Photoreactive oligoethylene glycol polymers - Versatile compounds for surface modification by thin hydrogel films

Article abstract of DOI:10.1039/c2sm26879e

Solid surfaces are modified using photo-crosslinkable copolymers based on oligo(ethylene glycol) methacrylate (OEGMA) bearing 2-(4-benzoylphenoxy)ethyl methacrylate (BPEM) as a photosensitive crosslinking unit. Thin films of about 100 nm are formed by spi

Full text of DOI:10.1039/c2sm26879e

Soft Matter
View Article Online
View Journal | View Issue
PAPER
Photoreactive oligoethylene glycol polymers – versatile
compounds for surface modification by thin hydrogel
films†
Cite this: Soft Matter, 2013, 9, 929
a
a
b
*
*
´
Jens Buller, Andre Laschewsky and Erik Wischerhoff
Solid surfaces are modified using photo-crosslinkable copolymers based on oligo(ethylene glycol)
methacrylate (OEGMA) bearing 2-(4-benzoylphenoxy)ethyl methacrylate (BPEM) as a photosensitive
crosslinking unit. Thin films of about 100 nm are formed by spin-coating these a priori highly
biocompatible copolymers onto silicon substrates. Subsequent UV-irradiation assures immobilization and
crosslinking of the hydrogel films. Their stability is controlled by the number of crosslinker units per
chain and the molar mass of the copolymers. The swelling of the hydrogel layers, as investigated by
ellipsometry, can be tuned by the crosslinker content in the copolymer. If films are built from the ternary
copolymers of OEGMA, BPEM and 2-(2-methoxyethoxy) ethyl methacrylate (MEO2MA), the hydrogel
films exhibit a swelling/deswelling transition of the lower critical solution temperature (LCST) type. The
observed thermally induced hydrogel collapse is fully reversible and the onset temperature of the
transition can be tuned at will by the copolymer composition. Different from analogously prepared
thermo-responsive hydrogel films of photocrosslinked poly(N-isopropylacrylamide), the swelling–
deswelling transition occurs more gradually, but shows no hysteresis.
Received 13th August 2012
Accepted 29th October 2012
DOI: 10.1039/c2sm26879e
www.rsc.org/softmatter
Crosslinking of polymers via benzophenone groups is a long
established strategy for the preparation of polymer networks.20
1 Introduction
Polymers composed of oligo(ethylene glycol) (OEG) macro-
monomers are attractive materials for bio-medical applications
due to their biocompatibility1 and ability to suppress non-
specic interactions with biomaterials.2–4 Moreover they have,
for instance, been implemented as constituents for sensing
applications,5–9 for “smart” surfaces with the capability of
controlling cell adhesion,10–12 and, if appropriately functional-
ized, they are even capable of reacting to the presence of bio-
It also offers the possibility of tethering polymers to
surfaces.21–24 Benzophenone excited by UV light undergoes a
rapid intersystem crossing to a reactive triplet state oen being
represented as a diradical. As such it easily abstracts aliphatic
hydrogens from virtually any organic compound. Therefore if
organic substrates are used, radicals are readily formed on the
surface and may subsequently react with the polymer layer on
the top. Consequently, the copolymer gets immobilized onto
the substrate. The application of polymer attached benzophe-
none units is thus especially convenient as radicals generated
on the copolymer simultaneously lead to crosslinking.
When efficient crosslinking is ensured, hydrogel layer prop-
erties can be conveniently controlled by the crosslinker content
and/or irradiation time.23 Yet the efficiency of crosslinking is
strongly related to the reactivity of hydrogens in the copolymer
toward abstraction.25 Considering this, OEGMA copolymers
represent an ideal material for benzophenone crosslinking: the
hydrogens of the a-methylene group next to the ether moieties
within the side chains are known to be particularly prone to
radical attack. Thus, radical formation should be facilitated at this
position.26–28
macromolecules.13
A particularly useful property of the
poly(OEG) copolymer system is the facile tunability of the phase
transition temperature. It can be adjusted at will by simply
varying the ratio of comonomers with different oligo(ethylene
glycol) chain lengths.14–16 Despite these interesting properties, it
seems that their temperature dependent swelling behaviour on
surfaces has exclusively been studied for brush architec-
tures,17–19 and there are no reports about nanoscale oligo-
(ethylene glycol) thin lms.
a
¨
Universitat Potsdam, Department of Chemistry, Applied Polymer Chemistry,
Karl-Liebknecht-Str. 24-25, 14476 Potsdam-Golm, Germany. E-mail: laschews@
uni-potsdam.de
bFraunhofer Institute for Applied Polymer Research, Geiselbergstrasse 69, 14476
In this paper we apply polymerizable 2-(4-benzoylphenoxy)
ethyl methacrylate (BPEM) as photocrosslinker (Fig. 1). BPEM and
its closely related derivatives have been investigated occasionally
as monomeric and polymeric photoinitiators.29,30 However,
Potsdam-Golm, Germany. E-mail: erik.wischerhoff@iap.fraunhofer.de
† Electronic supplementary information (ESI) available: Temperature dependent
evolution of ellipsometric angles of thin crosslinked hydrogel lms of
copolymer T-3 and of pNIPAM-x. See DOI: 10.1039/c2sm26879e
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 929–937 | 929

View Article Online
Soft Matter
Paper
investigations have so far mainly been focused on the photo-
In a second step, a mixture of 10.0 g of the as-prepared inter-
chemistry and photophysics of this compound,30,31 while reports mediate 2-(4-benzoylphenoxy)ethanol (41.2 mmol), triethylamine
dealing with the use as a crosslinker seem to be virtually absent. (11.4 mL, 82.5 mmol) and 4-dimethylaminopyridine (0.25 g, 2.1
ꢂ
We report here the synthesis and characterization of poly(OEG) mmol) in dichloromethane (225 mL) was cooled to 0 C. Meth-
methacrylates incorporating the benzophenone bearing mono- acryloylchloride (4.6 mL, 45.4 mmol) in dichloromethane (20 mL)
mer, BPEM, and their use for stabilized thermo-responsive coat- was added drop wise. Aer stirring for additional 2 h at 0 ꢂC, the
ings. The properties of such lms were investigated and reaction mixture was allowed to warm to room temperature and
relationships to experimental parameters were established.
stirred overnight. The mixture was then poured into a saturated
aqueous NaHCO3 solution. The organic phase was separated,
washed with additional saturated aqueous NaHCO3 and NaCl,
dried by Na2SO4 and ltered through a column lled with silica
gel and evaporated to give 7.5 g of monomer BPEM as colourless
crystalline solid (yield 60%).
2 Experimental
2.1 Materials
4-Hydroxybenzophenone (98%, Aldrich), 2-bromoethanol (98%
¨
for synthesis, Merck), potassium carbonate (Riedel-de-Haen),
2.3 Synthesis of copolymers
potassium iodide (Acros), dimethylformamide (DMF), triethyl-
amine, 4-dimethylaminopyridine, methacryloyl chloride (97%,
Aldrich) and ethanol (99.8%, Chem Solute) were used as
received. Dichloromethane (analytical grade, J.T. Baker) was
dried over calcium chloride prior to use. 2-(2-Methoxyethoxy)
ethyl methacrylate (MEO2MA, 95%, Aldrich) and oligo(ethylene
glycol) methylether methacrylate (OEGMA, Aldrich, average
Mr ¼ 475 g molꢀ1) were passed through a basic alumina column
before use to remove inhibitors. 2-20-Azobisisobutyronitrile
(AIBN, 98%, Acros) was recrystallized from methanol before use.
Single side polished silicon wafers with an oxide layer of
1.0 nm thickness (Si-Mat, Kaufering, Germany) were cut into
square pieces of 2 cm ꢁ 2 cm.
The copolymers were synthesised according to a standard
procedure. In a round bottom ask protected from ambient
light appropriate amounts of 2-(2-methoxyethoxy) ethyl meth-
acrylate (MEO2MA), oligo(ethylene glycol) methylether methac-
rylate (OEGMA, average molar mass 475 g molꢀ1) and 2-(4-
benzoylphenoxy)ethyl methacrylate (BPEM) were dissolved in
ethanol (15 wt% monomers for copolymers of moderate molar
mass; 20 wt% monomers for copolymers of elevated molar
mass). The initiator azobisisobutyronitrile (AIBN, 0.5 equiva-
lents) was added, the ask was sealed with a rubber septum and
the mixture was purged with argon for 30 min. The mixture was
heated to 60 ꢂC in an oil bath for 20 h. The reaction was stopped
by opening the ask and allowing contact with air. The mixture
was diluted with deionized water and subsequently puried by
dialysis against deionized water for at least 3 days (Roth,
ZelluTrans membrane, molecular weight cut off: 4000–6000).
The polymer was isolated by freeze drying. All operation and
purication steps were carried out under careful protection
from light.
2.2 Synthesis of 2-(4-benzoylphenoxy)ethyl methacrylate
(BPEM)
The benzophenone monomer was synthesized following a
strategy similar to that reported by Zvara et al.31 First 4-hydroxy-
benzophenone (10.0 g, 50.44 mmol) was dissolved in dry DMF
(150 mL). K2CO3 (13.82 g, 100 mmol), KI (3.98 g, 24 mmol) and
2-bromoethanol (5.3 mL, 74 mmol) were added and the solution
was stirred at 65 ꢂC overnight. As thin lm chromatography
showed incomplete conversion, two additional portions of K2CO3
(7.0 g, 50 mmol) and 2-bromoethanol (4.4 mL, 62 mmol) were
added and the solution was stirred for further 18 h. The yellowish
reaction mixture was poured into 1.5 L of ice-water and the
precipitate was separated by centrifugation. The collected product
was dried in a vacuum at room temperature over CaCl2 to give
10.7 g of colourless shiny platelets (86% yield).
2.4 Film preparation
Silicon substrates were washed and hydrophilised for 20 h in a
solution of 100 mg of potassium permanganate in 40 mL of
concentrated sulphuric acid. The substrates were washed three
times with water, then with ethanol and nally with toluene. To
provide an organic surface for the immobilization of the copoly-
mer, samples were immersed in 1 vol% solutions of (3-amino-
propyl) dimethylethoxysilane in toluene, resulting in the silani-
zation of the substrates. Aer 6 h the substrates were taken out of
the solutions, washed with toluene and ethanol and dried under
air ow. Solutions of the copolymers in ethyl acetate were spin-
cast onto the pre-treated substrates with a standard spin-coater
(KL-SCV, Schaefer technologies, Langen/Germany). About 5
droplets of a 1–2 wt% solution were dispensed on the wafer at zero
rotation, and then the copolymer was cast at 50–70 rotations per
second for 30 s.
2.5 UV radiation chamber
UV irradiation was carried out in a UVACUBE 100 radiation
¨
¨
chamber (Honle, Grafelng/Germany) with a 100 W iron doped
mercury vapour lamp equipped with sheet glass as a lter (cut off
Fig. 1 Chemical structures of the monomers used.
930 | Soft Matter, 2013, 9, 929–937
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
310 nm). The distance of the sample from the light source benzophenone via a Williamson ether synthesis, then the
was 20 cm.
resulting functional alcohol was coupled with methacryloyl
chloride. The UV-sensitive benzophenone group is thus
attached to a polymerizable methacrylic unit via an ethylene
glycol spacer. This specic design, differing from other benzo-
phenone functionalized monomers,23,25 was selected for two
reasons. First, compared to the direct coupling via a phenolic
ester bond, the aliphatic ester of the polymerizable moiety in
combination with the phenol ether link introduced by the short
ethyl spacer should provide superior stability of BPEM against
hydrolysis. Secondly, the introduction of the ethylene glycol
spacer results in close electronic and steric similarity of BPEM
to the other comonomers used. Accordingly, their copoly-
merisation behaviour should be close to the ideal azeotropic
case. This implies that the benzophenone monomers are
incorporated into the copolymer with nearly the same proba-
bility as the other comonomers, independent of monomer
2.6 Characterisation methods
Cloud points of the aqueous polymer solutions were measured in
3.0 g Lꢀ1 solutions in ultrapure water on a thermostatted TP1
photometer (Tepper, Mainz/Germany).32 Transmittance at 670 nm
was monitored as a functioꢀn1of temperature with heating/cooling
1
ꢂ
cycles at rates of 1 C min (cell path length 12 mm). H NMR
spectra were recorded in CDCl3 using a Bruker Avance 300 spec-
trometer. Size exclusion chromatography (SEC) was run at 50 ꢂC in
DMF (ow rate 1 mL minꢀ1) using a Spectra Physics Instruments
apparatus equipped with a UV-detector SEC-3010 and a refractive
index detector SEC-3010 from WGE Dr Bures (columns: guard (7.5
ꢁ 75 mm), PolarGel-M (7.5 ꢁ 300 mm)), and calibrated with linear
polystyrene standards (PSS/Germany). For samples with a thick-
ness exceeding 300 nm, a Dektak 150 Prolometer (Veeco Instru-
ments Inc.) was used for surface lm thickness measurements.
2.7 Ellipsometry
Ellipsometry used a Multiscope apparatus from Optrel GbR
(Kleinmachnow/Germany) equipped with a HeNe laser (632.8
nm wavelength), in a null ellipsometer conguration, with an
angle of incidence of 60ꢂ. Dry lm thicknesses were calculated
by the soware “Elli”, version 5.2 (Optrel GbR, Kleinmachnow/
Germany), using a four-layer model with the following param-
eters: layer 1: air (n ¼ 1.000, k ¼ 0), layer 2: organic layer (n ¼
1.4800, k ¼ 0), layer 3: SiO2 (d ¼ 1.0 nm, n ¼ 1.4580, k ¼ 0), layer
4: silicon (n ¼ 3.8858, k ¼ ꢀ0.0200). The samples for ellips-
ometry were prepared on pieces cut from silicon wafers.
2.8 Swelling experiments
Samples were placed into a uidic cell (Optrel GbR, Kleinmach-
now/Germany) lled with deionized water at ambient tempera-
ture. The basis for the calculation of the thickness and the
refractive index of the swollen lm by the soware “Elli” was a
4-layer system with xed parameters for layer 1: (water n ¼ 1.332,
k ¼ 0), layer 3: (SiO2 d ¼ 1.0 nm, n ¼ 1.4580, k ¼ 0) and layer 4:
(silicon n ¼ 3.8858, k ¼ ꢀ0.0200). The hydrogel layer (layer 2) was
considered as a homogenous mixed layer of polymer and water.
This layer was calculated in the “n variable” mode, in which the
soware searches for points on the Delta, Psi-trajectories,33 which
match best the experimentally determined values.
2.9 Temperature dependent measurements
The uidic cell was connected to a Julabo F32-MH refrigerated
heating circulator lled with water. Layer thicknesses were
ꢂ
determined at 5 C intervals and the cell was allowed to equil-
ibrate for 20 min aer each temperature step.
Fig. 2 UV-vis spectra of BPEM monomers and copolymers produced. (a) BPEM
monomer in n-hexane (--- 10-fold magnification) and (b) absorption of a
copolymer layer of 800 nm thickness on a quartz window after different periods
3 Results and discussion
3.1 Synthesis of the UV-crosslinker monomer
of UV-irradiation (solid line 0 min, --- 1 min, 2 min,
3 min, 4 min, --- 6 min,
The synthesis of the monomer was carried out in two steps.
10 min, and 16 min). Inset: decrease of the maximum absorption peak at
Briey an ethylene glycol group was rst attached to 4-hydroxy 287 nm with duration of UV-irradiation.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 929–937 | 931

View Article Online
Soft Matter
Paper
conversion. Consequently, the copolymers are expected to be markedly lower than those of analogous binary copolymers of
chemically more homogeneous than analogous materials the same MEO2MA : OEGMA ratio without BPEM,15 even if only
prepared with 4-methacryloyloxy-benzophenone, and their small amounts of benzophenone comonomer are incorporated.
composition should be close to the one of the monomer feed, For example, the cloud point of tercopolymer T-5 containing 1
with a fairly even distribution of crosslinking sites. The mol% of BPEM is about 8 ꢂC lower than that of the binary
UV-spectrum of puried BPEM exhibits the typical absorption copolymer of the same MEO2MA–OEGMA composition. For
pattern of a donor substituted benzophenone, namely an copolymer T-2 containing 0.5 mol% of BPEM, the cloud point
ꢂ
intense p / p*-transition at 277 nm and a much weaker n / still is lowered by 2 C compared to the equivalent copolymer
p*-transition at 342 nm (see Fig. 2). For photo-activation of the containing no BPEM. This is explained by the hydrophobic
BPEM units, both wavelength ranges are equally useful.34
character of the benzophenone moiety, decreasing the cloud
point of the ternary copolymers. In order to maintain a specic
phase transition temperature, this effect can be easily coun-
terbalanced by adjustment of the comonomer ratio, i.e. by
incorporating more OEGMA into the copolymer.
3.2 Synthesis of UV-crosslinkable copolymers
The BPEM monomer was used as a comonomer in free radical
copolymerization with 2-(2-methoxyethoxy) ethyl methacrylate
(MEO2MA) and oligo(ethylene glycol) methylether methacrylate
(OEGMA, Mn ¼ 475 g molꢀ1). The content of incorporated
3.3 Characterisation of the hydrogel layers
benzophenone moieties in the copolymer was calculated by For immobilisation of the copolymers onto silica
1H-NMR spectroscopy. The integrated signals in the aromatic substrates, silica wafers were treated with (3-aminopropyl)
range from 7.0 ppm to 7.9 ppm (m, 9H of aromatic group) were dimethylethoxysilane. This pre-treatment provides organic
compared to the integrated signals of the methylene protons at groups, which can react with the benzophenone moieties upon
the a-position to the ester group at 4.0 ppm (broad signal, 2H). UV-irradiation, ensuring the covalent immobilisation of the
Keeping in mind the effective accuracy of quantitative NMR polymer onto the surface.27 The pre-treated silica wafers were
spectroscopy, the calculated benzophenone content and the feed spin-coated with different copolymers. Film thicknesses from
ratio, within the precision range of the method, are the same.
10 nm to nearly 1 mm are accessible with this technique. The
Binary copolymers of photoreactive BPEM with MEO2MA thickness of the layers is adjusted by the choice of the copoly-
and ternary copolymers of BPEM, MEO2MA and moderate mer concentration and the spinning speed. The best results
fractions of OEGMA are thermosensitive in aqueous media, regarding homogeneity of the lms were obtained with ethyl
showing a lower critical solution temperature (LCST) type acetate as a solvent for coating. Other solvents such as toluene,
transition. The cloud point of these copolymers can be easily THF, chloroform, ethanol or water did either lead to inhomo-
adjusted by varying the comonomer composition.35 In contrast, geneous layers, or were only useful for layers thinner than
binary copolymers of BPEM with OEGMA did not show thermo- 100 nm. As the nal layer thickness aer rinsing is affected by
responsive behaviour (cloud point beyond 90 ꢂC). Cloud points the molar mass of the copolymers used, the formation of a given
of the different copolymers can be taken from Table 1. Char- thickness requires slightly different spin-coating conditions for
acteristically, the cloud points of such ternary copolymers are different molar mass analogues of the same copolymer. Higher
Table 1 Characterization of the copolymers synthesised by free radical polymerization
BPEM
Turbidity
Expected transition
OEGMA/
MEO2MA
in the feed (mol%)
BPEM
incorporated into
BPEM units
Cloud transition for analogous OEGMA/
in the feed the copolymera
Mn
per polymer pointe midpoint
MEO2MA
b
c
Polymer
(mol%)
(kg molꢀ1
)
DPn
chaind
(ꢂC)
(ꢂC)
copolymersf (ꢂC)
POEGMA
B-1
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
0
0
79
138
91
292
80
350
49
125
84
166
291
192
615
169
738
104
664
385
529
536
344
201
0
—
—
—
—
—
—
—
24.8
35.5
49.2
56.6
22.3
32.0
—
—
—
—
—
—
—
25.7
36.1
50.8
57.9
23.1
33.2
<100
<100
<100
<100
<100
<100
<100
28.0
38.4
49.8
62.3
31.0
—
0.1
0.25
0.25
0.5
0.5
1.0
0.5
0.5
0.5
0.5
1.0
0.5
0.1
0.26
0.25
0.5
0.8
1.3
0.7
0.6
0.7
0.7
1.0
0.5
0.3
0.5
1.5
0.9
3.7
1.0
3.3
1.9
2.6
2.7
3.4
1.0
B-2a
B-2b
B-3a
B-3b
B-4
T-1
T-2
T-3
T-4
0 : 100
10 : 90
21 : 79
33 : 67
3 : 97
130
147
68
T-5
PNIPAM-x NIPAM
23
a
b
Determined by 1H-NMR spectroscopy. Apparent number average molecular weight Mn determined by SEC in DMF, calibrated by polystyrene.
c
d
e
f
Calculated from SEC results and monomer feed. Average number calculated from NMR and SEC results. Onset of turbidity. The midpoint
of turbidity transition of equivalent binary copolymers with the same MEO2MA/OEGMA ratio but without BPEM, according to ref. 15.
932 | Soft Matter, 2013, 9, 929–937
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
molar masses require lower polymer concentrations. As a lm
thickness of 100 nm was targeted, according to the molar mass
of the copolymer, a copolymer concentration of 1 wt% in ethyl
acetate was used for high molar mass compounds B3-b and
B2-b and a copolymer concentration of 2 wt% in ethyl acetate
was used for all other lower molar mass compounds. The spin-
speed for every copolymer solution was 50 rotations per second.
The thicknesses of the silane monolayer and the spin-cast
copolymer layer were determined by ellipsometry, assuming a
refractive index of 1.48 for both compounds (the refractive index
for bulk POEGMA was determined as nD (ref. 20) ¼ 1.480). All
spin-coated layers showed a specic interference colour. As the
refractive index of POEGMA is quite similar to that of silicon
dioxide (silicon dioxide nD (ref. 20) ¼ 1.458 and POEGMA nD
(ref. 20) ¼ 1.480), the interference colours of layers of the same
thickness for both materials should match well. Thus with the
help of a colour chart for thermally grown silicon dioxide, the
colour of the coated copolymer layer could be attributed to a
certain thickness.36,37 The thickness values determined by
ellipsometric measurements were in good agreement with
thicknesses estimated with the colour chart for SiO2 coatings.
3.4 Kinetics of UV-crosslinking
The photoreaction of the benzophenone moiety is monitored via
UV-spectroscopy aer spin-coating a thin copolymer lm onto a
quartz window. In order to get a sufficiently high absorbance
value, the layer thickness was chosen to be relatively thick (900 nm
as determined by a prolometer). The absorbance maximum at
287 nm was measured for the coated windows irradiated with
UV-A radiation for distinct time intervals (see Fig. 2b). With on-
going photoreaction, the absorbance band decreases gradually.
Under the given conditions, the half life time of the benzophe-
none moiety is found to be in the range of 1 min by UV/Vis
spectroscopy. The reaction is almost complete aer 10 min, and
the residual absorption reaches a plateau value. Accordingly,
samples were irradiated for 15 min to ensure maximum conver-
sion of the photoreactive groups in the experiments.
Fig. 3 Stability of the UV-cured hydrogel layers against washing/extraction:
remaining layer thickness on silicon substrates after washing – (a) remaining layer
thickness for different benzophenone contents in the copolymer, and (b)
comparison of the remaining layer thickness for copolymers of different molar
masses and benzophenone contents. Striped grey bars: lower molar mass
copolymers (Mn ¼ 50–140 kg molꢀ1, copolymers B-1, B-2a, B-3a, and B-4),
checkered white bars: higher molar mass copolymers (Mn ¼ 290–350 kg molꢀ1
,
copolymers B-2b and B-3b).
3.5 Stability of the crosslinked layers
For stability tests, the copolymers were crosslinked and
Highly swollen hydrogel systems are desirable in some bio- immobilized by exposing the samples to UV-A radiation for 15
applications. In such systems, the high liquid content imitates min. The unbound copolymer was then rinsed off, and remaining
conditions close to the ones in biological uids. Furthermore, traces were removed by washing in water overnight. The thus-
the higher mesh size enables effective diffusion of small and leached hydrogel layer showed resistance to phosphate buffer
medium sized biological compounds, like proteins, inside the saline solution (PBS) with a pH of 7.4 and to various organic
network structure. The use of oligo(ethylene glycol)-based solvents, such as chloroform, ethyl acetate and ethanol. When
copolymers in this context seems to be quite suitable, as they immersed in these liquids for 1 week, the thicknesses of the
are known to suppress unspecic interactions and thus prevent polymer layers remained constant. The extent of the elutriated sol
fouling.
fraction was strongly related to the molar fraction of the cross-
According to Flory–Rehner theory,38 swelling is inversely linker in the copolymer (see Fig. 3a, binary copolymers B-1 to B-4).
proportional to the crosslinker content. Thus for highly swollen While the main part of spin-coated copolymer B-4 with 1 mol% of
systems, a low crosslinker content is desirable. Still an innite crosslinking benzophenone groups was robustly immobilized
reduction of the crosslinker content is not feasible, as it is onto the surface (85% layer thickness aer washing), most of the
limited by the practical preparation of a stable network struc- spin-coated copolymer B-1 containing just 0.1% crosslinker was
ture. In order to achieve a maximum swelling, we thus tested the washed away despite the UV-crosslinking step: only 32% layer
stability of crosslinked hydrogel layers with the minimized thickness remained aer all washing steps. This nding is easily
crosslinker content.
rationalised. Every polymer chain has to be connected to other
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 929–937 | 933

View Article Online
Soft Matter
Paper
chains by at least two crosslinking points for creating a stable
network. For a stable binding of a single copolymer chain to the
network, at least one connection per chain is necessary. For
benzophenone bearing copolymers, Hayward et al. investigated
the possible reaction routes towards these crosslinks.25 Aer UV
activation, benzophenone moieties abstract one hydrogen from
adjacent polymer chains and thus create two radicals, one at the
benzophenone site and one at the activated carbon. They found
that recombination of the benzophenone radical with the radical
on the polymer chain is the most probable reaction. Thus, every
benzophenone moiety can create one link at best. Consequently,
the absolute minimum for a completely interconnected system is
one benzophenone unit per polymer chain. Obviously in copoly-
mers B-1, B-2a and B-3a the number of BPEM units per chain is
below this critical value. Thus, not all copolymer chains can be
immobilized, and the stability of the spin-coated layers decreases
with decreasing content of benzophenone moieties. Though
copolymer B-4 on average bears one benzophenone unit per
copolymer chain, not all polymer chains may in fact contain a
Fig. 4 Swelling ratios SR determined by ellipsometric measurements under
water. Crosslinked films of binary copolymers with different BPEM contents are
examined (B-1, B-2a, B-3a, and B-4).
benzophenone moiety due to a statistical distribution of the 3.6 Swelling behaviour of the hydrogel layers at room
benzophenone in the copolymer chains. Moreover, due to loop temperature
formation, not every activated BPEM unit may contribute to
crosslinking. Consequently also in copolymer B-4 not every poly-
The thickness and the refractive index of the swollen hydrogel
lms in pure water were determined by ellipsometry in a ow
mer chain is covalently connected to the hydrogel. For a given
cell (Fig. 4). The optical values of water as surrounding medium
crosslinker content, more stable layers can be assembled with
were considered for the calculation of the swollen lm. The
refractive index of the swollen layers determined by ellipsom-
copolymers of an elevated molar mass. Longer polymer chains
contain more BPEM units than copolymers of lower molar mass,
etry was in a range between 1.345 and 1.355. Since water is the
and hence, crosslinking and immobilisation onto the substrate is
main constituent of the hydrogel in the swollen state, its
refractive index is near the value of pure water (which is 1.332).
more efficient. The highly effective crosslinking for higher molar
mass copolymers is readily demonstrated by the behaviour of the
In theory, it is possible to determine the volume fraction of
pure copolymer in solution. When exposed to daylight, the higher
the copolymer in the swollen hydrogel layer and thus the
weight copolymers rapidly gelled and could not be dissolved any
swelling ratio from the mixed refractive index of the layer, since
more, whereas the lower molar mass copolymers tolerated brief
the refractive indices of the pure copolymer and pure water are
contact with daylight.
known.39 However, in highly swollen systems even small devi-
ations in the measured refractive index result in large variations
in the swelling ratio. Thus, we determined the swelling ratio
exclusively from the thickness in the swollen and in the dry
state. As the lm may only swell in one dimension, the swelling
ratio SR can be calculated from the thicknesses according to
Hydrogel layers built of copolymers with different molar
masses also differ in their stability (see Fig. 3b, copolymers
B-2b and B-3b). In copolymer B-2a with a molar mass of 90
kg molꢀ1 containing 0.25 mol% benzophenone comonomer, a
polymer chain only contains 0.5 benzophenone units on
average; therefore, it is not surprising that almost 50% of the
layer thickness is washed away in the rinsing procedure. In a
SR ¼ dswollen/ddry
copolymer with a higher molar mass of 300 kg molꢀ1 and the
same benzophenone content (B-2b), every copolymer chain in which dswollen is the equilibrium layer thickness of the
contains 1.5 benzophenone units per chain. The photo-cross- swollen hydrogel layer under water at a given temperature and
linked layer is thus more stable. Indeed about 80% of the ddry is the dry layer thickness.
originally deposited material remains immobilized onto the
We investigated rst the swelling behaviour of binary oli-
substrate aer rinsing. Copolymers with a higher benzophe- go(ethylene glycol methacrylate) copolymers with different
none content of 0.5 mol% generate even more stable networks. molar fractions of BPEM. Copolymers of this kind are soluble in
Here, the higher molar mass copolymer B-3b contains 3.7 water at temperatures from 0–100 ꢂC. The layer thicknesses
benzophenone units per chain on average. This polymer determined under water were considerably higher than the dry
remains on the substrate almost completely aer rinsing, layer thicknesses. The swelling ratio SR of these swollen
whereas lms made of copolymers B-3a with the same cross- hydrogels was inversely proportional to the BPEM content, as
linker content but a lower molar mass of around 80 kg molꢀ1 expected.38 When decreasing the BPEM ratio from 1% to 0.1%,
still lose about 30% of the original layer thickness upon the swelling ratio increased by a factor of 4.5 (see Fig. 4).
rinsing. These ndings highlight the importance of using high Therefore, the volume fraction of water in the swollen hydrogel
molar mass compounds, when stable coatings with low can be easily adjusted in a wide range via the molar fraction of
crosslinking density are desired.
benzophenone units, incorporated into the copolymer.
934 | Soft Matter, 2013, 9, 929–937
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
Fig. 5 Thermo-responsive collapse of a film formed from T-2, followed by
ellipsometry. Filled squares (-) refer to the layer thickness and open triangles (D)
refer to the refractive index.
3.7 Thermo-responsive hydrogel layers
The thermal behaviour of thermo-responsive hydrogel layers
was studied by temperature dependent ellipsometric measure-
ments. To exclude systematic errors caused by temperature
effects, a control experiment was performed rst: a blank silicon
wafer with an oxide layer of the known thickness was mounted
in the ow cell and the optical values were determined to be in
the range from 20 to 60 ꢂC. In the whole temperature range,
values for D and J did not change more than 0.8ꢂ and 0.2ꢂ,
respectively. For the considered hydrogel layers, the addition of
these D and J values to the determined layer thickness at
elevated temperature would result in a slight overestimation of
the layer thickness (i.e. 4 nm for an 88 nm thick layer). However,
this uncertainty has only a minor inuence on the thickness
values and would, if at all, counteract the observed extent of
shrinking of the layer caused by the phase transition. In the
following, this baseline effect is therefore neglected.
Fig.
6
Thermo-responsive collapse of various copolymers in water. Filled
symbols: swelling ratios SR of crosslinked copolymer films in water determined by
ellipsometry; and open symbols: turbidity measurements of copolymers in
aqueous solution. (a) Ternary copolymers of MEO2MA, OEGMA and BPEM. Upside
triangles (:, D) for copolymer T-4, circles (C, B) for copolymer T-3, and downside
triangles (;, V) for copolymer T-1. (b) Swelling ratios SR (-) and turbidity (,) of
the binary copolymer PNIPAM-x (NIPAM and BPEM).
analysed by a quartz crystal microbalance (QCM).17 To verify
The temperature induced deswelling of a layer of the thermo- whether the broad transition is indeed a characteristic feature
responsive hydrogel is shown in Fig. 5. Upon heating, the of the oligo(ethylene glycol)-based copolymers, we synthesised
thickness of the hydrogel layer decreases markedly, while an analogous PNIPAM copolymer also bearing the benzophe-
simultaneously the refractive index increases. Similar to the none unit. The ellipsometric data of the photocrosslinked
hydrogel layers that do not exhibit thermoresponsivity, the PNIPAM were collected and mathematically treated in the same
refractive index of the highly swollen hydrogel is close to way as the data of the oligo(ethylene glycol)-based lms. In
the refractive index of pure water. During the phase transition, contrast to the OEG containing lms, the PNIPAM lms show a
more and more water is expelled from the hydrogel network.
sharp thermal transition (Fig. 6b). Similar results have already
Thus the volume fraction of organic polymer in the hydrogel been reported for the comparison of thin lms made of poly-
gets larger, and the averaged refractive index approaches the (diethyl acrylamide) (PDEAAm) and PNIPAM.40 Schmaljohann
value of the pure polymer of approximately 1.48. At tempera- et al. attributed the gradual transition in the PDEAAm-lms to
tures below the onset of the collapse, the layer thickness seems different hydrogen bonding properties of PDEAAm compared to
to increase by about 10%. As in this temperature range, the PNIPAM.
difference between the refractive index of the aqueous medium
Similar to PDEAAm-copolymers, OEG-based copolymers can
and of the hydrogel layer is low, the observed differences only act as hydrogen bond acceptors, whereas PNIPAM addi-
however may be still within the experimental precision and tionally acts as a hydrogen bond donor. Increased intra- and
therefore cannot be considered signicant.
interchain interactions may thus sharpen the transition for
The thermal transition of the oligo(ethylene glycol)-based PNIPAM, while OEG-based copolymers in accordance with this
lms is rather broad and covers a temperature range of about explanation show a gradual transition.41,42 In fact, physically
ꢂ
20 C. This nding is in agreement with a report on the same crosslinked thin polymer lms of hydrophobically end-capped
copolymer system immobilised in a brush architecture, as PNIPAM have been shown by GISAXS experiments to undergo a
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 929–937 | 935

View Article Online
Soft Matter
Paper
lm is enhanced with increasing OEGMA content. This nding is
well consistent with similar investigations for these types of bulk
hydrogels.47 Looking at the shape of the transition curve, it is
obvious that the collapse neither follows a sigmoidal curve, as is
typical for cloud point measurements of polymer solutions, nor a
steady decrease. For copolymers with higher OEGMA contents of
21% and 33%, the transition curve can be separated into two
distinct ranges. At temperatures below the transition temperature,
the hydrogel gradually repels water from the network, whereas
above a specic temperature, a steeper decrease of the water
content is visible. The corresponding onset temperature increases
with the OEGMA content: for the lm made of T-3, it is in the
range of 30 ꢂC, whereas for the lm made of T-4, it is around 40
ꢂC. The steeper decay is attributed to the thermo-responsive
collapse of the hydrogel lm. Consistent with this observation, a
minor decrease in swellability was also found for a lm of pure
OEGMA polymer B-3a. Upon heating from 20 ꢂC to 60 ꢂC, the
swelling degree drops from 5.9 to 5.6. This is putatively attributed
Fig. 7 Switching cycles of a thermo-responsive copolymer film prepared from
T-1.
much narrower thermal transition in a water-saturated atmo- to the degeneration of solvent quality for oligo(ethylene glycol)s in
sphere in comparison to thin lms of analogously designed water with temperature.48 Still, the hydrogel lm remains highly
polymers, which have a short oligo(ethylene glycol)methylether hydrated even at high temperatures.
side chain, namely poly(methoxy diethylene glycol acrylate).42–44
For the thermo-responsive copolymer lms, the phase tran-
Even if the chemical structures of our polymer hydrogel lms sition temperatures are qualitatively comparable to those of the
differ from these examples in various variables, and our study copolymers in solution. However, the onset temperature for the
has been done in contact with liquid water, the observed transition is shied towards lower temperatures. Possibly this is
distinctions between PNIPAM and OEG-side chain polymers due to the higher copolymer concentration in the hydrogel as
during the thermal transition suggest a general pattern.
the transition temperature depends notably on the weight
The different hydrogen bonding interactions of PNIPAM are fraction of the copolymer in water.41 Yet, for a given weight
also held responsible for the observed hysteresis between the fraction the transition temperature is easily adjustable by the
phase transition in heating and cooling cycles in PNIPAM-based copolymer composition, so that thin polymer hydrogel lms of
polymers. In contrast, OEG-based polymers virtually do not any desired transition temperature can be prepared.
exhibit a hysteresis.15,45,46 In accordance with this observation
for the linear polymers in solution, all investigated lms formed
from OEG copolymers did not reveal any notable differences in
4 Conclusions
heating and cooling cycles. Even aer multiple switching cycles Oligo(ethylene glycol) copolymers bearing benzophenone units
the lms swelling ratio did not change (see Fig. 7).
provide an easy and versatile access to biocompatible polymer
For some applications this quality and a more gradual coatings on organic surfaces. We demonstrate that application-
transition could be favourable over a sharp transition. For relevant lm parameters, such as thickness, crosslinking density,
instance a slower transition potentially could favour a rapid swelling and phase transition behaviour can be adjusted at will,
transport of water out of the hydrogel, as skin formation at the by appropriate choice of copolymer composition, crosslinker
lm surface is reduced. Moreover, a more gradual change may content, molar mass of the parent polymers and spin coating
be advantageous for different sensing applications as here oen conditions. Compared to the frequently used thin lms of PNI-
a quantitative statement is preferred to a simple yes/no signal. PAM, the thermally induced swelling–deswelling transition is
In any case, our results demonstrate that the precise chemical broadened, but shows no hysteresis. Thus, the new photoreactive
structure of thermo-responsive hydrogels has more implica- oligo(ethylene glycol) copolymers represent an interesting basis
tions than simply dening the transition temperature.
for coatings with potential application in the bio-medical eld.
A special interest in the use of oligo(ethylene glycol) based
copolymers lies in the easy tunability of the phase transition
temperature. Hence we also veried that the temperature
Acknowledgements
response of the prepared lms can be adjusted to certain The authors thank Matthias Rohn (Fraunhofer IAP) for help
temperatures. We therefore prepared lms of different como- during polymer synthesis and layer preparation, Christoph
nomer compositions (copolymers T-1, T-3 and T-4) and followed Wieland (University of Potsdam) for SEC experiments, and Frank
the phase transition by temperature dependent ellipsometric Bier (Fraunhofer IBMT) for stimulating discussions. Financial
measurements. Fig. 6a illustrates the transition curves for copol- support was provided by the initiative “Spitzenforschung &
¨
ymers with different ratios of longer oligo(ethylene glycol) side Innovation in den neuen Landern” of the German Federal
chains. The higher the OEGMA content, the higher the phase Ministry of Education and Research (BMBF) (grants 03IS2201B
transition temperatures. Likewise, the swellability of the hydrogel and 03IS2201C).
936 | Soft Matter, 2013, 9, 929–937
This journal is ª The Royal Society of Chemistry 2013

View Article Online
Paper
Soft Matter
27 G. Dorman and G. D. Prestwich, Biochemistry, 1994, 33,
Notes and references
5661–5673.
¨
1 J.-F. Lutz and H. G. Borner, Prog. Polym. Sci., 2008, 33, 1–39.
28 V. Kumar and D. Kalonia, AAPS PharmSciTech, 2006, 7, E47–
2 J. M. Harris, Poly(ethylene Glycol) Chemistry: Biotechnical and
Biomedical Applications, Plenum Press, New York, USA, 1992.
3 H. Ma, M. Wells, T. P. Beebe and A. Chilkoti, Adv. Funct.
Mater., 2006, 16, 640–648.
4 D. Sung, S. Park and S. Jon, Langmuir, 2009, 25, 11289–11294.
5 G. R. Hendrickson and L. A. Lyon, So Matter, 2009, 5, 29–35.
6 A. Hucknall, S. Rangarajan and A. Chilkoti, Adv. Mater., 2009,
21, 2441–2446.
E53.
29 D. Ruhlmann, K. Zahouily and J. P. Fouassier, Eur. Polym. J.,
1992, 28, 1063–1067.
30 A. M. Rufs, A. Valdebenito, M. C. Rezende, S. Bertolotti,
C. Previtali and M. V. Encinas, Polymer, 2008, 49, 3671–
3676.
ˇ
´ˇ
31 P. Hrdlovic, I. Lukac and I. Zvara, Eur. Polym. J., 1981, 17,
1121–1126.
7 J. Trmcic-Cvitas, E. Hasan, M. Ramstedt, X. Li, M. A. Cooper,
C. Abell, W. T. S. Huck and J. E. Gautrot, Biomacromolecules,
2009, 10, 2885–2894.
8 I. Tokarev and S. Minko, So Matter, 2009, 5, 511–524.
9 C. Pietsch, R. Hoogenboom and U. S. Schubert, Angew.
Chem., Int. Ed., 2009, 48, 5653–5656.
32 J. Weiss, A. Li, E. Wischerhoff and A. Laschewsky, Polym.
Chem., 2012, 3, 352–361.
33 H. G. Tompkins and W. A. McGahan, Spectroscopic
Ellipsometry and Reectometry: A User’s Guide, John Wiley &
Sons, New York, USA, 1999.
34 N. J. Turro, Modern Molecular Photochemistry, University
Science Books, Sausalito, USA, 1991.
¨
10 E. Wischerhoff, K. Uhlig, A. Lankenau, H. Borner,
A. Laschewsky, C. Duschl and J.-F. Lutz, Angew. Chem., Int.
Ed., 2008, 47, 5666–5668.
35 J.-F. Lutz, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 3459–
3470.
11 J. E. Gautrot, B. Trappmann, F. Oceguera-Yanez, J. Connelly,
X. He, F. M. Watt and W. T. S. Huck, Biomaterials, 2010, 31,
5030–5041.
12 E. Wischerhoff, N. Badi, J.-F. Lutz and A. Laschewsky,
So Matter, 2010, 6, 705–713.
13 J. Buller, A. Laschewsky, J.-F. Lutz and E. Wischerhoff, Polym.
Chem., 2011, 3, 352–361.
14 J. F. Lutz, O. Akdemir and A. Hoth, J. Am. Chem. Soc., 2006,
128, 13046–13047.
36 W. A. Pliskin and E. E. Conrad, IBM J. Res. Dev., 1964, 8, 43–
51.
37 S. Franssila, in Introduction to Microfabrication, John Wiley &
Sons, Ltd, New York, USA, 2010, pp. 503–504.
38 P. J. Flory and J. Rehner, J. Chem. Phys., 1943, 11, 521–527.
39 M. Born and E. Wolf, Principles of Optics: Electromagnetic
Theory of Propagation, Interference and Diffraction of Light,
Cambridge University Press, Cambridge, UK, 2011.
40 D. Schmaljohann, D. Beyerlein, M. Nitschke and C. Werner,
Langmuir, 2004, 20, 10107–10114.
15 J. F. Lutz and A. Hoth, Macromolecules, 2006, 39, 893–896.
16 C. Weber, R. Hoogenboom and U. S. Schubert, Prog. Polym.
Sci., 2012, 37, 686–714.
41 V. Aseyev, H. Tenhu and F. Winnik, Adv. Polym. Sci., 2011,
242, 29–89.
17 X. Laloyaux, B. Mathy, B. Nysten and A. M. Jonas, Langmuir,
2010, 26, 838–847.
42 Q. Zhong, E. Metwalli, G. Kaune, M. Rawolle, A. M. Bivigou-
Koumba, A. Laschewsky, C. M. Papadakis, R. Cubitt and
18 A. Synytska, E. Svetushkina, N. Puretskiy, G. Stoychev,
S. Berger, L. Ionov, C. Bellmann, K.-J. Eichhorn and
M. Stamm, So Matter, 2010, 6, 5907–5914.
19 H. Kitano, T. Kondo, H. Suzuki and K. Ohno, J. Colloid
Interface Sci., 2010, 345, 325–331.
¨
P. Muller-Buschbaum, So Matter, 2012, 8, 5241–5249.
43 W. Wang, K. Troll, G. Kaune, E. Metwalli, M. Ruderer,
K. Skrabania, A. Laschewsky, S. V. Roth, C. M. Papadakis
¨
and P. Muller-Buschbaum, Macromolecules, 2008, 41, 3209–
3218.
20 S. Tocker, US Patent 3062784, 1963.
21 O. Prucker, C. A. Naumann, J. Ruhe, W. Knoll and
44 Q. Zhong, W. Wang, J. Adelsberger, A. Golosova,
A. M. B. Koumba, A. Laschewsky, S. S. Funari, J. Perlich,
¨
C. W. Frank, J. Am. Chem. Soc., 1999, 121, 8766–8770.
22 N. Griep-Raming, M. Karger and H. Menzel, Langmuir, 2004,
20, 11811–11814.
¨
S. V. Roth, C. M. Papadakis and P. Muller-Buschbaum,
Colloid Polym. Sci., 2011, 289, 569–581.
¨
45 O. Akdemir, N. Badi, S. Pfeifer, Z. Zarafshani, A. Laschewsky,
¨
23 R. Toomey, D. Freidank and J. Ruhe, Macromolecules, 2004,
E. Wischerhoff and J.-F. Lutz, ACS Symp. Ser., 2009, 1023,
189–202.
37, 882–887.
24 M. E. Nash, W. M. Carroll, P. J. Foley, G. Maguire,
C. O. Connell, A. V. Gorelov, S. Beloshapkin and
Y. A. Rochev, So Matter, 2012, 8, 3889–3899.
25 S. K. Christensen, M. C. Chiappelli and R. C. Hayward,
Macromolecules, 2012, 45, 5237–5246.
46 A. Miasnikova and A. Laschewsky, J. Polym. Sci., Part A:
Polym. Chem., 2012, 50, 3313–3323.
´
47 R. Parıs and I. Quijada-Garrido, Eur. Polym. J., 2009, 45,
3418–3425.
48 E. A. Bekturov and Z. K. Bakauova, Synthetic Water-Soluble
26 J. W. McGinity, T. R. Patel, A. H. Naqvi and J. A. Hill, Drug
Dev. Ind. Pharm., 1976, 2, 505–519.
¨
Polymers in Solution, Huthig & Wepf, Basel, Switzerland,
1986.
This journal is ª The Royal Society of Chemistry 2013
Soft Matter, 2013, 9, 929–937 | 937