- Email: [email protected]
The effect of anisotropicity of Norway spruce (Picea abies) during two-body abrasion
Wear 272 (2011) 38–42
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
The effect of anisotropicity of Norway spruce (Picea abies) during two-body abrasion Pasi Karinkanta ∗ , Mirja Illikainen, Jouko Niinimäki Department of Process and Environmental Engineering, Fibre and Particle Engineering Laboratory, P.O. Box 4300, FIN-90014 University of Oulu, Finland
a r t i c l e
i n f o
Article history: Received 19 January 2011 Received in revised form 29 June 2011 Accepted 11 July 2011 Available online 20 July 2011 Keywords: Anisotropic Electron microscopy Norway spruce (Picea abies) Particle size and shape Surface analysis Two-body abrasion
a b s t r a c t The effect of anisotropicity of Norway spruce (Picea abies) during two-body abrasive wear was investigated by rubbing the wood with five different orientations while using constant surface pressure and a sanding belt with very fine abrasive grits. The anisotropic nature was found to affect the microstructure of the worn surface and the breakage mechanism of the surface. The properties of the particles that were released from the surface during abrasion were dependent on the grinding orientation and if the particle originated from early- or latewood. The wear process was influenced by the anisotropic nature of wood. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Norway spruce (Picea abies) is the most common wood species used by the northern European woodworking industry. In this industrial domain, two-body abrasion is typically found in the surface finishing stage. Due the anisotropic and inhomogeneous nature of wood, the resulting surface and released wood particles are influenced by the orientation of the wood when it is imposed on the abrasive wear. Ohtani with co-workers has studied the wear process and worn surfaces of different wood species during two-body [1,2] and threebody abrasion [2–4] considering the anisotropic nature. They have found out that during the two-body abrasion of katsura wood, the wear rate increases when higher load is applied and that the twobody wear is affected by the anisotropicity of the wood. In the case of Norway spruce, the effect of anisotropicity under the abrasive wearing process has not been studied. This study aimed to gain thorough understanding about the effect of anisotropicity of Norway spruce under the two-body abrasion considering the surface under wear, particles generated during wearing and the wear process. The effect of anisotropy of Norway spruce under abrasive wear was studied by carrying out the sanding of cubic wood pieces that were cut from the same tree. These pieces were ground so that
∗ Corresponding author. Tel.: +358 8 553 2578; fax: +358 8 553 2405. E-mail address: pasi.karinkanta@oulu.fi (P. Karinkanta). 0043-1648/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2011.07.004
physically three different surfaces were imposed on the abrasion. The effect of the abrasion on the wood was inspected from two points of view: by considering the properties of the surface of the worn wood and the properties of the wood particles released during wearing. The wear process was investigated by studying the wear rate and energy consumption. The results indicated that the anisotropicity of wood affects the resulting surface, released particle properties and wear rate as well as capability to resist wear when it is imposed on the abrasion. 2. Materials and methods 2.1. Wood samples The raw material used in the experiments was dried Norwegian spruce (Picea abies) with dry matter content approximately 94%. Cubic wood pieces with an average length of 34 mm (whose exact dimensions were measured) were prepared as far away as possible from the pith to neglect the effect of curvature of the growth rings making it possible to consider the wood as an orthotropic material instead of anisotropic. The prepared wood pieces contained six faces where opposing faces could be considered identical by the material behaviour due to the orthotropic nature. In this study three surfaces perpendicular to the axes L, R and T (see Fig. 1) were investigated. These three faces of the cubic sample can be stressed in two main directions which contain the largest differences between mechanical properties when considering abrasive wear. These form six totally different com-
P. Karinkanta et al. / Wear 272 (2011) 38–42
39
Table 1 The averaged wear rate and energy usage for different orientations.
Fig. 1. Schematic illustration of a cubic sample with the grinding directions and three different surfaces: Top, Side and Out.
binations to impose shear stresses on orthotropic wood sample: Out-L, Out-T, Side-L, Side-R, Top-R and Top-T. Out, Side and Top identify the sample’s face perpendicular to the axis R, T or L (see Fig. 1) towards the abrasive belt while the second letter refers to the moving direction of the abrasive belt. In this investigation the orientations Out-L, Out-T, Side-L, Side-R and Top-R were studied. Replicate runs were done for investigated orientations in order to estimate the confidence interval. 2.2. Grinding experiments Experimental setup for the abrasive grinding is illustrated in Fig. 2. A piece of wood cut from the Norway spruce i.e. a cubic wood sample was placed on the top of a belt sander. An obstacle was positioned alongside the abrasive belt of the grinding machine to prevent any movement of the wood sample while in addition to acting as a device to secure a weight, positioned on top of the wood sample, to prevent any horizontal movement. The weight was used to adjust the compressive load on the sample while the centre of the gravity of the weight was balanced with the wood sample to prevent the uneven abrasion of the wood. Before experiments took place, the wood samples were pre-ground using an abrasive belt with a grit size below 125 m (P120) to ensure an identical initial stage for every sample. The experiments were carried out with an abrasive belt containing corundum abrasives with a grit size below 68 m (P220). The wood powders produced during wearing were collected after the obstacle from the surface of the belt with an efficient hoover and gathered into small dust bags where they were later removed for analysis. The abrasive belt was replaced between samples. Grinding experiments were performed using a compressive loading of 36 N. The dry mass of the cubic samples was lower than 20 g before grinding, which means that their contribution to the surface pressure was negligible. The speed of the abrasive belt varied between 47 and 51 rpm during the trials. It was decided that the grinding process would be kept running as long as the weight
Orientation
Pressure [kPa]a
Average wear rate [mgb /s]
Average energy usage [J/gb ]
Side-L Out-L Side-R Out-T Top-R
30.3 30.7 30.1 30.9 31.2
81 81 86 70 6
896 894 548 934 10,050
a Pressure was calculated from the force imposed by the weight to the surface area of the cubic sample assuming that the whole area is in contact with the belt. b Oven dry mass.
was not touching the obstacle below (A in Fig. 2). However, due to an inefficient grinding process of the specimens in the Top-R orientation, the process was stopped even though weight remained far away from the obstacle. The mass of the wood samples was measured before and after grinding to obtain an idea of how much wood powder was produced by wearing. In addition, the grinding time and average power consumption of the grinding machine were recorded. The average wear rate and energy consumption during the grinding process were calculated using Eqs. (1) and (2), respectively. ˙ = m
(m0 − mg ) , tg
(1)
˙ is the wear rate, m0 is the mass of the wood sample before where m grinding, mg is the mass of the wood sample after grinding and tg is the time used in the grinding process. E=
(Pa − Pi ) , ˙ m
(2)
where E is the average energy consumption per gram, Pa is the average power consumption during the grinding process and Pi is the average power consumption during idle running. 2.3. Analysis of the worn surfaces Images of the worn surfaces were obtained with a Field Emission Scanning Electron Microscope (FESEM, Zeiss Ultra plus). Confocal Laser Scanning Microscope (CLSM, Zeiss LSM 5 Pascal) was used to measure the maximum and average peak to the valley values of the worn surfaces. 2.4. Analysis of the released wood particles Particle size distributions of the wood powders were measured with a laser diffraction instrument (Beckman Coulter LS 13320) using the Fraunhofer optical model. Particle size measurements were carried out in dilute water-wood suspension. Particle shape was investigated by analysing the optical images of the powders that were gathered with CCD-camera from a tube flow of the samples in very dilute water-wood suspension. In the imaging section the flow was introduced into cuvette. Images were used to gain information about the aspect ratio from the particles larger than 2 m. Over 30,000 particles were analysed per sample to obtain projected area based aspect ratio distribution. 3. Results 3.1. Average wear rate and energy usage
Fig. 2. Schematic illustration of a used experimental setup for surface grinding. The black arrow indicates the direction of the abrasive belt.
The average wear rate and energy consumption of the studied wear processes are presented in Table 1. The wear rate was the lowest, with less than ten percent of the other orientations, in the case of the Top-R orientation. In the case of orientation Out-T the
40
P. Karinkanta et al. / Wear 272 (2011) 38–42
Fig. 3. FESEM images of the ground wood surface from the sections of early- and latewood. Grinding was accomplished in the vertical direction.
wear rate was smaller than in the case of Side-L, Out-L and SideR. The remaining orientations possessed slightly similar values of wear rate. Energy usage values for the cubic wood piece showed the Top-R orientation to be the most difficult to grind whereas the Side-R orientation to be the easiest.
Table 2 Roughness values of the wood surface ground with different orientations. Variation of the average values is estimated as a confidence interval with 95% confidence level assuming a two-tailed normal distribution.
3.2. Worn surfaces FESEM images of the ground wood surfaces are presented in Fig. 3. Similar surfaces were seen also from the replicates. In the case of Side-L and Out-L orientations wood samples were ground along the longitudinal axis. Grinding with these orientations produced surfaces containing mainly long vertical slits. Partly separated particles on the surface of earlywood appeared elongated in shape. In both orientations the surface of latewood is smoother than the surface of earlywood. In the case of Side-R and Out-T orientations wood samples with the surfaces perpendicular to the tangential and radial axis were ground crosswise to the longitudinal axis. In this case the surfaces of earlywood appeared similar with broken cell walls that were still partly attached to the wood matrix. In both orientations the surface of the latewood appears to be smoother and does not have large broken parts of cell walls alike in the earlywood section. The earlywood surface produced by grinding the surface perpendicular to the longitudinal axis in the radial direction (Top-R) has a lot smoother microstructure compared with other earlywood surfaces (Fig. 3). In this case there are minimal differences between the surfaces of early- and latewood although earlywood is slightly coarser. The values of the surface roughness are listed in Table 2. According to Table 2, the surface ground in the Top-R orientation contains the smallest average and maximum peak to valley values and therefore possesses the smoothest surface compared with the other samples. The surface ground in the Side-L orientation contains a
Orientation
Average peak to valley [m]
Side-L Out-L Side-R Out-T Top-R
10.6 13.7 12.3 11 5.6
± ± ± ± ±
0.9 1.0 0.8 2 0.7
Maximum peak to valley [m] 17.8 21.1 20.8 18 11
± ± ± ± ±
0.7 1.2 1.3 3 3
smaller value of maximum peak to valley than in the Out-L and Side-R orientations and a smaller value of average peak to valley than in the Out-L orientation. Other surfaces contain the values of average and maximum peak to valley rather similar to each other and due the variation the differences are not significant between Out-L, Side-R and Out-T. 3.3. Particle size distribution Differential particle size distributions of the powders produced during the abrasion are shown in Fig. 4, whereas Table 3 presents four different particle sizes from the volume based cumulative distribution. According to Fig. 4 and Table 3 Side-L and Out-L orientations produced powders with the widest particle size distribution while the narrowest particle size distribution and the smallest median size were obtained by grinding in orientation Top-R. Orientations Out-T and Side-R produced powders with the narrower particle size distribution than in the case of Side-L but wider than in the case of Top-R. The median size of the wood powders produced using orientations Out-T and Side-R were smaller than with the powders produced using orientation Side-L but larger than powders produced with orientation Top-R. The median size of the
Table 3 Various particle size values obtained from the volume based cumulative size distribution of the produced wood powders. Variation is estimated as a confidence interval with 95% confidence level assuming a two-tailed normal distribution. Orientation
Median size [m]
Side-L Out-L Side-R Out-T Top-R
53 41 38.7 34 24.6
± ± ± ± ±
9 3 0.7 5 0.7
25% below size [m] 24 19.0 19.5 17 11.54
± ± ± ± ±
4 0.7 0.7 3 0.06
75% below size [m] 110 80 63.9 58 42
± ± ± ± ±
20 8 0.8 10 3
90% below size [m] 190 138 98.9 100 62
± ± ± ± ±
40 15 1.2 30 8
P. Karinkanta et al. / Wear 272 (2011) 38–42
41
Fig. 5. Projected area based aspect ratio distribution of the produced wood powders.
powders with the widest aspect ratio distribution are produced with Side-L and Out-L orientations. Powders gained during grinding with the orientations Out-T and Side-R contain similar distributions which are narrower than in the case of Side-L and Out-L but wider than in the case of Top-R. The powder with the narrowest aspect ratio distribution is obtained with orientation Top-R. To describe the appearance of the particles gained with different orientations several particles imaged with CCD camera are presented in Fig. 6. 4. Discussion
Fig. 4. Volume based differential particle size distributions of the produced powders from (a) the experiments and (b) the replicates.
powders produced with orientation Out-L is between median sizes gained with orientations Side-R and Side-L. 3.4. Particle shape analysis Aspect ratio distributions of the produced powders with different orientations are shown in Fig. 5. According to Fig. 5, the
The average wear rate and energy usage (Table 1) showed that the most difficult surface to wear was the Top-surface. Wood is very well known to be capable of resisting crack propagation across the grain better than in any other directions [5,6] and therefore it appears that small cracks involved in the fracture on the surface of wood can resist crack propagation better across the grain when sanding the surface perpendicular to the longitudinal axis than with other orientations. Grinding with Out-L and Side-L orientations produced worn surfaces in the earlywood section that appear to possess a lot of longitudinally broken tracheids that have left vertical slits on the surface shown in Fig. 3. This kind of breakage refers to the peeling off mechanism. This is supported by produced wood powders that contain relatively high amount of large particles (Table 3) and wide aspect ratio distribution (Fig. 5) compared with the other powders,
Fig. 6. Images of the particles gained with CCD-camera.
42
P. Karinkanta et al. / Wear 272 (2011) 38–42
which results in longitudinal wood pieces (Fig. 6). Grinding with Out-T and Side-R orientations produced earlywood surfaces that possess a number of partly attached rectangular particles where the slits, showing the broken edge of tracheids, are orientated nearly perpendicular to the grinding direction (Fig. 3). In this case the particles separated by grinding with these orientations contain narrower aspect ratio than powders ground along the longitudinal axis (Fig. 5) and the largest particles can be characterised as rectangular shaped parts of cell walls (Fig. 6). The mechanism in this case seems to be repeated compression and flattening with the subsequent breakage of the tracheid wall rather than peeling. Grinding with Top-R orientation produced very smooth surfaces (Fig. 3 and Table 2) and the separated particles were smaller and rounder compared with any other sanding orientations (Table 3 and Figs. 4 and 5). Latewood surfaces appeared to be smoother than earlywood surfaces in all of the sanding orientations (Fig. 3) after two-body abrasive wearing. Therefore it seems that only small particles are removed from this section keeping the produced surface rather smooth compared with earlywood. Thick cell walls and related low compressibility of the latewood is probably the main reason why only small particles are removed from the latewood section with the abrasive grinding. 5. Conclusion The anisotropic nature of Norway spruce under two-body abrasion leads to differences in the resistance of wear, resulting surface and powder properties. The lowest wear resistance is achieved with orientation Side-R whereas the Top-surface has the highest wear resistance. Top-surface is also the smoothest after the abra-
sion, whereas all other surfaces show rougher structure without clear differences. Orientations Out-L and Side-L produces larger and more elongated particles compared with the other orientations. Orientation Top-R produces wood particles of smaller size and roundish shape compared with the particles produced with other orientations. Worn surfaces originating from the latewood are always smoother than the surfaces originating from the earlywood. The particles separated from the latewood are smaller than the ones originating from the earlywood. Acknowledgements The authors gratefully acknowledge the financial support provided by the Graduate School in Chemical Engineering (GSCE) and the Finnish Cultural Foundation. The authors would also like to thank UPM for analysis and financial support as well as J. Karvonen, K. Kilpimaa and O. Laitinen for their help during laboratory analysis. References [1] T. Ohtani, The effects of mechanical parameters of the stress–strain diagram on wood abrasion, Wear 265 (2008) 1557–1564. [2] T. Ohtani, T. Yakou, S. Kitayama, Two-body and three-body abrasive wear properties of katsura wood, J. Wood Sci. 47 (2001) 87–93. [3] T. Ohtani, K. Kamasaki, Effect of microscopic tissue on three-body abrasion in cell structure of wood, Wear 262 (2007) 453–460. [4] T. Ohtani, K. Kamasaki, C. Tanaka, On abrasive wear property during three-body abrasion of wood, Wear 255 (2003) 60–66. [5] M.F. Ashby, F.R.S., K.E. Easterling, R. Harryson, S.K. Maiti, The fracture and toughness of woods, Proc. R. Soc. Lond. A 398 (1985) 261–280. [6] G. Prokopski, Investigation of wood fracture toughness using mode II fracture (shearing), J. Mater. Sci. 30 (1995) 4745–4750.
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
The effect of anisotropicity of Norway spruce (Picea abies) during two-body abrasion Pasi Karinkanta ∗ , Mirja Illikainen, Jouko Niinimäki Department of Process and Environmental Engineering, Fibre and Particle Engineering Laboratory, P.O. Box 4300, FIN-90014 University of Oulu, Finland
a r t i c l e
i n f o
Article history: Received 19 January 2011 Received in revised form 29 June 2011 Accepted 11 July 2011 Available online 20 July 2011 Keywords: Anisotropic Electron microscopy Norway spruce (Picea abies) Particle size and shape Surface analysis Two-body abrasion
a b s t r a c t The effect of anisotropicity of Norway spruce (Picea abies) during two-body abrasive wear was investigated by rubbing the wood with five different orientations while using constant surface pressure and a sanding belt with very fine abrasive grits. The anisotropic nature was found to affect the microstructure of the worn surface and the breakage mechanism of the surface. The properties of the particles that were released from the surface during abrasion were dependent on the grinding orientation and if the particle originated from early- or latewood. The wear process was influenced by the anisotropic nature of wood. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Norway spruce (Picea abies) is the most common wood species used by the northern European woodworking industry. In this industrial domain, two-body abrasion is typically found in the surface finishing stage. Due the anisotropic and inhomogeneous nature of wood, the resulting surface and released wood particles are influenced by the orientation of the wood when it is imposed on the abrasive wear. Ohtani with co-workers has studied the wear process and worn surfaces of different wood species during two-body [1,2] and threebody abrasion [2–4] considering the anisotropic nature. They have found out that during the two-body abrasion of katsura wood, the wear rate increases when higher load is applied and that the twobody wear is affected by the anisotropicity of the wood. In the case of Norway spruce, the effect of anisotropicity under the abrasive wearing process has not been studied. This study aimed to gain thorough understanding about the effect of anisotropicity of Norway spruce under the two-body abrasion considering the surface under wear, particles generated during wearing and the wear process. The effect of anisotropy of Norway spruce under abrasive wear was studied by carrying out the sanding of cubic wood pieces that were cut from the same tree. These pieces were ground so that
∗ Corresponding author. Tel.: +358 8 553 2578; fax: +358 8 553 2405. E-mail address: pasi.karinkanta@oulu.fi (P. Karinkanta). 0043-1648/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2011.07.004
physically three different surfaces were imposed on the abrasion. The effect of the abrasion on the wood was inspected from two points of view: by considering the properties of the surface of the worn wood and the properties of the wood particles released during wearing. The wear process was investigated by studying the wear rate and energy consumption. The results indicated that the anisotropicity of wood affects the resulting surface, released particle properties and wear rate as well as capability to resist wear when it is imposed on the abrasion. 2. Materials and methods 2.1. Wood samples The raw material used in the experiments was dried Norwegian spruce (Picea abies) with dry matter content approximately 94%. Cubic wood pieces with an average length of 34 mm (whose exact dimensions were measured) were prepared as far away as possible from the pith to neglect the effect of curvature of the growth rings making it possible to consider the wood as an orthotropic material instead of anisotropic. The prepared wood pieces contained six faces where opposing faces could be considered identical by the material behaviour due to the orthotropic nature. In this study three surfaces perpendicular to the axes L, R and T (see Fig. 1) were investigated. These three faces of the cubic sample can be stressed in two main directions which contain the largest differences between mechanical properties when considering abrasive wear. These form six totally different com-
P. Karinkanta et al. / Wear 272 (2011) 38–42
39
Table 1 The averaged wear rate and energy usage for different orientations.
Fig. 1. Schematic illustration of a cubic sample with the grinding directions and three different surfaces: Top, Side and Out.
binations to impose shear stresses on orthotropic wood sample: Out-L, Out-T, Side-L, Side-R, Top-R and Top-T. Out, Side and Top identify the sample’s face perpendicular to the axis R, T or L (see Fig. 1) towards the abrasive belt while the second letter refers to the moving direction of the abrasive belt. In this investigation the orientations Out-L, Out-T, Side-L, Side-R and Top-R were studied. Replicate runs were done for investigated orientations in order to estimate the confidence interval. 2.2. Grinding experiments Experimental setup for the abrasive grinding is illustrated in Fig. 2. A piece of wood cut from the Norway spruce i.e. a cubic wood sample was placed on the top of a belt sander. An obstacle was positioned alongside the abrasive belt of the grinding machine to prevent any movement of the wood sample while in addition to acting as a device to secure a weight, positioned on top of the wood sample, to prevent any horizontal movement. The weight was used to adjust the compressive load on the sample while the centre of the gravity of the weight was balanced with the wood sample to prevent the uneven abrasion of the wood. Before experiments took place, the wood samples were pre-ground using an abrasive belt with a grit size below 125 m (P120) to ensure an identical initial stage for every sample. The experiments were carried out with an abrasive belt containing corundum abrasives with a grit size below 68 m (P220). The wood powders produced during wearing were collected after the obstacle from the surface of the belt with an efficient hoover and gathered into small dust bags where they were later removed for analysis. The abrasive belt was replaced between samples. Grinding experiments were performed using a compressive loading of 36 N. The dry mass of the cubic samples was lower than 20 g before grinding, which means that their contribution to the surface pressure was negligible. The speed of the abrasive belt varied between 47 and 51 rpm during the trials. It was decided that the grinding process would be kept running as long as the weight
Orientation
Pressure [kPa]a
Average wear rate [mgb /s]
Average energy usage [J/gb ]
Side-L Out-L Side-R Out-T Top-R
30.3 30.7 30.1 30.9 31.2
81 81 86 70 6
896 894 548 934 10,050
a Pressure was calculated from the force imposed by the weight to the surface area of the cubic sample assuming that the whole area is in contact with the belt. b Oven dry mass.
was not touching the obstacle below (A in Fig. 2). However, due to an inefficient grinding process of the specimens in the Top-R orientation, the process was stopped even though weight remained far away from the obstacle. The mass of the wood samples was measured before and after grinding to obtain an idea of how much wood powder was produced by wearing. In addition, the grinding time and average power consumption of the grinding machine were recorded. The average wear rate and energy consumption during the grinding process were calculated using Eqs. (1) and (2), respectively. ˙ = m
(m0 − mg ) , tg
(1)
˙ is the wear rate, m0 is the mass of the wood sample before where m grinding, mg is the mass of the wood sample after grinding and tg is the time used in the grinding process. E=
(Pa − Pi ) , ˙ m
(2)
where E is the average energy consumption per gram, Pa is the average power consumption during the grinding process and Pi is the average power consumption during idle running. 2.3. Analysis of the worn surfaces Images of the worn surfaces were obtained with a Field Emission Scanning Electron Microscope (FESEM, Zeiss Ultra plus). Confocal Laser Scanning Microscope (CLSM, Zeiss LSM 5 Pascal) was used to measure the maximum and average peak to the valley values of the worn surfaces. 2.4. Analysis of the released wood particles Particle size distributions of the wood powders were measured with a laser diffraction instrument (Beckman Coulter LS 13320) using the Fraunhofer optical model. Particle size measurements were carried out in dilute water-wood suspension. Particle shape was investigated by analysing the optical images of the powders that were gathered with CCD-camera from a tube flow of the samples in very dilute water-wood suspension. In the imaging section the flow was introduced into cuvette. Images were used to gain information about the aspect ratio from the particles larger than 2 m. Over 30,000 particles were analysed per sample to obtain projected area based aspect ratio distribution. 3. Results 3.1. Average wear rate and energy usage
Fig. 2. Schematic illustration of a used experimental setup for surface grinding. The black arrow indicates the direction of the abrasive belt.
The average wear rate and energy consumption of the studied wear processes are presented in Table 1. The wear rate was the lowest, with less than ten percent of the other orientations, in the case of the Top-R orientation. In the case of orientation Out-T the
40
P. Karinkanta et al. / Wear 272 (2011) 38–42
Fig. 3. FESEM images of the ground wood surface from the sections of early- and latewood. Grinding was accomplished in the vertical direction.
wear rate was smaller than in the case of Side-L, Out-L and SideR. The remaining orientations possessed slightly similar values of wear rate. Energy usage values for the cubic wood piece showed the Top-R orientation to be the most difficult to grind whereas the Side-R orientation to be the easiest.
Table 2 Roughness values of the wood surface ground with different orientations. Variation of the average values is estimated as a confidence interval with 95% confidence level assuming a two-tailed normal distribution.
3.2. Worn surfaces FESEM images of the ground wood surfaces are presented in Fig. 3. Similar surfaces were seen also from the replicates. In the case of Side-L and Out-L orientations wood samples were ground along the longitudinal axis. Grinding with these orientations produced surfaces containing mainly long vertical slits. Partly separated particles on the surface of earlywood appeared elongated in shape. In both orientations the surface of latewood is smoother than the surface of earlywood. In the case of Side-R and Out-T orientations wood samples with the surfaces perpendicular to the tangential and radial axis were ground crosswise to the longitudinal axis. In this case the surfaces of earlywood appeared similar with broken cell walls that were still partly attached to the wood matrix. In both orientations the surface of the latewood appears to be smoother and does not have large broken parts of cell walls alike in the earlywood section. The earlywood surface produced by grinding the surface perpendicular to the longitudinal axis in the radial direction (Top-R) has a lot smoother microstructure compared with other earlywood surfaces (Fig. 3). In this case there are minimal differences between the surfaces of early- and latewood although earlywood is slightly coarser. The values of the surface roughness are listed in Table 2. According to Table 2, the surface ground in the Top-R orientation contains the smallest average and maximum peak to valley values and therefore possesses the smoothest surface compared with the other samples. The surface ground in the Side-L orientation contains a
Orientation
Average peak to valley [m]
Side-L Out-L Side-R Out-T Top-R
10.6 13.7 12.3 11 5.6
± ± ± ± ±
0.9 1.0 0.8 2 0.7
Maximum peak to valley [m] 17.8 21.1 20.8 18 11
± ± ± ± ±
0.7 1.2 1.3 3 3
smaller value of maximum peak to valley than in the Out-L and Side-R orientations and a smaller value of average peak to valley than in the Out-L orientation. Other surfaces contain the values of average and maximum peak to valley rather similar to each other and due the variation the differences are not significant between Out-L, Side-R and Out-T. 3.3. Particle size distribution Differential particle size distributions of the powders produced during the abrasion are shown in Fig. 4, whereas Table 3 presents four different particle sizes from the volume based cumulative distribution. According to Fig. 4 and Table 3 Side-L and Out-L orientations produced powders with the widest particle size distribution while the narrowest particle size distribution and the smallest median size were obtained by grinding in orientation Top-R. Orientations Out-T and Side-R produced powders with the narrower particle size distribution than in the case of Side-L but wider than in the case of Top-R. The median size of the wood powders produced using orientations Out-T and Side-R were smaller than with the powders produced using orientation Side-L but larger than powders produced with orientation Top-R. The median size of the
Table 3 Various particle size values obtained from the volume based cumulative size distribution of the produced wood powders. Variation is estimated as a confidence interval with 95% confidence level assuming a two-tailed normal distribution. Orientation
Median size [m]
Side-L Out-L Side-R Out-T Top-R
53 41 38.7 34 24.6
± ± ± ± ±
9 3 0.7 5 0.7
25% below size [m] 24 19.0 19.5 17 11.54
± ± ± ± ±
4 0.7 0.7 3 0.06
75% below size [m] 110 80 63.9 58 42
± ± ± ± ±
20 8 0.8 10 3
90% below size [m] 190 138 98.9 100 62
± ± ± ± ±
40 15 1.2 30 8
P. Karinkanta et al. / Wear 272 (2011) 38–42
41
Fig. 5. Projected area based aspect ratio distribution of the produced wood powders.
powders with the widest aspect ratio distribution are produced with Side-L and Out-L orientations. Powders gained during grinding with the orientations Out-T and Side-R contain similar distributions which are narrower than in the case of Side-L and Out-L but wider than in the case of Top-R. The powder with the narrowest aspect ratio distribution is obtained with orientation Top-R. To describe the appearance of the particles gained with different orientations several particles imaged with CCD camera are presented in Fig. 6. 4. Discussion
Fig. 4. Volume based differential particle size distributions of the produced powders from (a) the experiments and (b) the replicates.
powders produced with orientation Out-L is between median sizes gained with orientations Side-R and Side-L. 3.4. Particle shape analysis Aspect ratio distributions of the produced powders with different orientations are shown in Fig. 5. According to Fig. 5, the
The average wear rate and energy usage (Table 1) showed that the most difficult surface to wear was the Top-surface. Wood is very well known to be capable of resisting crack propagation across the grain better than in any other directions [5,6] and therefore it appears that small cracks involved in the fracture on the surface of wood can resist crack propagation better across the grain when sanding the surface perpendicular to the longitudinal axis than with other orientations. Grinding with Out-L and Side-L orientations produced worn surfaces in the earlywood section that appear to possess a lot of longitudinally broken tracheids that have left vertical slits on the surface shown in Fig. 3. This kind of breakage refers to the peeling off mechanism. This is supported by produced wood powders that contain relatively high amount of large particles (Table 3) and wide aspect ratio distribution (Fig. 5) compared with the other powders,
Fig. 6. Images of the particles gained with CCD-camera.
42
P. Karinkanta et al. / Wear 272 (2011) 38–42
which results in longitudinal wood pieces (Fig. 6). Grinding with Out-T and Side-R orientations produced earlywood surfaces that possess a number of partly attached rectangular particles where the slits, showing the broken edge of tracheids, are orientated nearly perpendicular to the grinding direction (Fig. 3). In this case the particles separated by grinding with these orientations contain narrower aspect ratio than powders ground along the longitudinal axis (Fig. 5) and the largest particles can be characterised as rectangular shaped parts of cell walls (Fig. 6). The mechanism in this case seems to be repeated compression and flattening with the subsequent breakage of the tracheid wall rather than peeling. Grinding with Top-R orientation produced very smooth surfaces (Fig. 3 and Table 2) and the separated particles were smaller and rounder compared with any other sanding orientations (Table 3 and Figs. 4 and 5). Latewood surfaces appeared to be smoother than earlywood surfaces in all of the sanding orientations (Fig. 3) after two-body abrasive wearing. Therefore it seems that only small particles are removed from this section keeping the produced surface rather smooth compared with earlywood. Thick cell walls and related low compressibility of the latewood is probably the main reason why only small particles are removed from the latewood section with the abrasive grinding. 5. Conclusion The anisotropic nature of Norway spruce under two-body abrasion leads to differences in the resistance of wear, resulting surface and powder properties. The lowest wear resistance is achieved with orientation Side-R whereas the Top-surface has the highest wear resistance. Top-surface is also the smoothest after the abra-
sion, whereas all other surfaces show rougher structure without clear differences. Orientations Out-L and Side-L produces larger and more elongated particles compared with the other orientations. Orientation Top-R produces wood particles of smaller size and roundish shape compared with the particles produced with other orientations. Worn surfaces originating from the latewood are always smoother than the surfaces originating from the earlywood. The particles separated from the latewood are smaller than the ones originating from the earlywood. Acknowledgements The authors gratefully acknowledge the financial support provided by the Graduate School in Chemical Engineering (GSCE) and the Finnish Cultural Foundation. The authors would also like to thank UPM for analysis and financial support as well as J. Karvonen, K. Kilpimaa and O. Laitinen for their help during laboratory analysis. References [1] T. Ohtani, The effects of mechanical parameters of the stress–strain diagram on wood abrasion, Wear 265 (2008) 1557–1564. [2] T. Ohtani, T. Yakou, S. Kitayama, Two-body and three-body abrasive wear properties of katsura wood, J. Wood Sci. 47 (2001) 87–93. [3] T. Ohtani, K. Kamasaki, Effect of microscopic tissue on three-body abrasion in cell structure of wood, Wear 262 (2007) 453–460. [4] T. Ohtani, K. Kamasaki, C. Tanaka, On abrasive wear property during three-body abrasion of wood, Wear 255 (2003) 60–66. [5] M.F. Ashby, F.R.S., K.E. Easterling, R. Harryson, S.K. Maiti, The fracture and toughness of woods, Proc. R. Soc. Lond. A 398 (1985) 261–280. [6] G. Prokopski, Investigation of wood fracture toughness using mode II fracture (shearing), J. Mater. Sci. 30 (1995) 4745–4750.