Quantitative and simultaneous non-invasive measurement of skin hydration and sebum levels.

We report a method on quantitative and simultaneous non-contact in-vivo hydration and sebum measurements of the skin using an infrared optical spectroscopic set-up. The method utilizes differential detection with three wavelengths 1720, 1750, and 1770 nm, corresponding to the lipid vibrational bands that lay "in between" the prominent water absorption bands. We have used an emulsifier containing hydro- and lipophilic components to mix water and sebum in various volume fractions which was applied to the skin to mimic different oily-dry skin conditions. We also measured the skin sebum and hydration values on the forehead under natural conditions and its variations to external stimuli. Good agreement was found between our experimental results and reference values measured using conventional biophysical methods such as Corneometer and Sebumeter.


Introduction
Skin hydration (moisture) and sebum (skin surface lipids) are considered to be important factors in skin health; a right balance between these components is an indication of healthy skin and plays a central role in protecting and preserving skin integrity [1]. Optimal balance between hydration and sebum levels provides the skin with a radiant, smooth texture and a natural pigmentation appearance, which is important from a cosmetic perspective. Hydration and sebum retaining ability of the skin is primarily related to the stratum corneum (SC). The SC plays the role of the barrier to water loss and is composed of the corneocytes and an intercellular lipid bilayer matrix. The water retaining property of the SC is dependent on these two major components. The presence of natural hygroscopic agents collectively referred to as natural moisturizing factor (NMF) and the SC intercellular lipids arrange orderly to form a barrier to prevent transepidermal water loss. Epithelium remains flexible when it contains 10-20% water, but becomes brittle, when it drops below 10% [2]. Stratum corneum receive hydration underneath skin layers [3] and to a lesser extent from the atmosphere [3]. Skin conditions such atopic dermatitis shows drop in skin hydration level reflecting in a drop of water holding capacity of the skin, increased transepidermal water loss (TEWL) and defect in barrier function [4][5][6]. The same symptoms are seen in individuals suffering from psoriasis, eczema and ichthyosis vulgaris [7,8]. This similarity in symptoms leads to complications with diagnostics, which often requires a biopsy, an invasive approach [9]. Nevertheless, these mentioned disorders show peculiar skin conditions with respect to the balance between hydration and oiliness. Eczema leads to minor water loss (few percent) combined with noticeable oiliness drop (~25%) [10,11], whereas psoriasis shows dramatic decrease of hydration (~70%) and oiliness (~40-70%) levels [12]. Ichthyosis vulgaris shows decrease of hydration level (~63%) while the level of superficial skin lipids does not vary significantly (~ ± 15%) [13,14].
Studies show that superficial lipids play an important role in the barrier function, creating a filter for interaction with the external environment. Skin health is associated with the stability of the functioning of the skin barrier, which depends on the continuity of the skin's superficial lipids structure [15,16]. Lipid phase behavior in the stratum corneum is considered to be crucial for the skin barrier function [17]. Skin superficial lipids have been found to serve as water modulator in the stratum corneum [18]. Thus, the water-sebum system determines the condition of the skin and can be used as an indicator of skin health.
Sebum is a mixture of fatty acids, triglycerides, proteins, and other molecules produced by the sebaceous glands in the dermis. Sebum keeps the skin smooth and flexible by sealing and preserving moisture in the corneal layer and preventing evaporation and bacterial infections. The sebum excretion rate (SER) reflects the amount of sebum production and is closely related to the physiological activities of the sebaceous glands. This is important information in the pathogenesis of sebaceous glands disorders and pimples and acne. Excessive sebum production can cause clogged pores possibly resulting in blemishes. Sufficient amount of skin hydration and sebum makes the skin appear smooth, soft and supple whereas the lack of moisture can cause the skin to look dull and cracked, appearing older [19]. The reduction in the efficiency of the barrier and moisture-maintaining functions of the skin results in easily dried, roughened skin which can be potentially more vulnerable to risk of infection [15].
The most well-established commercially available moisture detectors measure electrical properties such as capacitance and alternating current conductivity on the skin surface. Transepidermal water loss (TEWL) expressed in grams per square meter and per hour is used for studying the water barrier function of the human skin. However the method is very sensitive to environmental changes and requires several minutes to retrieve stable readings. The most commonly used skin hydration measuring devices are Skicon200, Corneometer CM820, Nova DPM 9003. These devices use rigid probes which must be in contact with the skin. Furthermore, the measurements are influenced by the amount of electrolytes, contact area, applied pressure and are sensitive to the external temperature and humidity. Furthermore these methods are not suitable for measuring changes in the hydration levels over time and to visualize the spatial distribution and heterogeneity of the skin moisturizing-ability of the whole face [20]. Near infrared multispectral imaging is an optical method that measures skin hydration based on the prominent water absorption peaks in the absorption spectrum. For methods which use shorter wavelengths, the absorption of water is very low while the scattering volume is high, resulting in a higher light scattering and influence of other skin chromophores on the measured hydration levels [21]. The measured values of hydration are influenced by the presence of other chromophores in methods using a single wavelength. In order to correct for the influence and artefacts arising from other chromophores, an analytic method based on the difference in absorbance of two NIR wavelength bands (1060 nm and 1450 nm) have been reported [22]. These methods use widely spaced wavelengths where the variation in wavelength dependent scattering effects also influences the measurement results.
In all optical methods reported above, the results are influenced by various factors such as wavelength dependent scattering effects, the presence of other chromophores and none of them are able to measure sebum and hydration simultaneously.
Presently available sebum measuring devices are based on grease-spot photometry and gravimetric analysis that are both tedious and time-consuming [23][24][25]. The most wellestablished commercially available devices measure optical properties such as skin gloss, and sebutape transparency. The gravimetric method provided more accurate results along with increased complexity of obtaining data [25]. The most commonly used skin surface lipids measuring devices are Sebumeter and Glossymeter. These devices need to be in contact with the skin and use non-disposable rigid probes. Furthermore, measurements can be influenced by contact area, applied pressure and time of applying. Furthermore, they are sensitive to skin pollution and can be sensitive to atmosphere humidity changes. Moreover, they are not suitable for measuring changes in the sebum levels over time and for visualizing the spatial distribution and heterogeneity of the skin oiliness of the whole face.
In short, in spite of many technological developments throughout the years, until now no non-contact devices and methods have been reported for the quantitative and simultaneous measurement of skin superficial lipids and water. Development of a non-contact method for measuring skin hydration and sebum simultaneously will enable to assess the balance between these factors related to skin health and to select the appropriate skin care treatment and products and will make it possible to monitor the progress during treatment.
In this manuscript, we report on quantitative and simultaneous non-contact in-vivo sebum and hydration measurements using a short wave infrared optical spectroscopic set-up using differential detection between three wavelengths 1720, 1750, 1770 nm. Initially, we measured the absorption properties of artificial sebum and water in the spectral range from 400 to 2000 nm and identified the spectral bands around 1720 nm corresponding to the lipid vibrational bands that lay "in between" the prominent water absorption bands. We built an experimental set-up that was employed to shine light at these three wavelengths to the skin and detect the light backscattered from the skin using a Ge photodiode. We have applied 20 µg/cm 2 of sebum-water mixtures in different volume fractions on the skin to mimic different oily-dry skin conditions. The estimated sebum and hydration levels were compared with conventional devices Corneometer CM825 (Courage & Khazaka electronic) and Sebumeter SM 815 (Courage & Khazaka electronic). Good agreement between experimental results and reference measurements were found.

Materials and methods
We measured absorption spectra of skin surface lipids (artificial sebum) and water using an integrating sphere spectrophotometer (PerkinElmer Lambda 900, 150 mm) and calculated the ratio of absorption coefficients ( Fig. 1) which shows a good agreement with the known spectrum of human adipose tissue [26,27]. Artificial sebum showed sufficiently high contrast and absorption peaks near 1210, 1728, 1760, 2306 and 2346 nm. This is in agreement with previous studies that report values of optimal wavelengths that potentially are able to target lipid rich tissue such as sebaceous glands and sub-cutaneous fat [26,27]. The spectral band for simultaneous sensing of hydration and sebum levels is optimally chosen to have high absorption coefficients of water and sebum and at the same time a large ratio of these absorption coefficients to obtain high contrast between the two chromophores. The spectral window around 1700 nm have high absolute values of the absorption coefficient and a high ratio of the absorption coefficient of sebum to the one of water and, simultaneously, a minimal influence of other skin chromophores such as melanin and blood. Fig. 1. Ratio of absorption coefficient of sebum to water measured in the shown spectral range between 800 to 2400 nm. Yellow bands represent the optimal spectral bands for simultaneous hydration and sebum sensing defined by high absorption coefficients of water and sebum and a large ratio of the absorption coefficients to obtain high contrast.
The experimental setup (Fig. 2) used for the skin hydration and oiliness level measurement comprises three quasi continuous laser sources, beam shaping optics and mirrors to guide the laser beam via the beam path. The laser sources (LD 1, LD 2, LD 3) were short wave infrared semiconductor lasers diodes emitting a wavelength of 1720 ± 4 nm, 1750 ± 5 nm, 1770 ± 20 nm emitting approximately 40 mW at each wavelength. The lasers are spatially combined along the same optical path using flipping mirrors (FM1, FM2). The beams are focused one by one through a central aperture in the mirror (M5) before it illuminates an area of approximately 12.6 mm 2 on the skin surface with a power of approximately 10 mW for each wavelength. This corresponds to 0.08 W/cm 2 , which is below the acceptable safety limit of 0.1 W/cm 2 in this spectral range. Light backscattered from the skin is collimated and reflected by the mirror (M5) and focused at the detector (PD) using a focusing lens (L10). The polarizers (P1 s and P2 p ) are set in crossed polarization configuration.
To mimic different natural dry and oily skin conditions, we applied mixtures of water and sebum in various volume fractions ranging from 0 to 100% on the skin of forearms of two healthy volunteers (skin type I-II). Water and sebum (wfk Testgewebe GmbH) were mixed using an emulsifier (Triton-X 100, 5%). The study was approved by the ethical committee and all volunteers gave written informed consent. Measurements were repeated five times for each volume fraction. Corneometer and Sebumeter was used for hydration and sebum reference measurements respectively. We calculated the volume fraction of sebum and water for all applied mixtures from the ratio of backscattered light to the incident light intensity for each wavelength. The amount of water (c w ) and lipids (c s ) were calculated from this ratio using an algorithm based on Beer-Lambert's law for light propagation in scattering media. The wavelengths 1720 nm and 1750 nm are used for estimating the sebum content and 1750 nm and 1770 nm for the water content.
( ) ( ) To measure the natural skin condition and its response to different stimuli, experiments were carried out on the forehead. In order to simulate different levels of oiliness and hydration of the conditions standard techniques were used [28,29]. High hydration level of the skin was reached by applying a wet wool fabric for 30 minutes. 70% iso-propanol was used for decreasing hydration and oiliness level of the skin. Natural levels of sebum on the T-zone for oily skin corresponds to 20 µg/cm 2 [29]. The condition dry skin with excessive oiliness was replicated by applying artificial sebum on the treated area. Measurements were performed using our experimental prototype device and with Corneometer and Sebumeter for reference measurements.

Results
The volume fraction of water and sebum measured for different sebum-water mixtures applied onto the skin are shown in Figs. 3 and 4 respectively. The vertical axis corresponds to the estimated amount of water (Fig. 3.) and sebum (Fig. 4)

Discussion
The experimental results of our non-invasive infrared spectroscopic method to simultaneously determine skin surface lipids and hydration volume fractions show good agreement with the commercial instruments Corneometer and Sebumeter. In our study, we have used wavelengths near 1720 nm because the absorption coefficient of sebum and water and the corresponding ratio of absorption coefficients are high in this spectral range. Also, the influence of other skin chromophores such as melanin and blood is expected to be lower in this spectral range compared to other shorter wavelengths. The lipid absorption bands around 1,210 nm and 2200 nm are also interesting based on the higher values of absorption contrast between sebum and water that varies significantly in narrow spectral band of 50 nm. Even though the other absorbers such as collagen [30], blood [31] and proteins [32] can also influence our experimental results in this spectral region around 1720 nm the absolute values of absorption coefficient and the amount of these chromophores present in the measurement volume of our set-up is relatively lower compared to water and sebum [11]. The maximum water content in the skin changes from 30% in stratum corneum to 70% in epidermis. The spectral band around 1700 nm is expected to be less sensitive for other chromophores when the sampling depth [33] of the optical set-up is chosen to be optimized for measuring skin barrier function, which in turn depends on the properties of stratum corneum. Nevertheless, the influence of protein absorption has to be considered in this spectral band around 1720 nm for the application focused on determination of stratum corneum hydration and oiliness. The ratio of sebum to proteins can be in the range of 1/3 to 2/3 for clinically healthy individual [34] and thus absorption losses can be primarily related to three chromophores: water, sebum and protein. Protein content is nearly constant for healthy individual, so it can be accounted as a baseline correction factor in the calculations. The strong absorption of lipids at the wavelengths near 1720 nm is caused by vibrational overtone modes, in particular from the C-H stretching [35][36][37]. Although artificial sebum has a chemical composition that differs from natural sebum, the absorption spectra of artificial and natural sebum shows absorption maxima at 1210 nm, 1728 nm, 2306 nm and 2346 nm [26,27]. In our calculations, we have used combination of the wavelengths 1720 nm and 1750 nm for estimating the sebum content and 1750 nm and 1770 nm for the water content. The same results can also be obtained by using the combination of λ s = 1720 nm, λ 0 = 1705 nm, λ w = 1694 nm. In general, for this application any set of sources with wavelengths corresponding to maximal (λ s ), minimal (λ w ) and equal (λ 0 ) ratio of absorption coefficient of sebum to water could be used. The accessibility of certain light sources was the final criterion in choosing the first set of wavelengths.
Even though our measurements show agreement with the reference measurements obtained with Corneometer and Sebumeter, direct comparison of our results with these commercial devices is difficult as the techniques sample different depths inside the skin. To investigate the measuring depth of our experimental set-up, experiments were performed on layers of sebum of various thicknesses applied to the skin. The experiments were performed by applying water-sebum emulsion (40% vol. of sebum & 60% vol. of water) in increasing layer of thickness from 0 to 1 mm in steps of 100 µm on a highly absorbing layer to avoid the possible long path length photons that may penetrate beyond the first layer. These experiments suggest that the imaging depth of our experimental set-up is approximately 350 µm and that the light backscattered from the epidermis also contributes to the measured values. Nevertheless the estimation of imaging depth can be influenced by the factors such as varying scattering properties of mixture, thickness of applied emulsion. The sampling volume can be optimized for various dermatological applications by changing the illuminationdetection geometry of our experimental set-up such as oblique incidence, source-detector separation [38].
The large error bars observed in our measurements are due to the non-homogeneity of the water-sebum mixture and variation in the thickness of the applied layer. Most detected light in this experimental configuration is coming from the layer applied on the skin and therefore can be highly dependent on the non-homogeneity of the mixture and layer thickness. Our approach to mix sebum and water using an emulsifier doesn't guarantee perfect uniformity and thickness of the applied layer. In addition to this, when sebum and water are mixed in various volume fractions, we observe significant differences in the scattering properties of the sample. This becomes prominent when the volume fractions of the individual components are comparable.
One of the potential advantages of our method is that it is insensitive to the presence and variation of other skin chromophores such as blood and melanin. Hence our optical method can be applied independent of skin type. Moreover, the probe does not need to be in contact with the skin so that the repeated measurements can be performed on the same location without changing the skin conditions. Further in-vivo studies were performed with this prototype device in a panel test to measure the inter-and intra-individual variations in skin sebum and hydration levels and its variations to external stimuli and to compare these results with standard devices. These results will be reported separately in the near future. Quantitative and simultaneous determination of these two biophysical parameters as demonstrated in this study will enable the clinicians to classify the skin types into Normal skin (N), Dry skin (D), Oily Skin (O), Oily-Hydrated skin (OH) and oily-dry skin (OD) and can provide personalized skin treatment solutions.

Conclusions
This study present a novel non-invasive short wave infrared spectroscopic technique for simultaneous measurement of oiliness and hydration levels of the skin. We have built an experimental set-up operating in the spectral region around 1720 nm, utilizing the lipid vibrational bands that lay "in between" the prominent water absorption bands. Sebum and water was mixed in various volume fractions using an emulsifier and was applied onto the skin to mimic different dry-oily skin conditions. The amount of sebum and water estimated from the backscattered light measurements from the skin using this short wave infrared spectroscopic set-up showed good agreement with the values measured with both the commercially available Sebumeter and Corneometer. Furthermore the method is, to a large extent, independent of the presence of other chromophores such as blood and melanin. The natural sebum and hydration levels measured on the T-zone of the forehead and its variations due to external stimuli as measured with our set-up are in good agreement with the reference measurements obtained using the above mentioned commercial devices. To summarize, the preliminary results demonstrate the feasibility of this novel noninvasive optical method for simultaneously measuring the hydration and sebum retaining ability of the skin.