In this study, 18 mm thick MDF and PB panels with uncoated surface manufactured by Starwood, Bursa, Turkey were used. Two different adhesives of polyurethane (PU) obtained from Soudal, Belgium and polyvinyl acetate (PVAc) obtained from Filli Boya, Istanbul, Turkey were selected to be applied into the pilot holes. All screws were Philips flathead sheet metal screws made from stainless steel and plated by zinc. The screw major diameters were 3,5 and 4,0 mm with their lengths of 45 mm.

A complete five-factor factorial experiment with 7 replications per combination was conducted to evaluate factors on direct withdrawal loads of screw driven into MDF and PB. The five-factors were material (MDF and PB), pilot-hole type (no pilot-hole and pilot-hole drilled), adhesives (PU and PVAc), screw orientation (face and side) and screw major diameter (3,5 mm and 4,0 mm), soaking type (non-water and water soaked).

Therefore, a total of 448 DSW tests were performed on 224 test blocks. As shown in Figure 1, each test block had nominal dimensions of 50 mm × 50 mm × 18 mm (length x width x thickness) (TS EN 13446 2005).

Figure 1:
(a) Configuration of face and (b) side test blocks for evaluating DSW tests.

All test blocks were cut along the length direction of full-sized MDF and PB panels and were controlled at 20 °C ± 2 °C and 65 % relative humidity for two weeks in accordance with TS EN 320 (2011). The test blocks were divided into 4 groups based on the pilot-hole types. One of these groups did not have any pilot hole drilled into the test blocks. The test blocks in the other three groups had pilot-holes drilled in 80 % of screw major diameter. The pilot-hole diameters were 2,8 mm and 3,2 mm for the screw major diameter of 3,5 mm and 4,0 mm, respectively and drilled into the center of the side and face of MDF and PB test blocks. In two of these three groups, the pilot-holes of the test blocks were applied by two different adhesives which were PVAc and PU with the amount of 1 drop by a 5 ml injector. The screws were driven into all test blocks after the test blocks had been applied by adhesives. Half of the test blocks in all groups were tested right away for DSW and then the other half were immersed in pure water and kept in it for two hours before testing. The DSW tests were carried out using a Shimadzu AGIC/20/50KN test machine according to TS EN 320 (2011) and TS EN 13446 (2005) standards. The determination of density profile of MDF and PB was performed on IMAL DPX200 test machine (Imal Pal Group, Italy) using 10 different test blocks.

Load-time curve of DSW test for MDF and PB samples has shown in Figure 3. The curves illustrate a linear region that gradually becomes non-linear as it approaches the maximum load. After the maximum load was reached, the applied load decreased steadily until the test was concluded when screw was withdrawn from the face of MDF and PB. In the case of side orientation, the curves showed a linear relationship between load and time until a load drop, after which the load reached a plateau for both materials.

Figure 3:
Load-time curves of DSW test for (a) MDF face, (b) MDF side, (c) PB face and (d) PB side.

Table 2 summarizes mean DSW values of MDF and PB materials. In general, the mean DSW values ranged from 1048 N to 2076 N for face orientation of non-water soaked MDF whereas the values ranged from 695 N to 1018 N for water soaked MDF. For the side orientation, non-water soaked MDF had the values ranged from 335 N to 1634 N while water soaked MDF had the values ranged from 79 N to 767 N. In the case of face orientation of PB material, non-water-soaked ones ranged from 948 N to 1646 N whereas the values ranged from 474 N to 1053 N for water-soaked ones. The values for side orientation of non-water soaked PB ranged from 476 N to 1313 N while the values ranged from 190 N to 704 N for the water soaked PB material.

Table 3 summarizes ANOVA results obtained from the GLM procedure performed for data set. The five-factor interaction was significant which suggested that the significant interaction for the data set should be analyzed further. In general, four main effects of the data set were all significant with their p values less than 0,0001. Comparing the F values of the main effects, soaking type had a much greater F value of 1768,63 than the orientation with and F value of 1116,73, pilot-hole diameter with an F value of 311,88, screw major diameter with an F value of 75,59 and material with an F value of 57,17. Consequently, it was shown that the soaking type, which has the highest F value was the main factor affecting DSW when all parameters were compared (Freund et al. 2010, Kuang et al. 2017).

Effects of other four factors on DSW values were analyzed by considering their significant five-factor interactions. A one-way classification of 64 treatment combinations was created for DSW data set to evaluate mean differences among those combinations using the protected Least Significant Difference (LSD) multiple comparison procedure. Table 2, Table 4 and Table 5 summarize mean comparisons of DSW values for material, screw major diameter, pilot-hole, screw orientation, and soaking type, respectively, using the single LSD value of 135,03 N.

Table 2:
Mean comparisons of DSW (N) for pilot-hole type within each combination of screw orientation, soaking type, screw major diameter and material.

Table 3:
Summary of ANOVA results on five-factors of DSW data set.

Table 2 indicated that in general, the pilot-hole diameter with PU had the highest mean DSW load than the other types of pilot-holes followed in all combinations. There were only two cases in which no statistical difference was found in mean DSW between the pilot-holes with PVAc and PU. These cases were in the face of water soaked PB test blocks and the side of water soaked MDF test blocks driven by screw with 3,5 mm major diameter. The mean lowest DSW values were found in all non-drilled MDF test blocks. The type of the adhesive have significant effects on DSW resistance of both MDF and PB (Ors et al. 1998, Conrad et al. 2004, Sackey et al. 2008).

Table 4 indicated that the general trend was that the mean DSW was higher in MDF than PB in most cases. In a study by McNatt (1986), the MDF had higher DSW load than PB since MDF has a more uniform vertical density profile than PB. This is thought to be one of the reasons why the DSW resistance in MDF is higher than the corresponding ones in PB (Wang et al. 2007). The internal bond strength and density profile of the boards which directly affect DSW resistance depends on parameters such as fiber/chip properties and adhesive ratio (Mcnatt 1986).

Table 5 indicated that mostly the mean DSW was higher when the material was driven by the screw with 4,0 mm major diameter than the one with 3,5 mm. The screw major diameter statistically affected the mean DSW when the PU applied in pilot-hole for both materials. The screw major diameter of 4,0 mm had statistically higher mean DSW than the corresponding one with 3,5 mm in all combinations except one case. There was no significant difference between the screw major diameters in the PB side test blocks soaked in water. There was a clear trend that no significant difference was found in mean DSW among the screw major diameters when no adhesive applied in the pilot-holes of PB test blocks. The same trend was followed when no pilot-holes were drilled in MDF test blocks in all combinations.

Table 4:
Mean comparisons of DSW (N) for MDF and PB within each combination of screw orientation, screw major diameter, material and soaking type.

Table 5:
Mean comparisons of DSW (N) for screw major diameter within each combination of pilot-hole diameter, screw orientation, material and soaking type.

The mean DSW was statistically higher when the material was not soaked in water than the ones soaked in water because of swelling of the particles in the materials in all combinations (Figure 4). This can be explained by the gaps between the particles because of water absorption in which reduces the mechanical properties of particles as shown in Figure 4 MDF-b, MDF-d, PB-b and PB-d. The ratios of DSW in the material not soaked in water to the one in water soaked were 1,5 for face orientation and 3 for side orientation when using both screws in MDF. In the case of PB, the ratios were 2 and 2,5 for face and side orientations using both screws, respectively. This indicates that the PB which has larger particles than MDF have been affected more in the case of water soaking.

Figure 4:
(a) MDF test blocks with non-water soaked before and (c) after testing, (b) water soaked before and (d) after testing; (a) PB test blocks with non-water soaked before and (c) after testing, (b) water soaked before and (d) after testing.

The mean DSW was statistically higher when the screws driven into the face of both materials than the corresponding ones driven into the side orientation in all combinations since the screw was penetrated into three layers of the materials. This situation is related to the overall density of the materials in where the surface density of the panels was higher than the core density (Mcnatt 1986). In the case of the DSW resistance in side orientation in MDF and PB depended only on the core density of the panels where the particles were larger and wider than the ones in surfaces. Hung and Wu (2010) found a correlation between the DSW resistance and core density and reported that the reason of it was the interfacial adhesion between binding agent and particles of bamboo plastic composites. Rajak and Eckelman (1993) also reported that one of the parameters affecting the DSW was the core density of the material when driving screws in the side of the wood-based materials. A proper pilot-hole size needs to be drilled into the side of the material in order to prevent the splitting in the sides of material.

The DSW ratios from face to side orientation in MDF were 2 and 4 for the test blocks non-soaked in water whereas the ones soaked in water driven by both screw major diameters, respectively whereas the corresponding ratios were 1,5 and 2 in PB. The reason of the high ratio of face to side orientation in MDF and PB materials soaked in water could be the fractural particle deformation around the screw driven in the core of the materials during screw driving process. In addition, MDF has twice higher ratio than PB with the reason of having more fractural deformation in MDF which has higher density in the core.

^♠Corresponding author: ebirinci@kastamonu.edu.tr