Associations between the physiological basis of fabric-evoked prickle,fiber and yarn characteristics and the Wool ComfortMeter value

Theoretical analysis

Using Euler’s buckling formula (Equation (2)), calculations were carried out to determine the length and diameter of the fibers capable of achieving a threshold buckling load of 0.74 mN. Numerous researchers have measured Young’s modulus for wool fibers, estimates of which have been shown to vary from 3.22 to 5.40 GPa.^2,3,17 In the present calculation two values for Young’s modulus, 3.5 GPa and 5.4 GPa, were used for wool fibers, and the results were graphed to identify the response surface at which the nerve response would be triggered.

Assuming the WCM is detecting and giving an approximate count of fibers capable of exerting a pre-buckling load of ∼0.74 mN or greater, these numbers can then be related via the relationship demonstrated by Garnsworthy and co-authors between stimulus and sensation intensity, as described in Equation (1).

A tensioned wire with a measurement length of 0.12 m is mounted in the WCM recording head. The distance the recording head travels over the fabric surface during a measurement pass is 0.20 m, and the recording head therefore interacts with fibres protruding from the fabric surface over an area of 0.024 m² (0.12 m × 0.20 m). The mean WCM value of the 48 fabrics in the wearer trials ¹⁰ associated with a scanning area of 0.024 m² was converted to incidents per 10 cm² (Ip). The predicted human sensation (pSp) was then calculated using Equation (1).

Wearer trials

The design and conduct of wearer trials and the full list of garments were exactly as described previously.^10,16 In brief, garments of standard sizes and known construction were evaluated using a set protocol over a range of controlled environments. The full list of garments is provided in Table 1. The test protocol consisted of five sequential stages as follows:

Pre-trial acclimatization, during which no measurements were made.

Wearers in the age group 25 to 35 were female and drawn from the local urban community. More than 1800 wearers were asked to score various sensations of the garments, including prickle, on a scale of 1 to 9 (1, not detected; 2, just detected/threshold; 3, slightly detected; 5, moderately detected; 7, very detected; 9, extremely detected). The weighted mean prickle scores from all wearer trials were used to analyze the data, and linked garments were used to remove any bias or drift which might have occurred over time or between wearer trials with variations in wearer cohorts. ¹⁶

Wool ComfortMeter assessment

WCM assessment was undertaken using the draft test method. ¹⁸ Five samples (30 cm × 30 cm) were cut from each fabric. Samples were then hung vertically and the technical rear or back (the side generally worn next to the skin) of the fabric was lightly and evenly steamed using vertical movements of a Personal Hand Steamer. The fabrics were conditioned at 20℃ and 65% RH for 24 hours prior to testing, and then placed face down on the WCM platform and the back of the fabric measured. Each sample was subjected to 10 passes of the recording head and the mean WCM value recorded.

Incidents counted by the WCM measurement wire

The WCM was designed to identify and count the fibers protruding from the fabric surface that had a fiber length and diameter capable of triggering the prickle response in wearers.^7,12 The length and diameter of fibers capable of achieving a threshold buckling load of 0.74 mN, based on Euler’s buckling formula (Equation (2)) and a Young’s modulus of 3.5 GPa and 5.4 GPa are shown in Figure 1. A minimum of 0.2 mm free fiber length was used for the calculation. Below 0.2 mm length the deformation of the skin is thought to lead to a bottoming out of the fabric against the skin, and as more fabric touches the skin this reduces the pressure at the fiber end. ¹ The lines in Figure 1 show the response surface of wool fiber diameters and the length of fibers projecting above the yarn surface that are capable of triggering the response. The buckling load is exceeded by the fibers if their diameters and the length of their protruding fiber ends fall below the curve at a particular modulus value, whereas values above the line do not exceed the buckling threshold. The yarn surface fiber length effect is such that as the fiber diameter is reduced below 40 µm, progressively shorter fibers are required to exceed the buckling threshold. Figure 1 shows that fibers as fine as 10 µm can reach the threshold force to cause prickling if they have a short enough free length projecting above the yarn surface. This finding differs substantially from earlier reports, as discussed below. OPEN IN VIEWER

The assumptions in the model are that the Euler column extends at an angle of 90° from a solid, non-deformable surface. ¹⁹ In fabrics the three-dimensional structures and yarn compressional behavior provide a moving extendable base for fibers that may project towards the skin. Furthermore, since they have variable diameter and cross-sectional shapes the fibers do not resemble rods or columns. Yarns also have thin and thick areas caused by the distribution of fibers in the roving that subsequently cause variations in yarn twist, resulting in variations in the forces holding the fibers in position.

The results show that a higher Young’s modulus results in an upwards shift of the response curve (Figure 1). This means that for any given fiber diameter, increasing the Young’s modulus results in fibers of length corresponding to the distance between the two curves for higher buckling loads. The potential therefore exists for more fibers where the wool has a higher Young’s modulus to trigger a response, but whether this occurs in practice depends upon the distribution frequency of the lengths of the protruding fiber ends, and also whether there is a difference in Young’s modulus in commercial batches during yarn construction.

In earlier studies, either the coefficient of variation of fiber diameter (CVD) or a measurement of the frequency of fibers coarser than about 30 µm were regarded as important determinants of the prickle sensation.^20,21 In more recent research, however, using a much larger range of fabrics and constituent mean fiber diameters (MFD), once MFD was included in the prediction model and the prickle response of wearers was quantified in relation to MFD and other significant fiber, yarn and fabric factors, the various measurements of fiber diameter distribution became insignificant. This included CVD (P value 0.87), and measurements of the coarse edge of the fiber diameter distribution, such as the percentage of fibers >27 µm or >30 µm, none of which were significant, with P values ranging from 0.34 to 0.97. ¹⁰

If it is assumed that fibers shorter than 1.0 mm can interact with the skin when the garment is skin-tight and there is movement of fabric relative to the skin, then fibers finer than 20 µm are capable of triggering the prickle response (Figure 1). This might be the reason that fibers >27 µm in diameter were not significant in the model for wearer prickle assessment in earlier investigations using more than 1800 wearer assessments. ¹⁰ It would also explain the observations of Naylor at al., who found that to achieve comfort equivalent to knitted fabrics a woven fabric needed to be made from wool with 3 µm finer MFD. It was concluded that the length of the protruding fiber ends in woven fabrics was less than that in knitted fabrics, and that the stiffness and capacity to excite nerve endings was therefore significantly increased. ²² Thus, as a result of the higher yarn twist and the structure of woven fabrics than in knitwear, yarns are held more tightly and the fabrics become stiffer. Shorter and finer fibers in woven fabrics have less freedom to move away from the skin and they behave more like firmly anchored rods. The findings illustrated in Figure 1 thus have important applications for both knitted and woven woolen fabrics.

Estimating human sensory perceptions

The WCM incidence per 10 cm² (Ip) against the predicted human response (pSp) calculated using Equation (1) is shown in Figure 2. The relationship between the wearer prickle response (Sp) and predicted human response (pSp) derived from Equation (2) (Garnsworthy’s model) is shown in Figure 3. pSp was strongly correlated with the average prickle rating (Sp) assigned by wearers of the garments, with low prickle scores (Sp) associated with low pSp, as expected. An Ip between 10 and 17 would equate to 2.4 < pSp < 3.4, equivalent to wearer prickle score between 2.0 and 2.5 when some slight prickle was detected. The data suggested that a fabric with around 10 incidents per 10 cm² (Ip = 10, pSp = 2.4) would equate to a wearer prickle response of 2.0, which is equivalent to no or barely detectable prickle. OPEN IN VIEWER

OPEN IN VIEWER

Figure 3.

Relationship between the prickle scores of wearers (Sp) and the predicted human response (pSp) of fabrics (n = 48, r = 0.81). Dash lines indicate: wearer prickle score of 2.0, equivalent to no detection or barely detectable prickle; and 2.5, equivalent to some slight detection of prickle.

In this sample of knitwear made using a variety of yarn and fabric structures and different fiber types, there is a linear relationship between predicted human prickle sensations (pSp) and average prickle rating (Sp). Garnsworthy et al. applied a 0 to 10 scale to forearm tests, in which participants were given calibration fabrics scoring 0 and 10. In the more extensive wearer trials used in the present study a 1 to 9 scale was used, without the use of an equivalent calibration fabric but with clear definitions of the perceptions relating to each score. Unfortunately, no technical details of the calibration nor the test fabrics used by Garnsworthy et al. have been indicated. ¹ It appears highly unlikely, however, that the calibration fabric ‘0’ would actually produce a score of ‘1’ in the more extensive later trials,^10,11 as the lowest mean scores for cotton or ultrafine wool across the linked trials always exceeded 1.0 (see Figure 3), and there were always some sensitive individuals within the trial population who reported scores greater than 1 for these fabrics. Indeed, Garnsworthy et al. reported that the least prickly fabric of the 13 tested scored about 0.4, and they too found some more sensitive individuals, even within their much smaller sample of 55 participants. ¹

Figure 4 shows the observed Ip (WCM incidents/10 cm²) against the 1–9 scale score for the wearer trial under two different environmental testing situations. Superimposed on this is the Garnsworthy et al. line of best fit (Sp ∼ 0.54 Ip^0.66, r = 0.91). ¹ On Garnsworthy’s scale (Figure 9 in their paper), Ip values of approximately 8 or less equate to Sp scores of less than 2. The non-linearity observed in their predicted human response (pSp), based on Stevens’s psychophysical power law (Figure 4) is therefore probably related to the difficulty of scoring fabrics found to give no prickle sensation when the frequency of prickle incidents are below 8. This may reflect their practice of not providing a calibration fabric-rated zero prickle during the actual forearm tests. ¹ This finding may also indicate that any evaluation of fabrics needs to have a large enough population to ensure that sensitive individuals are included in the sample. It should be noted that the prickle scale used by Garnsworthy and co-authors to calibrate the fabrics is likely to be different from the scales used in the present study. OPEN IN VIEWER

Our analysis suggests that the correlation between the prickle scores in all our test environments was similar to the predicted incidence of prickle in fabrics (Table 2). However, the prickle responses in our hot environment more closely overlaid the predicted responses obtained with the forearm test, as modeled by the Garnsworthy et al. line of best fit (Figure 4 and Table 2). This suggests that the sensitivity of the forearm test when the fabric is gently moved under slight pressure ¹ is similar to the effect of walking in the hot environment in our Period 11. The regression coefficients for the prickle ratings (Sp) in the cool environment, hot environment and the weighted prickle score are summarized in Table 2. For each of these linear regressions there was no evidence of curvilinearity (P > 0.1). The regression constants for each of the regressions (that is the predicted Ip when the WCM incidents/10 cm² equalled zero) differed little from 1, which was the rating defined for no perceived prickle (Figure 4, Table 2).

OPEN IN VIEWER

Table 2.

Regression and correlation coefficients along with the standard error (se) and residual standard deviation (RSD) for the relationships between the observed wearer prickle response (Sp) and the Wool ComfortMeter incidence per 10 cm² in fabrics during the test protocol for different testing environments. Period 3 was a cool environment with limited movement but no walking. Period 11 was a hot environment with walking. Weighted prickle response used data from all test periods. All correlation coefficients are significant at P < 0.001

Response variate	Regression constant (se)	Dependent variate	Regression coefficient (se)	r	RSD	P
Sp Period 3	1.19 (0.111)	WCM	0.054 (0.0060)	0.793	0.356	<0.001
Sp Period 11	1.60 (0.182)	WCM	0.093 (0.0099)	0.807	0.582	<0.001
Sp Weighted prickle all periods	1.33 (0.128)	WCM	0.067 (0.0069)	0.814	0.407	<0.001

This evidence suggests that when the incidence of prickle is less than about 3 in a cool testing environment and less than about 5 in a hot testing environment the Wool ComfortMeter values may be slightly higher than the response predicted using the Garnsworthy et al. line of best fit, based on Stevens’s psychophysical power law, However, over the entire range of possible prickle incidents and wearer responses the Wool ComfortMeter values provided a reliable prediction of the average human response in all testing environments, even at a low incidence of prickle (Figure 4 and Table 2).