Printer Friendly

Experimental and theoretical behavior of exterior wood coatings subjected to artificial weathering.

Abstract Several solvent- and water-borne exterior wood coatings were artificially weathered to study their performance behavior. Penetration and layer thickness were measured on unweathercd samples and compared to theoretical estimates by means of measured basic parameters of coating and substrate. Color, surface roughness, gloss, and adhesion were monitored during aging, and the latter two were also calculated and compared to their experimental values. Theoretical values of gloss, and especially those of adhesion, were less successful on an absolute scale, but were in accordance with practical values on a relative scale. The influence of solid content, drying speed, and viscosity on penetration depth manifested itself clearly both in theory and in practice. In general, solvent-borne coatings performed well, but some water-borne coatings also showed good performance. By measuring the characteristics of a coating as described in this article, it is possible to rank the coatings and follow their weatherability. Finally, calculation of theoretical values proved a promising method for initial screening purpose.

Keywords Exterior wood coating, Artificial aging, Gloss, Penetration, Adhesion

Introduction

Finishes for wooden surfaces are subjected to severe in-service environmental conditions during natural weathering--both physical and biological. Because of this, lab-based test methods that can accurately and rapidly predict the durability of coatings have been sought since the beginning of the modern coatings era. (1) With this in view, researchers in wood-coaling science mostly make use of devices that combine a set of controllable parameters such as temperature, UV irradiance, and humidity (spraying or condensation)--aiming at an acceleration of natural weathering phenomena. A lot of research has focused on the relationship between artificial and natural weathering (1-3) to optimize accelerated weathering cycles or devices and to increase a correlation with natural weathering. (4), (5) It is a given fact that the stress levels in an artificial weathering device have to be chosen carefully so that the acceleration does not lead to an aging process that is different from the one under service conditions. (4) Obviously, different artificial tests differing in UV light dose, sequence order of stress, and exposure duration can lead to different degradation phenomena. (6) Nowadays, artificial aging is generally accepted as a tool for evaluating the performance of a coating and hence will be used in this article with a specific weathering cycle to imitate natural weathering closely. Yet, assessment of performance is also a much-discussed topic in coating research. Many parameters are available for evaluating the degradation of the coating. In this work, both aesthetic as well as physicochemical parameters were looked at for use in weatherability ranking. In addition to practical measurements, theoretical values were also calculated, using some measured basic parameters of substrate and coating. This approach can provide additional value by ranking coatings according to a predicted value as an initial screening exercise.

Aesthetic aspects of the coating can be described by processing and interpretation of color and gloss data, and their change during weathering. These parameters can be related to physical and chemical changes of the coating on its surface as well as changes in the interior of a coating and substrate. In some cases, gloss loss is even used for service life prediction. (7) Calculation of gloss starting from a surface with known properties is a complex problem. In addition to surface roughness, the topographical orientation of the microfacets and refractive index of semi transparent samples influence gloss. (8) The latter indicates the added complexity concerning semi transparent coating systems. For that reason, the theoretical value of gloss was calculated only for the opaque coating samples based on the work of Hunt of al (9). For surface roughness measurements, confocal scanning laser microscopy (CSLM) was used. Surface roughness was characterized using standard and fractal parameters.

All aforementioned weathering values mainly describe the physical properties of the coating, whereas adhesion aims at more physicochemical processes within the coating and at the wood-coating interface. Measurement of the adhesive strength of the coating to the wood with a torque technique is discussed in detail in Van den Bulcke et al. (10) The theoretical approaches as elaborated by de Meijer (11) and Liptakova and Kudeia (12) were taken into consideration, and a comparison was made with measured values. In addition, penetration of the coating into the wood was measured using CSLM. A prediction was made using the straightforward Washburn model and a simplified version of the extensive method of de Meijer et al. (13) with a rhcological-based approach.

Experimental procedure

Materials and sample preparation

Eight opaque (O) and nine semitransparent (T) solvent- (S) and water-borne (W) coatings were manufactured at Akzo Nobel Decorative Coatings laboratories in Vilvoorde, Belgium, of which the composition is summarized in Table 1. These were used to create three layered coating systems--except for two systems that were composed of only two coats as recommended by the manufacturer (Table 2}--by applying them by brush to straight-grained Scots pine sapwood {Pinus sylvestris L.) boards measuring 40 cm x 5 cm x 1 cm with the tangent of the annual rings of the specimen close to an angle of 45[degrees] with the horizontal surface. After application of each coat, at least 24 h of drying time was allowed. The amount of coating applied on the wood specimen, as well as the weight of the sample during drying, was recorded with a Sartorius balance with a readability of 0.1 mg. Subsequently, the samples were conditioned at 20 [degrees]C and 65% RH for 3 weeks. Thereafter, the samples were cut to 32 cm lengths and prepared for artificial weathering (Atlas UV2000). The remaining 8 cm parts of the specimens were used for penetration and layer thickness measurements. Penetration, layer thickness, color, gloss, surface roughness, and adhesion were determined on all unweathered specimens.
Table 1: Composition of the coatings

Type Coating Resin Resin (%)
Opaque
Solvent-borne OS-Alk1 Alkyd 33
 OS-Alk2 Alkyd 28
 OS-Alk3 Alkyd 38
 OS-Alk4 Alkyd 36
Water-borne OW-PU1 Polyurethane 21
 OW-PU-Ac Polyurethane-acrylate 23
 OW-PU2 Polyurethane 24
 OW-Ac Acrylate 25
Semitransparent
Solvent-horne TS-Alk1 Alkyd 61
 TS-Alk2 Alkyd 65
 TS-Alk3 Alkyd 33
 TS-Alk4 Alkyd 43
Water-borne TW-Ac-Alk Acrylate-alkyd 26
 TW-Ad Acrylate 39
 TW-AC2 Acrylate 22
 TW-PU Polyurethane 35
 TW-PU-Ac Polyurethane-acrylate 30

Type Solids (%) PVC (%) Water (%]

Opaque
Solvent-borne 75 25 0
 69 30 0
 68 15 0
 57 12 0
Water-borne 51 24 42
 56 27 44
 53 20 45
 54 20 45
Semitransparent
Solvent-horne 69 5 0
 76 4 0
 33 - 0
 46 2 0
Water-borne 33 1 58
 41 2 48
 31 5 65
 38 - 60
 32 1 63

Table 2: Coating systems

System code Coat 1 Coat 2 Coat 3

OS-2Alk OS-Alk1 OS-Alk
OS-3Alk-1 OS-Alk2 OS-Alk3 OS-Alk3
OS-3Alk_2 OS-Alk4 OS-Alk4 OS-Akl4
OW-3PU OW-PU1 OW-PU1 OW-PU1
OW-3PU-AC OW-PU-Ac OW-PU2 OW-PU2
OW-3AC OW-Ac OW-Ac OW-Ac
TS-3Alk-1 TS-Alk1 TS-Alk2 TS-Alk2
TS-3Alk_2 TS-Alk2 TS-Alk4 TS-Alk4
TS-3Alk__3 TS-Alk3 TS-Alk3 TS-Alk3
TS-3Alk__4 TS-Alk3 TS-Alk4 TS-Alk4
TW-3Ac-Alk 1 TW-Ac-Alk TW-Ac-Alk TW-Ac-Alk
TW-3Ac-Alk_Ac TW-Ac-Alk TW-AC1 TW-Ac1
TW-2AC TW-AC2 TW-AC2
TW-3Ac-Alk-PU TW-Ac-Alk TW-PU TW-PU
TW-3Ac-Alk-PU-Ac TW-Ac-Alk TW-PU-Ac TW-PU-Ac


Artificial aging

Samples were artificially weathered in an Atlas UV2000 device with fluorescent lightbulbs with a wavelength of 310-340 nm. This weathering cycle (W) held an alternation of two subcycles ([W.sub.1] and [W.sub.2]). Figure 1 illustrates their light and water-spraying regime (wet = water spraying; dry = no water spraying) within the complete weathering cycle. Both subcycles were repeated for 144 h. [W.sub1] was followed by 24 h of storage in the deepfreeze and [W.sub.2] by 24 h of storage in the refrigerator. The same procedure was used by Van den Bulcke et al. (10) In all, samples were weathered up to 2000 h. After each 500 h, color and gloss were recorded. Adhesion and surface roughness were measured only every 1000 h for the former test due to the destructive nature of it.

[FIGURE 1 OMITTED]

Measurements and theoretical calculations

Penetration and layer thickness of the coating systems were determined with CSLM as outlined in Van den Bulcke et al. (14) Coated wooden cubes of approximately 1 [cm.sub.3], stained with safranin dye to change the auto-fluorescence of the wood, were mounted in a Biorad Radiance2000 microscope. The microscope, equipped with a l0x Plan Apo objective lens (NA = 0.45), focuses a laser beam onto the microtomed transverse section of the cube and excites the safranin fluoro-chromes. By proper selection of the excitation and emission wavelength filters, the coating can be visualized, and an image sized 1024 x 1024 pixels is captured with a 2x electronic magnification resulting in 1.11 um/ pixel. Subsequently, this image is processed to obtain the penetration of the coating in the earlywood as shown in Fig. 2. This way, results are obtained by scanning a cross section of the coated cube. More information on the detailed methodology used for penetration and layer thickness measurement can be found in Van den Bulcke et al. (14) An attempt was made to simulate the penetration of stain using the Washburn equation, as outlined by de Meijer et al. (13)

[FIGURE 2 OMITTED]

[dh/dt]=[2[[gamma].sub.1]rcos[theta]-[rho]g[pi]h[r.sup.2]]/8[eta]h (1)

with

h, penetration depth (m)

[gamma]1, surface tension of the liquid (N m-1)

r, capillary radius (m)

[theta], contact angle ([degrees])

[rho], density of the liquid (kg/[m.sup.3])

g, gravitational acceleration m [s.sup.-2])

[eta], dynamic viscosity (kg [m.sup.-1] [s.sup.-1])

It is assumed that wood capillaries are perfect cylinders, extractive free, and that cell walls do not swell during penetration. Surface tension was measured with the Kruss DSA-10 Drop Shape Analyzer using the pendant drop method. The contact angle of the coatings on the substrate was also measured with this device, using the sessile drop method. The permanent change of contact angle at the wood-coating interface, as a result of a combination of wetting and penetration, implies problems for equilibrium determination, but is overcome by using the method of Liptakova and Kudela. (15) A Micromeritics pycnometer was employed to determine the density of the coatings. Pore sizes and pore size distribution were not measured. The ordinary differential equation was solved in MATLAB[R] for several capillary radii, including the small-, average-, and large-sized pores of pine sapwood. (16) Although, increasing viscosity and decreasing surface tension x cos (contact angle) in function of the increasing mass fraction during drying were not measured explicitly, as mentioned in de Meijer et al. (13) This was implemented by changing their value in accordance with the change in mass fraction. Viscosity followed an exponential increase, whereas the surface tension x cos (contact angle) followed a linear decrease. The change of coating density during drying is negligible. Because it is unknown at which mass fraction viscosity approximates infinity and surface tension x cos (contact angle) drops to zero, values were interpolated between the measured viscosity and an increasing end value, corresponding to a mass fraction of 0.99. The penetration depth when mass fraction approaches 1 was determined in function of changing viscosity, surface tension x cos (contact angle), and capillary radius. As such, three-dimensional volumes of penetration depth can be rendered. Change in mass fraction of the polymer inside a wood capillary was determined as follows:

[[empty set].sub.cm](t)=0.5r[rho]f/0.5r[rho]-[q.sup.w]t(3)

with

r, capillary radius (m)

, density of the liquid (kg [m.sup.-3])

f, mass fraction solid of the coating (-)

[q.sup.w], the evaporation rate to the wood (kg [m.sup.-2] [s.sup.-1])

t, time (s)

Evaporation rales of solvent from the wood substrate were determined using mass loss measurements on glass and wood:

[q.sup.w]=[m.sub.p]+[m.sub.o]-[[m.sub.p]/[[phi].sub.m][t.sub.1]]/[t'.sub.1]A-[q.sup.a](2)

with

[m.sub.p], mass of polymer (kg)

[m.sub.o], mass of water or solvent at t = 0 (kg)

[[phi].sub.m](t), mass fraction of polymer at time t (-)

[t'.sub.1], end of first drying stage on wood (s)

[q.sup.a] the evaporation rate to the air (kg [m.sup.-2] [s.sup.-1])

A, surface [(m.sup.2)]

Mass loss was recorded for 24 h. Equilibrium was not reached after 24 h, yet this time period was sufficient to determine the end of the first drying stage, necessary for the calculation of [q.sup.w]. Details regarding the theory can be found in de Meijer (11) and dc Meijer et al. (13) The model is formulated for capillary penetration at the axial surface, but is assumed to be indicative for penetration in wood that has mixed directions and claims no exact prediction of depth. The authors are aware of the oversimplification of the problem due to the nonlinear (logarithmic) behavior of viscosity in function of mass fraction, and by equal treatment of the different coatings concerning the relationship between surface tension x cos (contact angle) and mass fraction. Nevertheless, evaporation rates and mass fraction changes are explicative for penetration data.

Color was determined before weathering and after every 500 h of weathering by measuring reflectance in the range of 400-700 nm with the Konica Minolta CM-2600d portable spectrophotometer in diffuse light mode--using an 8 mm aperture and a D65 light source. With the accompanying software, data were transformed into color values using the L*a*b color space representing color perception attributes. (17) Color changes due to weathering were monitored by calculating [DELTA]E, (18) which is the difference between the sample after x hours weathering (subscript x) and before weathering (subscript 0) using the formula:

[DELTA]E=[square root of([([L.sub.x]-[L.sub.o]).sup.2]+[([a.sub.x]-[a.sub.o]).sup.2]+[([b.sub.x]-[b.sub.o]).sup.2])] (4)

which is the Euclidean distance two points in the L*a*b color space with

L, lightness of color

a, position between magenta and green

b, position between yellow and blue

Gloss values were recorded at angles of 20[degrees], 60[degree], and 85[degrees] before weathering and after every 500 h of weathering using a Rhopoint Novo-gloss meter. Gloss is considered to be the specular component of reflection. Obviously, the gloss meter also records part of the diffuse reflectance as a result of the restraints of the optics.

Surface roughness of a coating is a complex yet interesting property, considering the link with gloss. Initially, the topography of the coating must be mapped. This was carried out by taking a multislice stack of confocal images and processing them to build a digital surface. The first image was scanned at the top of the surface, and subsequently a series of images was taken while lowering the focal plane with a fixed step size (1 [mu]m) until complete extinction. The excitation wavelength of the laser beam was 488 nm, and the same emission wavelength was captured with the detector, thereby using the microscope in reflection mode. The z-stack, consisting of several images of 256 x 256 pixels and 1.421 urn/pixel recorded with an electronic zoom of 3, was loaded in the software package MATLAB(R) to reconstruct the surface. Roughness was characterized by means of a set of time-dependent parameters:

1.[R.sub.a.sup.t]--Mean surface roughness

2.[R.sub.[sigma].sup.t]--Standard deviation of surface roughness

3.[R.sub.FD.sup.t]--Fractal dimension according to the triangular prism method. In short, the triangular prism method triangulates the topological map of surface roughness for different window sizes and sums up the surface of the triangles. As window size increases, surface decreases, and a line can be fit through these surface data. The slope is considered to be a measure of fractal dimension. More detailed information can be found in Van den Bulcke et al. (19) and the references given there.

These parameters were monitored during weathering and used to compare the different coatings. Furthermore, the relationship with gloss was sought. Linking gloss with surface roughness is a complex problem. More specifically, surface roughness influences gloss by scattering the light due to irregularities. Many researchers described theoretical models to approximate practical gloss values. The Beckmann model of light reflection was applied in this work. The mathematics of this model can he found in Hunt et al. (9) Considering the problems that can arise when illuminating at an angle of 20[degrees] or 85[degrees], the calculations were limited to an incident angle of 60[degrees] and compared with measured values. Furthermore, because semitransparent systems complicate the calculations of light reflection, only opaque systems were studied here.

Adhesion measurements on all coating systems before weathering and after 1000 and 2000 h of weathering were an indication of the strength of the polymeric coating system. An improved version of the torque device, as described in Van den Bulcke et al., (10) was used for adhesive strength measurements. Briefly, a metal dolly was glued on the coated surface using two-component epoxy glue. When the glue hardened, the dolly was pulled off the coating. The exerted force was a measure of the adhesion under dry conditions--referred to as dry adhesion. The configuration of the torque device is schematically drawn in Fig. 3a, and the torque part is illustrated in Fig. 3b. Manualexertion of force was automated to increase the couple with a constant rate by installation of a motorreductor--i.e., a combination of an electromotor and a reduction case. By regulation of the voltage, the couple was constantly raised. This voltage regulator was, in its turn, connected to a National Instruments NI-DAQ mx Base 6008 data acquisition device. In this way, using a Lab View application, the motorreductor was computer controlled via the USB NI-DAQ device (Ao = analog out) and the voltage regulator. Measurements of torque were also recorded in the computer using the same NI-DAQ device, by reading in the electric signal of the Wheatstone bridge (Ai = analog in).

[FIGURE 3 OMITTED]

Further improvement of the technique included the sandblasting of the metal dollies, because untreated surfaces often exhibited glue failure on the metal due to the circular patterns caused by manufacturing. The fact that torsion was applied on the samples, the specific circular shape of artifacts had a baleful influence. At least ten adhesion measurements were done on every sample according to the protocol as elaborated in Van den Bulcke et al. (10) The methods of Liptakova and Kudela (12) and de Meijer (11) were used to predict the adhesion of unweathered coating systems. Both methods were based on surface energy calculations of solid coatings and wood surfaces. To under-stand the wood-solid coating interaction, it is necessary to map the surface energy values of wood [[gamma].sub.S1] and solid coating [[gamma].sub.S2] .Applying the method of Liptakova, free surface energy was calculated by using only one liquid standard--i.e., water--by which possible errors due to more standards with different polarities are avoided. Disperse and polar parts of the solids are used for calculating adhesion:

[W.sub.a]=2[square root of([[gamma].sub.S1.sup.d][[gamma].sub.S2.sup.d])]+2[square root of([[gamma].sub.S1.sup.p][[gamma].sub.S2.sup.p])] (5)

where superscripts d and p indicate the polar and disperse part of the surface energy. The method expounded by de Meijer (12) relies on the Van Oss theory, using three fluids (diiodomethanc. water, and formamide) to determine the acid base, and Lifshitz-van der Waals component of coating and wood. The work of adhesion is computed as:

[W.sub.a]=2([square root of([[gamma].sub.S1.sup.LW][[gamma].sub.S2.sup.LW])]+[square root of([[gamma].sub.S1.sup.+][[gamma].sub.S2.sup.-])]+[square root of([[gamma].sub.S1.sup.-][[gamma].sub.S2.sup.+]))](6)

At last, comparison of the theoretical and practical results throws light on the usefulness of theoretical considerations concerning adhesion in dry conditions. Furthermore, the practical adhesion results are also judged in the perspective of the penetration measurements. Finally, correlations between all theoretical and practical values are calculated. Coating systems are ranked according to their theoretical and practical aging behavior to clearly view their interrelation.

Results

Penetration and layer thickness on Scots pine early-wood zones were measured using CSLM. The results for the coating primers are shown in Fig. 4. Penetrationin latewood was restricted. This is in close agreement with earlier research.(20), (22) It is clear that semitransparent coatings with low density and low viscosity have a higher average penetration than the pigmented opaque primers. Penetration of the hybrid alkyd-acrylic coating TW-Ac-Alk is remarkable, probably due to its low viscosity. (23) Thickness is significantly related to solid percentage (0.764*).

[FIGURE 4 OMITTED]

Simulation of the penetration is only reviewed here. Input data are not given, merely a visual representation in Fig. 5 for two coatings. Simulation was done for viscosity measurements at a shear rate of 0.1 s. It is clear that a change of viscosity and surface tension x cos (contact angle) in function of mass fraction polymer is of utmost importance, because these two parameters strongly influence the outcome of the differential equation. Furthermore, the size of the capillaries has an obvious influence on penetration, especially when dealing with the initial parametric values. When calculating penetration depth for average pore size (30 [mu]m), assuming that viscosity at mass fraction 1 is le5 and surface tension x cos (contact angle) approaches zero, correlation between measured and calculated data is clear: >0.8. High penetration values for the solvent-borne TS-Alk3 and water-borne hybrid TW-Ac-Alk clearly stand out. When comparing [q.sub.w] values (x) with penetration data in [mu]m (y), a negative exponential relation is found:

[FIGURE 5 OMITTED]

y=24.07+31.38[e.sup.(-x/4.75e-6)] (8)

Values of [q.sub.w], ranging from 3.3e-6 to 7.9e-5, are in close agreement with those reported by Eckersley and Rudin (24) and de Meijer. (11) Logically, the faster the coating dries through evaporation of the solvent, the faster mass fraction and viscosity increase, making penetration difficult. Mass fraction changes as a function of time prove that slower mass fraction increase results in deeper penetration. Again, the deep penetrating primers dry slower, resulting in a gradual increase of mass fraction. The measured parameters give an indication of the penetrating power of a coating. Using correlation analysis, penetration is mainly influenced by the changing mass fraction and the time involved for reaching maximum solid percentage. It should be noted that these calculations do apply for penetration in pine sapwood in axial direction, but are assumed to be explicative for mixed directions, too. Yet, the influence of extractives and the swelling of the cell wall due to solvent ingress are not taken into account. Extractives have only a minor influence on penetration. (11)

Color evaluation is presented in Fig. 6, showing the change in [DELTA]E for the different systems. It is evident that semitransparent systems experience more color change than opaque systems. Furthermore, solvent-borne semitransparent coatings perform slightly better than water-borne semitransparent systems. For the opaque systems, all coatings perform similarly and have values that are difficult to discern with the eye--certainly when gloss is the dominant factor influencing the assessment of color change.

[FIGURE 6 OMITTED]

Gloss loss for an angle of 60[degrees] is illustrated in Fig. 7. The high gloss semitransparent polyurethane coating performs well, even approaching gloss values reached by the opaque solvent-borne alkyds and the opaque water-borne polyurethanes. After 2000 h, the value of the polyurethane system retains its high value, performing better than the opaque finishes. This explanation holds for an angle of 60[degrees]. Gloss and gloss loss at an angle of 20[degrees] are different. First of all, gloss is lower. Second, the changes for most systems are more pronounced. For an angle of 85[degrees], all but two systems have increased gloss values after weathering--probably owing to a smoothing of sharp peaks. Similar gloss behavior is reported by Malshe and Waghoo. (25) Nonetheless, 60[degrees] gloss is assumed to be the best average reference, whereas 20[degrees] and 85[degrees] measurements are meant for high and low gloss coatings, respectively.

[FIGURE 7 OMITTED]

Surface roughness changes during weathering are outlined in Table 3, listing different roughness parameters. In general, the changes are relatively small, except for OW-3PU-Ac, which shows a steep increase in mean roughness.
Table 3: Surface roughness parameters and measured gloss in function of
weathering

 t [R.sub.a.sup.t] [R.sub.[SIGMA].sup.t]
 ([mu]m) ([mu]m)
OS-2Alk 0 21.31 1.82
 1000 19.02 1.83
 2000 17.84 2.25
OS-3Alk_1 0 11.30 0.97
 1000 11.06 1.12
 2000 10.69 0.64
OS-3Alk_2 0 14.10 1.58
 1000 14.55 1.13
 2000 15.79 1.36
OW-3PU 0 11.86 1.10
 1000 11.35 1.79
 2000 11.45 1.32
OW-3PU- 0 14.95 1.81
Ac 1000 16.89 2.16
 2000 16.41 1.68
OW-3AC 0 15.51 1.00
 1000 16.85 2.15
 2000 16.29 1.53

[R.sub.FD.sup.t] Gloss 60[degrees] [(%).sup.a]

OS-2Alk 2.37 18.68
 2.33 16.72
 2.32 13.38
OS-3Alk_1 2.09 78.02
 2.14 78.94
 2.13 68.10
OS-3Alk_2 2.24 64.92
 2.26 57.60
 2.29 43.64
OW-3PU 2.14 22.06
 2.14 20.61
 2.13 14.55
OW-3PU- 2.24 65.26
Ac 2.28 61.69
 2.27 49.20
OW-3AC 2.26 32.40
 2.28 22.27
 2.28 18.66

(a) Measured with Novo gloss meter


Straightforward Pearson correlation shows a strong link (0.97) between the parameters [R.sub.a.sup.t] and .[R.sub.FD.sup.t]. The link with the standard deviation is 0.70. The gloss of unweathcred samples has a negative correlation with [R.sub.a.sup.t] and [R.sub.FD.sup.t]--only the OW-3PU coating system deviates substantially from the relationship between gloss and roughness. However, 60[degrees] gloss after 1000 and 2000 h of weathering is also linked with the standard deviation. The Beckmann theory of light reflection uses the autocorrelation function and standard deviation of roughness as parameters, without taking into account the effect of refractive index, shadowing, and masking as a result of topography. The calculated results for 0 h of weathering are given in the bar graph in Fig. 8. in general, the gloss is not predicted perfectly because important aspects of reflection theory are neglected. Nevertheless, high values such as the OS-3Alk_l coating arc definitely the result of a low roughness and its influence on gloss.

[FIGURE 8 OMITTED]

Adhesion of water-borne coatings to wood is shown in Fig. 9. A comparison is made between the measured, values and calculated ones as determined by the method used by Liptakova and Kudela (12) and de Meijer (13) for dry adhesion of unweathered coated samples. It should first be mentioned that the two Y-axes have different units, so comparing both values is merely done in a relative way. The values are comparable with the data reported by Liptakova and Kudela. (12) They also found high values for polyure-thanes, as is the case with these results.

[FIGURE 9 OMITTED]

"Without detailed analysis of the graph in Fig. 9, it is obvious that adhesion is a complex concept and quite difficult to predict. When looking closely, results of the de Meijer method are quite variable, while measured data have more in common with the Liptakova values. Scattering of the data gives a correlation of more then 0.6 (not significant), while this is considerably less for the de Meijer approach. The distinction between specimens based on dry adhesion is not straightforward. Wet adhesion, however, possibly does much more damage and pushes the coating to the limit, sometimes resulting in complete delamination of coating from substrate. Therefore, wet adhesion will be studied in the future in accordance with the methods of de Meijer and Militz (26) and Hora (27)

When dry adhesion and penetration are related, a significant linear relationship between penetration and weathered adhesion values of the solvent-borne transparent coatings is observable (Fig. 10). The trend of deeper penetration resulting in stronger adhesion is not distinct for the other weathered coating systems. The positive relationship is also not obvious for the unweathered samples. Post-curing and removal of solvent remainders after weathering could have been beneficial to the degree of success of adhesion measurements. The glue failure on many solvent-borne systems may be the explanation for the scattering of unweathered adhesion values. A positive correlation was expected, as already demonstrated by Van den Bulcke et al. (10) and de Meijer, (11) for earlywood penetration. Apparently, mechanical anchoring is an important decisive factor in the adhesion of certain coating systems.

[FIGURE 10 OMITTED]

All these measures of degradation--experimental or theoretical--can be used to rank the coating systems from high to low resistance toward weathering. However, multicriteria ranking can only be done if certain weights are attached to the different parameters, expressing the relative importance toward service life. It will be assumed that all measured and predicted values have equal importance--i.e., w = 1. As a consequence, the coatings will be ranked several times and their mean position will be calculated, for the theoretical as well as practical values. In fact, deep penetration, a thicker coat, high adhesion, low [DELTA]E, and high gloss (60[degrees]) are considered as important parameters for service life, separating high from lower quality coatings. This ranking results in Fig. 11a for the opaque coatings and Fig. 11b for the transparent ones. Parameters measured during weathering are used to rank the coatings before (0 h) and after weathering (2000 h). Square symbols represent a mean ranking according to penetration, gloss, adhesion, and layer thickness before weathering and gloss, adhesion, and color change after 2000 h of weathering. The filled circles represent a ranking based on gloss and adhesion only. This ranking procedure leads to the general conclusion that water-and solvent-borne products have representatives in both the best- and worst-performing coating groups. The transparent solvent-borne systems score a little better in general, not withstanding the promising hybrid-polyurethane system TW-3Ac-Alk-PU. Average rank numbers of theoretical calculations of opaque systems based on penetration, gloss, and adhesion before weathering are useful, but naturally depend on the possibility of making correct estimates of the parameters. In particular, adhesion is a difficult parameter and hence the influence on ranking is obvious. It should be stressed that this ranking is based on a few parameters. When more parameters are available, the chance is less that one of them will outweigh the others.

[FIGURE 11 OMITTED]

Conclusions

Artificial aging is part of the standard set of tools at the disposal of the coating manufacturer to evaluate the resistance of organic coatings. All weathering conditions are designed to mimic aspects of the outdoor environment. A coating's weatherability is assessed in terms of loss of adhesion, loss of gloss, and color changes. (28) Several tests are available to screen the coatings and it was the purpose of this article to complete practical values with theoretical estimates. First, penetration and layer thickness were measured using confocal laser microscopy--a valuable technique for wood-coating interaction studies. (29) The solventborne alkyds especially penetrate deep and form a thick protectant coat on the substrate, which was in agreement with previous findings. (11) Water-borne alkyds penetrated less. The influence of viscosity, drying speed, and solid content was evident in the theoretical calculation of penetration depth. Furthermore, color and gloss are common measures defining the aesthetics of a coating. Naturally, opaque coatings show only slight color differences after weathering, whereas the semitransparent systems perform less. Gloss resulted in the prevailing water-borne polyurethane resin containing coatings, retaining high reflection values side-by-side with the solvent-borne alkyds. The prediction of reflection percentages for opaque systems on the basis of surface roughness measurements comes close to measured values for unweathered systems. A torque apparatus was improved for adhesion testing. Calculation of theoretical values gave relative indications of the strength of the coatings, but missed good correlations with the measured values. This corresponds to the conclusions of de Meijer and Militz. (26) Ranking of the coatings proves the promising evolution of some water-borne coatings and their resistance to weathering. In general, the water-borne finishes even shift to better positions at the expense of solvent-borne systems. The hybrid primers, combined with a polyurethane topcoat, give a good semitransparent system with proper weatherability.

Withal, a combination of measurement and calculation is a convenient means to test coatings and limit the time intensiveness of experimental tests. Ultimately, in combination with theoretical models, (30), (31) true service life prediction is near.

Acknowledgments The authors would like to thank Professor P. Van Oostveldt (Department of Molecular Biotechnology, Ghent University) for the use of the confocal scanning laser microscope. The essential help of Johan De Clercq (Department of Mechanical Construction and Production in Labo Soete, Ghent University) with the development of the torque device is also highly appreciated. Further, the authors owe their gratitude to E. Mol of the Laboratories of Akzo Nobel Decorative Coatings Vilvoorde in Belgium for the formulation of the coatings. Finally, the authors would like to thank Professor W. Steurbaut (Department of Crop Protection, Laboratory of Crop Protection Chemistry, Ghent University) for the use of the Drop Shape Analysis equipment.

References

(1.) Carlozzo, BJ, Andrews, J, Anwari, F, DiLorenzo, M, Glover, R, Grossman, S, Harding, C, McCarthy, J, Mysza, B, Raymond, R, Skerry, B, Slifko, PM, Stipkovich, W, Weaver, JC, Wilson, G, "Correlation of Accelerated ExposureTesting and Exterior Exposure Sites. 2. One-Year Results." J. Coat. Technol., 68 (858) 47-61 (1996)

(2.) Okkonen, EA, River, BH, "Outdoor Aging of Wood-Based Panels and Correlation with Laboratory Aging: Part 2." For. Prod. J.. 46 (3) 68-74 (1996)

(3.) River, BH, "Outdoor Aging of Wood-Based Panels and Correlation with Laboratory Aging." For. Prod. J., 44 (11-12) 55-65 (1994)

(4.) Brunner, S, Richner, P, Muller, U, Guseva, O, "Accelerated Weathering Device for Service Life Prediction for Organic Coatings." Polym. Test, 24 (1) 25-31 (2005)

(5.) Podgorski, L, Arnold, M, Hora, G, "A Reliable Artificial Weathering Test for Wood Coatings." Presented at Third International Woodcoatings Congress--Woodcoatings, Foundations for the Future, The Hague, NL, October 2002

(6.) Perrin, FX, Irigoyen, M, Aragon, E, Vernet, JL, "Evaluation of Accelerated Weathering Tests for Three Paint Systems: A Comparative Study of Their Aging Behaviour." Polym. Degrad. Stab., 72 (1) 115-124 (2001)

(7.) Guseva, O, Brunner, S, Richner, P, "Service Life Prediction for Aircraft Coatings." Polym. Degrad. Stab., 82 (1) 1-13 (2003)

(8.) Hebert, M, Emmel, P, Hersch, RD, "A Prediction Model for Reflection on Varnished Metallic Plates." Presented at First European Conference on Color in Graphics, Imaging and Vision, Piotiers F, April 2002

(9.) Hunt, FY, Galler, MA, Martin, JW, "Microstructure of Weathered Paint and Its Relation to Gloss Loss: Computer Simulation and Modeling." J. Coat. Technol., 70 (880) 45-54 (1998)

(10.) Van den Bulcke, J, Rijckaert, V, Van Acker, J, Stevens, M, "Adhesion and Weathering Performance of Water-Borne Coatings Applied to Different Temperate and Tropical Wood Species." J. Coat. Technol. Res., 3 (3) 185-191 (2006)

(11.) de Meijer M, Interactions Between Wood and Coatings with Low Organic Solvent Content, PhD Thesis, University of Wageningen, Wageningen, the Netherlands, 1999

(12.) Liptakova, E, Kudela, J, "Study of the System Wood-Coating Material. Part 2. Wood-Solid Coating Material." Holzforschung, 56 (5) 547-557 (2002)

(13.) de Meijer, M, van de Velde, B, Militz, H, "Rheological Approach to the Capillary Penetration of Coating into Wood." J. Coat. Technol., 73 (914) 39-51 (2001)

(14.) Van den Bulcke, J, Rijckaert, V, Van Acker, J, Stevens, M, "Quantitative Measurement of the Penetration of Water-Borne Coatings in Wood with Confocal Lasermicroscopy and Image Analysis." Holz Als Roh-und Werkst, 61 (4) 304-310 (2003)

(15.) Liptakova, E, Kudela, J, "Analaysis of the Wood Wetting Process." Holzforschung, 48 (2) 139-144 (1994)

(16.) Wagenfuhr, R, Scheiber, C, Holzatlas. Leipzig, Germany (1989)

(17.) Sanchez-Maranon, M, Soriano, M, Melgosa, M, Delgado, G, Delgado, R, "Quantifying the Effects of Aggregation, Particle Size and Components on the Colour of Mediterranean Soils." Eur. J. Soil Sci., 55 (3) 551-565 (2004)

(18.) Macleod, IT, Scully, AD, Ghiggino, KP, Ritchie, PJA, Paravagna, OM, Leary, B, "Photodegradation at the Wood-Clearcoat Interface." Wood Sci. Technol., 29 (3) 183-189 (1995)

(19.) Van den Bulcke, J, Van Acker, J, Stevens, M, "Image Processing as a Tool for Assessment and Analysis of Blue Stain Discolouration of Coated Wood." Int. Biodeterior. Biodegrad., 56 (3) 178-187 (2005)

(20.) de Meijer, M, Thurich, K, Militz, H, "Comparative Study on Penetration Characteristics of Modern Wood Coatings." Wood Sci. Technol., 32 (5) 347-365 (1998)

(21.) Nussbaum, RM, Sutcliffe, EJ, Hellgren, AC, "Microautora-diographic Studies of the Penetration of Alkyd, Alkyd Emulsion and Linseed Oil Coatings into Wood." J. Coat. Technol., 70 (878) 49-57 (1998)

(22.) Van Loon, J, "The Interaction Between Paint and Substrate." JOCCA, 49 844-867 (1966)

(23.) Scheikl, M, Dunky, M, "Measurement of Dynamic and Static Contact Angles on Wood for the Determination of Its Surface Tension and the Penetration of Liquids into the Wood Surface." Holzforschung, 52 (1) 89-94 (1998)

(24.) Eckersley, ST, Rudin, A, "Drying Behavior of Acrylic Latexes." Prog. Org. Coat., 23 (4) 387-402 (1994)

(25.) Malshe, VC, Waghoo, G, "Weathering Study of Epoxy Paints." Prog. Org. Coat., 51(4) 267-272 (2004)

(26.) de Meijer, M, Militz, H, "Wet Adhesion of Low-VOC Coatings on Wood--A Quantitative Analysis." Prog. Org. Coat., 38 (3-4) 223-240 (2000)

(27.) Hora, G, "Measuring Wet Adhesion of Wood Coatings by Pull-off Test." Presented at Fourth International Woodcoatings Congress--Woodcoatings, Developments for a Sustainable Future, The Hague, NL, October 2004

(28.) Dickie, RA, "Chemical Origins of Paint Performance." J. Coat. Technol., 66 (834) 28-37 (1994)

(29.) Singh, AP, Dawson, BSW, "Confocal Microscope--A Valuable Tool for Examining Wood-Coating Interface." J. Coat. Technol. Res., 1 (3) 235-237 (2004)

(30.) Hinderliter, B, Croll, S, "Monte Carlo Approach to Estimating the Photodegradation of Polymer Coatings." J. Coat. Technol. Res., 2 (6) 483-491 (2005)

(31.) Van den Bulcke, J, Van Acker, J, Saveyn, H, Stevens, M, "Modelling Film Formation and Degradation of Semi-Transparent Exterior Wood Coatings." Prog. Org. Coat., 58 (1) 1-12 (2007)
COPYRIGHT 2008 Federation of Societies for Coatings Technology
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2008 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Bulcke, Jan Van den; Acker, Joris Van; Stevens, Marc
Publication:JCT Research
Date:Jun 1, 2008
Words:6370
Previous Article:Enhancing exterior performance of clear coatings through photostabilization of wood. Part 2: coating and weathering performance.
Next Article:Assessment of UV-permeability in nano-ZnO filled coatings via high throughput experimentation.
Topics:


Related Articles
The wood coatings market: wood protective coatings and preservatives are poised for slow, steady growth in the years ahead.
Solid color deck stains: trends, substrates, and formulation study.
Celanese offers emulsion polymers.
Wood coatings demand to grow despite some challenges.
Ability of transparent coatings to retard leaching of borates in a Weather-Ometer[R] test.
Analysis of weathering effects in automotive coatings using thermoluminescence.
Hydroxyphenyl-s-triazines: advanced multipurpose UV-absorbers for coatings.
Considerations for characterizing moisture effects in coatings weathering studies.
Enhancing exterior performance of clear coatings through photostabilization of wood. Part 2: coating and weathering performance.

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |