1. Characterization of yield stress and slip behaviour of skin/hair care
gels using steady flow and LAOS measurements and their
correlation with sensorial attributes
S. Ozkan, T. W. Gillece, L. Senak and D. J. Moore*
Materials Science Group, Global R&D, Ashland Specialty Ingredients, 1361 Alps Road, Wayne, NJ 07470, U.S.A.
Received 29 June 2011, Accepted 19 December 2011
Keywords: gels, LAOS, rheology, sensory correlation, slip, yield stress
Synopsis
Gels made with three different polymers widely used as rheology
modifiers in cosmetic formulations (cross-linked poly(acrylic acid),
cross-linked poly(maleic acid-alt-methyl vinyl ether) copolymer and
cross-linked poly(acrylic acid-co-vinyl pyrrolidone) copolymer) were
characterized by rheological and sensory evaluation methods to deter-
mine the relationship between sensorial perception and corresponding
rheological parameters. Both conventional rheological characteriza-
tion methods and a more recent method, Fourier Transform Rheology
with Large Amplitude Oscillatory Flow data (LAOS), were utilized to
characterize the material with and without wall slip. Sensorial analy-
ses were implemented in vivo to evaluate the perceived ease of initial
and rub-out spreadability, cushion, pick-up and slipperiness attributes
of the gels. Results were statistically analysed by both variance (ANO-
VA) and principle component analysis (PCA). Sensorial panel testing
characteristics discriminated the three materials, and PCA analyses
revealed that sensory attributes could be well predicted by rheological
methods. Rheological experiments, without wall slip, revealed that gel
strength in the linear viscoelastic region (LVR) and yield stress of these
materials are similar, but exhibit significantly different wall slip and
thixotropy behaviour in the low shear rate region under wall slip con-
ditions. Above the critical shear rate, which corresponds to the yield
stress, all tested materials did not slip and behaved as conventional,
shear thinning polymeric fluids. In particular, the rheological parame-
ters and sensorial perception of the 1% cross-linked vinyl pyrrolidone/
acrylic acid copolymer were significantly affected by wall slip and/or
thixotropy-related shear banding phenomena.
Introduction
Sensory properties of personal care products contribute substantially
to overall consumer acceptance. Therefore, costly time-consuming
sensory evaluation techniques are applied to guide the formulator in
identifying and defining the sensory profile of a product. Further, by
correlating quantitative instrumental parameters with critical, yet
subjective, sensorial ratings, consumer perception may be better
understood.
Conventional rheological and mechanical testing methods, such
as dynamic oscillatory flow and steady torsional flow measure-
ments, measuring Young’s modulus or maximum normal force,
etc., have been widely utilized to characterize hydrogels in the
food industry for the purpose of correlating structural properties
with sensorial perception [1–6]. Even though a significant
amount of recent literature is available for this purpose, certain
aspects of hydrogel rheology, such as thixotropy, wall slip and
shear banding phenomena, were not addressed in these studies.
In a recent review [7], Fisher et al. mention that a description of
the rheology of gels and concentrated food systems, such as gel-
like glassy matter exhibiting ageing behaviour, is considered a
new approach and is not embraced as a method of choice by
food scientists [7].
From a rheologist’s point of view, characterization of hydrogels
and gel-like percolated suspensions/emulsions, and the determina-
tion of their accurate yield stress present special challenges. These
materials are associated with thixotropy, viscoplasticity and wall
slip. Ideally, the true yield stress should be determined directly as
the minimum values of the shear stress (1-D) or stress magnitude
(3-D) at which deformation is observed, giving special consideration
to the wall slip effect. Typically, the yield stress is determined from
shear stress vs. shear rate data by fitting the Bingham or Herschel–
Bulkley equations to the data. These methods are, however, prone
to experimental errors and wall slip effects, and not all materials
comply with these common methods. Consequently, measuring the
rheological parameters without paying attention to thixotropy and
wall slip, as well as correlating these parameters with sensorial
attributes, may not be optimal.
In this study, we conducted an extensive rheological study
addressing wall slip and thixotropy to investigate the correlations
between the rheological properties (conventional and Fourier trans-
form rheology with large amplitude oscillatory flow data (LAOS))
and the sensorial attributes of three commercially available cosmetic
rheological modifiers. The purpose of the study was to discern which
rheological methods and parameters are more appropriate for corre-
lation to specific sensorial attributes. To the extent of our knowl-
edge, a sensorial correlation study that combines conventional and
LAOS rheological parameters with and without wall slip has not
been published as it relates to cosmetics.
Correspondence: Seher Ozkan, Material Science Group, Global R&D, Ash-
land Specialty Ingredients, 1361 Alps Road, Wayne, NJ 07470, U.S.A.
Tel.: 9736283971; Fax: 9736283886; e-mail: sozkan@ashland.com
This work has been presented at the 82nd
Society of Rheology Annual
Meeting, Santa Fe, New Mexico, October 27, 2010 and MRS Fall meeting,
Boston, MA, December 4, 2009.
*Current address: TRI-Princeton, 601 Prospect Avenue, Princeton, NJ
08540, U.S.A.
International Journal of Cosmetic Science, 2012, 34, 193–201 doi: 10.1111/j.1468-2494.2012.00702.x
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie 193

2. The selected hydrogel rheology modifiers were as follows: cross-
linked poly(acrylic acid), cross-linked poly(maleic acid methyl vinyl
ether) copolymer and cross-linked poly(acrylic acid-co-vinyl pyrroli-
done) copolymer. The slip behaviour of these hydrogels was investi-
gated by compiling steady torsional flow data at different gap
openings and dynamic strain sweep data with smooth and rough-
ened surfaces (to allow or suppress wall slip). The yield stress, consis-
tency index and power law index of the hydrogels were obtained by
fitting slip-corrected shear stress vs. shear rate data to the Herschel–
Bulkley (H-B) model. In addition to conventional rheological charac-
terization methods, a recent method, Fourier transform rheology
with LAOS was utilized to characterize the material with and with-
out wall slip. Sensorial analyses were implemented in vivo to evalu-
ate the perceived ease of initial and rub-out spreadability, cushion,
pick-up and slipperiness attributes of the gels. The overall rheological
parameters were correlated with sensory panel test results.
Materials and methods
Three different rheology modifiers were studied: cross-linked poly
(acrylic acid) (CarbopolÒ
980; Lubrizol, Wickliffe, OH, U.S.A.),
cross-linked poly(maleic acid-alt-methyl vinyl ether) copolymer
(StabilezeÒ
QM; ASI, Wayne, NJ, U.S.A.) and cross-linked poly
(acrylic acid/vinyl pyrrolidone) copolymer (UltrathixÔ
P100; ASI).
We will use the following acronyms in the text from this point on
to reference the thickeners: PAA-XL for cross-linked poly(acrylic
acid); PVM/MA-XL for cross-linked poly(maleic acid-alt-methyl
vinyl ether) copolymer; and PAA/VP-XL for the cross-linked poly
(acrylic acid/vinyl pyrrolidone) copolymer.
Samples were prepared as 1% (w/w) gels in de-ionized water
and were neutralized and preserved following the protocols given
by the manufacturers. Air bubbles were removed by centrifuging
the samples. Prior to use, each sample was left unperturbed for at
least 24 h to rebuild the structure lost during preparation and han-
dling. The sample concentrations were confirmed by thermogravi-
metric analysis measurements (Hi-Res TGA 2950 from TA
Instruments, New Castle, DE, U.S.A.). During the study, all samples
were stored in sealed containers at room temperature.
Rheological characterization
The rheological properties of 1% gels were investigated using a
stress-controlled AR-G2 rheometer and a strain-controlled ARES
rheometer (TA Instruments). All tests were carried out at
25 ± 0.1°C. As sample loading conditions influence testing, a
meticulous and consistent routine for sample loading was followed
to promote reproducible results. Prior to data collection, a 5-min
delay was applied to ensure rebuilding of the gel structure that was
compromised during sample loading. Four different types of mea-
surement results are reported:
Dynamic oscillatory measurements
The strain and frequency dependency of the materials functions,
such as the magnitude of complex viscosity (g*), storage modulus
(G¢(x)), loss modulus (G¢¢(x)) and oscillatory stress, were measured
using dynamic testing. Generally, smooth-surfaced plates were used
for dynamic testing; however, strain sweep experiments were also
repeated at x = 1 rps using plates, the surfaces of which were cov-
ered with 400 grit, adhesive-backed, waterproof sandpaper (ARC
Abrasives, Inc., Troy, OH, U.S.A.) to eliminate wall slip. Fourier
transform analysis was applied to the large amplitude oscillatory
shear (LAOS) flow data that had been collected with smooth and
sandpaper-covered plates. The sinusoidal stress response signal col-
lected from the sample was separated into elastic and viscous stress
contributions using symmetry arguments [8]. Chebyshev polynomi-
als (closely related to the Fourier deconvolution) were utilized as
orthonormal basis functions to further decompose these stresses
into odd and even harmonic components having physical interpre-
tations [8]. Multiple steady-state wave forms were used for data
analysis (typically three cycles of data were collected, and the last
two cycles, where the data had equilibrated, were used) at each
coordinate pair (x, c0).
Steady torsional flow experiments
The AR-G2 rheometer with parallel disc fixtures of 20 mm diame-
ter was used for rotational viscometry. At each shear rate, a fresh
sample was used to avoid pre-shearing of the gels. Steady torsional
flow experiments with parallel plates were carried out at two differ-
ent gap heights, 1.0 and 1.5 mm, for each shear rate. Each condi-
tion was repeated at least three times with fresh samples. As will
be discussed in the results section, the results from these experi-
ments are indicative of the presence of wall slip in torsional flow
[9, 10]. The wall slip behaviour of the gels, characterized in terms
of the slip velocity vs. the wall shear stress relationship, was used
in correcting the rheological data for wall slip and calculation of
Navier’s slip coefficients following Kalyon et al. [11]. Herschel–
Bulkley-derived parameters, such as the yield stress, consistency
index and power law index, were determined from fitting slip-cor-
rected steady torsional flow data to the Herschel–Bulkley model
defined by the following equation (1),
s ¼ s0 þ m_cn
ð1Þ
where s is the shear stress (Pa), s0 is the yield stress (Pa), m is the
consistency index (Pa s1/n
), _c is the shear rate (1 s)1
) and n is
the power law index.
Continuous shear rate ramp tests
The AR-G2 rheometer, equipped with 20-mm parallel disc fixtures,
was used for stress/shear rate ramp tests by ramping up to 500 s)1
in 1 min at two different gap openings (1.0 mm and 1.5 mm) to
determine the shear rate range where wall slip influenced the mea-
surements.
Extensional tests
The AR-G2 rheometer, with 20 mm stainless steel parallel plates,
was used to carry out extensional force measurement tests. The
same pre-test protocols used in rotational experiments were used
for all extensional testing. The initial gap was set to 400 lm, and
the top plate was raised to a 3-mm gap with a speed of 3 mm s)1
.
Force vs. gap data were collected, and the maximum force reading
was reported. Each sample was tested at least six times, and the
results were reported as averages of maximum force values with
95% confidence intervals. One-factor analysis of variance (ANOVA)
test was applied to test the significance of differences.
Sensory evaluation
Primary skin-feel parameters, such as pick-up, cushion/body/firm-
ness, initial spreadability, and secondary skin-feel parameters, such
as rub-out spreadability, slipperiness/lubrication, were evaluated for
correlation against rheological parameters. We used the existing
lexicon to define these attributes and to determine the intensity
scales and reference values for each attribute [12].
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201194
Sensory correlation with rheology S. Ozkan et al.

3. A randomized complete-block experimental design was carried
out for evaluation of the samples, where panellists are the ‘blocks’
and the samples are the ‘treatments’. Each panellist evaluated all
three samples to form a ‘complete block’. This design type is effec-
tive when panellists are consistent in rating the samples, but may
be using different ranges of the scale to express their perceptions.
Primary and secondary skin-feel parameters were evaluated by six
selected untrained panellists. Each attribute was tested on separate
days. The panellists rated the pick-up, cushion/body/firmness, ini-
tial spreadability, rub-out spreadability and slipperiness/lubrication
character of the gels on a scale of 0–10. Reference standards were
available as given in Meilgaard et al. [12]. The same sample
batches were used for both sensory evaluation and rheological
characterization of 1% PAA-XL, 1% PVM/MA-XL and 1% PAA/VP-
XL. Samples were stored in sealed ointment jars at 25°C and equili-
brated at least 24 h prior to evaluation to let the internal structure
completely rebuild. All tests were conducted in a temperature- and
humidity-controlled environment. Panel members were informed
that the samples were composed of three different thickeners, but
the identity of individual samples was not disclosed. Members
worked individually, and no discussions took place during the ses-
sions.
A rectangle of silicone release paper was taped to the bench top,
and 0.3 grams of each sample and reference materials were deli-
cately and simultaneously applied to the substrate surface using a
measuring spoon to ensure the minimal disruption of the internal
structure of the gel network. Sample jars were three-digit coded
and were presented to panellists in random order. Panellists were
instructed to rotate their finger at a defined rotation rate and dura-
tion for each test. They were asked to use the same finger to test
each sample and were subsequently instructed to clean fingers
between samples. Only one direction of rotation was permitted for
each panellist during the experiment. Prior to probing the perfor-
mance of the studied gels, the panellists were instructed to first cal-
ibrate the scale of the sensorial measurement by examining the
control(s).
Data analysis
Data were analysed by a two-factor (assessor, sample) analysis of
variance (ANOVA) test using ExcelÒ
software. The mean rating
and Fisher’s least significant differences for each term were calcu-
lated by ANOVA. Principle component analysis (PCA) of the mean
rating for each sensory attribute was used to visualize the relation-
ship between variables and samples using XLSTAT (Addinsoft, New
York, NY, U.S.A.).
Results and discussion
Rheological characterization
Dynamic oscillatory measurements
The strain amplitude dependency of dynamic material properties
were investigated using smooth-surfaced fixtures at 1, 10 and 20
rps frequency, and rough-surfaced fixtures at 1 rps frequency.
Material functions, such as the magnitude of complex viscosity
(g*), storage modulus (G¢(x)), loss modulus (G¢¢(x)), shear stress
and elastic stress (product of storage modulus and magnitude of
strain amplitude) [13, 14], of the 1% hydrogels were measured
over a strain amplitude (c0) range of 0.1–500%. All three hydro-
gels exhibited linear behaviour for a range of strain amplitudes up
to 1% for both low and high frequencies. However, at higher strain
amplitudes of 1% (or at high shear stress), the dynamic response
becomes non-linear and G¢, G¢¢ and g* decrease dramatically. Strain
sweep results have indicated that at low and high frequencies, G¢ is
higher than G¢¢, showing typical physical gel behaviour.
Strain sweep experiments were repeated at x = 1 rps with
plates, of which surfaces were covered with 400 grit adhesive-
backed waterproof sandpaper to eliminate slip. Figure 1a shows
that the modulus values started to decrease at a lower critical
strain, when compared to those shown in Fig. 1b, because of the
onset of slip. This trend is exacerbated in the 1% PAA/VP-XL,
where the slip effect is more pronounced. The yield stress values of
the three gels for slip and no-slip conditions, determined from the
maximum value of the elastic stresses (product of storage modulus
and magnitude of strain amplitude) at different frequencies, are
given in Table I. Table I shows that the calculated maximum elas-
tic stress values are frequency dependent (shear rate dependent)
because of shear thinning and/or thixotropic behaviour of the gels.
These values are also surface roughness–dependent because of the
onset of slip. In the low shear rate region (x = 1 rps), there is a
significant difference between the maximum elastic stress values
measured using smooth surfaces. In contrast, no significant differ-
ence between the gels was noted when rough surfaces were used.
The maximum elastic stress values increased with increasing fre-
quency for all samples.
The storage modulus in the linear viscoelastic region (LVR), G¢,
is another important parameter that reflects the strength and/or
ratio of the interactions among the polymer chains or swollen
cross-linked domains. The G¢ values of all three samples in the LVR
are given in Table II. Table II shows that, in the LVR, there was no
significant effect because of slip and that there was only a slight
increase in G¢ with increasing frequency in all samples.
(a) (b)
Figure 1 Strain amplitude dependency of the
storage modulus (G¢ (dyn cm)2
)) at 1 rps using
smooth surface fixtures (a) and rough surface f-
ixtures (b).
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201 195
Sensory correlation with rheology S. Ozkan et al.

4. Fourier transform analysis was also performed on the LAOS data
collected with rough and smooth surfaces at 1%, 200% and 400%
strain amplitude values. The sinusoidal stress response signal col-
lected from the sample was decomposed into elastic and viscous
stress contributions using symmetry arguments following the meth-
ods given by Ewoldt et al. (2008) [8]. Chebyshev polynomials were
calculated using MITlaos software, and the results are shown in
Tables III and IV [8]. One percent of strain amplitude was chosen
to represent the LVR, 200% strain amplitude to represent transition
region from linear to non-linear and 400% strain amplitude to rep-
resent the non-linear region of the material. Table III shows that
1% gels exhibit shear thickening / strain softening behaviour in
the LVR (1% strain amplitude). PAA/VP-XL gel exhibits shear thin-
ning / strain stiffening behaviour, whereas 1% PVM/MA-XL and
1% PAA-XL exhibit shear thickening / strain stiffening in the tran-
sition region (200% strain amplitude). All three gels exhibited
shear thinning / strain stiffening behaviour in the non-linear
region (400% strain amplitude). For smooth surfaces, all 1% gels
showed shear thickening / strain stiffening behaviour in the transi-
tion region (200% strain amplitude) and shear thinning / strain
stiffening behaviour in the non-linear region (400% strain ampli-
tude). The absolute values of all Chebyshev coefficients that were
measured with rough surfaces were different than the correspond-
ing coefficients measured with smooth surfaces, indicating the pres-
ence of wall slip effects on the measurements.
Steady torsional flow experiments
Steady torsional flow experiments with 20-mm parallel plates were
carried out at two different gap heights (1.0 and 1.5 mm) for each
different shear rate value. Each condition was repeated at least three
times with fresh samples. The results show that the shear stress val-
Table I Yield stress values determined from maximum elastic stress calcula-
tions for different frequency and surface conditions
1% PAA/
VP-XL
1% PVM/
MA-XL 1% PAA-XL
Maximum elastic stress,
Pa (x = 1 rps), smooth surface
29 139 136
Maximum elastic stress,
Pa (x = 10 rps), smooth surface
159 206 209
Maximum elastic stress,
Pa (x = 20 rps), smooth surface
191 246 259
Maximum elastic stress,
Pa (x = 1 rps), rough surface
169 175 164
Table II Storage modulus, G¢, values in the linear viscoelastic region for dif-
ferent frequency and surface conditions
1% PAA/VP-XL 1% PVM/MA-XL 1% PAA-XL
G¢, Pa (x = 1 rps),
smooth surface
859 788 565
G¢, Pa (x = 10 rps),
smooth surface
932 830 626
G¢, Pa (x = 20 rps),
smooth surface
923 884 641
G¢, Pa (x = 1 rps),
rough surface
851 765 550
Table III Chebyshev coefficients, which are calculated from large amplitude oscillatory flow (LAOS) data using MITlaos software. Experiments were conducted
at 1 rps frequency using smooth surface fixtures at 1%, 200% and 400% strain amplitudes
Strain amplitude % G¢, Pa G¢L/G¢M e3, Pa tand gL/gM v3, Pa s Physical attribute
1% PAA/VP-XL 400 29.51 3.79 10.38 2.68 0.83 )5.55 Shear thinning, strain stiffening
1% PVM/MA-XL 400 34.27 4.46 12.48 2.59 1.03 )4.21 Shear thinning, strain stiffening
1% PAA-XL 400 41.10 1.86 8.48 1.80 1.04 )2.56 Shear thinning, strain stiffening
1% PAA/VP-XL 200 65.78 1.79 11.7 1.8 0.95 )2.82 Shear thinning, strain stiffening
1% PVM/MA-XL 200 81.00 1.93 14.87 1.59 1.24 1.25 Shear thickening, strain stiffening
1% PAA-XL 200 84.50 1.40 9.54 1.23 1.25 1.70 Shear thickening, strain stiffening
1% PAA/VP-XL 1 843.2 0.91 )4.5 0.07 2.07 3.44 Shear thickening, strain softening
1% PVM/MA-XL 1 704.97 0.94 )1.79 0.08 1.89 1.70 Shear thickening, strain softening
1% PAA-XL 1 529.09 0.94 )1.58 0.08 1.87 1.56 Shear thickening, strain softening
Table IV Chebyshev coefficients, which are calculated from large amplitude oscillatory flow (LAOS) data using MITlaos software. Experiments were conducted
at 1 rps frequency using rough surface fixtures at 200% and 400% strain amplitudes
Strain amplitude, % G¢, Pa G¢L/G¢M e3, Pa tand gL/gM v3, Pa s Physical attribute
1% PAA/VP-XL 400 34.38 3.74 11.45 2.11 1.14 )2.56 Shear thinning, strain stiffening
1% PVM/MA-XL 400 37.06 4.08 12.57 2.35 1.10 )3.15 Shear thinning, strain stiffening
1% PAA-XL 400 43.64 1.79 8.48 1.65 1.12 )1.61 Shear thinning, strain stiffening
1% PAA/VP-XL 200 89.22 1.78 15.16 1.32 1.43 3.06 Shear thickening, strain stiffening
1% PVM/MA-XL 200 86.96 2.07 14.79 1.44 1.31 2.63 Shear thickening, strain stiffening
1% PAA-XL 200 90.01 1.28 8.29 1.05 1.32 3.32 Shear thickening, strain stiffening
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201196
Sensory correlation with rheology S. Ozkan et al.

5. ues increase with increasing gap separation at a constant apparent
shear rate. The dependence of the data on the gap separation sug-
gests sensitivity to the surface-to-volume ratio of the geometry being
utilized. This behaviour indicates the presence of wall slip in tor-
sional flow. A number of recent studies, which could be considered
benchmarks for the characterization of these types of materials,
showed that the steady-state simple shear behaviour can be very
well represented by the Herschel–Bulkley model. The yield stress of
these viscoelastic materials can be determined by the extrapolation
of the shear stress vs. shear rate data to zero shear rate by fitting
the data to the model after Rabinowitsch and slip corrections
[15–18]. Apparent torque and shear rate data were corrected for
non-linearity using the Rabinowitsch correction. Apparent shear
rate data were also corrected for wall slip in the low shear rate
region using the modified Mooney method according to the existing
literature [9–11]. H-B parameters, such as s0 (yield stress (Pa)), m
(consistency index (Pa s1/n
)), _c(corrected shear rate (1 s)1
)) and
n (power law index), are determined from fitting the
Rabinowitsch and slip-corrected steady torsional flow data
to the equation (1). The results are given in Table V.
Results show that the yield stress values determined from the H-B
fit are in the same range and order with the yield stress values
obtained from the maximum elastic stress calculations using
dynamic strain sweep data collected with rough surfaces (see
Tables I and V). The maximum elastic stress values are, however,
slightly higher than the yield stress values determined by the model.
We compared the results for 1% PAA-XL with the existing litera-
ture. Roberts and Barnes reported a study with Carbopol dispersions
taking slip phenomenon into account [19]. They used vane geome-
try with a slender gauze basket inserted inside the outer cylinder to
suppress wall slip and reported that the yield stress of 1% PAA-XL
as 115 Pa by fitting the data to the Herschel–Bulkley model. In this
study, we used a 20-mm stainless steel parallel plate working
against a TeflonÒ
-coated surface and fitted the slip-corrected data
to the same model. This laboratory predicted the yield stress value
of 1% PAA-XL as 124 Pa, indicating consistency with the existing
literature. It should be emphasized that water quality, differences in
neutralization and mixing processes, measurement protocols and so
on will affect the quality and precision of the results. As per Piau
et al., ‘The accuracy of quantification with complex fluids is much
poorer than with polymer solutions, and an overall accuracy of
10% can usually be considered as very good indeed’ [16].
Continuous stress or shear rate ramp measurements to establish the
shear rate dependency of shear stress and viscosity at different gap
openings
Continuous shear rate ramp tests were performed by ramping from
0–500 s)1
in 1 min at two different gap openings (1.0 and
1.5 mm) to determine the impact of the sample gap on the mea-
sured shear stress and viscosity. The results are given in Fig. 2. Fig-
ure 2a shows the impact of wall slip effects on the measurements
in the shear rate range between 0.06 s)1
and 2 s)1
. Slip velocity
increases up to 0.5 s)1
and then decreases with increasing shear
rate until it reaches 2 s)1
, the critical shear rate, where the mate-
rial becomes fluid. These findings are also in agreement with the
results given in previous literature [19 and 16]. Figure 2b clearly
shows the onset of wall slip effects on the measurements for 1%
PVM/MA-XL in the shear rate range between 0.05 s)1
and 5 s)1
.
Slip velocity increases to 0.1 s)1
and then begins decreasing with
increasing shear rate until it reaches 5 s)1
, the critical shear rate,
where the material becomes fluid. Figure 2c indicates that 1%
PAA/VP-XL measurements are affected by wall slip over a much
wider shear rate range than the other samples. The material slips
in the shear rate range between 0.01 s)1
and 10 s)1
, and the slip
velocity increases with increasing shear rate up to 0.02 s)1
, but
then subsequently tapers.
Yield stress values, or the critical stress values corrected for wall
slip (Table I and V), show that the critical stresses necessary to dis-
rupt the entire internal structure for all three materials are similar.
In contrast, under wall slip conditions, the gel samples exhibit sig-
nificantly different behaviour below the critical shear rate. The
Table V Herschel–Bulkley model parameters and Navier’s slip coefficients
1% PAA/VP-XL 1% PVM/MA-XL 1% PAA-XL
s0, Pa 161.5 168.5 123.7
m, Pa s1/n
12 22.8 41.5
n 0.54 0.52 0.43
b, m (Pa s1/nb
)nb
0.0033 0.141 0.024
sb 1.07 0.35 0.43
(a)
(b)
(c)
Figure 2 Continuous shear rate ramp results at 1.5-mm and 1.0-mm gap
openings for (a) PAA-XL, (b) PVM/MA-XL and (c) PAA/VP-XL.
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201 197
Sensory correlation with rheology S. Ozkan et al.

6. results indicate that the PAA/VP-XL possesses the highest slip
velocity and wall slip onsets at much lower shear stresses and
shear rate as compared to PVM/MA-XL and PAA-XL (Table I,
Fig. 2).
Extensional tests
The maximum force data collected during extensional testing were
used to glean information about each material’s cohesiveness. As
the intent is to mimic the ‘pick-up’ sensory evaluation test, the ini-
tial gap was kept as low as 400 microns, and the top plate was
raised with a relatively high speed (3 mm s)1
). After sample load-
ing, the material was equilibrated for 5 min prior to testing in
order to rebuild the internal structure. Each sample was tested at
least six times, and the results are given in Fig. 3 as averages of
maximum force values with 95% confidence intervals for each
sample. One-way ANOVA results indicate that 1% PAA/VP-XL gen-
erated significantly lower maximum forces than 1% PAA-XL,
whereas the 1% PVM/MA-XL forces were only directionally >1%
PAA/VP-XL. The lower extensional force of 1% PAA/VP-XL may
be attributed to increases in the surface-to-volume ratio, possibly
due to wall slip effects, shear banding and/or thixotropy.
Sensory evaluation
Primary skin-feel parameters, such as pick-up, cushion/body/firm-
ness, initial spreadability and secondary skin-feel parameters, such
as rub-out spreadability and slipperiness / lubrication, were evalu-
ated by six selected naı¨ve panellists. The purpose of the study was
to investigate the correlation of sensory attributes with the mea-
sured rheological properties of the three thickeners. At the initial
stage of the study, gels were characterized at 25°C by placing the
sample on a skin simulant surface, rather than on the forearm of
the panellist. This was performed to eliminate transient structural
changes of the materials during testing, because of temperature,
electrolyte effects and pH differences of the skin surface. Even
though trained panellists were not used, very explicit instructions
were prepared to make sure that each panellist would work at the
same shear rate range during the evaluation of each attribute.
ANOVA analysis conveyed that variations between assessors were
significant for cushion, initial spreadability and slipperiness ratings,
but were not significant for rub-out spreadability and pick-up
ratings. On the other hand, differences between samples were
significant for all attributes except rub-out spreadability. ANOVA
analysis showed that panellists used different parts of the scale to
express their perceptions, but were consistent in ranking and differ-
entiating their differences in general.
In PCA of the sensory evaluation data for the three different gel
samples, the first two principle components accounted for 81% and
19% of the variance, respectively. Sensory ratings of the three gel
samples plotted for the first two principle components are shown in
Fig. 4. One percent PAA/VP-XL scored to the left side of F1, show-
ing low values of cushion, slipperiness and pick-up ratings, but
high values of initial and rub-out spreadability ratings. One percent
PAA-XL scored on the right side of F1, contrasting with 1% PAA/
VP-XL, showing high values of cushion, slipperiness and pick-up
ratings, but low values of initial and rub-out spreadability ratings.
One percent PVM/MA-XL scored low in F2, showing low values of
all attribute ratings. The correlation between sensory attributes
showed that cushion, slipperiness and pick-up are related, whereas
initial and rub-out spreadability are related, but are in contrast
with cushion, slipperiness, and pick-up.
Correlation of sensory ratings with conventional rheological
parameters
Principle component analysis analysis was applied to the rheologi-
cal parameters and sensory evaluation ratings data together, and
the results are in Fig. 5. The first two principle components
accounted for 76.01% and 23.99% of the variance, respectively.
The correlation between the sensorial and rheological material
parameters indicates that pick-up, slipperiness and cushion are
related to each other and also to the consistency index (m) from
the H-B fit, as well as the maximum normal force (MNF) values
obtained from extensional experiments. However, pick-up, slipperi-
ness and cushion are in contrast with the parameter group, includ-
ing the power law index of H-B fit (n), gel strength values in the
LVR measured with smooth surfaces at 1, 10 and 20 rps frequency
values (G¢ S (x = 1, 10, 20 rps) and rough surfaces at 1 rps fre-
quency (G¢ R(x = 1 rps)). Another related parameter group formed
by initial spreadability, rub-out spreadability and slip power low
index (s) is in the negative side of both F1 and F2. This group of
parameters is in contrast with the elastic stress values measured
with smooth surfaces (ESS (x = 1, 10 and 20 rps)) and the shear
viscosity values measured at 10, 100 and 500 s)1
shear rate val-
ues (SV (@10, 100 and 500 s)1
)).
Figure 3 Average maximum force values measured during extensional tests
for 1% PAA/VP-XL, 1% PVM/MA-XL and 1% PAA-XL (*P < 0.009). Figure 4 Sensory data: Principle component analysis.
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201198
Sensory correlation with rheology S. Ozkan et al.

7. Correlation of sensory ratings with large amplitude oscillatory
shear flow (LAOS) parameters (Chebyshev coefficients)
Principle component analysis was applied to the Chebyshev coeffi-
cients given in Tables III and IV and to the set of sensory evalua-
tion ratings data in Fig. 6. The first two principle components
accounted for 65.72% and 34.28% of the variance, respectively.
The correlation between the sensorial attribute ratings and Cheby-
shev coefficients demonstrates that pick-up, slipperiness and cush-
ion are related to the alternative modulus measured with rough
and smooth surfaces at 400% strain (G¢ CR4 and G¢ CS4) and to
the v3 coefficient calculated from the data measured with smooth
surfaces at 400% strain (v3 CS4). The parameter v3 is indicative of
the shear thickening or shear thinning behaviour of the material
(v3 > 0 shear thickening, v3 < 0 shear thinning) [8]. This parame-
ter group is in contrast with the related parameter group, including
the alternative modulus, G¢, measured in the LVR at 1% strain
with smooth surfaces (G¢ CS001) and the tand value measured in
the transition region with smooth surfaces at 200% strain (tand
CS2).
Another related parameter group includes initial spreadability,
rub-out spreadability and g¢L/g¢M values measured with rough sur-
faces in the transition and non-linear regions (g¢L/g¢M CR2 and g¢L/
g¢M CR4). These parameters are positioned on the positive side of
both F1 and F2. This group contrasts the g¢L/g¢M values measured
with smooth surfaces in the transition and non-linear region (g¢L/
g¢M CS2 and g¢L/g¢M CS4), G¢L/G¢M values measured in the LVR
using smooth surfaces at 1% strain amplitude (G¢L/G¢M CS001) and
e3 values measured in the LVR using smooth surfaces at 1%
strain amplitude (e3 CS001). The e3 values are indicative of strain
Figure 5 Sensory and conventional rheological
parameter data together: Principle component
analysis.
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201 199
Sensory correlation with rheology S. Ozkan et al.

8. stiffening or softening behaviour of the material (e3 > 0 strain stiff-
ening, e3 < 0 strain softening) [8]. The strain stiffening or strain
softening behaviour of the material represents the increase or
decrease of the resistance of the material to the deformation in the
extensional direction and, for some cases, may be different than the
material’s response to shear deformation. For instance, random-
coiled high-molecular polymer chains may exhibit much higher
extensional viscosity than their shear viscosity with increasing
shear or strain rates. Therefore, a material can be shear thinning
(exerts less resistance to shear flow with increasing shear rate), but
at the same time strain stiffening (exerts higher resistance to exten-
sional flow with increasing strain rate) [20, 21].
The case can be made that the wall slip is expected to show its
effect in the even harmonics of the LAOS data, which can be
extracted from the Fourier amplitude spectrum. For this reason, we
have included the relative intensity of the second harmonic (I2/I1,
the ratio of the second harmonic to the principle harmonic) to
investigate the effect of slip on the even harmonics and their corre-
lation with sensory attributes. The results are given in Fig. 7. Fig-
ure 7a shows the strong correlation between the slip velocity and
the even harmonics of the material. Figure 7b,c shows the strong
correlation between the relative intensity of the even harmonics
and spreadability ratings, indicating the effect of slip on the percep-
tion of spreadability of the material. These results also confirm the
effect of slip on the rheological parameters and sensory perception
of the material.
Conclusions
Rheological methods can be successfully applied to objectively and
quantitatively describe sensory attributes of thickeners if necessary
attention is paid in choosing the right rheological methods. The
occurrence of wall slip and thixotropy may contribute to the sen-
sory perception of hydrogel-based personal care products and
should be characterized. We have determined that a significant
correlation can be made between slip velocity and the initial and
rub-out spreadability of the hydrogels. The applied shear rate
range may contribute to the material’s response to a given defor-
mation and to the sensorial perception of the product. The correla-
tion between sensory attributes showed that cushion, slipperiness
and pick-up are related, whereas initial and rub-out spreadability
are related, but are in contrast with cushion, slipperiness and
pick-up. These trends are in good agreement with existing litera-
ture.
Using FT analysis in LAOS can be effective in correlating sensory
rating results in skin/hair gels. Results indicate that the surface
Figure 6 Sensory and LAOS analysis data
together: Principle component analysis.
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201200
Sensory correlation with rheology S. Ozkan et al.

9. roughness of the plates, and choosing to test in the linear, transi-
tion or non-linear region, will determine which LAOS analysis
parameters will correlate best with which sensory parameters. This
indicates that wall slip and thixotropy have to be taken into
account when correlating LAOS analysis parameters. With these
precautions, rheological techniques, and LAOS in particular, can be
considered an exciting way to make inroads into sensorial percep-
tion analysis.
Acknowledgements
We thank Dr. Dilhan M. Kalyon of Stevens Institute of Technology for
his discussions regarding slip correction, Dr. Gareth H. McKinley of
MIT and Dr. Randy H. Ewoldt of University of Illinois for providing the
MITlaos software and for their guidance regarding LAOS analysis,
and Dr. Aloyse Franck of TA Instruments for his valuable comments
and help with analysing the LAOS data. We are grateful to the review-
ers for their time and comments that helped improve this study.
References
1. Stanley, N.L. and Taylor, L.J. Rheological
basis of oral characteristics of fluid and
semi-solid foods: a review. Acta Psychol. 84,
79–92 (1993).
2. Chen, J. Food oral processing – a review.
Food Hydrocolloids 23, 1–25 (2009).
3. Ciron, C.I.E., Gee, V.L., Kelly, A.L. and Auty,
M.A.E. Effect of microfluidization of
heat-treated milk on rheology and sensory
properties of reduced fat yoghurt. Food
Hydrocolloids 25, 1470–1476 (2011).
4. Janhøj, T., Frøst, M.B. and Ipsen, R. Sensory
and rheological characterization of acidified
milk drinks. Food Hydrocolloids 22, 798–806
(2008).
5. Loret, C., Walter, M., Pineau, N., Peyron,
M.A., Hartmann, C. and Martin, N. Physical
and related sensory properties of a swallowable
bolus. Physiol. Behav. 104, 855–864 (2011).
6. Holm, K., Wendin, K. and Hermansson,
A.M. Sweetness and texture perception in
mixed pectin gels with 30% sugar and a
designed rheology. LWT – Food Sci. Technol.
42, 788–795 (2009).
7. Fischer, P. and Windhab, E.J. Rheology of
food materials. Curr. Opin. Colloid Interface
Sci. 16, 36–40 (2011).
8. Ewoldt, R.H., Hosoi, A.E. and McKinley,
G.H. New measures for characterizing non-
linear viscoelasticity in large amplitude
oscillatory shear. J. Rheol. 52, 1427–1458
(2008).
9. Yoshimura, A. and Prud’homme, R.K.
Wall slip corrections for couette and par-
allel disk viscometers. J. Rheol. 32, 53–67
(1988).
10. Kalyon, D.M., Yaras, P., Aral, B. and Yilm-
azer, U. Rheological behavior of a concen-
trated suspension: a solid rocket fuel
simulant. J. Rheol. 37, 35–53 (1993).
11. Kalyon, D.M. Apparent slip and viscoplastic-
ity of concentrated suspensions. J. Rheol.
49(3), 621–640 (2005).
12. Meilgaard, M., Civille, G.V. and Thomas
Carr, B. Sensory Evaluation Techniques, 3rd
edn. CRC Press, Boca Raton (1999).
13. Yang, M.C., Scriven, L.E. and Macosko,
C.W. Some rheological measurements on
magnetic iron oxide suspensions in silicone
oil. J. Rheol. 30, 1015–1029 (1986).
14. Pai, V.B. and Khan, S.A. Gelation and rheol-
ogy of xanthan/enzyme-modified guar
blends. Carbohydr. Polym. 49, 207–216
(2002).
15. Meeker, S.P., Bonnecaze, R.T. and Cloitre,
M. Slip and flow in pastes of soft particles:
direct observation and Rheology. J. Rheol.
48, 1295–1320 (2004).
16. Piau, J.M. Carbopol gels: Elastoviscoplastic
and slippery glasses made of individual
swollen sponges Meso- and macroscopic
properties, constitutive equations and
scaling laws. J. Nonnewton. Fluid Mech. 144,
1–29 (2007).
17. Seth, J.R., Cloitre, M. and Bonnecaze, R.T.
Influence of short-range forces on wall-slip
in microgel pastes. J. Rheol. 52, 1241–1268
(2008).
18. Coussot, P., Lanos, C., Chateau, X. and
Ovarlez, G. Macroscopic vs. local rheology of
yield stress fluids. J. Nonnewton. Fluid Mech.
158, 85–90 (2009).
19. Roberts, G.P. and Barnes, H.A. New mea-
surements of the flow curves for Carbopol
dispersions without slip artifacts. Rheol. Acta
40, 499–503 (2001).
20. Onck, P.R., Koeman, T., van Dillen, T. and
van der Giessen, E. Alternative explanation
of stiffening in cross-linked semiflexible net-
works. Phys. Rev. Lett. 95, 178102 (2005).
21. Gardel, M.L., Shin, J.H., MacKintosh, F.C.,
Mahadevan, L., Matsudaira, P. and Weitz,
D.A. Elastic behavior of cross-linked and
bundled actin networks. Science 304, 1301–
1305 (2004).
(a) (b) (c)
Figure 7 Linear regression fit for relative intensity (I2/I1) vs. slip velocity coefficient, s, collected with smooth surfaces at 400% strain (a), and sensory ratings
vs. relative intensity (I2/I1) data collected with smooth surfaces at 400% strain (b and c).
ª 2012 ISP Investments Inc
ICS ª 2012 Society of Cosmetic Scientists and the Socie´te´ Franc¸aise de Cosme´tologie
International Journal of Cosmetic Science, 34, 193–201 201
Sensory correlation with rheology S. Ozkan et al.