Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

580 Ind. Eng. Chem. Res.

2011, 50, 580–589

Evaporation of Water from Structured Surfactant Solutions


Paschalis Alexandridis,* Shushan Z. Munshi, and Zhiyong Gu†
Department of Chemical and Biological Engineering, UniVersity at Buffalo, The State UniVersity of New York
(SUNY), Buffalo, New York 14260-4200

Alkyl-propoxy-ethoxylate surfactant aqueous solution films are exposed to air of constant relative humidity,
and the water loss from the film is monitored over time, until equilibrium is reached. The surfactants self-
assemble into lyotropic liquid crystals with a structure that varies depending on the water concentration at a
given time. The water loss data are analyzed to investigate the factors affecting the drying rate of the alkyl-
propoxy-ethoxylate surfactants, such as the air relative humidity, microstructure in the surfactant film, and
the surfactant degree of hydration. Analytical solutions of the diffusion equation are used to extract the water
diffusion coefficient in the film. A diffusion model that accounts for varying film thickness and evaporation
at the surface has been used to follow the water loss and thickness data over time, in order to assess the
relative importance of diffusion and evaporation under various conditions.

Introduction to an increased interaction spectrum, that is, a more complex


hydrophilic/hydrophobic balance between the parts of the
Amphiphilic molecules, such as surfactants, lipids, and block
molecule that range from the very hydrophobic PE block to
copolymers, are able to self-assemble in various ordered
the hydrophilic PEO block through the moderately hydrophobic
structures in the presence of selective solvents.1,2 The structures
PPO block. Such surfactants with “graded” triblock composition
attained by an amphiphile depend on a “preferred curvature”,
are found useful in several applications.24-27 For example, nonyl
which is initially set by the amphiphile molecular architecture
phenyl-propoxy-ethoxylate surfactants with long PEO blocks
but is also affected by the type and amount of solvent and solutes
have been used for the stabilization of aqueous dispersions.24
present.1-5 The spontaneous formation of supramolecular
Surfactants of the alkyl-propoxy-ethoxy sulfate type have been
structures based on amphiphiles has attracted increasing interest
used in enhanced oil recovery applications, where their graded
due to its inherent beauty and its wealth of current and potential
nature offers an advantage in stabilizing oil-water interfaces
technological applications.1,2 The great majority of studies
over alkyl ethoxylates.25,26
published in the open literature address systems at thermody-
In this study, we investigate the drying mechanism of alkyl-
namic equilibrium. Yet, mass transport properties are of great
propoxy-ethoxylate surfactant solutions by gravimetrically
importance in the utilization of the amphiphilic molecules.6 For
monitoring the water loss in air of controlled relative humidity
example, mass transport is relevant in the surfactant dissolution
(see schematic of Figure 1). In particular, we examine the effects
process,7 and in drug delivery using carriers containing am-
of the drying environment, the surfactant hydrogel microstruc-
phiphilic molecules.8-10 Transport properties such as evapora-
ture, and the surfactant “graded” composition. Analytical and
tion rates are of interest in the assessment of hazards arising
numerical solutions of the diffusion equation are used to obtain
from volatile chemicals, in drying processes (e.g., during
the water diffusion in the surfactant hydrogel and to assess the
preparation of powders), and in the release of volatile active
relative importance of diffusion in the film and surface evapora-
species such as perfumes and flavors from commercial
tion. This is one of very few studies on evaporation from
products.11,12 Very few reports exist in the literature on the
structured surfactant solutions and the first study done on the
evaporation rate of water in structured surfactant solutions.11,12
transport properties of alkyl-propoxy-ethoxylate surfactants.
Previous work in our group has examined the drying mechanism
of hydrogels formed by poly(ethylene oxide)-poly(propylene
Materials and Methods
oxide)-poly(ethylene oxide) (PEO-PPO-PEO) amphiphilic block
copolymers.13,14 Materials. Three different alkyl-propoxy-ethoxylate surfac-
Block copolymers of the PEO-PPO-PEO type find numerous tants with varying number, 8.5, 17, and 34, of ethylene oxide
applications in industry and research due to the large variety of (EO) segments were used as received from Dow Chemical
physicochemical properties that can be attained by varying the Company, Midland, MI (from here on, the surfactants will be
PEO/PPO ratio and relative block length, as well as the solution denoted as “8.5 EO”, “17 EO”, and “34 EO”, respectively).
conditions.2,15-17 Lately, ABC-type triblock terpolymers have These surfactants can be considered as short ABC triblock
attracted attention due to their ability to adopt complex terpolymers, comprised of polyethylene, poly(propylene oxide),
architectures.18,19 Following a comprehensive investigation on
the thermodynamics, structure, and dynamics of the PEO-PPO
type block copolymers,2,3,5,14-17 we turn our attention to ABC-
type triblock terpolymers, in particular amphiphiles which
consist of short PEO, PPO, and polyethylene (PE) blocks.20-23
The triblock architecture of these surfactants is expected to lead
* To whom correspondence should be addressed. E-mail:
palexand@buffalo.edu.

Current Affiliation: Department of Chemical Engineering, Univer- Figure 1. Schematic of the aqueous surfactant solution film undergoing
sity of Massachusetts, Lowell, MA 01854. evaporation.

10.1021/ie100261u  2011 American Chemical Society


Published on Web 07/28/2010
Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 581
and poly(ethylene oxide) blocks. The “8.5 EO”, “17 EO”, and
“34 EO” surfactants have molecular weights of 1266, 1641, and
2390 g/mol, respectively, and PEO contents of 30, 46, and 63
wt %, respectively. On the basis of their molecular weights and
chemical composition, “8.5 EO”, “17 EO”, and “34 EO” can
be represented by the formulas (C)13(PO)12.2(EO)8.5, (C)13-
(PO)12.2(EO)17, and (C)13(PO)12.2(EO)34, respectively (where C
stands for methylene, PO for propylene oxide, and EO for
ethylene oxide segment). The densities of the three surfactants
are approximately 1.06 g/cm3. All salts and sample preparation
method used in this study are the same as reported by Gu and
Alexandridis.13,14
Drying Experiments. The driving force for the solvent
(water) transfer from one phase to another is the difference in
the solvent chemical potential between the different phases. In
our study, the water chemical potential, ∆µ, at each air relative
humidity condition examined is higher than that in the hydrogel
sample (initially, 20 wt % surfactant in water), thus drying is Figure 2. Equilibrium water concentration (wt %) as a function of water
achieved until equilibrium can be attained. ∆µ can be determined vapor activity (P/P0) for (O) “8.5 EO”, (4) “17 EO”, and (0) “34 EO”
by the water vapor pressure in the air, p, as follows28,29 surfactants.

∆µ ) RTln ()
p
p0
) RTln(RH) (1)
Different ordered structures are formed at different relative
humidity conditions. At the experiment temperature (24 °C),
the hexagonal structure forms at RH approaching 100%, lamellar
where p0 is the saturated water vapor pressure and RH is the structure at 94-97% RH, and high-polymer-content phase at
relative humidity. Air of constant RH can be generated by 8-84% RH for the “8.5 EO” hydrogel. For the “17 EO”
saturated aqueous salt solutions.30 Saturated solutions of LiCl, hydrogel, the micellar cubic phase forms at 100% RH, the
NaBr, NaCl, KCl, KNO3, and K2SO4 were used to generate air hexagonal phase at 97% RH, the lamellar phase at 75-94%
RH of 10.9, 57.2, 75.3, 84.5, 93.8, and 97.4%, respectively. RH, and the high-polymer-content phase at 8-64% RH. The
The methodology and the data acquisition method used in this “34 EO” hydrogel exhibits micellar cubic phase at 97-100%
study are the same as reported by Gu and Alexandridis.14 In RH, hexagonal phase at 85-94% RH, bicontinuous cubic phase
brief, 20% aqueous surfactant solution was poured into a dish at 81% RH, and semicrystalline pastelike phase at 8-75% RH.
to form a film with an initial thickness of 5 mm (much smaller The above correspondence between relative humidity and
than the dish diameter so that the drying can be considered one- structure is based on separate studies of the equilibrium phase
dimensional). The samples were placed in a chamber containing behavior and structure as a function of composition (fixed water
still air of constant RH (maintained by saturated aqueous salt content). The phase boundaries thus determined at equilibrium
solutions), and the sample weight was monitored as a function are denoted with dotted lines in Figure 3. While some hysteresis
of time. In order to minimize the disturbance to RH of opening may be possible during drying, when the local composition
the system (sealed dishes) to take measurements, the sealed changes drive a phase change, the time scale of the drying
dishes with the higher RH (75.3, 84.5, 93.8, and 97.4%) were experiments is much greater than the time scale of phase
placed in a Glovebag model X-37-37H (I2R, Cheltenham, PA) transitions in similar surfactant lyotropic liquid crystalline
that maintained a high RH environment. For the lower RH cases systems. During drying, the chemical and physical properties
(10.9 and 57.2%), the sealed dishes were kept in an air- change,14,32 and thus it is our interest to investigate the water
conditioned room with temperature at 24 °C ((1 °C) and transport properties within the surfactant hydrogel until equi-
ambient RH ranging from 30 to 60%. librium is reached.
Time Evolution of Water Loss. The hydration isotherms
presented above establish the water content of the hydrogel films
Results and Discussion
in contact with air for different RH conditions at equilibrium
Equilibrium Water Content: Hydration Isotherm. De- (infinite time). In order to understand the solvent (water in this
pending on the relative humidity, the equilibrium water case) transport process, we monitored the kinetics of water loss
concentration of the hydrogel is different and corresponds from the hydrogel film. Figure 3 shows the evolution of the
to different ordered structures. The physical and chemical alkyl-propoxy-ethoxylate surfactant concentration (average over
stability of the surfactant is influenced by the water activity the whole film) as a function of drying time at the 10.9, 57.2,
(and correspondingly by the air RH). Hydration isotherms 75.3, 84.5, 93.8, and 97.4% air RH. For all three surfactants
are of great relevance to the drying experiments because they considered here, at the very start of the drying process, a slow
assist in setting the end-point of the drying process and in increase of the surfactant concentration is observed, followed
determining the packaging and storage requirements for a by a rapid increase. As the surfactant hydrogels approached their
desired shelf life.31 equilibrium concentration, the increase slows down and then
The weight of the hydrogel that has been measured gravi- reaches a plateau, where the surfactant concentration did not
metrically was converted into water concentration in weight change further. As expected, the equilibrium is reached faster
percent and is presented in the form of an adsorption isotherm at the lower RH conditions.
in Figure 2. The water concentration values reported are the The total water weight loss at any given time is defined as
average values over the whole film. At lower RH, the surfactant the difference between the water weight of the hydrogel film at
hydrogel retains very low water content. As the air RH increases, the start of the drying process and the water weight of the
the equilibrium water concentration in the hydrogel increases. hydrogel at any given time. The water weight loss data shown
582 Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

Figure 4. Total water loss (gram/square meters) at any given time, plotted
Figure 3. Average surfactant concentration in the film as a function of drying
as a function of drying time at air RH of (O) 10.9%, (4) 57.2%, (0) 75.3%,
time at air RH of (O) 10.9%, (4) 57.2%, (0) 75.3%, (]) 84.5%, (3) 93.8%,
(]) 84.5%, (3) 93.8%, and (right pointing open triangle) 97.4% for the (a)
and (right pointing open triangle) 97.4% for the (a) “8.5 EO”, (b) “17 EO”,
“8.5 EO”, (b) “17 EO”, and (c) “34 EO” surfactants. The solid lines represent
and (c) “34 EO” surfactants. The dotted lines represent phase boundaries
the numerical solution to the diffusion model for the air RH conditions of
at 24 °C. L1 f homogeneous liquid solution; I1 f micellar cubic lyotropic
10.9-84.5%.
liquid crystal; H1 f hexagonal lyotropic liquid crystal; V1 f bicontinuous
cubic lyotropic liquid crystal; LR f lamellar lyotropic liquid crystal; L2 f
high-polymer content solution. is observed for all three surfactants at all the RH conditions
considered here.
in Figure 4 are calculated as water weight lost per surface area
of the hydrogel film exposed to the air (which is a constant Drying Rate. The drying rate is defined as the derivative of
value of 9.62 × 10-4 m2 for our experiments). The weight of water loss with respect to drying time. The drying rate is
water lost from the hydrogel film increases initially linearly with constant at the initial stages of drying (stage I), followed by a
the drying time, followed by a nonlinear increase, until it reaches drop (stage II of the drying process). This constant drying rate
a constant value. A two-stage mechanism13,33,34 can be discerned at each air RH is plotted in Figure 5. A linear trend is observed
in the water weight loss curve: the linear region of water weight when the drying rate at stage I is plotted as a function of the air
loss versus time corresponds to stage I, and the nonlinear region relative humidity for all three surfactants. The drying rate drops
corresponds to stage II of the drying. The two-stage mechanism by about 1/2 as the air RH increased from 10.9% to 57.2% RH.
Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 583

Figure 5. Drying rate of stage I as a function of air RH for the (O) “8.5
EO”, (4) “17 EO”, and (0) “34 EO” surfactants.

Stage I of the drying process, also referred to as the constant-


rate period, is the first major drying period.35 Since the initial
water content in the surfactant hydrogel film is very high (80
wt %), the surfactant film is a homogeneous liquid solution
where water molecules can move relatively freely. Thus, the
drying process in stage I is limited by the water evaporation
from the film surface.14,35 As shown in Figure 5, the external
condition (i.e., air RH) is the primary factor affecting the water
evaporation rate of the surfactant hydrogel, since the drying rates
of all three surfactants at state I almost overlap with each other
at each air RH. As the drying proceeds, the morphology of the
hydrogel film starts to change and both the surface water
evaporation and the water transport inside the structured
hydrogel would affect the drying rate, which corresponds to
stage II of drying.14
A two stage drying mechanism was observed in the drying
of hydrogel films formed by Pluronic P105 and F127 PEO-
PPO-PEO block copolymers13,14 and in other polymer solu-
tions.34 A linear drop in the drying rate at stage I as the air RH
increases has also been observed in the Pluronic F127-water
and Pluronic P105-water systems.14
Structural Effects on Drying Rate. Studies have shown that
the permeability or molecule diffusion is affected by ordered
morphologies formed by block copolymers36 or semicrystalline
polymers.37 The microstructure of the three alkyl-propoxy-
ethoxylate surfactants considered here may have an influence
on the drying rate. To examine this, we plot in Figure 6 the
drying rate as a function of the water content, which spans
several phases (and corresponding microstructure). For the case
of the “8.5 EO”-water system (Figure 6a), the drying rate Figure 6. Drying rate as a function of water content at air RH of (O) 10.9%,
(4) 57.2%, (0) 75.3%, (]) 84.5%, (3) 93.8%, and (right pointing open
remains constant in the homogeneous liquid solution and the triangle) 97.4% for the (a) “8.5 EO”, (b) “17 EO”, and (c) “34 EO”
hexagonal phase and decreases starting from the lamellar phase surfactants. The dotted lines represent the phase boundaries. The notation
until the high-polymer-content phase, where a greater drop in used for the different phases is the same as in Figure 3.
the drying rate is observed. The same trend is observed in the is consistent with our finding that the initial stages of the drying
case of the “17 EO”-water system (Figure 6b); however, the process are determined by the air RH, regardless of the surfactant
decrease in the lamellar phase is steeper than that in “8.5 EO”- type. In the intermediate stages of the drying, a trend is observed
water system. In the case of “34 EO”-water system, the drying where the “8.5 EO” surfactant system has the highest drying
rate remains constant in the homogeneous liquid solution and rate, followed by the “17 EO” system, and then the “34 EO”
the micellar cubic phase and decreases starting from the system. As shown in the hydration isotherm (Figure 2), there
hexagonal phase. is higher water content at the end of the drying process for the
To compare the surfactants at the same air RH, the drying higher PEO surfactant, and therefore there is less water diffusing
rates for all three surfactants at 57.2 and 84.5% RH are plotted out of the hydrogel film and a lower drying rate. Studies on
as a function of water content in parts a and b of Figure 7, PEO-PPO-PEO block copolymers have shown that the drying
respectively. At the initial stages of drying (high water content), rate is lower when the PEO content of the block copolymer is
the drying rates of the three surfactants are very similar. This higher.14
584 Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

Figure 7. Drying rate as a function of (a, b) water content, (c, d) “corrected” water content when assuming the hydrophobic domain to consist of both the
alkyl part and the PPO part, and (e, f) when assuming the hydrophobic domain to consist of only the alkyl chain, at (left panels) 57.2% RH and (right panels)
84.5% RH, for the (O) “8.5 EO”, (4) “17 EO”, and (0) “34 EO” surfactants.

As mentioned previously, the surfactant self-assembled intermediate solvent for poly(propylene oxide), and a bad solvent
structure may be a factor in determining the drying rate. In parts for the alkyl chain.21 Therefore the hydrophobic domains can
a and b of Figure 7, the drying rate for all three surfactants be considered to consist of either only alkyl chains or both the
starts to decrease at approximately the same water content, alkyl and PPO parts of the surfactant. The “corrected” water
implying that the water content or the hydration level of the content, H2O*, when assuming the hydrophobic domains to
surfactant may be important. To further examine this point, we consist of the alkyl and the PPO parts is defined as14
tried to decouple the self-assembled structure from the hydration
level. The intermolecular interactions in surfactant hydrogels H2O wt %
H2O* ) (2)
originate mainly from interactions between the hydrophilic (100 - H2O wt %)PEO wt % + H2O wt %
segments and water.38 Taking this into account, we attempt to
“correct” the water content by factoring out the hydrophobic and when assuming hydrophobic domains to consist of only
domains. Water is a good solvent for poly(ethylene oxide), the alkyl chains is defined as
Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 585

H2O )
* time expansions of this trigonometric series lead to eqs 5 and
H2O wt % 6, respectively.40,41
(3)

( )
(100 - H2O wt %)(PPO wt % + PEO wt %) + H2O wt % ∞
Mt -(2n + 1)2π2Dt
M∞
)1- ∑ (2n +8 1) π 2 2
exp
l2
(5)
where H2O wt % is the water content in the surfactant hydrogel n)0
and PEO wt % and PPO wt % are the PEO and PPO contents

( )

Mt
of the surfactants, respectively. The drying rates at 57.2% and 4 Dt 8
∑ nl
0.5

84.5% RH are plotted as a function of the hydration level or M∞


)
l π ( ) + (Dt)0.5
l n)1
(-1)n ierfc
2(Dt)0.5
(6)
“corrected” water content when assuming the hydrophobic
domains to consist of the alkyl and the PPO parts in Figure where Mt is the water lost per unit area of film at time t, M∞ is
7c,d and when assuming hydrophobic domains to consist of only the maximum water loss, and l is the film thickness. Equation
the alkyl chains in Figure 7e,f. 5 converges rapidly at long times, whereas eq 6 converges
If the hydration level were to play a more important role than rapidly at short times.40 In what follows we present methods
the surfactant microstructure in the hydrogel film, then the which are based on eq 5 or 6 for extracting a diffusion coefficient
drying rates of all three surfactants should overlap when plotted from water vapor sorption/drying data.
as a function of the “corrected” water content. Indeed, the drying Method 1. One method to estimate the water diffusion
rates at 57.2 and 84.5% RH overlap fairly well when the coefficient is from the measurement of the sorption/drying half-
hydrophobic domains is taken to consist of both alkyl and PPO time, t0.5, obtained from a two-term approximation of eq 5:40,41
parts (Figure 7c,d), indicating that the hydration level does affect
the drying rate. When the hydrophobic domains are assumed 0.04919l2
t0.5 ) (7)
to consist of only the alkyl chains, then no overlap is observed Dh
(Figure 7e,f). This indicates that PPO is acting as hydrophobic
rather than as hydrophilic. where the sorption/drying half-time is defined as the time where
Studies on drying rates of PEO-PPO-PEO block copolymers Mt/M∞ ) 0.5.
and PEO homopolymers have also shown that the hydration Method 2. At short times, (Mt/M∞ < 0.5), the following
level rather than the microstructure is the determining factor expression can be obtained from the first term of eq 6:40,41
for the drying rate (note that the PEO-PPO-PEO drying rate
data points overlapped very well with the PEO homopolymer
data points).14 A similar approach was used when plotting the
Mt
M∞
)
4 DIt
l π ( ) 0.5
(8)
osmotic pressure as a function of the PEO concentration, which
is calculated based on the block copolymer concentration and The water diffusion coefficient (DI) can be calculated from the
PEO content in the block copolymer molecules.38 The osmotic initial slope of a plot of Mt/M∞ vs t0.5.
pressure data points for the PEO-PPO-PEO block copolymers Method 3. At longer times (Mt/M∞ > 0.5), the following
overlapped very well with the PEO homopolymer data points. expression can be obtained from the one-term approximation
This indicates that the osmotic pressure originates mainly from of eq 5:40,41
interactions between PEO and water.38 The observation that the
microstructure does not play a dominant role in the transport
properties possibly emanates from the water-continuous topol-
ogy in the systems considered. Even in the anisotropic lamellar
(
ln 1 -
Mt
M∞ ) 8
) ln 2 -
π
π2DFt
l2
(9)

phase, structural defects may allow the movement of water The water diffusion coefficient (DF) can be calculated from the
perpendicular to the (hydrophobic) surfactant layers. limiting slope of ln(1 - Mt/M∞) vs t.
Water Diffusion (Analytical Solutions for Limiting Method 2 was used for fitting data sets at short times, whereas
Cases). As mentioned previously, the drying rate is affected method 3 was used for fitting data sets at longer times. The
by water transport (diffusion) inside the structured hydrogel. fitting of the analytical solution to the drying data using method
The water diffusion can be quantified by extracting the water 2 is shown in Figure 8. The water diffusion coefficient (DI)
diffusion coefficient from the experimental water loss data. The values of all three surfactants for 93.8% and 97.4% RH were
(one-dimensional) water transport in the surfactant hydrogel can calculated based on the M∞ from the equilibrium data (Figure
be described by Fick’s second law:39 2). The water diffusion coefficients at 10.9%, 57.2%, 75.3%,
and 84.5% RH for all three surfactants using methods 1, 2, and
∂C ∂ ∂C 3 were calculated based on the kinetics data (Figure 3). When
∂t
) D
∂x ∂x( ) (4)
using method 2, the first few data points were omitted because
they were at the initial stages of the drying process, and when
where t is the drying time, x is the direction of the water using method 3, the last few data points were omitted because
diffusion (normal to the surface of the film), D is the diffusion they were close to the equilibrium states. Table 1 reports the
coefficient of water, and C is the water concentration in the water diffusion coefficient values extracted via eqs 7-9. The
surfactant hydrogel film. The water diffusion coefficient can diffusion coefficient obtained for the “8.5 EO”, “17 EO”, and
be obtained using an analytical solution of eq 4.13,39,40 Several “34 EO” systems at each air relative humidity from methods 1,
assumptions are made to this end: (i) the drying process is one- 2, and 3 were fairly similar.
dimensional, (ii) the drying process is isothermal, (iii) the film With the comparison of the diffusion coefficients values of a
thickness is constant, (iv) the water diffusion coefficient is given surfactant at varying RH, a general trend is that the
constant, and (v) the concentration at the surface of the hydrogel diffusion coefficient decreases as the air relative humidity
film is the same as inside the film. On the basis of these initial/ increases. When the three surfactants are compared at the same
boundary conditions and assumptions, a trigonometric series RH, the diffusion coefficient values obtained from methods 1
analytical solution of eq 4 can be obtained. Long time and short and 2 were very similar for all three surfactants at each air RH.
586 Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

Table 1. Water Diffusion Coefficients in Surfactant Hydrogels


Obtained from Analytical Solutions
Diffusion coefficient (m2/s)
surfactant air RH (%) Dha DIb DFc
“8.5 EO” 10.9 4.5 × 10-12 1.1 × 10-11 1.4 × 10-10
57.2 2.2 × 10-12 4.2 × 10-12 1.0 × 10-10
75.3 1.3 × 10-12 2.1 × 10-12 2.4 × 10-11
84.5 7.6 × 10-13 1.5 × 10-12 5.9 × 10-12
93.8 5.6 × 10-13
97.4 2.0 × 10-13
“17 EO” 10.9 4.7 × 10-12 1.0 × 10-11 1.4 × 10-10
57.2 2.1 × 10-12 4.5 × 10-12 3.9 × 10-11
75.3 1.2 × 10-12 3.0 × 10-12 8.6 × 10-12
84.5 7.7 × 10-13 1.5 × 10-12 5.7 × 10-12
93.8 4.8 × 10-13
97.4 2.6 × 10-13
“34 EO” 10.9 4.7 × 10-12 1.1 × 10-11 1.0 × 10-11
57.2 1.9 × 10-12 4.0 × 10-12 7.8 × 10-12
75.3 1.2 × 10-12 2.8 × 10-12 7.7 × 10-12
84.5 8.1 × 10-13 1.3 × 10-12 5.0 × 10-12
93.8 5.1 × 10-13
97.4 1.1 × 10-13
a
Value extracted from eq 7. b Value obtained from fitting eq 8 to the
experimental results at shorter times. c Value obtained from fitting eq 9
to the experimental results at longer times.

2.3 × 10-12 m2/s at 94% RH when calculated using methods


1, 2, and 3.41 The order of magnitude of these diffusion
coefficients is comparable to the findings of the present study.
The utilization of analytical solutions discussed above has merit
on the basis of the simplicity in calculation and also as a means
of comparison of our study to several published studies that
utilize such analytical solutions. A numerical solution allows
the use of transport equations that are a more accurate
representation of physical phenomena, thus we have used such
approach in our work.
Water Diffusion (Numerical Solution). To better analyze
the evaporation process, we solved the diffusion equation with
a numerical model that accounts for the water diffusion in the
film and varying film thickness. If we consider the water vapor
pressure in contact with the film to be constant at all times and
the film thickness to decrease with time (i.e., moving boundary
condition), the boundary condition at the air-film interface is
obtained by using the jump mass balance:14,42
∂C dh(t)
-D -C ) R(C - C∞), at x ) h(t) (10)
∂t dt
where h(t) is the film thickness, the term C(dh(t)/dt) accounts
for the water flux due to the moving boundary, and R is a
constant that represents the product of the mass transfer
coefficient and the proportionality constant.14 The boundary
Figure 8. Fitting of the analytical solution (eq 8) at data obtained for air
RH of (O) 10.9%, (4) 57.2%, (0) 75.3%, (]) 84.5%, (3) 93.8%, and (right condition at the hydrogel film-dish interface is
facing open triangle) 97.4% for the (a) “8.5 EO”, (b) “17 EO”, and (c) “34
EO” surfactants. DI represents the diffusion coefficient value (meters ∂C
D ) 0, at x ) 0 (11)
squared/second) obtained at each RH. ∂x
In the case of diffusion coefficient values obtained from method
The initial condition is
3, a trend is observed where “8.5 EO” has the highest value,
followed by “17 EO”, and the lowest value was for “34 EO”. C ) C0, at t ) 0 for 0 > x > L0 (12)
Method 3 is calculation based on longer time scales, where the
three surfactants possess different microstructures. Therefore, where C0 is the initial water concentration and L0 is the initial
a difference in the apparent diffusion coefficient is observed. film thickness. The water diffusion coefficient is considered to
The methods mentioned above were used in obtaining the be a function of water concentration in the hydrogel film:14,43
water diffusion coefficient in other systems such as Pluronic
P105.41 For this PEO-PPO-PEO block copolymer, the diffusion D ) D0e-k/C (13)
coefficients were in the range 4.8 × 10-12 to 7.3 × 10-12 m2/s
at 38% RH, 2.9 × 10-12 to 5.2 × 10-12 m2/s at 58% RH, 1.2 where D0 is the pre-exponential factor and k is a constant with
× 10-12 to 2.4 × 10-12 m2/s at 85% RH, and 1.3 × 10-12 to the same units as the water concentration. This model (eq 4
Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 587
Table 2. Parameters Obtained from Fitting the Diffusion Model to
Experimental Water Loss Data
R (m/s)
surfactant 10.9% RH 57.2% RH 75.3% RH 84.5% RH
-9 -9 -9
“8.5 EO” 9.0 × 10 4.5 × 10 2.6 × 10 1.7 × 10-9
“17 EO” 9.0 × 10-9 4.4 × 10-9 2.6 × 10-9 1.8 × 10-9
“34 EO” 9.0 × 10-9 3.8 × 10-9 2.8 × 10-9 2.2 × 10-9

evolution. D0 was set at the self-diffusion coefficient of neat


water, 2.23 × 10-9 m2/s at 25 °C,44 and k was set (after several
trials) to an optimum value of 0.1. The R impacted more on
the fitting of the initial stages of the drying process (ap-
proximately 1/3 of the total drying time). As mentioned earlier,
the initial stage of the drying mechanism is governed by
evaporation at the surface of the film, and this can be confirmed
by observing the R values (R is directly related to the mass
transfer coefficient14) found in Table 2. For all surfactants
considered here, R decreases as the air RH increases. The R
values are the same (or very similar) for the three surfactants
at each air RH. A similar trend of R being a function of air RH
has also been observed in the PEO-PPO-PEO block copolymer-
water systems.13,14
The mechanism of solvent drying from amorphous polymers,
such as benzene,45,46 ethyl acetate,45 ethyl methyl ketone,47 and
ethyl benzene48 from polystyrene, has been reported. In these
systems, the drying curve resembles a viscoelastic (non-Fickian)
diffusion curve. For example, the drying of ethyl methyl ketone
in atactic polystyrene exhibited a sequence of “sigmoid” type
f “pseudo-Fickian” type f “two-stage” type f “pseudo-
Fickian”type f ”Fickian” type characteristics.47 The term
“sigmoid” describes an S-shaped behavior of solvent loss versus
square root of time. “Pseudo-Fickian” describes a linear increase
of solvent loss with the square root of time (as predicted by
Fick’s law) with a prolonged approach to equilibrium. The “two-
stage” behavior is when the solvent loss appears to reach an
equilibrium level rather quickly but subsequently relaxes upward
from the equilibrium value over a long time scale.48 A similar
viscoelastic drying mechanism is observed for the three sur-
factant systems considered here, where the water loss curves
can be described as following a “sigmoid” f “pseudo-Fickian”
f “two-stage” behavior. The Fickian diffusion model and the
experimental data overlap very well at the initial stages of the
drying, in particular at the “sigmoid” and “pseudo-Fickian”
regions. However, the numerical solution does not capture the
water loss data as the surfactant solution approaches equilibrium
(“two-stage” behavior region), indicating that viscoelastic-type
effects are relevant in this region.
The Biot number, Bi, often used in heat transfer where it
Figure 9. Film thickness as a function of drying time at air RH of (O) represents the ratio of conduction to convection heat resistance,
10.9%, (4) 57.2%, (0) 75.3%, (]) 84.5%, (3) 93.8%, and (right facing
open triangle) 97.4% for the (a) “8.5 EO”, (b) “17 EO”, and (c) “34 EO” is also relevant to drying processes. In this case, the Biot number
surfactants. The solid lines represent the numerical solution to the diffusion represents the ratio of the internal to external mass transfer
model for the air RH conditions of 10.9-84.5%. resistance:49,50
with the boundary conditions of eqs 10 and 11) was solved
RL0
numerically to fit the experimental data in order to extract the Bi ) (14)
water diffusion coefficient and the mass transfer coefficient. The D0
water loss as a function of time data were fitted using this
numerical model, and the film thickness as a function of time where L0 is the initial surfactant hydrogel film thickness and
was predicted using the parameters obtained from the fitting. D0 is the pre-exponential factor of the water diffusion coefficient.
The solid lines shown in Figures 4 and 9 are the fitting results When Bi , 1, drying is limited by external mass transfer
for water loss and film thickness, respectively, obtained from resistance, which is the water evaporation at the surface of the
the numerical solution. Table 2 summarizes the parameters surfactant hydrogel. When Bi . 1, drying is limited by internal
obtained from the numerical fit. mass transfer resistance, which is the water diffusion in the
Three parameters (D0, R, and k) are needed for the numerical hydrogel film. When Bi ≈ 1, drying is dictated by both external
model to predict the water loss data and film thickness time and internal resistance.49,50
588 Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011

The Biot numbers were found to be on average 0.0202 at (3) Ivanova, R.; Lindman, B.; Alexandridis, P. Modification of the
10.9% RH, 0.0095 at 57.2% RH, 0.0060 at 75.3% RH, and lyotropic liquid crystalline microstructure of amphiphilic block copolymers
in the presence of cosolvents. AdV. Colloid Interface Sci. 2001, 89-90,
0.0042 at 84.5% RH. The Biot number is less than unity for all 351–382.
three surfactant hydrogel films, indicating that drying is limited (4) Alexandridis, P.; Olsson, U.; Lindman, B. A record nine different
by the water evaporation at the surface. Note that these findings phases (four cubic, two hexagonal, and one lamellar lyotropic liquid
correspond to early stage drying (stage I). When the drying crystalline and two micellar solutions) in a ternary isothermal system of an
proceeds, especially when it reaches stage II, both evaporation amphiphilic block copolymer and selective solvents (water and oil).
Langmuir 1998, 14 (10), 2627–2638.
and water diffusion have an effect and the Bi number is expected (5) Svensson, B.; Olsson, U.; Alexandridis, P. Self-assembly of block
to increase to a relatively larger number (L0 will decrease but copolymers in selective solvents: Influence of relative block size on phase
D0 will decrease even further). behavior. Langmuir 2000, 16 (17), 6839–6846.
(6) Gradzielski, M. Kinetics of morphological changes in surfactant
systems. Curr. Opin. Colloid Interface Sci. 2003, 8 (4-5), 337–345.
Concluding Remarks (7) Chen, B.-H.; Miller, C. A.; Walsh, J. M.; Warren, P. B.; Ruddock,
J. N.; Garrett, P. R.; Argoul, F.; Leger, C. Dissolution rates of pure nonionic
The drying of aqueous solutions of three alkyl-propoxy- surfactants. Langmuir 2000, 16 (12), 5276–5283.
ethoxylated surfactants with varying PEO content has been (8) Yang, L.; Alexandridis, P. Physicochemical aspects of drug delivery
investigated here. Films of surfactant solutions were exposed and release from polymer-based colloids. Curr. Opin. Colloid Interface Sci.
2000, 5 (1-2), 132–143.
to air of known water vapor pressure until equilibrium was (9) Yang, L.; Alexandridis, P. Controlled release from ordered micro-
achieved and the drying kinetics monitored. A two stage drying structures formed by poloxamer block copolymers. ACS Symp. Ser. 2000,
mechanism with a constant drying rate region (stage I) followed 752, 364–374.
by a drop in drying rate (stage II) was observed. The drying (10) Narasimhan, B. Mathematical models describing polymer dissolu-
rate at stage I was a linear function of the air relative humidity tion: Consequences for drug delivery. AdV. Drug DeliVery ReV. 2001, 48
(2-3), 195–210.
but depended very little on the surfactant type. (11) Beverley, K. J.; Clint, J. H.; Fletcher, P. D. I. Evaporation rates of
These novel alkyl-propoxyl-ethoxylate surfactants are capable structured and non-structured liquid mixtures. Phys. Chem. Chem. Phys.
of self-assembling into a variety of lyotropic liquid crystal 2000, 2 (18), 4173–4177.
ordered structures, depending on the water content. To probe (12) Aranberri, I.; Binks, B. P.; Clint, J. H.; Fletcher, P. D. I. Evaporation
rates of water from concentrated oil-in-water emulsions. Langmuir 2004,
the effect of surfactant microstructure on the drying rate, the 20, 2069–2074.
drying rate was plotted as a function of water content. The (13) Gu, Z.; Alexandridis, P. Drying of Poloxamer hydrogel films.
drying rate started decreasing at approximately the same water J. Pharm. Sci. 2004, 93 (6), 1454–1470.
content for all three surfactants, suggesting that hydration level (14) Gu, Z.; Alexandridis, P. Drying of films formed by ordered
dictates the drying rate. The hydration effect was further poly(ethylene oxide)-poly(propylene oxide) block copolymer gels. Langmuir
2005, 21 (5), 1806–1817.
analyzed by plotting the drying rate as a function of the (15) Alexandridis, P. Poly(ethylene oxide)-poly(propylene oxide) block
“corrected” water content. The data for the three surfactants copolymer surfactants. Curr. Opin. Colloid Interface Sci. 1997, 2 (5), 478–
overlapped very well when plotted as a function of the 489.
“corrected” water content indicating that the hydration level (16) Alexandridis, P.; Andersson, K. Reverse micelle formation and
water solubilization by polyoxyalkylene block copolymers in organic
determines the drying rate when assuming that the hydrophobic solvent. J. Phys. Chem. B 1997, 101 (41), 8103–8111.
domain consists of both the alkyl and the PPO parts. In addition (17) Schmidt, G.; Richtering, W.; Lindner, P.; Alexandridis, P. Shear
to the importance of the hydration level, we can also conclude orientation of a hexagonal lyotropic triblock copolymer phase as probed
that the PPO part of the “graded” alkyl-propoxy-ethoxylate by flow-birefringence, small-angle light- and neutron scattering. Macro-
surfactants acts more hydrophobic than hydrophilic. molecules 1998, 31 (7), 2293–2298.
(18) Hadjichristidis, N.; Iatrou, H.; Pitsikalis, M.; Pispas, S.; Avgeropou-
Analytical solutions of Fick’s second law that describe the los, A. Linear and non-linear triblock terpolymers. Synthesis, self-assembly
water diffusion in the surfactant hydrogel have been fitted to in selective solvents and in bulk. Prog. Polym. Sci. 2005, 30 (7), 725–782.
the experimental water loss data in order to extract the water (19) Triftaridou, A. I.; Vamvakaki, M.; Patrickios, C. S.; Stavrouli, N.;
diffusion coefficient. The water diffusion coefficient values Tsitsilianis, C. Synthesis of amphiphilic (ABC)(n) multiarm star triblock
terpolymers. Macromolecules 2005, 38 (3), 1021–1024.
calculated from the half time, short time scales, and longer time (20) Sarkar, B.; Alexandridis, P. Alkyl propoxy ethoxylate “graded”
scales of the drying process are comparable. The diffusion surfactants in aqueous solutions: Micelle formation and structure. J. Phys.
equation was solved numerically to better assess the evaporation Chem. B 2010, 114 (13), 4485–4494.
at the high water content region. The mass transfer coefficient (21) Sarkar, B.; Lam, S.; Alexandridis, P. Micellization of alkyl-propoxy-
extracted from fitting the experimental data with the numerical ethoxylate surfactants in water-polar organic solvent mixtures. Langmuir
2010, 26 (13), 10532–10540.
method established that the initial stages of the drying are (22) Shusharina, N. P.; Alexandridis, P.; Linse, P.; Balijepalli, S.;
controlled by the drying conditions (i.e., the air RH). Gruenbauer, H. J. M. Phase behavior and structure of an ABC triblock
copolymer dissolved in selective solvent. Eur. Phys. J. E 2003, 10, 45–54.
(23) Shusharina, N. P.; Balijepalli, S.; Gruenbauer, H. J. M.; Alexan-
Acknowledgment dridis, P. Mean-field theory prediction of the phase behavior and phase
structure of alkyl-propoxy-ethoxylate “graded” surfactants in water: Tem-
Financial support from the National Science Foundation perature and electrolyte effects. Langmuir 2003, 19 (10), 4483–4492.
(NSF) is greatly appreciated. We thank Dr. S. Balijepalli and (24) Trochet-Mignard, L.; Taylor, P.; Bognolo, G.; Tadros, T. F.
Dr. H. J. M. Gruenbauer of Dow Chemical Co. for providing Concentrated coal-water suspensions containing nonionic surfactants and
the surfactants used in this study. polyelectrolytes. 2. Adsorption of nonyl phenyl propylene-oxide ethylene-
oxide on coal and the rheology of the resulting suspension. Colloids Surf.,
A 1995, 95 (1), 37–42.
Literature Cited (25) Milter, J.; Austad, T. Chemical flooding of oil reservoirs. 7. Oil
expulsion by spontaneous imbibition of brine with and without surfactant
(1) Evans, D. F.; Wennerstrom, H. The Colloidal Domain: Where in mixed-wet, low permeability chalk material. Colloids Surf., A 1996, 117
Physics, Chemistry, Biology, and Technology Meet; VCH Publishers: New (1-2), 109–115.
York, 1994. (26) Minana-Perez, M.; Graciaa, A.; Lachaise, J.; Salager, J.-L. Systems
(2) Alexandridis, P.; Lindman, B. Amphiphilic Block Copolymers: Self- containing mixtures of extended surfactants and conventional nonionics.
Assembly and Applications; Elsevier Science B. V.: Amsterdam, The Phase behavior and solubilization in microemulsion. 4th World Surfactants
Netherlands, 2000. Congress, 1996; Vol. 2, pp 226-234.
Ind. Eng. Chem. Res., Vol. 50, No. 2, 2011 589
(27) Bollmann, L.; Urade, V. N.; Hillhouse, H. W. Controlling interfacial (40) Balik, C. M. On the extraction of diffusion coefficients from
curvature in nanoporous silica films formed by evaporation-induced self- gravimetric data for sorption of small molecules by polymer thin films.
assembly from nonionic surfactants. I. Evolution of nanoscale structures in Macromolecules 1996, 29, 3025–3029.
coating solutions. Langmuir 2007, 23 (8), 4257–4267. (41) Gu, Z.; Alexandridis, P. Sorption and transport of water vapor in
(28) Tsao, Y.-h.; Evans, D. F.; Rand, R. P.; Parsegian, V. A. Osmotic amphiphilic block copolymer films. J. Dispersion Sci. Technol. 2004, 25
stress measurements of dihexadecyldimethylammonium acetate bilayers as (5), 619–629.
a function of temperature and added salt. Langmuir 1993, 9 (1), 233–241. (42) Price, P. E., Jr.; Cairncross, R. A. Optimization of single-zone drying
(29) Parsegian, V. A.; Rand, R. P.; Fuller, N. L.; Rau, D. C. Osmotic of polymer solution coatings using mathematical modeling. J. Appl. Polym.
stress for the direct measurement of intermolecular forces. Methods Enzymol. Sci. 2000, 78 (1), 149–165.
1986, 127, 400–416. (43) Ion, L.; Vergnaud, J. M. Process of drying a polymeric paint by
(30) Greenspan, L. Humidity fixed points of binary saturated aqueous diffusion-evaporation and shrinkage. Determination of the concentration-
solutions. J. Res. Natl. Bureau Stand., A: Phys. Chem. 1977, 81A (1), 89– dependent diffusivity. Polym. Test. 1995, 14, 479–487.
96. (44) Gillen, K. T.; Douglass, D. C.; Hoch, M. J. R. Self-diffusion in
(31) Sablani, S. S.; Myhara, M.; Mahgoub; Al-Attabi, Z. H.; Al- liquid water to -31°C. J. Chem. Phys. 1972, 57 (12), 5117–5119.
Mugheiry, M. M. Water sorption isotherms of freeze dried fish sardines. (45) Odani, H.; Kida, S.; Tamura, M. Diffusion in glassy polymers. III.
Drying Technol. 2001, 19 (3&4), 673–680. Temperature dependence and solvent effects. Bull. Chem. Soc. Jpn. 1966,
(32) Fortes, M.; Okos, M. R. Drying theories: Their bases and limitations 39 (11), 2378–2385.
as applied to foods and grains. AdV. Drying 1980, 1, 119–154. (46) Kishimoto, A.; Fujita, H.; Odani, H.; Kurata, M.; Tamura, M.
(33) Croll, S. G. Heat and mass transfer in latex paints during drying. Successive differential absorptions of vapors by glassy polymers. J. Phys.
J. Coat. Technol. 1987, 59 (751), 81–92. Chem. 1960, 64 (5), 594–598.
(47) Odani, H.; Hayashi, J.; Tamura, M. Diffusion in glassy polymers.
(34) Guigner, D.; Fischer, C.; Holl, Y. Film formation from concentrated
II. Effects of polymer-penetrant interaction; diffusion of ethyl methyl ketone
reactive silicone emulsions. 1. Drying mechanism. Langmuir 2001, 17,
in atactic polystyrene. Bull. Chem. Soc. Jpn. 1961, 34 (6), 817–821.
3598–3606.
(48) Billovits, G. F.; Durning, C. J. Linear viscoelastic diffusion in the
(35) McCabe, W. L.; Smith, J. C.; Harriott, P. Unit Operations of
poly(styrene)/ethylbenzene system: Differential sorption experiments. Mac-
Chemical Engineering; McGraw-Hill, Inc.: New York, 1993.
romolecules 1993, 26 (25), 6927–2936.
(36) Csernica, J.; Baddour, R. F.; Cohen, R. E. Morphological arrange-
(49) Cairncross, R. A.; Jeyadev, S.; Dunham, R. F.; Evans, K.; Francis,
ments of block copolymers that result in low gas permeability. Macromol-
L. F.; Scriven, L. E. Modeling and design of an industrial dryer with
ecules 1990, 23, 1429–1433.
convective and radiant heating. J. Appl. Polym. Sci. 1995, 58, 1279–1290.
(37) Ngui, M. O.; Mallapragada, S. K. Quantitative analysis of crystal- (50) Cairncross, R. A.; Durning, C. J. A model for drying of viscoelastic
lization and skin formation during isothermal solvent removal from polymer coatings. AIChE J. 1996, 42 (9), 2415–2425.
semicrystalline polymers. Polymer 1999, 40, 5393–5400.
(38) Gu, Z.; Alexandridis, P. Osmotic stress measurements of intermo- ReceiVed for reView February 2, 2010
lecular forces in ordered assemblies formed by solvated block copolymers. ReVised manuscript receiVed July 1, 2010
Macromolecules 2004, 37 (3), 912–924. Accepted July 12, 2010
(39) Crank, J. The Mathematics of Diffusion; Clarendon Press: Oxford,
U.K., 1970. IE100261U

You might also like