1155© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com

1. Introduction

Osmotic pressure is a crucial factor for determining the sta-

bility of microcapsules and biological systems enclosed by

lipid bilayers. The pressure difference between the interior

and exterior of such systems induces water fl ux through

Osmocapsules for Direct Measurement of Osmotic Strength

Shin-Hyun Kim , * Tae Yong Lee , and Sang Seok Lee

semipermeable membranes, defl ating or infl ating the com-

partments. [ 1 ] In particular, the high osmotic pressure dif-

ference frequently disintegrates delicate capsules and cells

by imposing signifi cant stress on the membrane. [ 2 ] There-

fore, measuring the osmotic strength is very important for

a wide range of research and applications that use semiper-

meable membranes, ranging from the fundamental study

of cell functions to drug delivery systems. To estimate the

strength, colligative properties of aqueous solutions, such

as freezing-point depression, vapor-pressure lowering, or

boiling-point elevation, have been measured and analyzed. [ 3 ]

However, such indirect measurements require elaborate

machines to measure the properties. Moreover, additional

analysis is sometimes required for non-ideal solutions such

Monodisperse microcapsules with ultra-thin membranes are microfl uidically designed to be highly sensitive to osmotic pressure, thereby providing a tool for the direct measurement of the osmotic strength. To make such osmocapsules, water-in-oil-in-water double-emulsion drops with ultra-thin shells are prepared as templates through emulsifi cation of core–sheath biphasic fl ow in a capillary microfl uidic device. When photocurable monomers are used as the oil phase, the osmocapsules are prepared by in-situ photopolymerization of the monomers, resulting in semipermeable membranes with a relatively large ratio of membrane thickness to capsule radius, approximately 0.02. These osmocapsules are buckled by the outward fl ux of water when they are subjected to a positive osmotic pressure difference above 125 kPa. By contrast, evaporation-induced consolidation of middle-phase containing polymers enables the production of osmocapsules with a small ratio of membrane thickness to capsule radius of approximately 0.002. Such an ultra-thin membrane with semi-permeability makes the osmocapsules highly sensitive to osmotic pressure; a positive pressure as small as 12.5 kPa induces buckling of the capsules. By employing a set of distinct osmocapsules confi ning aqueous solutions with different osmotic strengths, the osmotic strength of unknown solutions can be estimated through observation of the capsules that are selectively buckled. This approach provides the effi cient measurement of the osmotic strength using only a very small volume of liquid, thereby providing a useful alternative to other measurement methods which use complex setups. In addition, in-vivo measurement of the osmotic strength can be potentially accomplished by implanting these biocompatible osmocapsules into tissue, which is diffi cult to achieve using conventional methods.

Microcapsules

DOI: 10.1002/smll.201302296

Prof. S.-H. Kim, T. Y. Lee, S. S. Lee Department of Chemical and Biomolecular Engineering and KINC, KAIST Daejeon, 305–701 , Korea E-mail: kim.sh@kaist.ac.kr

small 2014, 10, No. 6, 1155–1162

S.-H. Kim et al.

1156 www.small-journal.com © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

full papersas concentrated solutions of polymers or biomaterials, which

can exhibit nonlinear behavior. [ 4 ] By contrast, the osmotic

pressure can be directly measured by a setup composed of

manometers and semipermeable membranes. [ 5 ] However, this

measurement requires relatively large quantities of samples.

Moreover, these conventional methods are not available for

in-vivo or in-situ measurement of the osmotic strength. There-

fore, a facile and direct measurement of the osmotic strength

using an injectable kit remains an important challenge.

Double-emulsion drops have provided useful templates

to produce microcapsules because of their core–shell geom-

etry. A variety of microcapsules have been prepared through

solidifi cation of the shell phases, resulting in a membrane

enclosing the cores. [ 2a , 6 ] The membranes can be semiperme-

able due to formation of very small pores that allow diffusion

of water molecules through it but not solutes. Microfl uidics

has enabled the production of such double-emulsion drops

with unprecedented controllability and very high effi ciency

of encapsulation. [ 7 ] Therefore, double-emulsion drops pro-

duced in microfl uidic devices can be useful

to make semipermeable microcapsules

that are sensitive to the osmotic pressure

difference across the membrane. [ 8 ]

In this paper, we report a microfl u-

idic approach to produce microcapsules

with ultrathin and semipermeable mem-

branes or osmocapsules, providing a facile

and direct measurement of the osmotic

strength. Using capillary microfl uidic

devices, we prepare water-in-oil-in-water

(W/O/W) double-emulsion drops with an

ultrathin middle phase, which transform

into polymeric microcapsules that contain

an aqueous solution with standard osmotic

strength upon solidifi cation of the middle

phase. Either UV-induced photopolymeri-

zation or evaporation-induced consolida-

tion of the middle phase is employed. The

resultant microcapsules are highly sensi-

tive to osmotic pressure differences due

to the semi-permeability and very small

thickness of the membrane. Therefore, a

small positive pressure can lead to buck-

ling of the capsules by an outward fl ux of

water through the membrane. Therefore,

when a mixture of distinctively labeled

microcapsules, containing different

standard osmotic solutions, are dispersed

in an unknown solution, the capsules are

selectively buckled when subjected to pos-

itive pressure, thereby enabling an estima-

tion of the osmotic strength. In addition,

the capsules potentially enable in-vivo

measurement of the osmotic strength by

injecting them. Moreover, the capsules

can be ruptured when they are subjected

to a high osmotic pressure difference,

providing the osmotic pressure-triggered

release of encapsulants.

2. Results and Discussion

2.1. Preparation of ETPTA Capsules

The microfl uidic device comprised two tapered cylindrical

capillaries that were coaxially aligned in a square capillary as

shown schematically in Figure 1 a, where the left cylindrical

capillary has a 70-µm diameter orifi ce and a hydrophobic

surface, whereas the right cylindrical capillary has a 190-µm

diameter orifi ce and a hydrophilic surface. A small tapered

capillary was inserted into the hydrophobic cylindrical capil-

lary. We injected the aqueous solution of poly(vinyl alcohol)

(PVA) and NaCl through the small tapered capillary to form

innermost drops and a photocurable monomer of ethoxylated

trimethylolpropane triacrylate (ETPTA) through the hydro-

phobic cylindrical capillary to form ultra-thin shells. These

two immiscible fl uids fl ow through a single capillary channel

of the hydrophobic injection capillary. Because of the hydro-

phobic nature of the capillary, ETPTA fl ows along the inner

Figure 1. a) Schematic illustration of the microfl uidic capillary device for preparation of water-in-oil-in-water (W/O/W) double-emulsion drops with ultra-thin shells. b) Optical microscopy image showing the generation of double-emulsion drops in a dripping mode. c) Confocal microscopy images of microcapsules with a diameter of 122 µm, of which the shell is made of ETPTA, containing an aqueous solution of PVA and NaCl with different osmolarities of 33 (red), 342 (orange), 649 (yellow), 950 (green) mOsm L −1 , respectively from left to right. d,e) Scanning electron microscopy images of the dried microcapsules. The cross-sectional image shows the thickness of the membranes, 1.05 (thin part) and 1.34 µm (thick part), as denoted with pairs of arrows, where the thick middle fi lm is dried PVA.

small 2014, 10, No. 6, 1155–1162

Osmocapsules for Direct Measurement of Osmotic Strength

1157www.small-journal.com© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

wall of the capillary, whereas the aqueous

solution fl ows in the shape of plug-like

drops without contacting the wall, forming

a discontinuous core–sheath stream. [ 6b , 9 ]

We injected the same aqueous solution

through the interstices of the hydrophobic

capillary and the square capillary to form

a continuous phase and this emulsifi ed the

core–sheath stream at the tip of the hydro-

phobic capillary into double-emulsion

drops with an ultra-thin shell, as shown

in Figure 1 b and Movie S1 of the Sup-

porting Information. [ 6b ] Although excess

ETPTA between the plug-like drops pro-

duced large oil blobs, separation of the

double emulsion was simply accomplished

by exploiting their density difference. This

discontinuous emulsifi cation generated a

small fraction of double-emulsion drops

with a relatively thick shell when both ends

of the plug-like drops were emulsifi ed at

the tip of the injection capillary; the frac-

tion therefore depends on the length of the

drops and is typically less than 10% of the

total double-emulsion drops. The ETPTA

monomers in the ultra-thin shell were

polymerized in a collection bath by UV

irradiation for 2 seconds over a 1 minute

interval. Using this in-situ polymeriza-

tion technique, all double-emulsion drops

generated in the microfl uidic device were

converted into stable microcapsules. The

resultant microcapsules with an average

diameter of 122 µm, containing the aqueous

solution of different osmolarity and dye,

are shown in Figure 1 c, where the capsules appearing in red,

orange, yellow, and green colors contain aqueous solutions of

33, 342, 649, and 950 mOsm L −1 , respectively. The thickness

of the ETPTA membrane was characterized by scanning elec-

tron microscopy (SEM), as shown in Figures 1 d and e where

we broke defl ated capsules by applying a mechanical force to

observe the cross-section of the membrane, and resulted in a

thickness of 1.05 µm and 1.34 µm for the thin and thick parts,

respectively. Therefore, the ratio of average thickness to cap-

sule radius, h/R , is 0.0196. The membrane can be designed to

be magneto-responsive by adding magnetic nanoparticles. The

magnetic capsules can be easily separated from solutions and

non-magnetic capsules by applying a magnetic fi eld as shown

in Figure S1 of the Supporting Information, where capsules

contained food-coloring pigments instead of fl uorescent dyes.

This magnetic response is useful for the separation of the cap-

sules after measurements.

2.2. Estimation of Osmolarity Using ETPTA Capsules

When semipermeable capsules are subjected to positive

osmotic pressure, the capsules shrink due to the outward

fl ux of water through the membrane. The capsules exhibit

an isotropic shrinkage for small volume reduction and they

begin to buckle by forming one indentation at the weakest

point for further volume reduction than a certain value,

ΔV * , as shown schematically in Figure 2 a. [ 8,10 ] Therefore, the

osmotic pressure difference that reduces the volume of the

capsules more than ΔV * can be confi rmed by the formation

of an indentation on the capsule membrane. In particular,

sets of such semipermeable capsules of which the aqueous

cores have different osmolarity can be used to estimate the

osmolarity of the continuous phase. We demonstrate this by

using semipermeable ETPTA capsules. The ETPTA mem-

brane is permeable for water molecules, whereas imperme-

able for PVA and sodium and chloride ions; the permeability

of water through the ETPTA membrane is estimated as 2 ×

10 −24 m 2 in previous work. [ 6a , 8 ] For example, a mixture of four

distinct ETPTA capsules, confi ning aqueous solutions with

33, 342, 649, and 950 mOsm L −1 respectively, was dispersed

in an aqueous solution with 475 mOsm L −1 . This resulted in

the selective buckling of two capsules, namely those with 33

and 342 mOsm L −1 , because of the positive osmotic pressure,

whereas the other two capsules, with 649 and 950 mOsm L −1 ,

remained unchanged in size and shape, as shown in Figures 2 b

and c. Although the latter capsules experienced a negative

pressure, they remained without infl ation or release of

Figure 2. a) Schematic illustration of the buckling of semipermeable capsules subjected to positive osmotic pressure; the capsules maintain their spherical shape during the initial shrinkage of ΔV* and then, they begin to buckle through further shrinkage. The buckling ceases when the osmotic pressure difference becomes negligible b,c) Confocal microscopy images of a mixture of four distinct microcapsules dispersed in an aqueous solution of 475 mOsm L −1 , whereby the red, orange, yellow, and green capsules contain aqueous solutions of 33, 342, 649, and 950 mOsm L −1 , respectively; it can be seen that the red and orange capsules are buckled due to the positive osmotic pressure, whereas the yellow and green capsules remain spherical because of the negative osmotic pressure. The images were taken within 12 hours after the mixture was dispersed in the aqueous solution.

small 2014, 10, No. 6, 1155–1162

S.-H. Kim et al.

1158 www.small-journal.com © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

full papersencapsulants because of the high modulus of ETPTA, E ETPTA

(as high as 600 MPa). [ 11 ] In a similar fashion, the mixture

was dispersed in aqueous solutions with 175, 767, and 1068

mOsm L −1 , resulting in a selective buckling of capsules

when subjected to positive osmotic pressure, whereas the

others remained, as shown in Figure S2 of the Supporting

Information. It is shown that only red capsules buckled in

aqueous solutions with 175 mOsm L −1 , whereas red, orange,

and yellow capsules buckled in aqueous solutions with

767 mOsm L −1 , and all capsules buckled in aqueous solutions

with 1068 mOsm L −1 . Therefore, we can estimate the osmo-

larity of the continuous phase by observing the buckling of

a set of ETPTA capsules containing aqueous solutions with

different osmolarity. However, buckling of the ETPTA cap-

sule is not highly sensitive to the osmotic pressure difference,

thus lowering the precision of this estimation. To evaluate the

sensitivity, four distinct ETPTA capsules with 412, 456, 492,

and 550 mOsm L −1 were prepared as shown in Figure S3 of

the Supporting Information, and they were dispersed in an

aqueous solution with 475 mOsm L −1 . Although both red and

orange capsules were subjected to positive osmotic pressure,

only the red capsules buckled as shown in Figures 3 a and b.

The orange capsules were subjected to an osmotic pressure

difference, Π (= Δ CRT) , of 47 kPa (Δ C ca. 19 mOsm L −1 )

through the membrane, which was not enough to buckle

the ETPTA capsules, whereas the red capsules were sub-

jected to 156 kPa (Δ C ca. 63 mOsm L −1 ), where R is the gas

constant and T is the temperature. Similarly, the mixture dis-

persed in an aqueous solution with 592 mOsm L −1 exhibited

buckling of the red, orange, and yellow capsules, as shown

in Figures 3 c and d, although all capsules were subjected to

positive osmotic pressure, the green capsules were only sub-

jected to 104 kPa (Δ C ca. 42 mOsm L −1 ). Therefore, we can

estimate the threshold osmotic pressure difference needed

to buckle the ETPTA capsules to be approximately 125 kPa

(Δ C ca. 50 mOsm L −1 ).

2.3. Preparation of PLA Capsules

For a spherical capsule with uniform thickness of the mem-

brane, the threshold pressure difference, Π * , for buckling is

proportional to E(h/R) 2 and Δ V * /V 0 is proportional to h/R ,

where V 0 is the initial volume of the capsule. [ 10 ] Therefore, if

the ratio of membrane thickness to radius of the capsule, h/R ,

is reduced, the capsule becomes more sensitive to the osmotic

pressure difference and buckling begins earlier. To accom-

plish this, we used evaporation-induced consolidation of the

middle phase, enabling shrinkage of the membrane during

evaporation. [ 6b , e , 12 ] In a fashion similar to that of the ETPTA

capsules, we produced double-emulsion drops as templates,

where 12 wt% chloroform solution of PLA was employed

as the oil phase, instead of ETPTA. The double-emulsion

drops were collected in a glass petridish and incubated at

35 °C for 1 hour to completely evaporate the chloroform.

During the evaporation, the middle phase shrunk and fi nally

became a solid membrane made of PLA which is biocompat-

ible and biodegradable; more than 95% of double-emulsion

drops generated were converted into stable PLA capsules

during evaporation. Monodisperse PLA microcapsules with

an average diameter of 125 µm are shown in Figure 4 a. We

dried the capsules and observed them with SEM, as shown

in Figure 4 b. During drying, all capsules defl ated on the sub-

strate, forming a thin fi lm with a weak trace of the capsules.

To characterize the thickness of the membrane, we made

scars on the fi lm using a razor blade, which enabled the

measurement of the membrane thickness (150 nm) as shown

in the inset of Figure 4 b. Therefore, h/R is 0.0024 which is

reduced by a factor of 8 in comparison with the h/R of the

ETPTA capsules, which was 0.0196. This shows that buckling

of the PLA capsules is much more sensitive to the osmotic

pressure difference than that of ETPTA capsules even with

the higher modulus of PLA (2050 MPa). [ 13 ] Although PLA

is biodegradable, the capsules remain intact for one month if

they are kept in an osmolarity-matched aqueous solution. [ 6b ]

The relative sensitivity is clearly shown in Figure 4 c, where

both PLA and ETPTA capsules were subjected to a positive

osmotic pressure of 22.3 kPa ( Δ C ca. 9 mOsm L −1 ) across

their membranes and approximately 100 capsules were used

to plot this time-dependence of the fraction of buckled cap-

sules. The buckling of capsules could be confi rmed by the

sudden formation of indentations during observation with an

optical microscope. Although the rate of buckling is not fast

under this low osmotic pressure difference, more than 70% of

the PLA capsules were buckled within 12 hours, whereas less

than 10% of the ETPTA capsules were buckled; we attribute

Figure 3. a–d) Confocal microscopy images of a mixture of four distinct microcapsules dispersed in aqueous solution with (a,b) 475 mOsm L −1 and (c,d) 592 mOsm L −1 , whereby the red, orange, yellow, and green capsules contain aqueous solutions of 412, 456, 492, and 550 mOsm L −1 , respectively. Although both red and orange capsules are subjected to positive osmotic pressure in (a,b), the orange capsules remain unbuckled because of the insuffi cient osmotic pressure difference; the buckled red capsules are denoted with arrows. Similarly, the green capsules remain without buckling, although all capsules are subjected to positive osmotic pressure in (c,d). The images were taken within 12 hours after the mixture was dispersed in the aqueous solution.

small 2014, 10, No. 6, 1155–1162

Osmocapsules for Direct Measurement of Osmotic Strength

1159www.small-journal.com© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

this buckling of ETPTA capsules to the concurrent forma-

tion of inhomogeneous capsules in small portions, making

them more sensitive to the osmotic pressure difference. [ 8 ]

We also attribute the distribution of the response time of the

PLA capsules under low osmotic pressure difference to the

inhomogeneity of the membrane thickness. Approximately

90% of the PLA capsules buckled during 1-day incubation,

whereas 10% remained unbuckled. This is because of the co-

generation of double-emulsion drops with a relatively thick

shell that convert into less-sensitive capsules to osmotic pres-

sure difference.

2.4. Estimation of Osmolarity Using PLA Capsules

We characterized the time- and osmotic pressure-depend-

ence of the buckling of PLA capsules by observing them

in various osmolarity conditions whereby the difference in

osmolarity was 0, 5, 9, 16, 30, and 43 mOsm L −1 , as shown in

Figure 5 . Images of the PLA capsules, taken at 10 minutes

and 210 minutes after they were subjected to the osmotic

pressure difference of 73.4 kPa (Δ C ca. 30 mOsm L −1 ), are

shown in Figure 5 a and this buckling behavior is shown in

Movie S2 of the Supporting Information. The permeability, k ,

of water through the PLA membrane in Darcy's law can be

estimated from the time series of dimple evolution in Movie

S2 according to previous work, [ 8 ] which provided a value of

2 × 10 −25 m 2 . For high osmotic pressure differences such as

39.6 kPa (Δ C= 16 mOsm L −1 ), 74.3 kPa (Δ C= 30 mOsm L −1 ),

and 106.5 kPa (Δ C= 43 mOsm L −1 ), 50% of the capsules were

buckled after 184, 113, and 42 minutes, respectively. For small

pressure differences such as 12.4 kPa (Δ C = 5 mOsm L −1 ) and

22.3 kPa (Δ C = 9 mOsm L −1 ), it takes 840 and 500 minutes,

respectively. Therefore, the precision of the osmolarity meas-

urement is smaller than 5 mOsm L −1 ; however, the measure-

ments take a long time, which can lead to problems, such as

evaporation, for practical applications and therefore the time

needs to be further reduced by making thinner membranes

with a higher permeability.

Folding of the PLA membrane with high curvature during

buckling can lead to rupture of the capsules, forming cracks

or holes on the membrane surface. Subsequently, PVA and

ions can diffuse through the scars, thereby eliminating any

osmotic pressure difference. This absence of stress enables

the deformed capsules to recover their spherical shape. For

example, the capsules subjected to high osmotic pressure dif-

ference of 106.5 kPa (Δ C = 43 mOsm L −1 ) initially buckled

by forming an indentation and then, recovered their spherical

shape, as shown in Figure 6 a. Although the scars are not vis-

ible under an optical microscope, we confi rmed the diffusion

of materials by weakening the optical contrast of the capsules

after they recovered their spherical shape, as shown in the last

image of Figure 6 a. Diffusion of materials leads to the reduc-

tion of the refractive index contrast between the interior and

exterior of the capsule, thereby resulting in a low optical con-

trast. The time-dependence of the buckling and rupturing

of the capsules, subjected to a positive osmotic pressure of

106.5 kPa (Δ C = 43 mOsm L −1 ), is shown in Figure 6 b. The

fraction of buckled capsules quickly increases as they are

Figure 4. a) Optical microscopy image of microcapsules with a diameter of 125 µm whose membrane is made of PLA. b) SEM image of the dried PLA capsules; the capsules are completely defl ated, forming a thin fi lm on the substrate. The inset shows the thickness of the membrane, 150 nm. c) Time-dependence of the fraction of buckled capsules, where both ETPTA and PLA capsules are subjected to a positive pressure difference of 22.3 kPa (Δ C = 9 mOsm L −1 ) across their membranes.

Figure 5. a) Optical microscopy images of PLA microcapsules, taken at the denoted times in the images after they were subjected to a positive osmotic pressure of 74.3 kPa. b) Time-dependence of the fraction of buckled PLA capsules, whereby the PLA capsules were subjected to a series of positive pressure differences of 0, 12.4, 22.3, 39.6, 74.3, and 106.5 kPa across their membranes; these pressure differences correspond to 0, 5, 9, 16, 30, and 43 mOsm L −1 differences, respectively.

small 2014, 10, No. 6, 1155–1162

S.-H. Kim et al.

1160 www.small-journal.com © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

full paperssubjected to the osmotic pressure difference, whereas the

fraction of ruptured capsules slowly increases because of the

high degree of membrane folding required to induce rupture.

The fraction of ruptured capsules was always smaller than the

fraction of buckled capsules and it decreased with decreasing

applied osmotic pressure difference.

The release of encapsulants from osmocapsules is not

always favored. To make inert osmocapsules without release

function, we prepared capsules using a middle phase of

10 wt% toluene solution of poly(lactic-co-glycolic acid)

(PLGA); PLGA is also a biocompatible and biodegradable

polymer but it is less brittle than PLA. Most PLGA capsules,

subjected to a positive osmotic pressure of 106.5 kPa (Δ C =

43 mOsm L −1 ), buckled and remained deformed. In addition,

the capsules retained the encapsulants during the buckling as

shown in Figure S4 of the Supporting Information.

The buckling behavior of PLA microcapsules is much

more sensitive to osmotic pressure differences because

of the smaller h/R compared to that of ETPTA capsules,

thereby providing higher precision in measurement of

osmotic strength. We demonstrate this with a set of PLA

capsules encapsulating aqueous solutions of 94, 118, 129, and

148 mOsm L −1 , as shown in Figures S5 and S6, where the cap-

sules are labeled with red, orange, yellow, and green colors

in the same manner to the ETPTA capsules. The mixture of

these four distinct capsules was dispersed into fi ve different

aqueous solutions with 92, 110, 124, 147, and 163 mOsm L −1 ,

as shown in Figure 7 . All PLA capsules subjected to positive

osmotic pressure were buckled, whereas the other capsules

remained spherical. Therefore, it proves that the mixture

can be used to estimate the osmotic strength of unknown

Figure 6. a) Optical microscopy images of a PLA microcapsule, taken at denoted times in the images after subjection to a positive osmotic pressure difference of 106.5 kPa; the capsule buckled fi rst (middle image) and then recovered its spherical shape (right image) after rupture of the membrane. The optical contrast becomes weaker upon rupture. b) Time-dependence of fractions of buckled (red line) and ruptured (black line) PLA capsules, where the PLA capsules are subjected to a positive pressure difference of 106.5 kPa across their membranes.

Figure 7. Confocal microscopy images of a mixture of four distinct PLA microcapsules dispersed in an aqueous solution of 92, 110, 124, 147, and 163 mOsm L −1 , whereby the red, orange, yellow, and green capsules contain aqueous solutions of 94, 118, 129, and 148 mOsm L −1 , respectively; all capsules subjected to positive osmotic pressure are buckled as denoted with arrows, whereas others remained spherical. Some capsules were ruptured under high positive pressure, appearing transparent due to the release of the dye.

small 2014, 10, No. 6, 1155–1162

Osmocapsules for Direct Measurement of Osmotic Strength

1161www.small-journal.com© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

aqueous solutions. Some PLA capsules dispersed in aqueous

solutions with high osmolarities of 147 and 163 mOsm L −1

exhibited rupture of their membranes because of the high

degree of buckling, and they appeared transparent without

color due to the release of the dye molecules, whereas most

PLA capsules when dispersed in an aqueous solution with

low osmolarity of 92, 110, and 124 mOsm L −1 retained the

dye. To estimate the osmolarity of an unknown solution, we

can apply two or three sets of these osmocapsules, where the

fi rst set can be used to fi nd a rough range of the osmolarity

and the second and third sets will be used to estimate the

value more precisely. For example, to cover a wide range of

osmolarity of 0–1000 mOsm L −1 , three steps of measurements

with fi ve distinct osmocapsules at each step can be used,

which will provide an error smaller than 8 mOsm L −1 . If 10

capsules are employed for each osmolarity to have enough

statistics and avoid mismeasurement, a total of 150 capsules

(ca. 100 nL) would be required per single measurement.

Although the capsules can buckle in a sample whose volume

is similar to that of the capsules, in practice a volume of 100

times that of the capsules (i.e. 10 µL), will be required; which

is still much smaller than the volume required for conven-

tional micro-osmometers.

3. Conclusion

In this work, we report a method to estimate the osmotic

strength of aqueous solutions using microcapsules with an

ultra-thin membrane. Microfl uidic emulsifi cation of biphasic

fl ows enables the production of monodisperse double-emul-

sion drops with an ultra-thin middle phase, providing a stable

template to produce microcapsules containing an aqueous

solution with standard osmotic strength. By employing pho-

tocurable monomers as the middle phase, microcapsules with

a ratio of membrane thickness to capsule radius of approxi-

mately 0.02 were prepared by UV-induced polymerization.

These resultant capsules exhibited buckling when subjected

to a positive osmotic pressure as high as approximately

125 kPa. Highly sensitive microcapsules to osmotic pressure

difference were prepared by evaporation-induced consolida-

tion of the middle phase. Because of shell shrinkage during

evaporation, the resultant capsules have a ratio of membrane

thickness to capsule radius of approximately 0.002. These sen-

sitive microcapsules buckled under positive osmotic pressures

as small as 12.5 kPa, thereby enabling the precise estimation

of the osmotic strength using a set of distinct microcapsules

containing aqueous solutions of different standard osmotic

strengths. This approach directly measures the osmotic

strength using semipermeable membranes without complex

setup or analysis. In addition, it requires a very small volume

of the sample liquid. Therefore, it has great potential as an

alternative method to measure the osmotic strength. Further-

more, this capsule-based method can be potentially used for

in-vivo measurement of the osmotic strength by implanting

biocompatible and degradable PLGA or PLA capsules into

tissue. At the same time, we expect that the PLA capsules can

be used as smart delivery vehicles that sense osmotic pres-

sure and release encapsulants by rupture of their membrane.

In addition, the capsules can be inserted into reactors or

microfl uidic channels, which enables in-situ monitoring of the

osmotic pressure without sampling the solution. The current

approach has almost no limitations on material selection for

the membrane, thereby potentially providing a wide variety

of properties and applicable areas for osmotic pressure

measurement.

4. Experimental Section

Materials : Aqueous solutions containing PVA (Mw 13 000–23 000, Sigma-Aldrich) and NaCl were used as the innermost and continuous phase of the double-emulsion drops. As oil phases, we used one of ETPTA containing 0.2 wt% photoinitiator (2-hydroxy-2-methylpropio-phenone, Sigma-Aldrich), 12 wt% chloroform solution of poly(lactic acid) (PLA, Mw 15,000, Polyscience, Inc.), and 10 wt% toluene solu-tion of poly(lactic-co-glycolic acid) (PLGA, Mw 120,000, 8515 DL High IV, Evonik). For some experiments, we dispersed iron oxide nano-particles (smaller than 50 nm, Sigma-Aldrich) to render our capsules magneto-responsivity and silica particles to improve the dispersion stability. [ 14 ] To distinguish the capsules containing aqueous solu-tions with different osmolarity, we dissolved red dye, sulforhodamine B (Sigma-Aldrich) and green dye, 8-hydroxyl-1,3,6-pyrenetrisulfonic acid, trisodium salt (Spectrum Chemicals) in the innermost phase with four different mixing ratios; confocal microscopy showed four dif-ferent colors of red, orange, yellow, and green in overlay images. For the same purpose, we also used common food-coloring pigments with red and green colors. [ 15 ] The osmolarity of aqueous solutions was measured by an osmometer based on depression of the freezing point (Model 3300, Advanced Instrument, Inc.) before injection.

Device Fabrication : Two cylindrical capillaries (World Preci-sion Instruments, Inc., 1B100–6) were tapered by a micropipette puller (Sutter Instrument, P-97) to have a 70-µm diameter orifi ce and 190-µm diameter orifi ce, respectively. The cylindrical capil-lary with smaller orifi ce was treated with n -octadecyltrimethoxyl silane (Aldrich) to render the surface hydrophobic, whereas the other cylindrical capillary with larger orifi ce was treated with 2-[methoxy(polyethyleneoxy)propyl] trimethoxyl silane (Gelest, Inc.) to render the surface hydrophilic. These two cylindrical cap-illaries were coaxially aligned in a square capillary and a small tapered capillary was inserted into the untapered opening of the hydrophobic cylindrical capillary.

Supporting Information

Supporting Information is available from the Wiley Online Library or from the author.

Acknowledgments

This work was supported by the KAIST HRHRP (Project No. N10130008), the International collaboration grant (No. Sunjin-2010–002) and the Industrial strategic technology development

small 2014, 10, No. 6, 1155–1162

S.-H. Kim et al.

1162 www.small-journal.com © 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

full papersprogram (No. 10045068) of the Korea Evaluation Institute of Industrial Technology funded by the Ministry of Trade, Industry, & Energy (MI, Korea). We dedicate this article to late Professor Seung-Man Yang for his lifelong contribution to colloid and inter-face science.

[1] K. Livingston , Science 1995 , 268 , 314 . [2] a) H. C. Shum , J. W. Kim , D. A. Weitz , J. Am. Chem. Soc. 2008 ,

130 , 9543 ; b) C. Martino , S. H. Kim , L. Horsfall , A. Abbaspourrad , S. J. Rosser , J. Cooper , D. A. Weitz , Angew. Chem. Int. Edit. 2012 , 51 , 6416 .

[3] a) H. G. Applegate , Nature 1960 , 186 , 233 ; b) N. P. Money , Plant Physiol . 1989 , 91 , 766 .

[4] V. L. Vilker , C. K. Colton , K. A. Smith , J. Colloid Interf. Sci. 1981 , 79 , 548 .

[5] H. Vink , Eur. Polym. J. 1971 , 7 , 1411 . [6] a) S. H. Kim , S. J. Jeon , S. M. Yang , J. Am. Chem. Soc. 2008 , 130 ,

6040 ; b) S. H. Kim , J. W. Kim , J. C. Cho , D. A. Weitz , Lab Chip 2011 , 11 , 3162 ; c) S. H. Kim , H. C. Shum , J. W. Kim , J. C. Cho , D. A. Weitz ,

J. Am. Chem. Soc. 2011 , 133 , 15165 ; d) M. H. Lee , K. C. Hribar , T. Brugarolas , N. P. Kamat , J. A. Burdick , D. Lee , Adv. Funct. Mater. 2012 , 22 , 131 ; e) F. Q. Tu , D. Lee , Langmuir 2012 , 28 , 9944 .

[7] a) S. Okushima , T. Nisisako , T. Torii , T. Higuchi , Langmuir 2004 , 20 , 9905 ; b) A. S. Utada , E. Lorenceau , D. R. Link , P. D. Kaplan , H. A. Stone , D. A. Weitz , Science 2005 , 308 , 537 .

[8] S. S. Datta , S. H. Kim , J. Paulose , A. Abbaspourrad , D. R. Nelson , D. A. Weitz , Phys. Rev. Lett. 2012 , 109 , 134 302 .

[9] S. H. Kim , D. A. Weitz , Angew. Chem. Int. Ed. 2011 , 50 , 8731 . [10] S. Knoche , J. Kierfeld , Phys. Rev. E 2011 , 84 , 046 608 . [11] T. H. Lin , W. H. Huang , I. K. Jun , P. Jiang , Chem. Mater. 2009 , 21 ,

2039 . [12] D. Lee , D. A. Weitz , Adv. Mater. 2008 , 20 , 3498 . [13] O. Martin , L. Averous , Polymer 2001 , 42 , 6209 . [14] S. H. Kim , J. Y. Shim , J. M. Lim , S. M. Yang , Angew. Chem. Int. Ed.

2010 , 49 , 3786 . [15] S. H. Kim , J. W. Shim , S. M. Yang , Angew. Chem. Int. Ed. 2011 , 50 ,

1171 .

Received: July 27, 2013 Revised: October 3, 2013 Published online: January 30, 2014

small 2014, 10, No. 6, 1155–1162