Drying Technology, 0: 1–10, 2015

Copyright # 2015 Taylor & Francis Group, LLC

ISSN: 0737-3937 print=1532-2300 online
DOI: 10.1080/07373937.2014.951451

Diffusion of Moisture from Hydrogel Scaffold with Induced

Porosity from Self-Assembled Bubbles
Subhajit Patra, Dharmendra Kumar Bal, and Somenath Ganguly
Department of Chemical Engineering, Indian Institute of Technology, Kharagpur, India

gel film using a fluidic arrangement.[2–4] This method has

Hydrogel scaffolds find use in the regulated release of biological
agents, and as three-dimensional supports for the formation of a
tissue matrix. This article addresses the diffusion of moisture from
the alginate and chitosan scaffolds under vacuum, and the resulting
evolution of pore structure. A fluidic arrangement was used to
induce a self-aligned monolayer of bubbles in the aqueous film prior
to crosslinking. The regular alignment of bubbles remained
unchanged as the film underwent crosslinking and subsequent
shrinkage due to the release of moisture. A mathematical model
was utilized to derive an effective diffusivity of the scaffold as a
function of the moisture content. The gradient in this profile and
the limiting value at zero moisture content were analyzed for scaf-
folds of different thicknesses in the presence or absence of bubbles.
The desaturation behavior of the bubble zone, the surface of the
scaffold, and the gel matrix were studied. The moisture profile
within the scaffold was also simulated using the mathematical
model.
Keywords Encapsulation; Polymer drying; Transport phenom-
ena; Vacuum drying
INTRODUCTION
There is great demand for highly ordered porous struc-
tures that are biocompatible and biodegradable and can be
used for delivering drugs and harvesting tissue-forming
cells. Hydrogels are known for their super-porosity. A gel
layer can trap a significant volume of biological agents in
the crosslinked polymer network. The biological agent
can diffuse easily to the surrounding medium. This unique
set of properties of hydrogels make them good candidates
for applications in controlled release and tissue regener-
ation. Uniform distribution of porosity and the pore-to-
pore distance are important in this regard.
There are methods known for the induction of voids in a
gel layer.[1] These methods are fiber bonding, solvent casting
or particulate leaching, emulsion freeze drying, gas foaming,
electrospinning, thermal phase separation, and use of super-
critical CO2. Recently, bubbles were introduced into the
Correspondence: Somenath Ganguly, Department of Chemical
Engineering, Indian Institute of Technology, Kharagpur 721302,
India; E-mail: snganguly@che.iitkgp.ernet.in
Color versions of one or more of the figures in the article can be
found online at www.tandfonline.com/ldrt.
the potential to provide better control of void size and
porosity. The bubbles created with this method are mostly
monodispersive. They rapidly self-assemble to provide an
ordered structure and a second level of porosity over and
above the intrinsic porosity of the gel film. In this method,
the gel is not exposed to any thermal or chemical treatment.
The method is not as expensive, as the solid is free from
fabrication.
Alginate and chitosan are well-known biopolymers for
controlled release applications. Alginate is a naturally
occurring polysaccharide. This polymer is sourced from
brown algae that grow in warm areas.[5–9] Chitosan is the
deacetylated form of chitin, which is commonly sourced
from shells of prawns, crabs, lobsters, shrimps, and the
exoskeletons of insects.[10–12] Other than for the regulated
release of drugs, these gel scaffolds are being extensively
studied for microencapsulation of natural antioxidants in
food, and for the removal of heavy metals and polyphenols
from waste water.
This article addresses the moisture transport process
within the alginate and the chitosan films, which have voids
induced through fluidic arrangements. The induction
device is comprised of two pulled microcapillaries, stacked
one inside another. The gas thread from the inner capillary,
when dragged by the co-flowing aqueous phase, disinte-
grated into mostly monodisperse bubbles. The monolayer
of bubble self-assembled on a petri dish, and retained the
alignment, as the polymer chains in the film were integrated
in a network through the addition of a crosslinker. The
final product was a robust and free-resting hydrogel film
with sub-millimeter voids, embedded uniformly in the film.
The trapped water from the gel network had to be removed
prior to the uptake of biological agents. This removal can
be accomplished by the application of vacuum.
At the time of drying, a dual porosity evolves within
the scaffold. One level of porosity is intrinsically avail-
able within the polymer network of gel matrix. The other
level of porosity is from the bubbles, induced by the
fluidic arrangement. The moisture removal from the
dual-porosity matrix and its manifestations needs careful

1
2 PATRA ET AL.
investigation, as it may be quite different from the
moisture transport in a single porosity medium. Also,
the changes in the effective diffusivity of moisture in
the scaffold with drying, and the intrinsic diffusivity at
the limit of complete desaturation, may provide useful
insight into the evolution of pathways for diffusion.
The way the scaffold thickness and the presence of bub-
bles affect the effective diffusivity profile may provide
information about the relative importance of diffusion
through tortuous pathways within the gel matrix, the free
diffusion within voids, and restricted diffusion, if any, at
the surface of the scaffold. This article attempts to
address these aspects through experiments and math-
ematical modeling. Two different gel systems that pre-
sumably have different levels of moisture diffusivity
were considered in this investigation.
MATERIALS AND METHODS
The scaffold was prepared from the aqueous suspension
of biopolymer. Bubbles were formed, and sustained by the
use of a surfactant. The crosslinker solution was sprinkled
onto the bubbles to develop a free-resting structure out of
the aqueous film.
Sodium alginate solution with concentration of 4%
(mass basis) and pH of 7.55 was prepared by adding
sodium alginate (Sigma Aldrich, St. Louis, USA) to dis-
tilled water (Merck Specialities Pvt. Ltd., Mumbai, India).
The solution was agitated in a magnetic stirrer (Tarson
India Pvt. Ltd., Kolkata, India) for 12 hours at 350 rpm.
Next, a mechanical mixer (Type: M56, Elecktrocrafts India
Pvt. Ltd., Mumbai, India) was employed for three hours at
3000 rpm. The surfactant solution containing 4% pluronic
F-127 (Sigma Aldrich, St. Louis, USA), with a pH of
6.58, was prepared by stirring at 350 rpm for one hour,
and then the solution was kept for 24 hours at 4 C in a
refrigerator. The alginate and the pluronic solutions were
mixed in a ratio of 1:1 and stirred at 100 rpm for 10 minutes
using a magnetic stirrer.
The chitosan beads (Sigma Aldrich) were agitated in
0.2 M acetic acid (Merck Specialities) solution using a
mechanical stirrer at 2500 rpm for 16 hours. The viscous
chitosan suspension containing 1.5 wt.% of chitosan was
filtered through using Whatmen filter paper of Grade 4.
The pH of the resulting solution was 4.5. Lutensol AT 25
(BASF, Ludwigshafen, Germany), which is a linear alcohol
ethoxylate. 1 wt.% of Lutensol was added to the chitosan
suspension to reduce the surface tension. The pH after
mixing of Lutensol AT 25 was found to be 5.5.
A rheometer (Anton Paar, Ashland, VA, USA) and
viscometer (Brookfield, Middleboro, Germany) were used
to determine the viscosity of the polymer solution. The sur-
face tension and the contact angle were recorded in a goni-
ometer (Raantec, Borgholzhausen, Germany) on glass and
parafilm, respectively. Immediately after the preparation,
the solutions were sonicated for half an hour in an
ultra-sonicator (Oscar Ultrasonics Pvt. Ltd., MC-101,
Mumbai, India). Then, the solutions were left undisturbed
for an hour prior to measurement.
Two pulled capillaries were assembled in the co-flow
device. The schematic diagram of the co-flow device is pre-
sented in Fig. 1. The aqueous polymeric solution flowed
through the outer capillary at constant flow rate of
5 mL=min with the help of a syringe pump (Harvard
Apparatus, Holliston, USA), while the nitrogen gas was
passed through the inner capillary. A mass flow controller
(Alicat Scientific, Tucson, USA) was installed to deliver the
nitrogen gas from a gas cylinder to the co-flow device at a
flow rate of 1 mL=min.
When the two phases merged at the tip of the inner
capillary, the gas thread was squeezed into bubbles. The
aqueous phase with bubbles was collected in a petri dish.
The image of the bubbles was captured under in-line illumi-
nation. A computer with image acquisition software and a
microscope (Labomed, Los Angeles, CA, USA) with a
camera were utilized for this purpose. Further analysis,
such as the addition of the scale overlay and implemen-
tation of an edge-detection algorithm, were done using
Davis software (Lavision GmbH, Goettingen, Germany)
to estimate the bubble size.
The aqueous film was transformed to gel scaffolds by
sprinkling the crosslinker solution—the 4% CaCl2 solution
for the alginate and the formaldehyde solution for the chit-
osan, respectively. Several sets of gel scaffolds with and
without induced voids were placed under an absolute press-
ure of 8000 Pa at constant temperature. The temperature
and initial condition for each scaffold are specified in
Table 1. The weight of the scaffolds was measured every
15 minutes until the weight reached a constant value.
The thickness of the scaffold at four preordained loca-
tions was also measured simultaneously with the weight
using a digital caliper (Moore and Wright, Maastricht,
Netherlands). The scaffolds were put back into the vac-
uum system after the measurements. The time taken for
the measurements was minimized to 30 seconds. A part
of the fully dried scaffold was processed further for taking
images under a scanning electron microscope (JEOL, JSM
5800, Tokyo, Japan).
FIG. 1. Schematic diagram of the co-flow device.
 DIFFUSION OF MOISTURE FROM HYDROGEL SCAFFOLD 3

TABLE 1
Description of the scaffolds used in the experiments
Tag Scaffold type Initial weight (g) Drying temperature Initial thickness (mm)
A Alginate without bubbles 3.759 50 C 2.02
B Alginate with bubbles 4.767 50 C 2.5275
C Alginate without bubbles 4.751 50 C 0.965
D Alginate with bubbles 5.083 50 C 0.9125
E Chitosan without bubbles 8.2 50 C 2.13
F Chitosan with bubbles 4.247 50 C 0.965
M Alginate without bubbles 3.554 30 C 0.975
N Alginate with bubbles 5.505 30 C 1.7975
X Alginate without bubbles 3.312 40 C 0.7025
Y Alginate with bubbles 2.772 40 C 0.6725

Mathematical Model
The release of moisture from the gel scaffold under vac- where M is moisture content in gm=gm dry scaffold and A
uum follows Fick’s law of diffusion. The moisture content and B are constants.With the knowledge of A and B,
within the scaffold is given in Eq. (1) Eq. (1) was discretized in finite difference form in accord-
ance with the Backward Euler method.
 
@M @ @M
¼ Deff for 0 < x < l ð1Þ DDt  iþ1 
@t @x @x Mjiþ1 ¼ Mji þ M jþ1  2M iþ1
j þ M iþ1
j1 ð7Þ
ðDxÞ2
Here, Deff is the overall moisture diffusivity within the
scaffold and l is the half-thickness of the scaffold. It is The resulting tridiagonal system of equations was solved
assumed that the moisture content at the surface is negli- using the Thomas algorithm.
gibly small. The initial and boundary conditions are listed A time step of 0.005 hour was considered. At each time
below. step, the thickness of scaffold was interpolated from the
experimental data using the correlation in Eq. (5). The
Mðt  0; x ¼ lÞ ¼ 0 ð2Þ change in grid size due to shrinkage was adjusted in Eq.
(7) at the end of each time step.
Mðt ¼ 0; 0 < x < lÞ ¼ M0 ð3Þ
Dx l eW2 þ f W2 þ gW þ h
  ¼ ¼ ð8Þ
@M Dx0 l0 l0
¼0 ð4Þ
@x x¼0
To arrive at the values of A and B in Eq. (6), an estimate
The complication in the solution of Eq. (1) arises from of the diffusivity as a function of moisture content was
two aspects. The Deff is a function of M. Secondly, the l required from the experimental data. A pseudo-analytical
changes due to the shrinkage of the scaffold with release treatment was utilized as per the procedure, described in
of moisture. The experimental data for scaffold thickness the following.
as a function of moisture ratio was gathered here. The The analytical solution of Eq. (1), based on constant
experimental data for thickness is plotted as a function thickness and effective diffusivity, is given in Crank,[14]
of moisture ratio, and is fitted through a polynomial func- and is reproduced in Eq. (9).
tion of the following form.
 
4M0 X1 1 2
2 p Deff t
3 2
l ¼ eW þ fW þ gW þ h ð5Þ M¼ exp ð2n  1Þ
p n¼1 ð2n  1Þ 4l 2
 px 
The effective diffusivity is known to vary with the moist- cos ð2n  1Þ ð9Þ
2l
ure content. The following equation describes a common
functional form reported in the literature.[13] The moisture gradient at the scaffold surface may be
obtained by taking a derivative of this expression. At each
Deff ¼ A  eBM ð6Þ time step of experimental measurements, the effective
4 PATRA ET AL.

diffusivity was estimated using the value of moisture gradi- 2 3W0

e2 W6 2ef W5 ðf 2 þ 2egÞW4
ent as per the following equation: 6 6 þ 5 þ 4 7
6 7
6 3 2 27
  46 2ðeh þ fgÞW ðg þ 2fhÞW 7
@M dM 6 þ þ 7
Deff a ¼ kc aðM  0Þ ¼  ð10Þ 4 3 2 5
@x x¼1 dt þ 2ghW þ hlnW W
Deff ¼ ð16Þ
p2 ðt  t0 Þ
Here, kc is the overall mass transfer coefficient, a is the sur-
face area per unit volume of scaffold ¼1l for the geometry,
and M is moisture content averaged over the entire thick-
ness. Upon integration of Eq. (9), RESULTS AND DISCUSSION
From viscometric studies, the aqueous solutions of
 
8M0 X1 1 2
2 p Deff t alginate and chitosan were found to be shear-thinning
M¼ exp ð2n  1Þ ð11Þ fluid. The power law constants (i.e., K and n of 2% sodium
p2 n¼1
ð2n  1Þ2 4l 2
alginate solution) were recorded as 0.17 Pa s0.96, respect-
Upon differentiation of Eq. (9), ively. The surface tension and contact angle of 2% alginate
solution on glass were noted as 60 mN=m and 32 , respect-
   
@M 2M0 X1 2
2 p Deff t
ively. The addition of 2% pluronic in the solution dimin-
¼ exp ð2n  1Þ ð12Þ ished both the surface tension and the contact angle to
@x x¼l l n¼1 4l 2
33 mN=m and 14 , respectively. The power law constants,
Combining Eqs. (10), (11), and (12), and neglecting K and n for 1.5 wt.% chitosan and 1% Lutensol were deter-
higher-order term mined to be 0.24 Pa s0.85, respectively. The viscosity was
reduced in the presence of Lutensol. Presumably, the
2Deff M0 8kc M0 deterioration in viscosity resulted from the elongation in
¼ ð13Þ the polymer chain caused by the formation of micelles.
l p2
The surface tension and contact angle of the chitosan
p2 Deff solution in the presence of Lutensol on glass were
kc ¼ ð14Þ measured to be 36.45 mN=m and 37.4 , respectively. The
4l
corresponding values without Lutensol were 56.15 mN=m
Combining Eqs. (10) and (14) and 43.74 , respectively. The existence of Lutensol in the
liquid phase degraded both the surface tension and the
dW p2 Deff contact angle. Over the parafilm surface, contact angles
¼ W ð15Þ of chitosan solution with and without Lutensol were
dt 4l 2
recorded as 74.4 and 98.2 , respectively.
Replacing l with the functional form of Eq. (5), and inte- The image of bubbles in alginate solution under a digi-
grating Eq. (15) between two experimental time steps with tal microscope is presented in Fig. 2. The circumference
the assumptions that Deff does not change over the small of the bubble was delineated implementing computer
time step, programs in Davis software from Lavison GmbH. The

FIG. 2. Bubbles in aqueous solution under microscope: (A) alginate and (B) chitosan.
 DIFFUSION OF MOISTURE FROM HYDROGEL SCAFFOLD
FIG. 3. SEM image of scaffold after drying: (A) alginate and (B) chitosan.
average diameter was found to be around 500 mm. In
Fig. 3, an SEM image of crosslinked gel film after
vacuum drying is shown. The decrease in the weight
and the thickness during the removal of moisture fea-
tured significantly in the shrinkage in the void space.
The weight of the scaffold generally dropped from 4
grams to 0.15 grams and the average thickness decreased
by six to seven times.
The experiments were conducted in several sets of
scaffolds. Each set comprises two scaffolds. One scaffold
had voids. The other did not have any void. The details
of the scaffolds and the conditions under which the
moisture was removed are listed in Table 1. The rate
of moisture removal was computed from the loss of
weight with time. Figures 4, 5, and 6 show the rate of
moisture removal in (gm=gm dry scaffold) min1 as a
function of moisture content for alginate and chitosan
scaffolds, respectively. The initial moisture content per
FIG. 4. Drying rate curve for alginate scaffolds at higher temperature.
5
gram of bone-dry scaffold was significantly higher when
the voids were present. Also, the initial moisture content
for chitosan scaffolds was much lower compared to that
of alginate. At higher temperature, the rate of moisture
removal was significantly higher. In almost all cases the
falling rate was observed, which further supports the
assumption of zero moisture at the scaffold surface. In
the case of scaffold E, an irregular behavior at the initial
stage of drying was observed, which may be caused by
the irregular formation and collapse of small vapor
packets prior to the crust formation.
Figure 7 shows the reduction in the thickness of the
scaffolds as the moisture content was reduced. The drop
in the thickness of the scaffold was significantly higher
at the initial stage for the thicker scaffold with bubbles.
When the voids were not present, the reduction in thick-
ness became significant at a later stage. This reflects the
difference in desaturation for the voids and the gel matrix.
FIG. 5. Drying rate curve for alginate scaffolds at lower temperature.
6 PATRA ET AL.
FIG. 6. Drying rate curve for chitosan scaffolds.
A collapse of the gel structure around the voids is
expected during the desaturation. The final shrinkage
ratio tends to be 0.2 for thinner alginate scaffold and
about 0.1 for thicker alginate scaffolds. In the case of chit-
osan, the final shrinkage ratio was about 0.3. The higher
shrinkage ratio indicates a dense structure. The shrinkage
ratio for scaffolds with bubbles was found to be always
lower than scaffolds without bubbles.
Figure 8, 9, and 10 present the effective diffusivity of the
alginate and chitosan scaffolds, respectively. The effective
diffusivity was computed from experimental drying rates
using the analytical model outlined in the previous section.
The model assumes the shrinkage as per the interpolation
FIG. 7. Effect of moisture ratio in thickness of alginate scaffolds with
polynomial fit.
FIG. 8. Effective diffusivity for alginate scaffolds at higher temperature.
of experimental data in polynomial form, and step-change
in diffusivity at each time step of the experiment. The
effective diffusivity always decreased at the initial stage of
moisture removal. The falling rate in Figs. 4, 5, and 6
had already suggested that the surface moisture was absent.
On further removal of moisture from the scaffold, the
capillary contacts inside the scaffold were broken, resulting
in loss of effective diffusivity. The exception was observed
for the thinnest scaffolds, particularly with bubbles. For a
thin scaffold, the bubble occupied a significant portion of
the thickness. Also, a hardened layer was likely to form
at the scaffold surface which constituted a major part of
the thickness. These two aspects influenced the exception.
The effective diffusivity at the initial stage (i.e., at 100%
FIG. 9. Effective diffusivity for alginate scaffolds at lower temperature.
 DIFFUSION OF MOISTURE FROM HYDROGEL SCAFFOLD
FIG. 10. Effective diffusivity for chitosan scaffolds.
saturation) was found to be an order of magnitude lower
than the self-diffusivity of water (2.3  109 m2=s).[15]
This difference reflects the hindrance to the diffusion of
moisture caused by the tortuosity of pathways within the
scaffold.
At the limit of zero moisture content, the effective
diffusivity increased slightly upon removal of the last
trace of moisture. At this stage, the entire pathway,
devoid of water fragment or ganglia, became available
for transport in the vapor phase. This limiting value of
effective diffusivity was found to be several orders of
magnitude lower than the self-diffusivity. This limiting
value of effective diffusivity is lower for thinner scaffolds.
The surface evaporation and consequent hardening of a
FIG. 11. Error surface for scaffold A.
7
thin layer next to the surface seemed to have influenced
the effective diffusivity more for the thinner scaffolds.
The introduction of voids in these scaffolds increased
the limiting value due to the contribution of free diffusion
in voids. This trend did not continue for the thicker scaf-
folds, where the voids occupied a small portion of the
overall thickness. The chitosan offered a slower diffusion
at the limit of zero moisture, and the shrinkage ratio was
also found to be higher. This is a clear indication of a
dense structure.
As mentioned in the Methods section, the calculation of
effective diffusivity is based on a mathematical treatment
of the experimental data. For solution of Eq. (1), a func-
tional form correlating effective diffusivity with moisture
content was required instead of discrete data. After a
review of experimental data and the functional forms
proposed in the literature,[16,17] the following equation
was utilized in this article.
Deff ¼ A  eBM ð17Þ
Here, M is the moisture content in the gm=gm dry scaf-
fold, and A and B are constants. The values of A and B
were computed by minimizing the root mean square error
(RMSE) between the experimental and predicted moist-
ure content at various stages of the experiment. The
RMSE was studied over wide ranges of values of A and
B. The error surface for Scaffold A is presented in
Fig. 11 as a typical case. The Nelder-Mead Simplex
method was used to converge at the final values of A
and B. Figures 12 and 13 compare the experimental and
the predicted moisture content with time for the alginate
and the chitosan scaffolds, respectively. Figure 14
8 PATRA ET AL.
FIG. 12. Experimental and predicted moisture content with time for
alginate scaffolds.
compares the effective diffusivity of scaffold by the two
methods, as a typical case.
Therefore, the release of moisture from the gel scaf-
fold was found to comply generally with Fick’s second
law. The shrinkage of the scaffold was observed pri-
marily in the thickness, and was incorporated in the
diffusion model. The effective diffusivity in the scaffold
was considered a function of moisture content in a dis-
crete manner in the analytical treatment and through a
continuous form of AeBM in the numerical model. Both
treatments gave similar estimate of diffusivity, and thus
FIG. 13. Experimental and predicted moisture content with time for
chitosan scaffolds.
FIG. 14. Comparison of two methods for prediction of effective diffusivity
(typical).
validated each other. The model was used further to
develop the spatial distribution of moisture within the
scaffold.
Figure 15 presents the spatial distribution of moisture
content within the scaffold after one hour of mass transfer.
The x-axis represents the distance from the centre, normal-
ized against the initial thickness. The diffusion rate was
much higher for thin scaffolds with voids, which is reflected
in the sharper drop in moisture for scaffold D compared to
the same for scaffold C. This aspect appeared in Fig. 16,
where the moisture content at the center of the scaffold is
plotted with time.
FIG. 15. Spatial distribution of moisture content within scaffold after an
hour.
 DIFFUSION OF MOISTURE FROM HYDROGEL SCAFFOLD 9

diffusion pathways. The exceptions with thinner scaf-

folds with voids resulted from surface evaporation—
hardening and dominance of void desaturation over
the matrix-level desaturation. A functional form of dif-
fusivity has been suggested in this article, and the moist-
ure content of the scaffold with time was simulated
using the mathematical model. The match with the
experimental data and the spatial profile of moisture
were discussed.

NOMENCLATURE
A Specific area, m2m3
D Diffusivity, m2s1
kc Mass transfer coefficient, ms1
L Half-thickness of scaffold, mm
M Local moisture content, gm=gm dry
scaffold
FIG. 16. Change in the moisture content at the center of the scaffold M Average moisture content, gm=gm dry
with time. scaffold
X Distance from the center of the scaffold,
mm
CONCLUSIONS T Drying time, min
Hydrogel scaffolds offer porous structure, where the Greek letters
porosity could be augmented by induction of a self-aligned W Moisture ratio at time t
monolayer of bubbles in the aqueous state through a fluidic Dx Grid size at time t, mm
arrangement. This article addresses the role of voids in Dt Time interval, sec
the diffusion of moisture out from the scaffold and the
Subscripts
evolution of the diffusion pathways. Two gel systems were
considered in this investigation. They are alginate, cross- 0 Initial
linked with CaCl2, and chitosan, crosslinked with formal- Eff Effective
dehyde. These gel systems are well-known for their Abbreviation
potential in absorbing heavy metals and polyphenols from RMSE Root mean square error
waste water, encapsulating natural antioxidants, drugs, SEM Scanning electron microscope
and matrix-forming cells for biological applications. The
regular alignment of bubbles remained unchanged as the
aqueous film underwent crosslinking and subsequent FUNDING
shrinkage due to the release of moisture. For the same The authors acknowledge the Department of Science &
scaffold thickness, the moisture content per unit weight Technology, Government of India, for financial support.
of dry scaffold was much higher when the bubbles were
present. When the voids were present, the desaturation rate REFERENCES
1. Chung, H.J.; Park, T.G. Surface engineered and drug releasing
was initially much faster. The rate curve indicated the
pre-fabricated scaffolds for tissue engineering. Adv. Drug Delivery
absence of surface moisture in almost all cases. The final Rev. 2007, 59, 249–262.
shrinkage ratio was highest in the case of chitosan, indicat- 2. Martynov, S.; Wang, X.; Stride, E.P.; Edirisinghe, M.J. Preparation
ing a dense structure. The ratio was lowest for thicker scaf- of micro-morous alginate gel using a microfluidic bubbling device.
folds with voids. Int. J. Food Eng. 2007, 6, 1556–3758.
3. Chung, K.; Mishra, N.C.; Wang, C.; Lin, F.; Lin, K. Fabricating scaf-
A mathematical model was utilized to derive the
folds by microfluidics. Biomicrofluidics 2009, 3, 22403.
effective diffusivity of the scaffold as a function of the 4. Wang, X.; Li, X.; Stride, E.; Huang, J.; Edirisinghe, M.; Schroeder,
moisture content. The diffusivity at 100% saturation C.; Best, S.; Cameron, R.; Waller, D.; Donald, A. Novel preparation
was typically an order of magnitude lower than the and characterization of porous alginate films. Carbohydr. Polym.
self-diffusivity of water. The decrease in diffusivity with 2010, 79, 989–997.
5. Draget, K.I.; Ostgaard, K.; Smidsrod, O. Homogeneous alginate gels:
moisture removal suggested breaking of capillary con-
A technical approach. Carbohydr. Polym. 1991, 14, 159–178.
tact. Near the limit of zero moisture content, this trend 6. Kuo, C.K.; Ma, P.X. Ionically crosslinked alginate hydrogels as
was reversed, suggesting less hindrance to diffusion due scaffolds for tissue engineering: Part 1: Structure, gelation rate and
to the complete removal of water fragments from the mechanical properties. Biomaterials 2001, 22, 511–521.
10 PATRA ET AL.
7. Sheridan, M.H.; Shea, L.D.; Peters, M.C.; Mooney, D.J. Bioabsorbable
polymer scaffolds for tissue engineering capable of sustained growth
factor delivery. J. Controlled Release 2000, 64, 91–102.
8. Rebecca, H.L.; David, H.A.; Frank, T.G. Transport characterization
of hydrogel matrices for cell encapsulation. Biotechnol. Bioeng. 1996,
54, 365–373.
9. Dohnal, J.; Stepanek, F. Inkjet fabrication and characterization of
calcium alginate microcapsules. Powder Technol. 2010, 200, 254–259.
10. Prashanth, K.V.; Tharanathan, R.N. Crosslinked chitosan: Prep-
aration and characterization. Carbohydr. Res. 2006, 341, 169–173.
11. Kim, I.Y.; Seo, S.J.; Moon, H.S.; Yoo, M.K.; Park, I.Y.; Kim, B.C.;
Cho, C.S. Chitosan and its derivatives for tissue engineering applica-
tions. Biotechnol. Adv. 2008, 26, 1–21.
12. Testouri, A.; Honorez, C.; Barillec, A.; Langevin, D.; Drenckhan, W.
Highly structured foams from chitosan gels. Macromolecules 2010, 43,
6166–6173.
13. Thuwapanichayanan, R.; Prachayawarakorn, S.; Kunwisawa, J.;
Soponronnarit, S. Determination of effective moisture diffusivity
and assessment of quality attributes of banana slices during drying.
LWT–Food Sci. Technol. 2011, 44, 1502–1510.
14. Crank, J. The Mathematics of Diffusion; Oxford University press:
Oxford, UK, 1975.
15. Mills, R. Self-diffusion in normal and heavy water in the range
1–45 C. J. Phys. Chem. 1973, 77, 685–688.
16. Boudhriouaa, N.; Bonazzi, C.; Daudin, J.D. Estimation of moisture
diffusivity in gelatin–starch gels using time-dependent concen-
tration-distance curves at constant temperature. Food Chem. 2003,
82, 139–149.
17. Ruiz-Cabrera, M.A.; Foucat, L.; Bonny, J.M.; Renou, J.P.; Daudin,
J.D. Assessment of water diffusivity in gelatine gel from moisture pro-
files.II. Data processing adapted to material shrinkage. J. Food Eng.
2005, 68, 221–231.