PAPER www.rsc.org/softmatter | Soft Matter

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

Solvent induced transition from wrinkles to creases in thin film gels withdepth-wise crosslinking gradients†‡

Murat Guvendiren,a Jason A. Burdick*a and Shu Yang*b

Received 2nd May 2010, Accepted 15th July 2010

DOI: 10.1039/c0sm00317d

We investigated solvent induced transition of surface instability from wrinkles to creases in

poly(2-hydroxyethyl methacrylate) (PHEMA) gels with depth-wise crosslinking gradients. The mode of

surface instability and morphology of surface patterns was found to be dependent on the equilibrium

linear expansion, which was a function of crosslinker concentration and the solvent–polymer

interaction. The maximum linear expansion was obtained when the PHEMA film was swollen in a good

solvent, which had the Hildebrand solubility parameter (ds) close to that of PHEMA gels, 26.6 to

29.6 MPa1/2. In a relatively poor solvent (e.g. water), wrinkling patterns were obtained and the

morphoplogy was determined by the concentration of the crosslinker, ethylene glycol dimethacrylate

(EGDMA). In a good solvent, such as alcohol and alcohol/water mixture, the equilibrium linear

expansion ratio increased significantly, leading to the transition from wrinkling to creasing instability.

In an ethanol/water mixture, we systematically varied the ratio between ethanol and water and

observed the transition from wrinkling to creasing when gradually adding ethanol to water, and the

reverse transition when adding water in ethanol. The onset of the linear expansion ratio for creasing

(ac,c) was again found dependent on EGDMA concentration: ac,c z 2.00 and 1.3, respectively, for gels

with 1 and 3 wt% EGDMA. Finally, we demonstrated confinement of the creases by combining

swelling and photopatterning.

Introduction

Polymer networks in the form of thin films are attractive for

many applications, such as sensors,1 microfluidic devices,2

responsive coatings,3,4,5 tissue culture substrates3,4,6,7 and bio-

adhesives.8–10 In many of the applications, the polymer gels

interact with a solvent. In a good solvent, the network is highly

swollen. The resulting large volume change could lead to unde-

sirable failures in the gels. Recently there has been interest in

harnessing swelling induced instability to create surface patterns

with controlled morphology, order, size, and complexity. It is

known that the degree of swelling, and hence the volume change,

is dependent on polymer chemistry, the crosslink density and the

solvent quality, i.e. the affinity between the solvent molecules and

the polymer chains.

When an isotropic gel is attached on a substrate, it swells

preferentially perpendicular to the rigid substrate, resulting in

anisotropic osmotic pressure that generates a biaxial compressive

stress on the film surface. When the stress or linear expansion

(a ¼ h/ho, the ratio of swollen film thickness to initial thickness)

aDepartment of Bioengineering, University of Pennsylvania, 210 S 33rdStreet, Philadelphia, PA, 19104, USA. E-mail: burdick2@seas.upenn.edu; Fax: +1 (215) 573-2071; Tel: +1 (215) 898-8537bDepartment of Materials Science and Engineering, University ofPennsylvania, 3231 Walnut Street, Philadelphia, PA, 19104, USA.E-mail: shuyang@seas.upenn.edu; Fax: +1 (215) 573-2128; Tel: +1(215) 898-9645

† Electronic Supplementary Information (ESI) available: Confocalmicroscopy images showing fracturea in the films crosslinked with 1wt% and 3 wt% EGDMA, respectively. See DOI: 10.1039/c0sm00317d/

‡ This paper is part of a Soft Matter themed issue on The Physics ofBuckling. Guest editor: Alfred Crosby.

This journal is ª The Royal Society of Chemistry 2010

exceeds a critical value, the gel surface buckles, generating

different modes of instabilities, including wrinkling and creasing

(Fig. 1).11–15 For an incompressible network, such as an elasto-

meric poly(dimethylsiloxane) (PDMS) coated with a thin hard

skin, the compressive stress generated under bending is typically

small, and formation of wrinkles with smooth and shallow

Fig. 1 (a,b) Schematic illustrations of swelling instabilities from wrin-

kling to creasing in a surface-attached gel when increasing the compres-

sive stress influenced by the solvent quality. The shallow undulations

(wrinkling) become localized sharp folds (creasing) with increasing

equilibrium linear expansion of the gel. (b) The dots that are on the free

surface in the case of wrinkles run into each other to form sharp folds in

transit to creasing. In the latter, the dots are no longer part of the free

surface. (c) Confocal microscopy depth profile images (scanned at 1 mm

steps) depicting wrinkling and creasing formation in a PHEMA film

(2 wt% EGDMA) with depth crosslinking gradient swollen by water (left)

and ethanol/water mixture (60/40 v/v%) (right).

Soft Matter, 2010, 6, 5795–5801 | 5795

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

surface undulations are often observed.16–21 When the network

becomes highly swollen, such as polyacrylamide gels in water, the

initially free surface begins to touch each other, forming cusps in

the gel in the form of localized sharp folds, or so called creasing

instability (Fig. 1b).22–24

Although the driving force for wrinkling and creasing is the same,

the two modes of instability occur at very different critical strains, and

hence, different critical linear expansions (ac). There have been a wide

range of critical values reported experimentally, 2.0–3.7 in different

materials systems.13–15,25,26 Hong et al.27 recently derived a theoretical

onset value for creasing, ac,c¼ 2.4, assuming the crease formation is

fast so that there is no solvent migration in the gel, thus, resembling

incompressible elastomers. This theoretical prediction, however, may

not always represent the behavior of a real swollen gel system. Much

work in the literature has focused on manipulation of swelling-

induced wrinkling instabilities in elastomeric films,28–34 and creasing

instabilities in hydrogel films15,22,27,35,36 to form complex patterns. Few

have studied the transition between these two modes of instability.37,38

Further, most of the material systems are homogeneous networks.

It will be appealing to design a new material system that will

continuously change the surface instability behaviors in a controlled

fashion, addressing often neglected factors, such as the solvent–

network interaction and solvent diffusion kinetics in the network, to

the onset of wrinkling and creasing, as well as their transition.

We have recently reported the formation of a wide range of

surface wrinkling patterns by swelling poly(2-hydroxyethyl

methacrylate) (PHEMA) hydrogel films confined onto a rigid

substrate simply by varying the concentration of the crosslinker,

ethylene glycol dimethacrylate (EGDMA).11 The ability to control

the morphologies of the surface patterns, ranging from random

worm-like structures to peanut shape, lamellar and highly ordered

hexagonal structures, lies in the creation of depth-wise modulus

gradient in PHEMA gels, allowing for fine-tuning the solvent

diffusion and swelling kinetics by the EGDMA concentration. The

resulting surface patterns are wrinkles not creases, which are often

observed in highly-swollen hydrogels.14,15 This is because water is

a relatively poor solvent for PHEMA gels, thus, a lower onset

linear expansion for wrinkling, ac,w ¼ 1.12, is obtained in our

system. It is possible to transform wrinkles to creases when

swelling the PHEMA gradient gels with a better solvent.

Here, we report creasing formation in the gradient PHEMA

gels swelling in different solvents. Our results indicated that the

equilibrium linear expansion ratio (ae), thus, the mode of the

instability, can be tuned by solvent quality. The initially shallow

undulations that appeared in water swollen films began to

expand and touch each other, forming sharp folds, with addition

of a good solvent, such as ethanol, in water. The onset for

transition of the wrinkling to creasing (ac,c) was found to

decrease significantly from 2.0 to 1.3 with increasing EGDMA

concentration from 1 to 3 wt%. The magnitude of the instability

and fold depth increased with further increasing the linear

expansion, and finally, the films fractured when the fold depth

reached the film–substrate interface.

Results and discussion

The preparation of gradient PHEMA gels followed the proce-

dure reported previously.11 Crosslinked PHEMA films were

photopolymerized from a precursor solution of partially

5796 | Soft Matter, 2010, 6, 5795–5801

polymerized PHEMA, photoinitiators (Darocur 1173) and

EGDMA. The precursor solution was cast onto a methacrylated

glass slide, and exposed to UV light in open air. During radical

polymerization, oxygen dissolved in the precursor scavenge the

radicals generated from the photoinitiators, creating a depth-

wise modulus gradient in the PHEMA film. The depth profile

could be varied by the film thickness and EGDMA

concentration.

When immersed in water, water diffused through the

gradient PHEMA film, creating an anisotropic osmotic pres-

sure; its profile was dependent on the crosslinking gradient.

The outer gel surface becomes viscoelastic and liquid-like

while the bottom layer near the substrate remains as an elastic

solid. Due to the glassy nature of the dry PHEMA, the

swelling-induced surface patterns are kinetically trapped upon

drying, allowing us to directly visualize the swollen patterns in

the dry states.

For a surface-attached film, the volumetric swelling ratio (V) is

approximate to linear expansion (a), which can be defined as:

a ¼ h

h0

¼ 1

ð1� fsÞ(2)

where h and h0 are the thickness of swollen and dry films,

respectively, and fs is the solvent fraction in the network, which

is dependent on the degree of crosslinking and solvent quality.

The solubility parameters (d) and the Flory–Huggins polymer–

solvent interaction parameter (c) are approximation of the

solubility of polymers in solvents. Therefore, it is possible to vary

the swelling ratio to observe both wrinkling and creasing insta-

bilities in one single system by simply increasing the solvent

quality without changing the materials characteristics, therefore,

attesting the experimentally obtained critical values versus the

theoretical prediction.

Previously, we found that the onset of wrinkling in water

swollen PHEMA gradient films is independent of EGDMA

concentration but decreases with increase of the initial film

thickness. The equilibrium linear expansion (ae), however,

decreases from 1.30 to 1.20 when increasing the EGDMA

concentration from 1 to 3 wt%, leading to the transit from

peanut-shaped patterns to highly ordered hexagonal structures.12

All these surface patterns are in nature smooth wrinkles, which

are fundamentally different from the creasing patterns with

sharp folds observed in the highly swollen polyacrylamide

hydrogels (Fig. 1a).14,15

However, when the gradient-PHEMA films were equilibrated

in various good solvents, including methanol (MeOH), dimethyl

sulfoxide (DMSO), ethanol (EtOH) and 1-propanol (NPA) for

72 h, very different pattern morphologies were observed, sug-

gesting creasing formation although the onset was dependent on

the EGDMA concentration (Fig. 2a). To understand the swelling

behaviors, we first calculated the solvating potential of solvents

in different polymer–solvent systems according to the Hilde-

brand theory:39,40

a

amax

¼ exphaa�ds � dp

�2i

(3)

where amax is the maximum linear expansion ratio, and a is

a constant. ds and dp are the solubility parameters for the solvent

and the polymer network, respectively.

This journal is ª The Royal Society of Chemistry 2010

Fig. 2 (a) Optical microscopy images of swelling induced surface

patterns of gradient PHEMA films with 1 and 3 wt% EGDMA equili-

brated in various solvents for 72 h. Scale bar: 100 microns. (b) Equilib-

rium linear expansion values (ae) of PHEMA films vs. the Hildebrand

solubility parameter of different solvents. The arrow indicates the

appearance of film fracture in DMSO.

Table 1 Hildebrand solubility parameter values for different solvents39

Solvent ds [(MPa)1/2]

1-Propanol (NPA) 24.5Ethanol (EtOH) 26.5Dimethyl sulfoxide (DMSO) 26.7Methanol (MeOH) 29.6Water 47.8

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

From eqn (3), it is clear that amax is reached when swelling the

network in a solvent with ds z dp, i.e. the highest affinity to the

polymer network. The solubility parameter values for different

solvents39 are summarized in Table 1. The dp of PHEMA gels is

calculated from dp ¼ fEGDMA(dEGDMA � dPHEMA) + dPHEMA,

where fEGDMA is the EGDMA fraction, dEGDMA¼ 19.23 MPa1/2

and dPHEMA ¼ 29.66 MPa1/2.41,42 The solubility parameters of the

PHEMA gels containing 1 wt% and 3 wt% EGDMA, are

26.5 and 29.6 MPa1/2, respectively.

As seen from Fig. 2b, amax was obtained from DMSO

[ds¼ 26.7 (MPa)1/2], where large cracks and film detachment from

the glass slide were observed within 48 h swelling. The equilib-

rium linear expansion decreased gradually in solvents with

ds < 26.58 (MPa)1/2 but decreased sharply in solvents with

ds > 26.5 (MPa)1/2. For PHEMA with 3 wt% EGDMA, a similar

This journal is ª The Royal Society of Chemistry 2010

trend was observed as that with 1 wt% EGDMA, although ae’s

were in general smaller: ae ¼ 2.06 � 0.21, 1.37 � 0.15 and 1.15 �0.11 in MeOH [ds ¼ 29.6 (MPa)1/2], EtOH [ds ¼ 26.5 (MPa)1/2],

and water [ds ¼ 47.8 (MPa)1/2], respectively, in comparison to

2.25 � 0.17, 1.84 � 0.13 and 1.36 � 0.05 from swelling PHEMA

with 1wt% EGDMA in the corresponding solvents.

To systematically study the solvent effect on linear expansion

and surface pattern formation, we fine-tuned the solvating

potential by using a two-component mixed solvent. We gradually

changed the solvent quality by addition of a good solvent to

a poor solvent or vice versa. A two-component solvent system,

which has one component with ds value higher and the other with

ds value lower than that of the solute (here, polymer network), is

often a better solvent than the individual ones.39 For this

purpose, ethanol with ds > dp (good solvent) and water with

ds < dp (poor solvent) with different volume ratios were mixed for

swelling PHEMA gels. The films were equilibrated in each

solvent mixture for at least 30 min. The resulting surface patterns

were found similar to those equilibrated for 72 h, therefore,

30 min swelling period was chosen here for the rest of the study.

The equilibrium surface patterns of gradient-PHEMA films

(with 1–3 wt% EGDMA) swollen in EtOH/water mixed solvents

are shown in Fig. 3a. Regardless of EGDMA concentration, the

initial pattern morphology changed significantly in the solvent

with 40 vol% EtOH and no major change of pattern size and

morphology was observed with EtOH vol% further increased up

to 80 vol%. For gels with 3 wt% EGDMA, the equilibrium

pattern morphology in water was hexagonal with long range

order, which gradually transformed to a mixture of hexagonal

pattern and peanut-shaped pattern (both smooth wrinkles) with

20 vol% EtOH, then to needle-shaped crease patterns in 40 vol%

EtOH, which remained up to 80 vol% EtOH (Fig. 3b). For 1 wt%

EGDMA, the original peanut-shaped patterns transformed to

needle-shaped patterns in 20 vol% EtOH and then into star-

shaped patterns, typical indication for creasing instability, in the

mixed solvent with 40 vol% EtOH. However, a decrease in

pattern size, and change of pattern morphology was observed in

pure ethanol, suggesting de-swelling of the film. These results

were in good agreement with the equilibrium expansion ratio

values obtained from the confocal depth profiles (see Fig. 4): ae

increased with increasing EtOH volume fraction. For gels with

1 wt% EGDMA, ae ¼ 1.3 water, and increased to ae ¼ 2 in

40 vol% EtOH, but somewhat plateaued afterwards at 2.1 in

60 vol% EtOH and 2.0 in 80% EtOH. ae decreased to 1.8 in 100%

EtOH. For gels with 3 wt% EGDMA, the ae followed the same

trend as that with 1 wt% EGDMA but the values were much

lower, in the range of 1.2 to 1.3. We believe that the plateau in ae

may be attributed to the crosslinking gradient in the PHEMA

films. As we reported before,11 the depth of crosslinking gradient

decreases with increasing crosslinker concentration: �34 mm for

films with 1 wt% EGDMA and � 27 mm for 3 wt% EGDMA.

Beyond this depth, the film becomes uniformly crosslinked and

the amount of solvent diffused through the film is significantly

decreased. Thus, even though the solubility parameter of the

mixed solvent continued to increase with addition of EtOH, ae

did not change much.

To futher confirm the instability mode (wrinkling vs. creasing)

in the swollen PHEMA films, we monitored the evolution of

surface instabilities in situ using confocal microscopy depth

Soft Matter, 2010, 6, 5795–5801 | 5797

Fig. 3 (a) Optical microscopy images showing the equilibrium swelling patterns for PHEMA gels (with 1–3 wt% EGDMA) in ethanol/water mixtures.

(b) Confocal microscopy images of PHEMA films (3 wt% EGDMA) scanned in xy- (left) and xz- planes (right). Dotted lines on the xy- scans correspond

to the location in the xz-scans. The dotted region on the depth profiles indicate the dynamic evolution from wrinkling (shallow undulation) to creasing

instability (formation of fold). Films were first immersed in water and ethanol was gradually added to the system. Films were equilibrated for at least

30 min. in each solvent composition before scanning.

5798 | Soft Matter, 2010, 6, 5795–5801 This journal is ª The Royal Society of Chemistry 2010

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

Fig. 4 The equilibrium expansion ratio (ae) of PHEMA gels (containing

1 and 3 wt% EGDMA) in ethanol/water mixtures. Films were equili-

brated in each solvent for 30 min.

Fig. 5 Sequence of in situ confocal fluorescent images showing the depth

profiles of the PHEMA gels in ethanol/water mixed solvent systems. Gels

were equilibrated in each solvent for 30 min. Images were taken from the

same location for each sample. Scale bars are 100 mm.

This journal is ª The Royal Society of Chemistry 2010

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

profile images (xz-scans) (Fig. 5). For films with 1 wt% EGDMA,

the amplitude of the patterns increased with increasing EtOH

vol%, and high fluctuations in the amplitude were observed as ae

approached �2.0 (40% EtOH). At this point, the linear expan-

sion was accompanied by the transverse expansion of the swollen

layer,22 which induced surface contacts, leading to the formation

of folds or cusps in the film (Fig. 1 and 5). This ae value is defined

as ac,c, the onset for transition from wrinkling to creasing. For

films with 3 wt% EGDMA, shallow folds appeared at ac,c z 1.3

(in 40 vol% EtOH) (Fig. 5 and Fig. 3b), whereas sharp folds were

only observed in DMSO when film became fractured at ae > 2. In

the reported swelling of homogeneous gels, the critical linear

expansion values for creasing are in the range of 2.0 to 3.7 and

independent of the crosslinker concentration.13–15,27 In our

PHEMA gels with depth-wise crosslinking gradient, the onset for

creasing was found to be dependent on the crosslinker concen-

tration and at lower linear expansion values.

Once the fold is formed in the creases, a cusp could be either

pushed out or the two neighboring cusps could merge to form

a single cusp.22,24,43 Tanaka et al.43 have argued that the only way

for the film to expand after cusp formation is by expansion of the

cusps outside of the film, leading to the disappearance of the

folds. When swelling PHEMA in DMSO for a prolonged time,

we observed that the fold depth increased with further expansion,

and disappeared into cracks when the compressible stresses

overcame the bonding strength between film and the substrate

such that the folds approached the film-substrate interface

(see Fig. S1, ESI†). The onset linear expansions for fracture af

was 2.82 � 0.41 and 2.09 � 0.14 for films with 1 wt% and 3 wt%

EGDMA, respectively.

To investigate the possibility of the reversibility of the pattern

morphologies between different wrinkle patterns and creases by

varying solvent quality, we first kept the PHEMA films (1 wt%

EGDMA) in water in the rubbery state, followed by gradual

increase of the EtOH volume fraction in the EtOH/water mixture

from 0% to 100 vol%, and then reversed the order from 100% to

0 vol% EtOH. The equilibrium morphology for each solvent was

captured after the film was kept in the solvent for 30 min. The

film initially formed smooth peanut-like wrinkles in pure water

and was in transit to creases with addition of ethanol as seen in

Fig. 3; the creasing instability was subsequently reversed by

addition of water into ethanol (Fig. 6). The near identical equi-

librium pattern morphologies during swelling (Fig. 3) and de-

swelling (Fig. 6) processes with identical EtOH vol% suggests

that the transition between wrinkling and creasing was almost

reversible except that the morphology of 0% EtOH during

deswelling was not completely recovered. This may be due to

residual solvent left in the film during deswelling process.

Nevertheless, the location of wrinkles or creases was highly

correlated from cycle to cycle. Such memory effect is in agree-

ment with the report by Trujillo et al. on crease formation in

homogeneously crosslinked polyacrylamide gels.15

To further demonstrate the utilization of creasing to create

complex patterns, we coupled photopatterning and solvent

swelling to direct the surface instability in confined regions. As

seen in Fig. 7, the PHEMA precursors were first crosslinked

through a photomask with square dot arrays (diameter¼ 100 mm,

pitch¼ 400 mm) (see insets in Fig. 7a,b) at the same intensity used

to create the gradient PHEMA gels, followed by pattern

Soft Matter, 2010, 6, 5795–5801 | 5799

Fig. 6 Confocal images of a PHEMA gel film with 1 wt% EGDMA showing the reversibility of pattern formation in xy- (top) and xz- (bottom) scans.

The PHEMA film was initially equilibrated in 100% ethanol, and water was gradually added into the system. Scale bar: 100 mm.

Fig. 7 Gradient PHEMA films (1 wt% EGDMA) crosslinked through

a photomask, followed by developing and swelling in various solvents.

When washed with methanol, the uncrosslinked regions were washed

away, whereas the crosslinked regions formed creasing patterns (a,b).

Inset: photomask. When the film in (b) was dipped into DMSO, cracks

started to appear (c), leading to the formation of square patterns for

extended periods (d). Scale bars: 50 mm.

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

development and swelling in methanol. The uncrosslinked regions

were washed away with methanol, during which swelling-induced

patterns developed simultaneously (Fig. 7a,b) in the exposed

regions. By using photomasks of opposite transparency, we

directed the crease formation either within the dots (like the flower

petals) or in the surrounding regions. Because light intensity

through the mask has a Gaussian distribution, the center of the

dots was more crosslinked than the edge of the dot, therefore, no

creases formed within the swelling time (30 min.). When the film

from Fig. 7b was dipped into DMSO, a much better solvent for

PHEMA networks, creases started to disappear and the film

surface became fractured (Fig. 7c), leading to the formation of

square frameworks, where the crosslinking density was the lowest,

over an extended swelling time (Fig. 7d).

5800 | Soft Matter, 2010, 6, 5795–5801

Conclusions

We investigated solvent swelling induced dynamic evolution of

surface instability from wrinkles to creases in PHEMA gels with

depth-wise crosslinking gradients. The maximum linear swelling

was obtained when the PHEMA films were swollen in a good

solvent with the Hildebrand solubility parameter, ds, close to that

of PHEMA, 26.6 to 29.6 MPa1/2. In a relatively poor solvent

(e.g. water), wrinkling appeared and the morphology was

determined by the crosslinker concentration. In a good solvent,

such as alcohol and alcohol/water mixtures, the linear expansion

ratio increased significantly, leading to the transition from

wrinkling to creasing instability. The onset of the linear expan-

sion ratio for creasing (ac,c) was found dependent on the

EGDMA concentration. The linear expansion reached another

critical value for fracture (af), where the depth of the folds

approached the film-substrate interface, leading to crack

formation and film detachement. The value of af was also found

strongly dependent on crosslinker concentration. Finally, we

demonstrated confinement of the creases by combining swelling

and photomasks.

Our study suggests that it is important to consider solvent–

polymer interactions and crosslinking density of the polymer gels

together with the cohesive strength of the gel to develop a full

picture of swelling induced instability and its dynamic pattern

transition from one mode to another. We believe the difference of

the reported critical linear expansion ratios for crease formation

in our system vs. the homogeneous and highly swollen gels, such

as polyacrylamide, will offer new insights for theoretical study of

the mechanism of instability of soft networks.

Experimental

PHEMA prepolymer was prepared by UV exposure (UVP Black

Ray, 8 mW cm�2) of 2-hydroxyethyl methacrylate (HEMA, 2mL,

98%, Alfa Aesar) as monomer and the Darocur 1173 (3 wt%,

60 mL, Ciba) as photoinitiator for 60 s. Another mixture of

Darocur 1173 and the crosslinker, ethylene glycol dimethacrylate

(EGDMA, Polysciences) (2 : 1 wt ratio), was added to the

This journal is ª The Royal Society of Chemistry 2010

Dow

nloa

ded

by U

nive

rsity

of

Penn

sylv

ania

Lib

rari

es o

n 03

Nov

embe

r 20

10Pu

blis

hed

on 3

1 A

ugus

t 201

0 on

http

://pu

bs.r

sc.o

rg |

doi:1

0.10

39/C

0SM

0031

7DView Online

viscous prepolymer to form the PHEMA precursor solution. A

series of precursor solutions without and containing EGDMA

(1–3 wt%) were prepared, which were stable for several months

when kept in dark. �100 mm thick PHEMA films were prepared

by coating the precursor solutions on glass slides using

a Micrometer Adjustable Film Applicator (S271727, Sheen

Instruments Ltd., Kingston, England). Precursor coated glass

slides were then exposed to UV light (Omnicure S1000 UV Spot

Cure System, Exfo Life Sciences Division, Mississauga, Ontario,

Canada) for 20 min. (10 mW cm�2, 365 nm) in an open air

environment. In order to prevent film delamination from glass

slide during swelling, the glass slide was functionalized with

3-(trimethoxysilyl)propyl methacrylate (TMS, Aldrich) before

casting the PHEMA precursor solution.12 The PHEMA films

were swollen in different solvents, including distilled water,

methanol (MeOH, Sigma), ethanol (EtOH, Decon Labs Inc.),

1-propanol (NPA, Fisher), dimethyl sulfoxide (DMSO, Sigma),

and EtOH/water mixtures for at least 30 min. For photo-

patterning, PHEMA precursors were exposed to UV light through

a photomask with square dot arrays (dot diameter ¼ 100 mm,

pitch ¼ 400 mm), (see Fig. 7). The dot size is larger than the

wrinkle wavelength of PHEMA gels (�80 mm) with the dry film

thickness of 100 mm.11 The photomask was placed in between the

lamp and the precursor solution in close proximity to the

precursor solution.

Optical microscopy was performed on an Olympus BX61

motorized microscope with Hamamatsu controller. Confocal

microscopy imaging was performed on a Nikon TE300 inverted

microscope fitted with a Bio-Rad Radiance 2000 MP3 system

(Bio-Rad Laboratories, Inc., Hercules, CA). Images were

observed through a 10� and/or 20� objective, and recorded with

Lasersharp 2000 software (Bio-Rad Laboratories, Inc.). A fluo-

rescent dye, methacryloxyethyl thiocarbonyl rhodamine B

(PolyFluorTM 570, Polysciences, Warrington, PA), was added

to the precursor solution to obtain contrast (gel appeared red

under fluorescent light). Depth profiles were obtained by col-

lecting xz-scans with 1 mm step size. All measurements were

performed using ImageJ (1.41n, National Institute of Health).

Acknowledgements

The research was funded in part by National Science Foundation

MRSEC DMR05-20020, an NSF CAREER award DMR-

0548070 (SY), and a Fellowship in Science and Engineering from

the David and Lucile Packard Foundation (JAB).

References

1 V. Kozlovskaya, E. Kharlampieva, I. Erel and S. A. Sukhishvili, SoftMatter, 2009, 5, 4077.

This journal is ª The Royal Society of Chemistry 2010

2 S. R. Quake and A. Scherer, Science, 2000, 290, 1536.3 O. Ernst, A. Lieske, M. Jager, A. Lankenau and C. Duschl, Lab Chip,

2007, 7, 1322.4 H. A. von Recum, S. W. Kim, A. Kikuchi, M. Okuhara, Y. Sakurai

and T. Okano, J. Biomed. Mater. Res., 1998, 40, 631.5 J. Kim, J. Yoon and R. C. Hayward, Nat. Mater., 2010, 9, 159.6 N. Mari-Buye, S. O’Shaughnessy, C. Colominas, C. E. Semino,

K. K. Gleason and S. Borros, Adv. Funct. Mater., 2009, 19, 1276.7 M. Guvendiren and J. A. Burdick, Biomaterials, 2010, 31, 6511.8 H. Lee, B. P. Lee and P. B. Messersmith, Nature, 2007, 448, 338.9 M. Guvendiren, D. A. Brass, P. B. Messersmith and K. R. Shull,

J. Adhes., 2009, 85, 631.10 M. Guvendiren, P. B. Messersmith and K. R. Shull,

Biomacromolecules, 2008, 9, 122.11 M. Guvendiren, S. Yang and J. A. Burdick, Adv. Funct. Mater., 2009,

19, 3038.12 M. Guvendiren, J. A. Burdick and S. Yang, Soft Matter, 2010, 6,

2044.13 E. Southern and A. G. Thomas, J. Polym. Sci. A, 1965, 3, 641.14 T. Tanaka, S. T. Sun, Y. Hirokawa, S. Katayama, J. Kucera,

Y. Hirose and T. Amiya, Nature, 1987, 325, 796.15 V. Trujillo, J. Kim and R. C. Hayward, Soft Matter, 2008, 4, 564.16 E. Cerda and L. Mahadevan, Phys. Rev. Lett., 2003, 90, 074302.17 X. Chen and J. W. Hutchinson, J. Appl. Mech., 2004, 71, 597.18 J. Groenewold, Phys. A, 2001, 298, 32.19 R. Huang, J. Mech. Phys. Solids, 2005, 53, 63.20 W. T. S. Huck, N. Bowden, P. Onck, T. Pardoen, J. W. Hutchinson

and G. M. Whitesides, Langmuir, 2000, 16, 3497.21 E. Sultan and A. Boudaoud, J. Appl. Mech., 2008, 75, 051002.22 T. Hwa and M. Kardar, Phys. Rev. Lett., 1988, 61, 106.23 A. Onuki, Phys. Rev. A: At., Mol., Opt. Phys., 1989, 39, 5932.24 H. Tanaka, H. Tomita, A. Takasu, T. Hayashi and T. Nishi, Phys.

Rev. Lett., 1992, 68, 2794.25 A. N. Gent and I. S. Cho, Rubber Chem. Technol., 1999, 72, 253.26 A. Ghatak and A. L. Das, Phys. Rev. Lett., 2007, 99, 076101.27 W. Hong, X. H. Zhao and Z. G. Suo, Appl. Phys. Lett., 2009, 95,

111901.28 E. P. Chan and A. J. Crosby, Soft Matter, 2006, 2, 324.29 N. Bowden, S. Brittain, A. G. Evans, J. W. Hutchinson and

G. M. Whitesides, Nature, 1998, 393, 146.30 E. P. Chan and A. J. Crosby, Adv. Mater., 2006, 18, 3238.31 D. P. Holmes and A. J. Crosby, Adv. Mater., 2007, 19, 3589.32 P. C. Lin, S. Vajpayee, A. Jagota, C. Y. Hui and S. Yang, Soft Matter,

2008, 4, 1830.33 P. C. Lin and S. Yang, Appl. Phys. Lett., 2007, 90, 241903.34 Y. Zhang, E. A. Matsumoto, A. Peter, P. C. Lin, R. D. Kamien and

S. Yang, Nano Lett., 2008, 8, 1192.35 Y. Klein, E. Efrati and E. Sharon, Science, 2007, 315, 1116.36 J. S. Sharp and R. A. L. Jones, Phys. Rev. E: Stat., Nonlinear, Soft

Matter Phys., 2002, 66, 011801.37 P. M. Reis, F. Corson, A. Boudaoud and B. Roman, Phys. Rev. Lett.,

2009, 103, 045501.38 D. P. Holmes and A. J. Crosby, Phys. Rev. Lett., 2010, 105, 038303.39 A. F. M. Barton, Chem. Rev., 1975, 75, 731.40 J. H. Hildebrand, J. M. Prausnitz and R. L. Scott, Regular and

Related Solutions, 3rd edn, Van Nostrand-Reinhold, Princeton,1970.

41 T. Caykara, C. Ozyurek, O. Kantoglu and O. Guven, J. Polym. Sci.,Part B: Polym. Phys., 2002, 40, 1995.

42 O. Okay and C. Gurun, J. Appl. Polym. Sci., 1992, 46, 401.43 H. Tanaka and T. Sigehuzi, Phys. Rev. E: Stat. Phys., Plasmas, Fluids,

Relat. Interdiscip. Top., 1994, 49, R39.

Soft Matter, 2010, 6, 5795–5801 | 5801