Soft hydrogels from tetra-functional PEGs using UV-induced thiol-ene coupling chemistry: A structure-to-property study

Article abstract of DOI:10.1039/c4ra04335a

In this work, photo-induced thiol-ene coupling (TEC) was used to produce well-defined poly(ethylene glycol) (PEG)-based hydrogels. PEGs of four different molecular weights (2k, 6k, 10k, and 20k) were functionalized with G1-allyl dendrons using anhydride chemistry to produce tetra-functional TEC crosslinkable PEGs. The tetra-functional PEGs were subsequently crosslinked with a tri-functional thiol in ethanol to form hydrogels. The synthesized hydrogels were characterized with respect to swelling behaviour, rheological properties and hydrolytic degradation. It was found that the molecular weight of the PEG chain greatly influences the final properties of the hydrogel, where a higher molecular weight of PEG gives an increased weight swelling ratio from 240% for PEG-2k hydrogels to 1400% for PEG-20k hydrogels, as well as decreased elastic moduli, with Young's moduli ranging from 106 MPa to 6 MPa, for PEG-2k and PEG-20k hydrogels, respectively. It was also found that the hydrolytic stability in alkaline conditions (pH 10) decreased when the molecular weight of PEG in the hydrogels increased. the Partner Organisations 2014.

Full text of DOI:10.1039/c4ra04335a

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Soft hydrogels from tetra-functional PEGs using
UV-induced thiol–ene coupling chemistry: a
structure-to-property study†
Cite this: RSC Adv., 2014, 4, 30118
*
Kristina Olofsson, Michael Malkoch and Anders Hult
In this work, photo-induced thiol–ene coupling (TEC) was used to produce well-defined poly(ethylene
glycol) (PEG)-based hydrogels. PEGs of four different molecular weights (2k, 6k, 10k, and 20k) were
functionalized with G1-allyl dendrons using anhydride chemistry to produce tetra-functional TEC
crosslinkable PEGs. The tetra-functional PEGs were subsequently crosslinked with a tri-functional thiol in
ethanol to form hydrogels. The synthesized hydrogels were characterized with respect to swelling
behaviour, rheological properties and hydrolytic degradation. It was found that the molecular weight of
the PEG chain greatly influences the final properties of the hydrogel, where a higher molecular weight of
PEG gives an increased weight swelling ratio from 240% for PEG-2k hydrogels to 1400% for PEG-20k
hydrogels, as well as decreased elastic moduli, with Young's moduli ranging from 106 MPa to 6 MPa, for
PEG-2k and PEG-20k hydrogels, respectively. It was also found that the hydrolytic stability in alkaline
conditions (pH 10) decreased when the molecular weight of PEG in the hydrogels increased.
Received 9th May 2014
Accepted 26th June 2014
DOI: 10.1039/c4ra04335a
www.rsc.org/advances
PEG hydrogels have traditionally been formed through free-
radical chain-growth polymerization of PEG with e.g. acrylic or
Introduction
methacrylic end-functionalities, where the polymerization of
monomers that have more than one functional group leads to
network formation. However, in chain-growth polymerization,
high molecular weight polymers are formed at low monomer
conversions, leading to early gelation. This furthermore leads to
restricted movement of monomers through the gel, which can
result in inhomogeneous regions in the structure of the
resulting network.
Better control over the network structure can be achieved
through step-growth polymerization using precursors with a
narrow molecular weight distribution. Step-growth polymeri-
zation thus enables the formation of hydrogels that are more
homogeneous in their structure compared to networks formed
using traditional chain-growth mechanisms.8 Using orthogonal
crosslinking chemistries, such as the copper-catalyzed azide–
alkyne cycloaddition (CuAAC) or thiol–ene coupling (TEC),
enables an even higher level of structural control over the
network.9
TEC follows a radical-mediated step-growth mechanism that
can proceed under mild reaction conditions in various solvents,
as well as in the presence of oxygen, with a low level of side
reactions.10,11 It has previously been successfully utilized in the
formation of PEG-based hydrogels of linear allyl-functional
PEG, where the TEC-crosslinking strategy was utilized as a
powerful tool to design traditional hydrogels as well as inter-
penetrating networks (IPNs) with tunable mechanical proper-
ties and swelling behaviour.12–14 Also, TEC has been used in the
crosslinking of norbornene-functionalized PEG with cysteine-
Hydrogels have attracted much attention for a variety of appli-
cations, including anti-fouling coatings, super absorbents,
contact lenses, and drug delivery. Due to their high water
content, hydrogels are of interest for many biomedical appli-
cations such as regenerative medicine, tissue engineering and
drug delivery.1–3 For these applications, good control over the
chemical structure of the material is necessary, as well as
adequate knowledge about structure–property relationships.
One material that has gained interest in this area is poly-
(ethylene glycol) (PEG) hydrogels. PEG is a non-toxic, biocom-
patible polymer, and PEG-based hydrogels have shown great
promise in the biomedical eld.1–5 The properties of hydrogels
can be tuned by varying the composition of the network and the
conditions under which it is formed. The properties can addi-
tionally be modied by functionalization of pendant functional
groups in the polymer or through copolymerization. As an
example, introduction of the peptide sequence Arg-Gly-Asp
(RGD) has been successful in promoting cell adhesion to
otherwise inert PEG hydrogels,6 and it has also been shown that
incorporation of cleavable moieties in the structure can be used
to tune the degradation behaviour.7
KTH Royal Institute of Technology, Department of Fibre and Polymer Technology,
Teknikringen 56-58, SE-100 44 Stockholm, Sweden. E-mail: andult@kth.se; Fax: +46
8 790 8283; Tel: +46 8 790 8268
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c4ra04335a
30118 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
containing peptides. Hydrogels of this type have e.g. been was used as an internal standard to correct for ow rate uc-
investigated as scaffolds for culturing nerve cells15 and as tuations and PSS WinGPC Unity soware (version 7.2) was used
biodegradable gels for enzyme-responsive protein delivery to to process obtained data.
inammation sites.16 Additionally, TEC has been successful in
Differential Scanning Calorimetry (DSC) was carried out on a
the accelerated synthesis of dendrimers17 and utilized in the Mettler Toledo DSC-1 equipped with the Gas Controller GC100.
formation of bioactive functional hydrogels from linear The analyses were run under nitrogen atmosphere (30 mL
dendritic hybrids of PEG and bis-MPA-based bi-functional minꢀ1) at a temperature interval of ꢀ65 ꢁC to 150 ꢁC with a
dendrimers.18 The highly branched and exact structure of den- heating/cooling rate of 10 ꢁC minꢀ1 and 2 min isotherms. STARe
drimers offers new possibilities in the formation of functional Excellence Soware was used for evaluation of data.
hydrogels and advanced 3D networks.19
Matrix-assisted laser-desorption ionization - time-of-ight
The aim of this study was to combine TEC with basic den- mass spectroscopy (MALDI-ToF MS) was conducted on a
drimer chemistry in order to gain a better understanding of how Bruker UltraFlex MALDI-ToF MS with a SCOUT-MTP ion
the properties of well-dened rst-generation allyl-functional source (Bruker Daltonics) equipped with a N2-laser (337 nm),
PEG hydrogels depend on the structure of the network. and a gridless ion source. Acquisition of spectra was done in
Hydrogels with a high level of structural control were synthe- reector mode, with the exception of spectra for PEG-20k-OH
sized from rst-generation tetra-functional PEGs via photo- and PEG-20k-G1-allyl which were acquired in linear mode, and
initiated TEC. The swelling and mechanical properties of these the samples were prepared using using HABA and NaTFA as
hydrogels were assessed and related to the network structure. matrix and counter ion source, respectively. SpheriCal™ cal-
Additionally, degradation of the hydrogels in aqueous solutions ibrants (Polymer Factory Sweden AB) were used to calibrate
with different pH values was evaluated.
the instrument and obtained spectra were analyzed using
FlexAnalysis (Bruker Daltonics, v. 2.2). Average molecular
ꢀ
weights (Mn) and dispersities (Đ) were acquired using Poly-
tools v. 1.0.
Experimental
Materials
Rheology measurements were carried out on hydrogels
swollen to equilibrium on an Ares RDA-III rheometer (TA
Instruments) at 25 ꢁC using a Ø 25 mm parallel-plate geometry.
Dynamic strain sweep tests (0–60% strain, 10 rad sꢀ1) and
dynamic frequency sweep tests (0–100 rad sꢀ1, 5% strain) were
used to identify the linear region, aer which dynamic time
sweep tests (5% strain, 10 rad sꢀ1) were used to determine the
storage and loss moduli of the hydrogels.
All chemicals were purchased from Sigma-Aldrich and used as
received, unless otherwise stated.
2,2-Bis(methylol)propionic acid (Bis-MPA) was kindly
supplied by Perstorp Holding AB, Sweden. HCl, MgSO4, and
N,N0-dicyclohexylcarbodiimide (DCC) were purchased from
Acros Organics, Fisher Scientic. Pyridine and ethanol (96%)
were purchased from VWR. The UV initiator Irgacure 2959
(I2959) was purchased from Ciba Speciality Chemicals, Inc. and
pH buffer solutions with pH 4, 7, and 10 were purchased from
Scharlau. Deuturated chloroform (CDCl3) for NMR analyses was
purchased from Cambridge Isotope Laboratories, Inc. Poly-
(ethylene glycol) (PEG) with molecular weights of 2 kDa, 6 kDa,
10 kDa and 20 kDa were purchased from Sigma-Aldrich and
freeze-dried or stripped with toluene before use.
Tensile tests were conducted on an Instron 5944 tensile
testing machine equipped with a standard video extensom-
eter. Measurements were done on dog-bone-shaped hydro-
ꢁ
gels swollen to equilibrium in deionized (DI) water at 23 C
using a cross-head speed of 5 mm minꢀ1 and a 500 N load
cell.
Synthesis of bis(allyl propionic acid), BAPA
Instrumentation
Synthesis of BAPA was accomplished in accordance with a
Nuclear magnetic resonance (NMR) spectroscopy was attained previously published procedure.17 Bis-MPA (50.00 g, 0.3728 mol)
on a Bruker Avance 400 MHz instrument. 1H NMR spectra were and NaOH (149.1 g, 3.728 mol) were dissolved in toluene (750
acquired at 400 MHz and 13C NMR spectra were acquired at 100 mL) and heated to 110 ꢁC. Allyl bromide (225.8 mL, 2.609 mol)
MHz. CDCl3 was used as solvent, unless otherwise stated, and was added and the reaction was carried out overnight. The
the residual solvent peak was used as internal standard.
reaction mixture was allowed to reach room temperature and
Raman spectra were acquired on a Perkin-Elmer Spectrum then acidied using concentrated HCl (37%, 150 mL). The
2000 NIR-Raman and obtained data was analysed using Spec- product was puried through extraction three times with 250
trum soware. All spectra were based on 32 scans and acquired mL DI water, aer which the toluene phase was dried over
using a laser power of 800–1500 mW.
MgSO4 and the solvent removed under reduced pressure to yield
Size exclusion chromatography (SEC) was performed on a BAPA as a colourless viscous liquid (72.65 g, 0.3391 mol, 90.97%
TOSOH EcoSEC HLC-8320GPC system from PSS GmbH equip- yield). 1H NMR (CDCl3, 400 MHz): 1.24 (s, –CH3), 3.56 (q,
ped with an EcoSEC RI detector and three columns: PSS PFG 5
, 4.00 (dd, –O–CH2–CH]), 5.24 (m, –CH]CH2),
mm; Microguard, 100 A; and Microguard 300 A (MW resolving 5.87 (m, –CH]CH2), 11.33 (br s, –OH). 13C NMR (CDCl3, 100
˚
˚
range: 100–300 000 g molꢀ1). DMF (0.2 mL minꢀ1) with 0.01 M MHz): 18.06 –CH3), 48.30 (
LiBr was used as the mobile phase at 50 ꢁC. A conventional 72.51 (
), 71.93 (–O–CH2–CH]),
), 117.02 (]CH2), 134.72 (–CH]CH2),
calibration method was created using PEG standards. Toluene 180.82 (HOOC–).
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 30118–30128 | 30119

View Article Online
RSC Advances
Paper
DCM in diethyl ether 3 times. (16.57 g, 2.593 mmol, 77.89%
yield). 1H NMR (CDCl3, 400 MHz): d 1.22 (s, –CH3), 3.57 (d,
), 3.65 (s, –CH2– (PEG)), 3.98 (dd, –O–CH2–CH]),
4.26 (t, PEG–CH2–OOC–G1), 5.15 (dd, –CH]CH2 (cis)), 5.25 (dd,
–CH]CH2 (trans)), 5.86 (m, –CH]CH2). 13C NMR (CDCl3, 100
Synthesis of bis(allyl propionic acid anhydride), BAPA
anhydride
BAPA anhydride 1 was synthesized via DCC coupling in accor-
dance with a previously published procedure20 and conducted
as follows; BAPA (71.47 g, 0.3336 mol) was dissolved in 100 mL
dichloromethane (DCM) and cooled to 0 ꢁC. DCC (34.41 g,
0.1668 mol), mixed with 10.00 mL DCM, was slowly added to the
reaction vessel and the reaction mixture was stirred overnight.
By-products were ltered off and the solvent was evaporated
under reduced pressure to give BAPA anhydride 1 as a viscous
liquid (53.87 g, 0.1312 mol, 78.68% yield). 1H NMR (CDCl3, 400
MHz): d 18.11 (–CH3), 48.56 (
), 63.82 (–PEG–CH2–OOC–
G1), 69.29 (–O–CH2–CH2–COO–G1), 70.74 (–CH2– (PEG)), 72.22
(–O–CH2–CH]), 72.41 (
), 116.62 (]CH2), 135.02
(–CH]CH2), 174.70 (PEG–OOC–G1). SEC: Mn ¼ 6266 g molꢀ1, Đ
¼ 1.04.
ꢀ
The same purication protocol was used for PEG-10k-G1-
allyl and PEG-20k-G1-allyl, with precipitation in diethyl ether
2–4 times, whereas PEG-2k-G1-allyl was puried through
subsequent extraction with NaHSO4 (10%) and NaHCO3 (10%),
aer which the organic phase was dried over MgSO4, ltered,
and the solvent removed under reduced pressure.
MHz): d 1.27 (s, –CH3), 3.58 (s,
CH]), 5.24 (m, –CH]CH2), 5.86 (m, –CH]CH2). 13C NMR
(CDCl3, 100 MHz): d 17.44 (–CH3), 49.95 ( ), 71.54 (–O–
CH2–CH]), 72.44 ), 116.92 (]CH2), 134.65
), 3.99 (d, –O–CH2–
(
(–CH]CH2), 169.64 (–OOC–O–COO–).
General procedure for the synthesis of PEG-G1-allyl
General procedure for hydrogel formation
ꢀ
PEG-OH with Mn ¼ 2 kDa, 6 kDa, 10 kDa, and 20 kDa were Hydrogels were produced by crosslinking PEG-G1-allyl 2 with
dendronized using BAPA anhydride 1 to give PEG-G1-allyl 2, trimethylolpropane tris(3-mercaptopropionate) (TMP-(SH)3)
Scheme 1a. In the following paragraph, the synthesis of PEG-6k- using an allyl-to-thiol ratio of 1 : 1 and a solid content of 50 wt%
G1-allyl will be described in detail. Detailed information in ethanol (96%), Scheme 1b. The formation of hydrogels from
regarding the syntheses of PEG-2k-G1-allyl, PEG-10k-G1-allyl, PEG-6k-G1-allyl will here be described in detail. Details
and PEG-20k-G1-allyl can be found in the ESI.†
regarding the formation of hydrogels from PEG-2k-G1-allyl,
ꢀ
Synthesis of PEG-6k-G1-allyl. PEG-OH (Mn ¼ 6 kDa) (19.97 g, PEG-10k-G1-allyl, and PEG-20k-G1-allyl can be found in
3.328 mmol), DMAP (163.0 mg, 1.334 mmol), and pyridine the ESI.†
(1.610 mL, 19.99 mmol) were dissolved in 15.00 mL DCM. Using
Formation of PEG-6k-G1-allyl hydrogels. PEG-6k-G1-allyl
an ice-bath, BAPA anhydride 1 (3.830 g, 9.330 mmol), dissolved (2.000 g, 0.3129 mmol) and TMP-(SH)3 (137.0 mL, 0.4159 mmol)
in 1.000 mL DCM, was added dropwise and the solution was were dissolved in ethanol (96%, 2.758 mL) along with the
allowed to react over-night. An additional amount of DMAP initiator I2959 (10.9 mg, 0.50 wt% of total solid content).
(80.0 mg, 0.666 mmol) was added to ensure complete reaction Hydrogels were formed between two glass slides via UV irradi-
and full conversion of OH-groups was conrmed using 13C NMR ation using a Blak-Ray® XX-15BLB UV Bench Lamp (UVP,
analysis. Purication was achieved through precipitation from Cambridge, UK, 365 nm, 28.5 mW cmꢀ2) for 20 min. Aer
Scheme 1 Synthetic procedures. (a) Synthesis of PEG-G1-allyl 2 via functionalization of PEG-OH using BAPA anhydride 1 and (b) hydrogel
formation via UV-induced TEC.
30120 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
curing, hydrogels were washed in DI water (overnight + 2 h + 2 h) anhydride. Additionally, the peak in the spectra that corre-
and subsequently in ethanol (96%, 2 h + 2 h) to remove residual sponds to the carbonyl carbon shied from 169 ppm for the
reactants and thereaer the hydrogels were dried in air over- anhydride to 173 ppm for the formed ester in PEG-6k-G1-allyl.
night followed by 12 h in a vacuum oven at 50 ꢁC. The gel The chemical shis for the other peaks were also in accordance
fraction (GF) was determined from the differences in dry weight with expected values, as were the chemical shis in the 1H NMR
of hydrogels before and aer washing according to eqn (1):
spectra in Fig. 1a.
SEC was used to verify the functionalization and to deter-
wd;washed
GF ¼
ꢂ100%
(1)
ꢀ
mine the number average molecular weights (Mn) of function-
wd;cured
alized as well as unmodied PEG materials. In Fig. 2a, an
increase in molecular weight can be seen between unmodied
and modied PEG, verifying the functionalization. The dis-
persities calculated from SEC were narrow for all materials with
values below 1.06. The obtained molecular weights, which can
be found in Table 1, were found to correlate reasonably well
with the expected values.
For both PEG-20k-OH and PEG-20k-G1-allyl, a broadening of
the signals from SEC could be seen, Fig. 2a. That this broad-
ening effect could be seen for both modied and unmodied
PEG20k, suggested that it was an artefact from the starting
material and not an effect of the functionalization.
Successful functionalization of PEG-OH to PEG-G1-allyl was
also conrmed by MALDI-ToF MS, Fig. 2b. If the 2k system is
studied as an example, it can be seen that the distance between
the two populations in the two spectra is 392.5 m/z, which
corresponds to the substitution of both hydroxyl end groups to
G1-allyl. Furthermore, as SEC results suggested, two pop-
ulations with different molecular weights can be seen for both
modied and unmodied PEG-20k. A closer look conrms
complete functionalization and the bimodal nature of the
samples is thus an artefact from the original synthesis of PEG-
20k-OH.
where GF is the gel fraction (%), wd,washed is the weight of the
dried hydrogel aer washing (g), and wd,cured is the weight of the
dried hydrogel aer curing (g).
Swelling experiments
Swelling experiments in DI water were conducted on thoroughly
washed and dried hydrogels at 23 ꢁC. Hydrogels were dried in a
ꢁ
vacuum oven at 50 C before the dry weight of each specimen
was measured. The swelling in water was measured at several
different swelling times for each specimen. Before weighing,
excess water was gently removed from the surface of the
hydrogels using a tissue paper. The swelling ratio (SR) was
calculated using eqn (2):
mw ꢀ md
SR ¼
(2)
md
where SR is the swelling ratio, mw is the weight of the swollen
hydrogel (g), and md is the weight of the dry hydrogel before
swelling (g).
Degradation study
Hydrolytic degradation of the hydrogels was studied in buffer
solutions from Scharlau with three different pH values, pH 4
(potassium hydrogen phthalate), pH 7 (potassium di-hydrogen
phosphate/di-sodium hydrogen phosphate), and pH 10 (sodium
carbonate/sodium hydrogen carbonate). Degradation times
were estimated from visual inspection and two replicas were
used for each hydrogel system. The study was terminated aer
three months, aer which the solutions were freeze-dried and
analysed using MALDI-ToF MS and NMR.
Thermal properties
The thermal properties of synthesized materials were assessed
using differential scanning chromatography (DSC). Results
from DSC in Fig. 3 show that the temperatures for crystalliza-
tion and melting, as well as the degree of crystallinity, are
reduced when PEG-OH is modied into PEG-G1-allyl. This can
be explained by a reduced ability to form hydrogen bonds
between the polymer chains and also that the G1-allyl end
groups introduce more free volume in the materials. It should
be noted that the melting enthalpy for 100% crystalline PEG-OH
Results and discussion
Synthesis of allyl-functional PEGs
PEG-OH (Mn ¼ 2000, 6000, 10 000, and 20 000 g molꢀ1) were has been used for the calculations of the degree of crystallinity
functionalized with BAPA anhydride 1 to give PEG-G1-allyl 2 for both modied and unmodied PEGs. The values for the
with four allyl groups available for UV-initiated TEC reactions. degree of crystallinity for the modied PEG materials should
The functionalization of PEG-OH was monitored with 13C thus only serve as an indication and not be considered to be
NMR to ensure complete conversion of the hydroxyl groups. The absolute values. It is however clear that an increased chain
13C NMR spectrum for the pure product of PEG-6k-G1-allyl can length of PEG-G1-allyl increases the transition temperatures of
be seen in Fig. 1b, along with spectra for BAPA anhydride and melting and crystallization, as well as the degree of crystallinity.
unmodied PEG-6k-OH. As a result of the functionalization, The inuence of the functionalization, i.e. the differences
some characteristic peaks in the spectra changed shi. This was between PEG-OH and PEG-G1-allyl, is less apparent when the
used to monitor all allylation reactions until complete conver- PEG chain becomes longer, probably because the inuence of
sion of OH-groups was observed. As can be seen when the chain ends is less prominent when the chain length
comparing the spectra in Fig. 1b, the peak corresponding to the increases. The crystallization temperature was e.g. lowered by 9
methylene carbon in PEG-OH closest to the OH-group changed degrees (from 27 ꢁC to 18 ꢁC) for the 2k system, while the
shi from 61 ppm to 63 ppm upon functionalization with BAPA decrease in crystallization temperature for the 20k system was
ꢀ
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 30118–30128 | 30121

View Article Online
RSC Advances
Paper
Fig. 1 (a) 1H NMR spectra and (b) 13C NMR spectra of BAPA anhydride 1, unfunctionalized PEG-6k-OH, and synthesized PEG-6k-G1-allyl 2. The
residual solvent peak of CDCl3 was used as internal reference.
Fig. 2 (a) Molecular weights for all functionalized and unfunctionalized PEG materials obtained by SEC in DMF calibrated against PEG standards.
(b) MALDI-ToF spectra for all modified and unmodified PEGs. Spectra for PEG-20k and PEG-20k-G1-allyl were acquired in linear mode.
30122 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table 1 Comparison of theoretical and measured molecular weights
and dispersities obtained from SEC and MALDI-ToF MS for unmodified
and modified PEG materials. (All molecular weights are reported in g
molꢀ1.)
In order to determine if there were unreacted allyl or thiol
groups le in the hydrogels aer curing, unwashed hydrogels
were studied using RAMAN spectroscopy. The RAMAN spectra
shown in Fig. 4 suggest that all hydrogels, including hydrogels
from PEG-2k-G1-allyl, were fully cured, as no traces of allyl or
a
b
Sample
Mn,th
Mn
Đa
Mn
thiol could be seen in the spectra, at 1730 and 2573 cmꢀ1
,
respectively. (The peak seen at 1734 cmꢀ1 for TMP-(SH)3
corresponds to the C]O stretch of the carbonyls in the struc-
ture of the crosslinker.) Effective curing, in combination with
high gel content, thus showed that UV-induced TEC was a
robust strategy for hydrogel formation.
PEG-2k
PEG-2k-G1-allyl
PEG-6k
PEG-6k-G1-allyl
PEG-10k
PEG-10k-G1-allyl
PEG-20k
2000
2393
6000
6393
10 000
10 393
20 000
20 393
1727
2147
5862
6266
11 278
11 571
21 676
21 915
1.04
1.03
1.02
1.04
1.03
1.03
1.05
1.06
2004
2434
6059
6426
11 047
11 573
—
PEG-20k-G1-allyl
—
Swelling properties
a
b
Values obtained from DMF-SEC using PEG standards. Values
obtained from MALDI-ToF MS data using Polytools soware. No
values could be obtained for PEG-20k and PEG-20k-G1-allyl due to
poor resolution of the spectra.
The high water content of hydrogels is one of the main properties
of interest for many biomedical applications. In accordance with
theory, there is a strong correlation between the swelling prop-
erties and the chain length between crosslinks.21 The kinetic
swelling study conducted in this work is summarized in Fig. 5a
and shows that the equilibrium swelling increases with the
length of the hydrophilic PEG chain, from 240% to 1400% for
hydrogels based on PEG-2k-G1-allyl and PEG-20k-G1-allyl,
respectively, as seen in Fig. 5b and Table 2. These swelling ratios
are lower than swelling ratios obtained for similar hydrogel
systems from linear di-functional PEG-allyl with the same
only 0.8 degrees (from 44.6 ꢁC to 43.8 ꢁC). There were even larger
differences in the melting temperatures, where the functional-
ization caused a decrease in melting temperature with 21
degrees for the 2k system and 5 degrees for the 20k system.
Hydrogel formation
Hydrogels of PEG-G1-allyl and TMP-(SH)3 were successfully molecular weights, where the swelling for 2k and 6k hydrogels
formed under irradiation with UV light. The hydrogels had a were found to be 380% and 750%,13 as compared to 240% and
uniform appearance without visible defects and a gel fraction 650%, which were obtained for 2k and 6k hydrogels in this study.
(GF) above 90% (96 ꢃ 4%, 91 ꢃ 1%, and 90 ꢃ 1% for hydrogels This difference is to be expected due to the higher number of
of PEG-6k-G1-allyl, PEG-10k-G1-allyl, and PEG-20k-G1-allyl, functional groups of the tetra-functional PEG-G1-allyls herein
respectively), with the exception of hydrogels made from PEG- presented. For hydrogels with the same molecular weight of PEG,
2k-G1-allyl, where the GF was 79 ꢃ 9%. The lower GF for the more of the lower molecular weight crosslinker, TMP-(SH)3, had
PEG-2k-G1-allyl hydrogels indicated that the curing process was to be added to the tetra-functional PEGs in order to keep the allyl-
not as effective for crosslinking PEG-2k-G1-allyl as for the longer to-thiol ratio at 1 : 1 in the hydrogels. This thus resulted in a
PEGs. This could be due to the relatively lower initiator-to- tighter network compared to the di-functional PEGs, and thereby
monomer ratio, leaving more functional groups unreacted lower swelling ratios. Additionally, it can be seen in Fig. 5a that
during the curing process in comparison with the other hydrogels with shorter PEG chains reach their equilibrium
hydrogel systems.
swelling ratio faster than hydrogels with longer chain segments.
Fig. 3 Thermal properties of modified and unmodified PEG materials obtained from DSC. (a) Transition temperatures for melting and crys-
tallization. (b) Degree of crystallinity. (Calculated from the enthalpy of melting during the second heating.)
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 30118–30128 | 30123

View Article Online
RSC Advances
Paper
Fig. 4 RAMAN spectra of unwashed hydrogels after curing, showing no signs of residual allyl or thiol functional groups.
Fig. 5 (a) Swelling ratio as a function of swelling time for hydrogels swollen in deionized water at 23 ꢁC and (b) equilibrium swelling ratio as a
function of molecular weight of the PEG chain in the hydrogel. (c) Photograph of hydrogels swollen to equilibrium in DI water at 23 ꢁC.
30124 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Table 2 Swelling ratio (SR), and mechanical properties, for hydrogels of PEG-G1-allyl crosslinked with TMP-(SH)3
PEG MW (g molꢀ1
)
SR (%)
G0 (kPa)
G00 (kPa)
tan d
Ea (kPa)
2000
6000
10 000
20 000
241 ꢃ 5
649 ꢃ 17
730 ꢃ 28
1390 ꢃ 26
35.4 ꢃ 4.4
20.7 ꢃ 3.6
13.4 ꢃ 0.7
1.98 ꢃ 0.11
19.5 ꢃ 4.7
13.4 ꢃ 3.0
6.7 ꢃ 3.3
0.55 ꢃ 0.15
0.65 ꢃ 0.18
0.50 ꢃ 0.25
0.19 ꢃ 0.09
106 ꢃ 13
62 ꢃ 11
40 ꢃ 2
0.37 ꢃ 0.18
5.9 ꢃ 0.3
a
Young's moduli calculated from the storage moduli (G0) obtained from rheology measurements using eqn (4).
Hydrogels made from PEG-2k-G1-allyl are close to equilibrium
swelling aer just three hours, whereas the swelling of hydrogels
made from PEG-6k-G1-allyl, PEG-10k-G1-allyl, and PEG-20k-G1-
allyl stagnates aer approximately 24 hours. It should also be
noted that the swelling kinetics are greatly affected by the drying
conditions. Freeze-dried hydrogel samples e.g. have a porous
structure where capillary forces give rise to an increased water
absorption rate compared to that of the vacuum dried samples
tested in this study, for which the main driving force of absorp-
tion is diffusion.
As can be seen from the swelling experiment, there are large
differences in water absorption for the different hydrogels. Fig. 5c
shows a photograph of the swollen hydrogels, where signicant
differences between the systems can be seen. Changing the
molecular weight of the PEG-G1-allyl that makes up the main part
of the hydrogel network can thereby be used as a means to easily
tune the swelling properties for a certain application.
Elastic properties
Mechanical analyses of the hydrogels were conducted using
tensile tests and rheology. Tensile tests revealed that the
hydrogels had elastic characteristics with strains at break
around 23–30%, where hydrogels based on PEG-2k-G1-allyl
displayed the highest strain at break with 30 ꢃ 4%. Due to large
standard deviations, there were no signicant differences in the
measured strains at break for hydrogels based on PEG-2k-G1-
allyl. PEG-6k-G1-allyl, and PEG-10k-G1-allyl. Hydrogels based on
PEG-10k-G1-allyl did however display a measurably lower stress
at break and tensile toughness than hydrogels based on PEG-2k-
G1-allyl and PEG-6k-G1-allyl. Hydrogels based on PEG-20k-G1-
allyl could unfortunately not be evaluated using tensile tests,
since the samples broke under their own weight during
clamping in the machine due to the high water content. Please
see Table S2 in the ESI for further details.†
Due to large variations in the data obtained during tensile
tests, rheology was found to be a more suitable method for
analysing the mechanical properties of the hydrogels. Rheology
measurements were conducted on fully swollen circular
hydrogel samples using parallel-plate geometry. The storage
modulus (G0) and the loss modulus (G00) were determined using
a dynamic time sweep test at 5% strain with a frequency of 10
rad sꢀ1. For each hydrogel system, a minimum of 5 replicas
were used.
The results from rheology shown in Fig. 6 reveal that less
dense hydrogels are more ductile than hydrogels with shorter
chain segments between crosslinks, which is in accordance with
values that have been previously reported.12,18 There is also a
tendency towards larger differences between storage and loss
moduli for hydrogels that are more densely crosslinked. This
implies that the viscous part of the material properties increases
as the hydrogels become less crosslinked and the water content
increases.
For all tested hydrogels, G0 was found to be higher than G00.
Therefore, G0 was estimated to be equal to the shear modulus
(G) which could then be used to calculate the Young's moduli
(E) for the hydrogels according to rubber elasticity theory using
eqn (4):
The obtained equilibrium swelling was moreover used to
ꢀ
calculate the average molecular weight between crosslinks (Mc)
ꢀ
for the synthesized hydrogels. Calculations of Mc were made
using the Peppas–Merrill model,22 which is derived from the
equation originally introduced by Flory,23 according to eqn (3):
"
#
ꢀ
ꢁ
ꢀ
ꢁ
1=3
1
2
n2;s
n2;r
n2;s
n2;r
rpV1
ꢀ
Mc ¼ ꢀ
(3)
2
lnð1 ꢀ n2;sÞ þ n2;s þ cðn2;s
Þ
where rp is the polymer density (1.125 g cmꢀ3 for PEG24), V1 is
the molar volume of the solvent (18 cm3 molꢀ1 for water), n2,s is
the volume fraction of polymer in the swollen gel, n2,r is the
volume fraction of polymer in the relaxed gel during gel
formation, and c is the Flory polymer–solvent interaction
parameter (0.426 for PEG–water25) For further details, please see
the ESI.†
ꢀ
The obtained values of Mc were 333, 2520, 3640, and 10 830 g
molꢀ1 for hydrogels based on PEG-2k-G1-allyl, PEG-6k-G1-allyl,
PEG-10k-G1-allyl, and PEG-20k-G1-allyl, respectively. The
ꢀ
obtained Mc was found to be much lower than expected for all
ꢀ
hydrogel systems. One explanation for this can be that Mc is
measured between kinetically active crosslinks and not only
between covalent crosslinking points. In all probability, the
relatively high solid content during gel formation in this study
(50 wt%) meant that the polymer chains were intertwined in the
solution during crosslinking, giving rise to physical entangle-
ments in the nal network structure.26–28 These entanglements
E ¼ 2G(1 + n)
(4)
where E is the Young's modulus (Pa), G is the shear modulus
(Pa) and n is Poisson's ratio. The calculations were made
assuming a Poisson's ratio of 0.5. The calculated Young's
moduli, which can be found in Table 2, range from 6 to 106 kPa,
ꢀ
can act as kinetically active crosslinks, and as a result, Mc values
will be lower than expected.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 30118–30128 | 30125

View Article Online
RSC Advances
Paper
Additionally, the loss factor, tan d, was signicantly lower for
PEG-20k-G1-allyl hydrogels than for the other hydrogels, indi-
cating that the viscous characteristics of the 20k hydrogels were
more pronounced than for the other hydrogel systems. This is
probably an effect of the large quantities of water in the PEG-
20k-G1 hydrogels.
Hydrolytic degradation
Hydrolytic degradation of the hydrogels was studied in buffer
solutions at three different pH values: pH 4, pH 7, and pH 10.
Visual inspection was used to assess the degradation and the
time from immersion until complete dissolution of the hydro-
gels was noted. Undissolved samples were continuously inves-
tigated throughout the study using tweezers in order to detect
large changes in mechanical stability. No large changes could
be discerned for the undissolved samples.
Fig. 6 Storage and loss moduli of formed hydrogels acquired from
rheology (time sweep measurements with 5% strain at 10 rad sꢀ1). All
measurements were conducted on hydrogels swollen to equilibrium.
It is known that the hydrolysis of ester bonds can be base
catalysed31 and it was therefore hypothesized that degradation
would be faster at pH 10 than at pH 4 and 7. Visible signs of
depending on the chain length of PEG-G1-allyl. These values are degradation could not be detected in buffer solutions of pH 4
in the same range as several so body tissues29 and also in the and pH 7 during the three months the degradation study was
same range as earlier published results for similar carried out. However, signs of degradation could be seen
hydrogels.18,30
already aer 48 h in buffer solution with pH 10. As shown in
Fig. 7 (a) Degradation time (days) for hydrogels hydrolysed in pH 10 buffer solution, time until complete dissolution into the medium assessed via
visual inspection. (b) MALDI-ToF MS of freeze-dried samples of PEG-6k-G1-allyl and the PEG-6k-hydrogel after degradation in pH 10 buffer
solution for three months. On top is a mass spectrum of unmodified PEG-6k-OH for comparison. (c) Potential degradation products after
hydrolysis of the ester bonds in the hydrogels, PEG-OH, TMP 3, and compound 4 (9-methyl-7,11-dioxa-3,15-dithiaheptadecane-1,9,17-
tricarboxylic acid). Degradation products from uncured thiol compounds that might be present in the hydrogels have not been taken into
account.
30126 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 7a, PEG-20k-hydrogels were completely dissolved aer 3 functional rst-generation PEG hydrogels. The key to control-
days, whereas PEG-6k and PEG-10k hydrogels were dissolved ling the properties of hydrogels lies in controlling the network
aer 8 and 10 days, respectively, while PEG-2k hydrogels were structure. Understanding the structure–property relationships
stable for 5 weeks. A previous study regarding the hydrolytic and being able to control the nal properties of the materials
degradation of bis-MPA based dendrimers conrms that are essential for the development of hydrogels for applications
hydrolytic degradation is faster at alkaline pH.32 The study in e.g. biomedicine.
showed degradation of the ester bonds in a generation four
The anhydride chemistry used in this work proved to be a
dendrimer already aer six hours in aqueous solution with pH facile approach for the functionalization of PEG materials, as
7.5 at 37 ꢁC. In the same study, it was also shown that the shown by NMR, SEC and MALDI-ToF MS. Characterization of
hydrolysis rate was slower at lower temperatures and that the the thermal properties of synthesized PEG derivatives showed
bis-MPA based dendrimer was stable for three weeks when that replacing the hydroxyl end groups with G1-allyl reduced the
subjected to a solution with pH 7.5 at 8 ꢁC. With those results in crystallinity of the materials and also lowered the transition
mind, it was interesting to notice that there were no visible temperatures for both crystallization and melting. The differ-
signs of degradation of the hydrogels synthesized in this study ences between PEG-OH and PEG-G1-allyl in terms of thermal
in buffer solutions of pH 7 at 23 ꢁC aer three months. An properties were additionally more pronounced at lower molec-
explanation for this can be that the ester bonds are more ular weights.
shielded in the crosslinked hydrogels than they are in a
dendrimer.
Hydrogels were easily formed using thiol–ene coupling
chemistry (TEC). TEC proved to be a robust strategy for the
The degradation study clearly showed that the degradation formation of well-dened crosslinked hydrogels with tunable
rate in aqueous buffer at pH 10 depends of the chain length of properties. Networks formed from PEG-G1-allyl with higher
the PEG. A reason for this can be that hydrogels made from molecular weights had a higher swelling degree, were more
longer PEG chains also exhibit a higher swelling degree, thereby ductile and more susceptible to hydrolysis than networks
exposing the ester bonds inside the structure to the alkaline formed from PEG-G1-allyl with lower molecular weights. By
solution. The hydrogels with longer chains between crosslinks changing the molecular weight, hydrogels were produced with
also contain a much lower relative amount of crosslinks in the swelling ratios ranging from 240 to 1400% and Young's moduli
structure, and therefore degradation of the ester bonds results ranging from 6 to 106 kPa. The molecular weight of the PEG
in degradation of the crosslinking points faster for these gels chain was also found to greatly affect the degradation time in
and thereby a more rapid dissolution. The degradation study pH 10 buffer, where the time until complete dissolution into the
was terminated aer 3 months, aer which the pH 10 buffer medium ranged from 2 days to 5 weeks.
solutions containing dissolved hydrogel samples were freeze-
In summary, it has herein been shown that TEC is a powerful
dried and analysed using MALDI-ToF MS. Analysis of the crosslinking strategy for the synthesis of well-dened network
acquired spectra of the degraded 6k hydrogel in Fig. 7b showed structures and that the properties of the nal hydrogels can be
that the molecular weight of the residues of the degraded easily tuned by varying the network constituents.
hydrogel was in accordance with that of unmodied PEG-6k.
This supports the hypothesis that degradation occurred as a
result of hydrolysis of the ester bonds in the structure, as this
Acknowledgements
would mean cleavage of the G1-allyl end groups. The freeze- Wilhelm Beckers Jubileumsfond is acknowledged for nancial
dried samples were furthermore analysed with NMR spectros- support.
copy in the attempt of identifying formed degradation products.
1H and 13C NMR spectra (Fig. S1–S4 in the ESI†) indicated both
trimethylol propane (TMP) 3 and compound 4 as products from
References
the degradation, as suggested in Fig. 7c. Peaks characteristic to
both 3 and 4 are present in the spectra whereas some peaks
characteristic for the starting materials, TMP-(SH)3 and PEG-G1-
allyl, cannot be detected. This further supports the theory that
the ester bonds in the hydrogels had been hydrolysed. The
resolution of the spectra and the presence of many different
compounds in the samples made it difficult to draw further
conclusions about the degradation products. NMR spectra for
PEG-6k-G1-allyl, TMP-(SH3), TMP, and Na2CO3, as well as esti-
mated NMR spectra for compound 4, can be found in the ESI,†
along with the NMR spectra for the degraded samples.
1 N. A. Peppas, J. Z. Hilt, A. Khademhosseini and R. Langer,
Adv. Mater., 2006, 18, 1345–1360.
2 N. A. Peppas, K. B. Keys, M. Torres-Lugo and A. M. Lowman,
J. Controlled Release, 1999, 62, 81–87.
3 B. V. Slaughter, S. S. Khurshid, O. Z. Fisher,
A. Khademhosseini and N. A. Peppas, Adv Mater, 2009, 21,
3307–3329.
4 C. C. Lin and K. S. Anseth, Pharm. Res., 2009, 26, 631–643.
5 N. A. Peppas, Y. Huang, M. Torres-Lugo, J. H. Ward and
J. Zhang, Annu. Rev. Biomed. Eng., 2000, 2, 9–29.
6 J. A. Burdick and K. S. Anseth, Biomaterials, 2002, 23, 4315–
4323.
7 A. M. Kloxin, A. M. Kasko, C. N. Salinas and K. S. Anseth,
Science, 2009, 324, 59–63.
Conclusions
The purpose of this study was to gain a better understanding of
the structure-to-property relationships of well-dened allyl-
8 C.-C. Lin, A. Raza and H. Shih, Biomaterials, 2011, 32, 9685–
9695.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 30118–30128 | 30127

View Article Online
RSC Advances
Paper
9 M. Malkoch, R. Vestberg, N. Gupta, L. Mespouille, P. Dubois, 22 N. A. Peppas and E. W. Merrill, J. Appl. Polym. Sci., 1977, 21,
A. F. Mason, J. L. Hedrick, Q. Liao, C. W. Frank, K. Kingsbury
and C. J. Hawker, Chem. Commun., 2006, 2774–2776.
10 W. H. Binder and R. Sachsenhofer, Macromol. Rapid
Commun., 2008, 29, 952–981.
11 A. B. Lowe, Polym. Chem., 2010, 1, 17–36.
12 T. Yang, H. Long, M. Malkoch, E. K. Gamstedt, L. Berglund
and A. Hult, J. Polym. Sci., Part A: Polym. Chem., 2011, 49,
4044–4054.
1763–1770.
23 P. J. Flory, Principles of Polymer Chemistry, Cornell University
Press, Ithaca, NY, USA, 1953.
24 D. W. van Krevelen and K. te Nijenhuis, Properties of
Polymers: Their Correlation with Chemical Structure; their
Numerical Estimation and Prediction from Additive Group
Contributions, Elsevier Science Publishers B.V., Amsterdam,
Netherlands, 3rd edn, 1990.
13 T. Yang, M. Malkoch and A. Hult, J. Polym. Sci., Part A: Polym. 25 E. W. Merrill, K. A. Dennison and C. Sung, Biomaterials,
Chem., 2013, 51, 363–371. 1993, 14, 1117–1126.
14 T. Yang, M. Malkoch and A. Hult, J. Polym. Sci., Part A: Polym. 26 N. Boucard, C. Viton and A. Domard, Biomacromolecules,
Chem., 2013, 51, 1378–1386. 2005, 6, 3227–3237.
15 D. D. McKinnon, A. M. Kloxin and K. S. Anseth, Biomater. 27 L. G. Cima and S. T. Lopina, Macromolecules, 1995, 28, 6787–
Sci., 2013, 1, 460–469. 6794.
16 M. W. Tibbitt and K. S. Anseth, Biotechnol. Bioeng., 2009, 103, 28 L. Vervoort, I. Vinckier, P. Moldenaers, G. Van Den Mooter,
655–663. P. Augustijns and R. Kinget, J. Pharm. Sci., 1999, 88, 209–214.
17 M. I. Montanez, L. M. Campos, P. Antoni, Y. Hed, 29 C. T. McKee, J. A. Last, P. Russell and C. J. Murphy, Tissue
M. V. Walter, B. T. Krull, A. Khan, A. Hult, C. J. Hawker
and M. Malkoch, Macromolecules, 2010, 43, 6004–6013.
18 K. Oberg, Y. Hed, I. J. Rahmn, J. Kelly, P. Lowenhielm and
M. Malkoch, Chem. Commun., 2013, 49, 6938–6940.
Eng., Part B, 2011, 17, 155–164.
30 S. Mongkhontreerat, K. Oberg, L. Erixon, P. Lowenhielm,
A. Hult and M. Malkoch, J. Mater. Chem. A, 2013, 1, 13732–
13737.
´
19 K. Olofsson, O. C. J. Andren and M. Malkoch, J. Appl. Polym. 31 R. P. Schwarzenbach, P. M. Gschwend and D. M. Imboden,
Sci., 2014, 131, 39876.
20 H. Ihre, O. L. P. De Jesus and J. M. J. Frechet, J. Am. Chem.
Soc., 2001, 123, 5908–5917.
21 G. Mun, I. Suleimenov, K. Park and H. Omidian, in
Biomedical Applications of Hydrogels Handbook, Springer,
New York, 2010, pp. 375–391.
in Environmental Organic Chemistry, John Wiley & Sons,
Inc., 2nd edn, 2003, pp. 513–535.
32 N. Feliu, M. V. Walter, M. I. Montanez, A. Kunzmann,
A. Hult, A. Nystrom, M. Malkoch and B. Fadeel,
Biomaterials, 2012, 33, 1970–1981.
30128 | RSC Adv., 2014, 4, 30118–30128
This journal is © The Royal Society of Chemistry 2014