ORIGINAL ARTICLE Understanding breakage in curly hair G.A. Camacho-Bragado, 1 G. Balooch, 2 F. Dixon-Parks, 1 C. Porter 1 and H. Bryant 2 1 The L Oreal Institute for Ethnic Hair and Skin Research, Chicago, IL, U.S.A. 2 L Oreal Research and Innovation, Clark, NJ, U.S.A Summary BJD British Journal of Dermatology Correspondence G. Alejandra Camacho-Bragado. E-mail: acamacho@rd.us.lorea.com Accepted for publication 11 June 2014 Funding sources This study was funded by L Oreal. Conflicts of interest All the authors are employees of L Oreal. DOI 10.1111/bjd.13241 Background In 2005, the L Oreal Institute for hair and skin research carried out a multiethnic study to investigate hair breakage in women residing in the U.S.A. In this study it was reported that a large percentage (96%) of the African-American respondents experience breakage. A combination of structural differences and grooming-induced stresses seem to contribute to the higher breakage incidence in the African-American group as the chemical composition of African-American hair is not significantly different from other ethnic groups. Some authors have proposed that the repeated elongation, torsion and flexion actions may affect the components of the hair fibre. However, considering the different properties of cuticle and cortex, one would expect a different wearing mechanism of each, leading to the ultimate failure of hair. Knowing in detail how each part of the structure fails can potentially lead to better ways to protect the hair from physical insults. Objective To investigate crack propagation and fracture mechanisms in African- American hair. Methods Virgin hair of excellent quality was collected, with informed consent, from a female African-American volunteer. A series of controlled mechanical stresses was applied to 10-mm hair sections using a high-resolution mechanical stage (20 mn) up to the fracture of the fibre. The surface was monitored using scanning electron microscopy imaging during the stress application. X-ray tomographic microscopy images were acquired and quantified to detect changes in energy absorption as a function of applied stress that could be linked to increase in crack density. Results Analysis of the mechanical response of hair combined with the two imaging techniques led us to propose the following mechanism of hair breakage: cuticle sliding; failure of the cuticle cortex interface; nucleation of intercellular cracks and growth of cracks at the cuticle cortex junction; and propagation of intercellular cracks towards the surface of the hair and final breakage when these cracks merge at the cuticular junction. Conclusions The combination of scanning electron microscopy and X-ray tomography provided new information about the fracture of hair. Mechanical damage from grooming and some environmental factors accumulate in hair creating internal cracks that eventually result in breakage at unpredictable sites and therefore a continuous care regimen for the hair throughout the life cycle of the fibres is recommended. 10
Breakage in curly hair, G.A. Camacho-Bragado et al. 11 Hair breakage has been identified as one of the main hair problems in women of African descent in the U.S.A. as indicated by an internet survey performed in 2005. 1 In this survey, it was found that 96% of the surveyed participants experience breakage and 23% indicated this condition as their main concern. Multiple authors have reported on the potential causes of increased fragility of curly hair vs. other ethnic types. In some cases, genetic hair shaft abnormalities can be the origin; however, this is not necessarily a cause exclusive to curly hair. 2 In most cases, an acceleration of hair degradation is linked to a combination of structural characteristics of curly hair 3 and grooming practices, 4 7 therefore highlighting the importance of selecting an appropriate routine that can help extend the life of the fibres keeping the structural integrity of the hair. In spite of all this information, there has not been a report on the structural mechanism and process of hair breakage from a materials science approach; this article discusses a combination of techniques that were applied to study the failure mechanism of virgin (nonrelaxed) hair. The hair was considered to be a reinforced, highly hierarchical natural fibre with a remarkable tensile toughness (130 180 MJ m 3 ). 8 The method is a combination of in situ tensile testing, scanning electron microscopy (SEM) and X-ray tomography (XTM). The scale and resolution of these techniques are appropriate for studying the interaction of the different mesoscopic components of hair and help describe their individual fracture steps leading to the catastrophic failure of the structure. Human hair is a biological material formed, at the nanoscale, by keratin a-helices coiled into rope-like structures, which are bound by keratin-associated proteins (KAPs) to form microfibrils. Microfibril bundles, held together by a proteinaceous matrix, constitute the macrofibrils. On the microscopic scale, the macrofibrils aggregate into cortical cells kept together by the cellular membrane complex (CMC) to form the cortex. The cortex is protected from direct interaction with the environment by the cuticle, which consists of 5 10 layers of keratinized cells in a roof-tile-like arrangement. 9 Several groups have attempted to elucidate the mechanical behaviour of hair based exclusively on its molecular or nanoscopic components (keratin coils, microfibrils and amorphous protein matrix). 10 Other authors 3,11 use the surface quality and macroscopic geometry (twists, kinks, etc.) of hair fibres as key parameters in describing the failure of hair under different stress conditions (uniaxial tensile testing, cyclic fatigue, flexabrasion, etc.). 3,10 13 However, the series of phenomena leading to failure cannot be fully established from the properties of the molecular elements or the presence of surface defects alone. In the present study it was observed, for example, that fibres tested in tension (Fig. 1a, b) showed a step fracture similar to that previously reported. 3 However, fatigued fibres resulted in bevelled fracture surfaces (Fig. 1c). (d) (c) Fig 1. Postmortem (after fracture) scanning electron microscopy images of hair subjected to tensile stress showing stepwise fractures at a twist and within a homogeneous region. (c) Postmortem micrograph of a fatigued fibre revealing a bevelled surface fracture, a different fracture mode compared with tensile stress, in a region away from a twist (homogeneous region). (d) Fractured fibres collected from panellists that self-identified as having breakage issues.
12 Breakage in curly hair, G.A. Camacho-Bragado et al. Moreover, the presence of twists and kinks did not seem necessarily to increase the frequency of fracture occurrence at these locations, as the observation of 20 postmortem (after breakage) specimens revealed only four fractures associated with a twist or constriction such as the one in Figure 1a. A similar distribution of at-the-twist and away-from-the-twist fracture surfaces was observed in hairs collected from panellists who complained of hair breakage (Fig. 1d). These findings led us to believe that some internal crack propagation must occur and that stress may be concentrated internally and not only at the obvious macroscopic constraints and heterogeneities. Scanning electron microscopy and XTM (Fig. 2a) were used for multimodal imaging of the surface and the interior of the fibres. These techniques combined with in situ mechanical testing provided a better understanding of the evolution of internal stress distribution upon application of an external tensile force and the ultimate hair fracture. The combination of SEM and XTM provides a unique way to analyse the response of hair to stress and the changes induced by the strain applied. SEM offered a high-resolution view of the outer surface of the fibre, while synchrotron XTM allowed the visualization of the interior of the hair, particularly the cuticle cortex interface both in a three-dimensional (3D) view and in virtual cross-sections (Fig. 2d). It must be mentioned, that resolving the cuticle cortex interface by XTM requires special conditions not achievable with bench-top systems. Synchrotron light is the most viable solution for performing hard XTM of materials such as hair. In addition, the monochromaticity of synchrotron light enables accurate quantification of X-ray absorption coefficients, which can be linked to localized protein and lipid content. Methods Virgin hair of excellent quality as indicated by amino acid analysis (cysteic acid = 02, lanthionine = 01 and tyrosine = 20 mg amino acid per 100 g total amino acids) was collected from a female African-American donor who had given written consent. The hair was washed with a 10% ammonium lauryl sulfate solution and the fibres attached to brass ferrules (n = 9). Tensile test A Deben microtensile tester, customized by Gatan Inc. (Warrendale, PA, U.S.A.), was used to strain 10-mm sections of hair at 1 mm min 1 with a 2 N load cell (20 mn accuracy). The strain was applied to each fibre in four stages. Firstly, the fibres were strained up to the onset of plastic deformation; the second strain stage was performed to a point within the plastic region approximately before post-yielding; the third stage was stopped within the post-yielding region (Fig. 2b); the fibres were then strained a final time up to fracture. The hair fibres were imaged during (SEM) and after (SEM and XTM) each stage of the stepwise tensile test using an FEI QuantaTM 400 FEG Environmental scanning electron microscope (FEI Company, Hillsboro, OR, U.S.A.). X-ray tomograms of the entire fibre after each strain stage were also recorded. The final curve for the stepwise test was constructed (c) (d) Fig 2. Schematic showing the combination of scanning electron microscopy (SEM) and X-ray tomography (XTM) to capture surface and internal information of the specimen under tensile stress. Stress strain curve indicating the stages chosen for the stepwise tensile test. Stage 1, at the onset of plastic deformation; stage 2, within the plastic zone; stage 3, within the post-yielding region; stage 4, fracture. (c) Stress strain graph showing the elastic (Ee) and plastic (Ep) strain portions of the curve. (d) Schematic showing typical XTM virtual cross-sections of hair and the corresponding location within the length of the hair fibre.
Breakage in curly hair, G.A. Camacho-Bragado et al. 13 by shifting the partial curves an amount equivalent to the plastic deformation, so the elastic strain was accounted for one time only. Figure 2c shows how the elastic and plastic portions of a partial curve were defined. Synchrotron X-ray tomography X-ray tomography imaging was performed at the Advanced Light Source on Beamline (8-3-2) at the Lawrence Berkeley National Laboratory (Berkeley, CA, U.S.A.). Monochromatic radiation was used to obtain 2D projections that represent X-ray attenuation maps. These maps were used to reconstruct the 3D data volume at a resolution of about 2 lm; the intensity of the signal can be correlated with the local density in the specimen. Enough slices were captured to reconstruct the entire hair length. A detailed description of synchrotron XTM can be found in the literature. 14 Results The SEM images taken during the different stages of the stepwise tensile test indicate that the applied stress led to increasingly higher cuticle lifting (endocuticle failure). However, upon stress release, the cuticle closely returned to its initial configuration. Thus, the surface of the fibre revealed little evidence of the overall mechanical history of the hair. Figure 3 shows sequences of SEM images taken at two locations (close to the root and close to the tip) after different amounts of applied stress. The proximal end (close to root) is expected to have about 1 month less of grooming history than the distal portion. The cuticle lifting became permanent after the final stress application; the effect is gradually more prominent towards the position of the final fracture as shown in the postmortem (after fracture) image (Fig. 3b). On the other hand, XTM 3D reconstructions showed voids developing along the cuticle cortex junction after the second application of stress (Fig. 4). Interestingly, the voids were more noticeable after stage 2 than at the unstrained stage but did not show significant change after a third stress cycle. This contrasts with the observed decrease in attenuation coefficient between the same three stages. The attenuation coefficient is related to the density of the material under analysis as it is a measure of how much of the incident beam is scattered or absorbed by the specimen. In this particular case, cracks and voids would decrease the attenuation coefficient as the empty spaces absorb less radiation. Quantitative analysis of the stress attenuation coefficient (Fig. 5a) showed a steady reduction in relative energy absorption with increased stress, especially after stages 2 and 3, where the decrease is statistically significant (P <005). Fig 3. Sequences of scanning electron microscopy images taken after stages 2 and 3 of the stress strain test compared with postmortem (after fracture). Proximal section. Distal section and location of failure; the defect enclosed in the square was used as a fiducial to backtrack the area of failure, the exact location of failure is marked by black arrows.
14 Breakage in curly hair, G.A. Camacho-Bragado et al. Fig 4. Tomographic reconstructions of a hair fibre after different stress application stages. X-ray tomography (XTM) three-dimensional reconstructions showing multiple internal cracks and voids (white arrows) formed following stage 2 as a result of applied stresses. (c) (d) Fig 5. X-ray tomography quantitative analysis and stress strain curves. Relative energy absorption as a function of stress stages (statistically significant differences are marked with stars, error bars correspond to standard deviation). Increase in low-density voids causes a decrease in the energy absorbed by the specimen. Cross-sectional area measurements at five regions along the length of the fibre after each of three stress cycles (the area values for the location of failure are marked by the arrow); these show a formation of a neck at the site of final fracture. (c) Stress vs. strain curves, the contrast between the maximum value reached in the green vs. the other curves indicates the material has lost mechanical integrity (weakening). (d) Young s modulus values showing the changes in elasticity of the hair as stress is accumulated. Changes in cross-sectional area and Young s modulus as a function of sequential stress applications were also studied. The cross-sectional area was measured from averaging XTM data at five different locations along the fibre (Fig. 5b). This parameter remained relatively constant between baseline and stage 2. However, after the fibre exceeded 30% strain (Fig. 5b, stage 3), the cross-sectional area decreased more noticeably, particularly in the section where the final fracture occurred. Figure 5c and d shows the stress strain curves per stage and the Young s moduli calculated from these curves; we observed that the first stress application caused an 18% decrease in Young s modulus, while this parameter increased 11% after the second cycle. An additional cycle caused only a slight decrease of the elastic modulus. The impact of the changes in this parameter in the overall fracture mechanism will be discussed later. Discussion In spite of the cuticle splaying and the changes in crosssectional area during the stress application, the status of the fibre surface did not allow the prediction of the location of failure before it occurred. This could indicate that, under pure tensile forces, minor surface defects including missing or chipped cuticle layers and macroscopic heterogeneities contribute but do not necessarily play a determining role in
Breakage in curly hair, G.A. Camacho-Bragado et al. 15 ultimate hair breakage. One must remember than in real life, hair is rarely exposed to pure tensile stress and that pulling is usually accompanied by some sort of surface friction or abrasion from grooming tools. Based on the lack of surface breakage indicators, we propose that the critical cracks form and migrate from the inside out and that catastrophic failure occurs after the fibre has undergone a relatively large strain. The behaviour of the cuticle is consistent with adhesion failure at the cuticle cuticle interface as a first stage leading to fracture. As described by Robbins et al., 15 straining hair at humidity < 65%, corresponding to the humidity regime inside the microscope chamber, causes fracture at the CMC between cuticle layers due to weak bonds between hydrophobic components, particularly between the side chains of the fatty acid 18-methyleicosanic acid and the contiguous fibrous protein layer. This initial partial fracture of the cuticle is manifested as the initial decrease in elastic modulus. The generalized internal void formation eventually leads to detachment of the cuticle from the cortex; failure of the cuticle cortex interface results in the load being transferred and carried mainly by the cortex. At these later stages, the Young s modulus is dictated by the relatively stiff cortex, which is consistent with the observation of an increase in elastic modulus at stage 3 (Fig. 5c, d). The last small change in elastic modulus could be linked to propagation of cracks along the cortical CMC. As the contribution of the CMC to the elastic modulus is rather small, the corresponding change is expected to be small, as observed (Fig. 5c, d). The increase in crack density within the cortex accounts for the decrease in energy absorption at large strains as the air-filled cracks have a lower density than the crack-free areas of the fibre. Crack propagation along the intercellular space would cause the decrease in cross-sectional area by allowing the cells to slide and the hair to contract radially; thus one can define a mesoscopic-level Poisson effect in hair linked to the degradation of its hierarchical structure. It is not until the fibre has undergone large strains (> 30%) that the load is directly transferred to the cortical cells and eventually to individual microfibrils. At these stages, one would expect that fibres with different cortical cell distribution such as straight vs. curly hairs, 16 would have a different mechanical response as the different cell geometry (ortho- vs. meso- vs. paracortical cells) and packing could lead to different distribution of areas of internal stress concentration. As the CMC continues to fracture and the load is transferred to individual cells, the load-bearing macrofibrils and microfibrils within the cell start failing. Two competing processes have been reported to occur during straining of a-keratin intermediate filaments: 17 molecular stretching and molecular sliding. Time-resolved small-angle X-ray scattering data 17 showed that at low humidity the sliding process is more likely to occur. Thus, the last stage leading to failure of hair would be the fracture of intracellular CMC, which allows the fibrils to slide and eventually break. From the aforementioned observations, the mechanism of hair breakage can be summarized as taking place in the following four steps (Fig. 6): cuticle sliding; failure of the cuticle cortex interface (Fig. 6a); nucleation of intercellular cracks and growth of cracks at the cuticle cortex junction (Fig. 6b); and propagation of intercellular cracks towards the surface of the hair and final breakage when these cracks merge at the cuticular junction (Fig. 6c). In summary, the use of X-ray photon and electron imaging combined with in situ tensile testing provided a new insight into how human hair breaks, in particular curly hair. It has helped reconcile previous observations at the molecular level and at a purely macroscopic statistical level allowing us to propose a fracture mechanism that takes into consideration all constituents of hair at different scales. In the particular case of curly hair, fibres may break in two distinct ways: (i) at macroscopic constrictions, as they act as points of stress concentration and are more susceptible to surface-initiated cracks due to their inhomogeneous macrostructure, or (ii) at sites of accumulated internal stress; as demonstrated by this study, internal cracks may accumulate in the hair due to excessive grooming force without showing surface evidence of weakened spots; the hierarchical structure of hair is able to deflect the cracks to extend the life of the fibre up to a critical crack density and size. In both cases, hair fibres would benefit from a hair care regimen that reduces the grooming forces and the friction between grooming tools and fibres and in between fibres. Acknowledgments The studies were supported by L Oreal Research and Innovation. XTM was performed at the Advanced Light Source at Lawrence Berkeley National Laboratory, supported by the Office of Science, U.S. Department of Energy (DE-AC02- (c) Fig 6. Schematic representation of the hair-breakage mechanism. Failure of the cuticle cortex interface; observe the crack separating cuticle from cortex. Formation of intercellular cracks. (c) Propagation of cracks further into the cortex and towards the surface.
16 Breakage in curly hair, G.A. Camacho-Bragado et al. 05CH11231). The authors appreciate the assistance of Candace Woodson in the preparation of the illustrations. References 1 Bryant H, Porter C, Yang G. Curly hair: measured differences and contributions to breakage. Int J Dermatol 2012; 51(Suppl. 1):8 11. 2 McMichael A. Hair breakage in normal and weathered hair: focus on the black patient. J Investig Dermatol Symp Proc 2007; 12:6 9. 3 Kamath YK, Hornby S, Weigmann D. Mechanical and fractographic behavior of negroid hair. J Cosmet Sci 1984; 35:21 43. 4 Rosenborough I, McMichael A. Hair care practices in African- American patients. Semin Cutan Med Surg 2009; 28:103 8. 5 Sinclair R, Flagler MJ, Jones L et al. The proteomic profile of hair damage. Br J Dermatol 2012; 166(Suppl. 2):27 32. 6 Khumalo NP, Stone J, Gumedze F et al. Relaxers damage hair: evidence from amino acid analysis. J Am Acad Dermatol 2010; 62:402 8. 7 Dawber R. Cosmetic and medical causes of hair weathering. J Cosmet Dermatol 2002; 1:196 201. 8 Wolfram LJ. Human hair: a unique physicochemical composite. J Am Acad Dermatol 2003; 48:S106 14. 9 Robbins C. Chemical and Physical Behavior of Human Hair, 4th edn. Berlin: Springer, 2002. 10 Hearle JWS. A critical review of the structural mechanics of wool and hair fibres. Int J Biol Macromol 2000; 27:123 38. 11 Swift JA. The mechanics of fracture of human hair. Int J Cosmet Sci 1999; 21:227 39. 12 Bhushan B, Che N. AFM studies of environmental effects on nanomechanical properties and cellular structure of human hair. Ultramicroscopy 2006; 106:755 64. 13 Naresh MD, Subramanian V, Arumugam V, Sanjeevi R. A study on the mechanism of failure in keratin. Colloid Polym Sci 1991; 269:590 4. 14 Kanitpanyacharoen W, Parkinson DY, De Carlo F et al. A comparative study of X-ray tomographic microscopy on shales at different synchrotron facilities: ALS APS and SLS. J Synchrotron Radiat 2013; 20:172 80. 15 Robbins C, Weigmann HD, Ruetsch S, Kamath Y. Failure of intercellular adhesion in hair fibers with regard to hair condition and strain conditions. J Cosmet Sci 2004; 55:351 71. 16 Bryson WG, Harland DP, Caldwell JP et al. Cortical cell types and intermediate filament arrangements correlate with fiber curvature in Japanese human hair. J Struct Biol 2009; 166:46 58. 17 Kreplak L, Franbourg A, Briki F et al. A new deformation model of hard a-keratin fibers at the nanometer scale: implications for hard a-keratin intermediate filament mechanical properties. Biophys J 2002; 82:2265 74.