gels

Article
Effect of Shear History on Rheology of
Time-Dependent Colloidal Silica Gels
Paulo H. S. Santos 1, * ID
, Marcelo A. Carignano 2 and Osvaldo Campanella 1, *
1 Department of Agricultural and Biological Engineering, Purdue University, 225 South University Street,
West Lafayette, IN 47907, USA
2 Qatar Environment and Energy Research Institute, Hamad bin Khalifa University,
P.O. Box 34110 Doha, Qatar; [email protected]
* Correspondence: [email protected] (P.H.S.S.); [email protected] (O.C.)

Received: 10 October 2017; Accepted: 16 November 2017; Published: 20 November 2017

Abstract: This paper presents a rheological study describing the effects of shear on the flow curves
of colloidal gels prepared with different concentrations of fumed silica (4%, 5%, 6%, and 7%) and a
hydrophobic solvent (Hydrocarbon fuel, JP-8). Viscosity measurements as a function of time were
carried out at different shear rates (10, 50, 100, 500, and 1000 s−1 ), and based on this data, a new
structural kinetics model was used to describe the system. Previous work has based the analysis
of time dependent fluids on the viscosity of the intact material, i.e., before it is sheared, which is a
condition very difficult to achieve when weak gels are tested. The simple action of loading the gel
in the rheometer affects its structure and rheology, and the reproducibility of the measurements is
thus seriously compromised. Changes in viscosity and viscoelastic properties of the sheared material
are indicative of microstructural changes in the gel that need to be accounted for. Therefore, a more
realistic method is presented in this work. In addition, microscopical images (Cryo-SEM) were
obtained to show how the structure of the gel is affected upon application of shear.

Keywords: gel; silica; rheology; thixotropy; microstructure; viscosity; JP-8; fuel

1. Introduction
For a wide range of complex fluids such as gels, their flow properties are not a function of shear
rate alone [1]. Depending on the type of gel, temperature, and concentration, among other conditions,
gels can exhibit either solid or liquid viscoelastic behavior, which is commonly found in materials,
notably colloidal gels, that can form networks due to the presence of non-covalent bonds. A challenge
during the testing of gels is that flow and deformation can induce reversible and irreversible structural
changes, making them time or deformation dependent fluids [2].
Changes in the viscosity or viscoelastic properties of colloidal gels as a function of time, flow,
and deformation are relevant in several industrial processes and product formulations. For example,
the texture and sensorial characteristics of some food products depend on the processing steps and
on the shear history to which the product is subjected during manufacturing [3–5]. The change
with time of the rheological properties of pharmaceutical formulations may influence the controlled
release of loaded drugs, defining their therapeutic efficacy and bioavailability [6]. In colloid science,
the decrease of viscosity of gels and suspensions of colloidal particles when shear is applied make
them key components in mining, ceramic, and coating industries [7–11]. Moreover, time dependent
colloidal gels are an emerging alternative in biotechnology for cell culture [12] and in rocket propulsion
systems [13–15].
The reversible time dependent rheological phenomenon of fluids in which the viscosity of these
materials decreases with time and/or deformation is referred to as Thixotropy, whereas the opposite

Gels 2017, 3, 45; doi:10.3390/gels3040045 www.mdpi.com/journal/gels

Gels 2017, 3, 45 2 of 12

phenomenon is known as Rheopexy. For Thixotropic fluids, their viscosity decreases with time when a
constant shear is applied; however, their viscosity may be fully or partially recovered when the shear is
ceased [16]. The microstructural changes and mechanisms that govern this transition are quite complex
and not very well understood, especially in colloidal systems [12–19]. Colloidal suspensions can
aggregate to form a percolated network or gel when the system is destabilized [20–23]. The resulting
network formed by weak, mostly non-covalent, attractive forces gives the system solid-like viscoelastic
properties. Most often, the thixotropic behavior of these fluids is related to changes in their network
which is “broken” by mechanical stresses and the rheology of the gel transitions to that of a viscoelastic
liquid or a paste. The reduction or absence of the applied shear often allows the system to re-build the
particulate network and consequently to recover its initial conformation and viscosity.
Irreversible or partial microstructure changes may also occur in colloidal systems [2]. Fumed silica,
which is applied in many industrial activities, can form distinctive 3D gel networks at relatively low
concentrations. The surface of these colloidal particles plays an important role in the formation of
gels: the presence of hydrophilic groups makes the particles capable of building hydrogen bonds
between themselves to create a percolated network [24]. Because of their hydrophilic properties,
a strong particle-particle interaction is observed in the presence of hydrophobic solvents, resulting in
the formation of a gel network at low concentrations of particles. In contrast, a higher particle
concentration is needed for gel formation if the solvent has hydrophilic characteristics. In this case,
there is competition between particle-particle and particle-solvent interactions. In addition to the nature
of the solvent, other parameters may affect the rheological properties of the system, including the
concentration of particles, degree of dispersion, and temperature. Previous reports showed that these
may also affect the rheological time dependence of colloidal gels [13].
Although time dependent fluids have been an object of study for many years, there are still
few extensive accurate data available in the literature, in particular, for gels with a very fragile
structure that are unquestionably disturbed when loading them in the rheological testing instrument.
Thus, the characterization and a suitable rheological representation of these systems are notoriously
challenging and experimental data in a wide range of conditions would be extremely helpful in solving
open issues in thixotropic systems [25–27].
The objective of this work was to investigate and describe the macroscopic behavior of time
dependent silica gels as a consequence of microstructural changes induced by shear. Viscosity data
as a function of time was used to model irreversible flow property changes by incorporating a newly
defined structural decay parameter that can be included in models used to describe time independent
flow curves of fluids; for example, when studying the extended Herschel-Bulkley rheological model
used in this work. In addition, qualitative data showing how the shear history affects viscoelastic
parameters such as normal stresses measured during steady shear flow, as well as cryo-scanning
electron micrographs (Cryo-SEM) of sheared materials, are presented.
Related work studying the rheology of these materials clearly shows that the shear history is a
key parameter to be considered in the analysis because it strongly affects the properties of the gels,
to the extent that it is recommended to pre-shear the material before testing [28]; this reinforces the
importance of the studies presented in this manuscript.

2. Theory
Different approaches have been used to incorporate time dependence into rheological models.
Basically, they can be divided into three classes: the continuum mechanics approach, and the use of
structural kinetics and microstructural models. The first approach is not used in this work; instead,
it focuses on a combined approach involving a structural kinetic model where the shear history effect
and the material breakdown is incorporated by a microstructural model defined by the parameter λ
that expresses the degree of structure of the material. Most models applied to time dependent fluids,
such as the Tiu and Boger [29] model, also use a structural parameter, denoted by λ. The value of
λ varies from a value of λ0 , indicating that the material is initially “intact” and its structure is fully
Gels 2017, 3, 45 3 of 12

developed at the initial time t = 0, to an equilibrium value (λe ) that corresponds to a completely
broken structure due to flow or deformation. In a more recent work, Denn and Bonn [30] proposed
the description of the thixotropic behavior of complex fluids, such as gels, by a parameter that is a
connectivity of the material structure with its flow properties.
Our preliminary work [13] has shown that for rheologically highly time dependent gels, such as
those studied in the present study, it is extremely difficult, if not impossible, to obtain a flow curve
that represents the rheology of the intact gel system. As the sample is highly affected by shear,
the flow curve obtained describes a transient state of the fluid that depends on the measurement and
the instrument loading conditions. As discussed, the simple action of loading the sample into the
rheometer is enough to disturb its structure, which can seriously compromise the reproducibility of
the measurements. To avoid that source of error, in the present study, the flow curve of the completely
broken gel is used as the reference curve to define the structural parameter in the constitutive equation.
In a way, the approach to assess the rheology of these gel materials is taking into consideration practical
advises concerning pre-shearing the material before testing. Thus, samples are pre-sheared to eliminate
time effects on the data and the flow curve of the broken system is measured and described by a more
suitable rheological model. The extended Herschel-Bulkley model was used to describe the flow curve
.
of the material expressed as shear stress τ as a function of the shear rate γ of the fully broken gel,
i.e., when the structure of the gel is broken, and the gel behaves as a rheologically time independent
material described by the following equation [13]:
.n .
τ = τ0 + K γ + η∞ γ (1)

Therefore, to describe a gel with time dependent rheological properties, the structural parameter
.

λ γ, t , which depends on the shear rate and time, is incorporated into the model as:

.
.
.n .
h i
τ γ, t = λ γ, t · τ0 + K γ + η∞ γ (2)

where τ0 is the fluid yield stress, K is the consistency index, n is the power law index, and η∞ is
the equilibrium viscosity, estimated as the viscosity of the gel when large shear rates are applied.
In this work, the viscosity was determined by performing a measurement in a high shear rate
capillary viscometer.
.

The structural parameter λ γ, t characterizes the changes of the gel structure with shear rate and
time, and it is assumed to follow a second order kinetic equation:
.

dλ γ, t .
 .
2
= −c γ λ γ, t − λe (3)
dt
.
.

where c γ is a shear rate depending parameter, whereas λ γ, t is a parameter that describes changes
of the gel structure with time at a specific shear rate. It varies from an initial value λ0 , representing
the initial structure of the gel, to an asymptotic value λe , which will eventually be equal to 1 when
the structure of the material is completely broken, i.e., when it can no longer be altered. However,
in subsequent derivations, it will be left as λe for the sake of generalization.
Equation (3) can be integrated from an initial time t = 0, when the structural parameter is λ0 , to a
time t under which the structural parameter reaches the value λ(t) to yield:
.

.
λ0 + λ e · c γ · t · ( λ0 − λ e )
λ γ, t = .
(4)
1 + c γ · t · ( λ0 − λ e )
.
.

Thus, provided that the parameter c γ is known, the variation of the structural parameter λ γ, t
.

of the gel can be determined by Equation (4). Determination of the parameter c γ can be performed
by measuring the viscosity of the gel system with a function of time at fixed shear rates.
Gels 2017, 3, 45 4 of 12

.
.
τ (γ,t)
From Equation (2) and the definition of an apparent viscosity, η γ, t = .
γ
, the following
equation is obtained:
.
.
.
η γ, t γ
λ γ, t = h .n .
i (5)
τ0 + K γ + η∞ γ
.

As noted in Equation (5), the apparent viscosity of the gel η γ, t is a function of the shear rate
. .

γ and time t and accounts for the changes in the structural parameter λ γ, t . By differentiating
Equation (5) with respect to time and combining it with Equation (3), the following relation is obtained:
.
.n
h i
.
c γ · τ0 + K γ + η∞ · [λ(t) − λe ]2


dη γ, t
= . (6)
dt γ

Equation (6) can be rearranged as:

.

dη γ, t .
 .
2
= − a γ η γ, t − ηe (7)
dt
where: .
.
.
c γ γ
a γ = h .n .
i (8)
τ0 + K γ + η∞ γ

and .

τ0 + K γn + η∞
" #
.
.
η γ, t = λ γ, t . (9)
γ
A typical schematic thixotropic behavior of a rheological time dependent gel after the application
of constant shear rates of 50 1/s and 100 1/s is depicted in Figure 1a. By integrating Equation (6)
between t = 0 and an arbitrary time t, the following equation is obtained:

1 1 .

. = +a γ t (10)
η0 − η e


η γ, t − ηe

1 .

A plot of . versus t, at a constant shear rate, gives a straight line with slope a γ ,
η (γ,t)−ηe
.

from which the value of c γ can be calculated from Equation (8), provided the rheological model of
the sheared unstructured gel is known. Once the relevant parameters are known, Equation (4) can be
.

used to estimate the variation of structural parameter λ γ, t at different times and shear rates and
thus predict the corresponding value of the viscosity by Equation (9).
Gels 2017, 3, 45 5 of 12
Gels 2017, 3, 45 5 of 12

2.0
Shear
shear rateRate
50 s -150 1/s

Viscosity (Pa.s)
1.0 Shear
shear raterate
100100s -1 1/s

a)

0.1
0 20 40 60 80 100 120

50
shear
shear rate
rate 50 50
s -1 1/s
40 Shear rate
100100
s -1 1/s
1/(η(t)-ηe)

shear rate
30
b)
20
10

0
0 20 40 60 80 100 120
time
Time (s)
(seconds)
Figure 1. (a) Typical thixotropic behavior of a rheological time dependent gel after the application
Figure 1. (a) Typical thixotropic behavior of a rheological time dependent gel after the application of
of constant shear rates of 50 1/s and 100 1/s; (b) schematic of the approach used to determine the
constant shear rates of 50 1/s and 100 1/s; (b) schematic
.
of the approach used to determine the
parameters defining the structural parameter λ γ, t from Equation (10). Data used are from a 5%
parameters defining the structural parameter λ (γ,t ) from Equation (10). Data used are from a 5%
silica gel.
silica gel.
3. Materials and Methods
3. Materials and Methods
3.1. Gel Preparation
3.1. Gel gels were prepared by adding fumed silica (CAB-O-SIL® , M-5, Cabot Corporation,
Preparation
Colloidal
Boston, MA, USA) with an average particle size of 0.2–0.3 µm and a surface area of 200 m2 /g to a
Colloidal gels were prepared by adding fumed silica (CAB-O-SIL®, M-5, Cabot Corporation,
hydrophobic solvent (hydrocarbon fuel, kerosene grade JP-8, NATO code F-34, minimum density of
Boston, MA, USA) with an average particle size of 0.2–0.3 µm and a surface area of 200 m2/g to a
775.0–840 kg/m3 at 15 ◦ C; minimum flash point of 37–38 ◦ C; boiling point 157–300 ◦ C; freezing point
hydrophobic solvent (hydrocarbon fuel, kerosene grade JP-8, NATO code F-34, minimum density of
of −47 ◦ C; maximum aromatics content of 25.0% v/v and maximum sulphur content of 0.30% mass).
775.0–840 kg/m3 at 15 °C; minimum flash point of 37–38 °C; boiling point®157–300 °C; freezing point
Samples were mixed using an acoustic mixer (Resodyn Resonant Acoustic LabRAM, Butte, MT, USA)
of −47 °C; maximum aromatics content of 25.0% v/v and maximum sulphur content of 0.30% mass).
to form the network. The acceleration of the mixer for the gel preparation was approximately 100 G
Samples were mixed using an acoustic mixer (Resodyn Resonant Acoustic® LabRAM, Butte, MT,
for 60 s. The mixing conditions were based on the flow properties and stability of previously prepared
USA) to form the network. The acceleration of the mixer for the gel preparation was approximately
gels [31]. The rheological experiments were conducted within 24–48 h after gels were made.
100 G for 60 s. The mixing conditions were based on the flow properties and stability of previously
prepared gels [31]. The rheological experiments were conducted within 24–48 h after gels were made.
3.2. Microscopy
After the application of a controlled shear in a rheometer, the gel was immersed into liquid
3.2. Microscopy
nitrogen and transferred to the Gatan Alto 2500 pre-chamber (cooled to −170 ◦ C). The samples were
After the application of a controlled shear in a rheometer, the gel was immersed into liquid
directly transferred to the microscopy unit within a 30 min period for testing. Samples were imaged
nitrogen and transferred to the Gatan Alto 2500 pre-chamber (cooled to −170 °C). The samples were
with an FEI Quanta 3D field emission scanning electron microscope (FEI company, Hillsboro, OR,
directly transferred to the microscopy unit within a 30 min period for testing. Samples were imaged
USA) using the ET (Everhart-Thornley) detector operating at a 5 kV accelerating voltage, ~6 mm
with an FEI Quanta 3D field emission scanning electron microscope (FEI company, Hillsboro, OR,
working distance, and 30 µm aperture. After fracturing the sample with a cooled scalpel to produce a
USA) using the ET (Everhart-Thornley) detector operating at a 5 kV accelerating voltage, ~6 mm
free-break surface, the sample was not sublimated since it did not contain water. The samples were
working distance, and 30 µm aperture. After fracturing the sample with a cooled scalpel to produce
sputter coated for 90 s with platinum and then transferred to the microscope cryostage (−130 ◦ C)
a free-break surface, the sample was not sublimated since it did not contain water. The samples were
for imaging. Time elapsed between the application of shear to the gels and the preparation of the
sputter coated for 90 s with platinum and then transferred to the microscope cryostage (−130 °C) for
micrographs (about 30 min) was sufficient to assume that the gel structure had not changed. This was
imaging. Time elapsed between the application of shear to the gels and the preparation of the
confirmed by measuring the sample viscosity after 30 min resting and finding that it had not changed,
micrographs (about 30 min) was sufficient to assume that the gel structure had not changed. This was
and the same was true for the sample structure.
confirmed by measuring the sample viscosity after 30 min resting and finding that it had not changed,
and the same was true for the sample structure.
Gels 2017, 3, 45 6 of 12

3.3. Rheological Characterization

3.3. Rheological Characterization
Rheological characterization was carried out in triplicate using a rotational AR-G2 Rheometer
Rheological
with Smart SwapTM characterization
Geometry (TA was carried outNew
Instruments, in triplicate using
Castle, DE, a rotational
USA). A cone andAR-G2plateRheometer
geometry
with
was used and all measurements were performed at a constant temperature of 25 °C. plate
Smart SwapTM Geometry (TA Instruments, New Castle, DE, USA). A cone and geometry
This geometry
was used anand all measurements ◦
provided almost uniform shearwerefield.performed at a constant
Slip is an artifact temperature
that is often of 25 during
unavoidable C. Thisrheological
geometry
provided an almost uniform shear field. Slip is an artifact that is often unavoidable
measurements; however, the samples were sticky and slippage did not appear to be a major problem during rheological
measurements;
with these samples. however, the samples
A parallel plate were stickycould
geometry and slippage
help to did not whether
assess appear toorbenot a major problem
slippage was
with these samples. A parallel plate geometry could help to assess whether or not slippage
present in the measurements, but since this geometry does not provide an uniform shear field, the was present
in
usethe
ofmeasurements,
the cone and platebut since this geometry
geometry does not
was preferred toprovide
eliminatean potential
uniform shear field, related
problems the use to
of the
the
cone
interference of the yield stress in the measurements [30], which could be significant in this type of
and plate geometry was preferred to eliminate potential problems related to the interference of
the yield stress in the measurements [30], which could be significant in this
gel. A capillary rheometer (RH2000, Bohlin Instruments, Cirencester, UK) was used to obtain type of gel. A capillary
rheometer
viscosities (RH2000,
at a shearBohlin Instruments,
rate >1000 s−1 and the Cirencester,
value of ηUK) . was used to obtain viscosities at a shear rate
∞
− 1
>1000 s and the value of η∞ .
4. Results
4. Results and
and Discussion
Discussion
Three flow
Three flow curves
curves were
were measured
measured for for each
each sample
sample in in aarange
rangeof ofshear
shearrates
ratesfrom
from11toto1000
1000ss−−11..
Viscosityvalues
Viscosity valueswere
were first
first obtained
obtained by applying
by applying an increasing
an increasing shear
shear rate rate
ramp, ramp, by
followed followed by a
a decreasing
decreasing
shear shear(1000
rate ramp rate ramp
to 1 s− (1000 to 1finally
1 ), and s ), and
−1 finally by
by another another increasing
increasing ramp (1 toramp
1000 (1
s−to 1000 Figure
1 ) test. s ) test.
−1
2
Figure 2 presents the data obtained for a 6% gel. Typical time dependent materials
presents the data obtained for a 6% gel. Typical time dependent materials exhibit different flow curves exhibit different
flow curvesa whenever
whenever history ofapre-shear
history ofexists,
pre-shear
andexists,
the areaandofthe
thearea of the hysteresis
hysteresis loop showsloopevidence
shows evidence
of time
of time dependence in the fluid. A significant difference is observed between the
dependence in the fluid. A significant difference is observed between the first ramp up and the ramp
first ramp up and
the ramp down measurements. Therefore, the shear forces applied to the sample during the first ramp
down measurements. Therefore, the shear forces applied to the sample during the first ramp up
up appear to be high enough to completely break down the gel network. No hysteresis loop was
appear to be high enough to completely break down the gel network. No hysteresis loop was observed
observed when the second ramp up and a third ramp down were compared. In addition, no viscosity
when the second ramp up and a third ramp down were compared. In addition, no viscosity recovery
recovery was observed. The gel structure is originally formed after the silica particles are highly
was observed. The gel structure is originally formed after the silica particles are highly dispersed
dispersed in the solvent, which allows the formation of hydrogen bonds between the particles leading
in the solvent, which allows the formation of hydrogen bonds between the particles leading to the
to the development of the percolated network. The shear applied may affect this network and
development of the percolated network. The shear applied may affect this network and consequently
consequently the gel rheology. Letting the system rest by itself did not recover its structure, at least
the gel rheology. Letting the system rest by itself did not recover its structure, at least within the time
within the time scale investigated (up to 30 min). This can be understood by considering that energy
scale investigated (up to 30 min). This can be understood by considering that energy is needed for the
is needed for the initial formation of the gel (applied by the acoustic mixer) and the thermal motion
initial formation of the gel (applied by the acoustic mixer) and the thermal motion of the particles was
of the particles was not enough to re-build the network.
not enough to re-build the network.
1000

Ramp up
100 Ramp down
Ramp up (2)
Viscosity (Pa.s)

10

0.1

0.01
1 10 100 1000
Shear Rate (1/s)

Figure 2.
Figure 2. Viscosity
Viscosity measurements
measurements ofof 6%
6% gel.
gel. Three
Three runs
runs were
were carried out using
carried out using the
the same
same sample:
sample: shear
shear
rate was
rate was increased
increased from
from11to
to1000 −
1000ss (ramp
1
−1
(ramp up),
up), varied from 1000
varied from 1000 to −
to 11 ss (ramp
−11 (rampdown),
down), and
and increased
increased
from 1 to 1000 s−−11(ramp up 2).
(ramp up 2).

The results indicate that the combined effect of time and shear induces irreversible changes for
all concentrations. This allows the determination of the flow curves of the “broken” sheared gel
Gels 2017, 3, 45 7 of 12

(Figure 3) and the use of a time independent rheological model to describe the system. For that, the
extended Herschel-Bulkley model
Gels 2017, 3, 45
τ = τ 0 + Kγ n + η∞γ was applied to describe their flow
7 of 12
characteristics and to obtain the parameters reported in Table 1. Note that η∞ has been experimentally
measured by a capillary rheometer at shear rates >10,000 s−1.
The results indicate that the combined effect of time and shear induces irreversible changes for all
Few observations should be highlighted to show the strong time and shear dependence of these
concentrations. This allows the determination of the flow curves of the “broken” sheared gel (Figure 3)
gels. First, significantly lower viscosity values were obtained for pre-sheared gels (up to 60 times
and the use of a time independent rheological model to describe the system. For that, the extended
lower). Second, the power law indexes . n for the
. pre-sheared gels were in the range 0 < n < 1, which
Herschel-Bulkley model τ = τ0 + K γ + η∞ γ was applied to describe their flow characteristics and
characterizes them as shear thinning fluids. Inconsistent negative power flow indexes are no longer
to obtain the parameters reported in Table 1. Note that η ∞ has been experimentally measured by a
observed, as previously reported [13].
capillary rheometer at shear rates >10,000 s−1 .

10

4%
5%
6%
7%
1
Viscosity (Pa.s)

0.1

0.01
10 100 1000
Shear Rate (1/s)
Figure 3.
Figure Viscosity (mean)
3. Viscosity (mean) as
as aa function
function of
of shear
shear rate
rate for
for 4%,
4%, 5%,
5%, 6%,
6%, and
and 7%
7% gels.
gels. These
These curves
curves
describe the
describe the flow
flow behavior
behavior of
of the
the sheared
sheared and
and time
time independent
independent fluid.
fluid. The
The Extended
Extended Herschel-Bulkley
Herschel-Bulkley
model was
model was applied
applied to
to describe the flow
describe the flow characteristics
characteristics of
of the
the gels. Measurements were
gels. Measurements were carried
carried out
out in
in
triplicate. Error bars shown in the plot were calculated using a 95% confidence interval determined
triplicate. Error bars shown in the plot were calculated using a 95% confidence interval determined
with
with the
the software
software OriginPro
OriginPro 2016
2016 (OriginLab
(OriginLab Corporation,
Corporation, Northampton,
Northampton, MA, MA, USA).
USA).
Table 1. Extended Herschel-Bulkley parameters (Equation (2)) for sheared colloidal gels.
Table 1. Extended Herschel-Bulkley parameters (Equation (2)) for sheared colloidal gels.
Herschel-Bulkley Parameters 4.0% 5.0% 6.0% 7.0%
Herschel-Bulkley Parameters 4.0% 5.0% 6.0% 7.0%
τ߬0 (Pa)
଴ (Pa)
n
0.6 ±±0.07
0.6 0.07 2.4 ± 0.2
2.4 ± 0.2 5.5 ±
5.5 ± 0.40.4 5.0 5.0 ± 0.3
± 0.3
K (Pa·s )n 0.24 ± 0.03 0.58 ± 0.04 0.98 ± 0.1 1.67 ± 0.1
K (Pa·s
n ) 0.24 0.71
± 0.03 0.58 ±0.57
0.04 0.98 ±0.45
0.1 1.67 ±0.29
0.1
η ∞ (Pan · s) 0.01
0.71 0.01
0.57 0.01
0.45 0.290.01
2 0.97 0.99 0.99 0.99
η R(Pa·s)∞ 0.01 0.01 0.01 0.01
* n is the power law index.
2
R 0.97 0.99 0.99 0.99
Few observations should be highlighted to show the strong time and shear dependence of these
* n is the power law index.
gels. First, significantly lower viscosity values were obtained for pre-sheared gels (up to 60 times lower).
Second, the power
To further law indexes
understand thefor the pre-sheared
rheological behaviorgelsofwere
suchingels,
the range
Figure0 < n < 1,created
4 was which characterizes
to show that
them as shear thinning fluids. Inconsistent negative power flow indexes are
the flow curve of the “intact” gel does not follow the typical behavior of a purely shear thinning no longer observed,
fluid
as previously reported [13].
where shear stress increases with shear rate, albeit not in a linear manner (Figure 4A). In this case,
whenTothe further
shear understand the rheological
rate is increased, behavior
the shear stress of such which
decreases, gels, Figure
would4 be wasancreated to show
impossible that the
rheological
flow curvefor
behavior ofathe “intact”that
material geldoes
does not
not undergo
follow thestructural
typical behavior of a purely
modifications upon shear thinning
shear. It also fluid
showswhere
that
shear stress increases with shear rate, albeit not in a linear manner (Figure
the material is changing during the measurement due to the application of shear. Complementary4A). In this case, when the
shearisrate
data is increased,
provided the 4B,
in Figure shear stress decreases,
showing which
that the first would
normal be an
stress of impossible
the “intact”rheological
gel decreases behavior
when
for a material that does not undergo structural modifications upon shear. It also shows
the shear rate goes up. The first normal stress is an indication of the elasticity of the material, that the material
is changingby
generated during
the gelthe measurement
network, that is due
beingtobroken
the application
down byofthe shear.
effectComplementary data is provided
of shear. After pre-shearing the
in Figure 4B, showing that the first normal stress of the “intact” gel decreases
gel, little dependence of the normal stress with shear is observed. Furthermore, the magnitude of thewhen the shear rate
goes up. The first normal stress is an indication of the elasticity of the material, generated by the gel
network, that is being broken down by the effect of shear. After pre-shearing the gel, little dependence
of the normal stress with shear is observed. Furthermore, the magnitude of the normal stress is smaller
Gels 2017, 3, 45 8 of 12
Gels 2017, 3, 45 8 of 12

normal stress is smaller than the values measured for the “intact” gel. This is strong evidence that
than the values measured for the “intact” gel. This is strong evidence that there are structural changes
there are structural changes in the gel during deformation.
in the gel during deformation.

"Intact Gel"
Shear Stress (Pa) Sheared Gel

100

10

0 200 400 600 800 1000

Shear Rate (1/s)

(A)
10000

"Intact Gel"
Sheared Gel
Normal Stress (Pa)

1000

100
10 100 1000
Shear Rate (1/s)
(B)
Figure 4. “Intact” and sheared gels: (A) shear stress versus shear rate of 6% gels. A typical curve of shear
Figure 4. “Intact” and sheared gels: (A) shear stress versus shear rate of 6% gels. A typical curve of
thinning fluid is only observed for the sheared gel (empty circle marks); (B) first normal stress versus
shear thinning fluid is only observed for the sheared gel (empty circle marks); (B) first normal stress
shear rate of 6% gels. Filled squares show the decrease in normal stress of the “intact” gel at different
versus shear rate of 6% gels. Filled squares show the decrease in normal stress of the “intact” gel at
shear rates. Empty squares show no variation of normal stress at different shear rates.
different shear rates. Empty squares show no variation of normal stress at different shear rates.

Sheared and
Sheared and“intact”
“intact”
gelsgels
werewere also analyzed
also analyzed by scanning
by scanning electron microscopy).
electron microscopy). The micrographsThe
micrographs
illustrated illustrated
in Figure in Figure
5 show 5 show
different different
patterns for thepatterns for theespecially
two systems, two systems,
at lowespecially at low
magnifications.
magnifications. Both gels did have a primary structure of small particles closely packed
Both gels did have a primary structure of small particles closely packed together that were visible together that
were visible at higher magnifications. However, in the low magnification pictures,
at higher magnifications. However, in the low magnification pictures, the unsheared sample seems the unsheared
sample
to have seems
a moretoorganized
have a more organized
structure with structure with what
what appeared to beappeared to bepattern.
a cross-hatch a cross-hatch pattern.
After shear is
After shear is applied, this texture is no longer present (right micrographs) and the
applied, this texture is no longer present (right micrographs) and the sample had a more plate-like sample had a
more plate-like
appearance. It isappearance.
important toIt mention
is important to mention
that the presencethatof athe presence
pattern of gel)
(intact a pattern (intact
or lack gel) or
of a pattern
lack of a pattern (sheared gel) observed was consistent over the entire fractioned sample area. This
Gels 2017, 3, 45 9 of 12

Gels 2017, 3, 45 9 of 12
(sheared gel) observed was consistent over the entire fractioned sample area. This difference in the
texture of in
difference both
thegels at larger
texture scales
of both gelscan be an additional
at larger scales can indication of the changes
be an additional in the
indication organizational
of the changes in
structure
the of the system
organizational caused
structure by shear.
of the system caused by shear.

Figure
Figure5.5.Cryo-SEM
Cryo-SEMmicrographs
micrographsofof6%
6%“intact”
“intact”(on
(onthe
theleft)
left)and
andpre-sheared
pre-shearedgels
gels(on
(onthe
theright).
right).
Micrographs at different magnifications: 40 and 20 µm (from top to the bottom).
Micrographs at different magnifications: 40 and 20 µm (from top to the bottom).

StructuralKinetics
Structural KineticsModel
Model
Byshearing
By shearingthe thematerial
materialatata agivengiven shear
shear rate,
rate, viscosity
viscosity values
values cancanbe be recorded
recorded as aasfunction
a function
of
of time.
time. In way,
In this this way,
time time and shear
and shear are no are no longer
longer coupled coupled
as theyasare
theyin are
the in
rampthe up
ramp
andup anddown
ramp ramp
down measurements
measurements shown in shown
Figurein 3.
Figure 3. Thehistory
The shear shear canhistory
thencan
be then be incorporated
incorporated in termsinofterms of a
a kinetic
kinetic parameter that relates the viscosity with the structural changes in the
parameter that relates the viscosity with the structural changes in the material, expressed by Equation material, expressed by
Equation
(4). (4). Therefore,
Therefore, viscosity measurements
viscosity measurements as a functionas a function of time
of time were were performed
performed at five different
at five different shear
shear(10,
rates rates
50,(10,
100,50, 100,
500, and500,1000 and
s−1)1000
for 4%s−1and
) for5% 4%silica
and gel.
5% silica gel. how
To show To show how theofviscosity
the viscosity the intactof
the intact gel changes as a function of time due to shear, Figure 6 illustrates
gel changes as a function of time due to shear, Figure 6 illustrates the data obtained for 4% silica. The the data obtained for
4% silica.
initial Theof
viscosity initial viscosity
the system of the system
is reduced over time is when
reduced theover
sheartime when
rate is the shear
applied. There rate is applied.
is a significant
.
(γ )
There is a significant
decay followed by a constant decay followed by a constant (equilibrium) value at long
(equilibrium) value at long times. The data was used to calculate times. The dataa was

.

used to calculate a γ using Equation (10) for different shear rates (see also Figure 1) and c γ from
using Equation (10) for different shear rates (see .
also Figure 1) and c (γ ) from Equation (8). A power
Equation (8). A power law dependence of c γ on the shear rate was observed with similar indexes
law dependence
describing the gels c (γboth
of for ) onconcentrations
the shear rate (Table
was observed with similar
2). The constants λ0 andindexes
λe weredescribing the gelsfrom
also calculated for
both concentrations
the tests performed by (Table 2). The constants
extrapolation of the viscosity λ0 and curvesλe atwere
shearalso
rate calculated
0 and by the from
ratiothe tests
between
the experimentally measured value of η ∞ in the capillary viscometer
performed by extrapolation of the viscosity curves at shear rate 0 and by the ratio between the and the viscosity calculated by
the flow curve of the broken fluid, respectively. Both constants also followed
experimentally measured value of η in the capillary viscometer and the viscosity calculated by the a power law dependence
∞

flow curve of the broken fluid, respectively. Both constants also followed a power law dependence
on the shear rate. A summary of the compiled data and constants obtained is presented in Table 2.
Gels 2017, 3, 45 10 of 12

Gels 2017, 3, 45 10 of 12

on the shear rate. A summary of the compiled data and constants obtained is presented in Table 2.
For calculation of the power law relationships reported in Table 2, average values of the rheological
For calculation of the power law relationships reported in Table 2, average values of the rheological
parameters describing the Extended HB model were used. As illustrated in Figure 3, the rheological
parameters describing the Extended HB model were used. As illustrated in Figure 3, the rheological
data exhibits a very small error.
data exhibits a very small error.

Figure 6. Viscosity
Figure 6. Viscosity of
of4%
4%“intact”
“intact”gel
gelasasa afunction
function
of of time
time at different
at different constant
constant shear
shear ratesrates
(50, (50,
100, 100,
500,
500, and 1000
−
and 1000 s ).1 s−1).

Table 2. Structural kinetic parameters to describe viscosity time decay for 4% and
and 5%
5% gels.
gels.

Structural Kinetic Parameters 4.0% R2 5.0% R2

Structural Kinetic Parameters 4.0% R2 5.0% R2
c.
(γ ) . 0.36γ
0.0028 0.36
0.99 0.0026γ. 0.33
0.33
0.98
c γ 0.0028γ 0.99 0.0026γ 0.98
λeλe γ γ
. −0.48 −0.48
7.17 γγ
. −−0.26
0.26
27.627.6 0.98
0.98 7.17 0.96
0.96
. −0.85 . −0.68
λ0λ 582.6 γ γ −0.85
582.6 0.99
0.99 4.1γγ −0.68
4.1 0.98
0.98
0

Experimental and fitted viscosity values for 4% and 5% silica gels are depicted in Figure 7A for
Experimental and fitted viscosity values for 4 and 5% silica gels are depicted in Figure 7A for
two shear rates (10 and 500 s−−11 ). The measured data shows good agreement with the fitted curve
two shear rates (10 and 500 s ). The measured data shows good agreement with the fitted curve
obtained by the approach developed in this work at high and low shear rates. For these curves, λe was
obtained by the approach developed in this work at high and low shear rates. For these curves, λe
obtained from the viscosity measurements as a function of time for different shear rates. It is important
was obtained from the viscosity measurements as a function of time for different shear rates. It is
to mention that for high shear rates, the equilibrium viscosity reaches values close to those provided by
important to mention that for high shear rates, the equilibrium viscosity reaches values close to those
the flow curve of the broken gel (Extended HB Model), which implies that λe is close to 1. In practice,
provided by the flow curve of the broken gel (Extended HB Model), which implies that λe is close to
the application of constant high shear rates for long times is also able to completely break down the
1. In practice, the application of constant high shear rates for long times is also able to completely
3D network formed by the colloidal particles. In this case, when λe = 1, Equation (4) is simplified to:
break down the 3D network formed by the colloidal particles. In this case, when λe = 1, Equation (4)
is simplified to: .

.
λ0 + c γ · t · ( λ0 − 1)
λ γ, t = λ +. c
(γ ) ⋅ t ⋅ ( λ − 1) (11)
λ (γ, t ) 1=+ 0c γ · t · (λ0 −01) (11)
1 + c (γ ) ⋅ t ⋅ ( λ0 − 1)
Taking λe = 1 appears to be a good assumption for high shear rates; however, it is not as good for
Taking λe = 1 appears to be a good assumption for high shear rates; however, it is not as good
lower rates. Figure 7B shows the fitted and experimental viscosity values of 4% gels assuming λe = 1
for lower rates. Figure
for two different shear rates7B shows the1000
(100 and fitteds−and experimental
1 ). The mechanicalviscosity values of
stresses applied to 4%
the gels
systemassuming
at high
λ =
shear
e 1 for two different shear rates (100 and 1000 s −1). The mechanical stresses applied to the system
rates seems to be high enough to completely break down the gel network and to bring its viscosity
at
downhightoshear
thoserates seems
values to befor
obtained high
theenough to completely
time independent break
system down
(see Tablethe
1).gel
Thenetwork and tobebring
same cannot said
its viscosity
about lower down to those
shear rates. values
There obtained
is still for the
a transient time independent
structure system
that affects the (see Tablegiving
gel properties 1). Thehigher
same
cannot bevalues
viscosity said about lower
to the fluid.shear
Theserates. There
results showis still
theaimportant
transient structure thathistory
role of shear affects for
the predicting
gel properties
the
giving higher
rheological viscosityofvalues
properties to the fluid.materials,
time dependent These results show colloidal
especially the important role of shear
gels. Moreover, history
they for
provide
predicting the rheological properties of time dependent materials, especially colloidal gels. Moreover,
they provide extensive experimental data that may help to elucidate the flow effects on the
microstructure of time dependent colloidal gel systems.
Gels 2017, 3, 45 11 of 12

extensive experimental data that may help to elucidate the flow effects on the microstructure of time
dependent
Gels 2017,colloidal
3, 45 gel systems. 11 of 12

Figure 7. (A)
Figure Experimental
7. (A) Experimental (symbols)
(symbols)and andpredicted viscosity(lines)
predicted viscosity (lines) data
data forfor
4%4%
andand 5% gels
5% gels at 10 at
and10 and
500 s− 1 . sThe
500 4%4%
−1. The data is is
data plotted
plottedininthe left Y-axis
theleft andisisrepresented
Y-axis and represented byby square
square symbols.
symbols. The The 5% is
5% data data is
plotted
plotted in the in right
the right Y-axis
Y-axis and
and is is representedby
represented by diamond
diamond symbols;
symbols; (B)(B)
calculated viscosity
calculated assuming
viscosity assuming
λe = λ1e for
= 1 4%for gels
4% gels at shear ratesofof100
100and
and1000
1000 ss−
−1.1
at shear rates .

5. Conclusions
5. Conclusions
The present article
The present investigates
article investigatesthethe
rheological
rheologicalbehavior
behavior of
of time dependentcolloidal
time dependent colloidal gels
gels made of
made
of hydrophilic
hydrophilic fumedfumed
silicasilica
and and a hydrophobic
a hydrophobic solvent.The
solvent. The results
results show
showhow howthetheviscosity of the
viscosity ofgels
the gels
is affected by the shear history. Findings suggest that the application of shear irreversibly
is affected by the shear history. Findings suggest that the application of shear irreversibly changes changes
the microstructure
the microstructure of the
of the colloidalgels,
colloidal gels,their
their 3D
3D network,
network,and andconsequently
consequently theirtheir
flowflow
properties.
properties.
Shear forces affect not only the viscosity of the gel system, but also its elasticity (changes in normal
Shear forces affect not only the viscosity of the gel system, but also its elasticity (changes in normal
stresses). These effects result in structural changes in the material with the breakdown of the gel
stresses). These effects result in structural changes in the material with the breakdown of the gel
percolated network and inter-aggregate links present in the gel system. Cryo-SEM shows different
percolated
patternsnetwork and and
for the intact inter-aggregate links present
sheared gels, which may be aninindication
the gel system. Cryo-SEM
of changes shows different
in the organizational
patterns
structure of the system caused by shear. After the application of high shear for a relatively organizational
for the intact and sheared gels, which may be an indication of changes in the long time,
structure of the system
time dependence caused
in the bylonger
gels is no shear.observed
After theandapplication of high
the shear history shear
becomes for a relatively
irrelevant. To model long
time,the
time dependence
structural changesin the gels
during flow,isthe
noTiu
longer observed
and Boger modeland
was the
usedshear history becomes
with modifications irrelevant.
to consider
the flow curve of the pre-sheared material as the constitutive model. Good
To model the structural changes during flow, the Tiu and Boger model was used with modifications to agreement between
experimental
consider and predicted
the flow curve data was observed
of the pre-sheared materialforasthese colloidal gelsmodel.
the constitutive in a relatively wide rangebetween
Good agreement of
shear rates. Therefore, this approach provides a more realistic method to analyze the rheological
experimental and predicted data was observed for these colloidal gels in a relatively wide range of
properties of time dependent systems than the currently used Tiu and Boger model.
shear rates. Therefore, this approach provides a more realistic method to analyze the rheological
Acknowledgement:
properties This worksystems
of time dependent was supported
than theby the U.S. Army
currently Research
used Tiu andOffice under
Boger the Multi-University
model.
Research Initiative (MURI) grant number W911NF-08-1-0171. We would like to thank Debra M. Sherman for the
Cryo-SEM support
Acknowledgments: andwork
This discussions.
was supported by the U.S. Army Research Office under the Multi-University
Research Initiative
Author (MURI)Paulo
Contributions: grantH.
number W911NF-08-1-0171.
S. Santos We would
designed and performed like to thank
the experiments; PauloDebra M. Sherman
H. S. Santos, Osvaldofor the
Cryo-SEM support and discussions.
Campanella and Marcelo A. Carignano conceived and wrote the paper.
Author Contributions: Paulo H. S. Santos designed and performed the experiments; Paulo H. S. Santos,
Conflicts of Interest: The authors declare no conflict of interest.
Osvaldo Campanella and Marcelo A. Carignano conceived and wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Barnes, H.A. Thixotropy a review. J. Non-Newton. Fluid 1997, 70, 1–33.
References
1. Mewis, J.; Wagner, N.J. Thixotropy. Adv. Colloid Interface Sci. 2009, 147, 214–227.
1. 2. Mokoonlall,
Barnes, A.; Nöbel,aS.;
H.A. Thixotropy Hinrichs,
review. J. Post-processing
J. Non-Newton. of 1997,
Fluid fermented milk to
70, 1–33. stirred products: Reviewing the
[CrossRef]
effects on gel structure. Trends Food Sci. Technol. 2016, 54, 26–36.
2. Mewis, J.; Wagner, N.J. Thixotropy. Adv. Colloid Interface Sci. 2009, 147, 214–227. [CrossRef] [PubMed]
3. Bazmi, A.; Relkin, P. Effects of processing conditions on structural and functional parameters of whipped
3. Mokoonlall, A.; Nöbel, S.; Hinrichs, J. Post-processing of fermented milk to stirred products: Reviewing the
dairy emulsions containing various fatty acid compositions. J. Dairy Sci. 2009, 92, 3566–3574.
effects
4.
on gel structure. Trends Food Sci. Technol. 2016, 54, 26–36. [CrossRef]
Barbucci, R.; Pasqui, D.; Favaloro, R.; Panariello, G. A thixotropic hydrogel from chemically cross-linked
4. Bazmi,
guarA.;gum:
Relkin, P. Effects
Synthesis, of processing
characterization andconditions
rheological on structural
behaviour. and functional
Carbohydr. Res. 2008,parameters of whipped
343, 3058–3065.
dairy emulsions containing various fatty acid compositions. J. Dairy Sci. 2009, 92, 3566–3574. [CrossRef]
[PubMed]
Gels 2017, 3, 45 12 of 12

5. Barbucci, R.; Pasqui, D.; Favaloro, R.; Panariello, G. A thixotropic hydrogel from chemically cross-linked guar gum:
Synthesis, characterization and rheological behaviour. Carbohydr. Res. 2008, 343, 3058–3065. [CrossRef] [PubMed]
6. Lee, C.H.; Moturi, V.; Lee, Y. Thixotropic property in pharmaceutical formulations. J. Control. Release 2009,
136, 88–98. [CrossRef] [PubMed]
7. Labanda, J.; Llorens, J. Effect of aging time on the rheology of Laponite dispersions. Colloid Surf. A
Physicochem. Eng. Asp. 2008, 329, 1–6. [CrossRef]
8. Bosma, M.; Brinkhuis, R.; Coopmans, J.; Reuvers, B. The role of sag control agents in optimizing the
sag/leveling balance and a new powerful tool to study this. Prog. Org. Coat. 2006, 55, 97–104. [CrossRef]
9. Phuoc, T.X.; Howard, B.H.; Chyu, M.K. Synthesis and rheological properties of cation-exchanged Laponite
suspensions. Colloid Surf. A Physicochem. Eng. Asp. 2009, 351, 71–77. [CrossRef]
10. Labanda, J.; Llorens, J. Influence of sodium polyacrylate on the rheology of aqueous Laponite dispersions.
J. Colloid Interface Sci. 2005, 289, 86–93. [CrossRef] [PubMed]
11. Nosrati, A.; Addai-Mensah, J.; Skinner, W. Rheology of aging aqueous muscovite clay dispersions.
Chem. Eng. Sci. 2011, 66, 119–127. [CrossRef]
12. Pek, Y.S.; Wan, A.C.; Shekaran, A.; Zhuo, L.; Ying, J.Y. A thixotropic nanocomposite gel for three-dimensional
cell culture. Nat. Nanotechnol. 2008, 3, 671–675. [CrossRef] [PubMed]
13. Arnold, R.; Santos, P.H.S.; Campanella, O.H.; Anderson, W.E. Rheological and thermal behavior of gelled
hydrocarbon fuels. J. Propuls. Power 2011, 27, 151–161. [CrossRef]
14. Rahimi, S.; Natan, B. Thixotropic effect of inorganic gel fuels. J. Propuls. Power 2000, 16, 1182–1185. [CrossRef]
15. Jyoti, B.V.; Baek, S.W. Rheological characterization of ethanolamine gel propellants. J. Energ. Mater. 2016, 34,
260–278. [CrossRef]
16. Steffe, J.F. Rheological Methods in Food Process Engineering, 2nd ed.; Freeman Press: East Lansing, MI, USA, 1996.
17. Rajaram, B.; Mohraz, A. Microstructural response of dilute colloidal gels to nonlinear shear deformation.
Soft Matter 2010, 6, 2246–2259. [CrossRef]
18. Labanda, J.; Marco, P.; Llorens, J. Rheological model to predict the thixotropic behaviour of colloidal
dispersions. Colloid Surface A Physicochem. Eng. Asp. 2004, 249, 123–126. [CrossRef]
19. Labanda, J.; Llorens, J. A structural model for thixotropy of colloidal dispersions. Rheol. Acta 2006, 45,
305–314. [CrossRef]
20. Santos, P.H.S.; Campanella, O.H.; Carignano, M.A. Effective attractive range and viscoelasticity of colloidal
gels. Soft Matter 2013, 9, 709–714. [CrossRef]
21. Santos, P.H.S.; Campanella, O.H.; Carignano, M.A. Brownian dynamics study of gel-forming colloidal
particles. J. Phys. Chem. B 2010, 114, 13052–13058. [CrossRef] [PubMed]
22. Lu, P.J.; Zaccarelli, E.; Ciulla, F.; Schofield, A.B.; Sciortino, F.; Weitz, D.A. Gelation of particles with short-range
attraction. Nature 2008, 453, 499–503. [CrossRef] [PubMed]
23. Zaccarelli, E. Colloidal gels: Equilibrium and non-equilibrium routes. J. Phys. Condens. Matter 2007, 19,
323101. [CrossRef]
24. CAB-O-SIL Untreated Fumed Silica: Propertiesand Functions, Mechanisms of CAB-O-SIL, Technical Bulletin.
2009. Available online: http://www.cabotcorp.com/ (accessed on 1 August 2011).
25. Dullaert, K.; Mewis, J. Thixotropy: Build-up and breakdown curves during flow. J. Rheol. 2005, 49, 1213–1230. [CrossRef]
26. Dullaert, K.; Mewis, J. A model system for thixotropy studies. Rheol. Acta 2005, 45, 23–32. [CrossRef]
27. Willenbacher, N. Unusual thixotropic properties of aqueous dispersions of Laponite RD. J. Colloid Interface Sci.
1996, 182, 501–510. [CrossRef]
28. Rueb, C.J.; Zukoski, C.F. Viscoelastic properties of colloidal gels. J. Rheol. 1997, 41, 197–218. [CrossRef]
29. Tiu, C.; Boger, D.V. Complete rheological characterization of time-dependent food products. J. Texture Stud.
1974, 5, 329–338. [CrossRef]
30. Denn, M.M.; Bonn, D. Issues in the flow of yield-stress liquids. Rheol. Acta 2011, 50, 307–315. [CrossRef]
31. Arnold, R.; Santos, P.H.; Kubal, T.; Campanella, O.; Anderson, W.E. Investigation of gelled JP-8 and RP-1
fuels. In Proceedings of the World Congress on Engineering and Computer Science, San Francisco, CA, USA,
20–22 October 2009; Volume 1.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).