Abstract

Testing materials

Black spruce wood was selected for this study. We kiln dried 270 flat sawn boards (860-mm length [L]) and stored them in a conditioning room at 20°C and 40 percent relative humidity (RH) until they reached 10 percent equilibrium moisture content (EMC). After conditioning, all sections were machined at 52-mm (tangential [T]) width and 22-mm (radial [R]) thickness. A 25-mm-long section was crosscut from each specimen to measure the average (455 kg/m³) and standard deviation (35 kg/m³) of basic density (ovendry mass over green volume). Specimens were then divided into nine groups having an average density of 455 kg/m³ each. Subsequently, each group underwent a surfacing treatment. After surfacing, samples were resectioned to prepare specimens for roughness (50 mm [L]), microscopy (25.4 mm [L]), wettability (160 mm [L]), and coating application (620 mm [L]). After coating application, two matched samples were sawn. One sample underwent an accelerated aging treatment before being submitted to the adhesion test, and the other remained untreated for comparison purposes.

Machining treatments

Peripheral straight-edge knife planing was used to surface the specimens. Previous cuts were carried out to level samples and run-in knives prior to each surfacing treatment. Planing was performed with a Weining Powermat 1,000 moulder equipped with a cutterhead having 53 mm of cutting radius. The cutterhead rotation speed was 10,000 rpm with three different feed speeds, 8, 10, and 13 m/min. This produced 0.8-, 1.0-, and 1.3-mm wavelengths, respectively. These wavelengths are small and usually recommended for finishing purposes. Hence, planing should result in good surface quality. Cutting depth was 1 mm, and three rake angles were studied (10°, 15°, and 20°). The clearance angle was kept constant at 20° for all three rake angles.

Microscopic evaluation

Ten-millimeter cubes were cut to observe tangential and end-grain surfaces. Tangential surfaces were used to evaluate the fibrillation level and open lumens. End-grain surfaces were used to analyze cell damage and coating penetration. One end-grain surface was carefully cut with a razor blade mounted onto a microtome. All cubes were then desiccated with phosphorous pentoxide (P₂O₅) for a week and mounted on standard aluminum stubs with silver paint. Environmental scanning electron microscopy micrographs were taken for two representative machined samples for each machining treatment.

Surface topography measurements

Roughness measurements were carried out on defect-free zones with a Micromeasure confocal microscope. A surface of 12.5 mm (L) by 12.5 mm (T) was analyzed for each sample. Data were collected with Surface Map 2.4.13 software using an acquisition frequency of 300 Hz and a scanning speed of 12.5 mm/s. Three-dimensional roughness parameters were determined using the Mountain software. A cutoff length of 2.5 mm combined with a Robust Gaussian filter (ISO 16610-31; International Organization for Standardization 2002) were used for calculations. Arithmetical mean deviation of the profile (S_A), maximum profile peak height (S_P), and maximum profile valley depth (S_V) were calculated according to ISO 4287 (ISO 1997). Core roughness depth (S_K), reduced peak height (S_PK), and reduced valley depth (S_VK) were calculated from the Abbot curve according to ISO 13565-2 (ISO 1996).

Wettability measurements

Wetting properties were evaluated within 8 hours following the machining treatments. Wetting analyses were performed with a FTÅ D200 imaging goniometer at 20°C. One drop (∼6 μL) of pure water was added to wood surfaces with an injection microsyringe. A frame grabber recorded changes in droplet profile during the first 120 seconds of wetting. The measurements were carried out fiberwise because wetting is more important and constant in this direction (Gardner et al. 1991, de Meijer et al. 2000). Contact angle was calculated as a mean of both sides of the drop in order to compensate for any horizontal variations. Wetting rate was calculated as Δθ/Δt to assess spreading and penetration of pure water over the first 30 seconds of wetting. For surface energy measurements, initial contact angle of formamide was measured. Nonpolar and polar components of the surface energy were calculated using the harmonic mean method according to Wu (1971). According to Piao et al. (2010), no procedure has been developed for determining absolute values of surface energy. However, several authors have used this method to describe physicochemical changes of wood surfaces produced by different treatments (Wang et al. 2007, Wolkenhauer et al. 2007).

Coating application procedure

Machined surfaces were coated within 8 hours after the machining treatments. During this period, samples were placed face against face in order to keep contamination at minimum levels. Two coats of an acrylic water-based coating Laurentide–Innocryl PF were air sprayed (150-μm wet thickness) at room temperature according to manufacturer's specifications. Coating was then cured in an infrared oven following a methodology established with the coating manufacturer.

Accelerated aging

One set of specimens underwent accelerated aging treatments in a Cincinnati SubZero environmental stimulation chamber (WM-906-MP2H-3-SC/WC) with temperature and RH precisions of ±1°C and ±3 percent, respectively. Treatment was based upon ASTM D3459 (American Society for Testing and Materials 1998) and consisted of four cycles of 48 hours at 50°C and 10 percent RH followed by 48 hours at 50°C and 90 percent RH. Prior to treatment, specimen ends were sealed with paraffin to reduce moisture exchange through cross section. After aging, specimens were conditioned at 20°C and 40 percent RH until they reached their initial EMC (10%).

Adhesion tests

Adhesion of aged and unaged films was evaluated by means of a pull-off test according to ASTM D4541 (ASTM International 2002). A MTS QT5 universal testing machine having a maximal load capacity of 5 kN and ±0.12 percent precision was used. Small 20-mm-diameter dollies were glued on the film surface with Araldite 2011 two-part epoxy resin. After 24 hours of curing at 20°C and 40 percent RH, the perimeters of glued dollies were carefully incised to prevent propagation of failures out of the tested area. Pulling was applied at 1 mm/min until separation of the dolly from the substrate. Maximal normal pull strength at rupture was recorded.

Statistical analyses

A factorial design with two factors, feed speed and rake angle, was established for the experiment. Data followed a randomized block design, and results from mechanical tests were analyzed with the mixed procedure in SAS as repeated measures. Surface topography, wettability, and surface energy were analyzed as a one-way analysis of variance, following the general linear model procedure. Mean difference comparison tests were performed at a 5 percent probability level when required. Simple correlations between surface quality parameters and the pull-off results were studied using the correlation procedure. Analysis was done on SAS statistical package, version 9.2 (SAS Institute Inc. 2007).

Microscopy

End-grain surfaces

Peripheral planing treatments used in this experiment induced slight subsurface damage in black spruce wood surfaces. When it occurred, damage was limited to the first row of cells in both earlywood and latewood (Figs. 1 through 3). Few differences were observed among samples planed with 10° and 15° rake angles regardless of feed speed (Figs. 1 and 2). However, samples planed at a 20° rake angle seemed to have suffered more damage when the wavelength was larger than 0.8 mm (Fig. 3). The waterborne coating did not penetrate into cells. This is in agreement with previous scanning electron microscopy observations of the interface of acrylic coatings and wood surfaces (Thay and Evans 1998). It only filled lumens that were available on the surface after planing. Because of its higher radial diameter, more lumens were available in earlywood than in latewood (de Moura et al. 2010). Cutting action occurred more in latewood, and failures, in that case, often took place in or close to the middle lamella. Consequently, filling of lumens by the coating was lower in latewood than in earlywood areas (Figs. 1 through 3).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Specimens planed with a rake angle of 10° and 15° were similar and characterized by superficial subsurface damage (Figs. 1 and 2), comparable with what was reported for peripheral-planed black spruce wood with a 12° rake angle (Cool and Hernández 2011a). However, little crushing, more or less severe, was locally observed in earlywood (Fig. 1B) and in latewood (Fig. 2B). In some areas, cell walls in latewood showed microruptures (Fig. 1C). These microruptures, as well as slight crushing, were only observed in samples planed with a rake angle of 10°. Given that normal cutting force is of compressive nature throughout the visible part of the cutting path (Iskra and Hernández 2012), and because Type III chips are probably produced with low rake angles in peripheral up-milling (Koch 1985), it can be assumed that lower rake angles could compress cells to a greater extent. As a result, samples showed occasional slight crushing and microruptures, which were not seen in samples planed with a 15° rake angle.

Samples planed with a rake angle of 20° and a wavelength of 0.8 mm were virtually free of subsurface damage (Fig. 3A). When wavelength increased to 1.0 and 1.3 mm, slight crushing was induced in earlywood (Figs. 3B and 3C) and latewood (Fig. 3D) areas. According to Singh et al. (2010), deformed cells can be observed even though cell walls appear to be intact. With a rake angle of 20°, a Type II chip should be principally induced in the visible part of the cutting path (Koch 1985). As wavelength increases, Iskra and Hernández (2012) reported little effect on the negative normal force component, responsible for cell deformation and crushing, in peripheral up-milling of sugar maple wood. However, increasing wavelength had an impact on the parallel force component. The resulting cutting force could partially be responsible for the observed cell damage. However, subsurface damage at these conditions was very slight and still lower than that reported for face-milled (Cool and Hernández 2011b) or sanded surfaces machined with a single-stage program (Cool and Hernández 2011c).

Tangential surfaces

All surfaces were characterized by open lumens of tracheids and rays, plateau-like areas, and fibrillation. The availability of lumens should favor filling and coating penetration and increase mechanical anchorage of coating (Figs. 4 through 6; Singh and Dawson 2004, 2006). Fibrillation should also increase coating mechanical anchorage. According to micrographs, little effect of the feed speed could be observed, whereas the rake angle had a higher impact on the planed surfaces.

View Figure

• Download as Powerpoint
• Download Figure

At 10° rake angle, feed speed had little effect on tangential surfaces. Regardless of feed speed, surfaces showed low fibrillation, which corresponded to torn portions of tracheid cell walls (Figs. 4A, 5A, and 6A). Plateau-like regions, similar to those obtained by oblique cutting and helical planing of black spruce (Cool and Hernández 2011b) and by oblique cutting of sugar maple (de Moura and Hernández 2007), red oak, and white pine (Pinus strobus) wood (de Moura et al. 2010), were also observed. As reported by de Moura et al. (2010) for oblique-cut surfaces, the cutting action in peripheral planing took place by peeling in or close to the middle lamella. Ashby et al. (1985) reported similar results with various woods when studying crack propagation in the same cutting plane of this study (radial–longitudinal plane).

View Figure

• Download as Powerpoint
• Download Figure

As rake angle was increased to 15°, the fibrillation for surfaces prepared with a wavelength of 0.8 mm was similar to that observed on specimens planed with a rake angle of 10° (Figs. 4A, 4B, 5A, and 6A). As wavelength was increased to 1.0 and 1.3 mm, fibrillation increased (Figs. 5B and 6B). However, little subsurface damage was observed in end-grain surfaces (Fig. 2). Consequently, the resulting surface quality was probably associated with formation of chip Types II and III. Because surface quality was similar for samples planed with a wavelength of 0.8 mm, the cutting action (chip formation) was probably the same as that of samples prepared with a 10° rake angle. However, increasing wavelength could have increased cutting forces and slightly modified the cutting action. Hence, it took place through cell walls, and it amplified fibrillation levels slightly.

View Figure

• Download as Powerpoint
• Download Figure

When planing with a rake angle of 20°, fibrillation was reduced, but an important level of tearing appeared (Figs. 4C and 5C) with wavelengths of 0.8 and 1.0 mm. This tearing could be associated with Type I chip formation, which is favored when a knife is traveling on local grain deviation areas (against the grain milling). As wavelengths reached 1.3 mm, the normal cutting force is increased, and slight crushing (Figs. 3B through 3D) as well as plateau-like regions were visible on the surfaces (Fig. 6C). These surfaces were similar to those planed with a rake angle of 10° as well as those planed with a rake angle of 15° with a wavelength of 0.8 mm. It appears that increasing the rake angle to 20° modified the cutting action of the knife. Planing still occurred by peeling in or close to the middle lamella, but microruptures or cracks developed and extended below the surface, as does a Type I chip. As a result, surfaces were characterized by a certain level of tearing.

Information gathered from micrographs indicates that surface features depend on the cutting parameters. More specifically, cutting action and chip formation could change with rake angle and feed speed and affect surface quality at the microscopic level.

Surface roughness

Slightly torn grain was observed near the knots in the samples planed with a rake angle of 15° and a wavelength of 1.3 mm and a rake angle of 20° with wavelengths of 1.0 and 1.3 mm. Previous authors have reported similar results when planing white spruce wood (Cantin 1967, Hernández et al. 2001). Based on ASTM D1666 (ASTM International 2004), these authors concluded that a higher proportion of defect-free surfaces was obtained at a 15° rake angle with a wavelength of 1.3 mm. However, Hernández et al. (2001) observed that torn grain was less important in samples planed with a 10° rake angle, even though the presence of fuzzy grain increased. Because wood density of black spruce wood is higher than that of white spruce wood (Jessome 1977) and studied wavelengths were equal to 1.3 mm or smaller, the impact of the rake angle on the occurrence of surface defects was diminished.

Feed speed and rake angle both had a significant effect on four surface roughness parameters, S_A, S_P, S_K, and S_PK, but not on S_V and S_VK, where no significant differences were detected between treatments (Table 1). The four roughness parameters were also significantly affected by the interaction between rake angle and feed speed. In agreement with observations made from Figures 4 through 6, rake angle appeared to affect surface roughness more. For S_A and S_K, smoother surfaces were obtained when using a rake angle of 20° with a wavelength of 1.0 or 1.3 mm and a rake angle of 15° with a wavelength of 0.8 mm. The lower fibrillation observed on these surfaces contributed to significantly minimize surface roughness (Figs. 4 through 6; Table 1). These samples were also characterized by plateau-like regions, which further reduced S_A and S_K.

Table 1. Three-dimensional surface roughness parameters (μm) of black spruce wood specimens prepared by peripheral planing at three rake angles and wavelengths.^a

View Fullscreen

The parameters S_P and S_PK both give information on peaks in the surface roughness profile. More specifically, S_PK was often related to fuzzy grain (Fujiwara et al. 2005, Gurau et al. 2005) or to the level of fibrillation (Cool and Hernández 2011a, 2011b, 2011c). Because none of the samples were characterized by fuzzy grain, S_PK could be related to the fibrillation levels observed in samples. According to Table 1, values of S_P and S_PK were lower for surfaces planed with a rake angle of 20° at wavelengths of 1.0 and 1.3 mm and for those planed with rake angles of 10° and 15° at a wavelength of 0.8 mm. These samples were characterized by smooth areas with low fibrillation levels (Figs. 4 through 6), which contributed to the decrease of both S_P and S_PK parameters.

Wettability tests

No effect of machining parameters was detected on measured contact angle or the spreading and penetration of oopure water (Fig. 7). However, machining parameters had a significant effect on surface energy (Tables 2 and 3). The nonpolar or disperse component of the surface energy was significantly affected by the interaction between rake angle and feed speed (Table 2). Smaller values were observed for surfaces machined with a rake angle of 20° at wavelengths of 1.0 and 1.3 mm, whereas the best value was obtained at this rake angle but at a wavelength of 0.8 mm. Consequently, more sites were available to form hydrogen bonds with the waterborne coating on the latter surfaces. Lower normal cutting force associated with shorter wavelengths could have generated less heat, which reduced surface inactivation compared with higher feed speeds. Polar component was significantly affected by rake angle and feed speed. However, the interaction between both cutting parameters was not statistically significant. Thus, mean comparisons were performed separately for each source of variation by pooling values of the other source of variation (Table 3). This table shows that a rake angle of 20° and a wavelength higher than 0.8 mm significantly increased the polar component of surface energy. As reported by Gindl et al. (2004), a more important polar component is related to more hydrophilic surfaces. Although a rake angle of 20° and wavelengths of 1.0 and 1.3 mm resulted in a high polar component, the disperse component was lower for these cutting parameters (Table 2). The total surface energy had a similar behavior to the disperse component (Table 2). As for the latter, the former is significantly higher for surfaces planed with a rake angle of 20° at 0.8-mm wavelength. This shows that the total surface energy is strongly influenced by the disperse component of surface energy.

View Figure

• Download as Powerpoint
• Download Figure

Table 2. Surface energy components determined by the harmonic mean method for black spruce wood specimens prepared by peripheral planing with three rake angles and wavelengths.

View Fullscreen

Table 3. Results of the mean difference comparison tests performed on the data of the polar component of surface energy determined by the harmonic mean method for black spruce wood specimens prepared by peripheral planing at three rake angles and three wavelengths.

View Fullscreen

Adhesion tests

Results of the pull-off tests are summarized in Table 4. Before aging, the interaction of feed speed and rake angle had a significant effect on pull-off strength. Based on these results, samples planed with a rake angle of 10° and a wavelength of 1.3 mm, a rake angle of 15° and wavelengths above 0.8 mm, and a rake angle of 20° and a wavelength of 0.8 mm had the highest pull-off strengths (Table 4). Pull-off strengths of planed samples were similar to those of sanded black spruce wood samples (Cool and Hernández 2011c). As mentioned previously, planed surfaces presented some fibrillation that could increase surface roughness (Figs. 4 through 6; Table 1). Therefore, mechanical adhesion was significantly favored. The chemical adhesion could also have been improved for the samples prepared with a 20° rake angle through an increase in disperse and polar components of surface energy (Tables 2 and 3). During accelerated aging treatment, pull-off strength was significantly reduced by about 20 percent (Table 4). This behavior is similar to that observed by Cool and Hernández (2011b) on coated black spruce wood samples machined with alternative processes. However, after aging, water-based coating adhesion was not affected by either feed speed or rake angle. Furthermore, there were no statistically significant correlations between pull-off strength and the surface quality parameters (roughness and wettability) that were evaluated. Therefore, the pull-off strength test is required to assess coating performance on black spruce wood.

Table 4. Pull-off strength, before and after an accelerated aging treatment, of a waterborne interior coating applied on black spruce wood specimens prepared by planing at three rake angles and wavelengths.^a

View Fullscreen

Loss in adhesion (percentage) during the accelerated aging treatment was statistically influenced by rake angle. This loss in adhesion was lower than that observed for sanded samples (Cool and Hernández 2011c). Loss in adhesion was lower for samples prepared with rake angles of 10° and 20° (Table 5). Thus, a surface characterized by low subsurface damage and a certain level of fibrillation could favor adhesion of the studied waterborne coating. As reported by Cool and Hernández (2011a), when gluing planed black spruce wood, it seems that low cellular damage prevented degradation of the wood-coating interface and that fibrillation level observed was sufficient to enhance mechanical anchorage. Surfaces planed with a rake angle of 10° did not present torn grain and were defined by higher surface roughness (Table 1) and lower surface energy (Tables 2 and 3), which was constant for the three studied wavelengths. On the other hand, quality of surfaces planed with a 20° rake angle was influenced by wavelength, and the surfaces were characterized by torn grain, lower surface roughness, and higher surface energy. Therefore, chemical adhesion could be as important as the mechanical adhesion depending on cutting parameters.

Table 5. Loss in adhesion of a waterborne interior coating applied on black spruce wood specimens prepared by peripheral planing at three different rake angles.^a

View Fullscreen

According to our results, a rake angle of 10° at a wavelength of 1.3 mm is recommended, because it produced surfaces characterized by little torn grain and good initial pull-off strength and low loss in adhesion during accelerated aging. Compared with a rake angle of 20° that induced a similar loss in adhesion, cutting forces should be more important. Because the clearance angle was constant for all rake angles, the knife angle was higher for a 10° rake angle and should compensate for the higher wear rate that could be expected at lower rake angles. Furthermore, samples planed with a rake angle of 20° induced torn grain when feed speed was superior to 0.8-mm wavelength. Hence, the use of a rake angle of 10° with a wavelength of 1.3 mm (13 m/min feed speed) would be appropriate.