Effect of yarn characteristics on surface properties of knitted fabrics

Abstract

Fabric surface properties are significant in terms of fabric handle, sensorial comfort, aesthetic and performance properties. Yarn properties are among the most important parameters that affect fabric surface properties. Besides, fiber type, fiber properties and spinning technology etc. directly affect the physical, mechanical and performance properties of yarns as well as fabric surface properties. In the scope of this study, effects of fiber type (raw material), fiber fineness and fiber length on the surface properties of fabrics were investigated. Also, properties of yarns were measured and their effects on fabric surface properties were analyzed. For this purpose, unevenness, optical unevenness, imperfections, structural properties (diameter, density, roughness and shape), hairiness and frictional properties of yarns were measured, and relationships between abrasion resistance, pilling and frictional properties of knitted fabrics were examined. Regression models were developed in order to predict fabric surface properties from yarn characteristics. Based on comprehensive data analysis, it was concluded that variation in yarn friction and yarn hairiness explains approximately 80–85% of fabric-to-fabric and fabric-to-skin (gazelle skin) friction coefficients. Furthermore, positive correlations between yarn hairiness and weight loss, and yarn hairiness and thickness change after abrasion test, were observed. Additionally, a new parameter, the optical contact index (OCI), based on an image analysis method, was suggested to determine the surface properties and roughness of fabrics. Relationships between the OCI and other tested fabric surface properties were statistically analyzed. Statistical analyses showed that high correlations exist between the new parameter and fabric friction and abrasion resistance at the 0.05 significance level.

Keywords

yarn friction yarn hairiness fabric friction fabric roughness abrasion resistance image analysis method

Fabric surface properties are crucial, because they are directly related to many properties of fabrics. As it is known, surfaces of fabrics are not completely smooth. They have regular and irregular roughness depending on both yarn characteristics and fabric properties. Yarn properties such as unevenness, hairiness, friction, roughness, shape, diameter, density, spinning technology and the raw material, and fabric properties such as fabric type and knitting structure, are the most important parameters that should be taken into consideration during the analysis of fabric surface properties. In recent years, with the increasing importance of comfort perception and fabric handle, many researchers have been dealing with the analysis of fabric surface properties.^1–13 Moreover, image analysis methods have been used to determine different properties of yarns and fabrics. Objective evaluation of fabric surface properties is an area where image analysis methods are used.^14–20

Ajayi and Elder¹ compared the relationship between yarn-to-yarn (YY) and fabric-to-fabric friction. They concluded that the coefficients of friction of fabrics are greater than those of their component yarns, and that yarns with higher frictional properties yield fabrics whose frictional properties are also higher. In another study, Ajayi and Elder² examined the relationships between fabric friction, handle and compression. The results showed that higher fabric compression causes larger difference between static- and kinetic-friction forces. Rankumar et al.³ examined the influence of structural variables on the frictional properties of 1 × 1 rib-knitted cotton fabrics. The results indicated that both the loop length and the yarn linear density influence the fabric-to-fabric frictional properties; they defined a new frictional constant (K) and established an empirical relationship between the K and structural variables. Bertaux et al.⁷ investigated the relationship between friction and the tactile properties of woven and knitted fabrics, and found high correlations between friction and tactile properties for knitted fabrics. They emphasized that other parameters such as hairiness, bending, unit weight and thickness play important roles in these relationships. Xu et al.¹³ developed a new profilometer to assess fabric smoothness appearance by using laser triangulation and image processing techniques. They found that there is a high correlation between the results of subjective tests and parameters obtained from the new profilometer. Ravandi and Ghane¹⁴ analyzed fundamental factors affecting fabric surface protrusion. The results indicated that the protruding yarn density is strongly influenced by the associated yarn-yarn interaction at the crossing points and yarn spacing. They stated that there is high correlation between the normal load, yarn spacing and protruding yarn density. Semmani et al.¹⁹ focused on measuring the roughness of weft-knitted fabrics using a non-contact method. The results obtained from the image analysis were compared with mean absolute deviation of surface roughness (SMD) values measured by the Kawabata method. They found strong and negative correlation between fabric roughness values measured by the two different methods. Moezzi et al.²⁰ analyzed the effect of unevenness of weft yarns on asperities on surfaces of woven fabrics by the image analysis method. It is concluded that an increasing amount of irregularity and <50% thick places cause an increase in the count of asperities on fabric surface. In addition, the relationship between the parameter angular power spectrum, defined by the image analysis method and <50% thick places, was found to be fairly strong.

This study aimed to analyze fabric surface properties and the effects of yarn characteristics on surface properties. For this purpose, single jersey and interlock fabrics were systematically produced from natural, regenerated and synthetic yarns. Yarn properties and fabric surface properties were measured by some objective test methods. Regression models were developed to predict fabric surface properties from yarn properties. Furthermore, the surface properties of fabrics were analyzed by using image analysis, and correlations between the parameter obtained from image analysis method and other tested fabric surface properties were statistically analyzed.

Materials and methods

Materials

In this study, single jersey and interlock-knitted fabrics were produced from natural, regenerated and synthetic ring yarns. In order to investigate the effect of the raw material on fabric surface properties, regenerated and synthetic fibers with the same fiber fineness (1.3 dtex) and fiber length (38 mm) were used. Furthermore, in order to analyze the effect of fiber fineness and fiber length, polyester yarns were used with different fineness (1.3 and 1.6 dtex) and different lengths (32 and 38 mm). Cotton fibers were used as reference fibers. Structural properties of the raw materials are given in Table 1. All yarns were produced systematically in a ring-spinning machine with 20 tex linear density and α_tex 34.4 twist factor, which is suitable for knitted fabrics. The fibers of all yarns were passed through the same production steps to analyze the effect of the raw material accurately. The cotton fibers were passed through the combing process.

Table 1.

Structural properties of raw materials

Raw material	Fiber fineness (dtex)	Fiber length (mm)
100% cotton	1.4	29.5
100% Modal	1.3	38
100% Tencel	1.3	38
100% acrylic	1.3	38
100% polyester	1.3	38
100% polyester	1.3	32
100% polyester	1.6	38

Knitted fabrics were produced by using the same machine settings and loop lengths for both single jersey and interlock structures. Structural properties of the fabrics are given in Table 2.

Table 2.

Structural properties of fabrics

Raw material	Knitting type	Course per cm (cpc)	Wale per cm (wpc)	Loop length (mm)	Unit weight (g/m²)	Thickness (mm)
100% cotton	Single jersey	13.7	18.7	2.68	144.48	0.60
100% cotton	Interlock	11.6	15.1	3.20	225.30	1.24
100% Modal	Single jersey	13.4	17.0	2.76	127.11	0.48
100% Modal	Interlock	11.6	14.8	3.22	230.46	1.04
100% Tencel	Single jersey	12.8	19.2	2.65	136.56	0.53
100% Tencel	Interlock	11.9	14.5	3.38	232.06	1.14
100% acrylic	Single jersey	13.2	19.5	2.63	140.40	0.53
100% acrylic	Interlock	11.3	15.1	3.30	240.34	1.11
100% polyester (1.3 dtex-38 mm)	Single jersey	14.2	18.2	2.64	140.37	0.54
100% polyester (1.3 dtex-38 mm)	Interlock	12.5	15.0	3.38	251.42	1.35
100% polyester (1.3 dtex-32 mm)	Single jersey	13.9	18.6	2.73	142.67	0.52
100% polyester (1.3 dtex-32 mm)	Interlock	13.3	15.0	3.32	261.27	1.34
100% polyester (1.6 dtex-38 mm)	Single jersey	13.7	18.8	2.62	130.56	0.51
100% polyester (1.6 dtex-38 mm)	Interlock	13.5	14.8	3.26	262.24	1.34

In order to avoid any change in the surface properties of yarns and fabrics, treatments (such as waxing, bleaching or dyeing) were not applied to yarns or fabrics before and during their production.

Methods

Measurement of yarn properties

The unevenness, optical unevenness, imperfections and other structural properties (diameter, density, roughness and shape) of yarns were measured by an Uster Tester 5 S800 considering ASTM (American Society for Testing and Materials) D1425/D1425M-14.²¹ Yarn hairiness properties were measured by an Uster Tester 5 S800 (UT5) and an Uster Zweigle Hairiness Tester 5 (UZHT5), with 5 cN pretension considering ASTM D5647-07.²² The hairiness value obtained from the UT5 is called H and the hairiness values obtained from UZHT5 are called S3 and S1 + 2. While the H value represents the total length of protruding hairs per centimeter, the S3 value is the sum of all protruding fibers ≥3 mm (cumulative) and S1+2 value is the total amount of the fibers <3 mm.^23,24

Measurement of YY, yarn-to-metal (YM) and yarn-to-ceramic (YC) friction were performed according to ASTM D3108/D3108M-13²⁵ and ASTM D3412/D3412M-13²⁶ by Lawson Hemphill CTT, and the Capstan method was used for all surfaces. In this study, the YY, YM and YC tests were performed by using 5 cN input tension, which is within the specified range for knitted fabrics^27,28 and is also equal to pretension in UZHT5.

Measurement of fabric properties

Friction attachment, which is adaptable to a tensile tester, was used for the measurements of fabric-to-fabric and fabric-to-skin friction. During the fabric friction measurements, five test specimens were tested both in wale and course directions. Face sides of specimens were tested with a speed of 50 mm/min for 3 minutes test duration. Normal load was selected as 1.2 g/cm², and kinetic and static friction coefficients (µk and µs) were calculated by using frictional forces. For fabric-to-material friction tests, gazelle skin was used as material. The average roughness value (Ra) of gazelle skin was measured as 23.44 µm by an Ambios XP-2 high resolution surface profilemeter. The average roughness value of gazelle skin is within the roughness value range for human skin, as specified by Derler and Gerhardt.²⁹

Abrasion resistance and pilling tendency tests were performed by a James H. Heal Nu-Martindale Abrasion and Pilling Tester. Five fresh specimens were tested for both abrasion resistance and pilling tests. The weight losses (mg) and changes of thickness (mm) of the samples were calculated at the end of 15000 cycles to measure the abrasion resistance of the fabrics considering ISO 12947-3:1998.³⁰ Pilling tendencies of fabrics were determined in accordance with ISO 12945-2:2002.³¹ In this method, ratings for samples tested were determined by comparing standard photographs. In this standard, a 1–5 rating scale is used, where ‘5’ indicates that there is no visible change and ‘1’ indicates intensive pilling on the whole surface of the fabric.

In this study, an image analysis method was also used to examine roughness and surface properties of the fabrics. Images of fabrics were captured by a digital camera integrated to a microscope (Olympus BX43) with a 10× magnifying lens. The image capturing system and steps of image analysis are shown in Figure 1. Figure 1(a) shows the captured image of fabric without a light source. In the second step, images were captured with a light source-emitting stationary light-emitting diode light beam (Figure 1(b)) to obtain more detailed surface images appropriate for digital processing. These images were converted into gray-level images (Figure 1(c)) and binary images (Figure 1(d)), respectively, by using the Matlab image processing toolbox. The aim of this process was to determine asperities of the fabric surface. Asperities cause roughness and occur contact points. A new parameter called the optical contact index (OCI), which is based on the image analysis method, was suggested to determine the surface properties of fabrics within the context of study. OCI was calculated by using a suitable threshold value, which is the same for all images. The value of the pixels higher than the defined threshold value was converted into white pixels and lower ones were converted into black pixels using the Otsu method.³² In binary images, white pixels represented contact points (asperities) and black pixels represented non-contact points. Finally, the OCI (%) was calculated as the ratio of white pixels to the total pixels of the image.³³

Figure 1.

Image capturing system and steps for determining the optical contact index by image analysis.

Results and discussion

Yarn properties

The mean values and SDs of unevenness, optical unevenness, imperfections, structural properties (diameter, density, roughness and shape), hairiness and friction for all tested yarns are given in Table 3.

Table 3.

Structural properties, hairiness and friction coefficient values of yarns*

Parameters	Raw material
Parameters	Cotton	Modal	Tencel	Acrylic	Polyester 1(1.3-38)	Polyester 2(1.3-32)	Polyester 3(1.6-38)
Unevenness (% CVm)	11.40	10.82	10.40	10.63	11.21	11.46	12.09
Unevenness (% CVm)	(0.15)	(0.18)	(0.15)	(0.13)	(0.26)	(0.30)	(0.29)
Optical unevenness (% CV2D 0.3 mm)	13.19	11.95	11.47	11.78	12.53	14.10	13.81
Optical unevenness (% CV2D 0.3 mm)	(0.29)	(0.25)	(0.23)	(0.36)	(0.15)	(0.25)	(0.31)
−50% (thin place) (count/km)	0.0	0.0	0.0	0.8	1.3	6.7	8.3
−50% (thin place) (count/km)	(0.0)	(0.0)	(0.0)	(0.9)	(0.8)	(4.4)	(3.8)
+50% (thick place) (count/km)	5.8	4.6	3.3	1.3	4.6	10.8	15.4
+50% (thick place) (count/km)	(3.8)	(4.3)	(3.0)	(1.3)	(1.9)	(2.0)	(3.3)
+200% neps (count/km)	13.3	12.9	14.2	0.8	13.8	14.2	20.8
+200% neps (count/km)	(2.6)	(3.7)	(3.0)	(2.0)	(7.4)	(7.4)	(4.1)
Diameter (2DØ mm)	0.23	0.21	0.21	0.22	0.20	0.21	0.21
Diameter (2DØ mm)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)	(0.00)
Density (g/cm³)	0.46	0.57	0.60	0.51	0.61	0.58	0.59
Density (g/cm³)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)	(0.01)
Shape (roundness)	0.83	0.84	0.86	0.85	0.85	0.84	0.85
Shape (roundness)	(0.00)	(0.00)	(0.00)	(0.01)	(0.01)	(0.01)	(0.01)
Roughness (CV FS %)	8.84	8.50	7.69	7.21	7.85	9.39	8.51
Roughness (CV FS %)	(0.25)	(0.24)	(0.16)	(0.48)	(0.27)	(0.17)	(0.19)
Hairiness (H)	5.96	5.84	5.94	5.91	4.79	6.31	5.68
Hairiness (H)	(0.20)	(0.21)	(0.21)	(0.28)	(0.11)	(0.08)	(0.13)
SD of hairiness	1.34	1.29	1.30	1.43	1.17	1.46	1.47
SD of hairiness	(0.06)	(0.03)	(0.04)	(0.05)	(0.05)	(0.04)	(0.06)
S1+2 (count/100 m)	26262.0	21277.8	22968.3	19326.3	18716.5	19995.3	19096.0
S1+2 (count/100 m)	(1756.9)	(871.8)	(1096.7)	(990.3)	(484.4)	(845.3)	(351.1)
S3 (count/100 m)	3500.4	3201.8	3358.7	2995.8	2393.4	2643.4	2450.0
S3 (count/100 m)	(273.9)	(100.1)	(63.8)	(114.4)	(153.7)	(85.6)	(73.8)
µ_yarn-to-yarn **	0.293	0.250	0.242	0.242	0.339	0.369	0.359
µ_yarn-to-yarn **	(0.003)	(0.004)	(0.004)	(0.004)	(0.005)	(0.007)	(0.007)
µ_{yarn-to-metal} **	0.284	0.272	0.235	0.270	0.201	0.274	0.264
µ_{yarn-to-metal} **	(0.002)	(0.001)	(0.002)	(0.003)	(0.002)	(0.004)	(0.004)
µ_{yarn-to-ceramic} **	0.294	0.287	0.272	0.286	0.267	0.316	0.311
µ_{yarn-to-ceramic} **	(0.003)	(0.002)	(0.003)	(0.001)	(0.004)	(0.002)	(0.002)

SDs are given in parentheses.

Capstan method was used for yarn friction tests on all surfaces.

As a result of the statistical analyses, it is concluded that the effect of the raw material is statistically significant for hairiness and the structural and frictional properties of yarns. When the capacitive and optical unevenness values of yarns with the same fiber fineness (1.3 dtex) and same cut length (38 mm) are examined, it is seen that polyester yarns have the highest values and Tencel yarns have the lowest values. When Uster statistics are examined, polyester yarns have higher values for the capacitive irregularity values of yarns produced from fibers having the same fineness and length. Since all raw materials have the same linear density (1.3 dtex), the measured values are different from each other even though the calculated theoretical limit irregularity values of all yarns are the same (CV_lim% = 8.06). The reason is that yarn irregularity is also influenced by other fiber properties such as friction, bending and static electricity, as well as fiber length and strength.^34,35 Acrylic yarns have the highest diameter values while Tencel yarns have the lowest. Acrylic yarns have the minimum density (g/cm³) and roughness (CV FS %-CV for fine structure of the yarn surface) values. The shape (roundness) values of tested yarns are close to each other and these values range from 0.84–0.86. For both hairiness devices, the highest hairiness values were measured for Tencel yarns and the lowest for polyester yarns. It is seen that the highest sH (SD of hairiness) values belong to acrylic yarns and the lowest sH values belong to polyester yarns, and it is observed that polyester yarns have the lowest YM and YC, and the highest YY, friction coefficient values.

ANOVA (analysis of variance) results showed that the effects of fiber fineness and length are statistically significant for all properties of polyester yarns. It is possible to say that yarns produced from thinner and longer fibers have lower unevenness, roughness, hairiness and friction values, in accordance with earlier studies.^36–41 While the highest capacitive unevenness values belong to 1.6 dtex-38 mm polyester yarns, the highest optical unevenness, roughness, yarn hairiness, and YY and YM friction values belong to polyester yarns produced from 1.3 dtex-32 mm.

Surface properties of fabrics

Mean values of fabric-to-fabric and fabric-to-skin static and kinetic friction coefficients for both fabric directions are given in Table 4. The static friction force and coefficient are momentary concepts that appear at the beginning of movement. Moreover, the kinetic friction force and coefficient come up while the object is in motion, and they become important while the surfaces move over each other.

Table 4.

Fabric-to-fabric and fabric-to-skin friction coefficients

	Knitting structure	Raw material	Fabric direction
			Wale direction		Course direction
			µs	µk	µs	µk
Fabric-to-fabric	Single jersey	Cotton	1.215	0.715	1.090	0.672
		Modal	1.150	0.600	1.150	0.550
		Tencel	1.069	0.673	1.234	0.679
		Acrylic	1.266	0.824	1.071	0.740
		Polyester 1.3-38	1.418	1.046	1.469	0.988
		Polyester 1.3-32	1.547	1.169	1.657	1.217
		Polyester 1.6-38	1.517	1.165	1.627	1.254
	Interlock	Cotton	1.234	0.713	1.162	0.687
		Modal	1.110	0.652	1.141	0.559
		Tencel	1.030	0.627	1.036	0.604
		Acrylic	1.225	0.863	1.109	0.723
		Polyester 1.3-38	1.421	1.054	1.492	1.083
		Polyester 1.3-32	1.566	1.277	1.561	1.238
		Polyester 1.6-38	1.442	1.229	1.549	1.142
Fabric-to-skin	Single jersey	Cotton	0.203	0.115	0.192	0.113
		Modal	0.213	0.099	0.191	0.093
		Tencel	0.190	0.105	0.180	0.115
		Acrylic	0.276	0.155	0.246	0.146
		Polyester 1.3-38	0.183	0.117	0.183	0.116
		Polyester 1.3-32	0.228	0.163	0.206	0.128
		Polyester 1.6-38	0.242	0.156	0.235	0.148
	Interlock	Cotton	0.214	0.119	0.203	0.098
		Modal	0.211	0.087	0.194	0.077
		Tencel	0.189	0.108	0.205	0.114
		Acrylic	0.233	0.140	0.235	0.147
		Polyester 1.3-38	0.213	0.116	0.172	0.100
		Polyester 1.3-32	0.231	0.174	0.217	0.170
		Polyester 1.6-38	0.264	0.173	0.249	0.153

Fabric-to-fabric static friction coefficients range from 1.03–1.49, while values of fabric-to-fabric kinetic friction coefficients range from 0.60–1.05 for the single jersey and interlock-knitted fabrics produced by ring-spun yarns made of the same linear density (1.3 dtex) and fiber length (38 mm) (Table 4).

When the effect of raw material is examined, it is seen that 100% polyester fabrics have the highest fabric-to-fabric kinetic and static friction coefficients for both knitting structures. In addition, it is remarkable that values of kinetic friction coefficients for fabrics with the same raw materials and different knitting structures are very close to each other, and that the results are parallel for both knitting structures. Values of fabric-to-fabric static friction coefficients range from 1.09–1.23 and values of fabric-to-fabric kinetic friction coefficients range from 0.67–0.72 for 100% cotton fabrics, which were used as a reference. Values of fabric-to-skin static friction coefficients of polyester, acrylic, Tencel and Modal fabrics range from 0.17–0.28, while values of fabric-to-skin kinetic friction coefficients range from 0.08–0.15 (Table 4). Values of static and kinetic friction coefficients for 100% cotton fabrics are between 0.19–0.21 and 0.10–0.12, respectively. When the results are examined, it is seen that fabric-to-fabric friction test results are generally parallel to YY friction test results, and that fabric-to-skin friction test results are in line with YM friction test results. A higher variation in fabric friction can be explained by the fact that the behavior of yarns in the fabric structure differs from the behavior of single yarns. It is thought that the yarn intersections that occur during the fabric formation may influence the fabric friction results. This situation is observed more clearly because of the fabric interaction during the fabric-to-fabric friction. The reason for getting lower fabric-to-skin friction coefficients than fabric-to-fabric friction coefficients could be the fairly low roughness value of the gazelle skin surface used as material. Literature shows that rubber, polymeric materials and wood etc. are used as material in fabric-to-material friction tests and results vary over a large scale.^42–47 Moreover, results show that 100% acrylic fabrics have the highest kinetic friction coefficients while 100% Modal fabrics have the lowest for both directions (wale and course) and both structures (single jersey and interlock). Fabric-to-fabric and fabric-to-skin static and kinetic friction coefficient values and 95% confidence interval plots of knitted fabrics are given in Figure 2 and Figure 3. Results showed that the effect of raw material is statistically significant for fabric-to-fabric and fabric-to-skin friction (p < 0.05). In addition, it is observed that fabrics made of synthetic yarns have the highest values, whereas fabrics made of regenerated yarns have the lowest values for both knitting structures and frictional surfaces.

Figure 2.

Coefficients of fabric-to-fabric friction and 95% confidence intervals.

Figure 3.

Coefficients of fabric-to-skin friction and 95% confidence intervals.

Figure 4.

Weight loss (mg) and thickness change (mm) values, and 95% confidence interval. Minus values indicate a decrease in thickness.

Results showed that the fabric-to-fabric kinetic friction coefficients vary between 0.99 and 1.28, while the fabric-to-skin kinetic friction coefficients vary between 0.10 and 0.17 for polyester fabrics with different fiber lengths and fiber fineness. As in the case of yarn friction, the lowest values for fabric friction belong to fabrics containing polyester fibers of 1.3 dtex-38 mm. Statistical analyses showed that fiber fineness and fiber length have statistically significant effects on both fabric-to-fabric and fabric-to-skin friction (p < 0.05).

The abrasion resistance of fabrics is not only one of the most important surface properties, but is also crucial for fabric performance. Abrasion affects the appearance of the fabric by creating some physical changes and it also triggers significant loss of fabric strength. In this study, the weight loss (mg) and fabric thickness change (mm) after 15000 cycles of abrasion resistance testing were evaluated. ANOVA results showed that the effects of raw materials are statistically significant for both weight loss and fabric thickness change values for both of the knitting structures (p < 0.05). Weight loss (mg), thickness change (mm) values and 95% confidence interval graphs are given in Figure 4. The 100% Tencel fabrics have the highest weight loss values, whereas 100% polyester fabrics have the lowest for both knitting structures. A possible reason for this situation could be the higher tenacity of polyester fibers. Moreover, the more hairy structure of Tencel yarns could be the reason for the higher weight loss of fabrics made of these yarns. The change in fabric thickness after the abrasion test was the highest for 100% Tencel and the lowest for 100% polyester for both single jersey and interlock fabrics. In polyester fabrics, a small number of fibers is removed from the surface due to the high abrasion resistance, therefore the fabric thickness has slightly changed. In the fabrics produced from Tencel yarns, due to high yarn hairiness, more fibers are removed from the fabric structure and the change in thickness is greater. When the weight loss and thickness change values of the fabrics produced from polyester fibers with different linear densities and fiber lengths are examined, the lowest values belong to fabrics produced from 1.3 dtex-38 mm polyester fibers for both fabric structures, and the effect of fiber fineness and fiber length is statistically significant (p < 0.05).

Pilling is an important problem, which causes poor appearance on a fabric’s surface, and it is the result of scrubbing action on loose fibers on a fabric’s surface. As seen in Table 5, the pilling tendencies of the fabrics produced by yarns made of fibers with the same linear densities and fiber lengths are pretty high and close to each other. Similarly, pilling tendencies of the fabrics produced by polyester fibers with different linear densities and fiber lengths are also pretty high and close to each other. In general, considering the knitting structures, it could be stated that interlock fabrics have approximately half- or one-degree lower pilling ratings than single jersey fabrics.

Table 5.

Pilling ratings of the fabrics

Knitting structure	Raw material	Pilling ratings
Single jersey	Cotton	1
	Modal	2–3
	Tencel	2
	Acrylic	2–3
	Polyester 1.3-38	2
	Polyester 1.3-32	2
	Polyester 1.6-38	2
Interlock	Cotton	1
	Modal	1–2
	Tencel	1
	Acrylic	1–2
	Polyester 1.3-38	2
	Polyester 1.3-32	1
	Polyester 1.6-38	1

In the scope of this study, the term OCI was defined via the analysis of fabric surface images and was suggested to objectively determine fabric surface roughness properties. In Figure 5, binary images of fabrics having different OCI values are shown. OCI mean values of fabrics are given in Table 6. As seen in Table 6, Tencel fabrics have the highest OCI values while polyester fabrics have the lowest for both knitting structures. The ANOVA results showed that differences between the OCI values of the fabrics produced by yarns made of the same fiber fineness and lengths are statistically significant for both knitting structures (p < 0.05). For polyester fabrics, fiber length has a statistically significant effect on OCI values (p < 0.05) and the highest OCI values belong to fabrics that is made of 1.3 dtex-32 mm fibers for both knitting structures.

Figure 5.

Binary images of fabrics with different optical contact index values.

Table 6.

Optical contact index mean values of the fabrics and p values of analysis of variance

Knitting structure	Raw material	Optical contact index (%)	p values
Single jersey	Cotton	12.03	0.000
	Modal	15.39
	Tencel	16.58
	Acrylic	15.32
	Polyester 1.3-38	9.96
	Polyester 1.3-32	10.99
	Polyester 1.6-38	9.88
Interlock	Cotton	13.72	0.000
	Modal	14.64
	Tencel	17.78
	Acrylic	14.92
	Polyester 1.3-38	9.77
	Polyester 1.3-32	10.94
	Polyester 1.6-38	9.77

In order to analyze relationships between the OCI and other measured fabric surface properties, Pearson correlation coefficients were investigated. As a result of statistical analyses, it is seen that there are significant correlations between OCI, weight loss and fabric-to-fabric friction (Table 7).

Table 7.

Pearson correlation coefficients of optical contact index values and measured surface properties of fabrics

Pearson correlation coefficients	All fabrics	Single jersey	Interlock
Optical contact index (%) – Weight loss after abrasion (mg)	0.931*	0.896*	0.962*
Optical contact index (%) – Thickness change after abrasion (mm)	0.721*	0.850*	0.959*
Optical contact index (%) – Fabric-to-fabric friction (wale)	−0.842*	−0.828*	−0.863*
Optical contact index (%) – Fabric-to-fabric friction (course)	−0.843*	−0.801*	−0.883*
Optical contact index (%) – Fabric-to-fabric friction**	−0.840*	−0.810*	−0.868*
Optical contact index (%) – Fabric-to-skin friction (wale)	−0.513	−0.433	−0.579
Optical contact index (%) – Fabric-to-skin friction (course)	−0.324	0.325	−0.338
Optical contact index (%) – Fabric-to-skin friction**	−0.418	−0.378	−.0452

Statistically significant for α = 0.05.

The overall results for fabric friction tests (without considering the direction).

The relationship between the OCI parameter and the abrasion resistance of fabrics and fabric-to-fabric kinetic friction values are shown in Figure 6 and Figure 7. There is strong positive correlation between the OCI and weight loss values after 15000 turn abrasion testing for single jersey and interlock fabrics. However, negative correlation was found between OCI and fabric-to-fabric kinetic friction values for wale and course directions, and for both knitting structures. The negative relationship between the OCI values and the fabric-to-fabric friction coefficient can be explained by the number of stick-slip movements and the contact area formed during the friction tests. As the OCI values increase, the contact area (surface area) increases and the number of stick-slips that occur during friction tests under low loads decreases (Figure 5). With the increase of the contact area, the load per unit surface (normal load, gf/mm²) will decrease and therefore a reduction in the coefficient of friction will occur. OCI values represent the protruding regions (asperities) on the fabric surface. When an evaluation is made in terms of abrasion, these regions cause weight loss with mechanical effect under the applied loads. For this reason, there is a positive and strong relationship between weight loss values and OCI values.

Figure 6.

Relationships between optical contact index and other surface properties for single jersey fabrics.

Figure 7.

Relationships between optical contact index and other surface properties for interlock fabrics.

Effect of yarn characteristics on fabric surface properties

In order to examine the effects of yarn properties on fabric surface properties and to predict fabric surface properties from yarn properties, multiple regression analyses were performed. Dependent variables used in regression models and abbreviations of these variables are given in Table 8.

Table 8.

Dependent variables and abbreviations

Dependent variables	Abbreviations
Coefficient of fabric-to-fabric kinetic friction (wale direction)	FF_W
Coefficient of fabric-to-fabric kinetic friction (course direction)	FF_C
Coefficient of fabric-to-fabric kinetic friction*	FF
Coefficient of fabric-to-skin kinetic friction (wale direction)	FS_W
Coefficient of fabric-to-skin kinetic friction (course direction)	FS_C
Coefficient of fabric-to-skin kinetic friction*	FS
Fabric weight loss after abrasion test (mg)	Weightloss_mg
Thickness change after abrasion test (mm)	Thickness_mm

The overall results for fabric friction tests (without considering the direction).

Regression models were developed to predict fabric-to-fabric and fabric-to-skin friction coefficients for each direction. To predict the fabric-to-fabric kinetic friction coefficient in wale direction, four models were developed. These models include YY friction coefficient, hairiness caused by short fibers (S12), sH and YC friction coefficient as independent variables. Coefficients of determination of regression models are given in Table 9.

Table 9.

Coefficients of determination for fabric-to-fabric kinetic friction (wale direction)

Model	R	R ²	Adjusted R²	Standard error of the estimate
1	0.905^a	0.819	0.804	0.1092047
2	0.963^b	0.927	0.914	0.0722944
3	0.982^c	0.964	0.953	0.0534673
4	0.991^d	0.982	0.974	0.0394689

Independent variables: (constant), YY.

Independent variables: (constant), YY and S12.

Independent variables: (constant), YY, S12 and sH.

Independent variables: (constant), YY, S12, sH and YC.

It is seen that 80.4% of the fabric-to-fabric friction coefficient (wale direction) can be explained by only the YY friction coefficient (Table 9). This ratio increases up to 97.4% by adding hairiness caused by short fibers (S12), sH and the YC friction coefficient as independent variables into the regression models. The regression equation for model 4 is given in equation (1).

\begin{matrix} FF_W = 0.256 + 4.202 * YY - 2.67 * 10^{- 5} S 12 \\ + 1.249 * sH - 5.918 * YC \end{matrix}

(1)

Upon the prediction of the fabric-to-fabric kinetic friction coefficient in the course direction by multiple regression analysis, three models were developed, which are explained by the YY friction coefficient (YY), sH and shape. The coefficients of determination of regression models are shown in Table 10 and the regression equation of the third model is given in equation (2). The YY friction coefficient has a considerable effect in predicting the change in fabric-to-fabric friction coefficient in the course direction (FF_C) as well as the fabric-to-fabric friction coefficient in wale direction, and 86.1% of FF_C could be explained by the yarn-to-yarn friction coefficient. Furthermore, correlation between YY and FF_C equals 0.934.

\begin{matrix} FF_C = - 9.077 + 5.049 * YY + 9.448 * Shape \\ + 0.331 * sH \end{matrix}

(2)

Table 10.

Coefficients of determination for fabric-to-fabric kinetic friction (course direction)

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.934^a	.872	.861	0.1010243
2	.979^b	.959	.951	0.0597074
3	.987^c	.974	.966	0.0502145

Independent variables: (constant), YY.

Independent variables: (constant), YY, shape.

Independent variables: (constant), YY, shape and sH.

When the regression equations are formulated to estimate the fabric-to-skin friction coefficient in the wale direction, -50% thin place (Thin50) and +200% neps (Neps200) are taken as independent variables by the models. Approximately 88% of the fabric-to-skin friction coefficient in this direction can be explained by these variables. The regression coefficients of the models are given in Table 11 and the regression equation for the second model is given in equation (3). The regression equations, which were developed to predict fabric-to-skin friction in the course direction, include the sH and YM friction coefficient as independent variables. The rate of explanation of fabric-to-skin friction with these variables is 79%. Equation (4) shows the regression model for the fabric-to-skin friction coefficient (course direction) and Table 12 shows coefficients of determination.

FS_W = 0.142 + 0.01 * Thin 50 - 0.003 * Neps 200

(3)

FS_C = - 0.146 + 0.336 * sH - 0.721 * YM

(4)

Table 11.

Coefficients of determination for fabric-to-skin kinetic friction (wale direction)

Model	R	R²	Adjusted R²	Standard error of the estimate
1	0.834^a	0.695	0.670	0.0166871
2	0.949^b	0.900	0.882	0.0099843

Independent variables: (constant), Thin50.

Independent variables: (constant), Thin50 and Neps200.

Table 12.

Coefficients of determination for fabric-to-skin kinetic friction (course direction)

Model	R	R²	Adjusted R²	Std. Error of the Estimate
1	.741^a	.549	.511	.0186952
2	.904^b	.818	.785	.0124047

Independent variables: (Constant),sH

Independent variables: (Constant), sH and YM

On the basis of the reached statistical results, it is possible to say that regression equations developed to predict fabric-to-fabric and fabric-to-skin friction coefficients through yarn properties mostly include yarn friction, yarn hairiness and hairiness change as independent variables. For this reason, the regression analyses were performed again according to the Enter method in order to estimate the fabric-to-fabric and fabric-to-skin friction coefficients using these parameters obtained from only two devices (yarn friction and yarn hairiness). The regression models and the coefficients of determination for these models are given in Table 13. It is seen that coefficients of determination (adjusted R²) vary between 0.77 and 0.90. For the knitted fabrics used in the scope of this study, it can be concluded that fabric friction, which is very important in terms of fabric surface properties, could be explained by some yarn characteristics obtained by two-test devices.

Table 13.

Regression models of fabric friction coefficient

Regression Models	Adjusted R²
$FF_W = - 0.549 + 3.261 * YY - 0.206 H + 1.232 * sH$	0.884
$FF_C = - 0.741 + 3.992 * YY - 0.157 H + 0.978 * sH$	0.896
${FF}^{a} = - 0.645 + 3.627 * YY - 0.182 H + 1.105 * sH$	0.896
$FS_W = - 0.146 - 0.432 * YM - 0.021 * H + 0.378 * sH$	0.799
$FS_C = - 0.136 - 0.655 * YM - 0.007 H + 0.344 * sH$	0.768
${FS}^{a} = - 0.141 - 0.543 * YM - 0.014 H + 0.361 * sH$	0.785

The overall results for fabric friction tests (without considering the direction).

When a regression analysis was performed to estimate the weight losses (mg) in the fabrics after fabric abrasion resistance testing, one model was obtained that included the YY friction coefficient as independent variable. Approximately 65% of the weight loss (mg) is explained by the YY friction coefficient (Table 14 and equation (5)). Although this ratio is fairly high for a model with one independent variable, it is considered that this equation will not be sufficient to explain the weight loss without using fabric structural parameters. There is a similar situation for fabric thickness change after abrasion testing, and a strong positive correlation exists between fabric thickness change and the S3 value (R = 0.680).

Weightloss_mg = 57.992 - 153.841 * YY

(5)

Table 14.

Coefficients of determination for weight losses after abrasion

Model	R	R ²	Adjusted R²	Standard of the estimate
1	0.813^a	0.661	0.633	6.22854

Independent variables: (constant), YY.

Conclusions

In this study, the surface properties of single jersey and interlock-knitted fabrics produced by natural (cotton), regenerated (Modal and Tencel) and synthetic (polyester and acrylic) ring yarns, and the effects of fiber and yarn characteristics on these properties, were investigated. Unevenness, optical unevenness, imperfections, structural properties (diameter, density, roughness and shape), hairiness and frictional properties of the yarns, as well as abrasion resistance, pilling and frictional properties of the fabrics, were analyzed. It is seen that the effects of raw materials and the structural properties of fibers (linear density and length) are statistically significant for fabric surface properties (p < 0.05). Our results show that polyester fabrics have the highest and Modal fabrics have the lowest fabric-to-fabric friction, and acrylic fabrics have the highest and Tencel and Modal fabrics have the lowest fabric-to-skin friction for both knitting structures and both directions. Tencel fabrics have the highest and polyester fabrics have the lowest weight loss (mg) values among the fabrics produced by yarns made of same the fiber linear density (1.3 dtex) and length (38 mm). For polyester fabrics in which the effect of fiber fineness and length is investigated, the lowest fabric-to-fabric friction, fabric-to-skin friction, weight loss and thickness change values belong to fabrics produced from thinner (1.3 dtex) and longer (38 mm) polyester fibers. When a comparison is made in terms of pilling tendency, it is observed that the pilling tendencies of the fabrics produced by yarns made of fibers with the same linear densities and fiber lengths are pretty high, and that they are close to each other. In the scope of this study, correlation analysis and regression models were used to analyze the effects of yarn characteristics on fabric surface properties, and to predict fabric surface properties from yarn characteristics. It is seen that there is a high positive correlation between YY friction and fabric-to-fabric friction for both knitting structures and both directions. Furthermore, there is a positive correlation between thickness change after abrasion resistance testing and the yarn hairiness (S3) value. When the regression models were analyzed, the obtained results demonstrated that variation in frictional and hairiness properties of yarns could explain approximately 80–85% of fabric-to-fabric and fabric-to-skin friction. These correlation analysis results and regression models may be helpful for further studies on the prediction of surface properties of fabrics, which are very important in terms of fabric handle and sensorial comfort properties.

In addition to the experimental studies within the content of study, a new parameter based on image analysis was also suggested to objectively determine the fabric surface roughness. The new parameter, called the OCI, was calculated to determine the contact area. Relationships between this new parameter and certain fabric surface properties (abrasion resistance and friction coefficient) were also analyzed statistically. Correlation analyses showed that there are strong correlations between the OCI and these fabric surface properties at the 0.05 significance level. For this reason, it is considered that the proposed method could be used for the objective evaluation of fabric surface properties through fabric images, and that it may help save time, cost and labor.

Footnotes

Acknowledgements

Thanks to TÜBİTAK (The Scientific and Technological Research Council of Turkey-2211 National Doctorate Scholarship Program) and Kipaş Holding A.Ş. for their contributions to the study.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

References

Ajayi

Elder

. Comparative studies of yarn and fabric friction. J Testing and Evaluation 1994; 22: 463–467.

Ajayi

Elder

. Fabric friction, handle and compression. J Text Inst 1997; 88: 232–241.

Rankumar

Leaf

GAV

Harlock

. A study of the frictional properties of 1x1 rib-knitted cotton fabrics. J Text Inst 2000; 91: 374–382.

Jeddi

AAA

Shams

Nosraty

et al.

Relations between fabric structure and friction: Part I: Woven fabrics. J Text Inst 2003; 94: 223–234.

Rankumar

Wood

Fox

et al.

Developing polymeric human finger sensor to study the frictional properties of textiles. Part II: Experimental results. Text Res J 2003; 73: 606–610.

Das

Kothari

Vandana

. A study on frictional characteristics of woven fabrics. AUTEX Res J 2005; 5: 133–140.

Bertaux

Lewandowski

Derler

. Relationship between friction and tactile properties for woven and knitted fabrics. Text Res J 2007; 77: 387–396.

Maatoug

Sahnoun

Sakli

. Statistical analysis of surface roughness parameters for weft knitted fabrics measured by the textile surface tester (TST). J Eng Fiber Fabr 2012; 7: 104–112.

Ramalho

Szekeres

Fernandes

. Friction and tactile perception of textile fabrics. Tribol Int 2013; 63: 29–33.

10.

Sülar

Öner

Okur

. The roughness and frictional properties of cotton and polyester woven fabrics. Indian J Fibre Text 2013; 38: 349–356.

11.

Akgun

. Assessment of the surface roughness of cotton fabrics through different yarn and fabric structural properties. Fiber Polym 2014; 15: 405–413.

12.

Becerir

Akgun

Alpay

. Effect of some yarn properties on surface roughness and friction behavior of woven structures. Text Res J 2016; 86: 975–989.

13.

Akgun

Becerir

Alpay

. Investigation of the relation between surface roughness and friction properties of polyester fabrics after abrasion. J Text Inst 2018; 109: 322–337.

14.

Cuminato

Keyes

. Evaluating fabric smoothness appearance with a laser profilometer. Text Res J 1989; 68: 900–906.

15.

Ravandi

HAS

Ghane

. Study of fundamental factors affecting fabric surface protrusion Part I. J Text Inst 2000; 91: 100–106.

16.

Sakaguchi

Wen

Matsumoto

et al.

Image analysis of woven fabric surface irregularity. Text Res J 2001; 71: 666–671.

17.

Bueno

Renner

. Comparison of a new tribological method for the evaluation of the state of a fabric surface with the KES-F surface tester. J Text Inst 2001; 92: 299–314.

18.

Semnani

Yekrang

Ghayoor

. Analysis and measuring surface roughness of nonwovens using machine vision method. Int J Mater Metal Eng 2009; 3: 528–531.

19.

Semmani

Hasani

Behtaj

et al.

Surface roughness measurement of weft knitted fabrics using image processing. Fibres Text East Eur 2011; 19: 55–59.

20.

Moezzi

Ghane

Shahsavari

. The effect of mass irregularities of weft yarns on fabric surface protrusion. J Text Inst 2016; 107: 645–650.

21.

ASTM D1425/D1425M-14. Standard test method for evenness of textile strands using capacitance testing equipment.

22.

ASTM D5647-07. Standard guide for measuring hairiness of yarns by the photo-electric apparatus.

23.

Uster Tester 5 S800 Application Handbook. Uster: Switzerland, 2006.

24.

Uster Zweigle Hairiness Tester 5 Application Handbook. Uster: Switzerland, 2009.

25.

ASTM D3108/D3108M-13. Standard test method for coefficient of friction, yarn to solid material.

26.

ASTM D3412/D3412M-13. Standard test method for coefficient of friction, yarn to yarn.

27.

Henshaw

. Cam forces in weft knitting. Text Res J 1968; 3: 592–598.

28.

Koo

Seok

. Yarn tension variation on the needle during the knitting process. Text Res J 2004; 74: 314–317.

29.

Derler

Gerhardt

. Tribology of skin: Review and analysis of experimental results for the friction coefficient of human skin. Tribol Lett 2012; 45: 1–27.

30.

ISO 12947-3:1998. Determination of the abrasion resistance of fabrics by the Martindale method – Part 3: Determination of mass loss.

31.

ISO 12945-2:2000. Determination of fabric propensity to surface fuzzing and to pilling – Part 2: Modified Martindale method.

32.

Gonzalez

Woods

. Digital image processing, 2nd ed. New Jersey: Prentice Hall, 2004.

33.

Balci Kilic G and Okur A. Determination of surface properties of knitted fabrics by image analysis method. In: Proceedings of the 90th Textile Institute World Conference, Poznan, Poland, 25–28 April 2016. Manchester: The Textile Institute.

34.

Linhart

. On Huberty’s coefficient of irregularity of yarns and its estimators. Text Res J 1965; 35: 1055–1060.

35.

Monfort F and Vroomen F/ Studies in modern yarn production. Manchester: The Textile Institute, 1968, pp.13–14.

36.

Zurek

. Fiber length distribution as a function of the number of fibers in a cross section of a yarn. Text Res J 1962; 32: 255–257.

37.

Subramanian

Grover

Salhotra

. Contribution to the study of the hairiness of ring-spun yarns. J Text Inst 1971; 62: 424–437.

38.

Srivastava

TVK

Onions

Townend

. Some factors affecting the hairiness of worsted spun yarns. J Text Inst 1976; 67: 447–455.

39.

Barella

. Yarn hairiness. Textile Progress 1983; 13: 1–57.

40.

Zeidman

Suh

Subhash

. A new perspective on yarn unevenness: Components and determinants of general unevenness. Text Res J 1990; 60: 1–6.

41.

Nair

Sheela

Vivekanandan

et al.

Studies on friction in cotton textiles: Part I-A Study on the relationships between physical properties and frictional characteristics of cotton fibres and yarns. Indian J of Fibre & Textile Res 2013; 38: 244–250.

42.

Yoon

Sawyer

Buckley

. Improved comfort polyester: Part II: Mechanical and surface properties. Textile Res J 1984; 54: 357–365.

43.

Kothari

Swani

Gangal

. Frictional properties of woven fabrics. Indian J Fibre Text 1991; 16: 251–256.

44.

Ajayi

. Fabric smoothness, friction, and handle. Text Res J 1992; 62: 52–59.

45.

Kenins

. Influence fiber type and moisture on measured fabric-to-skin friction. Text Res J 1994; 64: 722–728.

46.

Kothari

Gangal

. Assessment of frictional properties of some woven fabrics. Indian J Fibre Text Res 1994; 19: 151–155.

47.

Das

Kothari

Vandana

. A study on frictional characteristics of woven fabrics. AUTEX Res J 2005; 5: 133–140.