Abrasion Resistance of Thermally and Chemically Modified Timber

Wood modifi cation is an appropriate way of improving the natural durability and dimensional stability of wood without the use of biocides. Different thermal and chemical wood modifi cation processes are available for this purpose, very differently affecting the structural integrity of wood. In this study, thermally modifi ed, melamine resin treated, acetylated, furfurylated, and mDMDHEU treated wood underwent abrasion tests according to two different methods representing different loads in practice. The Taber Abraser method caused crosswise cutting into the wood surface, while the Shaker method challenged mainly the specimen edges with dynamic loads. Abrasion resistance of wood was affected by all types of cell wall modifi cation, but the effects were strongly dependent on the type of modifi cation and the applied load type. For characterising the suitability of wooden materials with respect to wear resistance under outdoor conditions, it is recommended to apply a set of methods rather than a single test procedure to fully refl ect the loads occurring in practice.

Wooden fl oorings are exposed to various physical, mechanical, chemical, and fi nally biological loads especially if they are exposed outdoors (Brischke, 2010).Due to its hygroscopic character, moisture dynamics fi nd a response in swelling and shrinking of fl oor boards.Hereby, the anisotropy of wood can lead to drastic deformations and warping of whole decks.Cell wall modifi cation of wood has the potential to reduce such dimensional changes (Hill, 2006).Furthermore, aesthetic impairments of wood surfaces can result from UV degradation combined with leaching of lignin fragments and colonization of dark coloured molds leading to more or less homogeneous graying of the surface.In addition, some wood materials, e.g.sapwood portions, are susceptible to discoloration by blue stain fungi (Huckfeldt, 2009).However, all these biophysical degradation processes have in common that they affect the optical appearance only; the functionality of the fl ooring will not be compromised.
In contrast, mechanical loads occur in terms of wear, abrasion, and erosion and can signifi cantly impact on the functional performance of wooden fl ooring.The wear effect of walking persons is often intensifi ed by abrasive particles, e.g.sand, dust, winter grit, and other more or less sharp-edged particles.The resistance against abrasion strongly depends on the material -anatomic differences become evident, e.g. in the form of washboard effect as a result of differences between earlywood and latewood (Sell and Feist, 1986).A standardized method for determining the abrasion resistance of solid wood is still lacking.Hence, the Taber Abraser method, as referred to by EN 438-2 (2016), ISO 9352 (2012), and ASTM D 1044ASTM D (2013)), has been occasionally used for this purpose, although it is originally intended for testing the abrasion resistance of high-pressure laminated papers (Militz et al., 2011).As previously shown by Welzbacher et al., (2009), the loads occurring during Taber Abraser tests provoked by rotating sandpaper under defi ned grinding pressure do not necessarily refl ect the exposure conditions of an outdoor exposed fl ooring, for example a terrace decking.As shown for timber structures in the marine environment by Brischke et al. (2005) and Williams et al. (2010), it is rather necessary to use a set of test methods to fully refl ect the insitu conditions to which wood is exposed outdoors.Brischke et al. (2014) applied the so-called Shaker method for testing the abrasion resistance of wet and dry wood specimens made from various wood species and found that the abrasion resistance of wood is generally increasing with higher densities.However, they also reported about exceptions such as Douglas fi r (Pseudotsuga menziesii Franco) that reveals high abrasion resistance at rather moderate density.Apparently, anatomical features such as the regularly alternating early and latewood sections of Douglas fi r positively affect its abrasion resistance, which is in line with a theory of shockabsorbance effects formulated by Williams et al. (2010).Similar tests using a concrete mixer and shingles were performed by Williams et al. (2010) to characterize the performance of different wood species used for groynes.
The Shaker method has been previously applied on thermo-mechanically densifi ed and thermally modifi ed wood, which showed signifi cantly increased abrasion compared to untreated wood (Wehsener et al., 2017).Earlier work by Baird (2007), Militz et al. (2011) and Mahnert (2013) indicated that different kinds of chemical wood cell wall modifi cation have an effect on the abrasion resistance of wood as well.The aim of this study was, therefore, to determine the abrasion resistance of differently chemically and thermally modifi ed timber in comparative tests using the Taber Abraser and the Shaker test method.

Materijali
Specimens were prepared from differently modifi ed Scots pine sapwood (Pinus silvestris L.) and European beech (Fagus sylvatica L.) as well as untreated reference specimens from both wood species as shown in Table 1.Thermal modifi cation was conducted at 230 °C by Timura Holzmanufaktur (Rottleberode, Germany) using the Vacu³-process.
Heat treated beech wood was additionally modifi ed with methylolated melamine formaldehyde resin (MMF) Madurit MW 840/75WA (Ineos Melamines GmbH, Germany).Boards were impregnated in an impregnation plant at -0.60 mbar for 0.5 h followed by a pressure phase of 12 bar for 2 h.Specimens were dried and cured in a laboratory oven at maximum temperature of 120 °C for 24 h.
Commercially sized boards of beech were acetylated in the plant of Accsys Technologies in Arnhem, the Netherlands, using an industrial process.Prior to and after the acetylation, the dimensions and weight of the boards were measured.Since the process was not adapted, the acetylated beech showed a high amount of surface checks, but in general the degree of modifi cation was uniform to high levels.
The modifi cation process of furfuylation was carried out at KEBONY, Skien, Norway.Boards of beech The resistance against abrasion was determined according to the Taber Abraser method (EN 438-2, 2005).The following modifi cations of the Taber Abraser test were made in order to allow testing of solid wood: Specimens of (100 (ax.) x 100 x 7) mm³ were prepared and conditioned at 20 °C/65 % RH.The tree rings of all specimens were oriented 45° to their cutting edges.After weighing and measuring the thickness at four points, the specimens (n = 5) were clamped into the Taber Abraser and abraded with sanding paper S-42 with approx.60 min -1 for 1,000 revolutions.Afterwards, the decrease in thickness by abrasion was determined.The percentage loss in thickness Δt was determined as a measure of abrasion according to the following Eq. 3 for each specimen and an average determined: (3) Δt -abrasion (%) t 0 -thickness, before abrasion (mm) t 1 -thickness, after abrasion (mm)

REZULTATI I RASPRAVA
The abrasion resistance of wood was signifi cantly affected by different modifi cation processes, but the respective effect was strongly dependent on the method applied for testing the resistance to abrasion.Abrasion according to the Taber Abraser method was expressed as percentage reduction in thickness of the specimens (Δt) and decreased with increasing oven-dry density of the wood material (Figure 1).Thermal modifi cation and modifi cation with mDMDHEU led to a decrease in abrasion resistance according to the Taber Abraser method compared to the untreated references, were impregnated with furfuryl alcohol and subsequently dried and cured in an industrial plant.
Scots pine sapwood boards were impregnated in an autoclave at -0.98 bar for 1 h and at 12 bar for 2 h with a solution of modifi ed 1.3-dimethylol-4.5-dihydroxyethyleneurea(mDMDHEU) in the following concentration: 36 % of mDMDHEU stock solution and 2 % magnesium nitrate hexahydrate relative to the mass of mDMDHEU as catalyst.After impregnation, the samples were cured in steam atmosphere at 120 °C.The mean weight percent gain (WPG) was 19.4 %.

Abrasion resistance tests -Shaker method 2.3. Ispitivanje otpornosti na habanje -metoda Shaker
The resistance against abrasion was determined according to the Shaker method described by Brischke et al. (2005).Five oven-dry specimens of (35 (ax.) x 8.5 x 8.5) mm³ were laid in polyethylene fl asks (V = 500 ml) together with 400 g stainless steel balls of 6 mm diameter and moved in an overhead shaker at 28 revolutions min -1 for 72 h.In total, fi ve by fi ve specimens of each material were tested.Distances d between opposite corners of the oven-dried specimens were measured to the nearest 0.01 mm before and after abrasion.The percentage loss in dimension Δd was determined as a measure of abrasion according to the following Eq. 2 for each block and an average determined: (2) Furfurylation and an impregnation with melamine resin did not affect the abrasion of wood signifi cantly, but a melamine treatment of heat treated beech led to an improvement of its abrasion resistance.Abrasion of specimens challenged by shaking together with abrasive steel balls according to the Shaker method was clearly not correlated with the oven-dry density of the material.While the lighter wood of Scots pine showed clearly higher abrasion in the Taber Abraser compared to beech wood, both wood species showed only insignifi cantly different abrasion in the Shaker test (Figure 2).Only, acetylation led to a slight increase in abrasion resistance of wood, all other modifi cation pro-cesses reduced its abrasion resistance.Scots pine treated with mDMDHEU, thermally modifi ed beech with and without subsequent melamine treatment suffered from the highest abrasion in Shaker tests.The effect of the different modifi cations on the abrasion resistance of wood coincided fairly well with previous fi ndings by Brischke et al. (2012) who reported that the structural integrity of wood was negatively affected by cell wall modifi cation in the following order: Furfurylation < melamine resin treatment < mDMDHEU treatment.
Consequently, it was expected that abrasion determined with the help of both test methods was not correlated as shown in Figure 3. Acetylated beech showed less abrasion, and mDMDHEU treated Scots pine the most, if considering both methods.The test material was differently challenged by the two test methods.In the Taber Abraser, wood specimens were subject to cutting

Shaker -Abrasion
/ Shaker -otpornost na habanje , % by aluminium oxide abrasive on the sanding paper, which is applied under pressure and a rotational movement.This particular type of wear becomes evident by crosswise cutting traces on the specimens' surfaces as shown in Figure 4.In contrast, the steel balls used for the Shaker test caused a multiple, but slight dynamic impact on the specimens, to which their longitudinal edges were particularly subjected and suffered from break offs.The rounding of the specimens' edges can be seen from their cross sections as illustrated in Figure 5.
In accordance with the percentage abrasion, two groups of materials can be distinguished by visual appearance: Untreated Scots pine and beech as well as melamine treated and acetylated beech showed only slight rounding of edges, while furfurylated, thermally modifi ed beech with and without melamine treatment as well as mDMDHEU-treated Scots pine had significantly rounded edges.
Wooden fl ooring in use under outdoor conditions can suffer from both scratching and dynamic loads transversally to the surface leading to break offs and rounding of edges.To characterize the suitability of wood-based materials, especially treated and modifi ed wood, consequently requires the application of different test methods refl ecting the full spectrum of loads occurring in practice.

ZAKLJUČAK
Wood cell wall modifi cation had generally a signifi cant and mostly negative effect on the abrasion resist- ance of wood, but was strongly dependent on the type of modifi cation and the test method applied.The Taber Abraser method caused crosswise cutting into the wood surface.In contrast, the Shaker method challenged mainly the specimen edges with dynamic loads.Both types of wear occur frequently in practice when wooden fl oorings are exposed outdoors.In addition, further parameters, such as static and dynamic hardness, as well as susceptibility to deformations and their possible impact on wooden fl ooring, have the potential to interact with abrasion resistance.Future research will, therefore, focus on comparative analysis of wear parameters obtained in laboratory tests and under real exposure conditions.