LABORATORY DECAY RESISTANCE OF WOODFIBER/THERMOPLASTIC COMPOSITES.

STEVEN VERHEY [*]

PETER LAKS [*]

ABSTRACT

There is limited information available on the susceptibility of woodfiber/thermoplastic composites to biodegradation. The objective of this paper is to report on the laboratory decay resistance of model composites. Polypropylene/pine

composites were made with 30, 40, 50, 60, or 70 percent wood contents and exposed to brown- or white-rot fungi. In addition, 50/50 woodfiber/thermoplastic composites were treated with zinc borate (ZB) preservative at loadings of 1,3, or 5 percent and evaluated against the brown-rot fungus. The effect of surface abrasion on decay was investigated by sanding the surface of some samples. Weight loss was proportional to wood content in the unpreserved composites. Substantial weight losses were observed for all unpreserved composites tested with the brown-rot fungus, while white-rot attack was only significant in the high wood content composites, particularly when the surface of the samples was sanded prior to exposure to the fungus. The incorporation of ZB into the composites provid ed protection from fungal attack at all loadings.

Woodfiber/thermoplastic composites are being marketed for exterior use as substitutes for traditional pressure- or dip-treated wood and treated wood composites. The hydrophobic nature of the plastic matrix and its inherent high biodegradation resistance are believed to provide sufficient resistance to fungal degradation without the use of preservative chemicals. If the plastic and wood are mixed in such a ratio that a continuous plastic phase exists in the composite, the wood particles should be encapsulated and protected from the effects of moisture and fungal attack. A general review of the literature concerning the degradation of woodfiber/thermoplastic composites reveals contradicting findings.

Morris and Cooper [8] have presented the clearest evidence of woodfiber/thermoplastic composites experiencing decay in field exposure. This is perhaps the most significant report since it concerns commercial products already in use. Fungal fruiting bodies were found growing on commercially manufactured composites used on a boardwalk in Florida. There have also been reports of studies that show some susceptibility to fungal attack [2,10].

Efforts to study the biodegradation of these composites in the laboratory have apparently met with limited success. Two studies where laboratory samples were subjected to fungal attack showed no real decay. Johnson et al. [3] reported low levels of brown-rot fungal degradation of composites made from polypropylene and straw filler. Khavkine et al. [4] found good fungal colonization on the surfaces of composite samples with high wood content, but no significant decay of woodfiber/thermoplastic composites containing wood contents as high as 70 percent by mass. However, Mankowski and Morrell [7] have actually observed the effects of decay on model composites through the use of scanning electron microscopy. No significant decay was observed in 50/50 wood-HDPE composites, but clear evidence of attack was observed in higher wood content composites.

The clear evidence supporting fungal decay susceptibility presented by Morris and Cooper [8] and Mankowski and Morrell [7], combined with the unsuccessful attempts to show significant decay in other laboratory experiments, left a question as to how susceptible woodfiber/thermoplastic composites are to fungal attack. It is likely that these materials need to be treated with a chemical preservative before being used in outdoor applications.

The dominant method of producing treated composites is to incorporate the preservative chemical into the product during the manufacturing process. A general review of wood composite preservation by Laks [5] found that the most common preservative chemical used for in-process treatment is zinc borate (ZB). Borates are effective against both insects and fungi. ZB is considerably less soluble in water than other borate salts such as disodium octaborate tetrahydrate (DOT). Its low water solubility makes ZB resistant to leaching [6]. Good biological efficacy, combined with low mammalian toxicity and leach resistance, makes ZB a good choice for incorporation into model woodfiber/thermoplastic composites.

Woodfiber/thermoplastic composites are typically produced by extrusion or compression molding. Compression molding may facilitate the formation of a thin surface layer of plastic on a final product. The removal of this surface layer by a finishing step in the manufacturing process or as an artifact of installation may increase the decay susceptibility of these materials when they are exposed to an outdoor environment. Extruded deck products are sometimes finished by machining a wood grain appearance into the surface. This surface abrasion, combined with end cuts and sanding during installation, may also result in a higher susceptibility to fungal attack.

The purpose of this work is to report our results on laboratory decay resistance evaluation of model woodfiber/thermoplastic composites made with varied wood contents. The effect of surface abrasion on decay susceptibility was investigated by comparing test results from sanded and unsanded samples. The effect of ZB as a preservative was also investigated. This paper is a preliminary report in a comprehensive study of the laboratory and field biological degradation resistance of these model composite materials.

EXPERIMENTAL

MANUFACTURE OF COMPOSITES

The composites used for this work were produced from 20 mesh ponderosa pine (Pinus ponderosa Dougl. ex Laws) furnish (American Wood Fibers, Schofield, Wisconsin) and polypropylene (Exxon PP3505G-E1). Five types of composites, consisting of 30, 40, 50, 60, and 70 percent wood content based on the total panel weight, were generated using a hot press. ZB-treated panels were produced at a fixed wood content of 50 percent. Appropriate amounts of powdered ZB (US Borax, Borogard ZB (R)) were blended with the wood and polymer to achieve borate concentrations of 1, 3, or 5 percent of the ovendry wood weight in the final panel.

Ovendry wood particles, polymer, and ZB (if appropriate) were blended for 5 minutes in a Papenmeier high-intensity mixer to produce a homogeneous blend. The compounded mixtures were transferred to a 230- by 400- by 6-mm form between two caul sheets for pressing. All of the blends were hot-pressed at 200 [degrees] C using a 450- by 450-mm laboratory scale press. The pressing pressure varied between 72.5 and 145 MPa depending on the wood content of the panels. Higher pressure was required to compress the high wood content panels to the desired 6-mm thickness.

The blend for each panel was partially melted prior to high-pressure pressing. The press was closed until the caul sheets came into contact with both heated platens. Settling and flow of the material occurred as the plastic melted. Closing the press just enough to maintain good contact between the platens and the caul sheets compensated for the settling. This heating/melting period spanned 5 minutes for each blend. Immediately following melting, the panels were pressed, under pressure, for 7 minutes more.

After hot-pressing, the form was transferred to a cold press and allowed to cool down to 50 [degrees] C under 72.5 MPa of pressure in a separate, unheated press. At this point, the plastic was solid. The final step of manufacturing was to transfer the panel to a cold plate and allow it to cool completely under pressure from a 1-inch thick steel plate. This final cooling step prevented any possible warping of the panel.

The resulting composite panels appeared to have a continuous plastic layer present at both surfaces where the hot platens contacted the caul sheets. In an effort to mimic the surfacing that would occur from machining commercial products, some samples were sanded lightly before fungal testing. These samples were hand-sanded by placing the sample face down on a sheet of 150 grit sandpaper and sanding in a polishing manner for 25 revolutions on each side. It was thought that the sanding would remove the continuous plastic layer and expose some of the surface-layer wood particles to fungal attack, thereby maximizing potential degradation.

FUNGAL EXPOSURE

Soil-block testing was carried out according to AWPA E10-91 (1) with the following modifications. Samples for soil-block analysis measured 20 mm wide by 32 mm long by 6 mm thick, rather than being the 20-mm cubes specified in the standard. In addition, the brown-rot inoculation was performed at the same time that the samples were introduced into the soil jars. According to the standard, the brown-rot feeder strips are supposed to be inoculated and the fungus grown for 3 weeks prior to sample introduction. We have found that the chances for contamination increase significantly upon re-opening the inoculated jars and that satisfactory decay results are achieved by the modified method. The white-rot test was unmodified. An outline of the procedure for preparation of the jars and fungal testing is presented in the following paragraphs.

Soil jar and sample preparation. -- Eight-ounce jars were filled with 36 [+ or -] 1 mL of distilled water and 100 [+ or -] 1 g of dried soil. A southern pine feeder strip (Pinus spp.) was placed onto the surface of the soil in each of the brown-rot jars. Paper birch (Betula papyrifera Marsh.) feeder strips for the white-rot fungal tests were autoclaved along with the samples. The jars were fitted loosely with plastic lids and then sterilized at 120 [degrees] C in an autoclave for 30 minutes. The jars were covered and allowed to cool undisturbed until sample introduction.

Samples for the white and brown-rot fungal testing were conditioned to constant weight in a 40 [degrees] C oven and weighed accurately. The samples were then placed into test subgroups in labeled, glass petri dishes and autoclaved for 30 minutes at 120[degrees]C. The samples were also allowed to cool undisturbed until introduction into the jars.

White-rot samples. -- Susceptibility to decay by the white-rot fungus Trametes versicolor (MAD 697) was determined according to the standard without modification. Samples were buried level with the surface of the soil in the jars and a strip of fungal inoculum (6 by 25 mm) was placed in contact with both the soil and the specimen. A sterile paper birch feeder strip was then placed in contact with the fungal inoculum and the soil to facilitate fungal growth. Lids were screwed lightly onto the jars, and the caps were wrapped with Parafilm [R] to help prevent mite infestation during incubation. Control blocks (20-mm cubes) of paper birch were tested along with the analytical samples to verify satisfactory decay activity of the particular fungus used for this test. The test jars were placed in a controlled temperature (27[degrees]C) and humidity (80% relative humidity) room for 12 weeks.

Brown-rot samples. -- Decay resistance to brown-rot fungi was evaluated by exposure to Gloeophyllum traebeum (ATCC 11539). Feeder strips of southern pine that lay on the soil surface were partially covered with a 6- by 25-mm strip of fungal inoculum. Care was taken to make sure that the inoculum contacted both the soil and feeder-strip surfaces. The test sample was placed on top of the fungus. Control samples (20-mm cubes of southern pine sapwood) were also evaluated to verify fungal decay activity. The jars were sealed as described previously. The incubation time of the brown-rot samples was also 12 weeks.

Leaching of preserved samples. -- Preservative-treated samples are subjected to a leaching process to help determine the permanency of the preservative system. Groups of five samples were weighted down with stainless steel weights and soaked in 1 L of distilled water for 14 days. The water was changed daily. The leached samples were conditioned to a constant weight at 40[degrees]C and sterilized in the autoclave prior to testing.

Evaluation of decay. -- At the end of the incubation period, the jars were removed from the incubation room and visually inspected for evidence of contamination. The samples were removed from the jars and loose soil and fungal mycelia were scraped from the surface. After conditioning to constant weight (1 week) in a 40[degrees]C oven, the samples were weighed and the mass loss was calculated. It was assumed that mass loss was due to fungal degradation of the wood only. The overall mass losses could then be converted to percent wood mass losses through consideration of the composite's original wood content.

Statistical analysis. -- Two types of analysis of variance (ANOVA) testing were used to help interpret the weight loss results and determine the effect of sanding on the test samples. One-way ANOVA analysis was performed using the Minitab [TM] software package (Release 13). Two-way ANOVA testing was performed using Design Expert [TM] Version 6.0.0. Differences were evaluated at the 95 percent confidence interval for both types of analyses.

RESULTS AND DISCUSSION

Decay results from the control samples and untreated composites are presented in Table 1. It is important to verify that each type of fungus used in the experiment is capable of producing the necessary enzymes to decay wood when evaluating the decay resistance of preserved samples. The southern pine blocks showed an average weight loss of 30.5 percent with a standard deviation among the five samples of 5.8 percent. Similarly, the birch blocks lost an average of 38.1 percent of their original mass with a standard deviation of 7.4 percent. These results demonstrate good fungal decay activity on non-durable wood species.

The white- and brown-rot decay results from the woodfiber/thermoplastic samples are presented in Figures 1 and 2. Results for sanded and unsanded samples are presented in each. In general, the white-rot fungus was less effective at decaying the pine furnish in the composites. This is commonly the case with white-rot attack of conifers [9].

While the white-rot fungus produced less decay in the composites, this test did show a significant effect due to surface abrasion. The unsanded composites with wood contents from 30 to 50 percent consistently showed an insignificant weight loss of 1 percent or less when exposed to the white-rot fungus. The unsanded 60 and 70 percent wood composites showed higher (although still relatively low) weight losses. When the test samples were sanded prior to exposure to the fungus, the weight loss in the 30 to 50 percent wood composites remained low (less than 1%), however, the weight loss from the sanded 60 and 70 percent wood composites increased dramatically. One-way ANOVA analysis showed the difference between the unsanded and sanded 60 percent results and the unsanded and sanded 70 percent results to be statistically significant.

Two-way ANOVA analysis of the white-rot results revealed two significant contributions to increased decay susceptibility: higher wood content and sanding (Table 2). The ANOVA analysis also revealed the presence of a statistically significant interaction between surface abrasion and increasing wood content that increased decay susceptibility.

When considering the test results from a physical standpoint, it appears that the high plastic content of the low wood content samples (30% to 50%) effectively encapsulated the wood particles in plastic, thereby preventing fungal attack. The fungal hyphae could have been physically blocked from contacting the wood particles, or it is possible that the encapsulated particles were prevented from attaining a moisture content that was high enough to support decay. Surface sanding might only have exposed a small portion of the wood particles. As a result, the mass lost to decay would not be significant when compared to the overall mass of the composite.

The brown-rot test shows decay of the composites at all wood loading levels. The 60 and 70 percent wood content blends show very high levels of weight loss to decay: approximately 40 and 54 percent, respectively. Unlike the high wood content white-rot samples, surface abrasion had no significant effect on the decay susceptibility of the composites according to the ANOVA testing. The results of two-way ANOVA analysis (Table 3) show that the only significant factor in increasing decay susceptibility was increasing the wood content. The lack of a statistically significant effect of surface abrasion, as was seen with the white-rot fungus, may be due to the combination of higher weight losses (compared to the white-rot test) and the size of the standard deviations of some of the results. In either case, the overall higher weight loss from these samples appears to overshadow the effects of surface abrasion.

Preservative-treated samples were evaluated with the brown-rot fungus only, because it produced the highest weight losses in the untreated composites. A wood content of 50 percent was chosen since it matched the wood loading in the composite that showed fungal attack in the Morris and Cooper [8] study. Figure 3 presents the test results for the ZB-treated samples. The exact weight loss numbers are presented in Table 4. It can be seen that incorporation of 1 percent ZB into the composite effectively prevents fungal attack on the wood component. The 3 and 5 percent loadings provide similar effects. Leaching of the preservative-treated samples did not produce a significant change when compared to the unleached blocks of the same composition. The small weight gain appearing in the ZB-treated blocks likely reflects weight gain due to residual water after conditioning to constant weight at 40[degrees]C.

The results of both the white- and brown-rot fungal tests clearly show that wood composites made from pine wood fibers and polypropylene are susceptible to fungal attack, especially at wood contents greater than 50 percent of the total dry panel weight. ZB has been shown to effectively protect the composites against fungal decay, in the laboratory environment, at loadings as low as 1 percent. Future work will investigate field exposure of these model composites and their possible strength loss due to fungal attack.

SUMMARY AND CONCLUSIONS

Due to the limited body of literature available on the decay susceptibility of woodfiber/thermoplastic composites, a simple soil-block experiment was designed to test the common assumption that particle encapsulation by plastic would provide a decay-resistant composite without the use of chemical preservatives. Some evidence in the literature does suggest that such materials are subject to fungal attack despite formulation with a relatively high plastic content. It may follow that a critical plastic loading would exist, below which significant decay of the wood component would be observed.

The results of both the brown- and white-rot decay tests showed that weight loss was directly proportional to wood content. In the case of the brown-rot fungus, decay was evident at all wood loadings. The white-rot samples did not show significant weight loss until a wood loading of 60 percent was incorporated into the composites.

The effects of surface abrasion differed with respect to the fungus used for the decay test. Sanding was found to have a significant effect on the weight loss that resulted from exposure to the white-rot fungus. Decay in the brown-rot test appeared to be unaffected by surface abrasion.

The incorporation of ZB into the composites provided suitable protection of the wood component of the composites. No significant weight loss was observed from samples treated with even the lowest loading. The ZB preservative has also been shown to provide good resistance to leaching from the composite. No significant weight loss was observed from the leached samples.

The authors are, respectively, Graduate Student, Professor, and Research Scientist II, School of Forestry and Wood Prod., Michigan Technological Univ., 1400 Townsend Dr., Houghton, MI 49931. The authors would like to thank Amy Stephens, Erik Keranen, Glenn Larkin, and Dr. Laurent Matuana for their help and expertise in manufacturing and testing the composites. The authors would also like to thank American Wood Fibers, Exxon Chemical Americas, and US Borax for the donation of materials. This paper was received for publication in August 2000. Reprint No. 9160.

(*.) Forest Products Society Member.

[c]Forest Products Society 2001.

Forest Prod. J. 51(9):44-49.

LITERATURE CITED

(1.) American Wood Preservers' Association. 1999. Standard method of testing wood preservatives by laboratory soil-block cultures. AWPA E10-91. AWPA, Granbury, TX.

(2.) Brooks, J.G. and C.L. Beatty. 1995. A review of research and field service data for woodfiber-polymer composites. In: Proc. Woodfiber-Plastic Composites. Forest Prod. Soc., Madison, WI. p. 95.

(3.) Johnson, D.A., J.L. Urich, R.M. Rowell, R. Jacobson, and D.F. Caufield. 1999. Weathering characteristics of fiber-polymer composites. in: Proc. Fifth Inter. Conf. On Woodfiber-Plastic Composites. Forest Prod. Soc., Madison, WI. pp. 203-209.

(4.) Khavkine, M., M. Kazayawoko, S. Law, and J.J. Balatinecz. 2000. Durability of wood flour-thermoplastic composites under extreme environmental conditions and fungal exposure. Inter. J. of Polymeric materials 46:255-269.

(5.) Laks, P.E. 1999. The past, present, and future of preservative-containing composites. In: Proc. 33rd Inter. Particleboard/Composite Materials Symp. Washington State Univ., Pullman, WA.

(6.) _____ and M.E. Manning. 1994. Inorganic borates as preservatives for wood composites. Second Pacific Rim Bio-Based Composites Symp. Faculty of Forestry, Univ. of British Columbia, Vancouver, BC, Canada.

(7.) Mankowski, M. and J.J. Morrell. 2000. Patterns of fungal attack in wood plastic composites following exposure in a soil block test. Wood and Fiber Sci. 32(3):340-345.

(8.) Morris, P.I. and P.A. Cooper. 1998. Recycled plastic/wood composite lumber attacked by fungi. Forest Prod. J. 48(1):86-88.

(9.) Rayner, A.D.M. and L. Boddy. 1988. Fungal Decomposition of Wood. John Wiley & Sons, Chichester, UK.

(10.) Schmidt, E.L. 1993. Decay testing and moisture changes for a plastic-wood composite. Abstracts of the Res. Symp. In: Proc. American Wood Preservers' Assoc. 89: (108-109.) AWPA, Granbury, TX.

[Graph omitted]

[Graph omitted]

[Graph omitted]

TABLE 1.
Summary of weight losses from untreated soil
blocks and solid wood control samples.
                                         Average wood mass loss
Test fungus         Wood content                 Sanded
                         (%)
White-rot                30                        1.1
                         40                        0.6
                         50                        1.0
                         60                        1.9
                         70                        3.0
                 Solid paper birch           Not applicable
Brown-rot                30                        8.0
                         40                       11.8
                         50                       14.4
                         60                       40.4
                         70                       54.2
             Solid southern yellow pine      Not applicable
Test fungus  Unsanded
White-rot       1.3
                0.7
                0.6
                5.2
               14.6
               38.1
Brown-rot       7.9
               11.4
               16.1
               33.3
               57.9
               30.5
TABLE 2.
Results of two-way ANOVA analysis (95% confidence level) of the
white-rot weight loss data. [a]
Source             Sum of squares  DF  Mean square  F-value
Model                  427.18       9     47.46     125.82
Surface treatment       49.30       1     49.30     130.69
Wood Content           254.15       4     63.54     168.43
Interaction            123.72       4     30.93      81.99
Pure error              15.09      40      0.38
Cor total              442.27      49
Source             Prob [greater than] F
Model                [less than].0001
Surface treatment    [less than].0001
Wood Content         [less than].0001
Interaction          [less than].0001
Pure error
Cor total
(a)DF = degrees of freedom; values of "Prob
[greater than] F" less than 0.05 indicate
model terms that are significant.
TABLE 3.
Results of two-way ANOVA analysis
(95% confidence level) of the brown-rot weight loss data. [3]
Source             Sum of squares  DF  Mean square  F-value
Model                  9754.57      9    1083.84     114.26
Surface treatment         0.04      1       0.04       4.13E-03
Wood content           9681.83      4    2420.46     255.16
Interaction              72.70      4      18.18       1.92
Pure error              379.45     40       9.49
Cor total             10134.02     49
Source             Prob [greater than] F
Model                [less than].0001
Surface treatment              0.9491
Wood content         [less than].0001
Interaction                    0.1265
Pure error
Cor total
(a)Values of "Prob [greater than] F" less than 0.05 indicate model
terms that are significant.
TABLE 4.
Summary of brown-rot weight losses from samples treated with ZB. All
samples contained 50 percent wood and 50 percent plastic by weight.
            Average wood mass loss
ZB content        Unleached         Leached
   (%)
    0                12.9            19.5
    1                +0.8            +0.4
    3                 0.1            +1.4
    5                 0.0            +1.6