Author's Accepted Manuscript

Effect of water desorption on the rheology

and dynamic response of human hair to a
non-contact impact
J. Jamart, M. Djaghloul, J.M. Bergheau, H.
Zahouani

www.elsevier.com/locate/jmbbm

PII: S1751-6161(15)00072-7
DOI: http://dx.doi.org/10.1016/j.jmbbm.2015.02.029
Reference: JMBBM1406

To appear in: Journal of the Mechanical Behavior of Biomedical Materials

Received date:7 January 2015

Revised date: 24 February 2015
Accepted date:
26 February 2015

Cite this article as: J. Jamart, M. Djaghloul, J.M. Bergheau, H. Zahouani,

Effect of water desorption on the rheology and dynamic response of
human hair to a non-contact impact, Journal of the Mechanical Behavior of
Biomedical Materials, http://dx.doi.org/10.1016/j.jmbbm.2015.02.029

This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early
version of the manuscript. The manuscript will undergo copyediting,
typesetting, and review of the resulting galley proof before it is published in
its final citable form. Please note that during the production process errors
may be discovered which could affect the content, and all legal disclaimers
that apply to the journal pertain.
 Eect of water desorption on the rheology and dynamic
response of human hair to a non-contact impact
J. Jamarta,b,∗, M. Djaghloula , JM. Bergheaua , H. Zahouania
a Université de Lyon, Ecole Centrale de Lyon, ENISE, LTDS, CNRS-UMR5513, 69134 Ecully, France
b Société Calor/Groupe Seb, Département Soins de la Personne, 38780 Pont Eveque, France

Abstract
Human hair is a non-homogeneous complex material made of keratin bers oriented
along the longitudinal axis which oer anisotropic mechanical properties. Nowadays, it
is possible to measure the mechanical properties of hairs with the classical tests, but
most often, these tests are destructive and make hard to measure the inuence of some
external factors or treatments on the behavior of a same hair ber. In the current paper,
vibrations induced by a non-contact impact have been utilized as a representative response
of the mechanical behavior of hair. The characteristics of the vibratory response allow
measuring the variation in the mechanical properties and the instantaneous eect of an
external factor on the properties of a same sample. First, load relaxation tests have been
performed on hair samples after moisturization and for dierent times of an air-drying
process in order to characterize the change in the visco-elastic behavior of hair during
the water desorption. Other hair samples have been tested with our non-contact impact
and vibration technique in order to observe the change in the vibratory response during
the water desorption. The vibratory response has then been correlated to the mechanical
properties of the hair ber.
Keywords: Human hair, Water desorption, Induced vibration, Impact test

1. Introduction outer part and presents a multilayer orga-

As a biological composite material, hu-
man hair has a complex multi-scale struc-
ture extremely sensitive to several exter-
nal factors such as humidity, temperature,
chemical treatments, etc. To better un-
derstand the eect of these factors, it is
important to study and characterize the
bio-mechanical properties of hairs. With
a diameter of 60 to 100 µm, human hair
is made of a sulfur rich protein, the ker-
atin molecule. It's structure is composed
of two concentric parts. The cuticle is the
∗
Email address: juliette.jamart@ec-lyon.fr
Phone number: +33 472186558
Preprint submitted to Journal of the Mechanical Behavior of Biomedical Materials
nization made of attened cells, the scales.
The inner part is called the cortex and rep-
resents more than 90 % of the whole hair
weight (Zviak, 1988; Bouillon and Wilkin-
son, 2005). While the cuticle plays a pro-
tection role, the cortex is responsible for
the mechanical properties of hair. Indeed,
the cortex is composed of bril proteins
organized in helicoidal polypeptide chains,
the alpha-chains. The shape of the alpha
chains is maintained by hydrogen bonds
located between the coils of the helices.
The alpha-chains are also linked together
by other types of bonds, the salt bridges
and stronger covalent bonds, the disulde
March 5, 2015
bridges, in order to form the rst bril
structure of hair : the intermediate la-
ments (IFs) (Jones et al., 1997). Then the
IFs are embedded in an amorphous ma-
trix with a high sulfur content to form the
macrobrils and the cortical cells (Leon,
1972; Wolfram, 2003). Previous studies
have shown that this multi-scale organi-
zation did not dier according to the type
of hair studied or the ethnic origin (Wilk
et al., 1995; Franbourg et al., 2003). The
ber-matrix composite structure and the
orientation of the IFs along the longitudi-
nal axis is responsible for the anisotropic
mechanical properties of hairs (Robbins,
2012). Moreover, the presence of dierent
types of bonds and particularly of weak
bonds such as the hydrogen bonds makes
hair extremely sensitive to environmental
factors such as temperature and humidity.
The eect of water on the structure and
the mechanical properties of keratin bers
have been the subject of various previous
studies. The sorption-desorption hystere-
sis curve shows that hair bers are able
to absorb a large quantity of water, reach-
ing a maximum water regain of 26 % for
100 % of relative humidity (Chamberlain
and Speakman Leeds, 1931; Barba et al.,
2010). Concerning the hair structure,
Popescu and Höcker (2007) have shown
that water sorption resulted in swelling
and emphasized an anisotropy of swelling
with a radial swelling much more impor-
tant than the longitudinal swelling. In
another study, Feughelman (1994) built a
two-phase model where the intermediate
laments are considered to be water im-
penetrable rods embedded in a hydrophilic
matrix. This model is in accordance with
the swelling behavior of hair and can rep-
resent an explanation to the dependence of
the mechanical behavior of hair to relative
humidity.
The variation in the mechanical behav-
ior of hair as a function of relative hu-
midity has also been studied in the liter-
2
ature. As for all the keratin bers such
as wool, the tensile properties of hair are
changing with relative humidity (Hearle
and Morton, 2008). When hair is wet,
the Yield region starts for a lower level of
stress due to the breaking of the hydro-
gen bonds. Kreplak et al. (2002) explain
the eect of water by showing the coex-
istence of two deformation mechanisms in
the cortex, the sliding of the alpha-chains
inside the microbrils and their stretching.
The presence of water in the matrix would
make the sliding of the IFs easier whereas
for a dry hair under standard conditions
of humidity, the macroscopic deformation
would be due to the combination of both
phenomena.
In this general context, one of the main
tasks is to better understand the change
in the mechanical behavior of hair bers
during a variation of water content. Var-
ious dierent techniques can be used to
determine the mechanical properties of
hair. The most common are the destruc-
tive techniques such as tensile tests (Ve-
lasco and Dias, 2009; Thibaut et al., 2010;
Benzarti et al., 2011) and nanoindentation
tests (Wei et al., 2005). Huck and Baddiel
(1971) used a non-destructive method and
measured the dynamic response of hair
bers using an oscillating beam method in
order to characterize the elastic and loss
moduli. The study presented here is ex-
perimental. It uses relaxation tests per-
formed on several segments of a hair ber
in order to determine the global variation
in the visco-elastic properties of hair dur-
ing the water desorption kinetic. Then,
a new mechanical approach based on the
propagation of a vibration on a hair ber
has been developed. This approach uses
a non solid-solid contact impact in order
to have a non-destructive solicitation and
perform several tests on a same sample un-
der dierent conditions. Another advan-
tage of this method is the extremely brief
solicitation time which allows measuring
the instantaneous mechanical behavior of
the ber.
The present paper is organized as fol-
low : in Section 2, the samples and the
dierent devices used for this study are
described. Section 3 presents the relax-
ation tests performed and the associated
results. Section 4 is dedicated to the vi-
bration study. Finally, some conclusions
are drawn in Section 5.
2. Materials
2.1. Hair samples
For this study, virgin Caucasian hair
samples of type 2 were used. Before any
measurement, hair samples were washed
with a clarifying shampoo, rinsed with dis-
tilled water and stored during 48 hours
in a room at a temperature of 22C and
40 %RH.
2.2. Tensile testing machine
Relaxation tests on single hair bers
were performed with a tensile testing ma-
chine. The machine is made of the same
components as the classical testing devices
with a sample holder, a translation module
to pull up the sample and a force sensor
to compute the applied force on the ber,
as described in Fig.1.
The translation module used is a Poly-
tec Instruments stepper motor and can
reach a maximum speed of 1.5 mm/s over
a distance of 25 mm with a very high res-
olution lower than 1/100 µm. The speed
and the displacement are controlled by a
card connected to the engine control. The
force sensor which can load a maximum
of 3N with a resolution of 1 mN enables
the measurement of the force. The ap-
plied stress on the sensor is transmitted to
a computer via a National Instrument cap-
ture card and an amplier. Finally, a pro-
gram, implemented on Labview, is used to
control the engine movement and the data
acquisition.
3
2.3. Vibration measurement system
A new mechanical system based on the
vibration signature of hair was employed.
This system consists in creating an im-
pact without any solid-solid contact on the
hair sample and measuring its dynamic re-
sponse.
Fig.2 gives a description of the vibration
measurement system. The association of a
compressed air container, a pressure regu-
lator and a solenoid valve produces open-
ing/closing cycles of 3 ms creating an air
ux at a pressure of 1 bar in the pneumatic
system. Then, the air goes through a noz-
zle with an exit diameter of 2 mm in order
to produce extremely brief impacts on the
sample to generate a vibration. The dis-
tance between the exit of the nozzle and
the sample has been set to 10 mm. The
laser (LK-G82 Keyence) is used to mea-
sure the instantaneous deection at the
impact point located at 20 mm from the
rst clamp. The vibration is followed by
monitoring the wave propagation with a
second laser (PDV-100 Polytec) placed at
40 mm from the impact point. A program
implemented on Labview is used to con-
trol the solenoid valve and the data acqui-
sition. To control the measurement, hair
samples are initially placed in the tensile
testing machine described previously (Sec-
tion 2.2) in order to apply a preload.
3. Eect of water desorption mea-
sured by load relaxation tests
3.1. Load relaxation tests
3.1.1. Measurement
Load relaxation tests are often used to
analyze the mechanical behavior of visco-
elastic materials. It consists in applying
and maintaining a constant deformation
and observing a decrease of the normal
force due to the viscosity of the mate-
rial. In order to observe the eect of wa-
ter desorption on the visco-elastic proper-
ties of human hair, load relaxation tests
were performed on 7 segments of 30 mm
of a same hair ber. First, the clos-
est segment to the root was placed be-
tween the clamps of the tensile testing
machine with an initial length between
the clamps of 10 mm. A rst relaxation
test of 600 seconds was performed at an
ambient temperature of 22C and a rela-
tive humidity of 40 %RH, in order to ob-
tain the properties under dry conditions.
The other segments were immersed in dis-
tilled water for 15 minutes and succes-
sively tested with the same relaxation test
to represent the dierent times of an air-
drying process, at intervals of 20 minutes.
The selected desorption times were :
0 minute (t0 ), 20 minutes (t1 ), 40 minutes
(t2), 60 minutes (t3), 80 minutes (t4) and
100 minutes (t5 ). For each test, the ap-
plied deformation was 2 % and the loading
speed was 1 mm/s. This experiment was
reproduced with 5 dierent virgin Cau-
casian hair samples.
3.1.2. Rheological model and estimated
parameters
A rheological model was employed to
analyze the experimental load relaxation
curves. Our previous work (Benzarti
et al., 2014) has shown that the general-
ized two-branches Maxwell model allowed
describing the visco-elastic mechanical be-
havior of hair. This model consists in the
association of a spring connected in par-
allel with 2 branches each containing a
spring and a viscous damper connected in
series (Fig.3).
The relaxation function Φ(t) resulting
from this model is :
 
Φ(t) = K0 + K1 exp
−K1
t also sharper than for the dry curve.

n1
 (1)
+K2 exp
−K2
n2
t
where K0, K1 and K2 are the coecients
4
of the springs and n1 and n2 the coe-
cients of the viscous dampers.
All the experimental load relaxation
curves were tted with the generalized
two-branches Maxwell model using the
simplex method. Fig.4 shows an example
of one experimental load relaxation curve
and its optimized t.
The initial stiness Kt=0 and innite
stiness Kt=∞ can be estimated using the
parameters obtained from the optimized
t, equation 2a and equation 2b.
Kt=0 = K0 + K1 + K2 (2a)
Kt=∞ = K0 (2b)
Another parameter can be estimated :
the relaxation time tr . This time corre-
sponds to the intersection point of the two
asymptotes of the relaxation curve and is
obtained from the optimized t and equa-
tion 3.
K1 + K2
tr = K12 K22
(3)
η1
+ η2
3.2. Results
An example of the experimental load re-
laxation curves obtained on the 7 segments
of a same hair ber in the dry state, just
after the wetting phase and for the dif-
ferent desorption times is represented in
Fig.5. This graph emphasizes the eect
of water content on the relaxation behav-
ior. The comparison of the dry and the
wet curve highlights a change in the initial
and nal force and of the curvature of the
curve. The initial force decreases from 249
to 137 mN from the dry to the wet state
and the nal force decreases from 181 to
111 mN. The curvature of the wet curve is
Besides, the curves obtained for the dif-
ferent desorption times are located be-
tween the dry and the wet curve. The
longer the desorption time is, the closer to
the dry state the relaxation curve is. This
means that there is a progressive increase
of both initial and nal force as the sam-
ple gets dry. The curvature becomes also
smoother.
This progressive recovery is also ob-
served on the stinesses Kt=0 and Kt=∞
estimated with the generalized two-
branches Maxwell model and on the re-
laxation time tr .
Fig.6 represents the variation in each
parameter along the air-drying process for
the 5 hair bers studied. The values of
each relaxation parameter normalized by
the values obtained for the dry sample are
represented on the graphs. For both sti-
nesses, the ratio of the wet (0 minute) to
the dry state is close to 0.6 whereas for the
relaxation time tr , it is equal to 0.1, show-
ing that the eect of the water content on
the latter is more signicant. These three
parameters increase with time. However,
their kinetic of change are dierent. The
relaxation time tr represented in Fig.6c is
the parameter reaching its maximum value
fastest. The value at 20 minutes is equal to
80 % of the dry value. In contrast, the in-
nite stiness Kt=∞ represented in Fig.6b
has the slowest variation. Fig.6a shows
that the change in the initial stiness Kt=0
can be split into two parts. As for the re-
laxation time, a fast variation is observed
between the wet state and 20 minutes then
a more progressive change is observed from
20 minutes to 100 minutes.
3.3. Discussion
During a load relaxation test, the de-
crease of the measured force is due to
the progressive rearrangement of the ma-
terial micro-structure. The several relax-
ation tests performed for dierent desorp-
tion times permit to observe the change
in the visco-elastic behavior of a hair ber
according to its water content. As the re-
laxation tests were performed in the elas-
tic region, the initial stiness Kt=0 and the
innite stiness Kt=∞ can be respectively
5
correlated to the elastic moduli Et=0 and
Et=∞ of the hair ber by equation 4, where
S is the hair ber section and l its length.
K(t) =
E(t).S(t)
(4)
l(t)
The stinesses depend on geometrical
(S and l) and mechanical (E ) parameters.
The swelling curve represented by Ben-
zarti et al. (2014) shows that the swelling
of the section S is less than 15 % from
the dry to the wet state. Moreover, the
swelling in the longitudinal direction is
less than 2 % (Popescu and Höcker, 2007).
Therefore, the change in both stinesses
is not only due to the swelling behavior
of hair but also to a change in the elas-
tic moduli. Taking into account a 15 %
swelling of the section and a 1 % longitudi-
nal swelling of the sample after the wetting
phase, the ratio between Ewet and Edry can
be estimated by equation 5.
Ewet
= 0.87 ∗
Kwet
(5)
Edry Kdry
The initial modulus Et=0 decreases by
48 % just after the wetting phase. This
result is in accordance with the results ob-
tained by Wolfram (2003); Bouillon and
Wilkinson (2005) who have shown that an
increase of relative humidity created a de-
crease of the elastic modulus due to the
breaking of the hydrogen bonds caused by
the presence of water in the ber. The
increase of Et=0 and Et=∞ during the air
drying process can also be explained by
the creation of new hydrogen-bonds due
to the water desorption.
The change in the relaxation time tr is
in accordance with the results found by
Wortmann et al. (2009) who studied hair
bending properties. The lower the wa-
ter content is, the slower the relaxation
is. This behavior variation emphasizes the
ability of the keratin molecules to rear-
range each other. The low relaxation time
obtained just after the wetting phase is
due to the fact that keratin molecules are
more able to rearrange, the sliding mech-
anism of these molecules being facilitated
by the presence of water inside the ber
(Kreplak et al., 2002).
The relaxation curves and parameters
obtained show that water content has a
big impact on the visco-elastic properties
of human hair. However, these tests were
performed on several segments of a same
hair ber and not on a same sample. In-
deed, the rearrangement of the keratin
molecules which occurs during the relax-
ation changes the micro-structure organi-
zation. Thus, the results obtained in this
rst part give only access to the global
change of the mechanical properties of hair
during the air-drying process.
4. Eect of water content on the
propagation of an induced vibra-
tion
4.1. Measurement
The eect of water desorption on the
vibration signature of hair bers was fol-
lowed by the impact test described in 2.3.
First, a segment of 60 mm of a hair sample
was xed between the clamps of the tensile
testing machine. A rst test consisting in
applying a preload of 30 mN and 20 air jets
to the sample was performed. Then, the
sample was moisturized with 4 impulses
of a distilled water spray and the same
test was performed on the sample just af-
ter moisturization (0 minute) and after the
same desorption times as in 3.1. This ex-
periment was reproduced on 5 dierent
virgin Caucasian hair samples of type 2.
4.2. Selection of parameters
The measurements of deection at the
impact point (A) and of the velocity mea-
sured at 40 mm from the impact point
(B) normalized by its maximum value are
given in Fig.7.
6
Four parameters have been selected and
estimated to characterize the vibration re-
sponse of the signal (Fig.7):
• The wave propagation delay WPD
corresponds to the time between the
maximal deection at the impact
point (A) and the start of the damp-
ing part of the induced vibration at
the point (B).
• The damping coecient α is deter-
mined by tting the envelope of the
damping part of the vibration sig-
nal with an exponential decay func-
tion. α corresponds to the exponen-
tial decay coecient. A peak de-
tection function is used on Matlab
and the associated exponential decay
function is found by the method of the
least square.
• The vibratory level Lv is estimated by
the following equation :
 
VRM S
Lv = 20 log10 (6)
V0
where VRM S is the root mean square
value of the attenuation part of the
vibration signal and V0 the reference
value of the vibratory level.
• The natural frequency fn of the signal
is obtained by performing the Power
Spectral Density PSD of the damping
part of the signal.
4.3. Results
An example of the deection measure-
ments at the impact point (A) and of the
vibration signal measured at the point (B)
are given in Fig.8 for the same hair sam-
ple under dry conditions and just after
the wetting phase. The dierence between
both vibration signals shows the eect of
water content on the propagation of the
induced wave.
 The graph shows that the wave propa-
gation delay WPD increases when the hair
sample gets wet meaning that the longitu-
dinal speed of the induced wave is slower.
The attenuation part is also dierent and
seems slower for the wet state than for the
dry state.
The variation over time of the four esti-
mated parameters for the 5 samples stud-
ied is represented in Fig.9. Figures 9a,
9b, 9c and 9d give respectively the change
in the WPD, fn, α and Lv . For all the
parameters, an important variation is ob-
served just after the wetting phase, then,
a progressive recovery to the initial value
is observed as the sample gets dry. The
change in the selected parameters can be
classied into two categories: the WPD
and the vibratory level Lv decrease when
the desorption time increases whereas the
damping coecient α and the natural fre-
quency fn increase. The natural frequency
fn is the only parameter which reaches its
maximal dierence from its dry value just
after the wetting phase, at 0 minute. For
the other three parameters, the maximum
dierence is reached 20 minutes after the
wetting phase.
4.4. Discussion
In this conguration, the hair sam-
ple can be considered to be a clamped-
clamped circular beam in simple bend-
ing. The propagation of the induced vi-
bration can therefore be correlated to the
mechanical properties of the sample using
the Oberst method (Oberst, 1952).
For instance, the WPD which is pro-
portional to the inverse of the longitudinal
propagation speed vL is directly linked to
the elastic modulus by equation 7 where ρ
is the ber density.
E = ρ.vL 2 (7)
The water sorption creates an impor-
tant increase of the linear density but the
7
important radial swelling that occurs dur-
ing the sorption creates a global constancy
of the density ρ (Popescu and Höcker,
2007). As an example, King (1926) have
shown that the density of wool bers in-
creases by only 3 % from 0 % to 100 %
of relative humidity . Thus, the change
in the WPD after the wetting phase and
after dierent desorption times gives di-
rectly the change in the elastic modulus
E . The estimated ratios are given in Ta-
ble 1.
At 0 minutes and 20 minutes, the WPD
increases. The lowest ratio between E(t)
and Edry is obtained at 20 minutes and is
equal to 0.52. This result is in accordance
with the results found with the previous
relaxation tests presented in Section 3.3.
The elastic modulus is also linked to the
natural frequency fn by the equation 8.
E∝
fn .ρ.l4
(8)
d2
This equation takes into account a ge-
ometrical parameter: the ber diameter
d. The radial swelling RS(t) can be es-
timated considering the previous values of
elastic moduli and the measured values of
natural frequencies with equation 9.

d(t) fn (t).Edry
RS(t) = −1  −1 (9)
ddry fndry .E(t)
The highest value of radial swelling
RS of 19.89 % is obtained at 20 minutes.
This value is close to the maximal radial
swelling of hair obtained at 100 % of rela-
tive humidity (Popescu and Höcker, 2007).
The maximum of radial swelling and the
minimal elastic modulus are not obtained
just after the wetting phase ( 0 minute) but
20 minutes after. This delay of 20 minutes
can be considered to be the time neces-
sary for the water to penetrate into the
hair ber. As the sample has been super-
cially wet, the water is rst in contact
with the cuticle and has not an instan-
taneous eect on the mechanical proper-
ties. When the diusion of the water in
the cortex begins, the swelling of the ma-
trix starts and the eect on the mechanical
properties gets visible.
The change in the damping coecient α
also conrms the values of the estimated
elastic moduli. The stier the material is,
the faster the damping of the induced vi-
bration is. The same interpretation can
be performed for the vibratory level Lv
that follows exactly the same tendency as
the wave propagation delay WPD. The de-
crease of the elastic modulus produces an
increase of the vibration amplitude.
5. Conclusion
In the present paper, an experimental
study has been performed to better under-
stand the change in the mechanical behav-
ior of hair bers during a variation of water
content due to an air-drying process. The
load relaxation tests performed hilight the
global eect of water content on the visco-
elastic properties of hair. The elastic mod-
ulus decreases when the hair ber goes
from the dry to the wet state and increases
progressively as the hair ber gets dry.
The same tendency is also observed for
the relaxation time. These results con-
rm the main eect of water on the micro-
structure of hair, breaking the weak hy-
drogen bonds located between the coils of
the alpha-helices and facilitating the slid-
ing of the keratin chains and their rear-
rangement. However, because of the re-
organization of the micro-structure dur-
ing these tests, load relaxation tests do
not permit to follow instantaneously the
variation in the mechanical properties of
a same hair sample. Thus, a new me-
chanical system based on the vibration
signature of hair was employed. The se-
lected and estimated parameters obtained
by analyzing the vibration signal can be
8
correlated to the mechanical properties.
The wave propagation delay, is directly
linked to the elastic modulus such as the
natural frequency which depends also on
the hair diameter. When the hair ber
gets wet, the hair diameter increases by
19.89 % and the elastic modulus decreases
by 48 %. This decrease does not occur just
after the supercial wetting phase high-
lighting the fact that several minutes are
necessary for water to go through the cu-
ticle and penetrate into the cortex. The
same tendency was also measured on the
radial swelling. Thus, the results obtained
by this new measurement system are in
accordance with the results obtained by
performing load relaxation tests conrm-
ing the reliability of this method. This
method also has several advantages: the
measurement is non-destructive and quick,
thus making it possible to have several in-
stantaneous responses for a same sample,
giving access to the controlled change of
its mechanical properties.
References
Barba, C., Martí, M., Manich, A., Carilla, J.,
Parra, J., Coderch, L., May 2010. Water ab-
sorption/desorption of human hair and nails.
Thermochimica Acta 503-504, 3339.
Benzarti, M., Pailler-Mattei, C., Jamart, J., Za-
houani, H., Jun. 2014. The Eect of Hydration
on the Mechanical Behaviour of Hair. Experi-
mental Mechanics.
Benzarti, M., Tkaya, M. B., Mattei, C. P., Za-
houani, H., 2011. Hair Mechanical Properties
Depending on Age and Origin. International
Science Index 5 (2), 400406.
Bouillon, C., Wilkinson, J., 2005. The science of
hair care.
Chamberlain, N. H., Speakman Leeds, J.
B. I., 1931. Über Hysteresiserscheinungen
in der Wasseraufnahme des Menschenhaares.
Zeitschrift für Elektrochemie und angewandte
physikalische Chemie 37 (7), 374377.
Feughelman, M., Apr. 1994. A Model for the Me-
chanical Properties of the alpha-Keratin Cor-
tex. Vol. 64. Textile Research Institute.
Franbourg, A., Hallegot, P., Baltenneck, F.,
Toutain, C., Leroy, F., Jun. 2003. Current re-
 search on ethnic hair. Journal of the American unique physicochemical composite. Journal of
Academy of Dermatology 48 (6 Suppl), S1159. the American Academy of Dermatology 48 (6
Hearle, J., Morton, W., 2008. Physical properties Suppl), S10614.
of textile bres. Wortmann, F. J., Stapels, M., Chandra, L., Sep.
Huck, P., Baddiel, C., 1971. The mechanical 2009. Humidity-dependent bending recovery
properties of virgin and treated human hair and relaxation of human hair. Journal of Ap-
bers; a study by means of the oscillating beam plied Polymer Science 113 (5), 33363344.
method. Journal of the Society of Cosmetic Zviak, C., 1988. Science des Traitements Capil-
Chemists of Great Britain 410, 401410. laires.
Jones, L. N., Simon, M., Watts, N. R., Booy,
F. P., Steven, a. C., Parry, D. a., Oct. 1997.
Intermediate lament structure: hard alpha- Force sensor Motor
keratin. Biophysical chemistry 68 (1-3), 8393.
King, A. T., Jan. 1926. The relation to swelling
and sorption in water and other liquids. Jour-
nal of the Textile Institute 17 (1), T53T67.
Kreplak, L., Franbourg, A., Briki, F., Leroy,
F., 2002. A New Deformation Model of Hard
α-Keratin Fibers at the Nanometer Scale: Sample Holders
Implications for Hard α-Keratin Intermedi-
Figure 1: Schematic representation of the tensile
ate Filament Mechanical. Biophysical journal
testing machine
82 (April).
Leon, N., 1972. Structural aspects of keratin
bres. Journal of the Society of Cosmetic
Chemists of Great Britain 445.
Oberst, H., 1952. Über die Dämpfung der Velocimeter Solenoid valve
Biegeschwingungen Dünner Blech durch Fest
Haftende Belage. Journal Acustica 2 (4), 180 Pressure
194. regulator
Popescu, C., Höcker, H., Aug. 2007. Hair the
most sophisticated biological composite mate-
Force
rial. Chemical Society reviews 36 (8), 128291. Air blast
Robbins, C. R., 2012. Chemical and Physical Be- sensor
havior of Human Hair. Springer Berlin Heidel- Tensile Machine
berg, Berlin, Heidelberg. Hair fiber
Thibaut, S., de Becker, E., Bernard, B. a., Huart,
M., Fiat, F., Baghdadli, N., Luengo, G. S.,
Leroy, F., Angevin, P., Kermoal, a. M., Muller, Figure 2: Schematic representation of the vibra-
S., Peron, M., Provot, G., Kravtchenko, S., tion measurement system
Saint-Léger, D., Desbois, G., Gauchet, L.,
Nowbuth, K., Galliano, A., Kempf, J. Y., Sil-
berzan, I., Apr. 2010. Chronological ageing of K0
human hair keratin bres. International Jour-
nal of Cosmetic Science, 422434.
K1
1
Velasco, M., Dias, T., 2009. Hair ber charac-
teristics and methods to evaluate hair physical
and mechanical properties. Brazilian Journal of K2
2
Pharmaceutical Sciences 45.
Wei, G., Bhushan, B., Torgerson, P. M., Nov.
2005. Nanomechanical characterization of hu- Figure 3: Generalized two-branches Maxwell
man hair using nanoindentation and SEM. Ul- model
tramicroscopy 105 (1-4), 24866.
Wilk, K., James, V., Amemiya, Y., 1995. The
intermediate lament structure of human hair.
Biochimica et Biophysica Acta 1245, 392396.
Wolfram, L. J., Jun. 2003. Human hair: A

9
 280
Experimental curve
Maxwell 2-branches model 1
260
Normal Force [mN]

0.9

dry
240

K t=0 / Kt=0
0.8
220
0.7

200 0.6

180 0.5
t0 t1 t2 t3 t4 t5
0 tr 100 200 300 400 500 600
Time [s]
(a)
Figure 4: Example of an experimental load relax-
ation curve (solid line), its optimized generalized
two-branches Maxwell model t (dashed line) and
the intersection point of its two asymptotes used
to determine the relaxation time tr

t0 t1 t2 t3 t4 t5

(b)
1

0.8
tr / trdry

0.6
250
0.4
Normal Force [mN]

200 0.2
dry
t5
t4 0
t3 t0 t1 t2 t3 t4 t5
150 t2
t1 (c)
t0
100 Figure 6: Change over time in the three estimated
0 100 200 300 400 500 600
Time [s] relaxation parameters normalized by there dry
values. (a) Change in the initial stiness Kt=0 ,(b)
Figure 5: Example of the experimental relaxation Change in the innite stiness Kt=∞ , (c) Change
curves obtained on the same sample in the dry in the relaxation time tr
state, just after the wetting phase and for dierent
desorption times

10
 0.65
A B
0.6
2
Deflection [mm]

WPD [s]
0.55
0

A 0.5
-2
0.45
-4
0 0.2 0.4 0.6 0.8 1 1.2 1.4
0.4
(1)
dry 0min 20min 40min 60min 80min 100min
Normalized Velocity

e-
tt (2)
1

0.5

B 0
450
-0.5

-1 400
0 0.2 0.4 0.6 0.8 1 1.2 1.4
time [s]

fn[Hz]
350
(a)
1
300
(3)
250
Spectral Magnitude

0.8

0.6 200
dry 0min 20min 40min 60min 80min 100min

0.4

0.2
9
0
0 200 400 600 800 1000 8
Frequency [Hz]
7
(b)
6

dry 0min 20min 40min 60min 80min 100min

α
182

fn
180
Lv[dB]

178
2
Deflection [mm]

0 176

-2 wet
dry dry 0min 20min 40min 60min 80min 100min
-4
0 0.2 0.4 0.6 0.8 1 1.2 1.4

0.5
[arb.units]

-0.5

-1
0 0.2 0.4 0.6
time [s]
0.8 1 1.2 1.4
fn
α
Lv
 State 0 min 20 min 40 min 60 min 80 min 100 min
W P D(t)/W P Ddry 1.29 1.41 1.26 1.07 1.07 0.99
E(t)/Edry 0.62 0.52 0.65 0.86 0.87 1.01
fn (t)/fndry 0.63 0.75 0.77 0.86 0.88 1.03
RS (%) 0.60 19.89 8.92 0.33 0.50 1.26
Table 1: Change over time in the WPD, E , fn and RS normalized by their dry values

12