Effects of surface structure and substrate color on color differences in textile coatings containing effect pigments

Materials and methods

Substrates

Three grades of plain weaves (Almedahl-Kinna AB, Sweden), made from poly(ethylene terephthalate) (PET) fibers, scoured and heat-set by the supplier, were used as substrates. The weaves had different structure characteristics depending on yarn type, yarn number (excluded as requested by the supplier), thread count and color (black or white), given in Figure 1. The plain weaves were uncoated and free from fluorescent whiteners. The uncoated samples were selected to fully characterize the influence of surface structure on color differences. Also, uncoated samples are a natural selection of material when taking faster processing times, energy savings and cost-minimizations for manufacturers into account. The black textile substrates were dyed black by the supplier. The Filament weave (F) was made of a continuous multifilament yarn; the other two weaves, Spun Open and Spun Dense (SO and SD, respectively), were both made of spun staple fiber yarn with the same yarn number but with different thread counts in the warp and weft directions, as shown in Figure 1. F had the largest thread count, followed by SD. SO had the lowest thread count in both warp and weft directions, and consequently the most open structure with the largest pore size. The SO and SD weaves were chosen as substrates due to their different surface roughness. A black and white laminated paper test chart (TC) (A4SBoA from Zehntner GmbH, Schweiz/Switzerland) was included as a smooth reference surface to color differences caused by EPs on the different surface roughness and substrate colors, that is, black and white. In total, seven different substrates were used for each of the two pigmented coating pastes.

OPEN IN VIEWER

Figure 1.

Specifications of effect-pigment-coated samples and TiO₂-pigment-coated samples.

Preparation of coating pastes

Two coating pastes were prepared, one containing EP and one containing TiO₂ pigment. The TiO₂ coating paste was used for comparison in order to show the color differences generated between the spherical TiO₂ pigments and the flat-shaped substrate-based EPs. The multicolor EP (Colorstream T10-05 Pacific Lagoon, kindly supplied by Merck AB, Sweden) had the following structure: a mica platelet coated with a layer of titanium dioxide, followed by a layer of silicon dioxide and finally a layer of tin oxide. The platelet diameter distribution was 10–60 µm with a mean particle size of 24 µm according to the supplier, and the thickness was below 1 µm, ¹⁰ as confirmed by scanning electron microscope (SEM) images (see Figure 2). The EP is categorized as an interference pigment with a multicolor effect, depending on observation angle, within the turquoise–blue–violet range.

OPEN IN VIEWER

Figure 2.

Scanning electron images, cross-sections of effect pigment (EP)-coated samples (a) and (e) on test chart (TC)/black/EP; (b) and (f) on Filament (F)/black/EP; (c) and (g) on Spun Dense (SD)/black/EP; and (d) and (h) on Spun Open (SO)/black/EP at magnifications 150 × and 1000×, respectively.

EP (5 wt%) was dispersed in 8 wt% water before being mixed with 86.7 wt% of a polyurethane-based binder (Performax 16297G from Diazo AB, Sweden) using a plain stirrer for five minutes at 800 rpm at room temperature of 23℃. The binder contained a thermoplastic aliphatic polyester–polyurethane aqueous dispersion and a cellulose-type thickener. When more water was added to the formulation, the viscosity decreased. This was compensated for by adding 0.3 wt% of a polyurethane-based, non-ionic liquid thickener (Borchi Gel L75N from OMG Borchers GmbH/IMCD, Sweden AB). All the ingredients were mixed using a plain stirrer for five minutes at 800 rpm.

The coating paste containing TiO₂ was prepared from a commercial TiO₂ dispersion (Nanoflam WP 12, kindly supplied by Diazo AB, Sweden). The dispersion contained 60 wt% rutile TiO₂ pigment, with a mean particle size of less than 10 µm, 1 wt% polyacrylate as dispersing agent and 39 wt% water. The TiO₂ dispersion was first diluted with water to a concentration of 40 wt% rutile TiO₂ pigment (after stirring with a saw-tooth disc for 10 minutes at 1300 rpm) before being mixed with the binder. The final TiO₂ coating formulation was then prepared by mixing 12.5 parts of the 40 wt% TiO₂ dispersion with 87.3 parts Performax 16297G and 0.2 parts of Borchi Gel L75N and stirring it for five minutes at 800 rpm.

For each coating paste a batch of 400 g was prepared. By using the theoretical solids content for each ingredient given in Table 1, both coating pastes were calculated to consist of approximately 35.5 wt% solids. This is in agreement with the solids content measured by weighing each coating paste before and after drying for 24 hours at 22℃ and 33% relative humidity (RH): both pigment coating pastes were determined to contain 35 wt% dry weights. The concentration of pigments in the dried coatings was calculated to be 14 wt%.

OPEN IN VIEWER

Table 1.

Coating paste composition in weight%

Ingredients (solids in wt%)	Composition wet in wt%
	Effect pigment coating paste	TiO2-pigment coating paste
Performax (35)	86.7	87.3
Effect pigment (100)	5	–
Water	8	–
Borchi Gel L75N (48)	0.3	0.2
TiO2 dispersion (40)	–	12.5

Coating procedure

The two coating pastes were applied to uncoated black and white substrates in the warp direction of the textiles with a lab-scale knife-over-table coating method. ⁶ Substrate samples in the size of 20 cm × 50 cm were taped to an even surface and the coating formulation was manually applied using an applicator (ZUA 2000, Zehntner), with a gap setting of 200 µm. The TC was coated using the same method. The mean coating speed was 0.02 m/s for all samples, corresponding to a shear rate of 100 s⁻¹. The coated samples were first dried on stenter frames for 9 minutes at 80℃ and then cured for 5 minutes at 150℃ in a laboratory dryer/stenter (Ernst Benz AG, Zürich).

Two replicas of each coating paste sample (2 × 2 × 7 samples) were prepared, resulting in a total of 28 samples.

Characterization

Angle-dependent color measurement

A multi-angle spectrophotometer (MA68II from X-Rite, USA) was used to measure the colors at five different detection angles: 15°, 25°, 45°, 75° and 110° from the specular angle (see Figure 3). ¹⁹ The instrument has a repeatability with a CIELAB total color difference of ΔE* = 0.02. ²⁰ Measurements were made on duplicate samples at five different positions on each sample and mean values and standard deviations for each sample were calculated. The differences between the duplicates were small. The data were evaluated with respect to the (CIE 1964) 10° standard observer and D65 (average daylight) as the standard illuminant. ¹⁵ Illumination was at a fixed angle of 45° relative to the normal. The illumination source and the detector were consistently placed along the weft direction of the samples. The standard deviation was in general low and the highest standard deviations of the measurements were, for EP-coated samples: L* = 2.63, a* = 1.38, b* = 0.66 and for TiO₂-coated samples: L* = 2.08, a* = 0.41, b* = 0.76. Considerations were taken regarding the standard deviations of all measurements and no conclusions have been drawn based on results where the standard deviations are close to overlap. It should be pointed out that the measurement data obtained in this study are specific for the standard illumination source D65 representing average daylight. Depending on the end use of the fabric, specific illumination sources should be used. If, for example, the textiles are being used indoors the CIE standard illuminant A, representing typical domestic tungsten-filament lighting, should be considered.

OPEN IN VIEWER

Figure 3.

Graphic displaying the angles used by the spectrophotometer for illumination and measurement. All angles are given relative to the specular angle.

A three-dimensional graphic of the CIELAB color space can be seen in Figure 4. In this three-dimensional space, the color of samples can be characterized using the coordinates L*, a* and b*. The coordinate system is based on three opponent axes: L* representing white–black, a* and b* representing redness–greenness and yellowness–blueness, respectively. In this study the L*, a* and b* coordinates were chosen to present the measured values.

OPEN IN VIEWER

Figure 4.

Graphic of the three-dimensional CIELAB color space.

To evaluate the color differences between a reference and a textile sample the ΔL*, Δa* and Δb* values were calculated using the following formulas: ¹⁵

Δ L * = Δ L reference * - Δ L sample *

(1)

Δ a * = Δ a reference * - Δ a sample *

(2)

Δ b * = Δ b reference * - Δ b sample *

(3)

As mentioned above, the paper TC is considered as the reference. However, the ΔL*, Δa* and Δb* values between the highest and the lowest detection angles were also calculated for each of the uncoated black and white textile samples and with different surface structures.

The total color difference, ΔE*, gives the distance between the reference and the textile sample in the three-dimensional CIELAB color space and was calculated according to the following:

Δ E * = Δ L * 2 + Δ a * 2 + Δ b * 2

(4)

Viscosity

Both coating pastes showed similar shear-thinning behavior with small viscosity differences of less than 0.08 Pa.s, as seen in Figure 5. Their similar viscosity profiles and viscosity values can be explained as the effect of adding 0.3 wt% of the thickener Borchi Gel L75N into the EP-coating paste and 0.2 wt% of the thickener into the TiO₂-coating paste. The amount of thickener was adjusted in order to prevent the coating paste from penetrating and bleeding through the woven substrates during the coating procedure, and since the SO weave, due to its larger pore size between the yarns, was the most liable to penetration of the coating paste, the viscosity profiles of the pastes were set to match the SO-woven sample.

OPEN IN VIEWER

Figure 5.

Viscosity versus shear rate of coating pastes based on the effect pigment and TiO₂ pigment.

Air permeability and coating add-on

The air permeability measurements (see Table 2) showed that the volume of air which passed through uncoated sample fabrics in one second at a pressure of 200 Pa increased in the order of: TC, F, SD and SO. The fabrics made of spun yarns with lower thread counts compared to the filament fabric showed values that were one and a half and three times higher for SD and SO, respectively. This suggests that F had the lowest fabric porosity followed by SD and SO, and that spun yarns and a lower thread count in warp and weft directions increased the fabric porosity. This result is in agreement with visual observation of the fabrics.

OPEN IN VIEWER

Table 2.

Uncoated and coated fabric characteristics

Test chart		Filament		Spun Open		Spun Dense
Uncoated characteristics
Air Permeability [l/m2s]	0		443 ± 12,1		1535 ± 14.2		653 ± 8,3
Thread count: warp/weft	–		30.5/24		24/17		27.5/21
Weave density [g/m2]	–		110		135		147
Coated characteristics	TC/EP	TC/TiO2	F/EP	F/TiO2	SO/EP	SO/TiO2	SD/EP	SO/TiO2
Solid coating add-on [g/m2]	41	41	56	60	67	80	74	72

The dense TC had the lowest add-on, followed by the filament weave, which had a high thread count in both warp and weft directions (see Table 2). The weaves made from spun yarns had the highest add-ons, which is attributed to their increased fabric porosity, as stated above. A weave with a larger pore size is more likely subjected to penetration, leading to a higher coating add-on. Therefore, increasing coating add-ons with increasing porosity in the order TC, filament and spun weaves were expected. However, with the EP, the SO samples showed a lower add-on than the SD ones. This was unexpected, since the SO had the more porous structure. On the other hand, the TiO₂-coated samples showed, as expected, a greater add-on of the SO than the SD. The coating add-on was generally higher for the TiO₂-coated samples compared to the EP-coated samples. The difference is suggested to be due to the easier penetration of TiO₂ particles into the pores, whereas the large platelets of the EPs probably lead to a blocking of the pores, resulting in lower add-on.

Influence of textile surface structure on color differences

Surface structure of uncoated and coated substrates

The surface structures of uncoated substrates and of EP-coated substrates are shown in the SEM images in Figure 6(a)–(d) and Figure 6(e)–(h) respectively. The uncoated surface structures had different degrees of surface roughness. A smooth surface of the uncoated TC substrate is apparent in Figure 6(a), whereas the uncoated textile substrates (see Figure 6(b)–(d)) appears to have rougher surfaces due to their structure characteristics in the form of yarn type, yarn number and thread count. The F weave clearly shows a smoother surface due to its continuous multifilament yarn and higher thread count compared to the spun weaves (SD and SO) composed of spun yarn and with lower thread counts. The spun yarns clearly contribute to irregularities on the surfaces of SD and SO in the form of protruding fiber loops, as seen in Figure 6(c) and (d). It is difficult to determine which of the spun woven samples had the roughest surface. When comparing the uncoated SD and SO samples it is, however, obvious that SD had a tighter fabric construction due to higher thread counts in the warp and weft directions.

OPEN IN VIEWER

Figure 6.

Scanning electron microscope images of uncoated substrates (a)–(d) and effect-pigment-coated samples (e)–(h) on black substrates. From top to bottom: test chart, Filament, Spun Dense and Spun Open weaves at magnifications 50×.

The coated substrates also showed surface roughness to different extents (see Figure 6(e)–(h)). In all cases, the substrate structure could clearly be seen after the coating. A clear increase in surface roughness was apparent between the smooth TC/black/EP and F/black/EP (see Figure 6(e) and (f)) due to the structure characteristics of yarn type, thread count and construction, which are visible on the coated surface. For spun woven samples (SO and SD), the spun fibers contribute to a more irregular surface due to fiber loops that are more pronounced on the coated surfaces compared to the uncoated surfaces. The irregular surface of the SO-coated sample is generated both by fiber loops (caused by the yarn structure) and by its larger pores between the warp and weft (related to the thread count). The plain weave construction is not as accentuated on the SD/black/EP coated surface as it is on the SO/black/EP one (see Figure 6(g) and (h)). On the other hand, the protruding fiber loops appear to be more accentuated on the SD/black/EP surface compared to the coated SO/black/EP surface, which also contribute to an irregularly coated surface.

SEM images of the cross-sections of all the EP-coated samples are shown in Figure 2(a)–(d) at magnifications of 150 × and in Figure 2(e)–(h) at 1000×. These SEM images show that the coating penetration varies to different extents depending on the structure characteristics of the substrates. As expected, the smooth TC/black/EP sample (see Figure 2(a) and (e)) had no coating penetration. In the SEM images of F/black/EP (Figure 2(b) and (f)), a coating penetration between the interstices of the outer filaments of the yarn can be seen. The SO substrate (Figure 2(d) and (h)), having larger pores between the warp and weft (due to lower thread count) than the filament substrate, gave a deeper coating penetration and as an effect of this a higher coating add-on (see Table 2). The protruding fiber loops of SD/black/EP and SO/black/EP created higher peaks and valleys (see Figure 2(c) and (d)) and thereby rougher coated surfaces compared to the TC/black/EP and F/black/EP. However, the SO/black/EP appears to have a smoother surface than SD/black/EP, which is more tightly woven with a larger thread count in warp and weft directions (see Figure 2(c) and (d)). When correlating SEM images in Figures 2 and 6, a clear increase in surface roughness can be stated in the order of TC < F < SD ≈ SO.

The SEM images in Figure 2(e)–(h) indicate the distribution of effect pigment platelets embedded in the coating. The cross-section for the smooth TC/black/EP in Figure 2(e) shows the pigment platelets aligned more or less parallel to the surface. In Figure 2(f), platelets with a diameter of around 10 µm can be distinguished on top of the single filaments, having a diameter of 25 µm. When comparing the images of F/black/EP, SD/black/EP and SO/black/EP, an increasing degree of disorder can be seen where pigments seem to have aligned themselves on top of the fiber surfaces (see Figure 2(f)–(h)). This suggests that (a) the platelets follow the curved shape of the weaves so that most platelets were oriented out of the parallel plane and (b) the fiber loops of the spun yarns contributed to the disorientation of the pigment platelets relative to the substrate surface even further, so that the EP platelets were less aligned in SD- and SO/black/EP than in F/black/EP. The suggestions are based on studies of five images of each coated sample.

Color differences of uncoated textiles and coated substrates

The L*-, a*- and b*-color coordinates from the multi-angle measurements of the uncoated textile substrates are presented in Table 3. A small influence of the uncoated surface structures on the color differences between the detection angles and the textile samples was noticed. Uncoated white substrates showed wider variations of a*- and b*-values between the different detection angles compared to the uncoated black substrates. Yet the color differences between the highest and the lowest detection angles for each textile substrate were less than Δa* = 0.11 and Δb* = −0.26 for black fabrics and Δa* = 0.44 and Δb* = 1.16 for white substrates. On the other hand, the lightness differences were higher for the black textile substrates with ΔL* = 9.88 compared to ΔL* = 3.9 for white substrates between the different detection angles. A trend of decreasing L*-values at 15–45° angles for the black textile substrates in the order F > SD > SO was seen. Decreasing L*-values with increasing detection angles within the measurements of the black and white textile samples and at detection angles 15–45° was also noticed.

OPEN IN VIEWER

Table 3.

The L*-, a*- and b*-color coordinates of the uncoated black and white textile substrates measured at different detection angles with a multi-angle spectrophotometer

Substrates	L*					a*					b*
	15	25	45	75	110	15	25	45	75	110	15	25	45	75	110
F/black	26.83	25.86	21.77	18.58	25.71	0.33	0.35	0.36	0.35	0.25	−1.84	−1.81	−1.83	−2.07	−2.03
SO/black	20.15	19.47	18.59	19.08	22.12	−0.16	−0.12	−0.12	−0.14	−0.14	−2.19	−2.14	−2.31	−2.79	−2.80
SD/black	22.68	21.91	20.15	16.95	18.34	0.57	0.64	0.57	0.58	0.60	−1.77	−1.76	−1.85	−2.34	−2.32
F/white	97.68	97.45	95.99	93.78	95.51	0.27	0.21	0.09	−0.17	0.15	−0.02	−0.09	−0.43	−0.84	−0.39
SO/white	94.75	94.34	94.08	95.96	96.95	−0.21	−0.34	−0.50	−0.63	−0.44	1.54	1.41	0.98	0.80	0.38
SD/white	94.63	94.33	94.02	95.90	95.40	−0.53	−0.65	−0.81	−0.95	−0.75	0.56	0.43	0.04	−0.13	−0.57

The results of the multi-angle measurements are shown in a CIE a*-b* diagram in Figure 7 on EP-coated black samples. Samples on black substrates were chosen as reference samples to show the influence of the textile surface structure on the color differences. The influence of substrate color is discussed later.

OPEN IN VIEWER

Figure 7.

Multi-angle spectrophotometer measurements of effect pigment (EP)-coated black samples, CIE a* and b* coordinates at various detection angles. TC: test chart; F: Filament; SD: Spun Dense; SO: Spun Open.

For each of the coated black substrates, the a*-b*-coordinates give a horseshoe-shaped curve within the blue–green quadrant, which clearly shows that the color varies between the different detection angles. The a*-values for the textile sample curves decreased (more towards green) with increasing angles from 15°, 25° to 45° and thereafter the a*-value increased. A similar trend was seen for the TC sample, although the a*-value increased at the angle of 45°. For each sample curve the b*-value generally increased (more towards yellow) with increasing detection angle from 15°, 25°, 45° to 75°, but at angles greater than 75°, the b*-values decreased slightly (towards blue) for the textile samples.

The greatest color-changing effect, as observed both visually and by using the spectrophotometer, was as expected seen when the EP coating was applied to the smooth TC/black surface, where the EP platelets were able to align themselves parallel to the substrate surface. This is in agreement with the results from the cross-section SEM images in Figure 2(e) and the smooth surface (see Figure 6(e)). Therefore, the a*- and b*-coordinates of the reference TC/black/EP clearly had a wider variation within the blue–green quadrant than those of the textile samples.

However, the measurements presented in Figure 7 confirm that the EP coating also gave a color-changing effect that differed in extent depending on the substrate's surface structures and on the detection angle. Correlating the SEM images in Figures 2 and 6 and the measured a*- and b*-coordinates in Figure 7 clearly shows that the color-changing effect decreased with increasing surface roughness. It is also obvious that the small color differences measured on the uncoated textile substrates (see Table 3) had little influence on the greater color changes shown in Figure 7.

Figure 8 shows that the lightness (L*) decreased with increasing detection angle for all the coatings on black substrates. The smooth TC/black/EP sample had a higher L*-value than the rougher textile samples at all detection angles. This is suggested to be related to the fact that the L*--value at a 15° angle is highly influenced by the amount of light reflected at the specular angle referred to as gloss, which is larger on smooth surfaces, whereas on rougher surfaces and at other angles the light is scattered from the surface in a diffuse manner.^15,20

OPEN IN VIEWER

Figure 8.

L*-values of effect pigment (EP)-coated black samples as a function of detection angle. TC: test chart; F: Filament; SD: Spun Dense; SO: Spun Open.

Below a detection angle of 45°, the F/black/EP textile sample showed the smallest lightness difference (ΔL*-value), with respect to the reference TC/black/EP, followed by SD/black/EP and the SO/black/EP. A similar trend in decreasing lightness with increasing surface roughness at detection angles below 45° was also seen on uncoated black textile samples in Table 3.

The total color differences (ΔE*) at the various detection angles between the reference TC/black/EP and the black textile samples with EP coating are presented in Figure 9. The results showed that the total color differences increased in the order of F/black/EP < SD/black/EP <SO/black/EP at angles of 15°, 25° and 45°. This confirms that the uncoated textile structures seen in the SEM images in Figure 6(a)–(d) created by yarn type and yarn number, as well as thread count in warp and weft directions, influenced the coated surface roughness and most definitely also the orientation of the pigment platelets in the coating distinguished in Figure 2(e)–(h). This in turn highly influenced the total color differences, which increased with increased surface roughness in that order.

OPEN IN VIEWER

Figure 9.

Total color differences ΔE* at various detection angles between test chart (TC)/black/effect pigment (EP) and black textile samples with EP coating. F: Filament; SD: Spun Dense; SO: Spun Open.

At 15° angle the ΔE*-value of SO/black/EP was almost twice as high compared to F/black/EP and at greater angles it differed even more. The color-shifting effect may therefore be reduced by half or even more when using a substrate constructed of spun yarn with low thread count compared to a substrate made of filament yarns with higher thread counts. At 15°and 25° angles, the ΔE*-values of SD/black/EP were lower than those of SO/black/EP; at higher angles it was even significantly lower, closer to the ΔE*-values of F/black/EP. This suggests that the color-changing effect may be more striking using a substrate constructed of spun yarn with higher thread counts (as the SD substrate) compared to one with lower thread counts.

Although the interference pigment used in this study should be able to provide color-changing effects within the turquoise–blue–violet region, the complete color change could not be measured with the multi-angle spectrophotometer used.^12,16–18 It was, however, visually detected by the human eye under diffuse illumination at various observation angles. It is said that an experienced colorist can distinguish color differences (in diffuse illumination) of approximately Δa* = Δb* = 0.15 and lightness differences of approximately ΔL* = 0.2–0.3.^19,20 This suggests that a total color difference of around ΔE* = 0.29 can be differentiated. It clearly puts in perspective the influence of the substrate on the color-shifting effect, which is almost two orders of magnitude higher than the color difference that is distinguishable to the human eye.

TiO₂ pigments in comparison with effect pigment

Table 4 shows the L*-, a*- and b*-color coordinates from the multi-angle measurements of the TiO₂-coated black substrates at various detection angles. Figure 10(a) presents the a*- and b*-coordinates in a diagram. The color differences seen in Table 4 and Figure 10(a) are, as expected, significantly smaller than those for the EP-coated samples in Figure 7, which all showed wide variations within the blue–green quadrant. The TiO₂-pigment coating was used to show how a diffusively reflecting pigment generates color differences compared to the interference pigments used in this study. The color differences between the highest and lowest detection angles within the curves were less than Δa* = 1 and Δb* = 3.5 on black substrates.

OPEN IN VIEWER

Figure 10.

CIE a*-and b*-coordinates (a) and L*-values (b) of TiO₂-coated black samples as a function of detection angle. TC: test chart; F: Filament; SD: Spun Dense; SO: Spun Open.

OPEN IN VIEWER

Table 4.

The L*-, a*- and b*-color coordinates of TiO₂-coated black samples measured at different detection angles with a multi-angle spectrophotometer

Samples	L*					a*					b*
	15	25	45	75	110	15	25	45	75	110	15	25	45	75	110
TC/black/ TiO2	90.33	86.82	86.04	86.34	84.87	−2.02	−2.22	−2.27	−2.30	−2.16	−2.05	−1.80	−1.95	−2.43	−2.42
F/black/TiO2	72.24	67.86	58.45	57.70	59.16	1.83	1.79	2.41	2.49	2.61	−2.62	−2.59	−3.31	−4.07	−4.33
SO/black/TiO2	43.49	43.61	44.74	48.84	52.67	1.81	1.82	1.92	1.92	2.07	−2.26	−2.44	−3.12	−4.17	−4.46
SD/black/TiO2	51.08	50.55	49.07	49.77	52.69	0.31	0.20	0.13	0.10	0.14	−5.64	−6.05	−7.00	−8.30	−8.86

TC: test chart; F: Filament; SD: Spun Dense; SO: Spun Open.

Figure 10(b) shows the lightness (L*) at different detection angles for TiO₂-coated black samples. With increasing angle from 15° to 45° the lightness decreased and thereafter with increasing detection angle up to 110°, the lightness increased slightly, possibly because TiO₂ pigments, due to their irregular shape, scatter the light in a more diffuse manner. The textile surface structure seemed to influence the lightness of the TiO₂-coated samples in a similar manner to the EP-coated samples (see Figure 8). The TC/black/TiO₂ showed the highest L*-values, followed by F/black/TiO₂, SD/black/TiO₂ and SO/black/TiO₂, in decreasing order regardless of detection angle.

Influence of substrate color on color differences

The results of the multi-angle measurements are presented in a CIE a* and b* diagram for EP-coated white samples in Figure 11(a). All the materials are characterized by s-shaped curves passing through the blue–green, red–yellow and yellow–green quadrants. A comparison with black substrates (see Figure 7) shows that the substrate color greatly influenced the color differences.

OPEN IN VIEWER

Figure 11.

CIE a*- and b*-coordinates (a) and L*-values (b) as a function of detection angle for effect pigment coatings on a white substrate. TC: test chart; F: Filament; SD: Spun Dense; SO: Spun Open.

In general, both the a*- and b*-values were higher for the white substrates, moving towards red and yellow, respectively. It is generally agreed that the color-changing effects of interference pigments are best seen if they are coated on a dark substrate, which absorbs the wavelengths complementary to the reflected color, emphasizing the interference color at the observation angles. On the white substrates, the complementary color is, however, diffusely reflected in all directions except at the glancing angle, where the interference color is reflected. ¹⁷ However, reflections from the white substrate still affect the interference color, which becomes less intense, thereby also harder for the human eye to distinguish.

A distinct difference was observed in the L*-values between the substrates coated on black and those on white substrates (see Figure 8 and Figure 11(b)). The L*-values were generally much higher on the white substrates than on the black substrates. Although the lightness decreased with increasing detection angle on both white and black substrates, the change in lightness was less pronounced on the white substrates. At the 15°, 25° and 45° detection angles, the lightness decreased slightly for the white textile samples and more or less leveled out at 75° and 110°, probably due to the diffuse scattering of light from the white substrates.

In Figure 12 the total color differences (ΔE*) between the reference and the textile substrates showed that the textile surface structure influenced the color differences on the white substrates as it did for the black substrates. In all cases, except for at the 45° angle, the influence of surface roughness on ΔE* can be seen in the following order: F/white/EP, SD/white/EP and SO/white/EP. Considering that the ΔE*-values for white textile substrates were almost half as low compared to the black substrates (see Figure 9), and that the L*-values for white substrates (see Figure 11(b)) were higher than 80 for all detection angles, suggests that the color-changing effects were not as accentuated on the white as on the black substrates. However, since a total color difference of ΔE* = 0.29 can be differentiated by the human eye, a small color-changing effect can still be distinguished on all of the white textile substrates in the different detection angles.

OPEN IN VIEWER

Figure 12.

Total color differences ΔE* at various detection angles between test chart (TC)/white/effect pigment (EP) and white textile samples with effect pigment coating. F: Filament; SD: Spun Dense; SO: Spun Open.