749

Thermophysics and Aeromechanics, 2013, Vol. 20, No. 6

Peculiarities of diffusion in gels
*

B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov

Moscow State University of Mechanical Engineering, Moscow, Russia

E-mail: pokusaev2005@yandex.ru
(Received April 30, 2013)
An optical method was applied to study the peculiarities of diffusion in gel: this method provides real-time
visualization of spreading of solutes brought into the gel. It was shown that spectral characteristics of reflected light
give additional information about nature of diffusive spreading of solutes and about state of the gel. Gels with different
densities and lifetime were studied. These parameters have strong influence on the velocity of diffusion. The study
demonstrated critical differences for diffusion process in gels with true solutions and with solutions with nanoparticles.
Experiments discovered the anisotropy in 3D diffusion of solutes in gels; physical explanation of this phenomenon was
proposed.
Keywords: gels, optical methods, diffusion, nanoparticles, aging, anisotropy.
Introduction
The usual definition for gels is a dispersive system with liquid dispersing medium, and
the dispersion phase makes up a spatial structured mesh due to intermolecular interaction
in the contact sites. Gels have such inherent qualities like a tendency to retain its structure, plas-
ticity, resilience, and also thixotropy, i.e., the quality of restoring the gel original structure after
mechanical destruction under isothermal conditions. All these properties, including the loss
of fluidity, are attained at low contents of dispersed phase in the gel: from fractions of per-
cent to several percents only. Gels are subjected to aging, i.e., their physical and chemical
properties do change due to recondensation and recrystallization, including the process
of releasing of liquid phase because of settling of the structured mesh [1, 2].
All these features create interesting applications of gels in industries and technologies.
Gels are widely used in production of polymers, catalysts, sorbents, membrane filters, service
wellbore fluids, cosmetic and pharmacy compositions [3, 4]. Presently, the promising appli-
cation of gels is the use in regenerative medicine for growth of different biological issues,
including stem cells in vitro [5, 6]. The gel capillary network can be used for transport
of nutrients to separate cells with culture broth and for removal of cell metabolism. One of variants
of problem solution is creation of artificial hydrogel matrix with complicated spatial system
of microchannels: this task assumes formulation and solving of fundamental problems for
study of micro- and nano-porous systems in hydrogel matrix [7].

*
Research was financially supported by the Education and Science Ministry of RF (Project code 7.7787.2013) and by
RFBR (Projects No. 11-08-00368_ and No. 12-08-31243 mol_a).
B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov, 2013
B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov
750
The processes of diffusion and thermal conductivity in macro-objects have been studied
in detail [810], but specifics of these processes in microporous and microchannel-filled bodies
requires more focused consideration [1116]. The problems similar in physical nature have
been studied for the case of heat transfer in structured porous bodies (see, e.g., [17]). The trans-
fer processes in gels have such features as nonstationarity and anisotropy: they are determined
by the nature, structure, and behavior of transfer medium [18]. For this type of systems,
the critical aspect is the effect of solute size in comparison with inner scales of microchannels
within the transfer medium [19].
The study of transfer process in microsystems is often based on optical methods [20].
The optical methods are instrumental in collecting of data about fluid dynamics for oil dis-
placement with water [21], for study of physical-chemical mechanisms of interfacial instability
during chemosorption, and for study of microflows in the brim meniscus of wetting [22, 23].
Optical tools help in visualization of transfer process in real time mode and are useful for
understanding the physical and chemical features of the process. They are useful for obtaining
of direct data on macrokinetic transfer coefficients; in some cases, these coefficients can be
related to parameters describing the intrinsic structure of the transport medium.
The goal of this paper was experimental study of diffusion in gels with optical methods.
That is why we studied optical properties of initial solutions and the produced gels with different
density and lifetime. One of objectives was experimental proof of relation between the solute
transfer rate and the local-time state of the gel, since the gel is a medium with time-varying
properties. Gel is microporous medium, therefore, it is natural to expect a difference for diffu-
sion velocity for true solutions and solutions with nanoparticles. However, this difference
in diffusion velocities of solutes in gels has not been studied yet. Finally, optical methods can
be used for measuring the gel structure anisotropy through visualization of maldistribution
of solute diffusional transfer, and this can offer physical explanation of this phenomenon.
Investigation objects and methods
For study of diffusion process in a plain diffusional layer, we used gels on the basis of sili-
cates. Gels of different density were produced. The stock solution of sodium silicate was
diluted with distilled water to the levels of density 1.04, 1.068, and 1.09 g/cm
3
. The gel-
generating mixture was prepared by adding the hydrochloric acid to the salt solution, and
the mixture pH was adjusted to the range 3.54.5. The density of silicate aquatic solution was
controlled by refractometry method through measuring the refractive index with the Abbe refrac-
tometer. As it is shown below, this minor difference in the initial density of sodium silicate solu-
tion produces differences in properties of final gels (optical and diffusion-related properties).
Since the gel structure might be time-variable, we used both fresh and aged gels (with life-
time more than 24 hours) for study of aging on transfer processes.
The study of gel microstructure influence on diffusion in a plain layer was conducted
with three substances with very different physical and chemical properties. One substance was
a water solution of potassium permanganate KMnO
4
, which is a true solution, being a mixture at
molecular level. Another sample was water-based ink, which is an aquatic suspension of nano-
particles with wide dispersion composition. The third sample was inkjet ink nanoparticle
suspension with a narrow dispersion available in methylethylketone (organic solvent). All
these substances are gel-inert additives, that is, diffusion of these solutes does not change gel
properties.
For study of diffusion in volume, we tested the sample of hydrogels based on hyaluronic
acid modified with a cross-linker; the cross-linker might be alfa-, omega-diamines. The cross-
linker provides additional chemical bonds between different polymer chains making a hydrogel
matrix, and this cross-linking creates a special 3D structure of gel. As an optical indicator
(the substance for transfer) we used a gel-neutral aquatic dye solution.
Thermophysics and Aeromechanics, 2013, Vol. 20, No. 6
751
The optical method was used for estimating the diffusion coefficient for a solute in a gel.
This choice of optical methods was supported by high accuracy of methods, contactless nature,
and informative value of tools. The experimental setup diagram is shown in Fig. 1. The setup is
a microscope with an illumination system equipped with a device for placement and movement
of a cell comprising a gel sample. The setup has an option of placing an additional (external to
the cell) box with immersion liquids for ameliorating the effect of cell wall curvature. The mi-
croscopes digital camera is focused on the boundary of transport substance (solute) traveling
in the gel volume. The setup was designed for recording a flat front of the moving boundary
of the transport substance or a spherical front from a point source. The latter version was needed
for estimating the properties of diffusion anisotropy in gels.
Determining of diffusion coefficients was based on measuring of distribution of solute
concentration over the distance, that is, we tracked down the movement of iso-concentration
planes. This method is known as a method of moving boundaries, and the calculation of diffu-
sion coefficient is based on the velocity of transport of a constant-concentration solute inside
a gel matrix [24]. To estimate the order of the diffusion coefficient, the following formula was
applied:
2
( ), D x t = (1)
where D is the diffusion coefficient, x is the iso-concentration plane displacement velocity,
t is time, and is the coefficient that accounts for interaction of diffusive substance with the me-
dium (in our calculations, we used = 6).
Results and discussion
For the first stage of experiment, the refractive index of sodium silicate solutions was
measured at different concentrations of corresponding gels. Measurements were conducted
at the temperature of 21 . This data helped to control reproducibility of gel state during mea-
surements. The results are shown in Fig. 2. One can see from this graph that the refractive in-
dex is higher for original solution than for derived gel. This effect is linked to a change
in the molecular structure of silicate sodium and development of structured mesh: this
structure gives us a gel. Besides, we noticed a slower increment of refractive index with
the density growth for a gel than for a solution.
The dependency of refractive index on gel density brings a conclusion about a difference
of reflection and transmission spectra for gel samples at different densities, and this phenome-
non can be a foundation for developing a spectral method for analysis of gel structure.
Figure 3 depicts a picture of a fragment (corresponding to the same time moment) of a posi-
tion of the concentration front in a plain cell for the case of vertical diffusive motion of solutes
through a sodium silicate gel with a concentration 1.04 g/cm
3
. The pictures from left to right show
the cases of diffusion of KMnO
4
solution, inkjet ink, and water ink. The process of diffusion


Fig. 1. Experimental setup diagram.
1 cell with gel sample, 2 moving boundary of the marker substance, 3 system for marker feeding,
4 the additional cell with immersion fluid, 5 light beams, 6 illumination system,
7 microscope, 8 digital camera.
B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov
752
was initiated in three cells simultaneously. Figure 3 demonstrates that diffusion of KMnO
4

solution in the gel matrix takes place with a high velocity; the diffusive smearing of inter-
face is visible, and the visual pattern of diffusion fits the classic idea of diffusion dyna-
mics. As for diffusion of two types of inks in gel, their diffusion velocities are close.
The smearing of interface for spreading water-based ink is visible. On the opposite,
the diffusion of inkjet ink in gels creates a distinct boundary; this might be explained by
presence of organic solvent in the ink which forms the interface due to surface tension.
Here we use the collected experimental data on diffusion velocity and the method of iso-
concentration planes for estimating the diffusion coefficient on the basis of relation (1) for all
tested substances: the diffusion coefficient for KMnO
4
in gel was 2.910
9
m
2
/s, inkjet ink
1.110
9
m
2
/s, and for water-based ink 0.9610
9
m
2
/s. The difference in diffusion
coefficients of true solution (KMnO
4
) and nanoparticle suspension (ink) is determined by
a difference in sizes of transported particles, which are critical for particle-matrix interaction.
The Einstein formula [10] for the diffusion coefficient of a Brownian particle (and nanopar-
ticles suspended in liquid belong to this class) tells that the diffusion coefficient is inversely pro-
portional to the hydrodynamic drag of a particle in viscous fluid. We know that the nanoparticle
size in inks is about the size of structural mesh of the gel (or possibly bigger), this drag force
may be rather high and it prevents spreading of particles in gel (on the molecular level for
KMnO
4
diffusion, this makes impact on the value of transfer coefficient due to reduction of mole-
cular mobility).
When diffusion of water-based ink takes place in gel, only nanoparticles of finest fraction
spread out. However, motion of the painted front of inkjet ink took place in out experiments.
The possible explanation is the following. We observe diffusion of free water molecules
from the gel into the organic phase. This transition of water from one phase to another phase
creates a local osmotic rarefaction. This, on the one hand, creates the additional motive transfer
force revealed as excessive pressure in organic phase which would promote interface (together
with colorant nanoparticles) inside the gel; on the other hand, there is a possibility of changing
the gel structure near the interface (the structured mesh becomes loose).
The regular visual observation for the condition of diffusive front is good for qualitative
study only. A more accurate approach is an expe-
riment on measuring the relative transmittance (or
absorbance) of white light passing through the gel
and the diffusing substance. Meanwhile the sub-
stances of the studied system have different


Fig. 2. Refractive indices for sodium silicate
solutions and corresponding silicate gels.
1 liquid solutions, 2 fresh gel with
the same density.


Fig. 3. Position of a concentration front
in a flat cell while vertical diffusive movement
of the solute in a gel.
Pictures from left to right: KMnO
4
, inkjet ink,
water-based ink.
Thermophysics and Aeromechanics, 2013, Vol. 20, No. 6
753
spectral absorption characteristics, so the data on light absorbance at different wavelength
of light are very informative. The data on KMnO
4
in the sodium silicate gel are plotted in Fig. 4,
where we have shown the dynamics of light transmittance for a diffusive component at differ-
ent points of the gel sample. These spectral studies demonstrated that measurement of diffusion
in gel with a moving boundary method for different wavelengths of probing light gave different
values for diffusion coefficient for the same component; this can be indication of selective inte-
raction between the component and structures of the silica gel. Therefore, the data for transmit-
tance spectra for silica gel samples might be used for study and control of sorption features
of the gel matrix (when the multiphase and multicomponent filtration flows take place).
As was mentioned above, the gel is not a steady structure. Processes of recondensation
with variation and thickening of the structured mesh occur in the gel; that is, the medium
of diffusion changes with time (they call this process as gel aging). Therefore, it is meaning-
less to tell about transfer coefficients in gels beyond the time limits. This fact was confirmed
in experiments: we measured the diffusion coefficient for KMnO
4

in silicate gels with different
densities and different lifetime (a fresh gel and one-day aged gel). These results are plotted
in Fig. 5. This graph shows that the diffusion coefficient decreases with the gel density (this is
predictable due to condensing of gel mesh) and with gel lifetime. When relaxation processes
take place, the structured mesh also changes and becomes denser.
Let us consider the process of diffusion in gel with the evolving structure. We assume
that recondensation in gel produces temporary local stagnation zones without any mass transfer
to the ambient zones; the solute concentration there becomes frozen for a time. This is a typical
time of gel relaxation , which describes the duration of gel restructuring processes.
For this kind of diffusion model, a 1D nonstationary diffusion equation for gel can be
written in the following form:
) (
2
2

t
C C K
x
C
D
t
C
, (2)
where is the transfer substance concentration, t is the current time, x is the longitudinal coor-
dinate, and D is the diffusion coefficient of the transfer substance in the liquid phase of gel.
Averaged over volume and time , the transfer coefficient between the typical and dense struc-
tured mesh zones K depends on the gel density (and this produces dependency on the current
time of diffusion process t), on the volumetric fraction of gel with dense structure, diffusion
coefficient D, and the scale of developed dense structures. Thus, the value of K, the same way
as , is a characteristic of the intrinsic structure of gel. Note here that the concentration
in the last term of equation (2) is taken at moment of t , i.e., since the formation of a stagnant
zone which has been decomposed till
this moment and released out the accu-
mulated solute.

Fig. 4. Spectral-revealed curves of solute
front movement in the gel.
The abscissa axis plots the distance H from the ini-
tial solute front while diffusion into the gel
(gel is on the right, and solvent is on the left);
the ordinate axis is the relative intensity A
of transmitted light at given wavelength
(the complete transparency corresponds to num-
ber 256, the complete absorbance stands for 0).
Position of diffusion front: 1, 2 initial;
3, 4 intermediate; 5, 6 final. Wavelengths:
1, 3, 5 red light; 2, 4, 6 green light.

B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov
754
If we assume the relaxation time to be a small
parameter (this is not a compulsory condition for
gels), we can decompose
t
into a series in t and
keep only two main terms of the series, then we
obtain:
+

t
C
t C C
t

) ( . (3)
Substitution of (3) in equation (2) gives us an equation
2
ef
2
,
C C
D
t
x

=

(4)
where
ef
1
D
D
K
=
+
. (5)
This relation for diffusion coefficient in gels was developed in paper [15] from phenomeno-
logical ideas. Relation (5) for diffusion coefficient in gels is included into a single complex
of values K, which describes the intrinsic structure of gel and can be calculated from direct
experimental data. Thus, the optical-spectral study of diffusion in gels gives information about
dynamics of gel intrinsic structure.
We should note that relation (5) was obtained for small relaxation times , but when this
condition is not fulfilled, the reverse problem of diffusion on the basis of equation (2) should
be considered. For this situation, the parameters K and are the functions of current time t, so
time interpolation of these parameters can be done with exponential laws. As the gel aging
process continues, the transfer coefficient decreases. We should emphasize that this type of time
dynamics is inherent in diffusion processes in true solutions, but for diffusion of nanoparticles
in gels the time evolution of coefficients might be more complicated.
Fig. 5. Dependency of diffusion coefficient
(D10
10
m
2
/s) in sodium silicate gel on its density
(, g/cm
3
) while gel aging.
1 fresh gel, 2 old gel.

Fig. 6. Process of diffusion and time dependence of boundary position.
On the left picture of colorant distribution in the gel volume while 3D diffusion; on the right time-
fixed (after two minute interval) positions of the colorant zone within the gel volume.
Thermophysics and Aeromechanics, 2013, Vol. 20, No. 6
755
One of interesting features of gels is anisotropy, which means a difference in diffusion
velocity for different directions of matter transfer. To study this phenomenon, we use a sample
of hyaluronic acid based gel thickened with a cross-linker. An inert colorant was injected into
this sample via a syringe. For this situation, a three-dimensional diffusion of die took place.
For every two minutes, the colorant boundary position was recorded with a video-camera; this
gave a 2D image of the iso-concentration surface vs. time. A picture of diffusion process and
image of time dependency of the boundary position is shown in Fig. 6. The experimental data
are presented in detail in paper [18].
After processing of experimental data for tested hydrogel samples with formula (1),
the following average values of diffusion coefficients have been obtained: in vertical direction
D
x
= 0.4110
9
m
2
/s, and for horizontal direction D
y
= 0.9810
9
m
2
/s. The latter value
of diffusion coefficient exactly fits the result published in [14] for this medium. The aniso-
tropy in diffusion is explained by anisotropy of gel properties developed due to features of gel-
ling regime: polymerization conditions for longitudinal direction are different from those for
transverse direction. The possible reason for this difference for impact of gravity: gravity force
enhances settling of polymerized gel particles and gel densification in the longitudinal direction.
References
1. B.V. Deryagin, The stability of colloidal systems. Theoretical aspects, Russian Chemical Reviews, 1979, Vol. 48,
No. 4, P. 363388.
2. Yu.G. Frolov, Course on Colloidal Chemistry. Surface Phenomena and Dispersion Systems, Khimiya, Moscow,
1988.
3. H.K. Henisch, Crystals Growth in Gels, Pennsylvania State University Press, 1969.
4. R. Hilfer, Transport and relaxation phenomena in porous media, Advances in Chemical Physics, 1996, Vol. 92,
P. 299334.
5. B.A. Westrin and A. Axelsson, Diffusion in gels containing immobilized cells: a critical review, Biotechnology
and Bioengineering, 1991, Vol. 38, P. 439446.
6. H. Fang, R. Wan, X. Gong, H. Lu, and S. Li, Dynamics of single-file water chains inside nanoscale channels:
physics, biological significance and applications, J. of Physics D: Applied Physics, 2008, Vol. 41, No. 10,
. 103002-1103002-16.
7. B. Amsden, Solute diffusion within hydrogels: mechanisms and models, Macromolecules, 1998, Vol. 31, No. 23,
P. 83828395.
8. S.S. Kutateladze, Fundamentals of Heat Transfer, Acad. Press, New York, 1963.
9. A.V. Lykov, Theory of Heat Conduction, Vysshaya Shkola, Moscow, 1967.
10. J. Philibert, One and half century of diffusion: Fick, Einstein, before and beyond, Diffusion Fundamentals, 2006,
Vol. 4, P. 6.16.19.
11. A.H. Muhr and J.M.V. Blanshard, Diffusion in gels, Polymer, 1982, Vol. 23, July (Suppl.), P. 10121026.
12. S. Tan, H. Dai, and J. Wu, Optical investigation of diffusion of levofloxacin mesylate in agarose hydrogel,
J. Biomedical Optics, 2009, Vol. 14, No. 5, P. 050503/1050503/3.
13. M. Fischer, M. Dietzel, and D. Poulikakos, Thermally enhanced solubility for the shrinking of a nanoink
droplet in a surrounding liquid, Intern. J. Heat and Mass Transfer, 2009, Vol. 52, P. 222231.
14. T. Kosztolowicz, K. Dworecki, and St. Mrowczynski, Measuring subdiffusion parameters, Physical Review E,
2005, Vol. 71, . 041105-1041105-11.
15. M. Lauffer, Theory of diffusion in gels, Biophysical Journal, 1961, Vol. 1, P. 205213.
16. B. Amsden, Solute diffusion within hydrogels. Mechanisms and models, Macromolecules, 1998, Vol. 31,
P. 83828395.
17. A.P. Yankovsky, Modeling of thermal conductivity in hybrid composites made from binding matrix, with ortho-
gonal reinforcement with tubes and with disperse-hardened hollow inclusions, Themophysics and Aeromechanics,
2011, Vol. 18, No. 1, P. 135150.
18. D.A. Kazenin, S.P. Karlov, D.A. Nekrasov, and B.G. Pokusaev, Express-analysis of diffusion properties of gels
for arrangement of filtration flows in soils, Ekologiya & Promyshlennost Rossii, 2011, No. 1, P. 1315.
19. S. Ozturk, Ya.A. Hassan, and V.M. Ugaz, Interfacial complexation explains anomalous diffusion in nanofluids,
Nano Letters, 2010, Vol. 10, P. 665671.
B.G. Pokusaev, S.P. Karlov, A.V. Vyazmin, and D.A. Nekrasov
756
20. B.G. Pokusaev, D.A. Kazenin, and S.P. Karlov, Immersion tomographic study of the motion of bubbles
in a flooded granular bed, Theoretical Foundations of Chemical Engng., 2004, Vol. 38, No. 6, P. 561568.
21. O.V. Vitovsky, V.V. Kuznetsov, and V.E. Nakoryakov, Stability of the displacement front and development
of fingering in a porous media, Fluid Dynamics, 1989, Vol. 24, No. 5, P. 739745.
22. B.G. Pokusaev, D.A. Kazenin, S.P. Karlov, and A.V. Vyazmin, Interfacial mass transfer in the liquid-gas sys-
tem: an optical study, Theoretical Foundations of Chemical Engng., 2001, Vol. 35, No. 3, P. 227231.
23. S.P. Karlov, D.A. Kazenin, and A.V. Vyazmin, The time evolution of chemo-gravitational convection on a brim
meniscus of wetting, Physica A, 2002, Vol. 315, No. 12, P. 236242.
24. A.Ya. Malkin and A.E. Chalykh, Diffusion and Viscosity of Polymers, Measurement Methods, Khimia,
Moscow, 1979.