95 1.90 Stage 2: Formaldehyde solution (37% formalin) 2.84 5.68 Dodecyl sodium sulfate salt 0.50 0.50 Total weight 100.50 100.50 Table 2. -- Ingredients of FPL-1 epoxy resin. Ingredients Amount of chemical (g) Epoxy resin (D.E.R. 331) 79.3 Benzyl alcohol 9.9 Hydrophobic fumed silica 4.0 Triethylenetetramine 8.8
|Printer friendly Cite/link Email Feedback|
|Author:||Eisenheld, Leopold; Gardner, Douglas J.|
|Publication:||Forest Products Journal|
|Date:||Sep 1, 2005|
|Previous Article:||Predicting the strength of sawn wood products: a comparison between x-ray scanning of logs and machine strength grading of lumber.|
|Next Article:||Use and cost of Best Management Practices (BMPs) and BMP-related Sustainable Forestry Initiative guidelines to Arkansas timber producers.|