2.1.

Hair Anatomy

The hair shaft is mainly composed of keratin (a protein). It is generated by the living portion of hair called the hair follicle, a complex miniorgan of the skin.¹² Figure 1 shows a schematic diagram of a hair follicle. The cuticle is the outermost layer of the hair shaft. It is translucent so allows light to penetrate into the inner cortex layer. The cortex is the bulk of the hair shaft, which contains longitudinally oriented cells.¹³ The medulla is a vertical column of horizontal cells that are filled with large air spaces, at the center of the cortex.¹³ There are two types of melanin found in the cortex of hair: eumelanin and pheomelanin. It is known that light-colored hair (e.g., red and blond) has predominantly pheomelanin whereas eumelanin is responsible for dark-colored hair (e.g., black and brown).¹⁴ Melanin can also be absent in gray hair. Melanin is responsible for the majority of light absorption, especially at shorter wavelengths (e.g., the UV region).¹⁵

Fig. 1

Schematic diagram of skin layers containing terminal and vellus hair follicle.

Vellus and terminal hairs have distinct morphologies. Vellus hair is short (less than 2 mm from the skin surface), fine ($\le 30\mu \mathrm{m}$ in diameter), and unpigmented or little pigmented.^16,17 As people experience puberty, some vellus hairs will develop into terminal hairs due to hormonal changes.¹⁶ Therefore, prepubertal children have a much higher proportion of vellus hair as compared to adults. In contrast to vellus hair, terminal hair is coarse and pigmented, with length and diameter longer than 2 mm and thicker than $60\mu \mathrm{m}$.^16,17 Unlike vellus hair consisting of only the cuticle and cortex, terminal hair also contains the medulla, which has high light scattering characteristics.¹⁸ Overall, the anatomy of hair suggests that UV radiation might penetrate in vellus hair more easily than terminal hair. Furthermore, the bulge in the outer root sheath of the hair follicle that houses the stem cells is located $\sim 1200\mu \mathrm{m}$ from the exposed skin surface in the terminal hair, which is more than three times deeper than the location of the bulge in the vellus hair, about $360\mu \mathrm{m}$,¹⁹ implying that the stem cells in the vellus hair follicles are less protected from sun exposure than terminal hair follicles.

2.2.

Human Hair and Melanoma

No significant difference was found in the UV sensitivity of children and adult’s skins, from the aspects of UV-induced skin reactions and anatomic structures.^20,21 However, a positive correlation has been found between the incidence of melanoma and the number of vellus hair follicles.⁸ As melanomas derive from melanocytes and the stem cells residing in the bulge region of hair follicles have the ability to differentiate into melanocytes,^22,23 vellus hairs have been hypothesized to transmit UV energy more efficiently than terminal hairs and to damage the stem cells in the hair follicles (due to their unique morphologies). The cumulatively damaged stem cells are proposed to have the potential to become melanocytes and melanomas in later life. This hypothesis is further supported by Fang et al. who found that melanomas contain a stem cell population.²⁴

As the first step to proving the hypothesis, the optical properties of human hair (vellus and terminal hair) need to be well understood. More specifically, the overall optical properties of vellus hair are thought to be responsible for deeper penetration depth of UV radiation than in terminal hair.

4.1.

Sample Preparation

To ensure the hair samples are not attached to contaminants (e.g., dirt and oil), which could influence the measured results, samples are cleaned with ethanol or shampoo before the measurement.^40,41 For microspectrophotometry of hair, the samples are placed between a microscope slide and a cover slip^43,44 or mounted on a triholed metal slide.⁴⁰ Some sample holders can also be used to mount the hair and adjust its position and orientation.^41,45 The refractive index of human hair is about 1.54, which is higher than that of air.⁴⁶ So some researchers used the refractive index-matching liquid to minimize the light loss from the surface scattering in measurement,^43–45 whereas others included the mismatch of refractive index on the measurement results.^40–42 When UV light is involved in the measurement, a quartz microscope slide needs to be used instead of a glass one as glass absorbs UV and could mask the spectral result. It is noted that the refractive index-matching liquid (mounting media) should not degrade the hair sample nor absorb UV.⁴⁰

The optical properties in the transversal (cross sectional) direction of hair were studied by most of the groups because of the relatively simple sample preparation procedures.^40–45 In contrast, the properties in the longitudinal direction of hair were rarely examined due to the difficulties of achieving a flat cross section of hair and positioning it accurately for illumination. However, one group achieved this by mounting the properly cleaved hair to a thin metal wire, wrapped with double-sided tape.⁴¹ Although there are several techniques for preparing the hair samples with neat cross sections through several steps, involving the use of mounting media (e.g., paraffin and resin) and microtome sectioning, they are technically more difficult and time consuming to implement.^47–52 In one case, it was found that hairs with different colors sectioned from a microtome were too short to give discrimination in the spectrophotometry readings.⁴⁴ Also, when paraffin is used as a mounting medium since the keratin of hair is harder than paraffin, the hair could split, break, or move within the block.⁴⁸

4.2.

Measurement Techniques

The optical properties of biological tissues are commonly measured using an integrating sphere.^31,53 However, human hair has never been studied using this method due to the difficulty of accurately focusing light onto hair inside an integrating sphere. Nevertheless, the optical characteristics of some relatively larger mammals’ hair have been investigated using these techniques.^54,55 Instead, human hair was primarily studied using a common confocal setup.^18,35–45

A confocal setup of collimated transmittance measurement, as shown in Fig. 2, was constructed by Kharin et al. ⁴⁵ to examine the attenuation coefficient of human scalp hair. Five laser sources with different wavelengths ranging from 409 to 1064 nm were used to examine the wavelength dependence of the optical properties of different hairs. In addition, for the first time in the literature, a multiorder wave plate was used in this study to rotate the light polarization for studying the sensitivity of the optical properties on polarization of light.

Fig. 2

Schematic diagram of the confocal setup used for the measurement of collimated transmittance through a hair: $\lambda /2$, a multiorder half-lambda waveplate; $F1$, a $20\times 0.4$ NA microscope objective; $F2$, a 40-mm-focal-distance lens; $F3$, a 100-mm-focal-distance lens. Reproduced figure with permission.⁴⁵

Constructing a confocal setup for optical measurements of such a small object could be costly and time consuming, if starting from scratch. Alternatively, given the advantages of accurately focusing light onto hair using an optical microscope, a video camera and a light source that are attached to an optical microscope has been commonly used in the past to study the optical properties of human hair.^40,42–44 The Craic microspectrophotometer and Jasco spectrophotometer are two modern instruments in the market that can be used to study the optical properties of microscale materials.^56,57 They are powerful but expensive devices that combine the techniques of microscopy and spectrophotometry, providing measurement of the UV–visible–NIR range of transmission, absorbance, and reflectance for sampling areas as small as $10\mu \mathrm{m}\times 10\mu \mathrm{m}$.⁵⁷ In the microspectrophotometry studies of hair, the incident light was focused down to a size smaller than the diameter of human hair to obtain the absorbance or transmission spectra of hair samples.^40,43,44 A dark scan and a reference scan were required during measurement to obtain accurate spectra graphs. At a given wavelength, transmittance, $T$, for example, is calculated using Eq. (1), where $I_D$ is the dark reading measured by a detector when the light source is turned off while $I_R$ is the reference reading when the hair sample is absent in the light path; $I_{\text{hair}}$ is the reading when the hair sample is in place. To reduce the uncertainties of the result on each hair, $I_{\text{hair}}$ was determined by averaging multiple measurements along the length of the hair shaft^40,43,45

From the transmittance measurement, the attenuation coefficients of hairs were estimated using Beer’s law, as shown in Eq. (2), where $L$ is the thickness of the material.^43,45 The diameters of the hair samples have been measured conveniently using a video camera that is fitted to the microscope^42,43,45

However, by assuming the measured transmitted light is collimated, the attenuation coefficients were underestimated. Another limitation of only conducting the transmittance measurement is the difficulty of distinguishing the absorption and scattering coefficients from the attenuation coefficient on its own. Rather than calculating the attenuation coefficient from Beer’s law, Bashkatov et al.⁴² took a similar approach using a digital video microscopic system. The images of hairs in the reflectance and transmittance modes of a microscope were acquired and analyzed. From the color images of hair shafts, their absorption and reduced scattering coefficients were estimated using the inverse MC method. Readings from white and black papers were used as a reference and a background signal in the reflectance measurement, while the readings from a transparent glass were used as a reference in the transmittance measurement.

The refractive index of scalp hair was determined by increasing the refractive index of the hairs’ mounting media in small intervals until they matched, in which case the hair became invisible under a transmittance microscope.^46,58 Wang et al.⁴¹ used the optical low-coherence reflectometry (OLCR) technique to investigate the refractive index of human scalp hair at a wavelength of 850 nm. Different single hair shafts were scanned both longitudinally and transversely. The refractive index of hair was calculated as a ratio of the scan length of the reference mirror in the OLCR result and the physical length of the hair shaft. Although the attenuation coefficients were not reported in this study, they were estimated by Kharin et al.⁴⁵ using Beer’s law based on the magnitude of the interference signals in OCLR graph.

4.3.

Hair Optical Characteristics

Wells observed that gray hair acts as a natural optical fiber, transmitting light along the hair shaft.¹⁰ The transmittance microscope image of hair in his work illustrated that the cortex of the hair behaved like the core of an optical fiber, allowing the passage of light. In contrast to the cortex, the air-filled medulla at the center of the shaft appeared dark, implying no light transmitted through this region. Wells also pointed out that the color of hair is the most important factor affecting the light transmission along the shaft: darkly pigmented brown hair transmitted little or no light. Although no corresponding optical properties of hair were reported, these findings were confirmed by other later studies that presented numerical values.^41–43,45 These studies show that highly pigmented brown and black hairs have, as expected, the highest absorption and attenuation coefficients, whereas little- or unpigmented blond and gray hairs have the least (and insignificant) values. Among these studies, Kharin et al.⁴⁵ demonstrated that the attenuation coefficient of the medulla was dramatically higher than that of the cortex, especially in blond and gray hairs. This explains why the cortex of the gray hair appeared extremely bright while the medulla appeared completely dark in Wells’ study. In contrast with other measurement and analysis methods, Bashkatov et al.⁴² discriminated between the absorption and reduced scattering coefficients.⁴² However, lightly pigmented blond and gray hairs showed significantly higher reduced scattering coefficient, which is inconsistent with Wang’s results.⁴¹ This great inconsistency of results may be due to the fact that hair samples with and without medulla were studied by the two research groups, respectively. Apart from the dependence of the optical properties of hair on its color and the presence of a medulla, the attenuation, absorption, and reduced scattering coefficients all increase with a decrease in the wavelength.^42,45 In addition, the difference between the optical properties in highly and lightly pigmented hair increases with a decrease in wavelength.^42,45 These relations can be seen from Kharin et al.’s results in Fig. 3.⁴⁵ This observed result is primarily due to the fact that the absorption and scattering coefficients of melanin become progressively more significant as the wavelength decreases.^59,60 Therefore, it is inferred that the difference between the optical properties (absorption and scattering coefficients) of highly and unpigmented hairs (e.g., terminal and vellus hairs) will peak in the UV wavelength region. This means that UV rays are expected to penetrate deeper in vellus hair than terminal hair.

Fig. 3

Attenuation coefficient of human hair with different colors. Data are obtained from Kharin et al.,⁴⁵ Lin et al.,⁴³ and Wang et al.⁴¹ The properties of the black and blond hair samples are measured in the longitudinal direction while the rest of the data are all measured in the transverse direction.

The attenuation coefficient is the most commonly reported property of human hair in the literature because it can be estimated conveniently by applying Beer’s law. Figure 3 illustrates a summary of the attenuation coefficient of different colored hairs at different wavelengths, obtained from three research groups.^41,43,45 Table 1 summarizes the varied experimental factors (different hair samples, experimental methods, directions of measurement, and refractive index-matching conditions) in these studies, which may result in the difference seen in their results. The attenuation coefficient of black hair at the wavelength of 700 nm determined by Lin et al.⁴³ is inconsistent with the wavelength-dependency of the black hair observed in Kharin et al.’s work.⁴⁵ This inconsistency is probably attributed to the presence of medulla since the high scattering effect of medulla can contribute to overall attenuation coefficient of hair. In addition, different morphologies of scalp and body hair samples may result in different optical properties. In contrast, the attenuation coefficient of black hair reported by Wang et al.⁴¹ again differs to the other two studies. One fact is that Wang et al.⁴¹ performed the measurement in air, in which the surface scattering effect is taken into the total attenuation coefficient. Furthermore, the black and blond hairs were measured in the longitudinal direction by Wang et al.,⁴¹ whereas the rest were measured in the transverse direction. The attenuation coefficient of red hair measured in the transverse direction is surprisingly much higher than that of black hair at the same wavelength, both determined by Wang et al.⁴¹ It is too early to draw a solid conclusion that the optical properties are different in the transverse and longitudinal directions since other variables such as measurement techniques and hair samples may all affect the published results.

Table 1

Summary of the experimental factors in the studies of the attenuation coefficient of human hair.41,43,45 The details of the hair samples used by Wang et al. are not reported.

The lowest attenuation coefficient is expected to be observed in the cortex of the blond hair, because it mainly contains pheomelanin. The attenuation coefficient of the cortex of hair in Kharin et al.’s study shows an exponential relation with the wavelength. Based on the current knowledge, the cortex of the blond scalp hair is considered to be mostly close to that of vellus hair. Therefore, we extrapolate the attenuation coefficient of the cortex of blond hair at a wavelength of 350 nm to be $26{\mathrm{cm}}^{-1}$. By assuming the absorption to be dominant to the scattering in the cortex of blond hair, we estimated its penetration depth (at 350 nm wavelength) to be $376\mu \mathrm{m}$, using Beer’s law, which is about six times deeper than that of fair Caucasian skin,⁶¹ implying that UV may penetrate much deeper in vellus hair than skin. The contribution of vellus hair in the UV energy delivery is to be verified by determining the UV optical properties of vellus hair.

As presented by Barrett et al.,⁴⁰ the absorbance spectra (200 to 700 nm) of scalp hairs with different colors peaked in a wavelength range of 300 to 400 nm. However, no numerical values of the optical properties of hairs were reported and the absorbance spectra graphs have no scale for comparison. In another spectrophotometry study of red and brown hairs, the absorbance spectra (400 to 800 nm) of hairs from red-haired people differed noticeably to those from brown-haired people, but both peaked at a wavelength of about 440 nm.⁴⁴ In contrast, the absorbance spectra of melanin isolated from red and black hairs continuously decreased in the wavelength range of 270 to 700 nm; while a more marked decrease was observed in red hair.⁶² Greenwell et al.⁴⁶ found that regardless of age, gender, and race, the refractive index of the cuticles of hairs varied slightly from 1.543 to 1.554. As noted in this study, females and children have similar values of the refractive index, which tend to be higher than that of males. In contrast, Wang et al.⁴¹ showed that the refractive indices of the hair cortical regions were in a range of 1.56 to 1.59.

Studies in the literature so far have mainly looked at the influence of color on the optical properties of scalp hair in the visible and infrared range of wavelength. In addition, the hair examined was primarily adult terminal hair, and very little investigation was done on children’s hair and none on vellus hair. All in all, current knowledge of the optical properties of hair cannot provide enough evidence to support our hypothesis.

4.4.

Statistical Analysis

The anatomy of human hair varies with color, age, gender, race, and body site. Achieving statistically significant optical properties requires a substantial amount of work. Statistically significant results of the refractive index of human scalp hairs (2529 hairs from 97 individuals) were obtained by Grenwell et al., taking into consideration different races, sexes, and ages.⁴⁴ The dispersion of the result was about 1%. Similarly, a large number of samples (1080 hairs from 54 brown-haired individuals, 420 hairs from 21 red-haired individuals) were used in Nicholls’ study to determine their absorbance spectra. Wang et al. also investigated the refractive index of the cortical region (cortex and medulla) in the longitudinal direction of scalp hairs.⁴¹ However, the identities of the hairs and the number of samples were not reported. Kharin et al. used five hair samples from single individuals with different hair color (black, light brown, gray, and blond).⁴⁵ In this study, the age and gender of the individuals were not specified. Over 30 hair samples each from 10 individuals were used in Bashkatov et al.’s study.⁴² The age, gender, and hair color were specified in the paper. Barrett et al. also used five natural scalp hair samples each from seven groups of color (each group consisting of 1 to 5 individuals) to study the variation in their absorbance spectra due to the different levels of brown and blond colors.⁴⁰ However, no other information about the characteristics of the hair samples was given. In Lin et al.’s study, about five body hair samples each from about 45 individuals were used to determine the attenuation coefficient of human body hair.⁴³ This study calculated the mean attenuation coefficient of the hairs with different colors (auburn, brown, and black), from different gender, body sites (back, forearm, thigh, lower leg, shoulder, etc.), and ages (21 to 61). However, due to the randomness of the samples measured and multiple colors of hair samples involved, the statistical dispersion is more significant than scalp hair, as it can be seen from Table 2. Table 2 presents a summary of the optical properties of black hair and their standard deviation, for wavelengths from 600 to 700 nm.

Table 2

Summary of the optical properties of black hair in the transverse direction and their standard derivation.42,43,45

In the literature, the influences of gender, age, race and, more importantly, hair type on the optical properties of hair have not been investigated deeply. Since there are vellus and terminal hairs existing at different sections of human bodies and their morphologies vary greatly, if human body hairs are to be examined, the color, type, and site of hair are the main categories to be considered in any future statistical analysis.