ANALYSIS OF DISTRIBUTION OF WATER IN HAIR 39 mode. This difference in the water uptake between desorption and sorption modes, known as sorption hysteresis, is characteristic of hydrophilic materials. The magnitude of this hysteresis depends on the detailed molecular structure and hydrophilicity. Water uptake measurements provide information about the kinetics of water absorption, but not about distribution of water within the different morphological features of hair. This information can be derived from small-angle neutron scattering (SANS) obtained with wa- ter replaced by its equivalent deuterium oxide (D2O) to achieve the scattering contrast. The scattered intensity at small scattering angles (2θ) contains information about large length- scale (10–100 nm, mesoscale) structures. Although structural information at these length scales can be obtained using X-rays (small-angle X-ray scattering, SAXS), neutrons provide unambiguous interpretation of the features that are accessible to water. This is because neu- trons are scattered differently by hydrogen and deuterium. By replacing hydrogen with deuterium on polymer segments or host molecules, without altering the structure of the polymer or the interactions between the polymer and the host molecules, the distribution of these “colored” molecules can be imaged at large length scales (6). SANS measurements are mostly carried out with polymers that are specifi cally synthesized by replacing some of the hydrogen atoms with deuterium, a demanding task. However investigation of the dif- fusion of deuterated solvents into the polymer matrix is relatively simple. In this instance the contrast is generated by the distribution of the commercially available deuterated sol- vents, in our case deuterium oxide or heavy water, into the nondeuterated polymer. Note that with X-rays the scattering contrast is due to differences in electron density. Therefore, in a complex structure such as hair, changes in the scattering intensities in SAXS are due to both the electron density changes resulting from the presence of water and the changes in the structural features of the matrix itself. In SANS, it is possible to assign the features in the scattering pattern solely to structures associated with water (D2O). Such work has been done on synthetic polyamides and starch (7,8). The data in Fig. 3, for instance, show the four broad diffraction spots generated by the presence of D2O in the interlamellar spaces of a hydrated nylon fi ber (9). A second feature in the SANS pattern, the diamond-shaped equa- torial streak at the center of the pattern, is attributed to the water present in the longitudi- nal channels as well as from the surfaces of the fi ber (10,11). We here carry out SANS Figure 2. Sorption–desorption isotherm obtained using a dynamic vapor sorption analyzer (Surface Mea- surements Systems, Alperton, Middlesex, U.K.). Diamond symbols (bottom curve) corresponds to absorption and the square symbols (top curve) to desorption.

JOURNAL OF COSMETIC SCIENCE 40 measurements on several carefully prepared hair samples to see if similar distribution of water exists in hair. Infrared (IR) spectroscopy data are used to investigate the hydrogen– deuterium (H–D) exchange that occurs during saturation with D2O. EXPERIMENTAL European dark brown hair (EDBH) obtained from International Hair Importers (Glendale, NY) was used in these experiments. Samples for SANS measurements were prepared by fi rst removing the water in the hair by keeping them under vacuum for 24 h, and then equilibrating them for several days in D2O atmosphere of desired humidity obtained us- ing saturated salt solutions. A second set of samples were prepared by equilibrating them fi rst with D2O at the desired humidity for 24 h and then keeping the sample under vac- uum at room temperature to remove free D2O from the sample. For experiments designed to study the effect of oil, hair soaked with oil was prepared by treating the two tresses with oil (0.5 ml/g hair) followed by combing to spread the oil uniformly. One of the tresses was heat-treated for 90 s at 180°C. Both tresses were left overnight and then im- mersed in hexane for 30 s to remove surface oil. They were then equilibrated to the re- quired humidity with solutions of salt and D2O. Cuticle-free (CF) hair samples were also prepared using the CF hair supplied as special samples by International Hair Importers. SANS data were collected on the CG-2 and CG-3 beam lines at the Oak Ridge National Laboratory (Oak Ridge, TN). The q-range was between 0.05 and 0.025 nm−1. The data were normalized to tress thickness. Hair aligned in the form of a thin tress (~2–3 mm in thickness and ~15 mm in width) was mounted on U-frames made of aluminum plate using superglue. These plates were placed in airtight cells to maintain the samples at the required humidity. Different humidity levels were maintained using saturated solutions of appropriate salts in water and D2O. In the case of D2O the RH is expressed as %RHD. Figure 3. (A) Schematic illustration of the distribution of water in a fi ber with crystalline and amorphous domains and (B) the resulting SANS pattern. Water in the interlamellar spaces gives rise to the meridional (fi ber-axis) and off-meridional scattering. Water in the longitudinal channels contributes to the scattering along the equator (perpendicular to the fi ber-axis).