Coatings for Use on Wooden Bicycle Frames – Applicability , Test Methods and Artificial Weathering Results

Seven different coating systems were tested regarding their suitability for application on bicycle frames made of wood. Since bicycles are ridden throughout the year, the coating system has to withstand different outdoor climate conditions and, especially in winter, the salt brines from the roads. For this reason, an artifi cial weathering test with an additional freezing step was performed, followed by a specially developed salt brine spray test in ambient outdoor winter climate. Another focus was on the applicability of the coatings. For the tests, a range of coatings, originally developed for wooden front doors, fl oors and for boat building, was selected. These coatings were expected to meet the main requirements for UV radiation permanence, weather resistance and saltwater fi rmness. The main constituents of the coatings were water-based polymer dispersions, two-component polyurethane systems or synthetic and natural oils. The layer thicknesses of the coatings, as well as colour and gloss data of the specimens, were recorded and rated. The adhesion capability was examined by carrying out a cross-cut-test. Surface damages were observed visually. It turned out that colour and gloss change served as good indicators for the decay of the coating. Three different coatings passed the accelerated weathering test. Salt spray had no signifi cant effect on any of the coatings.

Ključne riječi: drveni okvir bicikla, premaz za drvo, izdržljivost, simulirano izlaganje vremenskim utjecajima, slani sprej Since the beginning of mobility, wheeled vehicles as well as ships and planes were manufactured from wood (Childe, 1951;Piggott, 1969).For vessels, the protection of wood against UV exposure and salt water has been an issue (Laidlaw, 1952).Even today car manufacturers, as the Morgan Motor Company, use ash wood for their car frames.However, the hygroscopicity and limited durability of wood causes diffi culties if used outdoors.At Morgan, therefore, steel or aluminium coverings are used to protect the wooden frame against decay ("The Morgan Motor Company," n.d.).This procedure of wood protection, however, does not make sense on wooden bicycles as weight is an issue and the frame should be visible as a design element (Figure 1).Another way to protect wood is practiced by boat building companies: wooden boats are coated in order to protect them against infl uences of UV radiation and salt water.
As coatings for wooden bicycle frames also have to provide protection against moisture accumulation in order to prevent biological deterioration, against UVradiation and against deterioration caused by salt brineinduced corrosion, the need has become obvious to apply the wood protection method used in boat building to the bicycle frame building.Furthermore, coatings for wooden bicycle frames have to be fl exible but also to show decent bonding properties in order to allow movements of the frame.Baur et al. (2006) studied the weathering performance of different wood species by determining the timber-specifi c colour degradation.They suggested coating the specimen with clear coat and specifi c additives in order to strengthen the protection against sunlight degradation.Dawson et al. (2008) proposed to enhance the exterior performance of clear coatings through delignifying the surface cells of Pinus radiata.
Finally, the coating should be easy to apply and maintain.Consequently, coatings were chosen for the bicycle wood frame that met the described requirements.Beside the resistance to weathering of the cured coating, the applicability and maintenance of the coating was also important for its choice.
To prove the performance of coatings on wood, the testing method EN 927-6 (DIN EN 927-6, 2006), which describes the exposure of wood coating systems to artifi cial weathering using fl uorescent UV lights and moisture, was chosen.This method is described as suitable to test the performance of coatings on wood and gives evidence of durability in natural weathering (Grüll et al., 2014).
The main goal of the present study was to fi nd the best UV, weather and salt brine resistant and easiest to maintain coating for wooden frame bicycles.

Materijal
European ash wood (Fraxinus excelsior L.) lamellae were bound with polyurethane resin and pressed in moulds for the bicycle frame (Figure 2).Specimen with the dimension of 150 x 75 x 6 mm³ were prepared from the lamellae and then coated.As there are no coatings available, exclusively developed for wooden bicycle frames, alternative coatings, developed for similar purposes, were investigated.
Besides coatings for front doors, which are directly exposed to weathering, coatings for wooden fl ooring and interior coatings, developed to resist abrasion and stain, were selected, as well as coatings for wooden boats, made to withstand weathering and seawater (Table 1).

Nanošenje premaza
The coatings consisting of several layers were applied according to the manufacturer instructions.As the repair-friendliness of the coating was in the focus of the manufacturer of the wooden bike, the applicability and maintenance of the coating had to be taken into account as selection criteria.The WD and PU-coating systems could be applied using a low volume low pressure (LVLP) spray gun with 1.4 mm nozzle -at 150 kPa air pressure.The WD System consisted of fi ve components: a primer, an intermediate coat and a fi nish coat, which also needed a hardener and a solvent.The PU system consisted of a fi lling primer, a hardener and a clear lacquer as fi nish.All other coatings could be easily applied using a brush, also according to the manufacturer specifi cations.The oils consisted only of one component.In total, 6 specimen of each coating were produced.Since no matting agents were used, the coatings appear glossy on WD, PU and O1 specimens and semi-glossy, as defi ned in EN 927-1 (EN 927-1, 2013), on HW and O3 after application.
The coatings were applied in a paint booth at ambient room temperature between 18 and 22 °C and a relative humidity between 40 and 60 % RH.All specimens had to dry for 24 h in the same conditions between successive coatings.The untreated wood was prepared with 150 grit abrasive paper and sanded with 240 grit abrasive paper between the coatings.The total number of layers for each coating is listed in Table 2.
For comparison, reference samples without coating were produced.The wood lamellae were conditioned in standardized conditions at 20 °C and 65 % relative humidity for 28 days.Also, coated samples which were not exposed to artifi cial weathering were produced for the salt spray test and as control samples for the cross cutting test.
For characterization of the several coatings studied herein, the layer thickness of each coating was determined using a microscope at 10x and 20x magnifi cation.Three days after completing the preparation of the samples, different coatings were tested following the EN 927-6 (DIN EN 927-6, 2006), which describes the exposure of wood coating systems to artifi cial weathering using fl uorescent UV lights and moisture.As weathering tester, a Q-LAB QUV accelerated weathering testing machine was used.The machine simulates the effects of sunlight, dew and rain with UV-exposure, condensing humidity and water spray in alternating cycles.
As cold winter conditions also had to be simulated, a freezing step was inserted into the standard cycle, and the duration of the UV-exposure and spray subcycle was shortened from 48 to 32 (Table 3).For freezing, the samples were alternately removed from UV and from the spraying step at the end of each cycle.The samples were frozen in a laboratory freezer at a controlled temperature of -25 °C.In order to intensify the climatic stress, 14 cycles were performed instead of 12.

Measurement of colour and gloss changes 2.4. Mjerenje promjene boje i sjaja
Before each weathering cycle, the coating surface of each specimen was inspected by human eye for colour change, yellowing, fading, cracks and fl aking following DIN EN ISO 4628-4 and 5 (DIN EN 4628-4, 2004, DIN EN ISO4628-5, 2004).For an evaluation of colour differences with the naked eye, scales as shown in Table 4 are often used.Additionally, at six individual spots on each specimen, the colour, lightness and gloss data were measured and recorded at the beginning of each weathering cycle with a BYK-Gardner spectroguide 45/0 gloss spectrophotometer.The colour and lightness were determined in the CIE system (Commission Internationale de l'eclairage), according to DIN EN ISO 11664-4 (DIN EN ISO 11664-4, 2012), where the L*a*b* colour space is described as a treedimensional space, where L* stands for lightness, a* is a green to red coordinate and b* is a blue to yellow coordinate.
The ΔE * ab (Equation 1) indicates the colour and lightness differences of the tested wood coatings and colour change during test procedure.
The mean values of colour and lightness differences were measured at the beginning of each testing cycle and mapped in a diagram (Figures 5 and 6).The intensity of the refl ected light was measured as Gloss units (GU) together with colour and lightness at the beginning of each cycle.

Cross cutting test 2.5. Određivanje adhezije premaza metodom zarezivanja mrežice
DIN EN 927-6 requires a cross-cut test for the evaluation of adhesion capability of varnish on wooden surface.Two weeks after having passed the accelerated weathering cycle, a right angle lattice pattern was manually cut into the coating with a blade according to EN ISO 2409 (ISO 2409(ISO , 2013).
An average distance of 2 mm from each cutting line is required for soft substrates and layer thicknesses up to 60 μm.After cutting, the adhesion resistance of coatings to separation from the wood surface is tested using an adhesion tape applied over the grid.Five minutes after application, the tape is removed at an angle of 60°.The test results were classifi ed according to EN ISO 2409 from 0 (no visible changes) to 5 (the worst adhesion), As only the WD, O1 and O2 series samples withstood the weathering cycle with an unscathed surface, these were rated and the mean value of each sample series was calculated.For comparison purposes, unexposed samples were also subjected to the cross-cut test.For a more detailed evaluation, the samples were classifi ed using a microscope with 10x and 20x magnifi cation.In standards for the automotive industry other test methods are also described, in which the parts are not exposed to a continuous salt spray, but are directly sprayed with salt brine.An example for this is the GMW 14872 Cyclic Corrosion Laboratory Test procedure (GMW14872, 2006), which provides an accelerated corrosion test for parts and components under laboratory conditions.The test method provides a combination of cyclic load by moisture, drying and salt spray to accelerate corrosion.The DIN EN ISO 15710 (2006) is intended primarily to test the protective effect of coatings on aluminium and aluminium alloys against corrosion and is vastly used in the aerospace industry.In this test method, a coated sample plate, whose coating is scratched, is alternately immersed in a dilute salt solution and removed from it.
Various testing methods indicate that there are different approaches to examine the infl uence of corrosion on materials and surface coatings.For the present application, the resistance of a wooden bicycle frame against pre-wetted salt, used by winter services in Germany, was of particular interest.Based on test methods for metallic components, a test method was developed to examine direct contact of coated wood with salt and brine.The specimens with a size of 75 mm x 300 mm were mounted on a frame in a 20° inclination (Figure 3) in southern direction and exposed to a natural outdoor climate for eight weeks.Temperature and relative humidity were continuously recorded (Figure 4).
German winter services mostly use up to 40 g/m² of the so-called wet salts to defrost snow or ice covered roads.Wet salts consist of dry salts moistened with saline.Accordingly, saturated brine, consisting of 22 % sodium chloride and 78 % water was prepared.Through a period of 8 weeks 60 g/m² of the salt solution was sprayed on the samples twice a day.It can be noted that, regardless of the method of application, layer thicknesses ranging between 35 and 50 μm were obtained (Table 5).With an average layer thickness of 18.0 μm, the O4 product had the lowest layer thickness -due to the fl uidity and method of application (wood had to be «saturated» with the oil at least one day) -Furthermore, the thickest layer of 51.8 μm was measured for the O1 product.
Both methods of application have advantages and disadvantages over the other method.The coating applied using a spraying gun produces a thinner and more uniform fi lm.If repairs are necessary, preference must be given to the oils, because of the simpler application.Repairs of a multi-layered coating system can only be carried out by a qualifi ed workshop.

Colour and gloss change 3.2. Promjena boje i sjaja
Form the beginning of weathering, a colour change and darkening of all specimens could be noticed.As shown in the average ΔE* ab change diagram (Figure 5), the colour of all samples changed after the fi rst cycle.All coated samples showed the tendency to become darker (Figure 6) at as long as the surfaces were not damaged.A washout of the coating generally leads to fl aking and, therefore, to a loss of protection against bleaching by UV irradiation.After the end of the fourth cycle, colour changes on PU and O4 samples could be measured.Only one week later, the discolouration could be seen by the naked eye.
The beginning of bleaching of the coating of HW samples could be measured after the fi fth week by ΔE* ab data acquisition.On O3 samples, a visible fading could fi rst be detected four weeks later.As the coating was largely destroyed having passed the eleventh cycle, the PU (quantity fi ve, crack width parameter 3 according to DIN EN ISO 4628-4), O4 and HW (severe fading) coatings were removed from the artifi cial weathering cycle.
For all other samples, up to the 13 th cycle, no fl aking was detected.However, after the last cycle, incipient fading was observed also with the naked eye on the O3-samples.
The coated specimens became glossier after the fi rst cycle.During the following test cycles, the gloss decreased in all samples.At the end of the tenth cycle, a signifi cantly stronger decreasing of gloss of the O1, O2 and PU samples was detected (Figure 7).

Određivanje adhezije premaza metodom zarezivanja mrežice
For comparison, specimens of all coated but not weathered control samples were cross cut.It can be stated that all control samples had intact surfaces, so all of these samples could be categorized as "0" according to DIN EN ISO 2409.A cross cut of an intact HW surface with classifi cation "0" is shown in Figure 8. From the samples exposed to weathering, only the samples that withstood the 14 cycles weathering test with an obviously intact surface were tested by cross-cut.The cross-cut samples of O1 and O2 series could be classifi ed as 0-1 (Figure 9) after weathering.In case of the WD series, a cross-cut area between 5 % and 15 % showed fl aking (Figure 10) (Table 6).The self-developed salt-spray-test follows the conditions that prevail on wintry German roads, and no generally valid conclusions can be drawn on its basis.For this purpose, a standardized test would have to be developed, that would also take into consideration colour and gloss measurements.This salt-spray test showed that, at the end of twelve weeks in ambient moderate winter climate, all test specimens survived the test unscathed.In periods of rising temperatures, the brine formed salt crystals on the surface of the specimen.These could be easily washed off.The surfaces had no visible cracks or fl aking; neither the crystals nor the brine itself had an effect on the varnish.Corrosion appeared only on the stainless steel screw fasteners and washers (Figure 11).Table 5 shows the results of the single tests in a brief summary.

ZAKLJUČAK
Although recommended, the two-component PU coating system is not suitable for the intended use, as  only after fi nishing the fi fth cycle, fl aking of the coating could be noticed.The hard wax oil, which is intended for the use on wooden fl oors, bleached at the beginning of the 8 th week.Also, two boat building oils failed.Flaking appeared on the O4 specimen after only six cycles, due to leaching.The O3 oil faded at the end of the weathering cycle.The multi-component waterbased coating for exterior doors (WD) and two oilbased coatings from manufacturers of yacht coatings (O1 and O2) can be recommended for the year round use as a good protection for wooden bicycle frames.
Considering the results of the colour change and gloss measurements, it can be concluded that the tested coatings, exposed to accelerated weathering, have shown signifi cant differences in resistance to UV radiation and extreme climate conditions.It turned out that the measurement of colour changes in artifi cial weathering tests is -together with the cross cut test -an effective way to determine suitable coating products for wooden frames.The salt brine spray test showed no further degradation on the coating surface.
Taking into account the results from the colour and gloss measurements, the use of a multi-component water based coating, as well as some of the oil based paints for boat building, can be recommended for the intended application.
In order to create a stress scenario on wooden bikes that is closer to a real daily use, in future tests, the coated wooden lamellae could be exposed to mechanical forces before weathering.Additionally, the resistance of coatings against crushed stones from the road surface and their ability to withstand abrasion in daily use could be part of future testing.

Table 4
(Buchelt and Pfriem, 2010.)ferences (Buchelt and Pfriem, 2010) Tablica 4. Ljestvica ocjena promjene boje(Buchelt and Pfriem, 2010.) signifi cant difference in colour, which is rarely tolerated / znatna razlika boje koja se rijetko tolerira above 5.0 the difference is judged to be other colour / razlika boje je takva da se procjenjuje kao druga boja To test the suitability of the bike frame coating for use in winter, two months after the accelerated weathering cycles, a salt-spray-test was performed in addition to the weathering test.Most corrosion tests for (coated) metal are based on a salt spray test, in which metal sample coupons are exposed to the salt spray.