JOURNAL OF COSMETIC SCIENCE 180 Assuming a circular cross section of the fi ber sample, even for a fi ber sample with an el- liptical cross section, the diameter of swollen fi bers and any signifi cant decrease in ellip- ticity were measured under a microscope at approximately 20 places randomly selected over the 20-mm length of the fi ber. This procedure was performed at room temperature because the diameter of the swollen fi bers in the mixed solution was approximately con- stant irrespective of the temperature (17). The average cross-sectional area of swollen fi - bers was calculated from the average value of the swollen fi ber diameter, Ds. FORCE–EXTENSION CURVE FOR SWOLLEN HAIR FIBERS Mechanical tests for swollen fi bers were performed according to the method described in a previous paper (15). A swollen fi ber sample of ca. 20-mm length was set between sam- ple holders in a liquid cell. The force–extension properties were measured in a mixed solution of 8 M LiBr and BC in a 55:45 volume ratio at 50°C after the measurement of unstrained zero length at equilibrium, L0. Several load and unload extension cycles were performed to remove any contribution from fl ow segments, and then force–extension curves were measured until about 40% extension at an extension speed of 10%/min was attained. Finally, the extended fi ber was retracted until zero force. Next, the mixed solu- tion in the cell was replaced with water and held for about 5 min. The zero length in water at room temperature, Lw, was also measured in the unstrained state and the fi ber was released from the sample holders and dried in air in a P2O5 desiccator. Its diameter was then measured using a laser method (KL151A, Anritsu Co.) and the average diame- ter, Dd, was obtained. The volume fraction of keratin materials in the swollen fi ber sam- ple, v2, was calculated using v2 = (Dd/Ds)2(Lw/Lo). Here Lw was assumed to be equal to the dry length. Note that the stress–strain curves for swollen fi bers were constructed using the force values for the average cross-sectional area of swollen and unextended fi bers. FORCE–EXTENSION CURVE IN WATER The untreated and permanent-waved hair fi bers were immersed in water at 20°C over- night and then extended in water at 10%/min. The stress–strain curves were constructed on the basis of stresses as the forces per cross-sectional area measured in the unstrained state and under room conditions. The initial modulus in water, Ew, was defi ned as the slope of the initial straight region of the stress–strain curve. TWO-PHASE MODEL STRUCTURAL CONCEPT It has been demonstrated that α-helical chains can be randomized by immersing an NEM-treated keratin fi ber in a mixed aqueous solution of concentrated LiBr and BC (17). It has also been shown that the randomized chains can be subsequently recrystallized in water, and that these conformational changes are substantially reversible (19). This provides evidence that blocking the free SH groups inhibits SH/SS interchange reactions to stabilize

TGA-INDUCED STRUCTURAL CHANGES IN HAIR 181 the network structure without any change in the relative positions of SS cross-links dur- ing the swelling process. The extension behavior for swollen hair fi bers in the mixed solution of aqueous 8 M LiBr and BC is similar to that of rubbers and elastomers undergoing simple extension, showing (i) typical rubbery stress–extension curves, (ii) excellent elastic recovery, and (iii) lower energy losses and a signifi cant decrease in energy loss with increasing temperature. This characteristic mechanical behavior clearly indicates that the crystalline α-helical struc- ture in hair was changed to an amorphous network cross-linked with SS bonds. In fact, from the measurements of equilibrium force and temperature relationships in human hair at a higher extension range, the energy components in the retractive forces were analyzed, and it was concluded that there was essentially no energy component at all for such a swollen keratin system (17). It was also shown that the stress–strain curve in water for the deswollen hair fi ber, which has been relaxed by rinsing in water at 30°C for 24 h after swelling in the mixed solution of 8 M LiBr and BC, is very similar to that of untreated hair fi bers (17,19). Note that under these conditions, almost perfect reformation of the hair structure occurs in water from swollen aggregates composed of randomly deformed α-helical chains and cystine-rich globular matrix proteins. This suggests that the SS bonds between IF proteins play a role in the reformation of α-crystallites and that the SS bonds between globular matrix proteins (KAP) retain their relative positions not only on the surface but also within the globule. A two-phase model for the assembly of IF and KAP components was fi rst presented by Feughelman (20), who developed a zone model comprised of an uncross-linked X-zone and a covalent cross-linked Y-zone forming SS cross-links between IF and KAP mole- cules. A similar model was presented by Crewther (21), who proposed a hypothesis with no covalent cross-links between IF and KAP forming a network cross-linked with SS bonds between globular matrix proteins. However, at present, the detailed cross-linked structure of keratin fi bers remains uncertain. Figure 1 shows a two-phase model for (a) unswollen and (b) swollen keratin fi bers consist- ing of IF and KAP components. In the swollen fi ber, two different components change to Figure 1. Schematic representation of a two-phase model for (a) unswollen and (b) swollen hair keratin fi - bers. (a) Assembly of the intermediate fi lament (IF) proteins with eight tetramers and cystine-rich globular matrix proteins (KAP) (24). (b) Cross-linked structure model in the swollen state comprised of a densely SS cross-linked microdomain phase of KAP and a lightly cross-linked rubbery phase of IF (14). A reversible conformational change occurs alternately (b) in the mixed solution composed of 8 M LiBr and BC in a 55:45 volume ratio and (a) in water.