UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Reducing the surface checking of deck-boards exposed to natural weathering: effects of wood species and… Cheng, Kenneth Jenkye 2015

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_may_cheng_kenneth.pdf [ 10.92MB ]
Metadata
JSON: 24-1.0166219.json
JSON-LD: 24-1.0166219-ld.json
RDF/XML (Pretty): 24-1.0166219-rdf.xml
RDF/JSON: 24-1.0166219-rdf.json
Turtle: 24-1.0166219-turtle.txt
N-Triples: 24-1.0166219-rdf-ntriples.txt
Original Record: 24-1.0166219-source.json
Full Text
24-1.0166219-fulltext.txt
Citation
24-1.0166219.ris

Full Text

 REDUCING THE SURFACE CHECKING OF DECK-BOARDS EXPOSED TO NATURAL WEATHERING: EFFECTS OF WOOD SPECIES AND SURFACE PROFILING    by   Kenneth Jenkye Cheng   B.Sc. (Wood Products Processing)., The University of British Columbia, 2010    A THESIS SUBMITTED IN PARTIAL FULFILLMENT  OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in  The Faculty of Graduate and Postdoctoral Studies  (Forestry)    THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)   April 2015    ©Kenneth Jenkye Cheng, 2015  ii  Abstract  Surface checking is a defect in wood decking that is highly disliked by consumers. Surface checking can be reduced by selecting species that are more resistant to checking, or profiling deck-boards with a series of V(rib) or U(ripple) shaped grooves. Most literature on checking of decking focuses on species that are already used as exterior products. This provides an opportunity to investigate the use of other under-utilized species as deck-boards. Surface profiling has been applied to deck-boards to reduce checking, but there is little research on why it is effective. I hypothesize that both species and the geometry of surface profiles will significantly influence checking of deck-boards exposed to natural weathering. To test this hypothesis I exposed deck-boards made from 9 untreated softwoods and 8 untreated hardwoods outside for one year and measured the checking of the boards. None of the species performed as well as western red cedar and ipe, durable species that resist checking. However, some diffuse porous hardwoods performed quite well and further improvements might be achieved with chemical or physical treatments. Profilometry was used to classify and identify the geometry of commercially profiled deck-boards. The ratio of the surface grooves (R1) to those of peaks (R2) classified profiles into two categories mentioned in the literature (rib and ripple). A new category of profile (ribble) was also identified that had intermediate characteristics of both rib and ripple profiles. New profiles with various R1/R2 and height to width (H/W) ratios of profile peaks and grooves were tested to examine the effect of profile geometry on the checking of Pacific silver fir boards exposed to natural weathering. Profiling reduced the width of checks but increased cupping of the boards. There was no consistent trend of R1/R2 and H/W ratios on checking, but rib profiles were better than ribble or ripple profiles at restricting checks. Therefore, I conclude that species and profile geometry influence checking. Furthermore, some of the rib profiles could be used with diffuse porous hardwoods and some softwoods to enable them to compete more effectively with decking made from durable wood species or wood plastic composites.        iii  Preface    Parts of Chapter 4 were presented as a poster at the American Wood Protection Association 110th Annual Meeting held in Newport Beach, California in 2014 under the tile “Use of confocal profilometry to describe, classify, and identify profiled decking”. I conducted all of the experimental research described and discussed in the poster, designed the initial layout of the poster, and presented the results at the conference. My co-author Dr. Philip D. Evans helped with the design of the experiment described in the poster and also with the layout of the poster.   Parts of Chapter 5 were presented at the Canadian Wood Preservation Association 35th Annual Meeting held in Vancouver in 2014 under the title: “Optimizing profiling to reduce the checking of Pacific silver fir decking”. I conducted the experimental research, wrote the manuscript and presented the results at the conference. My co-author Dr. Philip D. Evans helped with the experimental design, statistical analyses and edited the final manuscript. The citation for the paper is: Cheng, K., Evans, P.D. (2014). Optimizing profiling to reduce the checking of Pacific silver fir decking. Proceedings of the Thirty-Fifth Annual Meeting of the Canadian Wood Preservation Association (Oct. 28-29), Vancouver, B.C., Canada, 18 pp.          iv  Table of Contents  Abstract ...................................................................................................................................... ii Preface ....................................................................................................................................... iii Table of Contents ....................................................................................................................... iv List of Tables ............................................................................................................................... x List of Figures ............................................................................................................................. xi Acknowledgments ..................................................................................................................... xx Dedication ................................................................................................................................ xxi Chapter 1 : General Introduction .................................................................................................1 1.1 Wood decking ...................................................................................................................1 1.2 Surface checking ................................................................................................................4 1.3 Aim, hypothesis and significance .......................................................................................5 1.4 Study outline .....................................................................................................................5 Chapter 2 : Literature Review ......................................................................................................7 2.1 Introduction ......................................................................................................................7 2.2 Checking ............................................................................................................................8 2.2.1 Definitions ..................................................................................................................8 2.2.2 Mechanism (why checks form) .................................................................................12 2.3 Factors affecting checking ...............................................................................................16 2.3.1 Wood microstructure ...............................................................................................16 2.3.2 Wood species ...........................................................................................................19 2.3.2.1 Softwoods ...................................................................................................................... 20 2.3.2.2 Hardwoods ..................................................................................................................... 22 2.3.3 Growth ring orientation ............................................................................................26 2.3.3.1 Radial, tangential, rift and double rift ............................................................................. 26 2.3.3.2 Concave and convex growth ring orientations ................................................................ 28 v  2.3.4 Wood quality (juvenile v mature wood/density and grain angle etc.) ........................29 2.3.4.1 Density ........................................................................................................................... 29 2.3.4.2 The pith juvenile/mature wood ...................................................................................... 30 2.3.4.3 Grain angle ..................................................................................................................... 31 2.3.5 Wood defects (knots, compression wood etc.) .........................................................32 2.3.6 Weathering ...............................................................................................................35 2.4 Attempts to reduce checking ...........................................................................................37 2.4.1 Wood selection .........................................................................................................37 2.4.2 Center-boring ...........................................................................................................37 2.3.3 Planing and machining ..............................................................................................38 2.4.3 Profiling ....................................................................................................................40 2.4.3.1 History of surface grooving ............................................................................................. 40 2.4.3.2 Surface profiling of deck-boards ..................................................................................... 47 2.4.4 Kerfing ......................................................................................................................53 2.4.4.1 Square posts and rails ..................................................................................................... 54 2.4.4.2 Logs and round poles...................................................................................................... 54 2.4.4.3 Kerfing used in drying ..................................................................................................... 57 2.4.4.4 Kerfing of deck-boards ................................................................................................... 58 2.4.5 Incising .....................................................................................................................59 2.4.6 Other physical and mechanical treatments ...............................................................62 2.4.6.1 Screws and nails ............................................................................................................. 62 2.4.6.2 Hooks, rings and fasteners .............................................................................................. 64 2.4.6.3 Beveling ......................................................................................................................... 66 2.4.7 Exterior finishes (coatings) ........................................................................................67 2.4.7.1 Opaque type coatings ..................................................................................................... 68 2.4.7.2 Semi-transparent coatings .............................................................................................. 69 2.4.7.3 Oils and resins ................................................................................................................ 70 2.4.8 Preservative treatments ...........................................................................................71 2.4.8.1 Oil solvent based preservatives ...................................................................................... 71 vi  2.4.8.2 Water based preservatives ............................................................................................. 72 2.4.8.3 Other preservatives ........................................................................................................ 74 2.4.9 Water repellent additives .........................................................................................74 2.4.10 End-grain sealants...................................................................................................75 2.4.11 Wood modification .................................................................................................76 2.4.11.1 Acetylation ................................................................................................................... 77 2.4.11.2 Cross-linking, bulking and impregnation modification ................................................... 78 2.4.11.3 Thermal treatments...................................................................................................... 81 2.4.12 Cupping and attempts to reduce cupping ...............................................................81 2.4.12.1 Species and mechanical treatments .............................................................................. 82 2.4.12.2 Wood modification ....................................................................................................... 83 2.5 Concluding remarks .........................................................................................................84 Chapter 3 : Response of Eight Hardwoods and Nine Softwoods to Natural Weathering ............86 3.1 Introduction ....................................................................................................................86 3.2 Materials and methods ....................................................................................................87 3.2.1 Experimental design .................................................................................................87 3.2.2 Wood species ...........................................................................................................88 3.2.3 Preliminary characterization of species and samples ................................................90 3.2.4 Preparation of decking samples ................................................................................92 3.2.5 Weathering trial........................................................................................................92 3.2.6 Measurement of checking, erosion and discolouration .............................................93 3.2.6.1 Checking......................................................................................................................... 93 3.2.6.2 Erosion and micro-checking ............................................................................................ 94 3.2.6.3 Assessment of discolouration ......................................................................................... 94 3.2.6.3(a) Visual observations (Microtek) ............................................................................... 94 3.2.6.3(b) Spectroscopy .......................................................................................................... 95 3.2.6.3(c) Scanning electron microscopy................................................................................. 96 3.2.6.3(d) Insect-induced discolouration................................................................................. 97 3.2.6.4 Statistical analyses of data and graphical summaries ...................................................... 97 vii  3.3 Results ........................................................................................................................... 100 3.3.1 Qualitative assessment of checking ........................................................................ 100 3.3.2 Quantitative assessment of checking ...................................................................... 105 3.3.2.1 Check numbers and sizes .............................................................................................. 105 3.3.2.2 Correlation between density and checking ................................................................... 108 3.3.3 Discolouration of deck-board samples .................................................................... 112 3.3.3.1 Visual observations ...................................................................................................... 112 3.3.3.2 CIE colour parameters .................................................................................................. 114 3.3.3.3 Fungal discolouration ................................................................................................... 119 3.3.3.4 Insect induced discolouration ....................................................................................... 125 3.3.4 Erosion and micro-checking .................................................................................... 128 3.3.4.1 Erosion ......................................................................................................................... 128 3.3.4.2 Micro-checking ............................................................................................................. 131 3.3.5 General appearance of deck-boards ....................................................................... 134 3.4 Discussion...................................................................................................................... 139 3.5 Conclusions ................................................................................................................... 144 Chapter 4 : Use of Confocal Profilometry to Describe, Classify, and Identify Profiled Decking . 146 4.1 Introduction .................................................................................................................. 146 4.2 Materials and methods .................................................................................................. 147 4.2.1 Commercially manufactured profiled decking samples ........................................... 147 4.2.2 Confocal profilometry of profile dimensions ........................................................... 150 4.3 Results and discussion ................................................................................................... 151 4.3.1 Visual characteristics of the profiles ........................................................................ 151 4.3.2 Quantification of the geometry of profiles in commercial deck-boards ................... 161 4.4 Conclusions ................................................................................................................... 164 Chapter 5 : Optimizing Surface Profiling to Reduce the Checking of Pacific Silver Fir Deck-Boards ............................................................................................................................................... 165 5.1 Introduction .................................................................................................................. 165 viii  5.2 Materials and methods .................................................................................................. 165 5.2.1 Experimental design ............................................................................................... 165 5.2.2 Design of new surface profiles and manufacture of customized tooling .................. 167 5.2.3 Manufacture of profiled pacific silver fir decking .................................................... 170 5.2.4 Measurements of wood properties, cupping and profile geometry......................... 172 5.2.5 Weathering racks and exposure of samples ............................................................ 173 5.2.6 Measurement of macro-checks............................................................................... 174 5.2.7 Visualizing micro-checks ......................................................................................... 175 5.3 Results ........................................................................................................................... 176 5.3.1 Surface macro-checks ............................................................................................. 176 5.3.2 Positions of surface checks in deck-board samples ................................................. 178 5.3.3 Changes in profile geometry after weathering ........................................................ 181 5.3.4 Appearance of boards after weathering.................................................................. 182 5.3.5 Cupping of deck-board samples .............................................................................. 184 5.4 Discussion...................................................................................................................... 185 5.5 Conclusions ................................................................................................................... 187 Chapter 6 : General Discussion, Conclusions, and Suggestions for Further Research ............... 188 6.1 General discussion ......................................................................................................... 188 6.2 Conclusions ................................................................................................................... 192 6.3 Further research ............................................................................................................ 193 References .............................................................................................................................. 195 Appendices ............................................................................................................................. 219 Appendix 1: Checking and cupping of profiled and treated pacific silver fir and lodgepole pine deck-boards exposed to the weather for one year. ............................................................. 219 Appendix 2: Checking and cupping of the thirteen best combinations of radiata pine and mountain ash panels treated with preservatives, water-repellents, and coatings exposed outdoors for two years. ....................................................................................................... 220 ix  Appendix 3: Woods used in the study ................................................................................. 221 Appendix 4: AutoCAD images of knives designed for the experiment described in chapter 5 ............................................................................................................................................ 222                        x  List of Tables  Table 2.1: Summary of previous studies by Morris and co-workers that have tested the ability of surface profiling to reduce the checking of deck-boards ...........................................................49 Table 3.1: Literature values for the specific gravity of wood species exposed as deck-boards to natural weathering (Chudnoff 1980, Miles and Smith 2009) ......................................................89 Table 3.2: Growth rate, grain angle, basic density and range, wood type and grain contrast of wood species used for the decking test .....................................................................................91 Table 3.3: Significant effects and correlations between density and checking of the different species exposed outdoors ....................................................................................................... 108 Table 3.4: Species ratings for different appearance criteria ..................................................... 134 Table 4.1: Origins and density of the collected profiled, commercial and experimental decking samples ................................................................................................................................... 148 Table 4.2: Geometry of the profiles in commercial deck-board samples .................................. 161 Table 5.1: Dimensions of the designed profiles applied to Pacific silver fir deck-boards .......... 167 Table 5.2: Growth rate, density and grain angles of parent Pacific silver fir boards that were cut and machined to produce deck-board samples ....................................................................... 173             xi  List of Figures  Figure 1.1: Market segments for different deck-board materials in 2013. Percentages are based on total annual consumption of 3065 million lineal feet of decking (redrawn from: Freedonia Group 2014a,b,c) ........................................................................................................................1 Figure 1.2: Composite score of decking material attributes perceived by home and deck builders. Categories are calculated on a 7-point Likert-like scale based on consumers’ perceptions of decking materials. Higher values on the Likert-like scale means the product is viewed more favorably by the respondents (redrawn from Ganguly and Eastin 2009) ................3 Figure 1.3: Images of deck-boards exposed at a test site in Charlotte, North Carolina: (a) Quarter-sawn California redwood; (b) Flat-sawn California redwood (photos, Dr. Alan Preston) .4 Figure 2.1: Images of various defects caused by checking: (a) fungal fruiting bodies growing within checks in a test deck in Tsukuba, Japan; (b) severe check in a western red cedar power pole in Vancouver, Canada; (c) erosion of a 19th century Norfolk Island pine (Araucaria heterophylla (Salisb.) Franco) roofing shingle (photos, Dr. Philip D. Evans) ..................................7 Figure 2.2: Different types of end checks at the end of a Japanese cedar (Cryptomeria japonica (L.f.) D.Don) log: (a) star check; (b) end check; (c) shakes (photo, Dr. Philip D. Evans)................10 Figure 2.3: Diagram showing different types of physical failures of wood caused by surface stresses .....................................................................................................................................11 Figure 2.4: Mechanism of check formation in an unrestrained flat-sawn board during drying (a-e) and weathering (f-j). Brown= dry, blue = wet (redrawn from: McMillen 1955, Evans 2008)...13 Figure 2.5: Surfaces of a profiled radiata pine (Pinus radiata D. Don) deck-board: R=Radial longitudinal surface; T=Tangential longitudinal surface; X=transverse surface ..........................17 Figure 2.6: Visual assessment of severity of checking in deck-boards exposed outside in Vancouver or Ottawa for 9 years. WH: Western hemlock; WS: Western white spruce (Picea glauca (Moench) Voss); LP: Lodgepole pine (Pinus contorta var. latifolia Douglas); AF: Subalpine fir (Abies lasiocarpa (Hooker) Nutt.); ES: Eastern white spruce (Picea glauca (Moench) Voss); JP: Jack pine (Pinus banksiana Lamb); BF: Balsam fir (Abies balsamea (L.) Mill.); RP: Red Pine (Pinus resinosa Sol. ex Aiton); PP: Ponderosa pine (Pinus ponderosa Douglas ex C.Lawson); SP: Southern pine; WRC: Western red cedar. For cities: VAN: Vancouver; OTT: Ottawa. Note: Check appearance is rated from 0=no change to 4=severe checking. Red pine was not tested in Ottawa (redrawn from Morris and Ingram 2002) ...................................................................................21   xii  Figure 2.7: Checking of ten tropical species exposed to accelerated weathering for 2400 hours. BT: Burmese teak; PT: Plantation teak (Tectona grandis L.f.); CRP: Curupau (Anadenanthera macrocarpa (Benth.) Brenan); JIQ: Jichituriqui (Aspidosperma cylindrocarpon Müll.Arg.); IPE: Ipe; SCP: Sucupira (Diplotropis purpurea (Rich.) Amshoff); STO: Soto (Schinopsis quebracho-colorado (Schldl.) F.Barkley and T. Meyer); CUC: Cuchi (Astronium urundeuva (Fr. and All.) Engl.); ROB: Roble (Amburana cearensis (Allemão) A.C.Sm. ); MOMO: Momoqui (Caesalpinia cf. pluviosa DC); SIR: Sirari (Guibourtia chodatiana (Hassl.) J. Léonard); CTA: Cuta (Phyllostylon rhamnoides (Poisson)). Check appearance is rated on a scale of 1 to 10 where 1=severe change and 10=no change (redrawn from: Williams et al. 2001) ................................24 Figure 2.8: Graphs showing the relationship between check appearance and: (a) density; (b) warping; (c) swelling of different tropical wood species exposed to accelerated weathering for 2400 hours. Check appearance is graphed at a scale of 1 to 10 where 1=severe change and 10=no change (redrawn from: Williams et al. 2001) ..................................................................25 Figure 2.9: Orientation of growth rings at sawn wood surfaces: (a) Flat-sawn board; (b) Quarter-sawn board; (c) Rift-sawn board; (d) Double rift-sawn board (redrawn from: Tenorio and Moya 2011) ........................................................................................................................................26 Figure 2.10: Scots pine boards oriented: (a) Bark-side-up (Convex); (b) Pith-side-up (Concave).28 Figure 2.11: Shelling of a profiled western larch deck-board oriented pith-side-up (photo, Dr. Philip D. Evans)..........................................................................................................................29 Figure 2.12: Yellow cedar (Cupressus nootkatensis D. Don) containing compression wood: (a) longitudinal section containing darker coloured compression wood. Note the presence of compression wood around a knot (arrow); (b) transverse surface containing compression wood, note the darker bands of latewood in the growth rings of compression wood (arrow). Scale Bar = 20 mm (photos, Dr. Philip D. Evans)........................................................................................32 Figure 2.13: Cross cracks at the tangential surface of a hard pine (Pinus sp.) containing compression wood. Scale Bar = 5 mm .......................................................................................33 Figure 2.14: Images of scanned knots from off-cuts of wood used in Chapter 3: (a) intergrown knot in lodgepole pine; (b) loose knot in Alaskan yellow cedar. Scale Bar = 20 mm ...................34 Figure 2.15: Mixture of 125 mm and 150 mm diameter ACQ treated slash and radiata pine posts center-bored with 25 mm, 35 mm, and 45 mm diameter holes: (a) after treatment; (b) after several years of exposure (photos, Dr. Philip D. Evans)..............................................................38 Figure 2.16: Early examples of surface grooving: (a) Hand carved linen-fold paneling; (b) decorative profiled wood panel (top view); (c) decorative wood panel (side view); (d) cross-section of v-shaped grooved wood used to make compound lumber; (e) cross-section of notched grooved wood used to make compound lumber (Sources: Brock 1878, Mankey 1884, and Woodcarvers Guild 2003) ...................................................................................................41 xiii  Figure 2.17: Grooved and handsplit western red cedar shakes: (a) Irregular grooving; (b) hand split ...........................................................................................................................................42 Figure 2.18: Visualizing the contours of striated plywood: (a) Balanced striated plywood panel. Scale bar = 100 mm; (b) confocal profilometry image of part of the balanced striated plywood shown in (a); (c) extracted profile line scan from (b) .................................................................43 Figure 2.19: Use of Weldtex striated panels: (a) Weldtex used as square cladding on a house in Port Coquitlam during the 1950s (photo, Dan Price); (b) Weldtex used as interior paneling (Eichler Siding. http://eichlersiding.com/) .................................................................................44 Figure 2.20: Diagrams of Elmendorf’s profiled Douglas fir plywood: (a) top view; (b) side view (Source: Elmendorf 1950)..........................................................................................................45 Figure 2.21: Wood plastic composites with profiled surfaces: (a) Diagram of an early wood plastic composite with a reeded surface (Source: Roy 1955); (b) Contemporary wood plastic composite with a profiled surface (Ecowood. http://ecowood.co.za/) ......................................46 Figure 2.22: Examples of decking products with anti-slip inserts: (a) GripDeck board (Source: GripDeck 2015); (b) Grooved board with RetroGrip antislip deck metal inserts (Southern Timber. http://www.southern-timber.co.uk/); (c) Deck-board with an anti-slip insert consisting of a mixture of dried silica and calcined flint with two shallow surface drainage grooves (Source: Hill and Moss 2000); (d) Another type of anti-slip insert made from a mixture of dried silica and calcined flint. Note the absence of ‘surface drainage grooves’ (Source: Hill and Moss 2000) .....47 Figure 2.23: Bongossi deck-boards grooved with: (a) thin rectangular grooves; (b) wide rectangular grooves (Source: Shida et al. 1992). Scale bar = 100 mm ........................................48 Figure 2.24: Rib and ripple profiles used by Morris and McFarling in their study of the checking of Pacific silver fir and lodgepole pine deck-boards exposed to natural weathering: (a) ripple-flat edge; (b) ripple to edge; (c) rib-eased edge; (d) rib-flat edge and center (Source: Morris and McFarling 2008). Note footnote to Table 2.1. ............................................................................50 Figure 2.25: Confocal image of profiles used by Evans et al. (2010) in their study of the ability of rib and ripple profiles to reduce the checking of Pacific silver fir and southern pine: (a) rib profile; (b) ripple profile (Source: Evans et al. 2010) ..................................................................51 Figure 2.26: Examples of kerfed wood products: (a) Kerfing of boards used as the top of an outdoor restaurant table in Stockholm, Sweden; (b) kerfing underneath a highway crash barrier in Kyushu, Japan; (c) kerfed and treated posts in Kyushu, Japan; (d) small boxed heart kerfed timber beams in Kyushu, Japan (photos, Dr. Philip D. Evans) .....................................................53 Figure 2.27: Kerfing treatments applied to western hemlock square guard rail posts: (a) no treatment; (b) single kerf, 2” (50.80 mm) deep; (c) single kerf, 3 3/4“ (95.25 mm) deep; (d) double kerf, 2” (50.80 mm) deep, 1 ½” (38.10 mm) from center of posts on opposite sides (redrawn from Chandler 1968) ..................................................................................................54 xiv  Figure 2.28: Kerfing treatments applied to radiata and slash pine posts: (a) single kerf, 30 mm deep; (b) single kerf, 45 mm deep; (c) single kerf, 60 mm deep; (d) double kerfs, 15 mm deep; (e) double kerfs, 22 mm deep; (f) double kerfs, 30 mm deep (redrawn from Evans et al. 2000) 56 Figure 2.29: Examples of sewari: (a) Kerfed sewari style posts used in Japanese homes (photo, Dr. Hiroki Sakagami, Kyushu University); (b) kerfed sewari style posts used as a parking barrier in Japan (photo, Dr. Philip D. Evans) ..........................................................................................57 Figure 2.30: Images of 10 mm deep kerfs used by Ratu et al. (2007) in their study of the kerfing of flat-sawn southern pine deck-boards: (a) one kerf; (b) two kerfs; (c) three kerfs (photos, Dr. Philip D. Evans)..........................................................................................................................58 Figure 2.31: Images of: (a) Double headed toothed rollers used to incise wood; (b) lodgepole pine, butt-incised, transmission pole (photos, Dr. Philip D. Evans) ............................................59 Figure 2.32: Types of fasteners used in the studies of Evans et al. (2003) and Urban and Evans (2005): (a) Nails and screws affixed to radiata pine deck-boards; (b) screws and fasteners affixed to southern pine deck-boards (photos, Dr. Philip D. Evans) .......................................................63 Figure 2.33: Images of various mechanical fasteners designed to reduce the end-splitting of wood: (a) Circular nail plates used to reduce the end-splitting of a spotted gum (Corymbia maculata (Hook.) K.D. Hill and  L.A.S. Johnson) power poles prior to treatment with CCA; (b) Square nail plates used to reduce the end-splitting of Eucalyptus railway ties; (c) Steel ring used in Japan to reduce checking of a railway tie; (d) S-hook designed to reduce the end-splitting of a railway tie in South Africa (photos, Philip D. Evans) ...................................................................64 Figure 2.34: Images of beveled and rounded posts: (a) Rounded post; (b) four-sided beveled post; (c) one-sided beveled post; (d) two-sided (v-shaped) beveled post (Source: Popular Mechanics Press 1942) ..............................................................................................................67 Figure 2.35: White spruce boards weathered for 24 weeks in Vancouver: (a) untreated; (b) treated with PF resin. Scale bar = 50 mm ..................................................................................80 Figure 3.1: Randomized block design used for the experiment that examined the checking, discolouration and erosion of 17 different wood species ..........................................................87 Figure 3.2: Weathering racks used in the exposure trial: (a) Dimensions of an individual weathering rack; (b) Photograph of some of the deck-board samples on the four weathering racks during the exposure trial ..................................................................................................93 Figure 3.3: Extracted profile from a confocal profilometry line scan of an eroded white oak sample. Note the micro-checks (arrowed) in the sample ...........................................................94 Figure 3.4: Arrangement of specimens on SEM stubs ................................................................96 Figure 3.5: Images of: (a) western hemlock; (b) white oak; (c) Pacific silver fir; (d) Douglas fir boards showing severe surface checking. Scale bars = 50 mm ................................................. 100 xv  Figure 3.6: Images of: (a) lodgepole pine; (b) ipe boards showing moderate surface checking. Scale bars = 50 mm ................................................................................................................. 101 Figure 3.7: Images of: (a) aspen; (b) balsa; (c) western red cedar boards which developed small surface checks. Scale bars = 50 mm ......................................................................................... 102 Figure 3.8: Images of: (a) western red cedar; (b) white spruce; (c) western larch boards showing surface checks in the center of the boards. Scale bars = 50 mm .............................................. 103 Figure 3.9: Image of a lodgepole pine board showing surface checks around and within a knot. Scale bar = 20 mm ................................................................................................................... 104 Figure 3.10: Image of a red oak board showing checks that look like large multiseriate rays. Scale bar = 20 mm ................................................................................................................... 104 Figure 3.11: Total check number in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the square root of check number. The Y2 axis contains values on the natural scale (n). The LSD bar can be used to estimate whether differences between means for the various species are statistically significant ....................... 105 Figure 3.12: Average length of ten largest checks in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the natural logarithms of check length. The Y2 axis contain values on the natural scale (ex) ..................................................... 106 Figure 3.13: Average width of ten largest checks in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the natural logarithms of check width. The Y2 axis contain values on a natural scale (ex) .................................................................... 107 Figure 3.14: Total check number plotted against basic density for: (a) all wood species; (b) all hardwood species; (c) diffuse porous hardwood species. Y1-axis refers to the square root of check number ......................................................................................................................... 109 Figure 3.15: Check length plotted against basic density for: (a) all wood species; (b) all hardwood species; (c) diffuse porous hardwood species. Y1-axis refers to the natural logarithms of check length ........................................................................................................................ 110 Figure 3.16: Check width plotted against basic density for: (a) all wood species; (b) diffuse porous hardwood species. Y1-axis refers to the natural logarithms of check width ................. 111 Figure 3.17: Images of deck-board samples before and after 1 year of natural weathering: (a) balsa; (b) ipe; (c) southern pine; (d) western red cedar; (e) Douglas fir; (f) white oak. Scale bar = 50 mm..................................................................................................................................... 113 Figure 3.18: Lightness of species before and after natural weathering .................................... 114 Figure 3.19: Redness of species: (a) before; (b) after natural weathering ................................ 115 Figure 3.20: Yellowness of species before and after natural weathering ................................. 116 xvi  Figure 3.21: Colour change of species during natural weathering ............................................ 117 Figure 3.22: Change in the colour of test species during natural weathering estimated using Photoshop............................................................................................................................... 118 Figure 3.23: Images of weathered boards showing uneven discolouration of the surface of the boards: (a) Douglas fir; (b) basswood; (c) southern pine. Scale bar = 50 mm ........................... 119 Figure 3.24: Images of weathered boards showing black specks on the surface of the boards: (a) balsa; (b) red alder. Scale bar= 20 mm .................................................................................... 120 Figure 3.25: SEM images showing heavy fungal colonization of the surfaces of: (a) Douglas fir; (b) Pacific silver fir; (c) southern pine; (d) western larch samples exposed to natural weathering for 1 year ................................................................................................................................ 121 Figure 3.26: SEM images showing heavy fungal colonization of the: (a) lodgepole pine; (b) western hemlock; (c) aspen; (d) western larch samples exposed to natural weathering for 1 year ............................................................................................................................................... 122 Figure 3.27: SEM images of fungi developing in the cracks at the surface of weathered: (a) ipe; (b) white oak; (c) red oak; (d) maple samples exposed to natural weathering for 1 year ......... 123 Figure 3.28: SEM images of fungi which appeared to have different morphology at the surface of weathered wood: (a) red alder; (b) basswood; (c) Pacific silver fir; (d) lodgepole pine samples exposed to natural weathering for 1 year................................................................................ 124 Figure 3.29: Wasps stripping weathered wood from the surface of: (a) red alder; and (b) balsa deck-board samples (photos, Clement Tilloy) .......................................................................... 125 Figure 3.30: Images of deck-board surfaces affected by wasps: (a) basswood; (b) red alder; (c) aspen; (d) balsa. Scale bar = 20 mm ........................................................................................ 126 Figure 3.31: Images of deck-board surfaces showing black spots concentrated in areas that have been stripped of weathered wood by wasps: (a) basswood; (b) red alder. Scale bar = 5 mm... 126 Figure 3.32: Number of wasps visiting the surface of wooden deck-board samples................. 127 Figure 3.33: Average time (sec) that wasps spent on the surface of wood during their visits to weathered deck-board samples .............................................................................................. 127 Figure 3.34: Images of weathered: (a) white oak (Scale bar = 50 mm) and (b) southern pine (Scale bar = 10 mm) samples, which were largely free of damage caused by wasps ................ 128 Figure 3.35: Images of masked and unmasked areas of samples exposed to natural weathering for 1 year: (a) trembling aspen; (b) balsa; (c) western red cedar; (d) western larch; (e) Douglas fir; (f) white oak ...................................................................................................................... 129 Figure 3.36: Erosion of test species exposed to natural weathering for 1 year ........................ 130 xvii  Figure 3.37: Linear regression of basic density versus erosion of test species exposed to natural weathering for one year .......................................................................................................... 131 Figure 3.38: Average numbers of micro-checks in the eroded areas of different wood species exposed to natural weathering for 1 year................................................................................ 132 Figure 3.39: Average widths of micro-checks in the eroded areas of the different wood species exposed to natural weathering for 1 year................................................................................ 133 Figure 3.40: Graphical classification of softwoods into moderate to poor categories defining their suitability for use as decking ........................................................................................... 135 Figure 3.41: Graphical classification of hardwoods into moderate to poor categories defining their suitability for use as decking ........................................................................................... 136 Figure 3.42: Graphical classification of softwoods into good categories defining their suitability for use as decking.................................................................................................................... 137 Figure 3.43: Graphical classification of hardwoods into good categories defining their suitability for use as decking.................................................................................................................... 138 Figure 4.1: Different morphological elements of deck-board samples ..................................... 149 Figure 4.2: Altisurf 500 confocal profilometer used to characterize the geometry of profiles in commercial deck-board samples ............................................................................................. 150 Figure 4.3: Images of profile geometry obtained using confocal profilometry of deck-board samples: (a) height and width of peaks; (b) groove and peak radii .......................................... 151 Figure 4.4: Parameters of the profiles that were measured using image analysis .................... 151 Figure 4.5: Geometry of surface profiles in wooden decking manufactured commercially in eight different countries .................................................................................................................. 162 Figure 4.6: Appearance of commercial profiles: (a) Profile 6, a short wide rib; (b) profile 8, a moderate sized rib; (c) profile 12, a tall rib; (d) profile 10, a short, thin rib; (e) profile 11, a moderate sized rib with a flat groove; (f) profile 16, a wide rib; (g) profile 20, a short, thin ribble; (h) profile 22, a moderate sized ribble with a flat peak; (i) profile 21, a tall ribble; (j) profile 28, a short ripple; (k) profile 26, a moderate sized ripple; (l) profile 27, a thin ripple. Scale bar = 2 mm ............................................................................................................................................... 163 Figure 4.7: Appearance of commercial profiles with similar H/W, R1/R2 ratios but differing in appearance: (a) Profile 11; (b) profile 8; (c) profile 7. Scale bar = 2 mm ................................... 164 Figure 4.8: Appearance of commercial profiles with similar H/W, R1/R2 ratios but differing in appearance: (a) Profile 13; (b) profile 12. Scale bar = 2 mm .................................................... 164 xviii  Figure 5.1: Randomized block layout of samples during the weathering trial. Each weathering rack represents a block. Profiles are designated as: Rib (Rb); Ribble (Rbl); Ripple (Rp). Short and tall variation are denoted by a -/+ thin and wide variations are denoted by a –w/+w ............. 166 Figure 5.2: Geometry of designed profiles used to manufacture profiled decking from Pacific silver fir. The names of the different rib (Rb), ribble (Rbl) and ripple (Rp) profiles and precise R1/R2 and H/W ratios of the profiles are listed in Table 5.1. The numbers refers to the commercial profiles listed in Chapter 4 ................................................................................... 168 Figure 5.3: Appearance of the different rib, ribble, and ripple profiles in profiled Pacific silver fir decking. Note that the profiles were machined using material that was not tested in the weathering trial and hence growth ring orientations are a mix of those found in flat-sawn and rift sawn boards ...................................................................................................................... 169 Figure 5.4: Appearance of the narrow and wide rib profiles designed and used with Pacific silver fir ............................................................................................................................................ 169 Figure 5.5: Custom designed knives for rib (Rb), ribble (Rbl), and ripple (Rp) profiles. Short and tall variations are denoted by a -/+; thin and wide variations are denoted by a –w/+w ........... 170 Figure 5.6: Great-Loc SG Positive Clamping Universal Tool System: (a) in-feed diagram; and (b) a photograph of the cutter-head................................................................................................ 171 Figure 5.7: Steel fence and dial gauge micrometer used to measure the cupping of deck-board samples ................................................................................................................................... 172 Figure 5.8: Appearance of the profiled boards after they were weathered for 6 months (refer to Figure 5.1 for profile type) ...................................................................................................... 174 Figure 5.9: Average width of ten largest checks in profiled and flat (unprofiled) Pacific silver fir boards exposed outdoors to natural weathering for 6 months. The labels on the x-axis refer to the profiles listed in Table 5.1. Y1-axis refers to natural logarithms of check width. The Y2 axis contain values on a natural scale (ex) ...................................................................................... 176 Figure 5.10: Average length of ten largest checks in profiled and flat Pacific silver fir boards exposed outdoors to natural weathering for 6 months ........................................................... 177 Figure 5.11: Average area of ten largest checks in profiled and flat Pacific silver fir boards exposed outdoors to natural weathering for 6 months ........................................................... 177 Figure 5.12: Confocal profilometry images of the surface topography of profiled Pacific silver fir decking samples exposed to natural weathering for 6 months: (a) Rib sample showing checks within grooves, and a large check on the top of a peak (far left); (b) Wide rib sample showing checks within grooves; (c) Ripple sample showing large and small checks within grooves; (d) Short ribble sample showing large and small checks within grooves and two diagonal-checks that cross profile peaks (far right) ........................................................................................... 178 xix  Figure 5.13: Areas of large checks in profiled samples that developed in grooves (valleys), on peaks or diagonally in Pacific silver fir deck-board samples exposed to natural weathering for 6 months. Note that the short ribble profile (Rbl-) developed some peak checks but they were small (0.5 mm2) and do not appear on this figure .................................................................... 179 Figure 5.14: Close-up of peaks and grooves in profiled Pacific silver fir samples after 6 months of weathering. Note the development of checks at the base of grooves in each of the profiled specimens: (a) Standard rib (Rb); (b) Standard ribble (Rbl); (c) Standard ripple (Rp); (d) Standard ribble (Rbl). Scale bar = 1 mm .................................................................................................. 180 Figure 5.15: Geometry of profiles after 6 months of outdoor exposure compared to their initial geometry ................................................................................................................................ 181 Figure 5.16: Pacific silver fir boards exposed to the weather for 6 months: a) control (flat); b) rib; c) ribble; d) ripple. Scale bar = 50 mm ............................................................................... 182 Figure 5.17: Profiled Pacific silver fir boards exposed to the weather for 6 months: a) ripple; b) tall ripple; c) short ripple. Scale bar = 50 mm .......................................................................... 183 Figure 5.18: Profiled Pacific silver fir boards exposed to the weather for 6 months: a) rib; b) wide rib; c) thin rib. Scale bar = 50 mm ................................................................................... 183 Figure 5.19: Difference in cupping of Pacific silver fir boards before and after they were exposed to natural weathering for 6 months. Results for profiled samples are averaged across all boards and compared with cupping occurring in the flat (unprofiled) controls ................................... 184 Figure 5.20: Cross-section of a board with a rib profile after it was weathered for six months. Note the pronounced cupping of the board ............................................................................ 184           xx  Acknowledgments  To my supervisor, Professor Philip D. Evans, thank you for everything. Without your help I would not have had the success I had as a graduate student. During my time as your student you have guided me through my work, presentations and conferences. You have shown great patience and dedication to my experimental work and my writing. Most important of all you have always provided opportunities for me and my fellow students to broaden our exposure to the world of wood surface protection. I am truly grateful to have been your student.    I also express my gratitude to Dr. Paul Morris and Dr. Gregory Smith for being on my supervisory committee. Dr. Morris and his group at FPInnovations, especially Dave Minchin, have gone above and beyond, in helping and advising me. I would also like to thank Malcolm Knapp Research forest and FPInnovations for providing me with wood samples for my research and the technicians in CAWP who helped me process the samples. Finally thank you to everyone who helped me obtain or translate literature that I used to write my literature review.   I would like to thank the student award services at UBC for providing me with a grant to travel to Japan to attend the 11th Pacific Bio-Based Composites Symposium (BIOCOMP) in 2012. I was honored to win the best student poster award at this conference. I would also like to thank the American Wood Protection Association (AWPA) for awarding me the AWPA best student poster award at their Annual Meeting in Newport Beach in 2014. I would also like to thank the Canadian Wood Preservation Association (CWPA) for honoring me with the ‘Robert W. Stephens Memorial Award’ (for the best paper presented by a student) at the CWPA 35th Annual Meeting in Vancouver in 2014.   I am lucky to have spent my time as a graduate student at UBC with so many nice people from all around the world. George, Siti, Will, Vinicius, Yuner, Arash, Vincente, Jonathan, Barend, Chunling, and Stephan thank you for making my days at the lab lively and interesting. I extend similar thanks to the visiting students, scholars, and technicians in my lab group. Thank you Ian, Hiroki, Clément, Nicolas, Julian, Katharina, and Riina for helping me and allowing me to learn about your countries and cultures.   xxi  Dedication To my family and friends for their support To Ingrid who made the difficult times seem silly To my mom who is the foundation of my family             1  Chapter 1: General Introduction 1.1 Wood decking  Wood is a valuable resource that has helped civilization to evolve in many different ways. Wood is stiff, strong and inexpensive and it is still widely used in contemporary buildings. For example, 86.1% of the market for exterior decks attached to residential houses is still retained by solid wood (Figure 1.1) (Freedonia Group 2014a). In North America the decking market is predicted to grow by 2.4% per year for the period 2013 to 2018. The residential deck market in the US represents 60.3% of the total demand for decking material, while the non-residential deck market account for 31.3% followed by non-building construction decks (bridges, piers etc.) at 8.5% (Freedonia Group 2014a). The largest segment of the market for wooden deck-board materials is pressure treated southern pine (Pinus sp.) lumber followed by California redwood (Sequoia sempervirens (D. Don) Endl.) and cedar, mainly western red cedar (Thuja plicata D. Don) (Figure 1.1).   Figure 1.1: Market segments for different deck-board materials in 2013. Percentages are based on total annual consumption of 3065 million lineal feet of decking (redrawn from: Freedonia Group 2014a,b,c)  Figure 1.1 also shows that composite products such as wood plastic composites (WPC) are a popular alternative to wooden decking in North America (Freedonia Group 2014a,b,c). 2  Demand for WPC’s is expected to grow by 8.6% annually to 2018, while wood decking is only expected to grow by 1.1 percent during the same period (Eder 2013) (Freedonia Group 2014a). Plastic and aluminum decking only represent 2.9% of the market for decking but demand for them is expected to grow 11% annually to 2018. North America is the largest producer of WPC’s followed by China and Europe (Carus et al. 2014). However, China has seen the largest growth in production of WPC’s from 300,000 tonnes in 2010 to 900,000 tonnes in 2012 (Carus et al. 2014). Sixty-seven percent of the WPC’s manufactured in Europe are used for decking and most of the WPC’s that China produces are used for the same product (Carus et al. 2014). In Europe 260,000 tons of WPC’s was produced in 2012 (Carus et al. 2014). Germany is the largest market for WPC’s in Europe (Carus et al. 2014, Witten et al. 2014). WPC’s share of the deck-board market in France grew from 5% to 22% in the period 2005 to 2012 (Eder 2013). In Japan, the value of the market for WPC’s in 2005 was 300 million dollars (Gardner et al. 2008). Overall, the total world production of WPC’s grew from 1.515 million tonnes in 2010 to 2.43 million tonnes in 2012 (Carus et al. 2014).   The emergence of WPC’s as a serious competitor for wood in the deck-board market is quite recent and needs explaining. Deck-boards are an appearance-based product rather than a structural commodity. However, when wooden deck-boards are used outdoors their appearance changes; their surfaces become discoloured and develop checks (Feist 1983) even though the board itself usually remains structurally sound. Consumers often replace deck-boards for aesthetic reasons (Bolin and Smith 2010).  Conversely, WPC decking checks, warps and discolours less than wooden decking when used outdoors and, as a result, it requires less maintenance although it is not maintenance free (Figure 1.2). For these reasons consumers are switching to WPC decking in preference to treated wood decking (Ganguly and Eastin 2009, Freedonia Group 2014a,b,c). Manufacturers of WPC decking have adjusted the surface texture of their products so that they more closely resemble solid wood (Majewski and Rajaraman 2013, Freedonia Group 2014c). This gives WPC a more natural wooden look while retaining its weather resistant properties. WPC decking is also available in a range of different colours and its resistance to weathering has been improved by creating products whose upper surfaces consist solely of plastic (Freedonia Group 2014c).  3   Figure 1.2 shows the advantage wood plastic composite decking has over competing solid wood products in terms of ease of maintenance, durability, and decay resistance (Ganguly and Eastin 2009). It seems likely that there will be further erosion in the market share for wooden decking unless steps are taken to improve the resistance of wooden decking to weathering or reduce its maintenance requirements.   Figure 1.2: Composite score of decking material attributes perceived by home and deck builders. Categories are calculated on a 7-point Likert-like scale based on consumers’ perceptions of decking materials. Higher values on the Likert-like scale means the product is viewed more favorably by the respondents (redrawn from Ganguly and Eastin 2009)     4  1.2 Surface checking    Surface checking of wooden decking, the lengthwise separation of the wood along the grain, is highly disliked by consumers. Currently no method is completely effective at preventing the surface checking of wooden deck-boards exposed outdoors. Chemical preservatives when they contain hydrophobic emulsion additives (oil/wax) restrict checking of deck-boards, but they do not eliminate checking completely, particularly in species that are more prone to checking, for example, southern pine (Zahora 1992, 2000, Cui and Zahora 2000).   Some naturally durable tropical woods are viewed favourably by consumers for their aesthetics, but there is little information on their susceptibility to checking when used as deck-boards (Figure 1.2) (Ganguly and Eastin 2009). A little more information is available on the checking of naturally durable softwoods such as western red cedar and California redwood (Yata 2001, Ratu 2009, Morris and Ingram 2002). They check significantly less than other softwoods particularly when quarter-sawn, which explains, in part, their continuing popularity with consumers (Figure 1.3) (Ganguly and Eastin 2009).    Figure 1.3: Images of deck-boards exposed at a test site in Charlotte, North Carolina: (a) Quarter-sawn California redwood; (b) Flat-sawn California redwood (photos, Dr. Alan Preston)  The supply of such naturally durable softwoods and hardwoods is limited and there is a need to examine the use of other species as deck-boards to determine their susceptibility to checking, as well as finding ways of reducing checking. One way of reducing checking is to profile deck-boards. Surface profiling creates a series of V (rib) or U (ripple) shaped grooves on the surface of deck-boards (McFarling and Morris 2005, 2008). Evans et al. (2010) examined the (a) (b) 5  ability of profiles to reduce the checking of Pacific silver fir (Abies amabilis Douglas ex J. Forbe) and southern pine. They found that the effectiveness of profiling at reducing checking depended on the type of profile and wood species. For example, a hemispherical rib profile restricted checking of Pacific silver fir deck-boards exposed to accelerated weathering better than it did with southern pine. Conversely, a wavy ripple profile was slightly better at reducing checking of southern pine than a rib profile. Their findings indicate that profile type or geometry influences the checking of profiled wooden decking. 1.3 Aim, hypothesis and significance  The aims of this thesis are twofold: First, to obtain a better understanding of the checking and weathering of different wood species that could be used as decking; second, to examine how the geometry of surface profiling influences the checking of Pacific silver fir, a species that has been used commercially to manufacture profiled decking in Canada. I hypothesize that both species and profile geometry will significantly influence checking. I anticipate that the results of my work could lead to the development of improved wooden decking, either by identifying under-utilized species that are less susceptible to checking and weathering than those that are currently used as deck-boards, and also by improving the ability of profiling to reduce the checking of Pacific silver fir. Therein lies the significance of my research. 1.4 Study outline  This chapter (Chapter 1) provides background information and reasoning (general hypothesis) for this thesis. Chapter 2 provides additional background information on check formation, causes of checking, species differences, and the methods that have been used to reduce checking. These methods include physical methods (profiling, incising, mechanical fixings etc.), chemical treatments, and coatings. Chapter 3 describes experiments performed to examine the susceptibility of different wood species to check, discolour, and erode when exposed to natural weathering. Chapter 4 describes a preliminary experiment that characterized the geometry of deck-board profiles that are used commercially in several 6  different countries. Chapter 5 describes an experiment that examined the effects of profile geometry on the checking and cupping of Pacific silver fir. Chapter 6 discusses the extent to which results from the different experimental chapters support my general hypothesis described above, draws conclusions, and makes suggestions for further research. 7  Chapter 2: Literature Review 2.1 Introduction  Wood exposed outdoors to the weather is subjected to a complex process of environmental degradation caused by light, water, heat, pollutants, and surface staining fungi (Feist 1990). Such degradation is termed weathering. One of the effects of weathering is surface checking of wood (McMillen 1955, Evans 2008). Surface checking can ruin the appearance of wood and it is associated with a variety of other problems, most notably strength losses, erosion and premature fungal colonization of wood (Figure 2.1).      Figure 2.1: Images of various defects caused by checking: (a) fungal fruiting bodies growing within checks in a test deck in Tsukuba, Japan; (b) severe check in a western red cedar power pole in Vancouver, Canada; (c) erosion of a 19th century Norfolk Island pine (Araucaria heterophylla (Salisb.) Franco) roofing shingle (photos, Dr. Philip D. Evans)  This chapter will review the literature on the checking caused by weathering and drying and different methods that have been used to reduce this defect.    (a)                                                (b)                                                 (c) 8  2.2 Checking  Checking, along with cupping and warping, is a physical defect of wood that occurs due to dimensional change and associated surface stresses that develop when wood is exposed outdoors to the weather or drying (Schniewind 1963, Feist and Mraz 1978). Like wrinkles in an aging person, checking is visually unattractive and it can also lead to other problems, as mentioned above, such as fungal decay and strength losses (Heebink 1959, Stanzl-Tschegg et al. 1996, Choi et al. 2001, 2003, 2004, Kurisaki 2004, Morris et al. 2004). For example, Choi et al. (2003) mentioned that fungal colonization of chromated-copper arsenate (CCA) deck-boards can occur in three ways: (1) through the treated surface; (2) through cut ends of boards; or (3) through checks that develop at the surface of boards. They found that checks were the easiest route for fungi to colonize treated boards because the checks trapped water and by-passed the boards’ treated shell. In addition, small checks that form during weathering or drying can develop into bigger cracks that can reduce the strength of load bearing structures (Heebink 1959, Stanzl-Tschegg et al. 1996). Checks can compromise strength even further if they occur at critical points around fasteners (Heebink 1959). Checks also allow moisture to penetrate deeper into the board further reducing stiffness and strength (Forest Products Laboratory (FPL) 1957). 2.2.1 Definitions  There are many terms used to describe the physical failure of wood caused by unbalanced surface stresses. The main ones are checks, cracks, and splits, and there are also many sub-categories of checks. This section tries to clarify and define these terms.  Lamb (1992) described splits and cracks as wood failures that occur as ruptures or separations of the wood grain. Conners (2008) separated checks and splits as follows: (1) Checks: Found on only one surface of the wood; (2) Splits: Found on two surfaces of the wood.  The literature defines a check as a lengthwise separation of the wood along the fiber (Bucur 2011, National Hardwood Lumber Association (NHLA) 2011, Yang and Normand 2012). Côté (1968) cited by Yang and Normand (2012) defined checking as “longitudinal openings at weak points in the wood” caused by shrinking stresses. Checks can be so small that they can 9  only be seen with a microscope, or they can be large enough to be easily seen with the naked eye (Bucur 2011).  The dimensions of a check or a split include width, depth, length, volume and various shape parameters (Evans et al. 1997). The widths of checks are highly correlated with their depth, i.e. the wider a check becomes; the deeper it penetrates into boards (Yata 2001). Bucur (2011) categorized checks by their position in the wood, their shape and how they are caused, as follows: a) End checks: checks occurring at the ends of boards. b) Heart check/pith check: a check that starts from the pith and grows across growth rings towards the surface without reaching the surface.  c) Roller check: checks caused by cupped wood being straightened by mechanical rollers. d) Star check: a heart check that occurs in more than one direction. e) Surface check: a check occurring at the wood surface that usually doesn’t extend through the entire board. f) Through-check: also known as a split is a check that runs all the way through the surface of the wood or from another surface to the opposite surface or an adjoining surface.  Yang and Normand (2012) described other types of checks and provided additional information on them: a) End checks: cracks at the ends of logs and lumber during storage, seasoning, or kiln drying (Figure 2.2). These checks are caused by rapid moisture loss at the ends of the boards. Log end-checks appear in the heartwood if the bark is kept on or in the sapwood if the bark is removed. End checks also develop in wide boards and can develop into splits or through checks especially if the board is thin.  b) Surface checks: as described by Bucur (2011) are checks of various sizes that occur at the surface of a board. These checks form when the surface of a board is dried faster than the center, as explained below (Section 2.2.2). Surface checks are also known as face checks, and often develop within rays.  10  c) Internal checks: checks that occur within boards. They are common in boards with wet pockets or in wood that collapses during drying or heat treatment. Internal checks are also known as honeycomb checks. They cannot be seen unless the board is cut into sections and can develop into splits if they coalesce with surface checks.  d) Micro-checks: checks that are hidden to the naked eye. These checks include surface checks that close up during drying, and checks that develop in cell walls when wood is exposed to the weather. If surface checks are exposed to cyclical wetting and drying, they can open up and develop into visible checks.  e) Mechanical splits: defined as a “lengthwise separation of the wood due to the tearing of the wood cells” (NHLA 2011). These are associated with mechanical damage and can cause surface checks, internal checks, wounds, and cell collapse to form around a split.  f) Shakes: longitudinal separation of standing trees caused by growth stresses or wind damage.   Figure 2.2: Different types of end checks at the end of a Japanese cedar (Cryptomeria japonica (L.f.) D.Don) log: (a) star check; (b) end check; (c) shakes (photo, Dr. Philip D. Evans)     (a) (b) (c) 11   The terms surface checks and cracks are sometimes used interchangeably. Flaete (2000) distinguished them using the depth to which they penetrate wood. Cracks were distinguished from surface checks if they penetrated more than 75% of the board’s thickness. The most common type of failure for lumber exposed to weathering is a surface check, followed by end-checks, internal checks and, rarely, cracks and splits (Figure 2.3).   Figure 2.3: Diagram showing different types of physical failures of wood caused by surface stresses            12  2.2.2 Mechanism (why checks form)  Wood changes dimensions by shrinking and swelling when it gains or loses moisture water according to prevailing environmental conditions (outdoors, in-service, or during air or kiln drying) (Eckelman 1998). The moisture content wood attains either indoors or outdoors is called the equilibrium moisture content (EMC), and is mainly affected by the temperature and the relative humidity of air surrounding the wood (Eckelman 1998). For every 1% of moisture loss or gain wood will shrink and swell 1/30th of its total shrinkage and swelling value (FPL 1957).  These dimensional changes cause stresses to develop in the wood (Eckelman 1998). These stresses are reduced when wood distorts or checks (Stamm 1965). Shrinkage anisotropy between the radial and tangential directions also contributes to the stresses that lead to the formation of checks (McIntosh 1955, Schniewind 1959, 1960).   More specifically checking of lumber is caused by moisture differences between the surface and the center of the board during seasoning or drying. During the later stages of drying, checks close up and become micro-checks making it difficult to see the extent of checking. These micro-checks can open-up during service (Yang and Normand 2012). The focus of this thesis is on surface checking of lumber, specifically deck-boards, exposed to natural weathering. However, the mechanisms responsible for checking of wood during kiln drying have received more attention than those responsible for the checking of wood exposed to the weather, although there are similarities between the two forms of checking.      13   Figure 2.4 (a-e) shows how a flat-sawn board develops checks during drying (McMillen 1955). The right hand column of the same figure (f-j) shows how a similar board checks when it is exposed to natural weathering (Evans 2008).  Figure 2.4: Mechanism of check formation in an unrestrained flat-sawn board during drying (a-e) and weathering (f-j). Brown= dry, blue = wet (redrawn from: McMillen 1955, Evans 2008)  First, looking at the checking of wood during kiln drying, the entire surface of the board is exposed to the kiln’s heat, and moisture is removed from the face and the ends of the board (Figure 2.4a). As the board dries the surface of the board tries to shrink and warp, but it is restrained from doing so by the relatively wet (green) wood, which creates a tension stress at the surface, while the core develops compression stresses to counteract surface forces (Figure 2.4b) (McMillen 1955). The flat-sawn board tries to cup upwards on the convex (bark) 14  side because the tangential side shrinks 1.5 to 2.5 times more than the radial side during kiln drying (Skaar 1972, Spear and Walker 2006). If the perpendicular tensile stress exceeds the tensile strength of the board, checks, splits, and other types of deformation will develop to relieve the tensile stress (Figure 2.4c) (McMillen 1955, Stamm 1964). These “defects” are usually more severe when the surface of the board is dried too fast and it shrinks much faster than the core (FPL 1957).   The size and occurrence of the checks that develop in boards during drying depend to some extent on the growth ring orientation of the board. The bark-side (convex) surface may develop larger checks initially than the pith-side (concave) because the board is trying to cup to the bark-side (convex) surface, but the wet core prevents the surface from cupping, thus creating large tensile stresses (Figure 2.4c). As drying continues the dry surface zone continues to grow towards the center of the board and the core starts to drop below the fiber saturation point (Figure 2.4d) (McMillen 1955). The wood at the surface may become ‘set’ in its dimensions or, in other words, case hardened (McMillen 1955). Without the wet core restricting surface movement, the wood will start to cup upwards towards the bark-side (convex) surface and the checks located on the surface will start to close up and become narrower (Figure 2.4d). As a result, the checks on the pith-side (concave) surface will be subjected to more tensile stress and may become wider (Figure 2.4d).  The core will continue to shrink and dry and this time the dried surface will restrain the core from shrinking which causes tension in the core, while the surface undergoes compression, which could lead to internal (honeycomb) checking (Figure 2.4e) (McMillen 1955).  Checks that develop during the drying stage may be micro-checks that can reopen during weathering (Mackay 1973). However, Perem (1971) mentioned that the correlation between checking that developed during seasoning and checking during service was low for sugar maple (A. saccharum Marshall) railway ties. He found that only 7 to 14% of checks that developed in service were formed from ones that were present in wooden ties after they were seasoned (Perem 1971).    15   Wood exposed outdoors is subjected to cyclical swelling and shrinking due to exposure to solar radiation and water (Evans et al. 2008). Figure 2.4 (f-j) modifies the mechanism used to describe the checking of wood during kiln drying to explain the checking of deck-boards exposed to the weather. There are however some differences between deck-boards exposed to the weather and dimension lumber undergoing kiln drying as follows: 1. Fasteners such as decking screws restrict distortion of boards from relieving stresses (Evans 2008) 2. One side of the board is heavily exposed to the weather and wetting and drying, whereas the underside, if it is poorly ventilated, may remain above the fiber saturation point (Evans 2008) 3. The board is exposed to cyclical wetting and drying at its upper surface (Sell and Wälchli 1969, Evans 2008)   When rain water is absorbed by the upper surface of a deck-board it will swell, but swelling will be restrained by fasteners to some extent (Figure 2.4f) (Evans 2008). Some of the swelling of the surface will also be restrained by the relatively drier core, which causes compression of the surface and crushing of fibers (FPL 1957). Exposure to the sun removes moisture, and the top surface starts to dry faster than the sub-surface (Figure 2.4g) (Evans 2008). The temperatures at wood surfaces exposed outdoors can be as high as 80◦C (Wengert 1966, Sell and Wälchli 1969). Shrinkage at the surface will be restrained by the wetter sub-surface (Figure 2.4h) and also the fasteners that fix the boards to the sub-frame (Figure 2.4i) (McMillen 1955, Evans 2008).  As a result, tension stresses build up (Figure 2.4j) (McMillen 1955, Stamm 1964, Evans 2008). Crushed fibers would tend to expand, but may also be restrained from doing so (FPL 1957), which would create both compression and tension forces. When the tension stress exceeds the tensile strength of the wood, surface checks will develop to reduce the tension (Stamm 1964, Evans 2008).     16  2.3 Factors affecting checking  The checking of wood develops from degradation and micro-checking of wood’s cellular elements. Therefore in this section, I examine how wood’s anatomy influence micro-checking and checking. I then examine how checking can be influenced by wood species, grain orientation, wood quality, wood defects, and environmental factors.  2.3.1 Wood microstructure  Softwoods mainly consist of vertical and horizontal tissues called longitudinal tracheids (~90%) and ray cells (~10%), respectively (Hoadley 1990). Longitudinal tracheids are elongated vertically along the direction of tree growth and are associated with other longitudinal cells including, longitudinal parenchyma, and vertical resin canals. Rays consist of ray parenchyma, ray tracheids and horizontal resin canals (Hoadley 1990). Pits or openings are found in the walls of most of these different cell types. The pits allow the transfer of fluids between cells (Hoadley 1990). These pits include simple, bordered and half-bordered pits (Hoadley 1990). Hardwoods, like softwoods, also contain vertical and horizontal tissues. However, their longitudinal cells consist of vessel elements/pores, longitudinal tracheids, longitudinal parenchyma, and fibers (Hoadley 1990). Rays in hardwoods mainly consist of parenchyma cells (Hoadley 1990).   Growth rings are layers of wood cells added around the cambium of the tree as it grows (Hoadley 1990). These layers may consist of a first formed portion (earlywood) and late formed portion (latewood) (Hoadley 1990). Cells in the earlywood and latewood have different dimensions and properties. For example, earlywood cells are often bigger with larger lumens and thinner walls while latewood cells are often smaller with thicker walls (Hoadley 1990). As a result, latewood in softwoods is generally denser than earlywood.  The vertical and radial arrangement of cells in wood and the occurrence of growth rings create a material that is anisotropic. Three different planes are recognized in wood: transverse, radial, and tangential. The transverse plane or cross-section is the area where the plane is perpendicular to the stem, wood’s grain and hence longitudinal cells (Figure 2.5) (Hoadley 1990). The radial plane is parallel to the stem and passes through the pith (Figure 2.5) (Hoadley 17  1990). This plane is perpendicular to the ends of the growth rings. The tangential plane is also parallel to the stem like the radial plane, but it does not pass through the pith and runs tangentially or parallel to the growth rings (Figure 2.5) (Hoadley 1990). Most deck-boards are commonly exposed with the tangential surface facing uppermost.   Figure 2.5: Surfaces of a profiled radiata pine (Pinus radiata D. Don) deck-board: R=Radial longitudinal surface; T=Tangential longitudinal surface; X=transverse surface  High tensile stresses develop in-between tracheids and wood rays when wood dries (Schniewind 1963). Furthermore, during drying and weathering, checks develop in or around the rays (Schniewind 1963). For example, Evans (1989) exposed small samples of radiata pine to the weather for up to two years and found that macroscopic checks formed from micro-checks, which developed as a result of degradation of thin walled parenchyma cells in rays. Evans et al. (2008) examined the checking of lodgepole pine exposed outdoors. They found that micro-checks developed near the rays and grew between tracheids as a result of separation of the middle lamella. These micro-checks coalesced with other micro-checks to create checks that were visible on the surface of deck-boards (Evans et al. 2008). Evans et al. (2008) also found that micro-checks that developed from larger rays (fusiform rays) that contained a resin canal 18  were easier to see than the checks that developed in uniseriate rays. Ribarits and Evans (2010) modelled the tensile strains that develop during wetting and drying. Their model suggested that tensile strains caused surface checks to develop along the rays in flat-sawn boards. These surface checks were able to combine with internal checks to create cracks (Ribarits and Evans 2010).   Hale (1957) mentioned that differences in the shrinkage of earlywood and latewood also influence checking. For example, transverse and tangential wood surfaces exposed outdoors often check at the boundaries between earlywood and latewood (Evans 1989, Sandberg and Söderstorm 2006).  The different shrinkage of earlywood and latewood as well as the thin walls of earlywood contribute to the formation of such checks (Evans 1989). ‘Wood structure’ can influence checking in other ways. For example, Coupe and Watson (1967) exposed Douglas fir (Pseudotsuga menziesii Franco), radiata pine, Scots pine (Pinus sylvestris L.), and western red cedar to artificial weathering. They observed that most of the checking in softwoods was in the latewood due to the development of differential stresses between the ray and non-ray tissues and the presence of vertical resin ducts in the latewood of Douglas fir, radiata pine, and Scots pine. In accord with this observation, Mackay’s (1973) study of CCA treated radiata pine found that the position of the vertical resin canals either on the surface or near the surface influenced checks in tangential cell walls. Fujita (1970) found that checks start to develop in the latewood between the rays and tracheid’s cell walls or between the tracheid’s cell walls in the transverse surface.   Openings in materials concentrate stresses and are often the sites where crack initiation occurs (Illston et al. 1979). For example, Miniutti (1964) found micro-checks developed in the orifices of bordered pits in the walls of earlywood and latewood tracheids. Such pit micro-checks are one of the first signs of degradation of unfinished wood during weathering (Miniutti 1964, Evans et al. 1994). Evans (1989) also found that micro-checks in radiata pine exposed to natural weathering formed at an angle to the S2 layers of the secondary wall i.e., orientated at an angle to the fiber direction. Minutti (1964) found that latewood tracheids tended to split in tangential sections. Such splits were absent from earlywood. Evans (1989) found that micro-19  checks in radiata pine were more likely to form in half-bordered pits due to their elliptical apertures rather than in bordered pits, which have circular apertures.  2.3.2 Wood species  Certain species are less susceptible to weathering and checking than others but the precise reasons for their superior behavior are unknown, although species that are dimensionally stable tend to check less than ones that are less stable. MacLean (1956) mentioned that if hardwoods were dried under the same conditions as softwoods with similar properties, they are likely to develop more pronounced checks than the softwoods.  Sell and Leukens (1971) exposed twenty species of hardwoods and softwoods of various densities to 1 year of natural weathering and found that some species checked more than others, but eventually all of the species had the same grey weathered appearance. Softwoods rather than hardwoods are preferred for exterior poles in North America because of hardwoods’ excessive checking, according to Kretschmann (2010). However, eucalyptus species are commonly used for poles in Australia where they perform well (Francis and Norton 2006).   Some softwoods however, are more susceptible to weathering and checking than hardwoods (Roux et al. 1988, Flaete et al. 2000, Mitsui and Tsuchikawa 2005, Nejad et al. 2013). Mitsui and Tsuchikawa (2005) found that the softwood Japanese cypress (Chamaecyparis obtuse Sieb. et Zucc.) photodegraded faster than Japanese beech (Fagus crenata Blume) when exposed to artificial sunlight. Flaete et al. (2000) weathered common aspen (Populus tremula L.) and Norway spruce (Picea abies (L.) Karst.) in an accelerated weatherometer for 112 days and found that the woods checked differently. Aspen developed larger numbers of shorter checks while Norway spruce had less checks, but they were larger and longer. Roux et al. (1988) applied various types of coatings to European beech (Fagus sylvatica L.), Norway spruce, Scots pine, Douglas fir, and dark red meranti (Shorea spp.) panels and exposed them at four different sites in Europe. They found that untreated beech cracked and discoloured badly after 3 months of weathering. Pine and spruce developed numerous checks and cracks after 6 months, and Douglas fir developed cracks in the growth rings after 9-12 months. The appearance of untreated meranti was largely unaffected by 12 months of exterior exposure. 20  Nejad et al. (2013) found that American beech (Fagus grandifolia Ehrh.) cupped and checked less than eastern hemlock (Tsuga canadensis L.) after they were heat-treated in an oil bath at temperatures of 180◦C for three hours.  2.3.2.1 Softwoods  Research on the checking of softwoods has focused on western red cedar, Pacific silver fir, western hemlock, Douglas fir, Scots pine and southern pine (Coupe and Watson 1967, McCarthy et al. 1982, Morris and Ingram 1996, 2002). Johnson and Bradner (1932) mentioned that western larch (Larix occidentalis Nutt.) was more susceptible to checking than Douglas fir and southern pine, which are both comparable in density to western larch.  As a result, they suggested that western larch was less suitable than Douglas fir or southern pine for use as pole cross-arms. Coupe and Watson (1967) found most softwoods did not develop pit micro-checks whereas, western red cedar was susceptible to this form of checking due to the morphology of its cross-field pitting. McCarthy et al. (1982) exposed radiata pine, mountain ash (Eucalyptus regnans F. Muell), Scots pine, Douglas fir, western red cedar and merbau (Intsia sp.) panels to the weather. They found that untreated, western red cedar showed the least amount of checking and cupping after the 24 month trial while untreated radiata pine showed the most. Sandberg (1997, 1999) and Söderstorm and Sandberg (2006) found that Scots pine boards developed longer checks than Norway spruce when boards of both species were exposed to either natural weathering or wetting and drying cycles.   Relatively few studies have examined the checking of softwood decking. Morris and Ingram (1996) exposed Pacific silver fir and western hemlock (Tsuga heterophylla Raf. Sarg.) deck-boards to 30 months of natural weathering. Both species were either untreated or incised and CCA treated. They found that untreated and treated Pacific silver fir performed better than western hemlock both in terms of appearance and check depth. Yata (2001) measured check lengths in wooden benches (Japanese cypress), deck-boards (Japanese cedar), rock climbs (western hemlock), and table tops (California redwood) exposed to natural weathering.  Redwood and Japanese cypress had longer checks than Japanese cedar and western hemlock.  21  Morris and Ingram (2002) exposed softwood decking samples cut from 11 different species at two different sites: Vancouver and Ottawa. Boards were either untreated, unincised and treated with CCA, or incised and treated with CCA. Boards were exposed to the weather for 9 years after which they were examined for surface checking and general appearance (Figure 2.6).   Figure 2.6: Visual assessment of severity of checking in deck-boards exposed outside in Vancouver or Ottawa for 9 years. WH: Western hemlock; WS: Western white spruce (Picea glauca (Moench) Voss); LP: Lodgepole pine (Pinus contorta var. latifolia Douglas); AF: Subalpine fir (Abies lasiocarpa (Hooker) Nutt.); ES: Eastern white spruce (Picea glauca (Moench) Voss); JP: Jack pine (Pinus banksiana Lamb); BF: Balsam fir (Abies balsamea (L.) Mill.); RP: Red Pine (Pinus resinosa Sol. ex Aiton); PP: Ponderosa pine (Pinus ponderosa Douglas ex C.Lawson); SP: Southern pine; WRC: Western red cedar. For cities: VAN: Vancouver; OTT: Ottawa. Note: Check appearance is rated from 0=no change to 4=severe checking. Red pine was not tested in Ottawa (redrawn from Morris and Ingram 2002)    After nine years of exposure most of the untreated boards, with the exception of western red cedar and subalpine fir (Abies lasiocarpa (Hook.) Nutt.), had decayed and needed to be replaced (Morris and Ingram 2002). Western red cedar checked and cupped less at both sites than subalpine fir, which developed moderate checking (Morris and Ingram 2002). Untreated ponderosa pine (Pinus ponderosa Douglas ex C.Lawson) had lower check ratings in Ottawa, while southern pine and western hemlock had poor check ratings at both sites (Morris and Ingram 2002) (Figure 2.6). The relationship between cupping, checking, and check depth of boards at both the Ottawa and Vancouver sites was examined, but there was little influence of cupping on checking (including appearance and depth of checks). Ratu (2009) exposed southern pine and western red cedar deck-boards in a weatherometer designed to accelerate surface checking. He found that checks in southern pine were larger than those in western red cedar. Checks in southern pine developed in the center of the boards, and became wider and longer 22  with increasing exposure. Furthermore, the center checks were usually the largest ones in the boards (Ratu 2009). Conversely, there was no clear pattern to the spatial arrangement of checks in western red cedar (Ratu 2009).  2.3.2.2 Hardwoods  Research on the checking of hardwoods has mainly focused on oak, beech, and ipe (Coupe and Watson 1967, Williams et al. 2001). Coupe and Watson (1967) exposed three hardwood species: European beech, opepe (Sarcocephalus diderrichii De Wild. and T.Durand), and English oak (Quercus robur L.) to artificial weathering in addition to the softwoods mentioned above. They classified checks into three types: a) longitudinal checks that were close to the middle lamella between the walls of different elements; b) longitudinal checks in element walls and; c) diagonal checks through pits aligned with microfibrils. They found that pattern of checks in hardwoods was different. Longitudinal checking in the walls (Type b) occurred in the large vessels of oak and opepe, which may have Type a influenced checking. In accord with this observation, Mackay (1972) noted that drying checks in messmate stringybark (Eucalyptus obliqua L’Herit) started within the vessels. Coupe and Watson (1967) found that some of the checking they observed at tangential surfaces formed in rays. For example, oak and beech both developed checks in their multi-seriate rays. Watson also noted that two of the hardwoods species in the study, European beech and opepe, were known to split badly (Coupe and Watson 1967).  There have been some studies of checking of hardwoods during drying which is a problem with certain species. Arnold et al. (1950) noted that red oak ties developed more checks than beech and hard maple ties during seasoning. Balfas (1994) applied four types of end-coats to several cat statues made from albizia (Paraserianthes falcataria (L.) Nielsen.) and chinaberry (Melia azedarach L.), two species that are susceptible to cracking. The checking of coated statues was compared with that of uncoated controls. He found that for both untreated and treated samples, albizia had higher moisture gradients and greater checking than chinaberry possibly because it was lighter and dried faster than chinaberry. 23   Hardwoods are commonly used as railway ties in North America but many of them are prone to checking and splitting. For example, Burt (1955) mentioned that 48.8% of 30,000 oak ties installed in-track in Illinois were removed from service due to splits and checks. Perem (1971) also mentioned that sugar maple (Acer saccharum Marshall) and beech (Fagus sp.) railway ties checked more than birch (Betula sp. L.) ties during service. Furthermore, the checks in beech were wider than those in maple (Perem 1971). He suggested that the reason for the lower susceptibility of birch to checking was due to its low tangential to radial shrinkage ratio, which resulted in low seasoning stresses. Even though there was no record of why the aforementioned ties were removed from service, Perem (1971) noted that the checking performance of these species was correlated with their failure in service. For example, only 18.7% of the birch ties were replaced compared to 35% of the maple ties and 73% of the beech ties.  There has not been a lot of research on the weathering and checking of hardwoods, other than that mentioned above. However, Kishino and Nakano (2004) weathered eight different hardwood panels for up to 600 hours in a weatherometer and noted that the severity of checking depended on wood species, and their structure. Keruing (Dipterocarpus sp.) had the largest checks while cumaru (Amburana acreana (Ducke) A.C.Smith) and swamp mahogany (Eucalyptus robusta Sm.) had the least checking after exposure. Kishino and Nakano (2004) also examined the checking of ipe (Tabebuia sp.) samples with high (1.14 g/cm3) and low (0.90 g/cm3) densities. The former developed slightly smaller checks than the lower density samples. Izekor and Fuwape (2010) exposed Burmese teak (Tectona grandis L.f.) outdoors for 12 months and concluded that teak is resistant to both end splitting and face checking. Williams et al. (2001) exposed panels of ten different tropical species to accelerated weathering for 2400 hours in a xenon arc weatherometer and rated the species in terms of the degree of checking that developed in the panels (Figure 2.7).     24   Figure 2.7: Checking of ten tropical species exposed to accelerated weathering for 2400 hours. BT: Burmese teak; PT: Plantation teak (Tectona grandis L.f.); CRP: Curupau (Anadenanthera macrocarpa (Benth.) Brenan); JIQ: Jichituriqui (Aspidosperma cylindrocarpon Müll.Arg.); IPE: Ipe; SCP: Sucupira (Diplotropis purpurea (Rich.) Amshoff); STO: Soto (Schinopsis quebracho-colorado (Schldl.) F.Barkley and T. Meyer); CUC: Cuchi (Astronium urundeuva (Fr. and All.) Engl.); ROB: Roble (Amburana cearensis (Allemão) A.C.Sm. ); MOMO: Momoqui (Caesalpinia cf. pluviosa DC); SIR: Sirari (Guibourtia chodatiana (Hassl.) J. Léonard); CTA: Cuta (Phyllostylon rhamnoides (Poisson)). Check appearance is rated on a scale of 1 to 10 where 1=severe change and 10=no change (redrawn from: Williams et al. 2001)  Williams et al. (2001) compared the checking of the different tropical hardwood species with those of naturally grown and plantation-grown teak. None of the ten species performed as well as teak, but Curupau (CRP), Jichituriqui (JIQ), Soto (STO), and ipe warped and checked less than some of the other species (Figure 2.7). Check appearance was plotted against density, warp, and swelling for the different species in this study (Williams et al. 2001) (Figure 2.8). There was a positive correlation between warping and checking (Figure 2.8a,b) and a negative correlation between checking and volumetric swelling (Figure 2.8c). In other words, boards with more warping and less swelling checked less.  25    Figure 2.8: Graphs showing the relationship between check appearance and: (a) density; (b) warping; (c) swelling of different tropical wood species exposed to accelerated weathering for 2400 hours. Check appearance is graphed at a scale of 1 to 10 where 1=severe change and 10=no change (redrawn from: Williams et al. 2001)  Iimura (2001) suggested using timbers such as bongossi (Lophira alata Banks ex Gaertn), jarrah (Eucalyptus marginata Donn ex Sm.), and ipe as the top layer for glue-laminated timber to reduce checking of decks and hand rails. Opoku (2007) exposed seven hardwood cross-pieces to outside weathering for 6 months and found that English oak had the longest and greatest number of checks. He also exposed five untreated Ghanaian hardwood deck-boards to the weather for six months and then examined the checking of the boards. Denya (Cylicodiscus gabunensis Harms) and Esa (Celtis mildbraedii Engl.) boards checked whereas the Iroko (Milicia excels (Welw.) C.C. Berg), African Mahogany (Khaya ivorensis A. Chev.), and Dahoma (Piptadeniastrum africanum (Hook.f.) Brenan) boards were free of checks.     (a)                                                    (b)                                                    (c)  26  2.3.3 Growth ring orientation  The orientation of growth rings to the surface of boards that are exposed to the weather influences checking. This effect occurs as a result of the anisotropy of wood and the different shrinkage and swelling of wood in the tangential and radial directions. 2.3.3.1 Radial, tangential, rift and double rift   Sawing wood into boards produces different growth ring orientations at sawn surfaces. Flat-sawn boards have rings that are oriented parallel to the widest sawn surface (Figure 2.9a). Quarter-sawn or edge grain boards have growth rings that are perpendicular to the widest sawn surface (Figure 2.9b), whereas rift-sawn boards have a mix or growth ring orientations (Figure 2.9c). Double rift boards consist of a tangential section as well as two radial sections. However, growth rings in the radial parts of double rift-sawn boards are not completely perpendicular to the surface (Figure 2.9d).  Figure 2.9: Orientation of growth rings at sawn wood surfaces: (a) Flat-sawn board; (b) Quarter-sawn board; (c) Rift-sawn board; (d) Double rift-sawn board (redrawn from: Tenorio and Moya 2011)  Several studies have shown that quarter-sawn boards check less during outdoor exposure than flat-sawn boards (Schniewind 1963, Sandberg 1997, 1999, Flaete et al. 2000, Yata 2001). Browne (1952) mentioned that quarter-sawn boards of the same species tend to check or crack less than their flat-sawn counterparts. Stamm (1965) mentioned that flat-sawn wood with broad latewood bands tend to check and crack more at exposed surfaces usually at the junction between earlywood and latewood. This is because latewood, which has a high specific gravity, shrinks and swells three times more than earlywood (Stamm 1965).   (a) (b) (c) (d) 27   Sandberg (1999) compared the checking of quarter-sawn (radial) and flat-sawn (tangential) Scots pine and Norway spruce boards after they were exposed outdoors for 33 months. He found that flat-sawn boards had more cracks and the cracks were wider and longer than those in quarter-sawn boards. He explained the greater checking in the former boards as being due to the higher internal stresses caused by both anisotropic and surface-to-sub-surface moisture gradients in the wood. Earlywood and latewood ‘move’ to different degrees when their moisture content changes (latewood shrinks and swells more than earlywood).  In the radial direction the earlywood and latewood layers are arranged in a series so the layers can shrink and swell independently of each other. But in tangential flat-sawn boards the earlywood and latewood are stacked on top of each other so movement of layers is restrained by other layers, which cause stresses to develop (Sandberg 1999). Sandberg (1999) also observed that at tangential surfaces, checks occur first in the latewood and then move towards the surface of the board while at radial surfaces cracks occur at the growth ring boundaries and move towards the earlywood. Flaete et al. (2000) briefly mentioned that part of the reason that sawn boards containing pith check less than boards cut from near the bark is because pith sawn boards contain a great proportion of wood with radial orientation of growth rings. Sandberg and Söderstorm (2006) found that flat-sawn Norway spruce and Scots pine boards exposed to the weather for five years had deeper and more checks than similarly exposed quarter-sawn boards.   There are very few studies of the checking of rift-sawn boards, but one study by Tenorio and Moya (2011) commented on the subject. They found that black wattle (Acacia mangium Willd.) boards with rift-sawn and double rift-sawn orientations had more wet pockets after drying. The wet pockets caused uneven moisture gradients and led to defects such as warping, splitting, and checking when the wood adjusted to outdoor weather conditions. However, Yata (2001) found that western red cedar deck-boards with diagonal orientation of growth rings, similar to the rift-sawn orientation, developed fewer and smaller checks than either flat-sawn or quarter-sawn boards.  28  2.3.3.2 Concave and convex growth ring orientations   Flat-sawn boards can have their growth rings oriented either concave or convex to the face that is exposed upper-most, unless growth rings are parallel to the wide faces (Figure 2.10, also see Section 2.2.2). Another way of expressing these growth ring orientations is to refer to them as being bark-side-up (convex) or pith-side-up (concave).  Figure 2.10: Scots pine boards oriented: (a) Bark-side-up (Convex); (b) Pith-side-up (Concave)  Several studies have found that boards oriented pith-side-up check less when they are exposed to the weather than bark-side-up boards (Yata 2001, Urban and Evans 2005, Sandberg 1996, 1997, 2005, Sandberg and Söderstorm 2006). Yata (2001) examined the surfaces of naturally weathered Japanese cypress, western red cedar, Japanese cedar, and western hemlock wood products. He found that for all species, with the exception of western red cedar, checks were noticeably shorter in boards oriented pith-side-up compared to boards oriented bark-side-up. Sandberg and Söderstorm (2006) exposed Norway spruce and Scots pine boards with the two different growth ring orientations to the weather for five years. They found that boards with pith-side-up (concave) arrangement of growth rings had shorter checks than boards with bark-side-up (convex) orientation of growth rings, but the difference in checking was not statistically significant. Conversely, Perem (1971) reported that sugar maple ties installed bark-side-up and exposed to the weather for 6 years developed smaller cracks than ties installed pith-side-up, but he forecasted that the difference would become less obvious with time.  Urban and Evans (2005) investigated the checking of southern pine deck-boards cut from close to the pith and close to the bark and oriented pith or bark-side-up. The boards were fixed to deck sub-frames using three different types of fasteners and exposed outdoors for 6 months. They found that boards oriented pith-side-up developed fewer and smaller checks (a)                                                                          (b) 29  than boards oriented bark-side-up. Morris and McFarling (2008) exposed profiled and unprofiled Pacific silver fir and lodgepole pine deck-boards to the weather for one year. They found that profiled boards had less checking when the boards were oriented pith-side-up as opposed to bark-side-up. However, the difference was not statistically significant for Pacific silver fir. They did not recommend orienting southern pine boards pith-side-up due to a concern about “shelling”, a separation between growth rings causing severe raised grain (Figure 2.11) (Morris and McFarling 2008, Williams and Knaebe 1995).           Figure 2.11: Shelling of a profiled western larch deck-board oriented pith-side-up (photo, Dr. Philip D. Evans) 2.3.4 Wood quality (juvenile v mature wood/density and grain angle etc.)  Apart from wood species and growth ring orientation, there are other wood properties and factors that affect the checking of wood. 2.3.4.1 Density  An early report by the U.S. Forest Products Laboratory in Madison, Wisconsin stated that higher density wood species shrink more across the grain than lighter ones and this causes defects such as checks to develop (FPL 1957). Browne (1952) mentioned that low to medium density woods usually develop fewer and less visible checks than higher density woods. Chafe (1994) found a strong positive correlation between density and internal checking of mountain ash during drying. The correlation wasn’t as strong for surface checking because such checks tended to close up after drying (Chafe 1994). Ratu (2009) found that checking of southern pine 30  exposed to accelerated weathering was more severe than in western red cedar possibly because southern pine is denser than western red cedar. However, Balfas (1994) found that statues made from lower density albizia wood developed larger drying checks than statues made from denser chinaberry wood. Furthermore, Sandberg (1999) observed that density did not affect the checking of pine and spruce samples weathered outdoors for 33 months. 2.3.4.2 The pith juvenile/mature wood  There is a relationship between the presence of pith in wood and the occurrence of juvenile wood because juvenile wood is formed when a tree is young and is located adjacent to the pith. Hence the zone that contains the pith is called the juvenile zone, but there is no distinct border between the juvenile zone and mature wood, according to Makovická and Čunderík (2006). Younger trees and trees grown in plantation usually contain more juvenile wood as a proportion of the trees’ total cross-section. There are studies that have looked at the checking of wood in relation to its position relative to the pith as well as the checking of boards containing pith. Flaete et al. (2000) mentioned that higher grain angles in juvenile wood might encourage the formation of checks. However, he also added that the change of grain direction from spiral to straight at the interface between juvenile and mature wood might prevent checks from deepening and growing larger.   Perem (1971) mentioned that untreated sugar maple railway ties with pith that was closer to the surface tended to develop severe checking during seasoning. However, the opposite effect was observed in creosote-treated birch, maple, and beech railway ties that had been in service for 31 to 32 years. For example, treated ties were less likely to split if the pith was between 25.4 to 50.8 mm from the broad face and the ties were installed bark-side-up (convex orientation) (Perem 1971).  Sandberg (1996, 1997) mentioned that juvenile wood influenced the number of cracks that developed in Scots pine and Norway spruce boards during drying. Both growth ring orientation and the presence of pith influenced checking. Wood sawn from near the pith or containing the pith usually developed longer checks during drying than boards sawn away from 31  the pith (Sandberg 1996, 1997, 2005). Morris and Ingram (1996) mentioned that deck-boards that contained box heart or pith had more cross-checking and more twist. In accord with this observation, Kirker et al. (2012) exposed nine different wood species to outdoor weathering for up to one year and found that some of the southern catalpa (Catalpa bignonioides Walter) boards with juvenile wood showed signs of cross-checking. Izekor and Fuwape (2010) compared checking of juvenile and mature Burmese teak wood and found that the latter was more resistant to surface checking than juvenile wood.   In contrast to the aforementioned finding, Flaete et al. (2000) found that for both common aspen and Norway spruce, boards cut from near the bark (mature wood) checked more than boards cut close to the pith. They mentioned that juvenile wood had a smaller influence on checking compared to growth ring orientation. Juvenile wood is the first wood to be converted into heartwood and the presence of juvenile heartwood and mature sapwood complicate experiments that try to compare the checking of juvenile and mature wood.  Opoku (2007) exposed denya decking-boards outdoors for six months and found that ‘sapwood’ boards had larger and more checks than heartwood boards. Urban and Evans (2005) found that southern pine deck-boards cut from close to the pith did not check more than deck-boards cut from mature wood when the boards were exposed to the weather for 6 months. However, boards cut from juvenile wood contained various amounts of heartwood, which may have reduced their susceptibility to checking because the heartwood of pine is more dimensionally stable than its sapwood (Urban and Evans 2005). 2.3.4.3 Grain angle  Spiral grain occurs when fibers in wood deviate from the longitudinal axis of a tree or the long sides of a piece of lumber. Spiral grain can influence checking (Koehler 1955). When the grain orientation of wood is not parallel to a log’s pith, diagonal splits or cracks develop at the log surface (Tiemann 1941). The sawing of logs with star checks and spiral grain can result in cut boards that have checks or cracks that run diagonally along the surface of the boards instead of along the grain (Harris 1989). Spiral grain is a problem with profiled boards because checks cross the surface of the profiles (Evans et al. 2010). Spiral grain can also cause distortion 32  in the form of bowing because the wood shrinks more along the fiber than across (Sandberg 2005).   Interlocked grain occurs where the grain changes and reverses its direction in adjacent growth rings or zones of wood.  Wood with interlocked grain is difficult to split.  Interlocked grain is common in some dense hardwoods such as lignum vitae (Guaiacum officinale L.) (Tiemann 1941). Harris (1989) mentioned that boards with interlocked grain resist checking even better than quarter-sawn boards. 2.3.5 Wood defects (knots, compression wood etc.)   Defects such as reaction wood and knots can influence checking. Tension wood and compression wood are forms of reaction wood that develop in hardwoods and softwoods, respectively (Chow 1947, Core et al. 1961). Tension wood and compression wood are formed in leaning stems and both exert forces that cause stems to regain their vertical orientation (Figure 2.12) (Chow 1947).     Figure 2.12: Yellow cedar (Cupressus nootkatensis D. Don) containing compression wood: (a) longitudinal section containing darker coloured compression wood. Note the presence of compression wood around a knot (arrow); (b) transverse surface containing compression wood, note the darker bands of latewood in the growth rings of compression wood (arrow). Scale Bar = 20 mm (photos, Dr. Philip D. Evans)  Compression wood has a higher lignin content, lower cellulose content and is denser and more brittle than ‘normal’ wood (Côté et al. 1967). Wood containing compression wood (a)                                    (b)                                                33  tends to warp and split more during seasoning due to differences in longitudinal shrinkage between the compression wood and normal wood (Figure 2.12) (Core et al. 1961). The location and amount of compression wood in a board has an effect on how the wood warps and checks (Timell 1986). For example, if the compression wood is surrounded by normal wood, unbalanced strains develop which cause the wood to crack across the grain (Timell 1986). In particular the high longitudinal shrinkage of compression wood is thought to be responsible for the cross cracks that develop at the tangential and radial longitudinal surfaces of compression wood when it dries (Figure 2.13) (Core et al. 1961).   Figure 2.13: Cross cracks at the tangential surface of a hard pine (Pinus sp.) containing compression wood. Scale Bar = 5 mm   Bodner et al. (1997) mentioned that cross cracks in compression wood usually occur in the resin ducts and rays. Compression wood is characterized by spiral checking in its tracheids, which are oriented at 30◦ to 50◦ to the fiber axis (Core et al. 1961, Bodner et al. 1997). Tension wood is also prone to checking. For example, Clark (1958) found that drying checks in eastern cottonwood (Populus deltoids W.Bartram ex Humphry Marshall) were larger in wood containing tension wood than in wood free of such wood. 34   One of the most common ‘defects’ in wooden deck-boards are knots. Knots are dense tissues in wood where secondary branches were once located (Folvik and Sandland 2005). Two of the more common types of knots are intergrown knots, where the knot’s wood is contiguous with the adjacent stem wood, and encased knots, where the knots are encased with bark making them loose and easy to remove from the wood (Figure 2.14) (Hoadley 1990).   Figure 2.14: Images of scanned knots from off-cuts of wood used in Chapter 3: (a) intergrown knot in lodgepole pine; (b) loose knot in Alaskan yellow cedar. Scale Bar = 20 mm  Knots are particularly prone to checking especially during drying and planing (Boutelje 1966, Folvik and Sandland 2005). Differences in the transverse and longitudinal shrinkage of knots and the adjacent stemwood, and also the presence of reaction wood encourage the formation of checks in knots (Boutelje 1966). Leavengood and Swan (1999) recommended choosing boards for drying lumber with knots that were less than 12.7 mm in diameter to avoid excessive cracking of boards.   Folvik and Sandland (2005) examined how different types of knots affected the checking of Norway spruce boards.  They dried wood containing encased knots and wood containing intergrown knots. They found that encased knots checked more than intergrown knots. The high radial shrinkage of knots compared to the longitudinal shrinkage of the boards explained their propensity to check, according to Folvik and Sandland (2005). In addition, the higher density of knots and their lower equilibrium moisture content compared to those of the surrounding wood also made them susceptible to checking (Folvik and Sandland 2005). Folvik and Sandland (2003) also mentioned that knots in Scots pine were less susceptible to checking that those in Norway spruce because the knots in the latter species shrank less than those in Scots pine (Boutelje 1966, Folvik and Sandland 2003).  (a)                                                   (b)               35   Knots also influence the checking of adjacent stemwood. Boutelje (1966) mentioned that knots can influence the grain angle of the surrounding stemwood which encourages the formation of checks, as mentioned above (Section 2.3.4.3). Sandberg (2005) mentioned that knots are associated with distortions of boards, which can increase checking.  2.3.6 Weathering  There have been several reviews on the weathering of wood including recent ones by Evans (2008, 2013). I do not repeat this information here, but instead focus on the studies that describe the effects of weathering on the checking of wood.  Solar radiation, particularly UV light, causes surface degradation of wood, which increases the severity of checking that develops when wood is exposed outdoors. In a study of the effects of end-coating on end checking, Osborne (1965) noted that wood facing the sun checked more than wood that was in shade. Yata and Tamura (1995) placed filters over samples of Japanese cypress and western hemlock boards and found that checks developed in the rays when boards were exposed to wavelengths of less than 500 nm. The effect of solar radiation on checking was examined experimentally by Evans and Urban (2007) and Evans et al. (2008). They exposed lodgepole deck-boards for 12 weeks under four different filters that blocked different wavelengths of UV and visible light (UVB (260-345nm), UVA (260-400nm), UV/visible (260-760nm) as well as a filter that blocked all of the wavelengths and a filter that transmitted all of them. They found that boards under the filter that transmitted all of the radiation developed three times as many checks compared to the control under the filter that blocked all radiation.  In addition, boards under the filters that blocked highly energetic UVB radiation or UVA or visible light developed less checks than boards under the filter that transmitted the full solar spectrum. However, there was little difference in the checking of boards under the filter that blocked both UV and visible light and boards under the filter that blocked all radiation. This suggests that infrared radiation had less of an effect on surface checking than UV or visible light (Evans and Urban 2007, Evans et al. 2008). Ratu (2009) incorporated UV light irradiation into one of his wetting and drying cycles and found that wooden deck-board samples exposed to the 36  cycle with UV light developed larger and more numerous checks than cycles that simply exposed wood to wetting and drying.  Apart from the effects of solar radiation on checking and the critical role that wetting and drying play in generating the stresses that cause checking, other factors may also influence checking. For example, exposure of wood to sub-zero temperatures causes water in wood to freeze which leads to shrinkage and cracking according to Coupe and Watson (1967). However, Ratu (2009) found that a freezing step in accelerated weathering had no significant effect on the checking of wooden deck-board exposed to wetting and drying. Conversely, increased wood moisture contents decrease the strength of the wood, which makes it more susceptible to cracking, according to Bariska et al. (1988)  Thermal stresses can occur in wood even when moisture-induced stresses are absent according to Bariska et al. (1988). Temperature fluctuation between day and night also influence moisture gradients within wood. For example, a temperature difference of 4◦C can cause cracks to develop (Bariska et al. 1988). Wood exposed outdoors in coastal areas or partially submerged in seawater can develop micro-checking as a result of the effects of salt on wood cell walls (Johnson et al. 1992). Salt crystals within cell walls causes wood to swell at lower relative humidities (Stamm 1964).  This effect and subsequent shrinkage create larger stresses in wood, which increases micro-checking (FPL 1957).   Melanized fungi such as Aureobasidium pullulans (de Bary) Arnaud colonize wood exposed outdoors and are responsible for the grey colour of weathered wood. These fungi colonize rays and may damage wood’s microstructure, but their effects on the checking of wood is unknown (Seifert 1964, Schoeman and Dickinson 1997, Doi and Horisawa 2001).   37  2.4 Attempts to reduce checking  Even though there are many factors that influence the checking of wood there are also many methods that have been developed to reduce checking. These methods could be as simple as selecting a species that is dimensionally stable and less susceptible to checking, as well as using physical methods to restrain checking or chemical treatments to dimensionally stabilize wood or make it water repellent. Some of these methods are employed during drying, but they can also be applied to wood that is exposed outdoors to the weather.  2.4.1 Wood selection   As mentioned above (Sections 2.3.1 to 2.3.5) careful selection of species can reduce the checking of wood products. Selecting suitable species for use outdoors is a simple and widely used method of ensuring that checking does not greatly affect the service life and performance of wood products. This approach to reducing checking is examined in detail in Chapter 3. 2.4.2 Center-boring  Some simple machining techniques can reduce checking, for example center boring, which involves drilling a hole through the center of poles or posts. Liese (1961) center bored 22 mm diameter and 15 cm deep holes in the ends of European beech logs in an attempt to reduce end-checking. However, the treatment was ineffective and the number of checks that developed at the ends of the treated logs was the same as that in the control (2.4 checks per area for treated v. 2.4 untreated). Mater (1972) developed a machine to center bore Douglas fir and southern pine poles. He found that center bored poles developed less surface checking when they were exposed outdoors. Graham (1973) also suggested that checking of poles might be prevented by drilling longitudinal holes through their centers. Goodell and Pendlebury (1991) found that center bored 2.5 meter long untreated red spruce (Picea rubens Sarg.) poles developed shallower and narrower checks than regular poles when they were exposed outdoors for 1 year. However, center boring had little effect on the checking of CCA-C treated poles (Goodell and Pendlebury 1991). Evans et al. (1997) bored a 32 mm diameter hole along the entire length of 100 mm diameter slash pine (Pinus elliottii Engelm.) posts prior to 38  preservative treatment with CCA. Center boring reduced the area of checks that developed in posts exposed to the weather for one year, but reductions in check length were not statistically significant. In a follow-up study Evans et al. (2000) drilled different diameter (center-bored) holes (25 mm, 35 mm, and 45 mm) in small and large-sized slash and radiata pine posts prior to treatment of the posts with alkaline copper quaternary (ACQ) (Figure 2.15). The posts were exposed outdoors to natural weathering. They found that increasing the diameter of the center bored holes decreased the width, but not the depth or number of checks that were present in the posts after 1 year of natural weathering. Recently, Yeo (2007) found that center boring greatly reduced the length and width of surface checks that developed in square and round Japanese larch (Larix leptolepis Gordon) posts during kiln drying, but it was more effective at reducing checking in the square posts.     Figure 2.15: Mixture of 125 mm and 150 mm diameter ACQ treated slash and radiata pine posts center-bored with 25 mm, 35 mm, and 45 mm diameter holes: (a) after treatment; (b) after several years of exposure (photos, Dr. Philip D. Evans) 2.3.3 Planing and machining  Sawn wood has been suggested as being more suitable than planed wood for outdoor use because the rough surface hides visible surface checks (Woodhead 1969). On the other hand, planing before kiln drying was suggested as a means of reducing checking during drying (a)                                                 (b)               39  because it removes the irregularities at the wood surface that are the sites where checks form (Leney 1964). For example, Leney (1964) observed differences in checking between rough and planed red oak (Quercus rubra L.) boards following kiln drying. He randomly planed one side of the boards and counted the number of checks per inch that developed after boards were kiln dried. He also classified the checks into the following depth classes: (1) 1 to 3 mm deep; (2) 3 to 5 mm deep; (3) 5 to 7 mm deep; and (4) 7 to 9 mm deep.  He found that within all of these categories, rough surfaced boards checked more than the planed ones. The total number of checks per inch (25.4 mm) was reduced by ~3.6 times by planing the lumber prior to kiln drying (Leney 1964). Similarly, Gaby (1963) investigated the effects of green planing, dull circular headsaw cutting, sharp circular headsaw cutting, and sharp vertical band resaw cutting on the checking of kiln dried white oak (Quercus alba L.). The treatments were applied to the bark-side surface of several 1 inch flat-sawn boards. Green planed boards checked less than the other ‘surfaced’ boards. For example, when compared with circular-saw cut boards, green planing reduced the number of surface checks per square foot (0.093 m2) by 93%. The effect of depth of green planing on checking was also tested by Gaby (1963). He found a small positive correlation between planing depth and reduction of checking, but the result was not significant (Gaby 1963). Levi et al. (1970) found that planing CCA water repellent treated softwood boards reduced their tendency to crack. Mackay (1972) also examined the effects of surfacing on the checking of kiln dried wood. He dried several test blocks of rough sawn and planed messmate stringybark and found that planing was better than sawing at reducing checking. He explained his findings by suggesting that sawing created fractures that led to increased checking.   Stehr and Östlund (2000) examined whether tip cracks (micro-checks occurring perpendicular to the growth rings) caused by machining encouraged surface checks to develop at wood surfaces during weathering. His test substrate was Scots pine veneer glued on top of a laminated pine board (with a spruce veneer on the other side). His machining treatments were bottom planing, top rotary planing and bottom planing, abrasive planing, sawing and sanding.  He found that on the pith-side of boards more tip cracks developed after the two planing treatments compared to sawing. He also observed more tip cracks at lower grain angles with 40  planing compared with sawing.  He suggested that a gentle planing should be applied to the pith-side of boards to reduce tip cracks (Stehr and Östlund 2000).  2.4.3 Profiling   Surface profiling is the machining of a series of u or v shaped grooves on the surface of wood (Evans et al. 2010). Profiling has been shown to be able to hide and reduce the checking that develops in deck-boards exposed to the weather, as mentioned in Chapter 1. 2.4.3.1 History of surface grooving  Early examples of grooving can be found in linen-fold paneling, a form of wainscot paneling. These grooves are in panels that were carved by hand to imitate folds in linen (Figure 2.16a) (Woodcarvers Guild 2003).  Another early example of grooving used a rotary cutter to create v-shaped tongue and grooves in lumber to improve the adhesion of lumber panels during the manufacture of composites (Figure 2.16b,c) (Brock 1878). Rotary cutting was later used to machine wood along the grain to produce profiled decorative panels for railway carriages and furniture (Figure 2.16e,d) (Mankey 1884).         41                                                                                        Figure 2.16: Early examples of surface grooving: (a) Hand carved linen-fold paneling; (b) decorative profiled wood panel (top view); (c) decorative wood panel (side view); (d) cross-section of v-shaped grooved wood used to make compound lumber; (e) cross-section of notched grooved wood used to make compound lumber (Sources: Brock 1878, Mankey 1884, and Woodcarvers Guild 2003)        (a)                                                           (b) (c) (d)                                                               (e) (c) 42   Irregular grooving was also applied to exterior shakes and shingles to simulate the rough surface of hand-split shakes (Figure 2.17) (Gilmer 1933, Gilmer 1934, Gilmer 1936, Norlander and Knowles 1970).  Figure 2.17: Grooved and handsplit western red cedar shakes: (a) Irregular grooving; (b) hand split  Surface grooving was first used to reduce the checking of plywood. Grooved plywood was patented by Donald Deskey and William Bailey, manufactured by the United States Plywood Corporation and called balanced striated plywood (Figure 2.18) (Deskey 1942, Bailey 1944a,b). Grooving was originally designed to enable the panel to be bent to create curves and contours (Deskey 1942). Grooving ran longitudinally along the grain and was created using irregular-profiled knives (Deskey 1942). The profile at the surface of the panel consisted of ridges and grooves of different depths, some of which almost penetrated the top veneer sheet and reached the glue line (Bailey 1944a,b). However, the spacing of the grooves was quite uniform (Bailey 1944 a,b). The grooved surface was used as the outside/visible side of the panel because of its ability to break-up bending stresses at the surface (Deskey 1942). Deskey (1942) mentioned that grooving reduced face checking and was able to hide checks along the grooves.  Increased cupping of striated plywood, however, was a concern, Bailey (1944a, b) reduced such cupping by increasing the thickness of the veneer on the opposite side of the panel to create a ‘balanced striated panel’. (a) (b) 43                   Figure 2.18: Visualizing the contours of striated plywood: (a) Balanced striated plywood panel. Scale bar = 100 mm; (b) confocal profilometry image of part of the balanced striated plywood shown in (a); (c) extracted profile line scan from (b)  Striated plywood was designed in the early 1940’s, but its popularity took off in the 1950’s especially for prefabricated housing (Ottoson 2009). Striated plywood was sold by the United States Plywood (USP) Corporation under the name “Weldtex” (Ottoson 2009). Weldtex was used as paneling for the inside of prefabricated houses and also for outdoor cladding (Figure 2.19) (Ottoson 2009). (a) (b) (c)  44    Figure 2.19: Use of Weldtex striated panels: (a) Weldtex used as square cladding on a house in Port Coquitlam during the 1950s (photo, Dan Price); (b) Weldtex used as interior paneling (Eichler Siding. http://eichlersiding.com/)  Weldtex became very popular and a similar product was manufactured by Georgia Pacific in 1955 (Georgia-Pacific Corporation v U.S. Plywood-Champion Papers Inc. 1971, Ottoson 2009). Georgia Pacific sold striated plywood using the Weldtex brand and as a result USP Corporation sued Georgia-Pacific for patent infringement. Initially a judge ruled in favour of Georgia-Pacific and deemed Deskey’s three patents invalid (Deskey 1942, Bailey 1944a,b, Ottoson 2009). Subsequently, this decision was overturned, and in 1958 USP Corporation won its case, on the basis that Weldtex was already an established product and the fact that striating plywood was a novel process to reduce surface checking by creating “grooves” at the surface of plywood (Ottoson 2009). Georgia-Pacific appealed against the ruling and the compensation that Georgia-Pacific had to pay USP Corporation was reduced (Georgia-Pacific Corporation v U.S. Plywood-Champion Inc. 1971, Ottoson 2009). However, by this time USP had stopped producing Weldtex because of declining sales, which followed their merger with Champion Papers Inc. (Ottoson 2009). When Weldtex was being manufactured, research was carried out to examine the properties it imparted to plywood. For example, Browne (1952) investigated the effects of the striations on paint performance. He found that the surface grooves improved paint performance by reducing the area of the ‘difficult-to-treat’ latewood (a)                                                   (b)     45  bands and allowing boards to absorb more paint (Browne 1952). On the other hand he found that dirt accumulated between the grooves of striated boards (Browne 1952).   Elmendorf (1950) developed plywood with more uniform patterned grooving, spaced 3.175 mm apart. Such profiling also reduced surface checking of plywood (Figure 2.20) (Elmendorf 1950, Elmendorf and Vaughan 1960). According to Elmendorf (1950) the grooves allowed the surface of the boards to freely shrink and swell thus reducing surface checking. Elmendorf et al. (1960) exposed profiled Douglas fir plywood to five years of outdoor weathering and found that the face checks were thinner at the surface of profiled panels than those at the surface of the unprofiled Douglas fir controls (Elmendorf and Vaughan 1960). Striated or profiled plywood disappeared from the market after USP stopped production, but recently another U.S. company Eichler Siding has started to manufacture it (Eichler Siding 2015).    Figure 2.20: Diagrams of Elmendorf’s profiled Douglas fir plywood: (a) top view; (b) side view (Source: Elmendorf 1950)     (a) (b) 46   Surface profiles are also applied to interior flooring and one of the earliest patents for wood plastic composites describes a reeded (profiled) tongue and groove flooring board (Figure 2.21) (Roy 1955). Contemporary wood plastic decking is sometimes manufactured with a profiled upper surface (Figure 2.21).   Figure 2.21: Wood plastic composites with profiled surfaces: (a) Diagram of an early wood plastic composite with a reeded surface (Source: Roy 1955); (b) Contemporary wood plastic composite with a profiled surface (Ecowood. http://ecowood.co.za/)             (a)                                                                              (b) 47  2.4.3.2 Surface profiling of deck-boards  Surface profiling was originally applied to deck-boards to encourage drainage of standing water from deck-board surfaces, making them less slippery (Hislop 2006). Different profiles have been developed with this end in mind. The grooves of some of these profiles can contain inorganic or metal inserts to further reduce the slipperiness of the surface (Figure 2.22) (Hill and Moss 2000, Hislop 2006, GripDeck 2015).      Figure 2.22: Examples of decking products with anti-slip inserts: (a) GripDeck board (Source: GripDeck 2015); (b) Grooved board with RetroGrip antislip deck metal inserts (Southern Timber. http://www.southern-timber.co.uk/); (c) Deck-board with an anti-slip insert consisting of a mixture of dried silica and calcined flint with two shallow surface drainage grooves (Source: Hill and Moss 2000); (d) Another type of anti-slip insert made from a mixture of dried silica and calcined flint. Note the absence of ‘surface drainage grooves’ (Source: Hill and Moss 2000)     (a)                                                                              (b) (c)                                                                               (d) 48   There is very little information in the scientific literature on anti-slip decking, with the exception of the work of Shida and colleagues in Japan (Shida et al. 1992). They tested the coefficient of slip resistance of 11 floating pier deck-boards. Bongossi deck-boards were grooved with two different types of rectangular grooves and compared with flat controls (Figure 2.23).  Figure 2.23: Bongossi deck-boards grooved with: (a) thin rectangular grooves; (b) wide rectangular grooves (Source: Shida et al. 1992). Scale bar = 100 mm   When the surface of the grooved boards was dry the slipperiness of the three types of boards was the same, but when the surface was wet both types of grooved boards were less slippery than the control (Shida et al. 1992). In addition, Shida et al. (1992) also tested the slipperiness of western hemlock containing trapezoidal shaped grooves. They found that boards containing trapezoidal grooves were less slippery in both the dry and wet states than unprofiled flat boards.        (a) (b) 49   Profiling of deck-boards is common and a great variety of profiled deck-boards are sold at retail outlets. The profiles that are commonly machined into the surface of deck-boards have been called rib and ripple profiles. Rib profiles have a hemispherical shape while ripple profiles have a wavy or dentate shape (Evans et al. 2010). The ability of profiling to reduce the checking of deck-boards was first examined by Morris and co-workers. Their findings have been published in a series of reports, conference papers and journal articles and are summarized in Table 2.1. Table 2.1: Summary of previous studies by Morris and co-workers that have tested the ability of surface profiling to reduce the checking of deck-boards Reference Profile Type*  Species Summary McFarling and Morris (2005) McFarling et al. (2009) Rib Subalpine fir:  (600-800 x 131 x 26 mm)   Rib profile was used with CA (copper azole), ACQ-D and CCA preservatives. Profiling reduced check depth and check length for both treated and untreated boards after 5, 9, 17, and 23 months of weathering. Overall appearance was improved. Checks were constrained to the grooves Morris and McFarling (2008) Rib and Ripple Pacific silver fir and Lodgepole pine:  (815 x139.7 x 31.75 mm)   Four profiles types were applied to treated boards with various grain orientations (concave and convex) and exposed outdoors for one year. All of the profiles were effective at reducing checks. Some profile types were better suited to Pacific silver fir or lodgepole pine. Evans et al. (2010) Rib and Ripple Pacific silver fir and Southern pine:  (400 x 133 x 23 mm) Rib and ripple profiles were machined into boards which were exposed to accelerated weathering. Rib profile was more effective at reducing check width in Pacific silver fir than in southern pine. Checks were longer in profiled boards. Spiral grain reduced the effectiveness of profiling.  *please note that later in the thesis that some of these profile types are redefined as being ribble profiles as a result of experimental work described in Chapter 4. 50   McFarling and Morris (2005) machined subalpine fir deck-boards to create a rib profile (McFarling and Morris 2005, McFarling et al. 2009). Profiled boards were treated with CA, ACQ-D, and CCA, air dried and exposed to the weather (Table 2.1). The rib profiles reduced check length, width, and depth, and also concealed checks at weathered deck surfaces. Morris and McFarling (2008) subsequently tested the ability of four different profiles to reduce checking of flat-sawn Pacific silver fir and lodgepole pine deck-boards (Figure 2.24). Each profile was applied to the surface of boards with concave and convex growth ring orientations and boards were treated with one of three preservatives: (1) ACQ-D; (2) Carbon-based preservative; and (3) Carbon-based preservative with a 0.87% transparent iron oxide blend.  All of the profiles reduced the checking of deck-boards made from both species. Morris and McFarlings’ (2008) results are shown in Appendix 1.   Figure 2.24: Rib and ripple profiles used by Morris and McFarling in their study of the checking of Pacific silver fir and lodgepole pine deck-boards exposed to natural weathering: (a) ripple-flat edge; (b) ripple to edge; (c) rib-eased edge; (d) rib-flat edge and center (Source: Morris and McFarling 2008). Note footnote to Table 2.1.  Lodgepole pine deck-boards with the ripple flat edge configuration had shallower and shorter checks (Appendix 1). Pacific silver fir deck-boards with the rib-eased edge configuration had shorter checks (Appendix 1). Boards without the flat areas in the center had a better appearance and check rating because most of the checks developed in the center of the boards (Morris and McFarling 2008). Lodgepole pine with rib-eased edge and ripple to edge profiles had a better appearance rating while Pacific silver fir boards had a higher rating when they had a rib-eased edge profile (Appendix 1). Morris and McFarling (2008) found that lodgepole pine boards with concave grain orientation had lower check length, depth, and width than boards with convex orientation (Appendix 1) (Morris and McFarling 2008). However, growth ring (a) (b) (c) (d) 51  orientation (concave, convex) had no effect on the checking of profiled and unprofiled Pacific silver fir boards. Pacific silver fir deck-boards had the narrowest checks when ribbed-eased edge configuration and concave growth ring orientation were combined (Appendix 1).   Evans et al. (2010) tested the ability of rib and ripple profiles to reduce checking of Pacific silver fir and southern pine deck-boards exposed to five days of accelerated weathering (Figure 2.25). They found that the rib profile was more effective at reducing surface checks in Pacific silver fir than with pine. They also found that boards with spiral grain tended to develop diagonal checks that crossed-over the profiles making the checks easier to see (Evans et al. 2010).    Figure 2.25: Confocal image of profiles used by Evans et al. (2010) in their study of the ability of rib and ripple profiles to reduce the checking of Pacific silver fir and southern pine: (a) rib profile; (b) ripple profile (Source: Evans et al. 2010)  Cheng and Evans (2012) tested the ability of a rib profile and a phenol formaldehyde (PF) resin treatment to reduce the checking of white spruce deck-boards exposed to 24 weeks of natural weathering. The combination of profiling and PF treatment greatly reduced the size of checks that developed in weathered boards. They also found that profiled boards had narrower checks than the flat untreated control. Akhtari and Nicholas (2014) examined the effect of surface profiling on the checking of southern pine boards treated with two different preservative (EcolifeTM and amine copper azole) and exposed to accelerated weathering for 24 days. The combination of profiling with both of these treatments greatly reduced the size and number of checks that developed in the boards when they were exposed to accelerated weathering, as well as the distortion of the deck-boards.  (a) (b) 52   Very little research has been done that provides insights into how profiling reduces checking of decking. McFarling and Morris (2005) observed that profiling results in checks following the grooves of the profile which helps to hide the checks (McFarling and Morris 2005, McFarling et al. 2009). However, profiling increases the cupping of boards which can increase or decrease check size depending on how tightly fasteners secure boards to the underlying sub-frame. Morris and McFarling (2008) observed that cupping was highest in boards with a flat edge configuration and convex orientation growth rings (Appendix 1). However, this observation may be explained by the ripple profile being lower than the flat edge (Morris and McFarling 2008). Lodgepole pine with a ripple and flat edge configuration had the smallest check sizes among the profiled boards whose growth rings had a convex orientation (Appendix 1) (Morris and McFarling 2008). Evans et al. (2010) suggested that profiling reduced checking because of a combination of reduced surface tension stress as result of a larger surface area for drying and reduced moisture gradients.  They found that profiled boards had longer, but narrower checks than the unprofiled controls. Most of the checks occurred in the grooves of the profiles. For example, the ripple profile had no checks that developed on the peaks of the profile ridges (Evans et al. 2010). Evans et al. (2010) suggested that the location of checks in grooves might encourage the checks to become longer, but not wider.   Mallett (2012) measured the strain development at the surface of profiled radiata pine deck-boards during drying and wetting using digital image correlation. He found that surface profiling increased strains at the surface of deck-boards and focused them in profile grooves. According to Mallett (2012), increased strains were caused by a combination of shrinkage, and steep moisture gradients. The large surface area of the ridges caused them to swell and shrink faster (Mallett 2012). When the ridges swell the groove is pushed inwards, but when they shrink the groove is pulled apart resulting in very high strains which cause checks to form at the base of grooves (Mallett 2012).   The type of profile may influence how easy it is to see checks.  Evans et al. (2010) observed that boards with ripple profiles did not develop checks on their peaks (Evans et al. 2010). It is easier to see such checks that develop on the top than the base of the groove. 53  However, checks can be more easily seen in the grooves of ripple profiles because the grooves are wide (Evans et al. 2010). Boards with rib profiles develop checks in their grooves, but the grooves are narrow, making the check difficult to see unless the checks are very large (Evans et al. 2010).  2.4.4 Kerfing  Kerfing is the term used to describe the practice of making one or two radial saw cuts in the outer surface of poles, posts or square rectangular beams (Figure 2.26). Kerfs can be of varying depth, and they usually extend for some length along the pieces. Kerfing significantly reduces the checking of round poles and posts as well as square and rectangular posts (Graham and Estep 1966, Helsing and Graham 1976, Chandler 1968, Graham 1973, Ruddick and Ross 1979, Morrell and Newbill 1986, Ruddick 1988, Evans et al. 1997, 2000, Kurisaki 2004).   Figure 2.26: Examples of kerfed wood products: (a) Kerfing of boards used as the top of an outdoor restaurant table in Stockholm, Sweden; (b) kerfing underneath a highway crash barrier in Kyushu, Japan; (c) kerfed and treated posts in Kyushu, Japan; (d) small boxed heart kerfed timber beams in Kyushu, Japan (photos, Dr. Philip D. Evans) 54  2.4.4.1 Square posts and rails  Chandler (1968) cut different kerfs in square western hemlock guardrail posts prior to treatment of the guardrails with a petroleum-creosote preservative. The dimensions and locations of the kerfs in the guard rails are shown below in Figure 2.27. All three kerfing treatments reduced check widths, when the post were in service outside for two years but there was little difference in the effectiveness of the three types of kerfing treatments (Chandler 1968).   Figure 2.27: Kerfing treatments applied to western hemlock square guard rail posts: (a) no treatment; (b) single kerf, 2” (50.80 mm) deep; (c) single kerf, 3 3/4“ (95.25 mm) deep; (d) double kerf, 2” (50.80 mm) deep, 1 ½” (38.10 mm) from center of posts on opposite sides (redrawn from Chandler 1968) 2.4.4.2 Logs and round poles  Liese (1961) cut a 15 mm deep saw kerf in the end grain of 10 untreated European beech logs. The kerfed logs developed 2.8 end-checks per cm2 compared to 3.4 end-checks per cm2 for unkerfed logs. Graham and Estep (1966) found that the combination of incising and kerfing reduced the checking of round Douglas fir spar cross-arms.  They observed an inverse 55  correlation between kerf width and check widths when cross-arms were dried. Kerfing did not reduce the number of smaller checks, but they suggested that it reduced drying stresses in the cross-arms (Graham and Estep 1966). Graham (1973) found that if three equally spaced kerfs were cut into Douglas fir poles, two of the kerfs would stop widening when the poles dried while the other would continue to widen and deepen to the center of the post. This finding led to the idea of applying one deep kerf to a pole rather than applying multiple kerfs (Graham 1973). Helsing and Graham (1976) found that kerfing reduced the size of checks in pressure treated Douglas fir poles when they were exposed outdoors for 5 to 11 years. However, checks still developed at the base of the kerfs in some poles, suggesting that the kerf might concentrate drying stresses (Helsing and Graham 1976). Ruddick and Ross (1979) investigated the effect of a 4 mm wide kerf on the checking of untreated (unincised) Douglas fir poles. Kerfing reduced by 58% both the average width and depth of the worst checks that developed when poles were weathered outdoors for 44 months. The kerfs increased in width from 4 mm to almost 10 mm during the weathering trial, but the growth of the widest checks was small (Ruddick and Ross 1979). Morrell and Newbill (1986) also found that kerfing reduced the width of drying checks, and also controlled check openings, in untreated sections of Douglas fir poles stubs treated with various preservatives. Ruddick (1988) found that kerfing reduced the width and average depth of the ‘worst checks’ in white spruce poles treated with ammoniacal copper arsenate (ACA) by a factor of 2. At the same time the width of the kerf increased by 75% (Ruddick 1988). Goodell and Pendlebury (1991) found that kerfing also reduced checking of CCA-C treated red spruce posts exposed to weathering for one year. The reductions in check width and depths were to 40% and 41% compared to those of checks in center bored posts.   Evans et al. (1997) investigated the effects of single and double kerfing on the checking of CCA-C treated slash pine posts exposed to the weather in Australia for up to 1 year. All kerfs were 3 mm wide. The single kerf was 50 mm deep, while the double kerfs were each 25 mm deep on opposing sides. Both types of kerfing were equally effective at reducing the number and size of checks in the weathered posts. However, the single kerf tended to open wider than the double kerfs and as a result Evans et al. (1997) suggested that double kerfing might be more acceptable for some end-uses of the posts (Evans et al. 1997). In a follow-up study Evans et al. 56  (2000) investigated the effects of various treatments including kerfing on the checking of 125 mm and 150 mm diameter ACQ treated radiata pine and slash pine posts (Figure 2.28).   Figure 2.28: Kerfing treatments applied to radiata and slash pine posts: (a) single kerf, 30 mm deep; (b) single kerf, 45 mm deep; (c) single kerf, 60 mm deep; (d) double kerfs, 15 mm deep; (e) double kerfs, 22 mm deep; (f) double kerfs, 30 mm deep (redrawn from Evans et al. 2000)  They found that increasing the depth of single and double kerfs reduced check sizes and number of checks that developed when posts were exposed to the weather for one year (Evans 2000). The effect of increasing the depth of kerf at reducing checking was more effective with the double than the single kerfs. Kurisaki (2004) also tested the effects of single and double kerfing on the checking of 100 mm diameter ACQ treated Japanese cedar posts. He exposed the kerfed posts horizontally outside for 17 months. Kerfing reduced check size and depth, but there was no statistical difference between the effects of single and double kerfing on checking. Kurisaki (2004) also observed that only 6% and 18% of the checks in single and double kerfed posts penetrated through the treated zone into untreated sapwood.   57  2.4.4.3 Kerfing used in drying  Kerfing has also been used to reduce the checking of round and square posts during drying. Kerfing in Japan is called sewari, and involves ‘placing a single kerf’ on the least visible part of the wood, prior to drying (Figure 2.29) (Goda 2005). Many Japanese papers mention kerfing.  Yamazaki and Atsushi (1996) kiln dried kerfed and unkerfed square sections of Sakhalin fir (Abies sachalinensis F. Schmidt) posts. Originally, prior to drying, the kerfs in the posts were 10 mm wide. The average width of the kerfs after drying was 10.7 mm, while the average width of the widest checks in unkerfed boards were 10.6 mm. Their findings suggest that unkerfed posts develop checks similar in width to that of the kerf in kerfed posts. Yamazaki and Atsushi (1996) also air dried another set of kerfed and unkerfed Sakhalin fir posts of similar dimensions for one year before they kiln dried them. The kerfed posts developed shorter checks than the unkerfed posts after air and kiln drying and the kerfs were only 10.2 mm wide. Saitoh et al. (2000) kerfed boxed heart Japanese cedar posts prior to kiln drying. They noted that after drying almost 57% of kerfed pieces were free of checks while only 6% of unkerfed posts were free of checks. Check area in unkerfed posts was twice as large as that in kerfed posts (Saitoh 2000). Sadoh (2001) recommended kerfing square timbers during drying to reduce checking.   Figure 2.29: Examples of sewari: (a) Kerfed sewari style posts used in Japanese homes (photo, Dr. Hiroki Sakagami, Kyushu University); (b) kerfed sewari style posts used as a parking barrier in Japan (photo, Dr. Philip D. Evans)    (a)                                                                            (b) 58   Subsequent studies have confirmed that kerfing reduces surface checking of wood during drying (Sadoh 2001, Yeo et al. 2007, Lee et al. 2010, 2011a, 2011b). Yeo et al. (2007) found that kerfing reduced the length and width of surface checks that developed in round and square Japanese larch (Larix leptolepis G.) posts during kiln drying. Similarly, Lee et al. (2010) found that kerfing reduced the length and area of surface checks that developed when square-section Japanese cedar posts were dried using a radio-frequency/vacuum drying process. Lee et al. (2010) also observed that kerfing reduced the area of surface checks by over 50%. Furthermore, the combination of kerfing and high temperature low humidity drying almost completely eliminated surface checks (Lee et al. 2010). In subsequent studies Lee et al. (2011a,b) found that kerfing reduced the length of checks in Japanese larch logs and Korean red pine (Pinus densoflora Siebold and Zucc.) box heart posts during radio frequency drying.  2.4.4.4 Kerfing of deck-boards  Despite the positive effects of kerfing on the checking of square posts described above, Ratu et al. (2007) found that kerfing did not have the desired effect of reducing checking of flat-sawn southern pine deck-boards exposed to the weather (Ratu et al. 2007). They cut one, two or three, 3 mm wide kerfs, into the underside of southern pine boards (Figure 2.30). The depths of the kerfs were 10, 13, and 20 mm, equating to 25, 33 and 50% of the boards’ thickness. Boards were fixed, pith-side-up on a weathering rack and weathered for one year. Kerfing did not reduce the number or size of checks that developed in the boards, and the kerfs became narrower during exposure. Ratu et al. (2007) speculated that kerfing might have allowed the boards to flex more during wetting and drying creating greater surface strains at the exposed surface of deck-boards.  Figure 2.30: Images of 10 mm deep kerfs used by Ratu et al. (2007) in their study of the kerfing of flat-sawn southern pine deck-boards: (a) one kerf; (b) two kerfs; (c) three kerfs (photos, Dr. Philip D. Evans) (a)                                                 (b)                                                 (c) 59  2.4.5 Incising  Incising as its name suggests creates small incisions or perforations at wood surfaces allowing impermeable wood species to be more deeply and consistently penetrated by preservatives (Morris et al. 1990, Morris et al. 1994). The incisions are commonly produced by toothed rollers (Figure 2.31), but they can be produced by needles, lasers, drill-bits or high-pressure water jets (Morris et al. 1994). Several studies have shown that incising can reduce the size of checks that develop when wood is dried or exposed to the weather (Harkom and Rochester 1930, Arnold at al. 1949, Graham and Estep 1966, Ruddick and Ross 1979, Morris et al. 1990). However, some studies have shown that incising has little effect on checking (Jermer and Bergman 1992, Evans et al. 1997).     Figure 2.31: Images of: (a) Double headed toothed rollers used to incise wood; (b) lodgepole pine, butt-incised, transmission pole (photos, Dr. Philip D. Evans)  There are many examples of the ability of incising to reduce the checking of hardwood and softwood railway ties during drying and outdoor exposure. Harkom (1928) found that incised hard maple railway ties had less noticeable checks than unincised ones after they were stacked and air dried for 16 months. The Great Northern Railway and Canadian Pacific Railway incised Douglas fir, larch and hardwood ties prior to seasoning (Duncan 1932). The incising helped reduce checking in all of the ties and incising was adopted as a treatment prior to the seasoning of railway ties.  Arnold at al. (1949) found that incising of hardwood railway ties prior to treatment with a 50/50 creosote-tar solution reduced the width of checks that developed in (a)                                                                                (b) 60  the ties when they were exposed outdoors for 7 years. Only 2.5% of the incised ties had wide checks (9.5 mm wide) whereas 15.4% of unincised ties had wide checks. Francoisi (1956) noted that most of the large checks in incised beech (Fagus sp.) railway ties developed at the ends of the ties whereas large checks also developed in the face of unincised ties. Henry (1970) examined the checking of unincised and incised elm (Ulmus sp.), maple (Acer sp.), and sweetgum (Liquidambar sp.) railways ties during outdoor exposure. Most of the unincised elm and maple ties checked badly and failed during exposure. For incised ties, elm had the highest number of large checks while sweetgum had the lowest. It was noted that the checking of incised railway ties increased less during outdoor exposure than that of unincised ties.   Some studies have also shown that incising can reduce the checking of poles and posts during air drying and also when the poles and posts are exposed outdoors. Dundas and White (1972) mentioned that the checking pattern on incised western hemlock poles during treatment was more uniform and the poles had smaller checks than unincised poles. Burnes (1923a) examined the effect of incising on checking of western red cedar and northern white cedar (Thuja occidentalis L.) poles that were exposed in the field for 18 months. The ends of the poles were incised and butt-treated. He noted that the widths of checks decreased in the incised area and checks originating from the unincised sections would not penetrate more than 50.8 mm (2”) to 76.2 mm (3”) into the incised area. In a follow-up study, creosote-treated western red cedar poles were butt-incised and exposed outdoors for 18 months. Burnes (1923b) noted that checks that originated in the unincised area tended to close up and become narrower when they enter the incised area of the pole (Burnes 1923b). These findings show that incising can be used to prevent existing checks from becoming bigger.   Relatively few studies have examined whether incising can reduce the checking of dimension lumber. Harkom and Rochester (1930) observed that incising helped reduce checking of creosoted Douglas fir beams during drying. Morris et al. (1990) examined the effect of double density incising on the checking of white spruce, lodgepole pine and subalpine fir (SPF) lumber during drying. They found that incising reduced the size of checks possibly because it creates large numbers of micro-checks rather than large checks. More recently 61  Listyanto et al. (2013) combined CO2 laser incising and a steam injection treatment and claimed that the dual-treatment was able to restrict the checking of Japanese cedar lumber during drying (Listyanto et al. 2013). However, their claim is based on visual observations of the samples after drying rather than measurements of the number or sizes of checks that developed after the lumber was dried.  Incising can also be combined with kerfing to reduce checking. Graham and Estep (1966) combined incising and kerfing and found that it greatly reduced the number of checks as well as the width and depth of checks in round Douglas fir cross-arms stored outside for 37 months. However, incising on its own, irrespective of the incising pattern, only had a small effect on checking (Graham and Estep 1966). Lindgren (1976) investigated two types of incising as well as a combination of incising and kerfing to see if they could prevent large cracks from exposing untreated wood in pentachlorophenol-treated Douglas fir poles exposed in the field for 10 years. One incising treatment involved incising the poles at depths of 6.35 mm (1/4”) to 12.7 mm (1/2”) and drying the poles in oil under vacuum (Boultonized). Another incising treatment deeply butt incised the poles at depths of 50.8 mm (2”) with a treated zone of 63.5 mm (2.5”) deep. The kerfing and incising combination treatment involved incising to a depth of 19.1 mm (3/4”) in combination with a series of 54 mm deep (3 1/8”), 6.34 mm (1/4”) wide kerfs cut 609.6 mm (2’) above and below the groundline. Saw cuts were spaced between 31.8 mm to 38.1 mm apart depending on the diameter of the poles. Ninety two percent (92%) of the poles that were deeply incised had checks that were deep enough to expose untreated wood. In contrast only 56% of the Boultonized and incised poles had checks that exposed untreated wood. However, only 50% of the poles which were incised and kerfed developed checks that exposed untreated wood.  Despite all of the studies, mentioned above, that show that incising can reduce checking, there are observations to the contrary. For example, in a panel on cross ties at the meeting of the American Railway Engineering Association Kistler mentioned that even though incising was effective at reducing checking and splitting of certain southern and swamp oaks (Quercus sp.) ties it was ineffective with red oaks (Collister 1956). For poles, Jermer and Berman 62  (1992) examined the surface checking of incised and unincised and preservative-treated Scots pine glulam poles. They found that the incised poles had more checks than the unincised controls. In accord with Jermer and Bergman’s findings Evans et al. (1997) found that incising was ineffective at reducing check sizes or numbers in CCA-C treated slash pine posts exposed to 1 year of weathering (Evans 1997). Kurisaki (2004) also found that incising did not reduce check width and depth in 100 mm diameter ACQ treated Japanese cedar posts when they were exposed outside for 17 months. Finally, Morris and Ingram (2002) investigated the effects of incising on the checking and cupping of CCA-treated deck-boards cut from seven Canadian wood species and exposed to the weather for 9 years. Unincised boards acted as controls. Checks were deeper in unincised than in incised boards for lodgepole pine, subalpine fir, eastern white spruce and jack pine boards exposed in Vancouver and also in lodgepole pine, subalpine fir, eastern white spruce, and balsam fir boards exposed in Ottawa. In terms of a qualitative rating for check appearance, all of the incised boards (with the exception of western hemlock exposed in Vancouver) rated better than the unincised boards, but the difference between the two was not statistically significant. 2.4.6 Other physical and mechanical treatments  There are some other physical and mechanical devices that can be used to reduce the checking of wood. Some of these devices are used to prevent existing checks from getting bigger and they are generally limited in their effectiveness. Nevertheless, they may be useful in reducing checking if combined with other check-reducing methods, such as profiling, kerfing or incising.  2.4.6.1 Screws and nails  Some wood species, especially high density ones, tend to split when they are screwed or nailed (Heebink 1959). This can be prevented by pre-drilling nail holes, but this is time consuming and costly (Cooke 1936). A solution to the problem of the splitting of wood by nails is described by Cooke (1936). He used blunt nails to fix dry white cypress pine (Callitris glaucophylla J. Thomps. & L.A.S. Johnson) boards. White cypress pine is a dense, brittle, softwood commonly used for flooring in Australia. Cooke (1936) found that the use of blunt 63  nails reduced the splitting of boards by 38 to 46 %. Evans et al. (2003) examined the effect of different fixings on the splitting of CCA-treated radiata pine deck-boards exposed to the weather for 1 year. Three different types of fasteners were used: (1) nails from a nail gun; (2) self-drilling screws; and (3) screw-nails (Figure 2.32a). They found that checks associated with fasteners were larger around screws than around the other two types of fasteners (Evans et al. 2003). They suggested that the screws may have fixed boards more firmly than nails, which may have prevented release of stress via cupping and led to greater release of stress through checking.  In a follow-up study, Urban and Evans (2005) applied three different fasteners to flat-sawn southern pine decking: (1) standard decking-screw; (2) decking-screw with a rubber washer; and (3) hidden fastener (Figure 2.32b). The fasteners had little effect at reducing the size of the checks that developed when boards were weathered for 6 months, but boards fixed with hidden fasteners had checks with a higher length/width ratio. In other words the checks were longer and thinner possibly making them more difficult to see. According to Urban and Evans (2005) the hidden fixing may have allowed some flexing of boards releasing stresses by cupping rather than checking.     Figure 2.32: Types of fasteners used in the studies of Evans et al. (2003) and Urban and Evans (2005): (a) Nails and screws affixed to radiata pine deck-boards; (b) screws and fasteners affixed to southern pine deck-boards (photos, Dr. Philip D. Evans)   Self-drilling screws Screw-nails Nail gun nails Decking-screw Decking-screw with rubber washer Hidden Fastener (a)                                                                                (b) 64  2.4.6.2 Hooks, rings and fasteners  Kubler (1987) in a review on growth stresses and strains in wood mentioned that several different metal and plastic fasteners have been used to reduce end splitting of logs (Figure 2.33). The fasteners that Kubler (1987) mentioned included S-hooks, C-irons, and steel rings. These fasteners are fixed to the ends of logs and reduce the size of end-checks that develop when the logs dry. The fasteners are thought to restrain the shrinkage of wood which causes checks to form (Lepitre and Mariaux 1965 cited by Kubler 1987).   Figure 2.33: Images of various mechanical fasteners designed to reduce the end-splitting of wood: (a) Circular nail plates used to reduce the end-splitting of a spotted gum (Corymbia maculata (Hook.) K.D. Hill and  L.A.S. Johnson) power poles prior to treatment with CCA; (b) Square nail plates used to reduce the end-splitting of Eucalyptus railway ties; (c) Steel ring used in Japan to reduce checking of a railway tie; (d) S-hook designed to reduce the end-splitting of a railway tie in South Africa (photos, Philip D. Evans)  Lepitre and Mariaux (1965) fixed S-hooks to the ends of Okoume (Aucoumea klaineana Pierre) veneer logs transported from Gabon to France. They found that the S-hooks prevented the growth of 5 mm wide splits that passed through the pith, but they were unable to prevent larger cracks from getting bigger during air drying. Furthermore, the insertion of S-hooks into 65  the ends of logs caused new checks to form around the S-hooks (Lepitre and Mariaux 1965). A similar effect was noted by Evans et al. (2003). They found that checks in CCA-treated decking tended to develop and get bigger where nails or screws were located in deck-boards. Lepitre and Mariaux (1965) mentioned that cracks formed from inserting S-Hooks into logs could relieve stresses and prevent larger splits from occurring. Tisseverasinghe (1967 cited by Kubler 1987) found that S-hooks did not restrict end checks or severe splits from developing during the air drying of flooded gum (Eucalyptus grandis W.Hill ex Maiden) transmission poles.    Shunk (1976 cited by Kubler 1987) mentioned the use of S-Irons, C-irons, Beegle irons and dowels to reduce checking of railway ties. Perem (1971) mentioned the effect of S-hooks on the end-splitting of hardwood (sugar maple, birch, and beech) railway ties during service. He reported that S-hooks had no effect on the size or the number of splits that developed when the ties were air dried or used in service. Kubler (1987) mentioned that steel rings inserted into the ends of logs have been used to reduce checking of end-grain. However, Liese (1961 cited by Kubler 1987) found that single ring, double rings, or S-hooks were unable to prevent checks from developing and getting bigger in the ends of European beech logs during air drying. In accord with Lepitre and Mariaux’s (1965) observations, Liese (1961) found that inserting S-hooks into the ends of logs caused small checks to develop (Liese 1961). Yang and Waugh (2001) mentioned the use of metal or plastic S or C shaped irons to prevent end-splitting of logs, but they pointed out that the wood may split when the fastener is removed. Furthermore, they go on to point out that these fasteners do not prevent large splits from developing in highly stressed logs.  Since most of the aforementioned fasteners were not very effective at controlling end splits they were gradually replaced by nail plates (Kubler 1987). Nail plates contain nail-like protrusion which fix the plates to the ends of logs. These nail plates (gang nails) are more commonly used to connect the members of wooden roof trusses, but are also used to reduce the end checking of railway ties and power poles (Jureit and Kushner 1968, Jureit and Kushner 1970, Jureit and Kushner 1977, Jureit and Kushner 1979). Tisseverasinghe (1967) applied gang nails to the ends of flooded gum poles and found that the gang nails controlled the size of the 66  splits that developed when the poles were air dried. Tisseverasinghe (1967) suggested that circular gang nails roughly the same diameter as the logs would be more effective than square gang nails at preventing splitting. Hardie (1974) suggested that gang nails should be used to restrain the splitting of flooded gum log poles in Zambia. Decena and Cruz (1977) fixed nail plates to the ends of 55-105 cm diameter almon (Shorea almon Foxw.), red lauan (Shorea negrosensis Foxw.), and tangile (Shorea polysperma (Blanco) Merr.) logs. The length and width of checks were measured twice: (1) after the tree was felled and the nail plate was applied; and (2) 3 days later after the nail plate was removed (Decena and Cruz 1977). They found that the nail plates reduced the widths of checks two-fold (Decena and Cruz 1977). A recent study by Love and Morrell (2012) found that end-plating reduced check numbers and sizes in pentachlorophenol treated Douglas fir cross-arms. End-plating restricted the opening of checks during drying and also the closing of the checks during wetting. Overall, end-plating reduced check width by 73% (Love and Morrell 2012).  2.4.6.3 Beveling  There are some less well-known methods that have been used to reduce checking. One such methods is to bevel or round off the corners of posts to reduce splitting and checking of end-grain (Martin 1892, Popular Mechanics Press 1942). Martin (1892) suggested beveling the sides of posts to reduce splitting when posts are hammered into the ground. An article in a Popular Mechanics garden book suggested that beveling or rounding off the tops of posts can help deflect water and reduce checking and rotting (Figure 2.34) (Popular Mechanics Press, 1942).   67   Figure 2.34: Images of beveled and rounded posts: (a) Rounded post; (b) four-sided beveled post; (c) one-sided beveled post; (d) two-sided (v-shaped) beveled post (Source: Popular Mechanics Press 1942)  Contrary to suggestions that shaping the ends of posts can reduce checking, Liese (1961) applied a v-shaped bevel to 10 European beech logs, but he found that the treatment was ineffective at reducing end-splitting of the logs.  2.4.7 Exterior finishes (coatings)  Feist (2006) classified wood finishes into two basic types: (1) finishes that leave a film, layer, or coating on the surface; and (2) finishes that penetrate wood’s surface without leaving a coat or film. Finishes in the first category are: paints, varnishes, solid colour stains, and overlays bonded to wood surfaces (Feist 2006). Finishes belonging to the second category include: oils, water repellants, solvent-borne stains, preservatives, and surface treatments.   Nejad and Cooper (2011a) mentioned that coatings prevent wood from absorbing moisture and protect the wood from UV degradation, both of which should restrict checking. Stamm (1964) suggested that the function of coatings is to reduce the shrinkage and swelling of wood, but he noted that most coatings do not adhere well to wood nor are they completely water-proof. Some papers recommend against using film forming coatings because they do not prevent wood from checking. Furthermore, the coatings may crack, peel, and delaminate as a result of dimensional changes in wood (Nejad and Cooper 2011a, Feist 2006).  Floyd (1983) mentioned that film-forming coatings such as latex paints do not prevent wood from cracking and the film itself becomes very stressed and may crack when checks develop in the underlying (a)                                    (b)                                (c)                                      (d) 68  wood. Ashton (1967) observed that clear coatings containing alkyd resins crack, delaminate and then break apart due to the formation of small voids in earlywood, and the poor adhesion of the coating to the latewood. Roux et al. (1988) also observed that film-forming stains crack and peel off from wood. Nevertheless, some studies have shown that coatings can restrict cracks and checks from developing at wood surfaces (Mackay 1972, Nejad and Cooper 2011a). 2.4.7.1 Opaque type coatings  Coatings can be applied to untreated wood and to wood that has been treated with preservatives such as CCA. The coating may reduce the dimensional instability of wood resulting in less checking (Ross et al. 1992, Nejad and Cooper 2011a). McCarthy et al. (1982) applied 74 different combinations of preservatives and coatings (15 preservatives, 21 coatings) to radiata pine and mountain ash and exposed the treated and coated wood outdoors for two years. They found that only 13 combinations of coatings and preservatives performed well, and only one preservative treatment was effective at restricting the checking of panels (Appendix 2). They concluded that, a combination of a preservative and coating is needed to protect the wood from weathering-induced checking and cupping.   McCarthy et al. (1982) found that some acrylic and alkyd paints restricted the size of checks, while boards treated with an opaque acrylic latex stain only had moderate cupping and checking. In a later study, Roux et al. (1988) applied four different “chestnut” brown finishes (2 coats of impregnating stain, film forming stain, acrylic latex, and alkyd paint consisting of 1 brush coat of preservative, 1 white coat, and 1 final brown coat) to European beech, Norway spruce, Scots pine, Douglas fir, and dark red meranti panels and exposed the finished panels to the weather at four different sites in Europe. They found that the panels coated with the acrylic latex and alkyd paints tended to crack less than the panels coated with the other finishes but their behavior was different. The acrylic latex paint’s elasticity allowed it to accommodate the movement of the underlying wood but it did not eliminate shrinkage, swelling and cracking of wood and coating (Roux et al. 1988). Alkyd paints were more resistant to moisture, protecting the wood from dimensional changes and limiting cracks in the paint film (Roux et al. 1988).  69   Finishes can also reduce the checking of plywood exposed outdoors, For example, McLaughlan (1991) exposed several finished radiata pine plywood panels outdoors for twelve years in Rotorua, New Zealand. He found that painting the panels with an alkyd-based primer and a top coat of white alkyd paint reduced the number of checks that developed at the surface of the finished plywood panels. Panels that were painted contained 5 checks per 100 mm compared with 18 checks per 100 mm for uncoated plywood. Trinh et al. (2012) applied three different types of coatings (acrylic paint, acrylic stain, and alkyd stain) to European beech plywood that had been chemically modified with N-methylol melamine or N-methylol melamine with wax. They found that all three coatings were effective at reducing surface checking of the modified plywood. Furthermore, the coatings were more effective on plywood treated with N-methylol melamine and wax because it was more dimensionally stable than plywood simply modified with N-methylol melamine.  2.4.7.2 Semi-transparent coatings  Semi-transparent stains are popular because they are only mildly pigmented and do not obscure wood’s attractive features. They do not form a thick surface coating and do not blister or peel under the influence of moisture (Nejad and Cooper 2011a and Feist 2006). Furthermore, they are able to protect wood from UV degradation, although they are less effective than opaque film-forming finishes that completely block the transmission of light (Nejad and Cooper 2011a and Feist 2006). Nejad and Cooper (2011a) tested 14 different kinds of brown-coloured commercial semi-transparent stains on southern pine sapwood treated with different preservatives (CCA, CA, ACQ, and untreated). They found that all of the 14 coatings reduced checking of wood (expressed as surface area of the checks/total area of wood) by up to 30 to 40%. Nejad and Cooper (2011a) found that thicker and more viscous coatings were the most effective coatings at reducing checking. Nejad and Cooper (2011b) examined the surface of coated and weathered samples using scanning electron microscopy (SEM) and found that the stains penetrated through the woods rays, which are where checks often develop. Therefore, they recommend the use of a surface coating with high flexibility, and good adhesion and cohesion to reduce the checking of treated southern pine (Nejad and Cooper 2011b). Xie et al. 70  (2008) applied two semi-transparent varnishes (oil-based and water-based acrylic) to untreated and methylated 1,3 dimethylol-4,5-dihydroxyl-ethyleneurea (mDMDHEU) impregnated flat-sawn Scots pine panels. After 18 months of exposure they found that the oil-based varnish had peeled off untreated boards while the water-based acrylic adhered to the wood even when the panels cracked. They found that both types of coatings reduced surface cracking of mDMDHEU-treated boards (Xie et al. 2008). In contrast, Roux et al. (1988) found that a thin coating of an impregnating stain was ineffective at restricting the checking of wood.  2.4.7.3 Oils and resins  Oils and resins are important components of water repellents and water repellent preservatives, both of which can reduce the checking of wood exposed outdoors (also see later section on these treatments). Borgin and Corbett (1970a) found that linseed oil and fish oils were the best at reducing the checking of radiata pine boards exposed outdoors for four weeks. The linseed oil treated wood showed no deterioration while the wood treated with hard light fish oils developed slight cracking (Borgin and Corbett 1970a). Opoku (2007) applied decking oil to hardwood deck-boards and teak oil to hardwood cross-pieces and exposed the treated wood outdoors for 6 months. He found that both oils were effective at reducing the length of checks that developed in the wood during weathering. Miklečic´ et al. (2010) applied three types of natural oils (universal oil, thermowood oil, and teak oil) to three thermally modified hardwoods and exposed them to 672 hours of accelerated (QUV) weathering. Two of the oils (thermowood, and teak oil) contained UV reflectors and organic pigments. All of the oils reduced surface checking (Miklečic´ et al. 2010). In contrast, Sandberg (1999) found that linseed oil was unable to reduce the size of checks that developed when Norway spruce and Scots pine were exposed outdoors.   Borgin and Corbett (1970b) in follow-up work to that described above, applied various resins to radiata pine and found that Scopol 55NM and Alkydal S65 resins were the best at protecting wooden boards from weathering. They concluded that neither oil, wax, nor resin provided good protection on their own and they recommended that a combination of these materials would be better at protecting wood (Borgin and Corbett 1970a,b). However, 71  Campbell (1971) applied a micro-crystalline wax to tangentially sawn jarrah joinery stock to prevent it from checking during air drying in the summer. He found that the wax reduced surface checking and he suggested applying more wax to the sapwood face of the joinery because the checks were more severe on that side.  Mackay (1972) found that treating green wood with a monomer could reduce the initiation of cracks. Mackay (1972) brush coated messmate stringybark samples with different concentrations of glycol methacrylate (GMA), and polyethylene glycol (PEG) before samples were kiln dried. Both treatments reduced the checking that developed when the wood was kiln dried. Yusuf et al. (1995) applied a polyurethane-resin type lacquer modified with cross-linking agents to radiata pine laminated veneer lumber (LVL) and exposed the treated LVL to accelerated weathering. As expected they found that the coating reduced checking of the LVL.  2.4.8 Preservative treatments  Preservatives are generally designed and used to protect wood from fungal decay and insect attack rather than protecting wood from physical degradation caused by weathering (Zahora 1991). However, sometimes water repellant additives are added to preservatives to protect wood from weathering (Zahora 1991). There are two main types of preservatives that are available commercially: (1) oil and solvent-based, and; (2) water-based (Groenier 2006). 2.4.8.1 Oil solvent based preservatives  Oil-based preservatives make wood less susceptible to checking because the oil acts as a water repellent (Groenier 2006, Zahora 1991)  However, some of these preservatives for example creosote smell, and make the surface of the treated wood oily (Groenier 2006). Furthermore, creosote residue at wood surfaces can harden to form solid deposits called crud. In addition to creosote, other solvent-based preservatives include: pentachlorophenol, copper naphthenate, oxine copper, and iodopropynyl butylcarbamate (Groenier 2006).   Creosote is still widely used in North America to treat ties and bridge timbers. Creosote decreases checking and delamination of southern pine bridge timbers exposed to the weather 72  (Selbo et al. 1965).  Rietz (1961) treated red oak crossties with various creosote and pentachlorophenol solutions and exposed them to accelerated weathering. He found that all of the treatments reduced checking, but there was no significant difference in the effectiveness of the different treatments. Gilfedder et al. (1968) found that creosote was more effective at reducing the size of splits in messmate stringybark than water-based salt-preservative because it reduced the rate of drying of the treated wood. McCarthy et al. (1982) found that only one preservative treatment, Haeger’s Royal Process (oil-based copper and pentachlorophenol solution, Appendix 2), significantly reduced the checking and cupping of radiata pine panels exposed outdoors for two years.   In contrast, Buchanan et al. (1990) found that a solvent-based preservative was unable to reduce the splitting of Pacific silver fir and western hemlock shingles exposed to accelerated weathering. Similarly, McLaughlan (1991) found that brush coating radiata pine plywood with copper naphthenate did not reduce face checking of the plywood when it was exposed outdoors for twelve years. In accord with this observation, Page (2002) found that copper napthenate in a light organic solvent (LOSP) and also a tributyltin-napthenate LOSP were ineffective at reducing checking of softwood boards exposed to one year of weathering. However, Placket et al. (1984) found that a copper naphthenate LOSP was more effective than CCA or alkyl ammonium (AAC) preservatives at restricting the checking of radiata pine shingles.  2.4.8.2 Water based preservatives  Most water-based preservatives resist leaching and leave a dry paintable surface, which makes them more suitable than oil-based preservatives for residential uses such as decks and fences (Groenier 2006). However, some water-based preservatives contain hydrophilic components which attract moisture to the wood and increase swelling (Nejad and Cooper 2011a). Water-based preservatives commonly used for residential applications in different parts of the world include chromated copper arsenate (CCA), alkaline copper quaternary (ACQ), ammoniacal copper zinc arsenate (ACZA), copper azoles (CBA), borates, and micronized copper preservatives (Groenier 2006). 73   CCA if used alone does not prevent checking of wood, and there are reports that suggest that it does the opposite (Plackett et al. 1984). Some studies have suggested that the salt or oxides in CCA, or redrying following treatment, is responsible for the increased checking of CCA- treated wood exposed to the weather (Mackay 1973, Bariska et al. 1988, Sandberg 1999, Evans et al. 2003).  Sandberg (1999) mentioned that wood impregnated with CCA may be more susceptible to cracking compared with untreated wood because of damage to the wood surface caused by the preservative. Bariska et al. (1988) examined the micro-structure of CCA treated flooded gum and radiata pine. He concluded that acid hydrolysis of wood by CCA weakened the growth ring boundary, the ray and fiber tissues, the middle lamella of single cells, and made the wood more brittle and susceptible to checking when treated samples were exposed to weathering (Bariska et al. 1988). The presence of moisture in the wood reactivates the sulphates in CCA salts and dissolves wood substance according to Bariska et al. (1988). Mackay (1973) treated radiata pine boards with CCA and then dried them. The treated boards checked more than wooden controls treated with water, suggesting that the CCA made the wood more susceptible to checking. Evans et al. (2003) suggested that some of the checks that developed in CCA-treated decking boards during weathering might have developed when the boards were re-dried after preservative treatment. This suggestion was based on their observation that the checking of CCA treated wood was similar to that of water-treated boards rather than untreated and undried controls.   Plackett et al. (1984) found that CCA-treated radiata pine checked badly. Sandberg (1999) treated both Norway spruce and Scots pine boards with linseed oil, CCA, and a combination of CCA and linseed oil and found that none of the treatments significantly reduced checking of boards exposed outdoors for 33 months. Crawford et al. (1999) exposed several incised and unincised softwood boards treated with CCA-C and ACA outside for ten years. Unincised white pine (Pinus strobus L.) treated with CCA and incised white pine treated with ACA developed large splits and checks. Nejad and Cooper (2011a), however, found that the coating of CCA treated wood significantly reduced surface checking of boards exposed to the weather for three years. The effect of the same coating on the checking of untreated, ACQ, or CA treated wood was less pronounced. 74  2.4.8.3 Other preservatives  Some more recently developed preservatives have been found to reduce checking of wood during service. Dahlen et al. (2008) treated several southern pine deck-boards with a 3% waterborne resin acid (Pamite 90TM), a by-product of tall oil rosin, and exposed them outdoors for 22 months. The treatment greatly reduced the checking of the deck-boards. Akhtari and Nicholas (2014) tested whether EcolifeTM, an organic water-borne wood preservative containing wax and oil, or an amine copper azole (CA) reduced the checking of southern pine boards exposed to accelerated weathering for 24 days. Not surprisingly they found that the hydrophobic EcolifeTM preservative was better than amine copper azole at restricting the size of surface checks and distortion of the treated boards exposed to accelerated weathering. 2.4.9 Water repellent additives  Water repellents, as their name suggests, are designed to repel or retard the movement of moisture into wood (Miniutti et al. 1961). Borgin and Corbett (1970a) described water repellants as “solutions consisting of a blend of waxes, oils, and resins”. Waxes and oils are hydrophobes, resins strengthen the wood, and the solvent or water act as a carrier (Borgin and Corbett 1970a). Water repellent preservatives are another class of water repellants that are produced by adding a fungicide or mildewcide to the water repellent (Williams and Feist 1999). This combination of water repellant and preservative offers protection to machined wood products, according to Levi et al. (1970). For example, Miniutti (1961) mentioned that dipping wooden window sashes in a water repellent preservative provided good protection when the sashes were exposed outdoors.  Belford and Nicholson (1969) used a hydrophobic water repellent CCA preservative (CCA-WR) to treat beech, Douglas fir and western hemlock. Treated boards were exposed for three weeks in a weatherometer and also outdoors at several sites across the world. The hydrophobic CCA restricted the checking of boards exposed to both artificial and natural weathering (Belford and Nicholson 1969). Levi et al. (1970) exposed CCA-WR treated southern pine and Scots pine in a weatherometer. They found that the CCA-WR treated boards cracked less than CCA treated samples. Fowlie et al. (1990) exposed CCA and CCA-WR treated southern 75  pine boards outdoors for three months. As expected they found that the CCA water repellent restricted checking of boards. Zahora (1992) also found that a CCA-WR treatment restricted checking of southern pine treated boards exposed outdoor for one year, but the boards cupped. The boards were re-examined after nine years of outdoor exposure (Zahora 2000). After this period the CCA-WR had lost some of its ability to repel water but treated boards still checked less than CCA treated boards. Furthermore, the inner zone of boards treated with the CCA-WR was still water repellent.   Evans et al. (2009) exposed radiata pine treated with CCA containing different levels of wax and oil additives to natural weathering. They found that all of the wax and oil additives reduced the size and number of checks that developed in the CCA treated boards.  However, boards treated with CCA containing the oil additives were more dimensionally stable after weathering, than those treated with CCA containing the wax additive (Evans et al. 2009). Water repellents additives have also been added to other preservatives with positive effects. For example, Cui and Zahora (2000) found that a water repellent additive was effective at reducing surface checking of ACQ-treated southern pine exposed outdoors for three years.  2.4.10 End-grain sealants  End-grain sealants are viscous hydrophobic coatings that are applied to the ends of logs to prevent them from splitting during air drying. Dost (1983) suggested the use of end-grain sealants to prevent the splitting of the ends of treated sawn wood. Peck (1953) mentioned that white oak used for ship building was usually salted and end-coated in storage to prevent end-splitting. Simpson and Schroeder (1980) mentioned that end-coating reduced the end splitting of hardwood furniture components during kiln drying. Peck (1953) brushed urea-aldehyde paste, wood sealer, a water repellent preservative containing copper napthenate, and salt on the surface and ends of white oak. Half of the pieces were stored outside and half of the pieces were kept indoors. They found that all of the treatments reduced the size of checks and also end checks in the dried wood. A paint (Navy Specification 52-P-46) and urea-aldehyde paste were the most effective treatments, but some of the urea aldehyde paste was leached from samples exposed outdoors (Peck 1953).  76   Osborne (1965) applied three different coatings: (1) petroleum grease and power kerosene (60:40 weight); (2) bituminous emulsion; and (3) dimethyl silicone lacquer to the ends of blackbutt (Eucalyptus pilularis Sm.), brush box (Tristania conferta (R.Br.) Peter G.Wilson and J.T.Waterh.), and red mahogany (Eucalyptus resinifera Sm.) timber and exposed the end-coated samples outside for three months. Coatings 1 and 2 were both effective at reducing end checks, but coating 1 was the most effective. However, coatings do not provide a long term solution to the problem of end checking, and other methods such as protecting the timber from UV light was recommended (Osborne 1965). Dost (1983) applied various sealers and primers to the ends of Douglas fir glulam beams and exposed them outdoors for nine years.  A polyurea resin sealer was the most effective finish at preventing the beams from splitting (Dost 1983). Balfas (1994) brush-coated four types of end sealant (resorcinol formaldehyde, polyvinyl acetate, wood filler and cement) to wooden cat statues and found that all the treatments were effective at restricting moisture movement in the longitudinal direction. He found that polyvinyl acetate was the best sealant at reducing cracking of the statues because it adhered better to green wood. Resorcinol formaldehyde resin also performed well, but its adhesion to wet wood was poorer than that of the polyvinyl acetate (Balfas 1994).  2.4.11 Wood modification  Wood modification changes the properties of wood cell walls and tissues, and improves wood’s decay resistance without the use of biocides (Hill 2006). There are four main classes of wood modification: chemical, thermal, surface, and impregnation modification (Hill 2006). Hill (2006) gives a brief definition of each type of modification. Chemical modification is the “reaction of a chemical reagent with the wood’s structure creating a covalent bond between the reagent and the wood” (Hill 2006). Thermal modification is “the application of heat to wood to improve the performance of the wood” (Hill 2006). Surface modification is the “change of the wood’s surface with any chemical, biological or physical methods” (Hill 2006). Impregnation modification is the “filling of the wood with an inert material” (Hill 2006). Most of the chemical modifications involve reacting or cross-linking OH group in the cell wall, which gives wood new 77  properties (Hill 2006). Not all chemical modification used fall into one of the four categories of wood modification defined by Hill (2006).  2.4.11.1 Acetylation  Acetylation is most commonly a process where wood is reacted with acetic anhydride, and hydroxyl groups in wood are replaced by acetyl groups (Tarkow and Stamm 1955). Acetylation converts wood into a permanently swollen state making it more dimensionally stable (Tarkow and Stamm 1955). Acetylation reduces the checking of wood exposed to the weather (Kiguchi and Evans 1998).  For example, Tarkow and Stamm (1955) exposed acetylated and unacetylated Douglas fir plywood outdoors for two years. They found that acetylation greatly reduced face checking of the plywood. Plackett et al. (1992) found that acetylation reduced surface checking of radiata pine veneers during accelerated weathering. Dunningham et al. (1992) in related work exposed acetylated radiata pine veneers outdoors for 28 weeks. They found that acetylation reduced checking of veneers, but did not prevent micro-checking of wood cell walls (Plackett et al. 1992, Dunningham et al 1992). Imamura (1993) examined the micro-structure of acetylated Japanese cedar samples weathered outdoors for 8 weeks. He noted that acetylated wood developed fewer diagonal micro-checks of bordered pits than the untreated and weathered controls. This microscopic observation suggests how acetylation reduces check formation. Beckers et al. (1998) applied various alkyd and acrylic paints to acetylated and unmodified Scots pine panels and exposed them to 30 weeks of accelerated weathering. They found that cracking and delamination of the paint was significantly lower on acetylated samples. Akhtari (2010) examined the microstructure of acetylated and untreated Norway spruce and oriental beech (Fagus orientalis Lipsky) samples exposed to accelerated weathering. He noticed that untreated samples developed large checks along the fibers whereas acetylated samples had smaller and more dispersed checks.   78  2.4.11.2 Cross-linking, bulking and impregnation modification  There are many different types of chemicals that have been used to modify wood through cross-linking, which is a process that involves chemically joining two or more molecules by a covalent bond (Alger 1997). Some of the best examples of cross-linking treatments are those that have been applied to reconstituted wood products such as plywood and laminated veneer lumber (LVL). For example, Yusuf et al. (1995) treated radiata pine veneers with four different cross-linking agents: (1) tetraoxane; (2) glutaraldehyde (GA); (3) glyoxal, and; (4) dimethylol dihydroxy ethylene (DMDHEU). The treated veneers were glued with resorcinol resin and pressed at room temperature to form 2-ply laminate veneer lumber (LVL). They were then exposed to artificial weathering. All of the treatments reduced surface checking.   Some cross-linking treatments have recently been applied to solid wood products and have been able to reduce surface checking. Xie et al. (2008) modified flat-sawn Scots pine panels with methylated DMDHEU (mDMDHEU) and exposed treated samples outdoors for 18 months. In accord with Yusuf et al. (1995) findings, the mDMDHEU stabilized the wood and reduced checking. Xiao et al. (2012) impregnated Scots pine boards with glutaraldehyde, and exposed treated samples outdoors for 18 months. In accord with Yusuf et al. (1995) findings, the treatment reduced checking particularly when used at higher concentrations. However, not all chemical treatments applied to solid woods have been successful. For example, Donath et al. (2006) treated Scots pine with various silane chemicals and exposed them to natural or accelerated weathering. The treatments did not reduce surface checking even though the hydrophobicity of the wood was increased by the silanes. Miklečic´ and Jirouš-Rajkovic´ (2011) impregnated European beech with an aqueous solution of citric acid and sodium-hypophosphite monohydrate (SHP). Treated samples were exposed to 52 days of accelerated weathering. However, the treatments did not reduce checking.  Bulking treatments fill up the voids in wood cell walls with molecules that are small enough to penetrate the nano-capillary network in the cell wall (Hill 2006). The treatments keep wood in a swollen condition and can reduce checking. Modification of wood with polyethylene glycol (PEG) is a common bulking treatment. Stamm (1959) found that PEG-1000 79  reduced face checking of Douglas fir plywood during drying. Kim (2012) found that impregnation of western hemlock logs with low molecular PEG-1000 reduced checking of the logs during drying. This agrees with Stamm’s (1959) findings, but Furuno (2001) mentioned that PEG leaches out of wood when the treated wood was exposed outdoors.  Another chemical that penetrates the nano-capillary network of the cell wall and is used to modify wood is phenol formaldehyde (PF) resin. There are many studies of the effects of PF resins on the weather resistance of wood. PF resin is a good choice for the treatment of wood because it is inexpensive and has the following desirable properties: (1) low molecular weight resins can be synthesized that are small enough to penetrate wood cell walls; (2) the resins are soluble in polar solvents and water, and; (3) they are sufficiently polar to interact with cell wall components (Stamm and Seborg 1939 cited by Hill 2006).   Phenol formaldehyde has been applied to the face-ply of plywood to reduce checking (Stamm and Seborg 1939). Stamm and Seborg (1939) vacuum-treated Douglas fir veneer sheets with PF resin and used them to make plywood. They exposed the treated plywood and untreated controls to six months of weathering and found that the PF resin reduced face checking of the boards. PF resin has also been used to manufacture paper overlays which are bonded to plywood’s surface (Bosshard and Futo 1963, Knight and Domain 1962). These overlays protected the plywood from weathering and also reduced face checking (Bosshard and Futo 1963, Knight and Domain 1962).  Stamm (1964), cited by Lloyd and Stamm (1958) mentioned that wood modified to a weight gain of 35% with phenol formaldehyde resin had no face checking when used indoors, and the treatment greatly reduced face checking of wood exposed outdoors. Lloyd and Stamm (1958) found that an impregnation treatment with PF resin (Impreg) was better at reducing face checking of softwood plywood than a treatment that involved impregnation with PF resin followed by compression (Compreg). However, the opposite was the case for hardwood plywood. Stamm (1964) mentioned that treatment of wood with thermosetting resins such as PF makes the treated wood brittle because of the high rigidity of resin-impregnated fibers. This is one of the disadvantages of some impregnation modification treatments. 80   Sudiyani et al. (2001) treated Japanese cedar blocks with PF resin and exposed the samples to accelerated or natural weathering. They found that the treatment reduced checking especially when weight gain due to treatment was increased (10.4%, 16.7%, 20.6%).  Liu (2005) pressure treated the heartwood of plantation grown hybrids of Japanese larch and Olga bay larch (Larix olgensis A.Henry) with solutions containing different percentages of low molecular weight phenol formaldehyde resin. They exposed the treated wood to repeated wetting and drying and observed that check sizes in treated wood decreased when the level of PF resin increased. However, the effectiveness of the resin at reducing checking did not increase significantly beyond weight gains of 40%.  Cheng and Evans (2012) treated flat and profiled white spruce deck-boards with low molecular weight phenol formaldehyde resin and exposed the boards outdoors for 24 weeks (Figure 2.35). They found that the PF resin treatment reduced the number of checks by half and significantly reduced the length of checks in both flat and profiled boards.   Figure 2.35: White spruce boards weathered for 24 weeks in Vancouver: (a) untreated; (b) treated with PF resin. Scale bar = 50 mm  There have been other studies that have examined whether impregnation modification using different resins can reduce the checking of wood. Hansmann et al. (2006) impregnated (a) (b) 81  Norway spruce and black poplar (Populus nigra L.) boards with a melamine resin and exposed treated boards to accelerated weathering. They found that most of the samples developed a lot of cracks due to the increased brittleness of the modified wood. Trinh et al. (2012) exposed European beech plywood impregnated with a N-methylol melamine resin or made from N-methylol melamine resin-impregnated veneers to 18 months of weathering. They found that both treatments reduced checking of the plywood. 2.4.11.3 Thermal treatments  Dubey et al. (2010) heat treated flat-sawn radiata pine samples in an oil bath and exposed them to accelerated weathering for 2100 hours. They found that most of the heat treated and weathered boards had no surface checks. In accord with this observation, Huang et al. (2012) also found no surface checks in heat treated and weathered jack pine.  In contrast, Brischke and Rapp (2004) heat treated Pacific silver fir and Norway spruce boards at different temperatures in an oil bath and exposed boards to accelerated weathering followed by eight weeks of natural weathering. They found that samples checked during weathering. Those treated at 180◦C had the longest checks, but not the highest number, while samples heat treated at 220◦C had the fewest checks. In addition, Welzbacher et al. (2009) mentioned that thermally modified wood used as decks or flooring is more brittle and prone to develop splinters and flaking than untreated wood. Similarly, Miklečic´ et al. (2010) found that three uncoated thermally modified hardwoods, English oak, European ash (Fraxinus excelsior L.) and European beech were more susceptible to checking than their unmodified counterparts.  2.4.12 Cupping and attempts to reduce cupping  The checking of wood during outdoor exposure is one way that moisture-induced stresses are released (Stamm 1965, Evans 2008). They can also be relieved by cupping or warping (Stamm 1965, Evans 2008). However, for wood used outdoors, fasteners restrain boards from cupping, and may exacerbate checking (Stamm 1965, Evans 2008). For this reason, reducing cupping is just as important as reducing checking. On the other hand Zahora (1992) found that cupping of CCA-WR treated southern pine was greater than its CCA-treated counterpart, but the width of checks in the former were smaller. Cupping is also a problem with 82  profiled boards because profiling tends to exacerbate cupping. There is little information on the effects of profiling on the cupping of deck-boards but after striated plywood was developed it had to be redesigned to reduce cupping. This was achieved by increasing the thickness of the striated veneer to balance the structure of the panels (Bailey 1944 a,b).  2.4.12.1 Species and mechanical treatments  One way of reducing cupping is through the careful selection and preparation of wood material. Browne (1952) mentioned that the width of boards in siding should not be more than eight times the thickness to avoid excessive cupping. Virta et al. (2005) suggested that boards used for exterior cladding (siding) should be at least 28 mm thick to avoid excessive cupping. They also suggested using species that are more dimensionally stable to further reduce cupping (Virta et al. 2005).  Kerfing has been suggested as a way to reduce cupping but some research suggests that it is not effective. McMillin (1969) mentioned the use of stability kerfs (grooves) 6.35 mm deep and 6.45 mm wide cut into the back of southern pine as a way of reducing cupping. Nystrom (1995) designed a deck-board with two round stress relieved channels on the underside to reduce cupping. However, there is no experimental evidence or references in the patent to support the claim that stress relief grooves reduce cupping. On the contrary, Turner et al. (2008) cut two 5 x 5 mm kerfs (grooves) into the back-surface of flat-sawn radiata pine weather boards. Kerfed boards and unkerfed controls were exposed to wetting and drying and cupping of boards was measured. They found that the grooves, irrespective of the side they were applied to (convex and concave), did not reduce cupping (Turner et al. 2008). Despite these findings, Stubbersfield (2009) designed a deck-board with a longitudinal groove on its underside. The groove was intended to reduce cupping, but like Nystrom’s (1995) patent he presented no experimental evidence that the grooves reduced cupping. Ratu et al. (2007), however, found that kerfing reduced cupping of flat-sawn southern pine deck-boards exposed outdoors. They found that boards with 1 and 3 kerfs were better at reducing cupping than 2 kerfs possibly because the latter kerfs were near the fasteners in the boards. In addition, they 83  found that a deeper 20 mm kerf was more effective at reducing cupping than shallower kerfs (10 mm or 13 mm).  Besides kerfing there are other mechanical treatments that can also influence cupping. Morris and Ingram (2002) mentioned that almost all the incised CCA-treated species they tested cupped more than the unincised CCA-treated boards when they were exposed outdoors for 9 years. Urban and Evans (2005) mentioned that a hidden fixing (Figure 2.32b) allowed some flexing of boards, which helped relieve stresses by cupping rather than checking. 2.4.12.2 Wood modification  Some wood modification treatments are effective at reducing cupping because they make wood more dimensionally stable.  For example, Dahlen et al. (2008) mentioned that a 3% waterborne resin acid treatment reduced cupping of southern pine boards by up to 0.8 mm when boards were exposed outdoors. Xie et al. (2008) mentioned that Scots pine boards treated with mDMDHEU cupped less than untreated boards during outdoor exposure. Cheng and Evans (2012) treated white spruce boards with a low molecular weight phenol formaldehyde resin and exposed treated boards outdoors for 24 weeks. They found that the impregnation treatment reduced the cupping of boards during weathering. However, the overall cupping of the PF treated and untreated boards was the same because the PF-treated boards cupped after treatment. Akhtari and Nicholas (2014) also found that EcolifeTM was able to reduce cupping of southern pine boards exposed to accelerated weathering for 24 days.     84  2.5 Concluding remarks   Checking is one of the most important surface defects affecting the appearance of wood used as deck-boards. Checking is cited by builders as the reason why wooden decks are replaced and it explains in part why consumers are switching to other materials such as wood plastic composites. Surface checks not only give wood an unattractive surface, but they can also encourage decay and cause strength losses (Heebink 1959, Stanzl-Tschegg et al. 1996, Choi et al. 2001, Choi et al. 2003, 2004, Kurisaki 2004, Morris et al. 2004). My review of the literature shows that although methods have been developed to reduce checking none of them completely eliminate it. The most common method of reducing checking of deck-boards is the use of surface coatings or preservatives containing water repellants. However, many of the methods that are successful at reducing checking of poles, ties and structural lumber have not been thoroughly tested with deck-boards. Hence, there are opportunities to more fully evaluate other approaches to reducing the checking of decking.  Selecting species that are less prone to checking is a tried-and-tested strategy. There have been some studies that show that checking varies between different species (Morris and Ingram 1996, Yata 2001, Opoku 2007, Morris and Ingram 2002, 2010, Ratu 2009). For example, the literature suggests that western red cedar is less prone to checking when used as deck-boards (Morris and Ingram 1996, Morris and Ingram 2002, Ratu 2009).  In contrast, some other species such as western hemlock and southern pine are much more susceptible to checking (Morris and Ingram 2002, Ratu 2009). Furthermore, some studies recommend using softwoods in preference to hardwoods for exterior applications because the latter check more than softwoods (Kretschmann 2010). However, relatively few studies have examined the susceptibility of hardwood decking to checking or compared their performance with that of softwood decking. Some hardwoods such as ipe are reported to be suitable for exterior decking, in part, because they are resistant to checking (Williams et al. 2000).   Physical methods can be used to reduce checking such as kerfing, center-boring, and incising of wood products. However, few studies have examined whether they can reduce the checking of deck-boards exposed outdoors. Furthermore, when such methods have been used 85  with deck-boards they don’t appear to be very effective at eliminating surface checks (Ratu et al. 2007). Surface profiling, another physical treatment, shows greater promise as a way of reducing checking of deck-boards (McFarling and Morris 2005, McFarling and Morris 2008, McFarling et al. 2009, Evans et al. 2010, Akhtari and Nicholas 2014). To date the research on profiling of deck-boards has examined its ability to reduce checking of Pacific silver fir, lodgepole pine, subalpine fir, southern pine, and white spruce (McFarling and Morris 2005, Morris and McFarling 2008, McFarling et al. 2009, Evans et al. 2010, Cheng and Evans 2012, Akhtari and Nicholas 2014). There is some evidence from these studies that the effectiveness of profiling at reducing checking varies with species and profile type (Morris and McFarling 2008, Evans et al. 2010). Such findings suggest that the effectiveness of profiling at reducing checking could be increased by modifying profile geometry and tailoring geometry to species of interest. First, however, there is a need to measure, identify and categorize the geometry of profiles in profiled decking. Such a classification system would pave the way for studies of how profile geometry influences checking of deck-boards (and the development of more effective profiles). 86  Chapter 3: Response of Eight Hardwoods and Nine Softwoods to Natural Weathering 3.1 Introduction   Wood has a long tradition of being used for external decks but its market share for this end-use has been eroded by other deck-board products such as wood plastic composites, as mentioned in Chapter 1 (Freedonia 2014). There are many wood species used as lumber that are available in sufficient quantities to meet the demand for deck-boards, but there are certain specific attributes that such species would need to meet before they could be considered to be suitable for use as deck-boards. The characteristics of deck-boards that are important are:  decay and insect resistance, aesthetic beauty, availability, price, and low maintenance (Ganguly and Eastin 2009). The aesthetic beauty of deck-boards can be ruined by weathering and surface checking, a by-product of weathering as mentioned in Chapter 1 (Evans 2008).   My literature review (Chapter 2) shows that only a handful of studies have examined the checking of wood species used as deck-boards. Most of the studies have focused on softwood deck-boards (Morris and Ingram 1996, 2002) and the few studies on hardwoods are on species that are not native to North America (Williams et al. 2001, Opoku 2007). These studies have shown that there are pronounced differences in the checking of the different species. Therefore, I hypothesize that there will be significant differences in the checking of North American hardwoods and softwoods when they are exposed outdoors (as deck-boards) to the weather. I test this hypothesis here. The aim of this chapter is to find a species that can perform as well as western red cedar and ipe, two species that are used as deck-boards and are known to resist checking (Morris and Ingram 1996, 2002, Williams et al. 2001). I rated the species by their resistance to checking (check number and dimensions) as well as their surface appearance (erosion, fungal discolouration, colour changes).   I also examined the relationship between wood density and checking. There are studies that show that denser woods tend to develop more checks than less dense woods (Ratu 2009). The reason for this, according to Browne (1952) and the U.S. Forest Products Laboratory (FPL 87  1957), is that denser woods tend to shrink and swell more than light to medium woods causing more checking (FPL 1957, Browne 1952). I hypothesize there will be a positive correlation between density and checking.  3.2 Materials and methods 3.2.1 Experimental design  The experiment was a randomized complete block design with one fixed factor (wood species). Four different blocks were used, each block containing 17 different wood species (Figure 3.1). Each sample in a block was cut from a different board to ensure replication at the higher (block) level. Randomization of samples in each block used the design function in the statistical software Genstat 5 (VSN International 2009). Samples were fastened onto a rack, representing a block, and weathered outside for 12 months (four racks in total). The size and numbers of surface macro and micro-checks, colour and erosion of samples was measured after 12 months.  13 15 6 3 12 17 8 14 1 10 16 2 11 4 7 5 912 7 2 17 11 9 16 13 4 1 10 15 5 6 3 8 1413 14 11 10 8 15 6 17 9 4 5 1 2 12 16 3 79 10 14 12 11 16 7 8 5 13 3 6 1 4 2 17 151: Aspen2: Balsa3: Basswood4: Douglas fir5: Ipe6: Larch7: Lodgepole pine8: Pacific silver fir9: Red alder10: Red oak11: Southern pine12: Western bigleaf maple13: Western hemlock14: Western red cedar15: White oak16: White spruce17: Yellow cedarBlock 1Block 2Block 3Block 4 Figure 3.1: Randomized block design used for the experiment that examined the checking, discolouration and erosion of 17 different wood species  88  3.2.2 Wood species  Eight hardwoods and nine softwoods were used for the experiment. All of the softwoods and hardwoods, with the exception of balsa and ipe, are North American species.  To ensure that each board was cut from a different tree, several suppliers were asked to provide samples from different trees or different suppliers were used for the same species. Samples of the softwood species, maple, and aspen were donated by FPInnovations and UBC’s Malcolm Knapp research forest. The remaining hardwoods were purchased from retail outlets in Vancouver and the United States (Appendix 3).   The species tested are commonly used in exterior applications with the exception of balsa, aspen, red alder, and maple.  The most common species used as deck-boards in North American is southern pine, as mentioned in Chapter 1. Naturally durable species such as ipe and western red cedar are also commonly used for decking (Shook and Eastin 2001, Ganguly and Eastin 2009, Schulze et al. 2008). Ipe was specifically chosen for this experiment due to its high density and the premium it commands in the market place (Chudnoff 1980, Schulze et al. 2008, Borst et al. 2012). In contrast, balsa is very light and also porous (Greil 1998, Gutiérrez and Santiago 2006), and was included because of my interest in examining the relationship between wood density and checking. Table 3.1 shows the book values for the specific gravity of each of the species tested in this chapter:           89  Table 3.1: Literature values for the specific gravity of wood species exposed as deck-boards to natural weathering (Chudnoff 1980, Miles and Smith 2009) Species Scientific Name Abbreviated Name Specific Gravity Trembling aspen (Populus tremuloides Michx.) ASP 0.38 Balsa (Ochroma pyramidale (Cav. ex Lam.) Urb.) BAL 0.10-0.17 American basswood (Tilia Americana L.) BASS 0.37 Douglas fir (Pseudotsuga menziesii Franco) DFIR 0.48 Ipe (Tabebuia sp.) IPE 0.85-0.97 Western larch (Larix occidentalis Nutt.) LAR 0.52 Lodgepole pine (Pinus contorta var. latifolia Douglas) LP 0.41 Pacific silver fir (Abies amabilis Douglas ex J. Forbe) PSF 0.43 Red alder (Alnus rubra Bong.) RA 0.41 Red oak (Quercus rubra L.) RO 0.63 Southern yellow pine (Pinus sp.) SYP 0.51-0.59 Western bigleaf maple (Acer macrophyllum Pursh) BM 0.48 Western hemlock (Tsuga heterophylla Raf. Sarg.) WH 0.45 Western red cedar (Thuja plicata D. Don) WRC 0.32 White oak (Quercus alba L.) WO 0.68 White spruce (Picea glauca (Moench) Voss) WS 0.40 Yellow cedar (Cupressus nootkatensis D. Don) YC 0.44  Boards ranging in size from 520-1092 (length) x 138-178 (width) x 38-76 mm (thickness), and mainly flat-sawn were selected for this experiment. Boards were stored in conditioning room at 20 ± 1◦C and 65 ± 5 % for a minimum of three weeks, before processing. A resistance-type moisture meter RDM3 (Delmhorst Instrument Co.) was used to measure the moisture content of the samples. When the moisture content of the samples reached a minimum of 15% they were cut to length using a pendulum saw (Stromab PS 50/F) and planed to thickness with a planer (Martin T44). The moisture contents of the samples varied from 7 to 15%.                            90  3.2.3 Preliminary characterization of species and samples  The following properties of the wood species were measured or assessed prior to the weathering trial: (a) basic density; (b) growth ring width; (c) grain contrast, and; (d) grain angle. Sixty eight boards were measured (four boards for each species). Growth rings were manually counted on the end-grain of 35 mm thick samples. Heartwood and sapwood were identified visually and separated for each sample (Wiemann 2010). Grain contrast of the growth rings were categorized into: gradual/abrupt contrast of earlywood and latewood in softwoods, and diffuse/ring porous for hardwoods (Hoadley 1990, Wiemann 2010). A protractor and ruler were used to measure the grain angle of finished samples (500 (length) x 135 (width) x 35 mm (thickness)). A water displacement test was used to measure the basic density of samples cut from the parent material used to make deck-board samples (Ofori and Brentuo 2010). Samples used for density measurements ranged in size from 33-55 (length) x 20-33 (width) x 5-20 mm (thickness). The samples were soaked in distilled water for a minimum of 5 days and their volumes were measured by water displacement. They were then oven dried for a minimum of 1 day at a temperature of 100 ± 5°C, and weighed with an analytical balance (Ohaus Adventurer ARC120). Basic density of the different species is expressed as oven dry mass (g)/ green volume (cm3). Table 3.2 lists the properties of the different wood species.            91  Table 3.2: Growth rate, grain angle, basic density and range, wood type and grain contrast of wood species used for the decking test Species Average Number of Growth Rings  (rings/cm) Average Grain Slope                            (◦) Average Basic Density (g/cm3) (STD) Basic Density Range (g/cm3) Wood type  Board blocks                 Heartwood(H)/Sapwood(S) Grain Contrast 1 2 3 4 Balsa 1.6 0.6 0.14  (0.046) 0.08-0.19 H,S H H,S* H Diffuse Western red cedar 2.8 1.1 0.30 (0.039) 0.26-0.34 H H H H Abrupt Pacific silver fir 5.6 1.4 0.35 (0.042) 0.29-0.39 S S S S Gradual Basswood 5.3 1.3 0.36 (0.036) 0.31-0.40 H S S H Diffuse White spruce 7.6 0.3 0.38 (0.023) 0.36-0.41 * * * * Gradual Yellow cedar 8.3 1.3 0.39 (0.033) 0.36-0.43 H,S H H H,S* Gradual Red alder 2.4 2.5 0.39 (0.024) 0.36-0.41 H,S H H H Diffuse Trembling aspen 3.8 2.4 0.40 (0.008) 0.39-0.41 S S S S Diffuse Lodgepole pine 4.5 1.3 0.41 (0.018) 0.39-0.43 H H H H Abrupt Western hemlock 6.1 0.6 0.42 (0.022) 0.39-0.44 H H H,S H Gradual Douglas fir 4.6 3.4 0.43 (0.053) 0.39-0.51 H,S H H,S H Abrupt Western larch 5.8 1.5 0.47 (0.047) 0.43-0.54 H,S* H,S* H,S* H,S* Abrupt Southern pine 3.1 0.8 0.49 (0.017) 0.47-0.51 H,S H,S H,S H,S Abrupt Bigleaf maple 3.9 3.6 0.50 (0.045) 0.46-0.56 H,S H,S H,S S Diffuse Red oak 3.4 2.4 0.54 (0.052) 0.49-0.59 H H H H Ring Porous White oak 4.3 2.1 0.63 (0.068) 0.53-0.70 H,S H H,S H,S Ring Porous Ipe 5.5 1.3 0.87 (0.042) 0.81-0.89 H H H H Diffuse *denotes difficult to distinguish sapwood from heartwood. S* denotes sapwood located on the corners of boards.    92  3.2.4 Preparation of decking samples   Boards were cut to length using a chop saw (Stromab PS 50/F), and then planed to thickness using a rotary planar (Martin T44). They were finally ripped to width using a table saw (Martin T75). The boards were cut to a final dimension of 500 (length) x 135 (width) x 35 mm (thickness). A bench drill (Delta 161/2) was used to predrill 4 holes (Ø=3.97 mm) in the ends of the samples to prevent end splitting when they were screwed to horizontally inclined weathering racks. The holes were positioned 40 mm from end grain and 23 mm from the edges of the boards. The upper surface of each board was sanded using an 80 grit belt sander (Optimat SKO 213) and their edges were eased using an 80 grit edge-belt sander (Progress). The end grain of the samples was sealed with three coats of epoxy resin (G2 Epoxy, System Three Resins) to reduce end splitting of board samples during the exterior weathering trial which has a greater impact on short test specimens than on full length deck-boards that are used commercially. A stainless steel mask (41 (length) x 14 (width) mm in size) containing an 11 mm diameter hole was screwed to the top left corner of each board to allow erosion measurements to be made at the end of the weathering trial (see section 3.2.6.2). The finished boards were stored in a conditioning room, as above, for two weeks before they were fixed to weathering racks.  3.2.5 Weathering trial  Four deck frames measuring 3760 (length) x 460 (wide) x 495 mm (high) were constructed from pressure treated lumber (Figure 3.2a). Galvanized 63.5 mm long Robertson square head (8 mm) deck screws were used to fasten boards to the frames. Boards were placed bark side up (convex) as is recommended for deck-boards (Williams and Knaebe 1995). Untreated western hemlock deck-boards of the same size as test samples were fastened to the ends of each rack to protect the test sample’s edges from the sun. No center bracing (blocks and joists) or legs were placed inside the racks, so the samples were not rigidly fixed to the sub-frame (Hislop 2006). However, two blank pieces of western hemlock, as above, were placed in the center of each rack between the eighth and ninth samples to increase the structural stability of the rack. There was a 6.35 mm gap between each test-sample which is the minimum 93  recommended for standard deck construction to accommodate swelling of boards and prevent water pooling between them (Hislop 2006). The racks were placed outside in a secure area in the FPInnovations test site on the Vancouver campus of UBC (lat., 49.3°, long., 123.3°) (Figure 3.2b). Samples were exposed to the weather for one year from May 6th 2011 to May 8th 2012.     Figure 3.2: Weathering racks used in the exposure trial: (a) Dimensions of an individual weathering rack; (b) Photograph of some of the deck-board samples on the four weathering racks during the exposure trial 3.2.6 Measurement of checking, erosion and discolouration 3.2.6.1 Checking  Weathered samples were conditioned at 20°C ± 1°C at 65% for at least seven days prior to measurement of checking.  Checks were counted one inch from the fastener to prevent counting of checks that developed around decking screws or originated from end-grain. Checks inside knots were not counted either. The number of visible checks in each board was counted and they were numbered with a soft pencil (3B). The length and width of each check was measured for each board. Only the average length and width of the ten largest checks are presented in the results section of this chapter because most of the checks were too small to reduce the visual appeal of the deck-board samples. A transparent plexiglass (perspex) ruler and calibrated optical loupe were used to measure the length and width of checks, respectively, as described by Evans et al. (2010). (a)                                                                                   (b)  94  3.2.6.2 Erosion and micro-checking  A bandsaw (Meber SR-500) was used to cut out a 20 x 50 mm cross-section from the corners of the boards containing a steel mask to assess erosion. The central eroded area of the sub-sample that was not masked was scanned with a white light confocal profilometer (Altisurf 500) using the following parameters: 3 mm probe; scan area of 225 mm2; gauge resolution of 0.333 nm; spacing of measurements of 10 µm. To accurately measure erosion depth a waviness profile algorithm was used to eliminate large checks from being included in the erosion measurements. Images were exported as TIFF files and some are presented in this chapter. A line profile was extracted from the center of the eroded area on each sample and this profile was used to identify and then measure the dimensions of micro-checks in the eroded areas (arrowed in Figure 3.3). Micro-check width and number was measured using image analysis software (Figure 3.3) (Altimet Premium, v. 6.2.6142).   Figure 3.3: Extracted profile from a confocal profilometry line scan of an eroded white oak sample. Note the micro-checks (arrowed) in the sample  3.2.6.3 Assessment of discolouration 3.2.6.3(a) Visual observations (Microtek)  Each sample was scanned using a deck-top scanner (Microtek Scanmaker i800) using the following parameters: 600 dpi resolution; 256 Colours; 100% scale. Samples were scanned prior to being exposed outside, and again after the one year weathering trial. The images of the samples before and after weathering were compared by eye. Qualitative assessment was made 95  of the loss of visual appeal of the exposed deck-boards based on the degree of discolouration, checking and streaking.  3.2.6.3(b) Spectroscopy  The colour of each unweathered sample was measured using a Minolta spectrophotometer (Model #CM-2600d) at randomly chosen locations on the surface of samples. Measurements of colour samples after weathering returned to the same locations as those used to measure the colour of unweathered samples. The spectrophotometer uses two xenon light flashes to record two different colour parameters: SCI (specular component included) and SCE (specular component excluded) (Konica Minolta 2014). The SCI values are used here because this parameter has generally been used in previous studies of the colour of weathered wood (Kilian et al. 2010). An 8 mm diameter aperture in the spectrophotometer was used to capture the colour of samples. Auto-averaging was used to obtain average colour values of three measurements. To assess the greying of samples during weathering, colour difference was obtained for each weathered sample. Colour is expressed using the CIE 1976 L*a*b* space system in which L* represents lightness (0 = black, 100 = white), a* represents green-red hues (-60 = green, +60 = red), and b* represents blue-yellow hues (-60 = blue, +60 = yellow) (Konica Minolta 2007).  Overall colour change (ΔE) was obtained using the CIE76 formula ΔE*ab=√((L*2- L*1)2+(a*2- a*1)2+(b*2- b*1)2).   The colour change of samples during weathering was also estimated using the image analysis software Adobe Photoshop CS6. The average unweathered and weathered L, a, and b values of each species were entered into the colour picker of Adobe Photoshop to find the species’ matching colour in the program. The values obtained represent the species initial colour and its colour after weathering. The colour of the sample before weathering was used in the foreground colour picker, while the colour of the sample after weathering was used in the background colour picker. Using the foreground and background colour, a gradient tool was applied to display the contrast between the samples before and after weathering in a 35 x 40 mm square space. This visual method of assessing colour is useful for assessing the overall greying of samples during weathering.  96  3.2.6.3(c) Scanning electron microscopy  Samples were cut from the masked areas of deck-board samples in block 1. Samples were reduced in thickness to produce specimens 2 x 5-10 mm in size, each containing an eroded (unmasked) and uneroded (masked) area. Samples were cut to thickness using a custom-made precision pneumatic twin-blade saw (Coleman et al. 2009). Seventeen specimens were secured to 12 mm diameter aluminum mounting stubs using double-sided adhesive tabs (Cat.#76760). Two specimens were mounted on each stub with the last stub containing three specimens. Eight mounting stubs were prepared in total and specimens on each stub were identified by notches on the edges of the stubs. The unique shape of each specimen also helped to identify them (Figure 3.4).  Figure 3.4: Arrangement of specimens on SEM stubs  Specimens were coated with an 8 nm layer of gold and examined using a Zeiss Ultraplus field emission scanning electron microscope with an accelerating voltage of 5kV and working distances of 14.1 to 15.0 mm. Selected secondary electron images of the weathered surfaces of specimens were saved as TIFF files. Images of weathered surfaces were compared with those in 97  other papers but it is not possible to identify fungi colonizing the surfaces of specimens from the images (Feng et al. 2014, Hong et al. 2014, Sing et al. 2014, Ziglio and Gonçalves 2014). 3.2.6.3(d) Insect-induced discolouration  After the one year trial and following the aforementioned measurements, the samples were returned to the weathering racks and exposed to the weather for an additional 36 days between June 21 and July 27 2012. During this period between 10 am and 4pm the number of wasps visiting the surface of deck-board samples and the time they spent on the samples was recorded. The aim of this research was to see if the wasps preferred certain species as a source of fiber for their paper-like nests (Edwards 1980). The number of wasps that visited the surface of each specimen and the time the wasps spent on the surface of each specimen was recorded.  3.2.6.4 Statistical analyses of data and graphical summaries  Analysis of variance (ANOVA) was used to examine the effects of fixed (species) and random factors on checking, colour and erosion of deck-board samples. Statistical computation was performed using Genstat using a significance level p < 0.05. Preliminary checks were performed to test whether data conformed to ANOVA’s assumptions. The effects of species type on the following variables were analyzed: (1) number of macroscopic checks found at the surface of boards (data was transformed into square roots); (2) average length and width of the ten largest checks in specimens (data was transformed into natural logarithms); (3) L*a*b* values measured before (i) and after weathering (ii) and the overall colour change (ΔE) value derived from (i) and (ii) for specimens; (4) average erosion (data was transformed into natural logarithms); (5) average number of micro-checks in eroded areas for all the specimens; (6) average micro-check width in eroded areas in specimens. Significant results are presented as graphs with least significant difference (LSD) bars that can be used to compare whether means of species are significantly different (p<0.05).  The correlation (r2) and significance effect of density on checking were compared using the statistical graphing program, R 2.14.1 (R Development Core Team 2011). I sought correlations between density and checking for: (1) all species tested; (2) softwoods; (3) 98  hardwoods; (4) diffuse porous hardwoods. The following checking parameters were used: (1) number of macroscopic checks found at the surface of the boards (data was transformed into square roots); (2) average length and width of the ten largest checks in each specimen (data was transformed into natural logarithms). The correlation (r2) between density and erosion was also examined.   A graphical technique was used to compare the deterious effects of weathering on the appearance of the deck-boards. Five criteria were selected to rate the 17 species: 1) check number; 2) check size; 3) colour; 4) fungal discolouration; 5) erosion. Check number, check size and colour criteria used data from assessment of check number; ten largest check areas (length mm x width mm) (mm2); and overall colour change (ΔE), respectively. Fungal discolouration was estimated qualitatively based on the streaking or blotchiness of the surface (excluding that due to insect damage) and by looking at the degree of fungal colonization of weathered surfaces from SEM photos. The erosion criterion used data from erosion measurements (µm). The five criteria are plotted for each species using radar graphs (Atanassova 2010). Radar graphs are a popular method to display multivariate data sets and independent variables (Saary 2007, Atanassova 2010). For example, according to Jagadamma and Lal (2010) radar graphs are an ideal way to display categories and compare them.  Negri et al. (2005) also suggested using radar graphs to provide a summary to compare wood quality between different species of wood. Higher values in my radar graphs indicate species have fewer and smaller checks, and less erosion, fungal activity, and colour change.       99     With the exception of the fungal criterion which was based on qualitative assessment of the weathered surface of the boards and SEM images, all data was calculated from percentages using the following formula:  %10010.1)()(1  categoryinvaluespeciesHighestvaluespeciesCurrent   The highest value for the species was multiplied by 1.1 to add 10% to the lowest value of the worst performing species in the group to prevent scores of 0%. Radar graphs are separated into hardwoods and softwoods. These graphs were used to classify species into high, moderate and poorly performing species. This final qualitative classification gave a greater weighting to checking.              100  3.3 Results 3.3.1 Qualitative assessment of checking  All of the boards exposed to natural weathering developed checks. Some western hemlock, white oak, Pacific silver fir, and Douglas fir boards checked badly: their exposed surfaces contained checks that were both long and wide. Figure 3.5 shows examples of the surfaces of such boards.     Figure 3.5: Images of: (a) western hemlock; (b) white oak; (c) Pacific silver fir; (d) Douglas fir boards showing severe surface checking. Scale bars = 50 mm (a)                                                            (b) (c)                                                            (d) 101   Ipe and lodgepole pine boards contained numerous checks, but the checks were not as large as those in the species mentioned overleaf (Figure 3.6). These species were classified as having moderate surface checking because, although their checks could be as long and numerous as those in severely checked boards in Figure 3.5, they were usually thinner making them more difficult to see with the naked eye.   Figure 3.6: Images of: (a) lodgepole pine; (b) ipe boards showing moderate surface checking. Scale bars = 50 mm         (a)                                                                (b) 102   Aspen, balsa, and western red cedar developed relatively few checks and the checks that did develop tended to be small (Figure 3.7).     Figure 3.7: Images of: (a) aspen; (b) balsa; (c) western red cedar boards which developed small surface checks. Scale bars = 50 mm    (a)                                                               (b)  (c) 103   Checks often developed in the center of boards particularly in the softwoods Douglas fir, western larch, lodgepole pine, Pacific silver fir, southern pine, western red cedar, white spruce, and yellow cedar (Figure 3.8). Red alder and red oak also developed checks in the center of boards.     Figure 3.8: Images of: (a) western red cedar; (b) white spruce; (c) western larch boards showing surface checks in the center of the boards. Scale bars = 50 mm         (a)                                                (b)                                                 (c) 104   Some lodgepole pine, Pacific silver fir, western hemlock, white spruce, yellow cedar, basswood, Douglas fir, and western red cedar samples contained knots. Invariably checks tended to form around such knots (Figure 3.9). Checks also formed within knots in some of the boards.  Figure 3.9: Image of a lodgepole pine board showing surface checks around and within a knot. Scale bar = 20 mm  Red oak and white oak developed numerous medium to large-sized checks (see Figure 3.12 and 3.13). These checks appeared to form within the rays of both species (Figure 3.10), and as a result they were evenly distributed across the surface of boards. An observer could easily mistake these checks for large multiseriate rays.   Figure 3.10: Image of a red oak board showing checks that look like large multiseriate rays. Scale bar = 20 mm 105  3.3.2 Quantitative assessment of checking 3.3.2.1 Check numbers and sizes  Analysis of variance indicated that there was a significant (p>0.001) effect of species on check number and size (length and width). The difference in check numbers and sizes are depicted in Figure 3.11-3.13. Species can be identified in each graph using the abbreviated species names in Table 3.1.    Figure 3.11: Total check number in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the square root of check number. The Y2 axis contains values on the natural scale (n). The LSD bar can be used to estimate whether differences between means for the various species are statistically significant  Balsa and basswood had significantly (p<0.05) fewer checks than all other species with the exception of aspen (Figure 3.11). Red oak had significantly (p<0.05) more checks than all of the other species with the exception of white oak and western hemlock.  Southern pine, Pacific silver fir, ipe, Douglas fir, lodgepole pine, western hemlock, white oak, and red oak had significantly (p<0.05) more checks than balsa, basswood, aspen, bigleaf maple, red alder and western red cedar. 106   Figure 3.12: Average length of ten largest checks in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the natural logarithms of check length. The Y2 axis contain values on the natural scale (ex)  Checks in basswood and balsa were significantly (p<0.05) shorter than those in the other species (Figure 3.12). Checks in maple and aspen were significantly shorter (p<0.05) than those in all other species except for basswood, balsa and red alder. Checks in western hemlock were significantly longer (p<0.05) than those in red alder (Figure 3.12).    107   Figure 3.13: Average width of ten largest checks in deck-board samples. The labels on the x-axis refer to the species listed in Table 3.1. Y1-axis refers to the natural logarithms of check width. The Y2 axis contain values on a natural scale (ex)  Checks in aspen, basswood, maple, ipe, and balsa were significantly (p<0.05) narrower than those in white spruce, yellow cedar, Douglas fir, white oak, Pacific silver fir, and western hemlock (Figure 3.13). Western hemlock had significantly wider checks than all of the other species. Checks in white oak were significantly wider (p<0.05) than those in red oak. Checks in red oak were significantly (p<0.05) wider than those in aspen.            108  3.3.2.2 Correlation between density and checking  The relationship between basic density and checking of the different wood species was analyzed and graphed. Table 3.3 summarizes the significant effects and correlations between density and checking for the species exposed to natural weathering.  Table 3.3: Significant effects and correlations between density and checking of the different species exposed outdoors Experimental Factors Significance Correlation (R2) No. of checks  (Sq. root n) Width of ten largest checks  (ln mm) Length of ten largest checks  (ln mm) Number of Checks  (Sq. root n) Width of ten largest checks  (ln mm) Length of ten largest checks  (ln mm) All Species *** * ** 0.239 0.071 0.128 Softwoods NS NS NS 0.052 0.0003 0.031 Hardwoods *** NS *** 0.456 0.094 0.414 Diffuse porous Hardwoods *** ** ** 0.632 0.268 0.359 *=p<0.05; **p<0.01; ***=p<0.001; NS= not significant (p>0.05)   Regression analysis showed highly significant (p<0.001) effects of density on the number of checks in all of the different species, but the correlation coefficient was low (0.239). The correlation coefficients improved if regression analysis was performed only on the hardwoods (0.456) or diffuse porous hardwoods (0.632). There was no significant effect of or good correlation between density and checking in the softwoods. There was a significant (p<0.05) effect of density on check width in all of the different species, but the correlation coefficient was low (0.071). The correlation improved to 0.268 if regression analysis was performed only on diffuse porous hardwoods. There was no significant effect of or good correlation between density and width of checks in softwoods or all hardwoods. There was a significant (p<0.01) effect of density on check length in all of the different species, but again the correlation coefficient was low (0.128). The correlation improved when regression analysis was performed only on the hardwoods (0.414) or diffuse porous hardwoods (0.359). There was no significant effect of or good correlation between density and checking in the softwoods.    109   The relationship between density and check numbers can be seen in Figure 3.14. There are positive correlations between density and check number for all species, hardwoods and diffuse porous hardwoods. In other words higher density woods developed greater numbers of checks than lower density woods.  Figure 3.14: Total check number plotted against basic density for: (a) all wood species; (b) all hardwood species; (c) diffuse porous hardwood species. Y1-axis refers to the square root of check number (a)                                                                                  (b) (c)                                                                                          110   The relationship between density and check length can be seen in Figure 3.15. There are positive correlations between density and check length for all species, hardwoods and the diffuse porous hardwoods. In other words higher density woods develop longer checks than lower density woods.  Figure 3.15: Check length plotted against basic density for: (a) all wood species; (b) all hardwood species; (c) diffuse porous hardwood species. Y1-axis refers to the natural logarithms of check length   (a)                                                                                 (b) (c)                                                                                          111   The relationship between density and check width can be seen in Figure 3.16. There are negative correlations for species and diffuse porous hardwoods. Since the correlation for all species is not strong, the only finding worthy of mention is that higher density diffuse porous woods tended to develop thinner checks.  Figure 3.16: Check width plotted against basic density for: (a) all wood species; (b) diffuse porous hardwood species. Y1-axis refers to the natural logarithms of check width         (a)                                                                                 (b) 112  3.3.3 Discolouration of deck-board samples 3.3.3.1 Visual observations  The discolouration of wood during weathering is well known (Feist 1983). Wood exposed outdoors initially yellows or darkens, depending on species, before eventually becoming grey (Browne and Simonson 1957, Derbyshire and Miller 1981, Hon 1981). The greying is caused by colonization of the wood surface by black staining fungi (Schoeman and Dickinson 1997, Doi and Horisawa 2001).   The deck-board samples retained their natural colour during the first two months of the trial which started in May 2011. Around the third and fourth month (July to August) yellowing became noticeable in the softwoods while the hardwoods became darker. Ipe only darkened slightly.  Around the beginning of the fifth month (September) most of the samples, with the exception of the ipe and balsa, started to become grey. Ipe on the other hand, started to lighten and balsa was only slightly grey. At the end of the trial all of the samples had turned into various shades of grey. However, some samples of ipe still retained traces of their original brown colour or yellow colour. This greying of some of the species exposed to natural weathering can be seen in Figure 3.17.      113         Figure 3.17: Images of deck-board samples before and after 1 year of natural weathering: (a) balsa; (b) ipe; (c) southern pine; (d) western red cedar; (e) Douglas fir; (f) white oak. Scale bar = 50 mm        (a)                                                 (b)                                                 (c) (d)                                                 (e)                                                 (f) 114  3.3.3.2 CIE colour parameters  There was a significant effect (p<0.001) of ‘species’ on L, a, and b CIE colour values, both before and after weathering. Before weathering, as expected, ipe was significantly (p<0.05) darker than all of the other species whereas aspen (ASP) was significantly lighter than all of the other species (Figure 3.18). White oak (WO) was significantly (p<0.05) darker than red alder (RA) both before and after weathering. All wood species became darker during the weathering trial with the exception of ipe. Western red cedar (WRC) was significantly (p<0.05) lighter than all of the other species whereas aspen was darker than most of the other species.    Figure 3.18: Lightness of species before and after natural weathering      115   Unweathered ipe was significantly  (p<0.05) redder than all of the species except for western larch, western red cedar, red oak, red alder, southern pine, and Douglas fir (Figure 3.19). Unweathered aspen was significantly greener than all of the species except for balsa. All of the species became greener when they were exposed to natural weathering. In other words they lost their redness. However, at the end of the weathering trial, ipe and balsa were significantly (p<0.05) redder than all of the other species.     Figure 3.19: Redness of species: (a) before; (b) after natural weathering  (a) (b) 116   Unweathered southern pine was significantly (p<0.05) yellower than all of the species except for western larch (Figure 3.20). Unweathered balsa was significantly (p<0.05) bluer than all of the species. After weathering, balsa became bluer, but was now significantly (p<0.05) yellower than all other species except for red alder. All of the species became bluer when they were exposed to natural weathering. In other words they lost their yellow colour. However, at the end of the weathering trial balsa was significantly (p<0.05) yellower than all of the other species except for red alder. Red alder and ipe were also significantly (p<0.05) yellower than all of the remaining species except for Douglas fir.   Figure 3.20: Yellowness of species before and after natural weathering        117   Changes in the individual CIE Lab colour parameters are reflected in a significant effect (p<0.001) of species on the overall colour change of boards exposed to natural weathering (Figure 3.23). At the end of the trial all of the samples were grey, but the species exhibited different shades of grey. The colour change of ipe was significantly (p<0.05) smaller than those of all the other species. In contrast, the colour change of aspen was significantly (p<0.05) greater than those of all other species. The colour change of western red cedar was significantly (p<0.05) lower than those of all of the other softwoods. Southern pine exhibited the greatest colour change among the softwoods, and its colour change was significantly (p<0.05) greater than all of the softwoods except for lodgepole pine, white spruce, and western larch.   Figure 3.21: Colour change of species during natural weathering  Placing the average species’ L*a* and b* values in Photoshop’s colour matcher provides a visual representation of the colour change that occurred when the different species were exposed to natural weathering (Figure 3.22). Before weathering, ipe, western larch, western red cedar, red oak, red alder, southern pine, and Douglas fir were browner than the other species. After the samples were weathered all of them turned grey making it difficult to estimate colour differences. However, western red cedar appeared to have the lightest shade of grey among the species.  Balsa and basswood also appear to be a lighter shade of grey than the other species. Visually, ipe, southern pine, and white oak appear to show the largest colour 118  change while balsa and western red cedar had the lowest. Although ipe and white oak had relatively low colour change (ΔE) according to the CIE system, the loss of their red/brown colour makes the changes in their colour quite obvious.                    Figure 3.22: Change in the colour of test species during natural weathering estimated using Photoshop 119  3.3.3.3 Fungal discolouration  The greying of wood during weathering (above) is due to the delignification of wood surfaces and their colonization by dark-coloured staining fungi (Seifert 1964, Sell and Wälchli 1969, Evans 2013). In addition to such discolouration the majority of the samples developed other forms of uneven discolouration that appeared to be microbial in origin. Such discolouration was noticeable in Douglas fir, lodgepole pine, southern pine, aspen, basswood, maple, red alder, white spruce, Pacific silver fir, and yellow cedar. Images showing uneven discolouration in Douglas fir, basswood and southern pine can be seen below (Figure 3.23).    Figure 3.23: Images of weathered boards showing uneven discolouration of the surface of the boards: (a) Douglas fir; (b) basswood; (c) southern pine. Scale bar = 50 mm  (a)                                                (b)                                                 (c) 120   All of the species exposed to natural weathering had black specks on their surface that appeared to be of microbial origin, although most of the specks were difficult to see with the naked eye. This form of discolouration was most pronounced in balsa, basswood, aspen, and red alder. The images below show black specks on the surface of balsa and red alder samples (Figure 3.24).    Figure 3.24: Images of weathered boards showing black specks on the surface of the boards: (a) balsa; (b) red alder. Scale bar= 20 mm    (a)                                              (b)                                              121   Scanning electron microscopy (SEM) confirmed that fungi had colonized the surface of the weathered samples. All of the different species were examined and surface fungi were found in all species. SEM is not capable of quantifying the extent to which fungi colonized the different species, but in some species fungal colonization appeared to be particularly heavy (Figure 3.25).      Figure 3.25: SEM images showing heavy fungal colonization of the surfaces of: (a) Douglas fir; (b) Pacific silver fir; (c) southern pine; (d) western larch samples exposed to natural weathering for 1 year     (c)                                                                            (d) (a)                                                                            (b) 122   Fungi were also present on uneroded surfaces of lodgepole pine, western hemlock, balsa, western larch, southern pine, western red cedar, and aspen that were covered by the metal mask during the weathering trial. In some cases fungal colonization of uneroded surfaces (arrowed) was heavier than on the adjacent eroded surfaces (Figure 3.26).      Figure 3.26: SEM images showing heavy fungal colonization of the: (a) lodgepole pine; (b) western hemlock; (c) aspen; (d) western larch samples exposed to natural weathering for 1 year       (a)                                                                            (b) (c)                                                                            (d) 123   Fungi preferentially colonized cracks at weathered wood surfaces or the open ends of rays. Figure 3.27 shows fungal colonization of cracks and rays in ipe, white oak, red oak, and maple.     Figure 3.27: SEM images of fungi developing in the cracks at the surface of weathered: (a) ipe; (b) white oak; (c) red oak; (d) maple samples exposed to natural weathering for 1 year       (a)                                                                          (b) (c)                                                                           (d) 124   There appeared to be differences in the morphology of fungi colonizing weathered wood surfaces (Figure 3.28), but it is not possible to determine the types of fungi colonizing wood surfaces, for example, whether they are surface staining fungi, decay fungi or soft rot fungi.      Figure 3.28: SEM images of fungi which appeared to have different morphology at the surface of weathered wood: (a) red alder; (b) basswood; (c) Pacific silver fir; (d) lodgepole pine samples exposed to natural weathering for 1 year      (a)                                                                            (b) (c)                                                                             (d) 125  3.3.3.4 Insect induced discolouration  The weathered decks were visited by wasps that stripped weathered wood from the surface of the deck-board samples. Figure 3.29 shows images of wasps stripping wood from the surface of red alder and balsa samples.     Figure 3.29: Wasps stripping weathered wood from the surface of: (a) red alder; and (b) balsa deck-board samples (photos, Clement Tilloy)  This stripping of weathered wood from deck-boards gave the weathered surface a variegated or striped appearance. The appearance of 32% of the deck-boards were affected by the activities of wasps. Basswood, red alder, aspen, maple and balsa were particularly affected by wasps. Figure 3.30 shows images of basswood, red alder, aspen, and balsa boards ‘attacked’ by wasps. The appearance of red alder (Figure 3.30b) appeared to be more heavily affected by wasps than the other wood species.     (a)                                                                           (b) 126        Figure 3.30: Images of deck-board surfaces affected by wasps: (a) basswood; (b) red alder; (c) aspen; (d) balsa. Scale bar = 20 mm  The black spots mentioned above were very obvious in striped areas of weathered wood surfaces. Figure 3.31 shows such black spots at the surfaces of weathered basswood and red alder samples.   Figure 3.31: Images of deck-board surfaces showing black spots concentrated in areas that have been stripped of weathered wood by wasps: (a) basswood; (b) red alder. Scale bar = 5 mm (a)                                                                            (b) (c)                                                                            (d) (a)                                               (b) 127   The more prominent stripy appearance of basswood, red alder, aspen, maple, and balsa samples accord with observations in Figure 3.32 of the frequency that wasps visited the different samples. From this figure it is clear that wasps visited hardwood deck-boards samples more frequently than softwood samples. No wasps visited Pacific silver fir, Douglas fir, or western larch samples. In addition to their preference for certain species wasps also preferred to remove weathered wood from the edges rather than from the faces of samples.   Figure 3.32: Number of wasps visiting the surface of wooden deck-board samples  The average time that the wasps spent on the surface of the deck-boards was also recorded (Figure 3.33). Wasps tended to spend longer on wood species that they frequently visited suggesting the wasps may have preferred the weathered wood of these species for their nests.   Figure 3.33: Average time (sec) that wasps spent on the surface of wood during their visits to weathered deck-board samples 128   Surprisingly, despite the preference of wasps for white oak, the surface appearance of this species was largely unaffected by their activities (Figure 3.34a). Perhaps the wasps removed insufficient weathered wood to reveal the underlying unweathered wood. Southern pine was visited by wasps and it ranked sixth in terms of the length of time the wasps spent on its surface. However the appearance of southern pine was unaffected by the wasps except in one of the samples where there was some wasp damage adjacent to surface checks (Figure 3.34b).   Figure 3.34: Images of weathered: (a) white oak (Scale bar = 50 mm) and (b) southern pine (Scale bar = 10 mm) samples, which were largely free of damage caused by wasps 3.3.4 Erosion and micro-checking 3.3.4.1 Erosion  The wood of all species was eroded during natural weathering. Balsa and western red cedar were the most heavily eroded hardwood and softwood species, respectively (Figure 3.35b,c). Some softwoods samples had a corrugated surface because earlywood eroded more than latewood. Such an effect was most pronounced in western red cedar, western larch, (a)                                      (b) 129  western hemlock, white spruce, southern pine, lodgepole pine, and Douglas fir. However, western red cedar and western larch (Figure 3.35c,d) had the most distinct difference in erosion between earlywood and latewood. Hardwood species with diffuse porous arrangements such as maple (Figure 3.35e) eroded quite evenly. Ring porous samples like white oak (Figure 3.35f) and red oak developed roughened surfaces because numerous checks developed in the rays.    Figure 3.35: Images of masked and unmasked areas of samples exposed to natural weathering for 1 year: (a) trembling aspen; (b) balsa; (c) western red cedar; (d) western larch; (e) Douglas fir; (f) white oak   (a) (b) (d) (c) (e)                                                                                   (f)  130   As expected there was a significant (p<0.001) effect of ‘species’ on the erosion of deck-board samples during weathering (Figure 3.36). Balsa eroded significantly (p<0.05) more than all of the other species, followed by red alder and western red cedar. The erosion of the latter two species was significantly (p<0.05) greater than all of the other species except for basswood, Pacific silver fir, aspen, red oak, and Douglas fir.  Figure 3.36: Erosion of test species exposed to natural weathering for 1 year  There is a negative relationship between erosion and basic density (Figure 3.37). The denser species tended to erode less than the lower density species. Accordingly, there was a significant (p<0.001) correlation between density and erosion, but the correlation coefficient was quite low (R2=0.3518). The graph below suggests that the relationship between erosion and density is not linear possibly because the erosion of balsa is relatively high.   (e) (f) 131   Figure 3.37: Linear regression of basic density versus erosion of test species exposed to natural weathering for one year 3.3.4.2 Micro-checking  The eroded areas at deck-board surfaces contained micro-checks. These micro-checks were not visible to the naked eye but they were counted and their dimensions measured. There were significant effects of ‘species’ on check width (p<0.023) and number (p<0.001), but not check depth (p<0.075). Red oak had significantly (p<0.05) more micro-checks than all of the species except for ipe and aspen (Figure 3.38). Western red cedar had significantly less (p<0.05) micro-checks than red alder, white oak, Pacific silver fir, yellow cedar, maple, ipe, aspen, and red oak.  132   Figure 3.38: Average numbers of micro-checks in the eroded areas of different wood species exposed to natural weathering for 1 year  Micro-checks were significantly (p<0.05) wider in red oak than in all of the other species except for ipe, aspen, white oak, Pacific silver fir, and Douglas fir (Figure 3.39). Micro-checks in western red cedar were significantly (p<0.05) narrower than those in ipe, aspen, white oak, Pacific silver fir, Douglas fir, and red oak. Micro-checks in Douglas fir were significantly (p<0.05) wider than those in balsa, yellow cedar, white spruce, southern pine, basswood, lodgepole pine, and western red cedar.  133   Figure 3.39: Average widths of micro-checks in the eroded areas of the different wood species exposed to natural weathering for 1 year  There was not a strong relationship between micro-checking (numbers and width) and corresponding macroscopic checking that developed in weathered deck-board specimens, except for balsa and red oak which showed low and high numbers of micro and macro-checks, respectively.          134  3.3.5 General appearance of deck-boards  Numerical ratings for checking, discolouration, fungal colonization and erosion for each species can be found in the table below (Table 3.4). The highest and lowest value for each category is shown in green and red colour respectively. A high rating indicates that the species performed well in a particular criterion.  Table 3.4: Species ratings for different appearance criteria Species Softwood/ Hardwood Check Number Check Size Colour Fungi Erosion Mean Score RO Hardwood 9.1% 89.8% 30.2% 50.0% 79.4% 51.7% WO Hardwood 24.8% 72.2% 44.5% 30.0% 89.3% 52.2% BAL Hardwood 99.2% 97.1% 44.0% 50.0% 9.1% 59.9% RA Hardwood 86.6% 92.6% 47.3% 20.0% 69.8% 63.2% ASP Hardwood 90.5% 99.1% 9.1% 50.0% 78.5% 65.4% BM Hardwood 88.3% 98.3% 34.4% 40.0% 87.6% 69.7% BASS Hardwood 98.6% 97.0% 39.2% 40.0% 81.9% 71.3% IPE Hardwood 58.6% 94.2% 65.5% 65.0% 87.3% 74.1% WH Softwood 39.7% 9.1% 29.6% 35.0% 83.4% 39.4% DFIR Softwood 50.8% 40.7% 34.4% 30.0% 77.0% 46.6% PSF Softwood 59.9% 52.6% 31.2% 10.0% 80.8% 46.9% SYP Softwood 61.9% 82.1% 20.7% 35.0% 86.9% 57.3% LP Softwood 47.8% 90.7% 21.5% 50.0% 87.3% 59.5% LAR Softwood 75.1% 79.3% 28.4% 40.0% 88.1% 62.2% WS Softwood 73.0% 81.5% 22.5% 55.0% 82.4% 62.9% YC Softwood 69.0% 65.1% 31.5% 70.0% 86.5% 64.4% WRC Softwood 85.7% 90.5% 50.2% 75.0% 72.5% 74.8%  By combining appearance criteria and plotting them in radar graphs it is possible to classify softwoods and hardwood species into broad categories indicating their suitability for decking. Using Ganguly and Eastins’ (2009) method for averaging their Likert-like scores on deck-board materials within each category, each species was given a mean score based on their five ratings. The average score of all of the species was 60.1%. Species that had a mean score below average were defined as poor and moderate. Species that were above 60.1% were defined as good.  135   Southern pine, Douglas fir, lodgepole pine, Pacific silver fir and western hemlock are classified as moderate to poor (Figure 3.40). The latter two species were classified as poor. Western hemlock had the lowest rating for check size whereas Pacific silver fir had the lowest rating for fungal discolouration. Douglas fir also checked quite badly and it was not resistant to fungal discolouration. Lodgepole pine had small checks and a reasonable fungal discolouration rating but its checks were numerous and its change in colour change was high. Southern pine scored well in terms of checking but its colour change  and resistance to fungal staining was poor.  Figure 3.40: Graphical classification of softwoods into moderate to poor categories defining their suitability for use as decking    136   For the hardwoods, only red and white oak were classified as performing poorly (Figure 3.41). Red oak developed the most checks of all the species, although they were quite small. White oak performed poorly overall, but it was the least eroded of all the species. On the other hand, balsa had almost no checking, but it eroded very badly and for obvious reasons it is not suitable for decking (the surface is very soft and can be easily dented or roughened).   Figure 3.41: Graphical classification of hardwoods into moderate to poor categories defining their suitability for use as decking      137   Western red cedar, yellow cedar, white spruce, and western larch were classified as ‘good’ softwood species in terms of the performance criteria used to assess the suitability of wood species for use as decks (Figure 3.42). White spruce scored well, with the exception of its colour rating, which was similar to that of lodgepole pine. However, white spruce rated better in terms of check number than lodgepole pine. Western red cedar seems to be superior to yellow cedar and it developed a nice silver colour when it weathered.  However, its erosion rating is lower than those of yellow cedar, white spruce, or western larch. Both red and yellow cedar had the best fungal staining ratings among all the species tested here. Western larch barely scored above the mean score on average. Western larch scored fairly well in checking, but its colour change  and resistance to fungal staining, like southern pine, were fairly poor.   Figure 3.42: Graphical classification of softwoods into good categories defining their suitability for use as decking  138   Aspen, bigleaf maple, ipe, basswood, and red alder were rated as good (Figure 3.43). Aspen had a poor rating for colour change, but it scored better in the other categories. Ipe had the most checks within the group of better performing hardwoods, but it developed numerous long and narrow checks. Ipe unlike, most of the hardwoods in this category, scored fairly high in terms of fungal staining even though SEM images showed extensive fungal colonization of the micro-checks that developed at its weathered surface. However, scanned images showed little evidence of uneven discolouration at the surface of ipe samples. In addition, ipe had a good colour change rating, and its dark grey colour masked checks and fungal discolouration. Red alder barely scored above the mean score average. Red alder scored fairly well in terms of check size and number, but there was significant wasp and fungi discolouration to the surface of the red alder deck-boards.  Figure 3.43: Graphical classification of hardwoods into good categories defining their suitability for use as decking  139  3.4 Discussion  The overall aim of this thesis as described in Chapter 1 and mentioned in the introduction is to identify approaches to improving the performance of wooden decking so that it can compete more effectively with wood plastic composites, which check and discolour less than wooden decking when exposed outdoors. One approach to achieving this aim is to manufacture decking from wood species that are less susceptible to checking and discolouration than species that are currently used for decking. In this chapter I exposed deck-boards manufactured from 17 wood species and examined their resistance to checking, discolouration and erosion. Among the species that were tested were ones that are reputed to be good at resisting checking such as western red cedar and ipe and others that have a poor reputation (southern pine and western hemlock). I also tested some softwoods and hardwoods that are not traditionally used for decking. Even though I tested 17 species the number of potential wood species that could be used to manufacture decking is much larger so I sought to find whether there is a relationship between wood density and checking, which might allow the identification of other species as potential candidates for use as decking. I hypothesized that there would be significant differences between the species in their resistance to checking and also discolouration.  My results confirm this hypothesis. First, the species that have a good reputation in the market place such as western red cedar and ipe performed better than most of the other species. Overall, western red cedar performed the best out of the softwood species that were tested. It was less susceptible to checking and discolouration than the other softwoods although, due to its low density, it was more easily eroded than all of the other softwood species. Morris and Ingram (2002) also found that western red cedar performed better than southern pine, western hemlock, lodgepole pine, and western spruce during their tests on the weathering of Canadian decking species. Western red cedar developed fewer and smaller checks than southern pine, a finding that accords with those of Ratu (2009). Western red cedar has a reputation for “weathering to a pleasing silvery grey colour” (Stirling and Morris 2012). My results showed that western red cedar weathered to a lighter grey colour than the other 140  species, which lends some support to the former statement. Ipe was one of the hardwoods that performed well. It had the smallest colour change of all species, and some samples had not gone completely grey even after they were exposed outdoors for a year. Ipe had numerous checks, but they were quite thin (average width: 1.07 mm). Its micro-checks were colonized by fungi, but microbial discolouration of ipe was lower than that of most of the other species. Furthermore, its dark colour masked some of the discolouration and also the thinner checks that developed. Findlay (1938) noted that darker coloured woods may mask surface discolouration caused by staining fungi. The good performance of ipe here accords with previous studies that have tested its weathering and microbial resistance (Miller et al. 2003, Williams et al. 2001, Izekor and Fuwape 2010).  None of the species tested in this chapter exceeded western red cedar and ipe in terms of their performance, but a number of them were categorized as performing well. The softwoods that performed well were yellow cedar, white spruce, and western larch. The hardwoods that performed well were aspen, bigleaf maple, basswood, and red alder. Yellow cedar had good resistance to fungal staining, but its resistance to checking was not the best among the softwoods in this category. Its ability to resist surface checking needs to be improved. Furthermore, yellow cedar is not as readily available as western red cedar (Ganguly and Eastin 2009). More than two thirds of the home builders that were surveyed by Ganguly and Eastin (2009) indicated that they have never used yellow cedar to build decks. In addition, yellow cedar is grown at higher elevations making logging expensive and difficult (Malpass 1991). Hence, yellow cedar is not as readily available in the marketplace as western red cedar. White spruce ranked as well as yellow cedar in all categories except for discolouration. As white spruce is not durable, unlike western red and yellow cedar, it would need to be treated and finished before being used as decking, both of which would tend to improve its resistance to discolouration. White spruce and yellow cedar checked more than western red cedar and both species may benefit from physical methods such as surface profiling to reduce the size and number of checks that develop when they are exposed to natural weathering. Similarly, the suitability of lodgepole pine and western larch for decking, which were the softwoods that performed the best after the two cedar species and white spruce, would also be enhanced by 141  measures to increase their resistance to surface checking and staining. The hardwoods aspen, balsa, bigleaf maple, and basswood ranked well in terms of their resistance to checking, but their staining resistance was poor. Balsa is obviously too soft for decking although it could be hardened by impregnating it with polymer. Both balsa and basswood are in demand for specialty uses and supplies of them are too limited for their use as commercial decking. However, bigleaf maple and aspen are more plentiful and these species could be suitable for decking if treated and finished to improve their resistance to decay and staining.   My results also confirmed findings in the literature about the poor performance of some softwoods and hardwoods exposed to natural weathering. The poor resistance of western hemlock and southern pine to checking is well known (Morris and Ingram 2002, Ratu 2009), and my results also show that these species check badly. Morris and Ingram (1996) found that Pacific silver fir checked less than hemlock during weathering and my results concur with their findings.   Oak is used for decking, particularly in Europe (Opoku 2007), but previous studies of decking exposed outdoors have shown that oak checks badly. For example, Opoku (2007) mentioned that European oak exposed outdoors for six months had the longest and most checks of the seven species he tested. Furthermore, tests on the resistance of hardwood railway ties to checking have also shown that oak is very susceptible to checking. For example, Burt (1955) mentioned that 48.8% of 30,000 oak ties removed from service was caused by splits and checks, while Arnold et al. (1950) noted red oak ties checked more than beech and hard maple during seasoning.  My findings accord with this previous research. Checks at wood surfaces develop in the rays (Schniewind 1963) and this was particularly apparent with the two oak species tested here. Rays develop tensile stresses when wood surfaces absorbs moisture and these stresses initiate cracks (Ribarits and Evans 2010).  The presence of large and numerous ray ends at the tangential surfaces of deck-boards probably explain why red oak and white oak developed long and numerous checks. The difference in density of earlywood and latewood in oak as a result of their ring porosity may also have caused differential stress to occur at wood surfaces that encouraged checks to form (Schniewind 1959).  142   The influence of oaks’ anatomical structure on checking reduced the positive correlation between density and checking because these oaks checked more severely than predicted by the regression of density and check parameters. However, the correlation between density and checking of diffuse porous woods was stronger. Woods with higher specific gravity tend to shrink more across the grain compared with lighter ones which may be the reason why the denser diffuse porous hardwoods developed more and longer checks (FPL 1957). Density was not a good predictor of checking of softwoods, suggesting (like oak) that structural features, possibly the percentages of rays and differences in earlywood and latewood density may play an important role in addition to density in encouraging the formation of checks.   In addition to discussing the main aims of this chapter there are a number of other findings worthy of discussion: (1) discrepancies between my findings and results or comments in the literature; (2) activities of wasps at weathered wood surfaces.   Flaete et al. (2000) found that aspen had shorter cracks than Norway spruce during accelerated weathering, but they were more numerous. I found that check number and check length for trembling aspen were both less than in white spruce The reason for this discrepancy might be associated with species differences (Norway spruce v. white spruce) or the mode of weathering employed by Flaete et al. (2000) and that used here. Maclean (1956) mentioned that hardwoods are more likely to develop more pronounced checks than softwoods of similar properties. Hardwoods are known for their excessive checking in poles according to Kretschmann (2010). However, the resistance of lower density hardwoods to checking in this study was comparable to or greater than that of most of the softwoods I tested. Bradner (1932) mentioned that western larch was more prone to splits and checks than southern pine and Douglas fir. However, the western larch tested here performed better in terms of check number (larch: 61, southern pine: 94, Douglas fir: 85), and length (western larch: 79 mm, southern pine: 100 mm, Douglas fir: 85 mm), than Douglas fir and southern pine. Checks were also narrower in western larch (1.23 mm) than those in Douglas fir (1.44 mm). It has been suggested that the formation of smaller checks can relieve and prevent larger checks from forming (Lepitre and Mariaux 1965, Morris et al. 1990, Plackett et al. 1992). This trend was not seen in most of the 143  species tested here. For example, red oak had numerous micro-checks, but it developed numerous and large macro-checks. The wide micro-checks in red oak may develop within the voids of rays after they have been degraded by UV light and water (Evans 1989).  The inverse relationship between wood density and erosion of wood exposed to natural weathering was first described by Feist and co-workers (Feist and Mraz 1978, Sell and Feist 1986). My results accord with their findings because denser species such as ipe, southern pine, and white oak eroded less than lower density species. In addition, as expected, lower density species such as western red cedar eroded far more than southern pine, a much denser species, in accord with Feist and Mraz’s (1978) observations that western red cedar with a density of 0.30 g/cm3, eroded more than ponderosa pine (0.55 g/cm3). Sell and Feist (1986) noted that wood samples will erode more/less exponentially if the basic density is less than 0.3 or higher than 1. My results also accord with this suggestion because balsa with an average basic density of 0.143 g/cm3 eroded up to five times as much as basswood which had an average basic density of 0.356 g/cm3. However, some of my results do not follow the expected relationship between density and erosion. Hardwoods here eroded more than softwoods with similar density (yellow cedar, white spruce, red alder and aspen). Williams et al. (2001) mentioned that vessel size plays a role in erosion and that ring porous species erode faster due to their large vessels and air space (Williams et al. 2001). Red oak eroded significantly more than maple even though their densities were similar. However, this was not the case for white oak which eroded the least of all the species including ipe. White oak is known to possess an abundant amount of tyloses which are membranous outgrowths that form in the vessels and reduce wood permeability (Hoadley 1990). The lower permeability of white oak compared to red oak may explain why it eroded less than red oak. The presence of vessels in hardwoods may explain why hardwood of similar density to softwoods eroded more than the softwoods. Accordingly, the ring porosity and permeability of red oak here may explain why it eroded slightly more than some hardwoods which were less dense.   Weathered wood attracts wasps that remove wood fragments and use them to build their nests (Edwards 1980). The removal of wood fragments exposes the underlying 144  unweathered wood and causes the wood to develop a streaky appearance (Uzunovic et al. 2008). My finding confirms that wasps frequently visit weathered wood surfaces. In addition my results indicate that wasps show preference for certain wood species as a source of ‘fiber’ for their nest-building activities.  The wasps when given a choice appeared to show a preference for hardwoods that had a white weathered appearance, rather than softwoods. Some wasps have a preference for different coloured materials, such as ones with a reddish, or yellowish hue and some prefer wood with a light grey colour for their nests (Edwards 1980). Differences in the colour of the weathered woods may have influenced wasps’ species preferences as three species that were favoured by the wasps, balsa, red alder, and basswood, weathered to a light grey colour. They are all low density woods but density does not appear to be the only factor affecting wasps’ species preferences because they preferred oak over other species and spent just as much time residing on ipe as balsa (Figure 3.33). Balsa and red alder had spots of black staining fungi on the streaky surface created when wasps stripped weathered wood from deck-board surfaces. To my knowledge this interactive effect of insects and the microbial colonization of weathered wood has not been reported previously. 3.5 Conclusions  The aim of this study was to determine which wood species checked less and were best at maintaining their appearance when used for outdoor decking, and to test the hypothesis that density influences checking. I conclude that woods that are favoured for use as decking such as western red cedar, yellow cedar, and ipe performed well in terms of their resistance to checking, discolouration and erosion. However, certain species of hardwoods that are not traditionally used as decking such as aspen, basswood, and big leaf maple also performed well. Further research should be performed to further examine the suitability of these species for decking, which would include the use of preservatives, or stains to reduce their susceptibility to discolouration. Some softwoods that are not traditionally used for decking such as white spruce and western larch also performed well, but would require physical treatments such as profiling and possibly kerfing to reduce their susceptibility to checking.  145   I found a positive correlation between wood density and the numbers and length of checks that developed in diffuse porous woods exposed to the weather (0.632, 0.359), but the correlation between density and check width was not as strong (0.268). For example, ipe which was the densest hardwood developed narrow checks. Nevertheless, there may be merit in testing the suitability of diffuse porous hardwoods with similar densities to aspen and big leaf maple for use as decking. These species could include paper birch (Betula papyrifera Marshall), black cherry (Prunus serotine Ehrh.), southern magnolia (Magnolia grandiflora L.), sweetgum (Liquidambar styraciflua L.), yellow poplar (Liriodendron tulipifera L.) and black cottonwood (Populus trichocarpa Torr. and A.Gray) which are available as lumber and underutilized (Alden 1995, Wiemann 2010).146  Chapter 4: Use of Confocal Profilometry to Describe, Classify, and Identify Profiled Decking 4.1 Introduction  Consumers have been choosing wood plastic composites over solid wood decking due in part to the surface checking that develops at wooden deck-board surfaces during weathering, as mentioned in Chapters 1 and 2 ( Ganguly and Eastin 2009, Freedonia Group 2014a). To overcome this problem, wood can be treated with preservatives that contain hydrophobes (waxes and oils) or naturally durable species that are less susceptible to checking can be used in preference to treated wood (Zahora 1992, Morris and Ingram 2002). Chemically treated woods are perceived as being environmentally unfriendly in some countries, and some naturally durable wood species are sourced from tropical forests that are not sustainably managed (Schulze et al. 2008). Their use for decking is unacceptable. Alternative methods of reducing the checking of decking are physical modifications such as surface profiling or incising (Morris et al. 1990, McFarling and Morris 2005, Evans et al. 2010). Such methods have shown promise in hiding and reducing the size and number of checks that develop at wood surfaces exposed outdoors (Morris et al. 1990, McFarling and Morris 2005, Evans et al. 2010). However, surface profiling was not originally developed to reduce checking, but to create an anti-slip surface. For example, some profiled deck-boards are used in combination with a rough inorganic cement to produce a surface with high friction (Hill and Moss 2000, Hislop 2006, GripDeck 2013). There are many types of commercial profiles, but little information is available on which surface profiles are the most effective at reducing checking with the exception of the research by Evans et al. (2010), which showed that a rib profile was significantly better at reducing the checking of Pacific silver fir than a ripple profile.   An experiment, which is described in the next chapter, was designed to manufacture and expose different types of profiled boards outdoors.  This experiment builds on the previous research by McFarling et al. (2009) and Evans et al. (2010) and more closely examines the effects of profile geometry on the checking of profiled deck-boards. However, preliminary work 147  was needed to better understand the types of profiles that are being used commercially to modify the surface of deck-boards. Twenty eight different commercial and experimental profiled deck-board samples were collected from various parts of the world. Each profiled sample was scanned using confocal profilometry and the dimensions and geometry of the different profiles were measured using confocal profilometry. The research sought to develop a better method of classifying profiled decking than the subjective classification that has been used to-date. This preliminary research is described in this chapter.   4.2 Materials and methods 4.2.1 Commercially manufactured profiled decking samples  Deck-board samples of various dimensions (20 to 150 mm in length, 50 to 140 mm wide and 18 to 47 mm thick) from commercially profiled timber decking were obtained from seven countries (Table 4.1). Most of the profiled decking samples came from Australia with nine profiles, followed by Canada with seven profiles. Twenty softwood and eight hardwood decking samples were collected. The density of each sample was measured by water displacement and oven drying (overnight at 105 ± 5°C) (Ofori and Brentuo 2010). The density of samples ranged from 0.31 to 0.94 g/cm3. The growth ring number, growth ring width, and grain angle of each sample was measured using a protractor and ruler (Evans et al. 2010). The width and thickness of samples and kerf dimensions were measured using a digital caliper (Mitutoya CD-12”CP).    The species identity of the commercial decking samples was confirmed using light microscopy of planar sections cut from the samples. 10 mm x 10 mm samples were cut from each sample and soaked in distilled water for 5 days. Radial, tangential, and transverse sections (20 µm thick) were cut from samples that were held securely on the stage of a microtome. A microtome blade holder containing a disposable blade (Type S35, Feather Safety Razor Co.) was used to cut sections. Sections were briefly dipped in fresh (industrial grade) ethanol to dehydrate them and they were then transferred into a saturated solution of safranin (BDH Chemical Ltd) in ethanol. After two days stained sections were washed with toluene (Fisher Scientific). A droplet of dibutyl phthalate xylene (DPX) mountant (Fluka Analytical) was applied 148  to glass slides (76 mm x 26 mm x 1 mm VWR Micro Slides) and the sections were placed in the DPX and then covered with a glass cover slip (22 mm x 40 mm x 0.20 mm Fisher Finest Premium Cover Glass, Fisher Scientific). Prepared slides were dried at room temperature for one day. The sections were examined using a light microscope (Carl Zeiss) at various magnifications, and their microscopic wood anatomy was described and used to identify parent wood samples to the generic level.  Table 4.1: Origins and density of the collected profiled, commercial and experimental decking samples No. Species Latin Name Country/Region of Origin Country of Use Density (g/cm3) 1 Douglas fir P. menziesii Canada Canada 0.47 2 Spruce Picea sp. Norway United Kingdom 0.31 3 Merbau Intsia sp.  Southeast Asia Australia 0.67 4 Radiata pine P.radiata  Australia Australia 0.46 5 Merbau Intsia sp.  Southeast Asia Australia 0.75 6 Radiata pine P. radiata  Australia Australia 0.49 7 Radiata pine P. radiata Australia Australia 0.37 8 Lauan Shorea sp. Southeast Asia Japan 0.81 9 Pine Pinus sp. Australia Australia 0.54 10 Mangium A. mangium  Southeast Asia Japan 0.78 11 Larch Larix sp.  Germany Germany 0.53 12 Ipe Handroanthus sp.  South America Germany 0.92 13 Scots pine P. sylvestris Finland Finland 0.35 14 Radiata pine P.radiata  Australia Australia 0.33 15 Ipe Handroanthus sp. South America Japan 0.94 16 Pine Pinus sp. Finland Finland 0.41 17 Radiata pine P. radiata New Zealand Netherlands 0.54 18 Pacific silver fir A. amabilis Canada Canada 0.33 19 Merbau Intsia sp.  Southeast Asia Australia 0.67 20 Pacific silver fir A. amabilis Canada Canada 0.47 21 Pine Pinus sp.  Canada Canada 0.44 22 Cypress pine Callitris sp.  Australia Japan 0.60 23 Radiata pine P.radiata  Australia Australia 0.44 24 Spruce Picea sp. Germany Germany 0.47 25 Pine Pinus sp.  Canada Canada 0.50 26 Grapia Apuleia sp. South America Germany 0.82 27 Pine Pinus sp.  Canada Canada 0.41 28 Hemlock Tsuga sp.  Canada Canada 0.49  The end grain of collected samples was sanded with a belt sander and 150 grit abrasive paper (Progress). The end grain of sanded samples was scanned at a resolution of 1200 dpi to better see the growth ring orientations and the geometry of each sample’s profile (Microtek Scanmaker i800). These images are shown in the results and discussion section accompanied by 149  a description of the profiled sample. Country flags indicate the country where the deck-board was manufactured followed by symbols to identify softwoods or hardwoods. Most of the natural durability ratings for species are derived from Scheffer and Morrell (1998), and refer to both laboratory and field-tested durability. For three species, Pacific silver fir, apuleia and mangium, durability data was obtained from other sources (Greenwood and Tainter 1980, Alden 1997, Suprapti 2010). The durability of the species are ranked as very durable, durable, moderately durable, and non-durable. Different types of physical modifications to the boards are summarized in Figure 4.1. Profiles with an ‘eased edge top and underside’ (Figure 4.1) are sub-divided into four categories: slight, moderate, heavy rounded, and heavy straight.    Figure 4.1: Different morphological elements of deck-board samples  Close-up images of the geometry of profiles were taken at 3x magnification using a Canon EOS 7D camera equipped with a Canon MP-E 65 mm 1 to 5 x macro lens. Samples were located 50 mm away from the camera. The camera was attached to a modified vertical sliding stand (Polaroid MP4 Land Camera Stand).      150  4.2.2 Confocal profilometry of profile dimensions  Profiled samples were placed individually on the stage of a confocal profilometer (Figure 4.2) (Altisurf 500). A 37 mm long line scan, running at right angles to the direction of the grooves and peaks was obtained. This line scan contained topographical information on the profiles (Figure 4.3).   Figure 4.2: Altisurf 500 confocal profilometer used to characterize the geometry of profiles in commercial deck-board samples  The confocal profilometry parameters used when scanning samples were: 3 mm probe at a gauge resolution of 0.333 nm and a spacing of 10 µm. Topographical information for each profile was imported into image analysis software to obtain the dimensions of the profiles (Figure 4.3)(Altimet Premium, v. 6.2.6142).     151   Figure 4.3: Images of profile geometry obtained using confocal profilometry of deck-board samples: (a) height and width of peaks; (b) groove and peak radii  Four parameters were obtained for each profile: the height, width, and groove (R1) and peak radii (R2) (Figure 4.4). Depending on the width of the profiles, 5 to 10 measurements were obtained for each sample and averaged.    Figure 4.4: Parameters of the profiles that were measured using image analysis 4.3 Results and discussion 4.3.1 Visual characteristics of the profiles  This section starts with a description of each sample that was characterized using confocal profilometry. (a)                                                                                    (b)                                                                152   Profile Description               1             (a) Species, Douglas fir  (b) Width, 133 mm (c) Thickness, 38 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 8 (f) Bottom kerf: Depth, 4 mm; Max width, 11 mm; Min width, 4 mm (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section: Edge, 14 mm (j) Treatment, Heat-treated (k) Natural durability class, Moderately durable              2               (a) Species, Spruce (b) Width, 142 mm (c) Thickness, 26 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 8 (f) Bottom kerf: Depth, 3 mm; Max width, 14 mm; Min width, 3 mm (g) Eased edge top, Slight (h) Eased edge bottom, Heavy straight (i) Unprofiled section: Edge, 4 mm; Center, 14 mm (j) Treatment, ACQ (k) Natural durability class, Non-durable               3             (a) Species, Merbau (b) Width, 90 mm (c) Thickness, 19 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section: Edge, 10 mm (j) Treatment, No (k) Natural durability class, Very durable                4             (a) Species, Radiata pine (b) Width, 90 mm (c) Thickness, 22 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 1 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Non-durable   Softwood =        Hardwood =   153    Profile Description              5               (a) Species, Merbau (b) Width, 90 mm (c) Thickness, 20 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 8 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section: Edge, 10 mm (j) Treatment, No (k) Natural durability class, Very durable               6             (a) Species, Radiata pine (b) Width, 90 mm (c) Thickness, 22 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 2 (f) Bottom kerf, No (g) Eased edge top, No (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Non-durable               7             (a) Species, Radiata pine (b) Width, 69 mm (c) Thickness, 21 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 2 (f) Bottom kerf, No (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, ACQ (k) Natural durability class, Non-durable               8             (a) Species, Lauan (b) Width, 87 mm (c) Thickness, 20 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Moderately durable   154   Profile Description               9             (a) Species, Pine (b) Width, 71 mm partial width (c) Thickness, 21 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 2 (f) Bottom kerf, No (g) Eased edge top, No (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, ACQ (k) Natural durability class, Non-durable              10              (a) Species, Mangium (b) Width, 143 mm (c) Thickness, 20 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm: Board 1, 5; Board 2, 7; Board 3, 1 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, Edge glued laminate (k) Natural durability class, Durable              11              (a) Species, Larch  (b) Width, 143 mm (c) Thickness, 27 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 6 (f) Bottom kerf: Depth, 4 mm; Max width, 14 mm; Min width, 4 mm (g) Eased edge top, Slight (h) Eased edge bottom, Heavy straight (i) Unprofiled section: Edge, 11 mm (j) Treatment, No (k) Natural durability class, Moderately durable              12              (a) Species, Ipe (b) Width, 139 mm (c) Thickness, 20 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 14 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section: Edge, 11 mm (j) Treatment, No (k) Natural durability class, Very durable  155   Profile Description            13              (a) Species, Scots pine (b) Width, 117 mm (c) Thickness, 26 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section: Edge, 11 mm (j) Treatment, Heat-treated (k) Natural durability class, Non-durable              14              (a) Species, Radiata pine (b) Width, 89 mm (c) Thickness, 21 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 1 (f) Bottom kerf, No (g) Eased edge top, No (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Non-durable              15              (a) Species, Ipe (b) Width, 101 mm (c) Thickness, 20 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf, No (g) Eased edge top, No (h) Eased edge bottom, Slight (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Very durable              16              (a) Species, Pine (b) Width, 131 mm (c) Thickness, 26 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 2 (f) Bottom kerf: Depth, 2 mm; Max width, 13 mm; Min width, 1 mm (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section: Edge, 22 mm (j) Treatment, Heat-treated (k) Natural durability class, Non-durable   156   Profile Description              17              (a) Species, Radiata pine (b) Width, 140 mm (c) Thickness, 26 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf: Depth, 4 mm; Max width, 13 mm; Min width, 4 mm (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section: Edge, 9 mm (j) Treatment, Acetylated (k) Natural durability class, Non-durable            18              (a) Species, Pacific silver fir (b) Width, 50 mm partial width (c) Thickness, 23 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 16 (f) Bottom kerf, No (g) Eased edge top: Profile along eased edge (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Non-durable              19              (a) Species, Merbau (b) Width, 90 mm (c) Thickness, 19 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 16 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section, 10 mm (j) Treatment, No (k) Natural durability class, Very durable               20               (a) Species, Pacific silver fir (b) Width, 135 mm (c) Thickness, 27 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 3 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section, 16 mm (j) Treatment, ACQ (k) Natural durability class, Non-durable  157   Profile Description            21              (a) Species, Pine (b) Width, 50 mm partial width (c) Thickness, 23 mm (d) Growth ring orientation, Rift-sawn (e) Number of growth rings per cm, 6 (f) Bottom kerf, No (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Non-durable            22              (a) Species, Cypress pine (b) Width, 65 mm (c) Thickness, 20 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 10 (f) Bottom kerf, No (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, No (k) Natural durability class, Moderately durable            23              (a) Species, Radiata pine (b) Width, 89 mm (c) Thickness, 23 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 4 (f) Bottom kerf, No (g) Eased edge top, No (h) Eased edge bottom, Moderate (i) Unprofiled section, No (j) Treatment, CCA (k) Natural durability class, Non-durable              24              (a) Species, Spruce (b) Width, 140 mm (c) Thickness, 26 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 2 (f) Bottom kerf: Depth, 4 mm; Max width, 16 mm; Min width, 2 mm (g) Eased edge top, Moderate (h) Eased edge bottom, Heavy straight (i) Unprofiled section: Edge, 10 mm (j) Treatment, No (k) Natural durability class, Non-durable   158   Profile Description              25             (a) Species, Pine (b) Width, 93 mm (c) Thickness, 28 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 13 (f) Bottom kerf, No (g) Eased edge top, Slight (h) Eased edge bottom, Slight (i) Unprofiled section, 14 mm (j) Treatment, No (k) Natural durability class, Non-durable              26              (a) Species, Grapia (b) Width, 136 mm (c) Thickness, 25 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 5 (f) Bottom kerf: Depth, 4 mm; Max width, 13 mm; Min width, 3 mm (g) Eased edge top, Slight (h) Eased edge bottom, Heavy straight (i) Unprofiled section: Edge, 9 mm (j) Treatment, No (k) Natural durability class, Durable              27                (a) Species, Pine (b) Width, 135 mm (c) Thickness, 24 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 5 (f) Bottom kerf, No (g) Eased edge top, Heavy rounded (h) Eased edge bottom, Heavy rounded (i) Unprofiled section: Edge, 29 mm; Center, 23 mm (j) Treatment, ACQ (k) Natural durability class, Non-durable              28              (a) Species, Hemlock (b) Width, 140 mm (c) Thickness, 38 mm (d) Growth ring orientation, Flat-sawn (e) Number of growth rings per cm, 4 (f) Bottom kerf, No (g) Eased edge top, Moderate (h) Eased edge bottom, Moderate (i) Unprofiled section, 15 mm (j) Treatment, Heat-treated (k) Natural durability class, Non-durable  159   The width of the deck-boards samples ranged from 65 to 143 mm and their thicknesses varied from 19 to 38 mm. Growth ring width ranged from 1 to 16 growth rings per cm. Twenty one of the commercial boards were flat-sawn, the rest (7) were rift-sawn (Profiles 1,5,10,12,18-19,21).  Seven samples (25%) contained a groove/notch on the bottom, usually six large kerfs (Profiles 1,2,11,16,17,24,26). The dimensions of these grooves varied from: 2 to 4 mm in depth; 11-16 mm wide at the base; and 1 to 4 mm wide at the top. These large kerfs are designed to counter the tendency of boards to cup, according to Hislop (2006). Surface profiling was originally developed to allow plywood to bend towards the grooved side (Deskey 1942), but cupping is a problem with profiled deck-boards (Bailey 1944a,b, Cheng and Evans 2012). Cupping may be reduced by bottom kerfs, which relieve tension stresses, but experimental studies to demonstrate this are lacking (Hislop 2006, Stubbersfield 2006). However, bottom kerfs have other benefits. For example, they may allow better air circulation and drying of the underside of the board, reducing the possibility of decay (Stubbersfield 2006). Another under-surface modification to deck-boards is easing of the edges of the undersides of the boards (Hislop 2006). Fourteen boards (50%) were slightly eased on the bottom while the remaining fourteen boards (50%) had moderate to heavy eased edges. Four of the profiles had heavy straight eased edges (Profiles 2,11,24,26) and another one had heavy rounded edges (Profile 27).   Fifteen boards (54%) contained a flat unprofiled edge on the top while the rest of the boards were profiled to the edge. Profile 18 had a unique eased edge in which the profile was part of the eased edge. In addition, to flat unprofiled edges, boards 2 and 27 contained an unprofiled center. The reason for this modification is not clear. The difference in shrinkage between the tangential and radial directions creates strains and stresses in deck-board that are highest in the center of a flat-sawn board (Schniewind 1959). Hence, this region of a deck-board is more susceptible to checking (Schniewind 1959). Therefore, it follows that the center of boards should be profiled to reduce checking. On the other hand, the edges of flat-sawn boards 160  may not need to be profiled because the stresses are lowest at the edges of boards (Sandberg 1999, Ribarits and Evans 2010).   Twelve profiled deck-boards (43%) were treated with preservatives or modified to improve their durability or dimensional stability (Profiles 1,2,7,9,10,13,16-17,20,23,27- 28). Five boards were treated with alkaline copper quaternary (ACQ) (Profiles 2,7,9,20,27), four boards (14%) were heat treated (Profiles 1,13,16,28), and one (4%) was treated with chromated copper arsenate (CCA) (Profile 23) or acetylated (Profile 17). Another board (4%) was edge glued (Profile 10). The remaining deck-boards were mainly naturally durable hardwoods or softwoods. Seven boards (25%) were non-durable species obtained prior to treatment (Profiles 4,6,14,18, 21,24,25).              161  4.3.2 Quantification of the geometry of profiles in commercial deck-boards  The geometries of the profiles in the different commercial decking samples are shown below in Table 4.2. The profiles varied from: 0.708 to 2.29 mm in height; 2.61 to 6.99 mm in width; 0.09 to 1.38 mm for groove radius; and 0.68 to 4.45 mm for peak radius; 21 to 57 peaks per 15 cm.  Table 4.2: Geometry of the profiles in commercial deck-board samples No. Height (mm) Width (mm) Groove radius (mm) Peak radius (mm) No. Peaks /15 cm 1 1.16 4.00 0.09 2.54 37 2 2.12 5.99 0.12 3.23 25 3 1.59 4.89 0.10 2.53 30 4 1.64 4.98 0.12 2.56 30 5 2.19 5.03 0.12 2.37 29 6 1.33 4.98 0.15 2.94 30 7 2.00 4.92 0.14 2.28 30 8 1.74 4.31 0.14 2.09 34 9 1.54 4.83 0.17 2.49 31 10 0.71 2.61 0.10 1.46 57 11 1.98 4.97 0.16 2.01 30 12 2.12 4.79 0.18 2.17 31 13 2.06 4.53 0.19 2.08 33 14 0.71 5.22 0.47 4.45 28 15 1.8 4.76 0.24 2.18 31 16 1.85 6.99 0.24 2.10 21 17 1.53 5.01 0.29 2.37 29 18 1.25 3.89 0.27 1.82 38 19 2.29 4.96 0.45 1.83 30 20 1.5 3.92 0.44 1.07 38 21 2.08 4.84 0.73 1.49 30 22 1.82 4.99 0.50 0.97 30 23 1.68 6.18 1.25 2.38 24 24 1.85 5.87 1.38 1.91 25 25 1.79 4.82 0.97 1.21 31 26 2.01 5.84 1.3 1.51 25 27 1.72 4.14 0.59 0.68 36 28 0.85 3.63 1.32 0.99 41  To separate and classify the different deck-board samples I plotted the height (H) to width (W) ratios of the profiled peaks, and the radii of their surface grooves (R1) divided by the radii of their peaks (R2) (Figure 4.5). Fourteen of the profiles (50%) had H/W ratios between 35 to 45%. Thus their profiles are fairly even in height and width. Fifty percent of profiles had H/W ratios of less than 35%; which I term short or wide profiles or both (Figure 4.5). None of the 162  profiles had H/W ratios of greater than 50%. But profile 19 had a H/W ratio of 46% because the ribs in the profile were tall, and relatively thin. In contrast, profile 14 had a H/W ratio of 14% because the ribs in the profile were short and wide.  Figure 4.5: Geometry of surface profiles in wooden decking manufactured commercially in eight different countries  The profiled deck-board samples could be separated into three board categories based on their R1/R2 ratios. Two of the categories: (1) rib profiles, which have hemispherical peaks and small R1/R2 ratios, less than 30% (profiles 1 to 19 in Figure 4.5); (2) ripple profiles, which have undulating or wave-like dentate peaks (profiles 24 to 28), and R1/R2 ratios greater than 60% have been mentioned previously, although their geometries were not defined (Morris and McFarling 2008, Evans et al. 2010). In addition to these profiles, a third profile whose geometry fell between those of the rib and ripple profiles was evident. I term profiles with these intermediate characteristics (profiles 20 to 23) ribble profiles to distinguish them from rib and ripple profiles. Ribble profiles had R1/R2 ratios from 30 to 60%.  163   Most of the commercial profiles (1-19) were rib profiles. The appearance and geometry of the different rib, ribble and ripple profiles can be seen in Figure 4.6. The profiles in this figure represent the types of profiles that were encountered within the broad rib, ribble and ripple categories.                     Figure 4.6: Appearance of commercial profiles: (a) Profile 6, a short wide rib; (b) profile 8, a moderate sized rib; (c) profile 12, a tall rib; (d) profile 10, a short, thin rib; (e) profile 11, a moderate sized rib with a flat groove; (f) profile 16, a wide rib; (g) profile 20, a short, thin ribble; (h) profile 22, a moderate sized ribble with a flat peak; (i) profile 21, a tall ribble; (j) profile 28, a short ripple; (k) profile 26, a moderate sized ripple; (l) profile 27, a thin ripple. Scale bar = 2 mm  There are certain morphological characteristics of profiles that R1/R2 or H/W ratios cannot capture such as slope and peak width for profiles that have a flat top (Profile 16), or the width of grooves for profiles that have a flat groove instead of a round groove (Profile 11). For example, profile 11, which forms a cluster with profiles 8 and 7 in Figure 4.5, looks quite different from the latter profiles (Figure 4.7).  (d) short thin                              (e) moderate sized flat groove (f) wide (g) short thin                               (h) moderate sized flat peak    (i) tall (j) short                                        (k) moderate sized                    (l) thin Ribble Ripple (a) short wide                             (b) moderate sized                      (c) tall  Rib 164     Figure 4.7: Appearance of commercial profiles with similar H/W, R1/R2 ratios but differing in appearance: (a) Profile 11; (b) profile 8; (c) profile 7. Scale bar = 2 mm  Similarly, even though they have similar R1/R2 and H/W ratios, profile 13 appears to have a larger groove than 12 (Figure 4.8). Profile 12 appears to have a straighter peak rather than the curved one in profile 12.      Figure 4.8: Appearance of commercial profiles with similar H/W, R1/R2 ratios but differing in appearance: (a) Profile 13; (b) profile 12. Scale bar = 2 mm   However, in general R1/R2 and H/W ratios separate profile deck-boards quite well and hence these parameters could be used to design new profiles which systematically vary H/W, R1/R2 ratios to examine the effects of these parameters on the checking of profiled deck-boards. 4.4 Conclusions  Before I carried out this work there was little information on the geometry of profiles in profiled deck-boards and other morphological modifications to boards. I have shown that there are many variations to the morphology of profiled deck-boards including bottom kerfs, top and bottom eased edges, and unprofiled sections in addition to the variation in the geometry (height, width, peak radius, groove radius) of the profiles. I was able to separate and classify the profiles of the commercial deck-boards using a combination of confocal profilometry and image analysis. I conclude that profiled boards can be separated into 3 broad classifications based on R1/R2 ratios: rib, less than 30%; ribble, 30 to 60%; and ripple, greater than 60%.  However, the collected profiles only represent a small portion of the profiles currently being used all around the world. The effect of these profile parameters, and the height and width of the profiles on the checking of profiled boards exposed to natural weathering is examined in the next chapter.  (a)                                                 (b)                                                 (c)  (a)                                                  (b)                                      165  Chapter 5: Optimizing Surface Profiling to Reduce the Checking of Pacific Silver Fir Deck-Boards 5.1 Introduction   Surface profiling of deck-boards shows promise as a method of reducing the size and number of surface checks that develop when boards are exposed to natural and artificial accelerated weathering (Morris and McFarling 2005, 2008, McFarling et al. 2009, Evans et al. 2010). Profiling helps to hide checks at wood surfaces and it also creates an attractive anti-slip surface. Furthermore, it does not alter wood’s natural colour. Chapter 4 demonstrated that there is great diversity in the geometry of the commercial profiles applied to deck-boards. Furthermore, one study has already shown that the effectiveness of profiling at reducing the checking of Pacific silver fir depends on the type of profile (rib v. ripple) (Evans et al. 2010).  Therefore, I hypothesize profile geometry as defined in Chapter 4 will influence the ability of profiling to reduce the checking of Pacific silver fir. The aim of this chapter is to test this hypothesis and develop a profile that is better than those that have been used in the past to reduce the checking of Pacific silver fir deck-boards. 5.2 Materials and methods  5.2.1 Experimental design  The experiment was a randomized block design with six blocks (Figure 5.1). Each block consisted of 12 samples cut from a single parent board. The 12 different profiles were applied to the samples in each block. In other words there were six replicates for each profile type. Genstat 5 was used to randomly allocate profiles to samples and also the positions of samples in each block (weathering rack) (VSN International 2009, O’Neill 2010).  Data was collected on the size and number of surface checks in samples during weathering and the cupping of samples. Analysis of variance (ANOVA) was used to examine the effect of profile type on the checking and cupping of samples after weathering, with a significance level (α) of 0.05 (VSN International 2009, O’Neill 2010). A sub-routine (convstrt) 166  within the statistical program Genstat was used to contrast the difference between the checking and cupping of the unprofiled control and all 11 profiled samples, as well as between the profiled samples (VSN International 2009). Tests were carried out on data to determine if they conformed to ANOVA’s assumptions (normality, additivity and equality of variances). The effects of profile type on the following numerical indicators of checking and cupping were analyzed: (1) average width, length and area of the ten largest checks in each specimen (data was transformed to natural logarithms); (2) average area of the largest checks in each specimen (as for 1) that were located in the base of grooves, solely on profile peaks or crossing a groove and a peak; (3) number of microscopic checks in the base of grooves divided by the total number of grooves in each specimen; (4) cupping measured before (i) and after weathering (ii), and the difference in cupping (ii-i). Significant results (p<0.05) are presented in graphs and least significant difference (LSD) bars on graphs can be used to compare the means of the control and profiled samples and to estimate whether differences are statistically significant (p<0.05).  Rib+Rbl-Rb-Rp-RpRb-wRbFlatRbl+Rb+wRblRp+Rib+Rp+Rbl+Rb-FlatRblRbRb+wRb-wRp-RpRbl-Rib+Rb-Rp-FlatRpRbRbl+Rb+wRp+Rbl-RblRb-wRib+RblRb-Rb+wRb-wFlatRbRpRbl-Rp-Rp+Rbl+RpRbl+Rbl-Rb+wRbRb-Rb-wRp-FlatRblRib+Rp+Rp-Rbl-RpRbl+Rp+RblRb-Rb+wFlatRib+RbRb-wBlock 1 Block 4Block 2 Block 5Block 3 Block 6 Figure 5.1: Randomized block layout of samples during the weathering trial. Each weathering rack represents a block. Profiles are designated as: Rib (Rb); Ribble (Rbl); Ripple (Rp). Short and tall variation are denoted by a -/+ thin and wide variations are denoted by a –w/+w       167  5.2.2 Design of new surface profiles and manufacture of customized tooling  Eleven profiles were designed across the profile categories defined in Chapter 4 (rib, ribble, ripple) (Table 5.1).  Table 5.1: Dimensions of the designed profiles applied to Pacific silver fir deck-boards Profile type Peak height, mm Peak width,  mm Groove radius, mm Peak radius, mm Peaks/ 15cm Rib (Rb) 2.00 5.00 0.16 2.40 30 Rib+ (Rb+) 2.50 5.00 0.15 2.20 30 Rib- (Rb-) 1.50 5.00 0.16 2.40 30 Rib+w (Rb+w) 2.00 6.25 0.13 2.00 24 Rib-w (Rb-w) 2.00 3.75 0.11 1.70 40 Ribble (Rbl) 2.00 5.00 0.65 1.30 30 Ribble+ (Rbl+) 2.50 5.00 0.65 1.30 30 Ribble- (Rbl-) 1.50 5.00 0.65 1.30 30 Ripple (Rp) 2.00 5.00 1.00 1.20 30 Ripple+ (Rp+) 2.50 5.00 1.00 1.20 30 Ripple- (Rp-) 1.50 5.00 1.00 1.20 30  The geometries of the different designed profiles and those of the closest commercial profiles are plotted in Figure 5.2. The profile that was used commercially to manufacture profiled Pacific silver fir is circled in red in Figure 5.2. It was originally classified as a rib in Morris and McFarling’s (2005) study, but results in Chapter 4 indicate that it should be classified as a ribble profile (Morris and McFarling 2005, McFarling et al. 2009).   168   Figure 5.2: Geometry of designed profiles used to manufacture profiled decking from Pacific silver fir. The names of the different rib (Rb), ribble (Rbl) and ripple (Rp) profiles and precise R1/R2 and H/W ratios of the profiles are listed in Table 5.1. The numbers refers to the commercial profiles listed in Chapter 4              169   Figure 5.3 shows the rib (R1/R2 = 6.67%), ribble (R1/R2 = 50%) and ripple (R1/R2 = 83.3%) profile categories. The H/W ratios of profile peaks were modified to create tall (+25%), standard (0%, baseline) and short (-25%) versions of each profile. The H/W ratios as well as the height and width of three of the profiles variation are similar between the three profile categories (rib, ribble, and ripple).            Figure 5.3: Appearance of the different rib, ribble, and ripple profiles in profiled Pacific silver fir decking. Note that the profiles were machined using material that was not tested in the weathering trial and hence growth ring orientations are a mix of those found in flat-sawn and rift sawn boards   In addition to the three profiles within each of the rib, ribble and ripple categories, two additional rib profiles were created in which the widths of the peaks were changed by ± 25% (Figure 5.4).     Figure 5.4: Appearance of the narrow and wide rib profiles designed and used with Pacific silver fir  Accurate engineering drawings of each of the profiles were produced using AutoCAD (Autodesk 2009) and cutter knives capable of machining each profile were manufactured by a specialist tooling company. AutoCAD images of each profile can be seen in Appendix 4. Two knives, 150 mm wide, with a body diameter (Ø) of 125 mm and hook angle of 15° were manufactured for each profile (Figure 5.5). Short ribble (Rbl-)                Standard ribble (Rbl)                  Tall ribble (Rbl+) Short rib (Rb-)                      Standard rib (Rb)                         Tall rib (Rb+) Short ripple (Rp-)                  Standard ripple (Rp)                  Tall ripple (Rp+) Narrow rib (Rb-w)                 Standard rib (Rb)               Wide rib (Rb+w) 170   Figure 5.5: Custom designed knives for rib (Rb), ribble (Rbl), and ripple (Rp) profiles. Short and tall variations are denoted by a -/+; thin and wide variations are denoted by a –w/+w 5.2.3 Manufacture of profiled pacific silver fir decking  Pacific silver fir was selected because there is commercial interest in the manufacture of profiled decking from this species, and previous studies have shown that profiling reduces checking of Pacific silver fir decking (Chapter 2). Six Pacific silver fir boards measuring 4877 x 140 x 40 mm were donated by FPInnovations. Flat-sawn boards were selected for experimentation, although there were deviations of growth rings along the length of the boards.  The six parent boards were cut into twelve 40 cm long samples using a pendulum saw (Stromab PS 50/F), and the resulting samples were rotary planed to a thickness of 38 mm (Martin T44). The boards were then profiled with a moulding machine (Weinig Profimat 26 Super). Samples from the first parent board were assigned at random to the different profile types as mentioned above, including the unprofiled control. The two relevant profile knives for the first selected profile were inserted into a 125 mm diameter, two-wing, cylindrical rotary 171  cutter head with a hook angle of 15° (Great-Loc SG Positive Clamping Universal Tool System) (Figure 5.6).    Figure 5.6: Great-Loc SG Positive Clamping Universal Tool System: (a) in-feed diagram; and (b) a photograph of the cutter-head   The two knives were secured in place by applying even pressure to nine clamp screws. The cutter head was placed on the machine spindle of a moulding machine, aligned and then secured in place. The decking sample was then machined using a feed speed of 13 m/min and a spindle speed of 6000 rpm to produce a profiled sample with the selected profile. The samples were profiled on their bark side (convex orientation). This process was then repeated for the second assigned profile and so on until all twelve samples from the first parent Pacific silver fir board were profiled. Then samples from the second parent board were profiled, as above, followed by samples from boards 3, 4, 5 and 6 until all 72 samples (6 boards x 12 samples) had been profiled. The final dimensions of the profiled boards were 400 (length) x 135 (width) x 31.75 mm (thickness).  A bench drill (Delta 161/2) was used to predrill 4 holes (Ø=3.97 mm) in each of the samples to prevent end splitting from occurring when profiled deck-boards were screwed to the sub-frame of their respective weathering racks. The holes were positioned 40 (a)                                                                                                                    (b) 172  mm from end grain and 23 mm from the edges of samples. The end grain of the samples was sealed with three coats of epoxy resin (G2 Epoxy, System Three Resins) to reduce end-splitting of samples during exposure to the weather. Samples were air-dried in a conditioning room at 20 ± 1°C and 65 ± 5% relative humidity (r.h.) for two months before they were exposed to the weather. 5.2.4 Measurements of wood properties, cupping and profile geometry  Several measurements were performed to characterize the samples. The width of growth rings and grain angle of wood in each sample were measured using a protractor and ruler (Evans et al. 2010). Sub-samples cut from parent boards were subjected to a water displacement test and then oven dried overnight at 105 ± 5°C to determine their basic density as described in Chapter 3. The growth rate, grain angle, and basic density of samples are shown in Table 5.2.   To measure cupping of samples, defined as “a deviation from flatness transversely across the face of a board” (Knight 1961), each conditioned sample was placed flatwise on a steel fence and planer deviations were measured three times using a dial gauge micrometer (Mitutoyo Digimatic Indicator ID-S1012EB) attached to a machined steel square (Hao and Avramidis 2004) (Figure 5.7). The three measurements were averaged and expressed in mm from the center of the board to the highest point of distortion.   Figure 5.7: Steel fence and dial gauge micrometer used to measure the cupping of deck-board samples 173   The samples were weighed using a balance and their moisture contents were measured using a resistance-type moisture meter RDM3 (Delmhorst Instrument Co.). Moisture contents of individual boards varied from 10.7 to 18.2%.   Each sample was scanned using a deck-top scanner (Microtek Scanmaker i800) using the following parameters: 600 dpi resolution; RGB Colours; 100% scale. A sharpening filter and shadow detailing within Adobe Photoshop were used to highlight the difference between checks and wood. Confocal profilometry, using a 3 mm probe at a gauge resolution of 0.333 nm and a spacing of 10 µm, was used to obtain high quality images of an area measuring 185 mm2 at the surface of profiled samples before and after weathering. A 90 mm line scan obtained with the same setting, as above, was employed to obtain numerical values for the height, width, groove radius, and peak radius of profiles in each sample after weathering. Table 5.2: Growth rate, density and grain angles of parent Pacific silver fir boards that were cut and machined to produce deck-board samples Board No. Growth rings per cm (STD) Basic density (g/cm3) Grain angle (◦) (STD) 1 9 (0.62) 0.38 1.2 (0.96) 2 24 (3.16) 0.38 1.5 (1.07) 3 30 (1.44) 0.41 1.1 (1.03) 4 27 (2.77) 0.40 2.0 (1.76) 5 8 (0.80) 0.33 0.33 (0.33) 6 23 (2.45) 0.37 1.0 (0.96) Average 20 0.38 1.2  5.2.5 Weathering racks and exposure of samples  Profiled samples and the matching flat controls cut from each of the six parent boards were screwed to separate wooden sub-frames made from pressure-treated 2’ x 6’ lumber to create six mini-decks. No center bracing or legs were used in the center of the racks, so that the samples were influenced by the distortion of sub-frame when they were exposed outside (Hislop 2006). Boards were fastened at each corner to the sub-frames using 63.5 mm long, Robertson (8 mm) galvanized decking screws. A gap of 6.35 mm was left between each of the 12 boards in each rack to allow water to drain between boards (Hislop 2006). Unprofiled Pacific silver fir boards, measuring 400 x 50 x 31.75 mm, were screwed to each end of the row of 12 174  boards on each rack to prevent the sides of adjacent test samples from being exposed to the weather. The weathering racks were exposed outdoors in the FPInnovations test site on the Vancouver campus of UBC, for 6 months from February 20th to August 20th 2012 (Figure 5.8).   Figure 5.8: Appearance of the profiled boards after they were weathered for 6 months (refer to Figure 5.1 for profile type)  5.2.6 Measurement of macro-checks  Samples were conditioned at 20 ± 1°C and 65 ± 5% r.h. for exactly five days. Visible checks in all the samples (profiled and flat) were counted except for those located 25.4 mm from fasteners to exclude cracks originating from decking screws. Checks were marked with a soft pencil (3B), and the number of checks and their lengths and widths were recorded for each board.  The length and width of visible checks were measured using a transparent plexiglass (perspex) ruler and calibrated optical loupe (Evans et al. 2010). Since the checks occurred in different areas of the profile, the position of each check within grooves and peaks were recorded. An assignment of V for “valley” or P for “peak” was assigned to each of the checks. If the check crossed over multiple valleys and peaks it was designated as a “diagonal-check”.    175  5.2.7 Visualizing micro-checks  After the positions of checks were recorded I noted that a lot of small cracks were located within the valleys of the profiles. To more clearly see such micro-checks, sections measuring 25 x 135 x 31.75 mm were cross-cut from every weathered and conditioned sample. Sections were re-conditioned at 20 ± 1°C and 65 ± 5% r.h. for 3 days. A brown gel stain (0.2 mL, Varathane 601 Golden Oak) was applied within the grooves of each cross-cut specimen using a spatula. The stain was wiped off with a cotton cloth after 300 seconds. The cross-sections were sanded on their transverse face using an edge sander (Akhurst PMC-150) and a 150 grit abrasive belt. Cracks that were highlighted by the gel stain on the sanded surface were viewed by a magnifying glass and the presence or absence of checks in each groove was recorded. Images of the section’s transverse faces were taken at 4 x magnification using a Canon EOS 7D camera equipped with a Canon MP-E 65 mm 1 to 5 x macro lens. Samples were located 150 mm away from the camera and illuminated with a 4.5 Watt external light source (Litepanels Micro™). The camera was attached to a sliding plate (Manfrotto 454 Micrometric Positioning Sliding Plate) on a ball joint holder (Sirui K-20X) to obtain sharply focused images of checks within the grooves of samples.         176  5.3 Results 5.3.1 Surface macro-checks  Checks in profiled boards when averaged across all profiles were significantly (p = 0.004) narrower than those in the unprofiled control (Con). However, the effectiveness of profiling at restricting check width varied significantly (p = 0.022) with profile geometry (Figure 5.9).  Boards with the rib (Rb) and wide rib (Rb+w) profiles had the narrowest checks, but the width of large checks in boards with the tall rib (Rb+), ribble (Rbl), ripple (Rp) and short ripple (Rp-) profiles were not significantly different (p>0.05) from that of checks in the unprofiled control.   Figure 5.9: Average width of ten largest checks in profiled and flat (unprofiled) Pacific silver fir boards exposed outdoors to natural weathering for 6 months. The labels on the x-axis refer to the profiles listed in Table 5.1. Y1-axis refers to natural logarithms of check width. The Y2 axis contain values on a natural scale (ex)  Boards with rib profiles, with the exception of those with a short rib profile (Rb-), had the shortest checks, but the lengths of checks in boards with these rib profiles were not significantly (p>0.05) different from that of checks in the unprofiled control (Figure 5.10). Checks in boards with the short rib (Rb-), ribble (Rbl), short ribble (Rbl-), ripple (Rp) and short ripple (Rp-) profiles were all significantly longer (p<0.05) than those in the unprofiled control (Con).  177    Figure 5.10: Average length of ten largest checks in profiled and flat Pacific silver fir boards exposed outdoors to natural weathering for 6 months  The area of large checks was significantly (p<0.05) smaller in boards with the rib (Rb), wide rib (Rb+w) and narrow rib (Rb-w) profiles compared to that of checks in the unprofiled control (Figure 5.11), which reflects the effects of these profiles on check width and length (Figure 5.9 and 5.10). I sub-divided data for check area into three categories according to whether checks were located within grooves, on peaks or whether they crossed a peak (see following section).   Figure 5.11: Average area of ten largest checks in profiled and flat Pacific silver fir boards exposed outdoors to natural weathering for 6 months 178  5.3.2 Positions of surface checks in deck-board samples  The positions of surface checks in profiled sample can be observed in Figure 5.12. This figure shows topographical maps of the surface of rib, wide-rib, ripple and short ribble boards after they were exposed to the weather. Checks in grooves (arrowed) are prominent in samples with the ripple (Rp) and short ribble (Rbl-) profiles (Figure 5.12c,d). Checks that occur on the peaks of profiles (arrowed) can be seen on the left-hand sides of samples with the rib (Rb) and short ribble (Rbl-) profiles (Figure 5.12a,d). The latter board type also contains a prominent diagonal-check (arrowed) that crosses a profile peak (right-hand side of Figure 5.12d).   Figure 5.12: Confocal profilometry images of the surface topography of profiled Pacific silver fir decking samples exposed to natural weathering for 6 months: (a) Rib sample showing checks within grooves, and a large check on the top of a peak (far left); (b) Wide rib sample showing checks within grooves; (c) Ripple sample showing large and small checks within grooves; (d) Short ribble sample showing large and small checks within grooves and two diagonal-checks that cross profile peaks (far right)  179   Profiles with different geometries (rib, ribble and ripple) developed all three types of checks, with the exception of boards with the tall ripple (Rp+) profile, which didn’t develop checks on profile peaks (Figure 5.13). Boards with the standard ripple (Rp), standard ribble (Rbl) and short ribble (Rbl-) profiles had the smallest area of checks on their peaks whereas boards with the rib profiles had the most, with the exception of boards with a short rib (Rb-) profile. The area of checks on peaks in samples with the standard ripple (Rp), standard ribble (Rbl) and short ribble (Rbl-) profiles were significantly (p<0.05) smaller than those of checks in samples with the tall rib (Rb+), wide rib (Rb+w) and narrow rib (Rb-w) profiles. In contrast, the area of checks in grooves in samples with the ripple and ribble profiles were all significantly (p<0.05) greater than those of samples with the rib profiles, with the exception of samples with the short rib (Rb-) profile. The area of large diagonal-checks that crossed profile peaks was significantly (p<0.05) larger in samples with the short rib (Rb-) and short ripple (Rp-) profiles and lowest in samples with the tall ripple (Rp+) and narrow rib (Rb-w) profiles.   Figure 5.13: Areas of large checks in profiled samples that developed in grooves (valleys), on peaks or diagonally in Pacific silver fir deck-board samples exposed to natural weathering for 6 months. Note that the short ribble profile (Rbl-) developed some peak checks but they were small (0.5 mm2) and do not appear on this figure  180   In addition to the aforementioned large checks, numerous smaller checks developed in the grooves of profiles. This became apparent when I used high resolution confocal profilometry to image the surface of profiled samples (Figure 5.12). For example, if you look closely at the grooves in the profilometry images in Figure 5.12 you can see a small check in every groove. The percentage of grooves that contained checks that were visible to a hand lens was very high and varied from a maximum of 98.6% for boards with a rib profile to a minimum of 93.0% for boards with a tall rib profile. Differences in the percentages of grooves in the different profiled boards that contained small checks were not statistically significant (p = 0.408). These small checks developed at or near the bottom of grooves and propagated radially into board samples (Figure 5.14). Most of the checks could only be seen with a hand lens and therefore they didn’t affect the appearance of the boards.  Figure 5.14: Close-up of peaks and grooves in profiled Pacific silver fir samples after 6 months of weathering. Note the development of checks at the base of grooves in each of the profiled specimens: (a) Standard rib (Rb); (b) Standard ribble (Rbl); (c) Standard ripple (Rp); (d) Standard ribble (Rbl). Scale bar = 1 mm   181  5.3.3 Changes in profile geometry after weathering  Most of the profiles had a lower height/width ratio after weathering (Figure 5.15). In addition, the R1/R2 ratios of most of the profiles increased because grooves became larger after weathering. The wider grooves in ribble and ripple profiles might be due to checks opening up the grooves. The R1/R2 ratios of rib profiles did not increase as much, possibly because the cracks were not wide as those found in the grooves of samples with ribble and ripple profiles.   Figure 5.15: Geometry of profiles after 6 months of outdoor exposure compared to their initial geometry      182  5.3.4 Appearance of boards after weathering  All three types of profiles were effective at masking checks, as expected (Figure 5.16). Ribble and ripple boards were less effective than rib profiles at masking checks because checks could be more easily seen at the base of their grooves, and in addition some boards with ribble and ripple profiles developed more diagonal-checks than boards with rib profiles (Section 5.3.2).   Figure 5.16: Pacific silver fir boards exposed to the weather for 6 months: a) control (flat); b) rib; c) ribble; d) ripple. Scale bar = 50 mm        183   Checks were more difficult to see in boards with the tall ripple profiles than in boards with the short and standard ripple profiles because they developed relatively few diagonal-checks and no checks on their peaks (Figure 5.17).   Figure 5.17: Profiled Pacific silver fir boards exposed to the weather for 6 months: a) ripple; b) tall ripple; c) short ripple. Scale bar = 50 mm  Rib profiles including the wide and narrow rib profiles were good at hiding checks (Figure 5.18). The numerous peaks and grooves in samples with narrow ribs made it particularly difficult to see checks.    Figure 5.18: Profiled Pacific silver fir boards exposed to the weather for 6 months: a) rib; b) wide rib; c) thin rib. Scale bar = 50 mm 184  5.3.5 Cupping of deck-board samples  Profiling significantly (p = 0.02) increased the cupping of boards exposed to the weather (Figure 5.19), but had no significant (p = 0.457) effect on the initial cupping of boards after they were machined and conditioned. Results in Figure 5.19 for profiled boards are averaged across the different profiles because there was no significant (p = 0.265) effect of profile type on cupping.   Figure 5.19: Difference in cupping of Pacific silver fir boards before and after they were exposed to natural weathering for 6 months. Results for profiled samples are averaged across all boards and compared with cupping occurring in the flat (unprofiled) controls  The cupping of profiled boards that developed during natural weathering was quite pronounced as can be seen in the figure below (Figure 5.20) which shows a cross-section of a sample with a rib profile after it was exposed to the weather for six months (Figure 5.20).   Figure 5.20: Cross-section of a board with a rib profile after it was weathered for six months. Note the pronounced cupping of the board 185  5.4 Discussion  Profiling is interesting because it reduces the negative effects of checking on the appearance of deck-boards exposed outdoors. This effect of profiling on appearance has been explained by the reduced size of checks that develop in profiled boards, and also the difficulty of seeing checks that develop within grooves (McFarling et al. 2009). Therefore an ideal profile should restrict the size of checks that develop when deck-boards are exposed outdoors and confine visible checks to the grooves of profiles. None of the profiles I tested possessed exactly this desirable combination of properties, because none of them confined all large checks to profile grooves. Nevertheless, it seems likely that two of the rib profiles (rib and wide rib) would be better at restricting the size of checks in Pacific silver fir than the ribble profile (profile 20 in Figure 4.5 and 5.21 and Table 4.2) that was used commercially in Canada to profile Pacific silver fir. The rib profile (Rb) tested here is very similar to one that is being used in Germany to profile larch (Larix sp.) deck-boards (profile 11 in Figure 4.5 and Table 4.2), but none of the commercial profiles are identical to the wide rib profile I developed, which also performed very well.  I hoped that my research would reveal universal trends about the relationship between profile geometry (R1/R2 ratio and H/W ratio) and the ability of profiling to restrict the checking of Pacific silver fir, but no entirely consistent trends of these geometric parameters on check sizes emerged. For example, it is clear that rib profiles (low R1/R2 ratios) were generally better at restricting the size of checks than profiles with higher R1/R2 ratios (ribble and ripple profiles). However, checking of boards with a short rib profile was similar to that of boards with ribble or ripple profiles, despite the fact that the R1/R2 ratio of boards with short rib profiles was similar to that of boards with the other rib profiles.   The effect of profile geometry on the locations of checks was also inconsistent. For example, short rib and short ripple profiles had H/W ratios and groove depths of 30% and 1.5 mm, compared to 40% and 2.0 mm and 50% and 2.5 mm for the corresponding standard and tall profiles, respectively. Both of these short profiles encouraged the formation of diagonal-checks that crossed profile peaks, but the same effect was not observed in boards with a short ribble profile. Diagonal-checks develop more readily in Pacific silver fir with spiral 186  grain (Evans et al. 2010), a defect that is unavoidable in wood used to make decking. Diagonal-checks that cross profile peaks are very easy to see (Evans et al. 2010) and hence profiling should try to reduce their formation. Our results suggest that this desirable outcome may be achieved in Pacific silver fir by avoiding rib and ripple profiles with a low H/W ratio (<30%) and grooves less than 1.5 mm deep. Rib and ripple profiles with these characteristics were commonly encountered during our industry survey (profiles 1, 6, 10, and 28 in Figure 4.5 and Table 4.2) and hence it would be worthwhile conducting experiments to see if diagonal-checks can be reduced in other wood species by increasing the H/W ratio and depth of profiles machined into boards.  Profiling increased the tendency of boards to cup when they were exposed outdoors. The same undesirable tendency was noted in profiled (striated) plywood by Bailey (1944a,b). Bailey solved the problem of cupping of striated plywood by increasing the thickness of the striated veneer to create a balanced panel which equalized stresses in opposing veneers (Bailey 1944a,b). The same approach is clearly not suitable for wooden decking, but it’s possible that stress relief grooves that are machined into the undersides of some deck-boards (Nystrom 1995) in combination with a water-repellant preservative treatment and coating might reduce the tendency of profiled boards to cup when they are exposed outdoors. Further research would be needed to test this hypothesis.  My results showed that profiling restricted the width of checks that developed when Pacific silver fir deck-boards were exposed outdoors to the weather, but was ineffective at restricting large checks from becoming longer. These findings support previous observations (McFarling et al. 2009, Evans et al. 2010). I also found checks at the base of almost every groove in profiled boards, irrespective of profile type.  These checks were microscopic and could only be seen with a hand lens. Hence, they did not influence the appearance of boards. Very narrow checks (‘slits’) also developed at the base of grooved plywood exposed outdoors and it was suggested that they prevented the creation of long wide cracks (Elmendorf 1950). Lepitre and Mariaux (1965) also mentioned that small cracks prevented larger splits from developing in the ends of logs that contained S-hooks. The small microscopic checks I observed may not account 187  for differences in the ability of profiles to restrict the development of large checks because they were found in all of the different profiled boards. However, their formation and the increased tendency of profiled boards to cup suggest that profiling alters in a very fundamental way the stresses and strains that are responsible for the checking of wooden decking exposed outdoors to weathering. A recent study has confirmed that this is the case using digital image correlation of profiled boards exposed to artificial weathering (Mallett 2012). Using this technique Mallett (2012) explained why cracks are usually located at the bottom of the grooves. He suggested that when the two peaks dry and then shrink, the area between them (groove) is pulled apart causing checks. This suggestion explains why micro-checks and most macroscopic checks develop at the base of the grooves, and also the changes in profile geometry that occurs after samples are weathered.  5.5 Conclusions  This chapter compared the checking of profiled Pacific silver fir decking with different R1/R2 ratios and height to width (H/W) ratios of profile peaks after profiled boards were exposed outdoors for 6 months. In conclusion, there was no consistent trend of these geometric parameters on checking, but I found that profiles with small R1/R2 ratios (rib profiles) and grooves deeper than 1.5 mm were more effective at restricting checking of Pacific silver fir than profiles with larger R1/R2 ratios (ribble and ripple profiles). Furthermore, the development of unsightly checks that traverse profile peaks in boards with rib and ripple profiles could be reduced by creating grooves that are deeper than 1.5 mm. Profiling increased the undesirable cupping of deck-boards exposed outdoors and research is urgently needed to solve this problem. The research in this chapter demonstrates that it is possible to reduce surface checking in profiled Pacific silver fir decking by changing the geometry of surface profiles. In principle, the same approach could be used with other commercially important wood species to reduce the negative effects of surface checking on the appearance of wooden decking used outdoors.   188  Chapter 6: General Discussion, Conclusions, and Suggestions for Further Research 6.1 General discussion  In Chapter 1 I hypothesized that both wood species and profile geometry would significantly influence the checking of wood decking exposed to the weather. My experimental results in Chapters 3 and 5 clearly support this hypothesis. Finding alternative wood species for use as decking is important and several criteria need to be considered. Results in Chapter 3 showed that the different species I tested vary in their susceptibility to checking. There has only been a handful of research papers published on checking of different wood species, and even fewer on the checking of decking. The published research on the checking of decking has mainly focused on softwoods. This research has confirmed that western red cedar, a popular choice of species for decking, is resistant to checking (Morris and Ingram 2002, Ratu 2009). The tropical hardwood, ipe has been found to resist checking when exposed outside (Miller et al. 2003, Williams et al. 2001, Izekor and Fuwape 2010). It is also a popular decking material in North America and Europe (Schulze et al. 2008).  However, both ipe and western red cedar are not as widely available or as cheap as pressure treated lumber (Schulze et al. 2008, Eastin and Ganguly 2009). This provides an opportunity to explore the use of other wood species, as decking. None of the softwoods and hardwoods I tested performed as well as ipe and western red cedar in terms of checking, but their performance might be enhanced by treatments that can dimensionally stabilize the wood and reduce checking. Such check reducing treatments could include the use of water-repellant preservatives and finishes, or physical modification such as profiling or a combination of both. The species I examined that might benefit from such treatments are: aspen, bigleaf maple, basswood, red alder, yellow cedar, white spruce, and western larch. Aspen, basswood, and bigleaf maple performed the best out of these species in terms of checking. All three are diffuse porous hardwoods and their densities range from 0.36 to 0.40 g/cm3.  They are available as lumber, although aspen is mainly used as material for oriented strand board (Wiemann 2010). 189   Results in Chapter 5, which used the profile categories defined and classified in Chapter 4, showed that certain profiles such as the rib profile are very effective at reducing checking of Pacific silver fir. This species is quite prone to checking, as my results in Chapter 3 confirmed, but its susceptibility to checking can be reduced by profiling (Morris and McFarling 2005, 2008, McFarling et al. 2009, Evans et al. 2010). The study by Morris and McFarling (2005) showed that surface profiling was able to reduce check width in untreated Pacific silver deck-boards exposed outdoors for 23 months by a factor of two. I obtained comparable results as the ‘standard’ and the ‘wide’ rib profiles were able to reduce check width by ~3.1 and ~2.5 times, respectively. My results also show that the average width of the ten largest checks in Pacific silver fir deck-boards was ~3.5 times greater than those in western red cedar when boards were exposed outside for one year. Pacific silver fir had the second worse check rating among the 17 species I tested. Therefore, it is possible that profiling of species that performed better than Pacific silver fir in terms of checking, might produce a deck-board that could perform as well or better than unprofiled western red cedar deck-boards. Checks in ipe, on the other hand, were ~7 times narrower than those in Pacific silver fir making it less likely that profiling would create deck-boards that could match the performance of ipe. However, profiling species such as aspen, and maple might create decking whose performance comes close to that of ipe. Profiling is also a good treatment in species such white spruce, and larch that are difficult to penetrate with preservatives and have treated zones that can be breached by checks (Alden 1995, 1997).   These positive comments on profiling need to be tempered by some other considerations. The positive effects of profiling and also incising on checking vary with wood species (Collister 1956, Evans et al. 2010). The study by Evans et al. (2010) found that a rib profile was more effective at reducing checking in Pacific silver fir than it was at reducing checking of southern pine. Southern pine is a denser species than Pacific silver fir and it is possible that rib profiles may be better suited to the lower density hardwoods that performed well here. Conversely, ripple and ribble profiles may be better suited to more dense species such as larch. Certainly results in Chapter 5 suggest that profile geometry can be optimized for an individual species. Furthermore, the wide rib profile that is unique to this study, performed 190  just as well as the standard rib, which is similar in geometry to many profiles that are used commercially.   It is important to point out some of the disadvantages of profiled deck-boards. Profiled boards tend to cup excessively after they have been profiled and exposed outdoors, irrespective of profile type (Chapter 5). Such cupping may exacerbate checking if the board is very securely restrained by fasteners (Stamm 1965, Evans 2008). Furthermore, the raised edges of cupped deck-boards are a tripping hazard and may cause water to pool in the center of boards. In addition, profiled boards are more difficult to clean, resand and refinish than unprofiled boards. The latter problems might be reduced by altering the geometry of the profiles. For example, wider grooves could be used to facilitate cleaning and refinishing of profiled deck-boards. There have been suggestions that the grooves of the profiles (and the checks that develop in them) may encourage the ingress of moisture and micro-organism into boards. Therefore it is important to use profiling in combination with preservative treatment or durable species. Profiling was originally developed to make deck-boards less slippery (Hislop 2006). Boards that contain flat areas or wide peaks such as the wide rib might require an anti-slip coating. However, ripple or ribble profiles may not be suited to low density species because the smaller peaks of the profiles may be more easily damaged than wider ones. A wide profile such as the wide rib with its larger peak radius could be more suited to lower density species. In addition, profiling the convex side of flat-sawn boards might reduce shelling and separation of profiles in species with a large contrast in the density of earlywood and latewood (Morris and McFarling 2010, Williams and Knaebe 1995). The surface of profiled decking should be comfortable to bare feet and the wide rib design may be more comfortable than the tall ripple or ribble designs. The appearance of profiled decking is also very important. Diagonal-checks in particular mar the appearance of decking. Taller ripple and ribble profiles were better at restricting diagonal checking of Pacific silver fir because they constrained checks in the grooves unlike some rib profiles. Hence, tall profiles may be better suited to species that are particularly prone to diagonal checking because they contain wood with high grain angles. 191   Some other interesting observations were made in the course of my investigations that are relevant to the development of improved wooden decking. For example, the grey colour of the different species varied significantly after they were weathered. Such differences in grey colour might be related to how much of the surface has been eroded, because lower density species such as balsa and western red cedar tended to darken less than higher density species such as ipe. Ipe’s dark colour after it was weathered, however, may have helped mask surface checks.   Wasps frequently visited almost all of the hardwoods I tested, particularly white oak, red alder, and aspen. The wasps stripped wood from weathered deck-boards revealing unweathered wood which created a stripped surface. The striped areas in basswood and red alder were rapidly colonized by staining fungi. There is no literature on this subject, but my observations suggest that insects may act as vectors for the re-colonization of wood surfaces by staining fungi.  To date, profiling has mainly used a single profile type on each board, although some areas on boards can be left unprofiled. It can be argued, however, that combinations of profiles might be more suitable with some species. For example, most of the softwoods checked badly in the center of boards, an area where growth rings were parallel to the wide surface of flat-sawn boards. Boards with these characteristics could be profiled with a combination of profiles, for example a standard rib profile could be used in the center and a more widely spaced profile or flat surface at the edges. Accordingly some of the commercial profiles tested in Chapter 4 had unprofiled edges (Profiles 1-3,5,11-13,16-17,19-20,24-27). None of the profiles tested in Chapter 5 were exactly the same as the profiles tested by McFarling and Morris (2005) or Evans et al. (2010), but the standard rib profile was similar to some of the commercial profiles I characterized in Chapter 4 (Profiles 7, 8, 11). Some of the commercial profiles had unique shapes such as a flat groove, similar to a small kerf, or a flat top (Profiles 11, 22). These modifications may have evolved to make the surface more comfortable to walk on or easier to clean.   192  6.2 Conclusions  This thesis clearly demonstrates that wood species and profiling influences checking. Popular deck-board species such as western red cedar and ipe performed very well in terms of checking. None of the other species performed as well. However, aspen, basswood, and big leaf maple performed almost as well as western red cedar and further improvement might be possible with effective chemical or physical treatments. Profiling reduced checking, of Pacific silver fir, as expected, but certain profiles such as the rib profiles performed significantly better than others. Species such as western larch and white spruce, which rated poorly in terms of checking, but scored well in terms of other criteria could benefit from surface profiling. If profiling was applied to aspen, basswood, or maple they might perform as well as western red cedar. Rib and wide rib profiles performed better than all of the profiles used in the study, but the taller ripple and ribble profiles were better at constraining checks to the grooves. Therefore, I conclude that the rib or wide rib profiles are the best profiles to use with Pacific silver fir which is very susceptible to checking. Profiling might further improve the performance of western red cedar and ipe and improve their ability to compete with wood plastic composites. However, profiling increased the cupping of boards and a solution to this problem is urgently needed.  Finally I conclude that some species that are not commonly used for decking (aspen, basswood, and maple) show promise for this application provided they can be obtained in large volumes at a competitive price. If this was the case it would be worth exploring their use as deck-boards and developing appropriate treatments and profiles for them. I was able to design a new profile, the wide rib profile, which performed as well as the ‘standard’ rib profile that has been used by industry and better than ribble profiles that were similar to one that was used commercially in Canada to profile Pacific silver fir. This finding illustrates that the development of a way of categorizing and designing profiles provides an opportunity for companies to design new profiles which might be better than those that are currently used.  193  6.3 Further research  There is great scope to improve the performance of wooden decking so that it can compete more effectively with wood plastic composites. My review of the literature shows the research that is being done around the world to improve the performance of wooden decking. It is not my intention in this section to list or describe the latter research. Instead I mainly focus on new questions that have arisen from my research, or the research that is needed to broaden or deepen the significance of my findings.   The outstanding finding arising from my study of the weathering and checking of different wood species is that some lower density hardwood species show promise as decking. Lower density species such as basswood, and red alder performed well in terms of checking. This finding could be confirmed by examining the surface checking of other lower density hardwoods such as buckeye (Aesculus flava Sol.), black cottonwood (Populus trichocarpa Torr. and A.Gray), and black willow (Salix nigra Marshall) (Alden 1995). All of these species have fairly low density (0.35-0.39 g/cm3), and black willow is an interesting choice because unlike cottonwood and buckeye it is semi-diffuse porous (Hoadley 1990, Alden 1995). In addition, to this research on hardwoods, research is needed to better understand why wasps prefer hardwoods to softwoods when mining weathered wood for use as nest building material.   Chapter 4 characterized the geometry of profiles in commercially profiled boards. It is the most comprehensive study of its kind, but only 28 different profiles were characterized and no doubt many more profiles are being used commercially. In comparison to rib profiles, not a lot of ripple or ribble profiles were characterized. Therefore, research should be carried out to search for more commercial profiles with unique shapes including more examples of ribble and ripple profiles. If this research was carried out it might be possible to see if certain profiles are more commonly associated with particular species or wood types.   Chapter 5 examined the effect of the geometry of the profiles on the surface checking of Pacific silver fir deck-boards exposed to the weather. This research is the first serious attempt to tailor profiles to reduce surface checking of decking. Although two of the profiles performed 194  better than others, it is difficult to draw conclusions about the effect of profile geometry on the ability of profiles to reduce checking. Nevertheless, it would be worthwhile to carry out research to see if the profiles that were good at restricting checking of Pacific silver fir here, work well with other species. In addition, Chapter 5 showed that the geometry of the profiles changed after the profiled deck-boards were weathered. Mallett (2012) also noted reversible changes in profile geometry when profiled deck-boards were subjected to wetting and drying. Additional studies on the changes in profile geometry during weathering may provide insights into why some profiles were more effective at reducing checking than others. In addition, my discovery that micro-checks developed at the base of grooves in profiled decking is interesting and raises questions about whether such micro-checks might have redistributed the stress that causes large checks to form (Lepitre and Mariaux 1965, Morris and Ingram 2002). Further research to test this hypothesis is needed.  In addition, to all of the aforementioned research there is great scope to improve the performance of the different wood species and profiled decking that was tested here. For example, the use of hydrophobic wood preservatives and coatings could improve the performance of hardwood decking and also profiled Pacific silver fir decking. The cupping of profiled Pacific silver fir for decking might be improved through the use of under-surface kerfs. All of these areas are worthy of further research to enable wood decking to compete more effectively with wood plastic composite decking.   195  References  Akhtari, M. (2010). Study of microscopy properties of modified wood during weathering process. The International Research Group on Wood Protection, IRG-WP/10-40517.  Akhtari, M., Nicholas, D. (2014). Effect of profiling and preservative treatments on the weathering characteristics of southern pine deck boards. European Journal of Wood and Wood Products, 72(6), 829-831.  Alden, H.A. (1995). Hardwoods of North America. USDA Forest Service General Technical Report FPL-GTR-83. Forest Products Laboratory North Central Forest Experiment Station, Madison, W.I., U.S.A.  Alden, H.A. (1997). Softwoods of North America. USDA Forest Service General Technical Report FPL-GTR-102. Forest Products Laboratory North Central Forest Experiment Station, Madison, W.I., U.S.A.   Alger, M.S. (1997). Polymer science dictionary. Chapman & Hall, London, U.K.  Arnold, W.P. (1949). Effect of incising on checks and splits in mixed hardwood ties. Report on Assignment 7. Incising Forest Products. In: Proceedings of the Forty-Eighth Annual Convention of the American Railway Engineering Association, 385-386.  Arnold, W.P. (1950). Effect of incising on checks and splits in mixed hardwood ties; Great northern railway test; Minneapolis, St Paul & Ste. Marie railway test. Report on Assignment 7. Incising Forest Products. In: Proceedings of the Forty-Ninth Annual Convention of the American Railway Engineering Association, 356-357.  Ashton, H.E. (1967). Clear finishes for exterior wood field exposure tests. Journal of Paint Technology, 39(507), 212-224.  Atanassova, V. (2010). Representation of fuzzy and intuitionistic fuzzy data by radar charts. Notes on Intuitionistic Fuzzy Sets, 16(1), 21-26.   Autodesk. (2009). AutoCAD® 2010.  Autodesk inc., San Rafael, C.A., U.S.A.  Bailey, W.C., inventor, United States Plywood Corporation, assignee. (1944a). Process of making balanced laminated panels. United States Patent US 2,363,492.   Bailey, W.C., inventor, United States Plywood Corporation, assignee. (1944b). Balanced striated plywood panel. United States Patent US 2,363,927.   196  Balfas, J. (1994). The effect of end-coating treatments on cracking reduction of wood statues. Jurnal Penelitian Hasil Hutan, 12(5), 169-174.  Bariska, M., Pizzi, A., Conradie, W.E. (1988). Structural weakening of CCA-treated timber. Holzforschung, 52(5), 339-345.  Beckers, E.P.J., Meijer, M. de, Militz, H., Stevens, M. (1998). Performance of finishes on wood that is chemically modified by acetylation. Journal of Coatings Technology, 70(878), 59-67.  Belford, D.S., Nicholson, J. (1969). Emulsion additives for CCA preservatives to control weathering. Proceedings of the Sixty-Fifth Annual Meeting of the American Wood-Preservers’ Association, 38-51.  Bodner, J., Schlag, M.G., Grüll, G. (1997). Fracture initiation and progress in wood specimens stressed in tension. Part II. Compression wood specimens stressed parallel to the grain. Holzforschung, 51(6), 571-576.  Bolin, C.A., Smith, S. (2010). Life cycle assessment of ACQ-treated lumber with comparison to wood composite decking. Journal of Cleaner Production, 19(6-7), 620-629.  Borgin, K., Corbett, K. (1970a). The stability and weathering properties of wood treated with various oils. Plastics, Paint and Rubber, 14(1/2), 69-72.  Borgin, K., Corbett, K. (1970b). The stability and weathering properties of wood treated with various resins. Plastics, Paint and Rubber, 14(5/6), 61-66.  Borst, K. de, Bader, T.K., Wikete, C. (2012). Microstructure–stiffness relationships of ten European and tropical hardwood species. Journal of Structural Biology, 177(2), 532–542.  Bosshard, H.H., Futo, L.P. (1963). Weathering resistance of plywood bonded and coated with phenol formaldehyde. Forest Products Journal, 13(12), 557-561.  Boutelje, J.B. (1966). On the anatomical structure, moisture content, density, shrinkage, and resin content of the wood in and around knots in Swedish pine (Pinus sylvestris L.) and in swedish spruce (Picea abies Karst.). Svensk Papperstidning, 69(1), 1-10.  Brischke, C., Rapp, A. (2004). Investigation of the suitability of silver fir (Abies alba Mill.) for thermal modification. The International Research Group on Wood Protection, IRG/WP/04-40275.  Brock, W.E., inventor, Thompson,  J., assignee. (1878). Manufacture of compound lumber. United States Patent US 206,534.   197  Browne, F.L. (1952). Wood siding left to weather naturally. Architectural Record, 112(5), 197-199.  Browne, F.L., Simonson, H.C. (1957). The penetration of light into wood. Forest Products Journal, 7(10), 308-314.  Buchanan, B., Groot, R.C. de, Boyette, J. (1990). Resistance of shakes and shingles made from western hemlock and Pacific silver fir to wetting-drying degrade accelerated tests. Forest Products Journal, 40(9), 29-34.  Bucur, V. (2011). Terms for delamination in wood science and technology. In: Delamination in Wood, Wood Products and Wood-Based Composites. Springer, New York, U.S.A., 17-32.  Burnes, J.D. (1923a). Experiment to determine the effect of perforations upon the checking of northern white cedar and western red cedar after treatment. Proceedings of the Nineteenth Annual Meeting of the American Wood-Preservers’ Association, 263-268.  Burnes, J.D. (1923b). Inspection of treated poles set in the rice street lead of the St Paul Gas Light Company. Proceedings of the Nineteenth Annual Meeting of the American Wood-Preservers’ Association, 273-282.  Burt, C.S. (1955). How Illinois central extends life of its cross ties. Cross Tie Bulletin, 36(5), 11-14.  Campbell, G.S. (1971). Control of seasoning degrade in jarrah joinery stock. CSIRO Forest Products News Letter, 386, 1-2.  Carus, M., Eder, A., Dammer, L., Korte, H., Scholz, L., Essel, R., Breitmeyer, E. (2014). Wood-plastic composites (WPC) and natural fibre composites (NFC): European and global markets 2012 and future trends. Nova Institute for Ecology and Innovation, Huerth, Germany. Retrieved from: bio-based/eu/markets  Chafe, S.C. (1994). Preheating green boards of mountain ash (Eucalyptus regnans F. Muell) II. Relationships amongst properties. Holzforschung, 48(2), 163-167.  Chandler, C.R. (1968). Experimental guardrail posts final report on a method of reducing post checking. Technical Report 68-4. Oregon State Highway Department Materials and Research Division, Oregon, U.S.A.  Cheng, K.J., Evans, P.D. (2012). Weathering performance of white spruce decking treated with low molecular weight phenol formaldehyde resin. Proceedings of the Eleventh Pacific Rim Bio-based Composites Symposium (Nov. 28-30), Shizuoka, Japan, 575-577.  Choi, S.M., Ruddick, J.N.R., Morris, P.I. (2004). Chemical redistribution in CCA-treated decking. Forest Products Journal, 54(3), 33-37. 198   Choi, S.M., Ruddick, J.N.R., Morris, P.I. (2001). The possible role of mobile CCA components in preventing spore germination in checked surfaces, in treated wood exposed above ground. The International Research Group on Wood Protection, IRG/WP/01-30263.  Choi, S.M., Ruddick, J.N.R., Morris, P.I. (2003). Fungal colonization of CCA-treated decking. The International Research Group on Wood Protection, IRG/WP/03-10491.  Chow, K.Y. (1947). A comparative study of the structure and chemical composition of tension wood and normal wood in beech (Fagus sylvatica L.). Forestry, 20(1), 62-76.  Chudnoff, M. (1980). Tropical timbers of the world. USDA Forest Service Forest Products Laboratory, Madison, W.I., U.S.A.  Clark, W.P. (1956). Effect of tension wood on seasoning and machining of eastern cottonwood. Forest Products Journal, 8(3), 109-112.  Coleman, H.D., Yan, J., Mansfield, S.D. (2009). Sucrose synthase affects carbon partitioning to increase cellulose production and altered cell wall ultrastructure. Proceedings of the National Academy of Sciences of the United States of America, 106(31), 13118-13123.  Collister, L.C. (1955). Panel discussion on cross ties. Proceedings of the Fifty-Fourth Annual Convention of the American Railway Engineering Association, 1059-1066.   Conners, T.E. (2008). Producing and inspecting railroad crossties. University of Kentucky Cooperative Extension Service For-108. University of Kentucky College of Agriculture, Lexington, K.Y., U.S.A. Retrieved from: http://www2.ca.uky.edu/forestryextension/Publications/Railroad%20tie%20pub.pdf  Cooke, P.T. (1936). End splitting of cypress pine. Australian Forestry, 1(2), 39-40.  Core, H.A., Côté, W. A., Jr., Day, A. C. (1961). Characteristics of compression wood in some native conifers. Forest Products Journal, 11(8), 356-362.  Côté, W.A. Jr. (1968). Defects and abnormalities of wood. In: F.P., Kollmann, W.A. Jr., Côté (Eds.), Principles of Wood Science and Technology I. Solid Wood. Springer-Verlag, Berlin, Germany, 79-96.  Côté, W.A. Jr., Day, A.C., Kutscha, N.P., Timell, T.E. (1967). Studies on compression wood. V. Nature of the compression wood formed in early springwood of conifers. Holzforschung, 21(6), 180-186.   Coupe, C., Watson, R.W. (1967). (2) Fundamental aspects of weathering. Record of seventeenth annual convention-British Wood Preservers’ Association, 37–60. 199   Crawford, D. M., Groot, R.C. de, Gjovik, L.R. (1999). Ten-year performance of treated northeastern softwoods in aboveground and ground-contact exposures. USDA Forest Service Research Paper FPL-RP-578. Forest Products Laboratory, Madison, W.I., U.S.A.  Cui, F., Zahora, A.R. (2000). Effect of water repellent additive on the performance of ACQ treated decks. The International Research Group on Wood Protection, IRG/WP/40168.  Dahlen, J., Nicholas, D.D., Schultz, T.P. (2008). Water repellency and dimensional stability of southern pine decking treated with waterborne resin acids. Journal of Wood Chemistry and Technology, 28(1), 47-54.  Decena, A.S., Cruz, R.Z. de la. (1977). Study on controlling end splits in logs. Philippine Lumberman, 23(1), 18-22.  Derbyshire, H., Miller, E.R. (1981). The photodegradation of wood during solar irradiation. Part I: Effects on the structural integrity of thin wood strips. Holz als Roh- und Werkstoff, 39(8), 341-350.  Deskey, D., inventor, United States Plywood Corporation, assignee. (1942). Plywood panel. United States Patent US 2,286,068.   Doi, S., Horisawa, S. (2001). Fungi isolated from the surface of sugi (Crytomeria japonica) heartwood lumbers exposed at six sites of Japanese islands. In: Y., Imamura (Ed.), High-Performance Utilization of Wood for Outdoor Uses. Wood Research Institute, Kyoto University, Kyoto, Japan, 45-53.  Donath, S., Militz, H., Mai, C. (2007). Weathering of silane treated wood. Holz als Roh- und Werkstoff, 65(1), 35-42.  Dost, W.A. (1983). Preventing decay in exposed beams. Forest Products Journal, 33(10), 69-72.  Dubey, M.K., Pang, S.S., Walker, J. (2010). Color and dimensional stability of oil heat-treated radiata pinewood after accelerated UV weathering. Forest Products Journal, 60(5), 453-459.  Duncan, H.R. (1932). Incising of forest products material. Proceedings of the Thirty-Third Annual Convention of the American Railway Engineering Association Vol. 33, 549-552.  Dundas, K.B., White, R.R. (1972). The development of a treating procedure for western hemlock poles. Proceedings of the Sixty-Eighth Annual Meeting of the American Wood-Preservers’ Association, 169-174.   200  Dunningham, E.A., Plackett, D.V., Singh, A.P. (1992). Weathering of chemically modified wood natural weathering of acetylated radiata pine: preliminary results. Holz als Roh- und Werkstoff, 50(11), 429-432.  Eckelman, C.A. (1998). The shrinking and swelling of wood and its effect on furniture. Forestry and Natural Resource 163. Purdue University, West Lafayette, I.N., U.S.A. Retrieved from: https://www.extension.purdue.edu/extmedia/FNR/FNR-163.pdf  Eder, A. (2013). Conference charts the progress of WPCs. Compounding World, 5, 27-36. Retrieved from: http://www.beologic.com/_uploads/documents/Compounding%20World%20-%20Beologic.pdf  Edwards, R. (1980). Life history I. Nest initiation. In: Social wasps their biology and control. Rentokil Limited, East Grinstead, U.K, 25-61.  Eichler Siding (n.d.). Weldtex, combed plywood. Retrieved February 24, 2015 from: http://eichlersiding.com/wpblog/weldtex-combed-or-striated-plywood-now-available/  Elmendorf, A. (1950). Plywood panel. United States Patent US 2,514,318.  Elmendorf, A., Vaughan, T.W. (1960). Means for reducing the checking of Douglas fir plywood. Forest Products Journal, 10(1), 45-47.  Evans, P.D. (1989). Structural changes in pinus radiata during weathering. Journal of the Institute of Wood Science, 11(5), 172-181.  Evans, P.D. (2008). Weathering and photoprotection of wood. In: T.P., Schultz, H., Militz, M.H., Freeman, B., Goodell, D.D., Nicholas (Eds.), Development of Commercial Wood Preservatives, ACS Symposium Series 982. American Chemical Society, Washington, U.S.A., 69-117.  Evans, P.D. (2013). Weathering of wood and wood composites. In: R.M., Rowell (Ed.), Handbook of Wood Chemistry and Wood Composites. CRC Press, New York, U.S.A., 151-216.  Evans, P.D., Cullis, I., Morris, P.I. (2010). Checking of profiled southern pine and Pacific silver fir deckboards. Forest Products Journal, 60(6), 501-507.  Evans, P.D., Donnelly, C., Cunningham, R.B. (2003). Checking of CCA-treated radiata pine decking timber exposed to natural weathering. Forest Products Journal, 53(4), 66-71.  Evans, P.D., Kiguchi, M. (1998). Weathering and photoprotection of chemically modified wood. Proceedings of the Fourth Pacific Rim Bio-based Composites Symposium (Nov. 2-5), Bogor, Indonesia, 271-279. 201   Evans, P.D., Pirie, J.D.R., Cunningham, R.B., Donnelly, C.F., Schmalzl, K.J. (1994). A quantitative weathering study of wood surfaces modified by chromium VI and iron III compounds. Part 2. Image analysis of cell wall pit micro-checking. Holzforschung, 48(4), 331-336.  Evans, P.D., Urban, K. (2007). The effect of solar radiation on the surface checking of wood. The International Research Group on Wood Protection, IRG/WP/07-40356.  Evans, P.D., Urban, K., Chowdhury, M.J.A. (2008). Surface checking of wood is increased by photodegradation caused by ultraviolet and visible light. Wood Science and Technology, 42(3), 251-265.  Evans, P.D., Wingate-Hill, R., Barry, S.C. (2000). The effects of different kerfing and center-boring treatments on the checking of ACQ treated pine posts exposed to the weather. Forest Products Journal, 50(2), 59-64.  Evans, P.D., Wingate-Hill, R., Cunningham, R.B. (1997). The ability of physical treatments to reduce checking in preservative-treated splash pine posts. Forest Products Journal, 47(5), 51-55.  Evans, P.D., Wingate-Hill, R., Cunningham, R.B. (2009). Wax and oil emulsion additives: How effective are they at improving the performance of preservative-treated wood? Forest Products Journal, 59(1/2), 66-70.  Feist, W.C. (1983). Weathering and protection of wood. Proceedings of the Seventy-Ninth Annual Meeting of the American Wood-Preservers’ Association, 195-205.  Feist, W.C. (1990). Outdoor wood weathering and protection. In: R.M., Rowell, R.J., Barbour (Eds.), Archaeological Wood: Properties, Chemistry and Preservation. American Chemical Society, Washington, U.S.A., 263-298.  Feist, W.C. (2006). Exterior wood finishes. In: A.A., Tracton (Ed.), Coatings Technology Handbook, Third Edition. Taylor and Francis Group, Boca Ranton, F.L., U.S.A.   Feist, W.C., Mraz, E.A. (1978). Comparison of outdoor and accelerated weathering of  unprotected softwoods. Forest Products Journal, 28(3), 38-43.  Feng, J., Shi, Q.S., Chen, Y.B., Huang, X.M. (2014). Mold resistance and water absorption of wood/HDPE and bamboo/HDPE composites. Journal of Applied Sciences, 14(8), 776-783.  Findlay, W.P.K. (1938). The natural resistance to decay of some empire timbers. Empire Forestry Journal, 17(2), 249-597.  202  Flaete, P.O., Høibø, O.A., Fjærtoft, F., Nilsen, T.N. (2000). Crack formation in unfinished siding of aspen (Populus tremula L.) and Norway spruce (Picea abies (L.) Karst.) during accelerated weathering. Holz als Roh- und Werkstoff, 58(3), 135-139.  Floyd, F.L., (1983). Predictive model for cracking of latex paints applied to exterior wood surfaces. Journal of Coatings Technology, 55(696), 73-80.  Folvik, K., Sandland, K.M. (2003). Occurrence of checks in knots of Norway spruce timber during drying. The effect of various drying schedules. Proceedings of the Eighth International Union of Forest Research Organization Wood drying conference, Brasov, Romania.   Folvik, K., Sandland, K.M. (2005). Various wood properties influencing the development of checks in knots during drying. Rapport 58. Norsk Treteknisk Institutt (Norwegian Institute of Wood Technology), Oslo, Norway, 67-74.  Forest Products Laboratory. (1957). Shrinking and swelling of wood in use. U.S. Products Laboratory Report 736. USDA Forest Products Laboratory, Madison, W.I., U.S.A.  Fowlie, D.A., Preston, A.F., Zahora, A.R. (1990). Additives: an example of their influences on the performance and properties of CCA-treated southern pine. Proceedings of the Eighty-Sixth Annual Meeting of the American Wood-Preservers’ Association, 11-21.  Franciosi, G.F. (1956). Effect of incising beech cross ties, Forest Products Journal, 6(7), 264-270.  Francis, L., Norton, J. (2006). Australian timber pole resources for energy networks. A review. Technical Report. Department of Primary Industries and Fisheries, Queensland, Australia.  Freedonia Group. (2014a). Wood & competitive decking: United States. Freedonia Focus Reports US Collection. Freedonia Group, Cleveland, O.H., U.S.A.  Freedonia Group. (2014b). Wood decking: United States. Freedonia Focus Reports US Collection. Freedonia Group, Cleveland, O.H., U.S.A.  Freedonia Group. (2014c). Wood-plastic composite decking: United States. Freedonia Focus Reports US Collection. Freedonia Group, Cleveland, O.H., U.S.A.  Fujita, S. (1970). Microscopic observation of checks occurring under restrained drying. USDA Forest Service Translated Reports FPL-704. USDA Forest Service, Washington, D.C., U.S.A.  Furuno, T. (2001). Histochemical study on wood deterioration by UV irradiation and prevention of deterioration using Poly PEGMA. In: Y., Imamura (Ed.), High-Performance Utilization of Wood for Outdoor Uses. Wood Research Institute, Kyoto University, Kyoto, Japan, 71-84.  203  Gaby, L.I. (1963). Surface checking of white oak as related to mechanical processing. Forest Products Journal, 13(12), 529-532.  Ganguly, I., Eastin, I.L. (2009). Trends in the US decking market: a national survey of deck and home builders. The Forestry Chronicle, 85(1), 82-90.  Gardner, D.J., Han, Y., Song, W. (2008). Wood Plastic Composites Technology Trends. Proceedings of the Fifty-First International Convention of Society of Wood Science and Technology (November 10-12), Concepcion, Chile, 1-6.   Georgia-Pacific Corporation v U.S. Plywood-Champion Papers Inc., 415;416. (1971). United States Court of Appeals 446 F.2d 295, 1-10.  Gilfedder, J., Keating, W.G., Robertson, I. (1968). Influence of certain preservatives on pole splitting. Forest Products Journal, 18(1), 28-30.  Gilmer, E.E. (1934). Processing shingle shake. United States Patent US 1,943,597.   Gilmer, E.E. (1936). Shingle planing machine. United States Patent US 2,065,801.   Gilmer, E.E., inventor, Waters, R.W., assignee. (1933). Shingle planing machine. United States Patent US 1,910,895.   Goda, T. (2005). Japanese carpentry se-wari kerfing. Journal of Japanese Gardening, 46(7/8), 16-17.  Goodell, B., Pendlebury, A.J. (1991). Preservative treatment and field test monitoring of spruce pole stock: pressure and diffusible chemical treatments. Forest Products Journal, 41(4), 39-44.  Graham, R.D. (1973). Preventing and stopping internal decay of Douglas fir poles. Holzforschung 27(5), 168-173.  Graham, R.D., Estep, E.M. (1966). Effect of incising and saw kerfs on checking of pressure treated Douglas fir spar crossarms. Proceedings of the Sixty-Second Annual Meeting of the American Wood-Preservers’ Association, 155-160.  Greenwood, B.F de, Tainter, F.H. (1980). Comparative decay resistance of Paraguayan woods. Turrialba, 30(2), 137-139.   Greil, P., Lifka, T., Kaindl, A. (1998). Biomorphic cellular silicon carbide ceramics from wood: II. Mechanical properties. Journal of the European Ceramic Society, 18(14), 1975-1983.  204  GripDeck (n.d.). Anti-slip decking softwood board. Colchester, U.K. Retrieved February 24, 2015 from: http://www.gripdeck.co.uk/anti-slip-decking-softwood-board/  Groenier, J.S. (2006). Preservative-treated wood and alternative products in the forest service. Technology and development program report no. TE42G01. USDA Forest Service, Missoula, M.T., U.S.A.  Gutiérrez, M.V., Santiago, L.S. (2006). A comparison of sap flow measurements and potometry in two tropical lowland tree species with contrasting wood properties. Revista de Biología Tropical, 54(1), 73-81.  Hale, J.D. (1957). The anatomical basis of dimensional changes of wood in response to changes in moisture content. Forest Products Journal, 7(4), 140-144.  Hansmann, C., Deka, M., Wimmer, R., Gindl, W. (2006). Artificial weathering of wood surfaces modified by melamine formaldehyde resins. Holz als Roh- und Werkstoff, 64(3), 198-203.  Hao, B.Y., Avramidis, S. (2004). Annual ring orientation effect and slope of grain in hemlock timber drying. Forest Products Journal, 54(11), 41-49.  Hardie, A.D.K. (1974). Defects in the wood of fast-grown Eucalyptus grandis in Zambia. Commonwealth Forestry Review, 53(4), 310-317.  Harkom, J.F. (1928). Data on the checking of hard maple ties incised and non-incised treated at 50% m.c. and stacked in open piles for sixteen months. Report ID 12756. Forest Products Laboratory Forestry Branch Deptartment of the Interior, Canada.  Harkom, J.F., Rochester, G.H. (1930). Strength tests of creosoted Douglas fir beams. Circular 28. Forest Services Forest Products Laboratory Department of the Interior, Canada.  Harris, J.M. (1989). Spiral grain and wave phenomena in wood formation. Springer-Verlag, Berlin, Germany.  Heebink, T.B. (1959). Load-carrying capacity of deck boards for general purpose pallets. U.S. Products Laboratory Report 2153. USDA Forest Products Laboratory, Madison, WI, U.S.A.  Helsing, G., Graham, R.D. (1976). Saw kerfs reduce checking and prevent internal decay in pressure-treated Douglas-fir poles. Holzforschung, 30(6), 184-186.  Henry, W.T. (1970). Effect of incising on the reduction of serious checking in incised vs. unincised hardwood crossties. Proceedings of the Sixty-Sixth Annual Meeting of the American Wood-Preservers’ Association, 171-178.  205  Hill, C.A.S. (2006). Modifying the properties of wood. In: Wood modification chemical, thermal and other processes. John Wiley & Sons Limited, West Sussex, England, 19-44.   Hill, M.J., Moss, A.C., inventors, Stairway & Entrance Safety Products Ltd., assignee. (2000). Wooden decking element. UK Patent GB 2339437 (A).   Hoadley, R.B. (1990). Identifying wood accurate results with simple tools. The Taunton Press Incorporated, Newton, C.T., U.S.A.  Hon, D.N.S. (1981). Photochemical degradation of lignocellulosic materials. In: N., Grassie (Ed.), Developments in Polymer Degradation. Applied Science Publishers, London, U.K., 229-281.  Hong, J.H., Lee, J.J., Min, M.H., Ryu, S.M., Lee, D.H., Kim, G.H., Kim, J.J. (2014). 6-Pentyl-α-pyrone as an anti-sapstain compound produced by trichoderma gamsii KUC1747 inhibits the germination of ophiostomatoid fungi. Holzforschung, 68(7), 769-774.  Huang, X.N., Kocaefe, D., Kocaefe, Y., Boluk, Y., Pichette, A. (2012). Study of the degradation behavior of heat-treated jack pine (Pinus banksiana) under artificial sunlight irradiation. Polymer Degradation and Stability, 97(7), 1197-1214.  Iimura, Y. (2001). Maintenance of timber structures (including joint areas) in Japan. In: High-Performance Utilization of Wood for Outdoor Uses. Wood Research Institute, Kyoto University, Kyoto, Japan, 171-177.  Illston, J.M., Dinwoodie, J. M., Smith, A. A., (1979). Concrete, timber and metals. The nature and behaviour of structural materials. Van Nostrand Reinhold Company Ltd., Wokingham, U.K.  Imamura, Y. (1993). Morphological changes in acetylated wood exposed to weathering. Wood Research, 79, 54-61.  Izekor, D.N., Fuwape, J.A. (2010). Evaluation of weathering performance of teak (Tectona grandis L. F.) wood in outdoor application. African Journal of General Agriculture, 6(4), 219-227.  Jagadamma, S., Lal, R. (2010). Distribution of organic carbon in physical fractions of soils as affected by agricultural management. Biology and Fertility of Soils, 46(6), 543-554.  Jermer, J., Bergman, O. (1992). Glued laminated poles-progress report after 12 years’ testing. The International Research Group on Wood Protection, IRG/WP/3685-92.  Johnson, B.R., Ibach, R.E., Baker, A.J. (1992). Effect of salt water evaporation on tracheid separation from wood surfaces. Forest Products Journal, 42(7/8), 57-59.  206  Johnson, R.P.A., Bradner, M.I. (1932). Properties of western larch and their relation to uses of the wood. Technical Bulletin No. 285. U.S. Department of Agriculture, Washington, U.S.A.  Jureit, J.C., Kushner, B.H., inventor, Automated Building Components Inc., assignee. (1968). Apparatus for forming reinforced wooden members. United States Patent US 3,419,205.   Jureit, J.C., Kushner, B.H., inventor, Automated Building Components Inc., assignee. (1970). Method of forming reinforced wooden members. United States Patent US 3,540,107.   Jureit, J.C., Kushner, B.H., inventor, Automated Building Components Inc., assignee. (1977). Connector plate. United States Patent US 4,031,803.   Jureit, J.C., Kushner, B.H., inventor, Automated Building Components Inc., assignee. (1979). Connector plate. United States Patent US 4,165,672.  Kilian, T., Engelsing, K., Heidemeyer, P., Bastian, M. (2010). Innovation in the analysis of weathering induced color changes in plastics. Material Testing Product and Technology News, 40(86), 4-10.  Kim, C.H., Kang, C.W., Kang, S.G., Kang, H.Y. (2012). Effect of pretreatments on reducing surface cracks of heat-treated western hemlock roundwoods. Journal of the Korean Wood Science and Technology, 40(5), 343-351.  Kirker, G.T., Blodgett, A.B., Lebow, S.T., Clausen, C.A. (2012). Above Ground Field Evaluation and GC-MS Analysis of Naturally Durable Wood Species. The International Research Group on Wood Protection, IRG/WP/12-10764.  Kishino, M., Nakano, T. (2004). Artificial weathering of tropical wood. Part 1: Changes in wettability. Holzforschung, 58(5), 552-557.  Knight, E. (1961). The causes of warp in lumber seasoning. Western Dry Kiln Clubs 13th Meeting, Medford, O.R., U.S.A. Retrieved from:  https://ir.library.oregonstate.edu/xmlui/handle/1957/4870  Knight, R.A.G., Doman, L.S. (1962). Reducing the surface checking of plywood exposed to weather. Journal of the Institute of Wood Science, 10(12), 66-73.  Koehler, A. (1955). Guide to determining slope of grain in lumber and veneer. U.S. Products Laboratory Report 1585. USDA Forest Products Laboratory, Madison, W.I., U.S.A.     207  Konica Minolta. (2007). Colors also look different according to the subject conditions and environment. In: Precise colour communication color control from perception to instrumentation. Konica Minolta, Japan. Retrieved from: http://www.konicaminolta.com/instruments/knowledge/color/pdf/color_communication.pdf  Konica Minolta. (2014). Spectrophotometer CM-2600d simply expands the boundary in colour control. Konica Minolta, Japan. Retrieved from: http://www.konicaminolta.com/instruments/download/catalog/color/pdf/cm2600d_catalog_eng.pdf  Kretschmann, D.E. (2010). Chapter 6: Commercial lumber, round timbers, and ties. In: R.J., Ross (Ed.), Wood handbook - wood as an engineering material, USDA General Technical Report 190. Forest Products Laboratory, Madison, W.I., U.S.A.  Kubler, H. (1987). Growth stresses in trees and related wood properties. Forestry Abstracts, 48(3), 131-189.  Kurisaki, H. (2004). Effect of kerfing on preventing the surface check and reducing inner decay risk in sugi post. Proceedings of the Third International Symposium on Surfacing and Finishing of Wood, 336-341.  Lamb, F.M. (1992). Splits and cracks in wood. Western Dry Kiln Association 43rd Meeting, 16-24. Retrieved from:  https://ir.library.oregonstate.edu/xmlui/handle/1957/5302  Leavengood, S., Swan, L. (1999). Western juniper drying project summaries, 1993-96. General Technical Report - Pacific Northwest Research Station PNW-GTR-475. USDA Forest Service, Portland, O.R., U.S.A.  Lee, N.H., Li, C.Y., Zhao, X.F., Park, M.J. (2010). Effect of pretreatment with high temperature and low humidity on drying time and prevention of checking during radio-frequency/vacuum drying of Japanese cedar pillar. Journal of Wood Science, 56(1), 19-24.  Lee, N.H., Zhao, X.F., Shin, I.H., Park, M.J., Park, J.H., Park, J.S. (2011a). Characteristics of the radio-frequency/vacuum drying of heavy timbers for post and beam of Korean style housings Part I: For Japanese larch round logs with 150 mm and 210 mm in diameter and 2500 mm in length. Journal of the Korean Wood Science and Technology, 39(2), 125-131.  Lee, N.H., Zhao, X.F., Shin, I.H., Park, M.J., Park, J.H., Park, J.S. (2011b). Characteristics of the radio-frequency/vacuum drying of heavy timbers for post and beam of Korean style housings Part II: For Korean red pine heavy timbers with 250×250 mm, 300×300 mm in cross section and 300 mm in diameter, and 3600 mm in length. Journal of the Korean Wood Science and Technology, 39(2), 132-139. 208   Leney, L. (1964). Checking of planed and rough red oak during kiln drying. Forest Products Journal, 14(3), 103-105.  Lepitre, C., Mariaux, A. (1965). Splits in gaboon wood logs: influence of S-clamps and anti-split products. Bois et Forets des Tropiques, 104(11/12), 31-51.   Levi, M.P., Coupe, C., Nicholson, J. (1970). Distribution and effectiveness in Pinus sp. of a water repellent additive for water-borne wood preservatives. Forest Products Journal, 20(11), 32-37.  Liese, W. (1961). Cost and success of methods of protecting beech logs stored in the forest. Forstarchiv, 32(11), 234-236.  Lindgren, P. (1975). Inspection of experimental Douglas-fir poles. Dorena tap line, Cottage Grove Oregon. In: G. L., Downey, H. E., Troxell (Eds.), Proceedings of the Sixth Wood Pole Institute Conference. Colorado State University, Fort Collins, C.O., U.S.A., 41-64.  Listyanto, T., Ando, K., Yamauchi, H., Hattori, N. (2013). Microwave and steam injection drying of CO2 laser incised sugi lumber, Journal of Wood Science, 59(4), 282-289.  Liu, Y.L. (2005). The treatment of wood surface for artificial larch: Treatment by alcohol solution of phenolic resin with low molecular weight. Journal of Northeast Forestry University, 33 (2), 89-90.  Lloyd, R.A., Stamm, A.J. (1958). Effect of resin treatment and compression upon the weathering properties of veneer laminates. Forest Products Journal, 8(8), 230-235.  Love, C.S., Morrell, J.J. (2012). Effects of end plates on reducing checking of pentachlorophenol-treated Douglas-fir crossarms. Forest Products Journal, 62(2), 121-123.  Mackay, J.F.G. (1972). The occurrence, development and control of checking in Tasmanian Eucalyptus obliqua. Holzforschung, 26(4), 121-124.  Mackay, J.F.G. (1973). Surface checking and drying behaviour of Pinus radiata sapwood boards treated with CCA preservative. Forest Products Journal, 23(9), 92-97.  MacLean, J.D. (1956). Seasoning and checking of timbers before and after preservative treatment. Proceedings of the Fifty-Second Annual Meeting of the American Wood-Preservers’ Association, 35-48.  Majewski, T., Rajaraman, H., inventors, PolyOne corporation, assignee. (2013). Low gloss capstock. United States Patent US 2013/0303644 A1.   209  Makovická, P.J., Čunderík, I. (2006). Chosen physical properties of spruce juvenile wood (Picea abies Karst.). In: S., Kurjatko, J., Kúdela, R., Lagana (Eds.), Wood Structure and Properties ’06. Arbora Publishers, Zvolen, Slovakia, 93-97.  Mallett, J. (2012). Full field strain mapping of profiled radiata pine decking boards subjected to accelerated weathering (Bachelor of Engineering thesis). The Australian National University Canberra, Australia.  Malpass, G. (1991). Product Delivery. In: J.D., Lousier (Ed.), Yellow cypress: Can we grow it? Can we sell it?, FRDA report 171. Forestry Canada, Victoria, B.C., Canada, 5-7.  Mankey, F. (1884). Surface cut lumber and process of making the same. United States Patent US 309,071.   Martin, G.A. (1892). Making and setting poles. In: Fences, gates and bridges a practical manual. Orange Judd Company, New York, U.S.A., 98-99.  Mater, M.H. (1972). Boring 40-foot-long utility poles for conduit passage. Forest Products Journal, 22(6), 28-31.  McCarthy, R., Chin, C.W., Greaves, H. (1982). Above ground test of preservative treatments and surface coatings for timber. Architectural Science Review, 25(1), 1-9.  McFarling, S.M., Morris, P.I. (2005). High performance wood decking. Proceedings of the Twenty-Sixth Annual Meeting of the Canadian Wood Preservers’ Association, 99–108.  McFarling, S.M., Morris, P.I., Knudson, R.M. (2009). Extracting greater value from subalpine fir: Profiled decking. Forest Products Journal, 59(3), 24-28.  McIntosh, D.C. (1955). Shrinkage of red oak and beech. Forest Products Journal, 5(2), 355-359.   McLaughlan, J.M. (1991). Properties of treated and untreated Pinus radiata plywood after 12 years weathering. New Zealand Journal of Forestry Science, 21(1), 96-110.  McMillen, J.M. (1955). Drying stresses in red oak. Forest Products Journal, 5(1), 71-79.   McMillin, C.W. (1969). Performance evaluation of stained, rough-sawn southern pine siding. Forest Products Journal, 19(2), 51-52.  Miklečic´, J., Jirouš-Rajkovic´, V. (2011). Accelerated weathering of coated and uncoated beech wood modified with citric acid. Drvna Industrija, 62(4), 277-282.   210  Miklečic´, J., Jirouš-Rajkovic´, V., Pervan, S., Grujic´, S. (2010). Oils usage in finishing of thermally modified wood in outdoor aplications. Proceedings of the international symposium wood is good: transfer of knowledge in practice as a way out of the crisis. University of Zagreb Faculty of Forestry, Zagreb, Croatia, 89-96.  Miles, P.D., Smith, B.W. (2009). Specific gravity and other properties of wood and bark for 156 tree species found in North America. USDA Forest Service Research Note NRS-38. U.S. Forest Service, Newtown Square, P.A., U.S.A.  Miller, R.B., Wiedenhoeft, A.C., Williams, R.S., Stockman, W., Green, F. (2003). Characteristics of ten tropical hardwoods from certified forests in Bolivia. Part II. Natural durability to decay fungi. Wood and Fiber Science, 35(3), 429-433.  Miniutti, V.P. (1964). Microscale changes in cell structure at softwood surfaces during weathering. Forest Products Journal, 14(12), 571-576.  Miniutti, V.P., Mraz, E.A., Black, J.M. (1961). Measuring the effectiveness of water-repellent preservatives. Forest Products Journal, 11(10), 453-462.  Mitsui, K., Tsuchikawa, S. (2005). Low atmospheric temperature dependence on photodegradation of wood. Journal of Photochemistry and Photobiology, 8(2), 84-88.  Morrell, J.J., Newbill, M.A. (1986). Kerfing to prevent decay of Douglas fir poles: an update. Forest Products Journal, 36(5), 46-48.  Morris, P.I., Ingram, J.K. (1996). Dimensional stability of amabilis fir and western hemlock decking. Report preprared for project 17-10-O-009. Forintek Canada Corporation, Vancouver, B.C., Canada.   Morris, P.I., Ingram, J.K. (2002). Field testing of wood preservatives in Canada. XI: Nine-year inspection of the CITW decking test. Proceedings of the Twenty-Third Annual Meeting of the Canadian Wood Preservers’ Association, 156-169.  Morris, P.I., Ingram, J.K., Ruddick, J.N.R., Choi, S.M. (2004). Protection of untreated wood by adjacent CCA-treated wood. Forest Products Journal, 54(3), 29-32.  Morris, P.I., Lam, F., McKay, J.F.G. (1990). Production of higher-value treatable SPF lumber. Report FRR 119. Forintek Canada Corporation, Vancouver, B.C., Canada.  Morris, P.I., McFarling, S. (2008). Field testing of wood products in Canada XVII: High-performance profiled wood decking. Proceedings of the Twenty-Ninth Annual Meeting of the Canadian Wood Preservers’ Association, Campbellville, O.N., Canada, 72-82.   211  Morris, P.I., Morrell, J.J., Ruddick, J.N.R. (1994). A review of incising as a means of improving treatment of sawnwood. The International Research Group on Wood Protection, IRG/WP/94-40019.  National Hardwood Lumber Association. (2011). Rules for the measurement & inspection of hardwoods & cypress. National Hardwood Lumber Association, Memphis, T.N., U.S.A.  Negri, M., Pollini, C., Tessadri, B., Gaeti, N., Sandak, J., Teischinger, A. (2005). Measuring the quality of large timber: an integrated approach. Lignovisionen, 13, 77-82.  Nejad, M., Cooper, P. (2011a). Exterior wood coatings. Part-1: Performance of semitransparent stains on preservative-treated wood. Journal of Coatings Technology and Research, 8(4), 449-458.  Nejad, M., Cooper, P. (2011b). Exterior wood coatings. Part-2: Modeling correlation between coating properties and their weathering performance. Journal of Coatings Technology and Research, 8(4), 459-467.  Nejad, M., Shafaghi, R., Ali, H., Cooper, P. (2013). Coating performance on oil-heat treated wood for flooring. BioResources, 8(2), 1881-1892.  Norlander, N.E., Knowles, R.A., inventors, Weyerhaeuser Company, assignee. (1970). Method and apparatus for making simulated hand split shakes. United States Patent US 3,512,562.   Nyrud, A.Q., Roos, A., Rødbotten, M. (2008). Product attributes affecting consumer preference for residential deck materials. Journal of Forest Research, 38(6), 1385-1396.  Nystrom, E.E. (1995). Board for use in constructing a flooring surface. United States Patent US 5,474,831.   Ofori, J., Brentuo, B. (2010). Green moisture content, basic density, and shrinkage characteristics of the wood of Alstonia boonei, Antrocaryon micraster, Bombax buonopozense, Dialium aubrevillei and Sterculia rhinopetala. Ghana Journal of Forestry, 26, 41-49.  O’Neill, M. (2010). ANOVA & REML a guide to linear mixed models in an experimental design context. Statistical Advisory & Training Service Pty Ltd., Blackheath, Australia. Retrieved from: http://www.stats.net.au/dropbox/ANOVA%20&%20REML%20manual.pdf  Opoku, F.Y. (2007). An investigation into the suitability of selected lesser utilized Ghanaian hardwoods for use as outdoor furniture and decking (Master of Philosophy Thesis). Buckinghamshire Chilterns University College, High Wycombe, U.K.  212  Osborne, L.D. (1965). Reducing end-splitting of sawn timber. CSIRO Forest Products News Letter, 315, 6-8.  Ottoson, M. (2009). Weldtex: The plywood panel that grows old gracefully. Retrieved from: http://www.heritagerevival.com/biography.html  Page, D. (2002). Wood preservation. Brush on 'preservatives' - do they prevent timber deterioration? Wood Processing Newsletter, 32, 2-4. Wood Technology Group Forest Research, Rotorua, New Zealand.  Peck, E.C. (1953). Reducing checking in heavy white oak shipbuilding material during storage and construction. Journal of the Forest Products Research Society, 3(4), 22-23.  Perem, E. (1971). Checking and splitting of hardwood rail ties in seasoning and service. Canadian Forestry Service No. 1293. Canadian Forestry Service, Canada.  Plackett, D.V., Chittenden, C.M., Preston, A.F. (1984). Exterior weathering trials on Pinus radiata roofing shingles. New Zealand Journal of Forestry Science, 14(3), 368-381.  Plackett, D.V., Dunningham, E.A., Singh, A.P. (1992). Weathering of chemically modified wood. Accelerated weathering of acetylated radiata pine. Holz als Roh- und Werkstoff, 50(4), 135-140.  Popular Mechanics Press. (1942). Hints on building sturdy fences. Popular Mechanics Garden Book. Popular Mechanics Press, Chicago, U.S.A, 43.  R Development Core Team (2011). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.   Ratu, R.N. (2009). Development and testing of a weatherometer to accelerate the checking of wood (Master of Science Thesis). University of British Columbia, Vancouver, B.C., Canada. Retrieved from: https://circle.ubc.ca/  Ratu, R.N., Weizenegger, J., Evans, P.D. (2007). Preliminary observations of the effect of kerfing on the surface checking and warping of flat sawn southern pine decking. The International Research Group on Wood Protection, IRG/WP/07-20360.  Ribarits, S.G., Evans, P.D. (2010). Finite element modelling of the checking of wood exposed to accelerated weathering. The International Research Group on Wood Protection, IRG/WP/10-20459.  Rietz, R.C. (1961). Accelerated weathering of red oak treated with various preservatives used to treat crossties. Forest Products Journal, 11(12), 567-575.  213  Ross, A., Bussjaeger, S., Carlson, R., Feist, W.C. (1992). Professional finishing of CCA pressure-treated wood. American Painting Contractor, 69(7), 107-114.  Roux, M.L., Wozniak, E., Miller, E.R., Boxall, J., Böttcher, P., Kropf, F., Sell, J. (1988). Natural weathering of various surface coatings on five species at four European sites. Holz als Roh- und Werkstoff, 46(5), 165-170.  Roy, H.G. (1955). Artificial lumber products and their manufacture. United States Patent US 2,717,420.   Ruddick, J.N.R. (1988). Kerfing reduces checking in ACA-treated Western white spruce poles. The International Research Group on Wood Protection, IRG/WP/3477.  Ruddick, J.N.R., Ross, N.A. (1979). Effect of kerfing on checking of untreated Douglas-fir pole sections. Forest Products Journal, 29(9), 27-30.  Saary, J.M. (2008). Radar plots: a useful way for presenting multivariate health care data. Journal of Clinical Epidemiology, 61(4), 311-317.  Sadoh, T. (2001). Claim Ga Ohi Curui – Doh Shite Fugesu. In: “Ki Ga Wakaru”, Gakugei Shuppan-Sha. Kyoto, Japan, 116-119.  Saitoh, S., Nakajima, Y., Gensai, H., Iwatani, I. (2000). Kiln drying test of full scale sugi (Cryptomeria japonica) boxed heart timber. Transaction of the Annual Meeting of Kanto branch of the Japanese Forestry Society, 52, 173-176.  Sandberg, D. (1996). The influence of pith and juvenile wood on proportion of cracks in sawn timber when kiln dried and exposed to wetting cycles. Holz als Roh- und Werstoff, 54(3), 152.  Sandberg, D. (1997). Radially sawn timber. The influence of annual ring orientationon crack formation and deformation in water soaked pine (Pinus sylvestris L.) and spruce (Picea abies (L.) Kast.) timber. Holz als Roh- und Werkstoff, 55(3), 175-182.  Sandberg, D. (1999). Weathering of radial and tangential wood surfaces of pine and spruce. Holzforschung, 53(4), 355-354.  Sandberg, D. (2005). Distortion and visible crack formation in green and seasoned timber: influence of annual ring orientation in the cross section. Holz als Roh- und Werkstoff, 63(1), 11-18.  Sandberg, D., Söderström, O. (2006). Crack formation due to weathering of radial and tangential sections of pine and spruce. Wood Material Science and Engineering, 1(1), 12-20. 214   Scheffer, T.C., Morrell, J.J. (1998). Natural durability of wood: A worldwide checklist of species. Research contribution 22. Forest Research Laboratory, Oregon State University, Corvallis, O.R., U.S.A.   Schniewind, A.P. (1959). Transverse anisotropy of wood: a function of gross anatomic structure. Forest Products Journal, 9(10), 350-59.  Schniewind, A.P. (1963). Mechanism of check formation. Forest Products Journal, 13(11), 475-480.  Schoeman, M., Dickinson, D. (1997). Growth of Aureobasidium pullulans on lignin breakdown products at weathered wood surfaces. Mycologist, 11(4), 168-172.  Schulze, M., Grogan, J., Vidal, E. (2008). Forest certification in Amazonia: standards matter. Oryx, 42(2), 229-239.  Seifert, K. (1964). Changes in the chemical composition of wood by the blue stain fungus A. pullulans. Holz als Roh- und Werkstoff, 22(11), 405-409.  Selbo, M.L., Knauss, A.C., Worth, H.E. (1965). After two decades of service, glulam timbers show good performance. Forest Products Journal, 15(11), 466-472.  Sell, J., Feist, W.C. (1986). Role of density in the erosion of wood during weathering. Forest Products Journal, 36(3), 57-60.  Sell, J., Leukens, U. (1971). Investigations on weathered wood surfaces. II. Weathering phenomena on unprotected timbers. Holz als Roh- und Werkstoff, 29(1), 23-31.  Sell, J., Wälchli, O. (1969). Changes in the surface texture of weather-exposed wood. Material und Organismen, 4(2), 81-87.  Shida, S., Ono, H., Mikami, T., Takahashi, H. (1992). Utilization and evaluation of exterior wood IV. Slipperiness of wood decks for floating piers. Mokuzai Gakkaishi, 38(9), 835-840.  Shook, S.R., Eastin, I.L. (2001). A characterization of the U.S. residential deck material market. Forest Products Journal, 51(4), 28-36.  Shunk, B.H. (1976). Drying southern hardwood timbers, crossties, and posts. Forest Products Journal, 26(4), 51-57.  Simpson, W.T., Schroeder, J.G. (1980). Kiln-drying hardwood dimension parts. USDA Forest Service Research Paper FPL-388. Forest Products Laboratory, U.S.A.  215  Sing, N.N., Zulkharnain, A., Azman, H.R., Assim, Z., Husaini, A. (2014). Bioremediation of PCP by trichoderma and cunninghamella strains isolated from sawdust. Brazilian Archives of Biology and Technology, 57(6), 811-820.   Skaar, C. (1972). Hygroscopic shrinking and swelling of wood. In: Water in wood. Syracuse University Press, Syracuse, U.S.A., 73-125.  Spear, M., Walker, J.C.F. (2006). Dimensional instability in timber. In: J.C.F., Walker (Ed.), Primary wood processing: principles and practice. Springer-Verlag, Heidelberg, Germany, 95-120.  Stamm, A.J. (1959). The effect of polyethylene glycol on the dimensional stability of wood. Forest Products Journal, 9(10), 375-381.  Stamm, A.J. (1964). Wood and cellulose science. Ronald Press Company, New York, U.S.A.  Stamm, A.J. (1965). Modification of wood for improved finishing. Official Digest, 37(6), 707-717.  Stamm, A.J., Seborg, R.M. (1939). Resin-Treated Plywood. Industrial and Engineering Chemistry, 31(7), 897-902.  Stanzl-Tschegg, S. E., Tan, D. M., Tschegg, E. K. (1996). Mode II fracture tests on spruce wood. Mokuzai Gakkaishi, 42(7), 642-650.  Stehr, M., Östlund, S. (2000). An investigation of the crack tendency on wood surfaces after different machining operations. Holzforschung, 54(4), 427-436.  Stirling, R., Morris, P.I. (2012). Treatments to minimize extractives stain in western red cedar. BioResources, 7(2), 2461-2468.   Stubbersfield, E.M. (2009). A board for decking. Australian Patent AU 2003204845.  Sudiyani, Y., Ryu, J.Y., Hattori, N., Imamura, Y. (2001). Phenolic resin treatment of wood for improving weathering properties. In: Y., Imamura (Ed.), High-Performance Utilization of Wood for Outdoor Uses. Wood Research Institute, Kyoto University, Kyoto, Japan, 85-96.  Suprapti, S. (2010). Decay resistance of 84 Indonesian wood species against fungi. Journal of Tropical Forest Science, 22(1), 81–87.   Tarkow, H., Stamm, A.J. (1955). Acetylated wood information reviewed and reaffirmed. U.S. Products Laboratory Report 1593. USDA Forest Products Laboratory, Madison, W.I., U.S.A.  216  Tenorio, C., Moya, R. (2011). Kiln drying of Acacia mangium Willd wood: considerations of moisture content before and after drying and presence of wet pockets. Drying Technology, 29(15), 1845-1854.  Tiemann, H.D. (1941). What is 'spiral grain'? Southern Lumberman, 63(11), 55-57.  Timell, T.E. (1986). Compression wood in Gymnosperms. Volume 3. Ecology of compression wood formation, silviculture and compression wood, mechanism of compression wood action, compression wood in the lumber and pulp and paper industries, compression wood induced by the balsam woolly aphid, opposite wood. Springer-Verlag, Berlin, Germany.  Tisseverasinge, A.E.K. (1967). A preliminary report on the prevention of splitting in poles of Eucalyptus grandis using gang nails. Ceylon Forester, 8(1/2), 20-22.  Trinh, H.M., Militz, H., Mai, C. (2012). Modification of beech veneers with N-methylol melamine compounds for the production of plywood: natural weathering. European Journal of Wood and Wood Products, 70(1), 279-286.  Turner, J., Dawson, B., Sellier, D., Kimberley, M., Harrington, J. (2008). Do stability grooves in weatherboards reduce cupping in service? Wood Processing Newsletter, 41, 1-3. Scion, Rotorua, New Zealand.   Urban, K., Evans, P.D. (2005). Preliminary observations of the effect of growth ring orientation on the surface checking of flat sawn Southern pine decking. The International Research Group on Wood Protection, IRG/WP/05-20313.  Uzunovic, A., Byrne, T., Gignac, M., Yang, D.Q. (2008). Wood discolourations & their prevention with an emphasis on bluestain. Forintek Special Publication SP-50. Forintek Canada Corporation, Vancouver, B.C., Canada.  Virta, J., Koponen, S., Absetz, I. (2005). Cupping of wooden cladding boards in cyclic conditions - a study of boards made of Norway spruce (Picea abies) and Scots pine sapwood (Pinus sylvestris). Wood Science and Technology, 39(6), 431-438.  VSN International. (2009). GenStat for Windows 12th Edition. VSN International, Hemel Hempstead, U.K.   Welzbacher, C., Brischke, C., Rapp, A., Koch, S., Hofer, S. (2009). Performance of thermally modified timber (TMT) in outdoor application - durability, abrasion and optical appearance. Drvna Industrija, 60(2), 75-82.  Wengert, E.M. (1966). Parameters for predicting maximum surface temperature of wood in exterior exposure. USDA Forest Service Research Paper FPL 62. Forest Products Laboratory, Madison, W.I., U.S.A. 217  Wiemann, M.C. (2010). Chapter 2: Characteristics and availability of commercially important woods. In: R.J., Ross (Ed.), Wood handbook - wood as an engineering material, USDA General Technical Report 190. Forest Products Laboratory, Madison, W.I., U.S.A.  Williams, R.S., Feist, W.C. (1999). Water repellents and water-repellent preservatives for wood. USDA Forest Service General Technical Report FPL-GTR-109. Forest Products Laboratory, Madison, W.I., U.S.A.  Williams, R.S., Knaebe, M. (1995). The bark-side/pith-side debate. The finish line a forest products laboratory finishing factsheet. Forest Products Laboratory, Madison, W.I., U.S.A.   Williams, R.S., Miller, R., Gangstad, J. (2001). Characteristics of ten tropical hardwoods from certified forests in Bolivia. Part I: Weathering characteristics and dimensional change. Wood and Fiber Science, 33(4), 618-626.   Witten, E., Kruas, T., Kühnel, M. (2014). Composites market report 2014: Market developments, trends, challenges and opportunities. Carbon composites, Germany; AVK – federation of reinforced plastics, Germany. Retrieved from:  http://www.carbon-composites.eu/sites/carbon-composites.eu/files/anhaenge/14/10/28/ccev-avk-marktbericht_2014_english.pdf  Woodcarvers guild (2003). The woodcarvers guild book of wall panelling. Woodcarvers guild, Bridgwater, U.K. Retrieved from: http://www.wallpanellingcorp.com/Drama%20Production%20and%20Book/Book%20of%20Wall%20Panelling.pdf  Woodhead, W.D. (1969). Natural finishes on wood. CSIRO Forest Products News Letter, 365, 2-3.  Xiao, Z.F., Xie, Y.J., Adamopoulos, S., Mai, C. (2012). Effects of chemical modification with glutaraldehyde on the weathering performance of Scots pine sapwood. Wood Science and Technology, 46(4), 749-767.  Xie, Y.J., Krause, A., Militz, H., Mai, C. (2008). Weathering of uncoated and coated wood treated with methylated 1,3-dimethylol-4,5-dihydroxyethyleneurea (mDMDHEU). Holz als Roh- und Werkstoff, 66(6), 455-464.  Yamazaki, N., Atsushi, N. (1996). Todomatsu kanbatsuzai Kara no syou kakuzai (1). Rinsanshi dayori journal, 5, 11-15. Retrieved from: http://www.fpri.hro.or.jp/rsdayo/20987002001.pdf    218  Yang, D.Q., Normand, D. (2012). Element 4: Hardwood initiative - Development of new processes and technologies in the hardwood industry. Best practices to avoid hardwood checking. Part I. Hardwood checking-the causes and prevention. Transformative Technologies Program Project No. TT4.3.03 201005167. FPInnovations, Quebec, Q.C., Canada.   Yang, J.L., Waugh, G. (2001). Growth stress, its measurement and effects. Australian Forestry, 64(2), 127-135.  Yata, S. (2001). Occurrence of drying checks in softwood during outdoor exposure. In: Y., Imamura (Ed.), High-Performance Utilization of Wood for Outdoor Uses. Wood Research Institute, Kyoto University, Kyoto, Japan, 65-70.  Yata, S., Tamura, T. (1995). Histological changes of softwood surfaces during outdoor weathering. Mokuzai Gakkaishi, 41(11), 1035-1042.  Yeo, H., Shim, K.B., Lee, D.S., Park, J.S., Eom, C.D., Han, Y.J., Lee, H.W., Lee, N.H., Smith, W.B., Park, J.H., Park, M.J. (2007). Effects of center boring and kerf treatment on kiln-drying of larch square and round timbers. Forest Products Journal, 57(11), 85-92.  Yusuf, S., Imamura, Y., Takahashi, M., Minato, K. (1995). Weathering properties of chemically modified wood with some cross-linking agents and its decay resistance after weathering. Mokuzai Gakkaishi, 41(8), 785-793.  Zahora, A.R. (1991). Interactions between water-borne preservatives and emulsion additives that influence the water repellency of wood. The International Research Group on Wood Protection, IRG/WP/2374.  Zahora, A.R. (1992). A water repellent additive’s influence on field performance of southern yellow pine lumber. Proceedings of the Eighty-Eighth Annual Meeting of the American Wood-Preservers’ Association, 148-159.  Zahora, A.R. (2000). Long-term performance of a wax type additive for use with water-borne pressure preservative treatments. The International Research Group on Wood Protection, IRG/WP/40159.  Ziglio, A.C., Gonçalves, D. (2014). On the use of capsaicin as a natural preservative against fungal attack on Pinus sp. and Hymenaea sp. woods. Materials Research, 17(1), 271-274.   219  Appendices Appendices can be found in the DVD attached to this thesis. Appendix 1: Checking and cupping of profiled and treated pacific silver fir and lodgepole pine deck-boards exposed to the weather for one year.  Concave  Convex  Species/Profile check length (mm) check depth (mm) check width (mm) cupping (mm) check length (mm) check depth (mm) check width (mm) cupping (mm) Lodgepole Pine/ Rib-eased edge  513 7.3 0.3 0.3 971 10.5 0.5 0.5 Lodgepole Pine/ Rib-flat edge and center  532 5.3 0.3 0.2 1266 10.3 0.7 0.3 Lodgepole Pine/ Ripple to edge  356 9.2 0.5 0.1 936 12.2 1.1 0.4 Lodgepole Pine/ Ripple flat edge  286 4.3 0.3 0.6 634 6.7 0.5 1.2  Pacific Silver Fir/ Rib-eased edge  263 5.1 0.4 0.1 320 4.0 0.2 0.2 Pacific Silver Fir/ Rib-flat edge and center 526 5.6 0.6 0.2 591 4.3 0.6 0.4 Pacific Silver Fir/ Ripple to edge  483 4.0 0.3 0.4 668 4.3 0.4 0.4 Pacific Silver Fir/ Ripple flat edge  711 6.8 0.6 0.5 693 6.0 0.5 0.8 Note: Green colour denotes good performance for each check or cupping parameter category. Red colour denotes poor performance. (redrawn from McFarling and Morris 2008)  220  Appendix 2: Checking and cupping of the thirteen best combinations of radiata pine and mountain ash panels treated with preservatives, water-repellents, and coatings exposed outdoors for two years. Treatment Radiata Pine Preservative Coating Checks Cupping Hager's Royal Process No treatment No checking Slight to moderate cupping No treatment Primer, British Paints Four Seasons (2 coats) Slight checks bottom end No cupping No treatment Dulux Timbercolour (2 coats) Slight to moderate check Slight to severe cupping 3 min dip in WRP Primer, undercoat, Berger Ext. Full Gloss Slight to moderate check Slight cupping 3 min dip in WRP Primer, British Paints Four Seasons (2 coats) Very slight checks bottom end No cupping CCA pressure treated Primer, undercoat, Berger Ext. Full Gloss Slight checks bottom Very slight cupping CCA pressure treated+ 3 min dip in WRP Primer, undercoat, Berger Ext. Full Gloss Slight checks Very slight cupping CCA pressure treated+ 3 min dip in WRP Primer, British Paints Four Seasons (2 coats) Very slight checks bottom end No cupping Treatment Mountain Ash Preservative Coating Checks Cupping No treatment Primer, undercoat, Berger Ext. Full Gloss Slight checks bottom end No cupping No treatment Primer, British Paints Four Seasons (2 coats) Slight checks bottom end Very slight cupping No treatment Dulux Timbercolour (2 coats) Slight check (1 severe check along growth ring) Slight cupping 3 min dip in WRP Primer, undercoat, Berger Ext. Full Gloss Very slight checks bottom end No cupping 3 min dip in WRP Primer, British Paints Four Seasons (2 coats) Very slight checks bottom end Very slight cupping Note: WRP stands for Water Repellant Preservative (Cellavi WR 113). (redrawn from McCarthy et al. 1982)  221  Appendix 3: Woods used in the study  Species Company Location Douglas Fir FPInnovations Vancouver, BC Lodgepole Pine FPInnovations Vancouver, BC Pacific Silver Fir FPInnovations Vancouver, BC Southern Pine FPInnovations Vancouver, BC Trembling Aspen FPInnovations Vancouver, BC Western Hemlock FPInnovations Vancouver, BC Western Larch FPInnovations Vancouver, BC White Spruce FPInnovations Vancouver, BC Yellow Cedar  FPInnovations Vancouver, BC Western Bigleaf Maple Malcolm Knapp Research Forest Maple Ridge, BC Ipe  PJ White Vancouver, BC Basswood  Windsor Plywood Vancouver, BC Red Oak Windsor Plywood Vancouver, BC White Oak  Windsor Plywood Vancouver, BC Balsa Keim lumber  Millersburg, OH Exotic lumber  Gaithersburg MD Red Alder  Malcolm Knapp Research Forest Maple Ridge, BC Targowoods  Bellingham, WA Western Red Cedar FPInnovations Vancouver, BC Malcolm Knapp Research Forest Maple Ridge, BC     222  Appendix 4: AutoCAD images of knives designed for the experiment described in chapter 5                                                                                                       Thin rib                                                                           Wide rib                           Short rib                                                                         Tall rib Rib 223                                                                             Ripple                                                                              Ribble                       Short ripple                                                                 Short ribble                        Tall ripple                                                                     Tall ribble 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0166219/manifest

Comment

Related Items