Synthesis and application of new photocrosslinkers for poly(ethylene glycol)

Article abstract of DOI:10.1071/CH11485

Photocrosslinking of polyethylene glycol (PEG) using exogenous agents is a convenient way to produce branched PEG from commercial sources thus avoiding the tricky synthesis of new reactive and functional polymers. In this study, we synthesized two series of new photocrosslinkers, i.e. bis-fluorophenyl azide and bis-trifluoromethyl diazirine, which under soft UV-irradiation produce reactive species (i.e. nitrene and carbene respectively) that insert into the CH bond of the polymer backbone, building new bridges between macromolecular chains. These photocrosslinkers are different in terms of behaviour under irradiation and affinity for the target substrate (i.e. PEG). Thus, practical conditions for photocrosslinking of a 10-kDa PEG were studied and followed by NMR and size-exclusion chromatography. In particular, we investigated irradiation in bulk or in solvent, at different irradiation times, with several concentrations of PEG and photolinkers. Finally, we were able to design a procedure to obtain soluble crosslinked PEGs of 300kDa.

Full text of DOI:10.1071/CH11485

CSIRO PUBLISHING
Aust. J. Chem. 2012, 65, 193–201
Full Paper
http://dx.doi.org/10.1071/CH11485
Synthesis and Application of New Photocrosslinkers
for Poly(ethylene glycol)
A
A
A
Hakim Mehenni, Vincent Pourcelle, Jean-Franc¸ois Gohy,
A B
,
and Jacqueline Marchand-Brynaert
A
Institut de la Matie`re Condense´e et des Nanosciences (IMCN), Universite´ catholique
de Louvain, Baˆtiment Lavoisier, Place Louis Pasteur L4.01.02,
B-1348 Louvain-la-Neuve, Belgium.
B
Corresponding author. Email: jacqueline.marchand@uclouvain.be
Photocrosslinking of polyethylene glycol (PEG) using exogenous agents is a convenient way to produce branched PEG
from commercial sources thus avoiding the tricky synthesis of new reactive and functional polymers. In this study, we
synthesized two series of new photocrosslinkers, i.e. bis-fluorophenyl azide and bis-trifluoromethyl diazirine, which under
soft UV-irradiation produce reactive species (i.e. nitrene and carbene respectively) that insert into the C–H bond of the
polymer backbone, building new bridges between macromolecular chains. These photocrosslinkers are different in terms
of behaviour under irradiation and affinity for the target substrate (i.e. PEG). Thus, practical conditions for photo-
crosslinking of a 10-kDa PEG were studied and followed by NMR and size-exclusion chromatography. In particular, we
investigated irradiation in bulk or in solvent, at different irradiation times, with several concentrations of PEG and
photolinkers. Finally, we were able to design a procedure to obtain soluble crosslinked PEGs of 300 kDa.
Manuscript received: 16 December 2011.
Manuscript accepted: 19 January 2012.
Introduction
Therefore, the direct crosslinking of linear PEGs using an
exogenous agent that avoids specialist handling could be an
attractive way. Moreover, when a commercially available
inexpensive PEG can be used as starting material instead of
a,v-heterotelechelic PEGs, tricky syntheses are avoided.
Photolinkers such as arylazides and diazirines, generating
in situ nitrenes and carbenes respectively, are well-known
reagents[9] for photoaffinity labelling (PAL) of proteins[10–12]
and for the covalent surface modification of (bio)materials.[13–15]
Nevertheless, few examples of photoreactions performed in
bulk have been described[16] and only one publication reports
the use of a bis-diazirine reagent as a photocrosslinker for
PEG.[17]
Polyethylene glycol (PEG) is a very popular polymer that has
found numerous applications in e.g. drug-delivery systems[1]
and hydrogels for tissue engineering,[2] thanks to its biocom-
patibility, chemical inertness and solubility in both organic and
aqueous media. Branched and hyperbranched PEGs are desir-
able in particular cases for reducing the degree of crystallization
of the polymer, increasing its loading capacity, and making
copolymers, hydrogels, etc.[3] Among the strategies developed
to produce (hyper)branched PEGs, most are multistep proce-
dures. For instance, crosslinking methods usually require first
the synthesis of original new macromers such as multi-arm or
star-shaped PEGs with reactive chain ends,[4] PEG acrylate[5] or
a,v-heterotelechelic PEG[6,7] to mention but a few, and second,
the conception of new and original reactants or crosslinker
agents that will react with these macromers.[8] For instance, in a
very recent publication, Phelps and coworkers[4] produced, with
good control of crosslinking, a PEG hydrogel with ligands
incorporated. For that purpose, they synthesized four-arm PEGs
with maleimide reactive endings that were crosslinked using a
dithiol peptide.
In this paper, we describe the synthesis of original bis-
arylazides and bis-diazirines as potential crosslinking agents
of polymers under mild conditions. Their reaction with PEG was
studied in bulk and in solution, with the aim of forming defined,
soluble, branched PEGs under light activation.
Results and Discussion
Synthesis of the Photocrosslinking Agents
However, few strategies offer a simple and direct pathway.
Wilms and coworkers[3] present an elegant direct synthesis by
random copolymerization of ethylene oxide and glycidol to
produce, in a controlled manner, hyperbranched PEG-co-
polyglycerol. However, this methodology needs specific
equipment for the manipulation of hazardous gases (i.e. ethylene
oxide).
Photoreactive compounds 1[18] and 2[19] are well-known
photolinkers used in several applications ranging from biology
to materials science. These two heterobifunctional molecules
feature, on one side, an N-hydroxysuccinimidyl (NHS) ester that
is prone to react with amino compounds, and on the other side,
a photoreactive function, i.e. a perfluorophenyl azide (PFPA) for
Journal compilation Ó CSIRO 2012
www.publish.csiro.au/journals/ajc

194
H. Mehenni et al.
F
F
F
F
O
O
O
O
N
N
O
O
O
O
N3
N
F3C
N
2
1
3a, n ϭ 7
3b, n ϭ 12
O
4a, n ϭ 3
4b, n ϭ 5
C
H2
NH2
H2N
H2N
NH2
n
n
F
F
O
O
F
F
F
N
H
C
H2
N
H
5a, n ϭ 7
5b, n ϭ 12
n
N3
N3
F
F
F
F
O
O
F
F
O
F
N
H
N
H
n
6a, n ϭ 3
6b, n ϭ 5
N3
F
F
N3
O
O
F
F
N
H
C
H2
N
H
N
N
7a, n ϭ 7
7b, n ϭ 12
n
N
N
CF3
CF3
O
O
O
N
H
N
H
n
N
N
8a, n ϭ 3
8b, n ϭ 5
N
N
CF3
CF3
Fig. 1. Perfluorophenyl azide (PFPA) and trifluoromethyl phenyl diazirine (TPD) photocrosslinkers.
1 and a trifluoromethyl phenyl diazirine (TPD) for 2. In this
contribution, the synthesis of novel bis-PFPA and bis-TPD
photocrosslinkers is realized by reacting the NHS ester of 1 and
2 with a series of diamino spacers (3a, 3b, 4a and 4b) to obtain
eight compounds (Fig. 1) differing by the length and nature
(lipophilic or hydrophilic) of the spacer.
PFPA leads to a reactive nitrene under irradiation at wave-
lengths of ,250 nm. PFPA is preferred to simple aryl azides[20]
because it undergoes ‘singlet-like insertion’ into C–H bonds
mainly (i.e. in a concerted mechanism), instead of the stepwise
reactions expected from triplet nitrene produced by irradiation
of phenyl azide.[21] PFPA derivatives are thus very attractive
photolinkers, easily prepared, stable under storage and giving,
under light excitation, very fast and efficient reactions. One
major drawback is their excitation wavelength, which could
damage some UV-sensitive substrates.
Under excitation at ,350 nm, TPD gives a very reactive
singlet carbene that is able to insert into any X–H bonds in a
concerted mechanism.[9] Depending on the conditions (mostly
the concentration of the photolinker), the photodecomposition
of TPD can follow two pathways: (i) the direct elimination of
dinitrogen leading to a singlet carbene, or (ii) the rearrangement
into a diazo intermediate that either decomposes spontaneously
into the desired carbene, or persists until it is subjected to
supplementary irradiation at a wavelength of 300 nm.[11,12] This
phenomenon may sometimes preclude the use of TPD despite all
its other advantages (i.e. high singlet reactivity, non-damaging
irradiation wavelength to various substrates).
Under light excitation, the bis-PFPA and bis-TPD linkers
designed herein are able to give two simultaneous insertions into
C–H or X–H bonds of molecules present in their close vicinity.
Consequently, when mixed with polymers, they can create

Photolinkers for Crosslinking of PEG
195
CH2
O
CH2
O
O
F
O
F
F
F
F
F
H C
CH
F
2 F
2
O
F
F
F
F
F
O
F
N
H
X
N
H
F
F
O
F
F
F
O
F
HC
CH
F
F
F
F
O
hν ϭ 254 nm
ϪN2
N
H
O
N
H
X
N
H
N
H
N
H
X
N
H
CH2
CH2
N3
N3
N
N
MeO
H
n
O
O
O
O
O
CH2
O
CH2
O
hν ϭ 350 nm
N
H
X N
H
O
O
N
H
X
N
H
CH2
N
N
H2C
HC
ϪN2
N
H
X N
H
N
N
CH
CF3
CF3
CF3
CF3
O
O
CH
HC
CF3
CF3
CH2
CH2
X ϭ CH2-CH2- or CH2-CH2-O-
Fig. 2. Photocrosslinking of polyethylene glycol (PEG) induced by bis-PFPA (perfluorophenyl azide) and bis-TPD (trifluoromethyl phenyl diazirine).
(a)
junctions between two different macromolecular chains, as
depicted in Fig. 2 in the case of PEG. In this respect, they may
provide a convenient way to obtain crosslinked polymers.
The use of different kinds of spacers to separate the two
photoreactive centres gives us the opportunity to control the
behaviour of the photolinker in terms of solubility and affinity
with the substrate. Indeed, because the species formed after
irradiation (i.e. nitrenes and carbenes) are highly reactive, they
would react in an unspecific way with the closest chemical bond;
this means that, to be efficient, the photoreactive group has to be
located just next to the targeted bond of the substrate. For that
purpose, we used bis-amino aliphatic or bis-amino oligoether
spacers in order to vary the hydrophilic–lipophilic balance of
our compounds.
1.2
1
0 s
10 s
20 s
30 s
40 s
50 s
60 s
0.8
0.6
0.4
0.2
0
200
250
300
350
400
450
λ [nm]
(b) 0.2
0
Photochemical Properties
The PFPA compounds 5 and 6 exhibit the characteristic phenyl
azide UV absorption band at ,255 nm. Under irradiation in
acetonitrile solution, a decrease of the absorption at 255 nm was
observed (Fig. 3a) following apparent first-order kinetics, as
shown in Fig. 3b for 5a. From the slope of the curve, the rate
constant could be calculated. We found a half-life of 18 s, and an
irradiation time for total disappearance of the azide peak of
196 s. The decrease of the azide peak can also be observed by
infrared spectroscopy thanks to the characteristic stretching
band at ,2100 cmꢀ1 (Fig. 4a). The ratio of the absorbance
corresponding to the N3 stretching band (2125 cmꢀ1) to the
absorbance of the CO amide band (1654 cmꢀ1) of compound 5a
gave a curve (Fig. 4b) comparable with the one observed in the
UV spectrum. A total disappearance of the azide band was
observed after 180 s, which is in agreement with the result cal-
culated from UV measurements (Fig. 3).
Ϫ0.2
Ϫ0.4
Ϫ0.6
Ϫ0.8
Ϫ1
0
5
10
15
20
25
30
Irradiation time [s]
Fig. 3. (a) UV-vis spectra of 5a (0.17 mM in CH3CN) at different times of
irradiation. (b) Logarithmic plot of the decrease of the absorption maximum
at 255 nm of 5a (0.17 mM in CH3CN) consistent with first-order kinetics of
the photolysis.
The TPD compounds 7 and 8 show the characteristic peak of
TPD absorption at ,355 nm (Fig. 5a). The photolysis of these
compounds at 350 nm led to a decrease of the diazirine
absorption peak with apparent first-order kinetics and a half-
life of 1.1 min (Fig. 5b), in good agreement with published
results.[19,22] These observations prove that the introduction of
bis-amino spacers (3a, 3b, 4a and 4b) does not affect the
behaviour of PFPA and TPD moieties under UV irradiation.
architecture of the final polymers. In our case, the parameters
that might be interesting to investigate are: polymer : crosslinker
ratio, irradiation time, neat solid or solvent media, and affinity
between the crosslinker and the substrate. As very reactive
species are created during the irradiation process, it is crucial to
avoid side reactions as much as possible. So, in a first approach,
it seems obvious that the reaction should be carried out in neat
mixtures, as pure as possible to avoid insertion of the nitrenes
into undesired molecules (i.e. impurities, solvent molecules).
However, the reaction yields are expected to be low because, to
be efficient, the double insertion process requires two reactive
species in close proximity to C–H bonds of the substrate at the
Photocrosslinking of PEG
Post-polymerization crosslinking with the help of an additive
has the advantage, in comparison with other strategies,[23] of
offering several parameters that can be used to control the

196
H. Mehenni et al.
(a)
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
(a)
0 min
0.5 min
1 min
1.5 min
2 min
2.5 min
3 min
3.5 min
4.5 min
5.5 min
6.5 min
300
320
340
360
λ [nm]
380
400
420
440
3650
3150
2650
2150
1650
1150
650
Wavenumber [cmϪ1
]
(b)
0
(b) 0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Ϫ0.5
Ϫ1
Ϫ1.5
Ϫ2
Ϫ2.5
Ϫ3
0
1
2
3
4
Irradiation time [min]
0
50
100
Irradiation time [s]
150
200
Fig. 5. (a) UV-vis spectra of 7a (0.2 mM in CH3CN) at different times of
irradiation. (b) Logarithmic plot of the decrease of the absorption maximum
at 355 nm of 7a (0.2 mM in CH3CN) consistent with first-order kinetics of
the photolysis.
Fig. 4. (a) IR spectra of 5a at 0 s (upper line) and 180 s (lower line) of
irradiation (3 mM in CH3CN). (b) Absorbance ratio of the azide band to the
amide band of 5a (3 mM in CH3CN) at different irradiation times at 255 nm.
same moment; otherwise they miss their target, leading to
several uncontrolled side reactions (i.e. rearrangement, self-
coupling of reactive species, hydrogen abstractiony).
In order to evaluate the efficiency of our photocrosslinkers,
they were tested on PEG either in bulk or in solution.
Photocrosslinking in Solution
PFPA molecules, initially developed for PAL applications by
biologists, are currently employed in aqueous media. So we
tested the photocrosslinking reaction in an aqueous solution of
PEG and the crosslinker 6a. After 90 min of irradiation, the
formation of a precipitate was observed in the solution. This
indicated the formation of highly crosslinked PEG particles as
this precipitate could not be resolubilized in aqueous or in
organic solvents. Therefore, it was not possible to characterize
this precipitate in more detail by solution techniques. Indeed,
Blencowe et al.[17] also obtained insoluble crosslinked PEGs by
using bis-diazirines as photocrosslinkers. The swelling proper-
ties of their crosslinked PEGs were measured in a wide range of
solvents including water, dimethylformamide and tetrahydro-
furan. They demonstrated that the extent of swelling decreased
as a function of the amount of photocrosslinkers used as a result
of the increased crosslinking density. Unfortunately, they were
neither able to quantify the crosslinking density nor the mean
distance between crosslinking points. In this study, we
employed higher PEG photocrosslinker ratios and longer PEG
chain length than Blencowe et al.; consequently, we decided to
focus on the soluble part of the crosslinked PEGs that could be
analysed by SEC and so, in the following discussion, only sol-
uble fractions are considered. As can be seen from the 19F NMR
spectra (Fig. 8), several compounds were obtained after 90 min
of irradiation. In SEC (Fig. 9), a shoulder can be seen next to the
peak of the starting PEG and peaks of higher molar masses
(100 to 1000 kDa) appeared, indicating a successful crosslinking
of PEG chains. Unfortunately, we did not observe a sharp peak
Photocrosslinking in Bulk
We first attempted the photocrosslinking in bulk of neat
mixtures of PEG (10 kDa) and crosslinker 6a that were obtained
after solvent casting and careful removal of solvent traces under
high vacuum. As can be seen on Fig. 6 from 19F NMR, the
conversion of the PFPA groups was complete after 25 min of
irradiation and the two new peaks that appeared at ꢀ143 and
ꢀ161 ppm could be attributed to covalently linked products. The
peak at ꢀ76 ppm originates from the CF3COOH used as internal
standard. After irradiation, the crude mixture (soluble in the
analysis solvent) was analysed by size-exclusion chromato-
graphy (SEC), revealing a peak at ,10 kDa (Fig. 7) that
corresponds to the initial PEG sample. For the sake of com-
parison, pure PEG samples were also irradiated under the same
conditions (i.e. 250 nm, 25 min) and no other peaks than the one
at 10 kDa appeared in the SEC elution curve, meaning that no
chain scission occurred. One explanation of this unsuccessful
photocrosslinking could be that, during solvent evaporation,
PEG chains crystallize, leading to the expulsion of ‘impurities’
(i.e. the photocrosslinkers) from the spherulites to form segre-
gated domains. So, under irradiation, the reactive species would
tend to react among themselves in those domains instead of
reacting with PEG chains.

Photolinkers for Crosslinking of PEG
197
(a)
(b)
(c)
Ϫ75 Ϫ80 Ϫ85 Ϫ90 Ϫ95 Ϫ100 Ϫ105 Ϫ110 Ϫ115 Ϫ120 Ϫ125 Ϫ130 Ϫ135 Ϫ140 Ϫ145 Ϫ150 Ϫ155 Ϫ160 Ϫ165
Chemical shift [ppm]
Fig. 6. 19F NMR spectra (287 MHz, CDCl3, 258C) of the bulk mixture of 10-kDa PEG (polyethylene glycol) with 6a (16 : 1,
6a : PEG) at different irradiation times (254 nm): (a) 0 min; (b) 15 min; (c) 25 min.
parameters. Several factors, i.e. irradiation time, ratio of PEG to
photocrosslinker, concentration of PEG, were looked at in term
of molar mass obtained in the soluble fraction.
(a)
(b)
From Fig. 10, it is obvious that for a given PEG concentration
and a given photocrosslinker ratio, the irradiation time has no
effect, as after 5 min of irradiation, no further changes are
observed in SEC chromatograms. From Fig. 11, it can be seen
that, for a given irradiation time and a given PEG concentration,
higher amounts of photocrosslinker lead to a decrease in the
molar masses. This can be explained by a loss of insertion yield
due the higher probability of reaction of reactive species
between themselves. Consequently, it seemed logical that, at a
low PEG : photocrosslinker ratio, the insertion yield could be
improved with more concentrated PEG solutions. Surprisingly,
the increase of PEG concentration (at given ratio of crosslinker
and irradiation time) led to a drop in the molar mass (Fig. 12).
The trends discussed here were similarly observed with the
other PFPA photocrosslinkers 5a, 5b and 6b. This exploration of
the different parameters that govern the photocrosslinking
finally enabled us to elaborate a reproducible protocol to obtain
soluble crosslinked PEG of ,300 kDa with PFPA photocros-
slinkers. The conditions were as follows: 10 min of irradiation at
254 nm, PEG solution of 2 mg mLꢀ1, photolinker : PEG molar
ratio of 3 : 2, solution in H2O/CH3CN (3 : 1 v/v).
25
30
35
40
Elution time [min]
Fig. 7. Size-exclusion chromatography (SEC) chromatograms of (a) 10-kDa
PEG (polyethylene glycol) and (b) the bulk mixture of 10-kDa PEG with 6a
(16 :1, 6a : PEG) irradiated (254nm) for 25 min.
The TPD crosslinkers 7–8 were tested under the same
conditions (i.e. in bulk mixtures and in aqueous solvent) but in
all cases, no photocrosslinked PEG product was detected but
only low molar masses species certainly due to the coupling of
photocrosslinkers between themselves. The TPD series is thus
not useful for making crosslinked PEGs.
with a clear distribution of molar masses, meaning that cross-
linking was happening in an uncontrolled way. Consequently, in
order to obtain a controlled and reproducible protocol, several
parameters needed to be investigated.
Effect of the Photografting Conditions of 6a
Conclusions
on the Molar Mass of the Soluble Crosslinked PEG
Two series of photocrosslinkers have been synthesized and
tested for their capacity to produce soluble crosslinked PEG
We hypothesized that controllable architecture of photo-
crosslinked PEG may be obtained with optimized experimental

198
H. Mehenni et al.
(a)
(b)
Ϫ75 Ϫ80 Ϫ85 Ϫ90 Ϫ95 Ϫ100 Ϫ105 Ϫ110 Ϫ115 Ϫ120 Ϫ125 Ϫ130 Ϫ135 Ϫ140 Ϫ145 Ϫ150 Ϫ155 Ϫ160 Ϫ165 Ϫ170
Chemical shift [ppm]
Fig. 8. 19F NMR spectra (287 MHz, CDCl3, 258C) of crosslinked 10-kDa PEG (polyethylene glycol) with 6a (16 : 1, 6a : PEG)
at different irradiation times (254 nm, H2O : CH3CN solution (3 : 1 v/v)): (a) 0 min, and (b) 90 min.
(a)
(b)
360
310
260
210
160
110
60
10
0
5
10
15
20
25
30
35
[min]
Fig. 10. Molecular mass of soluble crosslinked PEG (polyethylene glycol)
for different times of irradiation at 254 nm in H2O : CH3CN (3 : 1 v/v)
solution with a 6a : PEG molar ratio of 4 : 1 at a PEG concentration of
2 mg mLꢀ1
.
25
30
35
40
Elution time [min]
predictions, the reactivity of compounds 5–8 was examined by
trial and error. In our hands, the bis-diazirines 7–8 could not be
used to crosslink PEG; we surmise that self-condensation of the
photolinkers and side reactions are the major processes what-
ever the nature and length of the spacer connecting the two light-
sensitive endings. The bis-arylazides 5–6 showed globally the
same behaviour and enabled the photocrosslinking of PEG
in aqueous acetonitrile solution, but not in bulk owing again
to self-coupling of the photolinkers, as seen by other groups.[25]
Best results were found by using 6a (i.e. two 4-azido-2,3,5,6-
tetrafluorobenzoyl motifs connected with a tetraethyleneglycol
spacer via amide bonds), allowing the production of soluble
crosslinked PEG of ,300 kDa, together with highly branched
insoluble PEG that we were not able to characterize further.
Fig. 9. Size-exclusion chromatography (SEC) chromatograms of (a) 10-kDa
PEG (polyethylene glycol), and (b) the crosslinked 10-kDa PEGwith 6a (16: 1,
6a : PEG) obtained after 90 min of irradiation at 254 nm in H2O:CH3CN
(3 : 1 v/v) solution.
starting from a 10-kDa PEG. Considering the previous litera-
ture, the diazirine derivatives should be more efficient than the
aryl azides in insertion reactions into C–X bonds under light
activation.[9,14] However, at high concentration, diazirines can
lead to persistent diazo compounds,[11,12] thus reducing the
yields of insertion. Moreover, the behaviour of the reactive
intermediates (nitrenes, carbenes) changes when reactions are
performed in diluted or concentrated solution, in the bulk or at
the surface of a solid material.[14,16,24] In the absence of reliable

Photolinkers for Crosslinking of PEG
199
360
310
260
210
160
110
60
broad-band proton decoupling. Spectra were obtained in CDCl3,
D2O or (CD3)2CO at room temperature. Chemical shifts d are
reported in ppm and are calibrated on the solvent signal at 7.26,
4.79 or 2.05 ppm for CDCl3, D2O or (CD3)2CO respectively.
Coupling constants J are given in Hz. UV-visible spectra were
recorded on a UV-vis-NIR Varian–Cary spectrophotometer
(l given in nm). Gel permeation chromatography (GPC) was
carried at 458C at a flow rate of 1 mL minꢀ1 out on a system
˚
composed of two PSS Gram columns (100 and 1000 A) con-
nected to a Waters 410 differential refractometer and a Waters
UV detector, with dimethylformamide as the carrier solvent.
Polystyrene standards were used for calibration. Low-resolution
mass spectra (MS) were acquired using a Thermo Finnigan LCQ
spectrometer in positive-APCI mode (atmospheric pressure
chemical ionization), positive-CI mode (chemical ionization) or
positive-ESI mode (electrospray ionization). Melting points
were recorded with a calibrated Bu¨chi melting point B-540
apparatus. Flash chromatography was performed on Merck
60 (40–63 mm) silica gel. The reacting and rinsing solutions
containing the polymer samples were shaken (200 rpm) with an
Edmund Bu¨hler stirrer (model KL-2).
10
1.5:1
4:1
8:1
12:1
16:1
Molar ratios
Fig. 11. Molecular mass of soluble crosslinked PEG (polyethylene glycol)
for different 6a : PEG molar ratios after 10 min of irradiation at 254 nm in
H2O : CH3CN (3 : 1 v/v) solution at a PEG concentration of 2 mg mLꢀ1
.
360
310
260
210
160
Synthesis of Precursors
O-succinimidyl-4-azido-2,3,5,6-tetrafluorobenzoate 1 was syn-
thesized according to a literature protocol.[18] The synthesis of
O-succinimidyl-4-(1-azi-2,2,2-trifluoroethyl) benzoate 2 was
carried out in seven steps as described elsewhere (V. Pourcelle
et al., unpubl. data).
a,v-diaminotetraethylene glycol 4a and a,v-diamino-
hexaethylene glycol 4b were synthesized from a,v-dihydroxy-
tetraethylene glycol and a,v-dihydroxyhexaethylene glycol
respectively, according to procedures published elsewhere.[26]
110
60
10
0
5
10
[mg mLϪ1
15
]
20
25
Fig. 12. Molecular mass of soluble crosslinked PEG (polyethylene glycol)
for different PEG concentrations at a 6a : PEG molar ratio of 1.5 : 1 after
10 min of irradiation at 254 nm in H2O : CH3CN (3 : 1 v/v) solution.
General Procedures for the Synthesis
of Photocrosscoupling Agents 5a, 5b
Into a flame-dried round-bottomed flask and under dimmed
light was introduced a solution of 2 equiv. of tetra-
fluoroarylazide 1 in dry acetonitrile (10 mL). Then, at room
temperature under stirring and an argon atmosphere, a solution
containing 1 equiv. of diamine spacer arms 3a or 3b in dry
acetonitrile (10 mL) was added dropwise. After 2 to 4 h stirring,
the mixture was filtered and the white precipitate obtained was
dissolved in distilled water and extracted with chloroform. The
organic layer was washed with water, dried over MgSO4 and
evaporated under vacuum. Compounds 5a and 5b were purified
by crystallization in a minimum of a mixture of CHCl3/hexane.
It is not obvious why in our case bis-arylazides were more
efficient than bis-diazirines as the answer to this question would
necessitate a dedicated study.[14] One explanation could be that
nitrenes are more persistent than carbenes.[9,16] As we need a
‘double insertion process’, requiring two reactive species in
close proximity to two C–H bonds at the same instant, nitrenes
might be superior because they would have more time to
undergo this double process, whereas in the same time,
carbenes would lead to several uncontrolled side reactions
(i.e. rearrangement, self-coupling of reactive species, hydrogen
abstractiony). Another explanation could be the above-
mentioned formation of persistent diazo species when diazirines
are reacted at high concentration in solution.[11]
Synthesis of 4-Azido-N-[7-[(4-azido-2,3,5,6-
tetrafluorobenzoyl)amino]heptyl]-2,3,5,6-
tetrafluorobenzamide 5a
The title compound (70 mg, 70 %) was obtained as a white
solid from 1 (118 mg, 0.36 mmol) and diamine spacer arm 3a
(23.1 mg, 0.18 mmol). mp 127.88C. nmax/cmꢀ1 3294, 2923,
Experimental
General
Solvents were of analytical grade and obtained from Aldrich–
Fluka, Acros or ROCC. 1,7-Diaminoheptane 3a, 1,12-
diaminododecane 3b, a,v-dihydroxytetraethylene glycol,
a,v-dihydroxyhexaethylene glycol and a-methoxy-v-hydroxy-
PEG (10000 g molꢀ1) were purchased from Aldrich. 1H
(300 MHz), 13C (75 MHz) and 19F (282 MHz) NMR spectra
were recorded on a Bruker Avance II 300 spectrometer. H
(500 MHz) and 13C (125 MHz) NMR spectra were recorded on
a Bruker AM-500 spectrometer. Spectra were obtained with
2854, 2125, 1654, 1554, 1485, 995, 724. lmax/nm (e/Mꢀ1 cmꢀ1
)
260 (67017). dH (300 MHz, CDCl3) 1.40 (m, 6H, –CH2–), 1.62
(m, 4H, NHCO–CH2–CH2–), 3.46 (dt, J 6.6, 4H, NHCO–CH2–),
6.13 (s, 2H, NHCO). dC (125 MHz, CDCl3) 158.1 (C¼O), 144.4
(ddd, 1JC–F 250, 2JC–F 14, 3JC–F ,5, CF), 140.8 (dd, 1JC–F 250,
2
2JC–F 14, CF), 122.1 (m, ArC–C¼O), 112 (t, JC–F 19, CN3),
1
40.2 (NHCO–CH2–), 29.3 (NHCO–CH2–CH2–), 28.3 (–CH2–),
26.5 (–CH2–). dF (282 MHz, CDCl3) ꢀ141.1 (dd, J 23, 12, 4F),
ꢀ150.6 (dd, J 23, 12, 4F). m/z (ESI) 587 (100 %, MNaþ).

200
H. Mehenni et al.
Synthesis of 3-Azido-N-[12-[(4-azido-2,3,5,6-
Synthesis of 4-[3-(Trifluoromethyl)diazirin-3-yl]-N-
tetrafluorobenzoyl)amino]dodecyl]-2,4,5,6-
tetrafluorobenzamide 5b
[2-[2-[2-[2-[2-[2-[4-[3-(trifluoromethyl)diazirin-3-yl]
benzoyl amino]ethoxy]ethoxy]ethoxy]ethoxy]ethoxy]
ethyl]benzamide 8b
The title compound (66 mg, 57 %) was obtained as a white
solid from 1 (120 mg, 0.36 mmol) and diamine spacer arm 3b
(36.3 mg, 0.18 mmol). mp 140.28C. nmax/cmꢀ1 3315, 2921,
The title compound (78 mg, 60 %) was obtained as yellowish
oil from 2 (120 mg, 0.37 mmol) and diamine spacer arm 4b
(51.4 mg, 0.18 mmol). nmax/cmꢀ1 3326, 2914, 2871, 1645, 1548,
1186, 1153. lmax/nm (e/Mꢀ1 cmꢀ1) 350 (277). dH (300 MHz,
CDCl3) 3.62–3.58 (m, 24H, –CH2CH2O–) 7.22 (d, J 8, 4H,
ArH), 7.31 (s, 2H, NHCO), 7.87 (d, J 8, 4H, ArH). dC (125 MHz,
CDCl3) 166.7 (C¼O), 136.1 (ArC–CNN), 132.4 (ArC–C¼O),
2852, 2123, 1652, 1544, 1475, 991. lmax/nm (e/Mꢀ1 cmꢀ1
)
260 (38725). dH (300 MHz, CDCl3) 1.39–1.28 (m, 16H,
–CH2–), 1.61 (m, 4H, NHCO–CH2–CH2–), 3.45 (dt, J 6.3, 7.2,
4H, NHCO–CH2–), 5.95 (s, 2H, NHCO). dC (125 MHz, C2D6O)
157.3 (C¼O), 143.4 (ddd, 1JC–F 246, 2JC–F 14, 3JC–F ,5, CF),
1
2
2
1
143.8 (dd, JC–F 248, JC–F 17, CF), 121.5 (t, JC–F 11, ArC–
128.1 (ArCH), 126.8 (ArCH), 122.3 (q, JC–F 273, CF3),
2
C¼O), 113.6 (t, JC–F 20, CN3), 40 (NHCO–CH2–), 29.76
70.9 (–CH2CH2O–), 70.5 (NHCO–CH2–CH2–), 70.1 (NHCO–
2
CH2–CH2–), 40.3 (NHCO–CH2–), 28.7 (q, JC–F 40, CN2).
(–CH2–), 29.74 (NHCO–CH2–CH2–), 29.4 (–CH2–), 26.9
(–CH2–). dF (282 MHz, CDCl3) ꢀ141.0 (4F, dd, J 23, 12),
ꢀ150.5 (4F, dd, J 20, 9). m/z (APCI) 635 (100 %, MHþ).
dF (282 MHz, CDCl3) ꢀ65 (s, CF3). m/z (APCI) 705 (30 %,
MHþ).
General Procedures for the Synthesis
of Photocrosscoupling Agents 6a, 6b, and 8b
General Procedures for the Synthesis
of Photocrosscoupling Agents 7a, 7b, and 8a
In a flame-dried round-bottomed flask and under dimmed light
was introduced a solution of 2 equiv. of tetrafluoroarylazide 1,
or trifluorophenyl diazirine 2, in dry acetonitrile (10 mL). Then,
at room temperature under stirring and an argon atmosphere,
a solution containing 1 equiv. of diamine spacer arms 4a or 4b in
dry acetonitrile (10 mL) was added dropwise. After 2 to 4 h
stirring, distilled water (20 mL) was added and the mixture was
extracted with CHCl3. The organic layer was washed with dis-
tilled water, dried over MgSO4 and evaporated under vacuum to
obtain compounds 6a, 6b and 8b with no further purification.
In a flame-dried round-bottomed flask, under dimmed light and
an argon atmosphere, 2 equiv. of trifluorophenyl diazirine 2,
1 equiv. of diamine spacer arms 3a, 3b or 4a and 6 equiv. of Et3N
were introduced in dry dichloromethane or dry DMF (4 mL).
After 20 h stirring at room temperature CHCl3 (50 mL) was
added. The organic layer was washed with water, dried over
MgSO4 and evaporated under vacuum. Compounds 7a, 7b and
8a were chromatographed on flash silica gel with hexane/ethyl
acetate (1 : 1, 2 : 1, 18 : 1 respectively).
Synthesis of 4-[3-(Trifluoromethyl)diazirin-3-yl]-
N-[7-[4-[3-(trifluoromethyl)diazirin-3-yl]benzoyl]
aminoheptyl]benzamide 7a
Synthesis of 4-Azido-N-[2-[2-[2-[2-[(4-azido-2,3,5,6-
tetrafluorobenzoyl)amino]ethoxy]ethoxy]ethoxy]ethyl]-
2,3,5,6-tetrafluorobenzamide 6a
The title compound (41 mg, 82 %) was obtained, after chro-
matography on silica gel with hexane/ethyl acetate (1 : 1)
followed by crystallization in a mixture of CHCl3/hexane, as a
white solid from 2 (59 mg, 0.18 mmol) and diamine spacer arm
3a (11.7 mg, 0.09 mmol). mp 1348C. nmax/cmꢀ1 3311, 2923,
2852, 1637, 1548, 1191, 1155. lmax/nm (e/Mꢀ1 cmꢀ1) 350
(506). dH (300 MHz, CDCl3) 1.38 (m, 6H, –CH2–), 1.60
(m, 4H, NHCO–CH2–CH2–), 3.43 (dt, J 7, 13, 4H, NHCO–
CH2–), 6.39 (m, 2H, NHCO), 7.39 (d, J 8, 4H, ArH), 8.03 (d, J 8,
4H, ArH). dC (125 MHz, CDCl3) 166.7 (C¼O), 136.2 (ArC–
CNN), 132.6 (ArC–C¼O), 127.7 (ArCH), 126.9 (ArCH), 122.3
The title compound (84 mg, 74 %) was obtained as a white
solid from 1 (120 mg, 0.36 mmol) and diamine spacer arm 4a
(34.7 mg, 0.18 mmol). mp 96.78C. nmax/cmꢀ1 3292, 2923, 2854,
2125, 1652, 1558, 1488, 1101, 993. lmax/nm (e/Mꢀ1 cmꢀ1) 260
(62170). dH (300 MHz, CDCl3) 3.64–3.62 (m, 16H,
–CH2CH2O–), 6.74 (s, 2H, –NHCO–). dC (125 MHz, CDCl3)
158.2 (C¼O), 145.4 (ddd, 1JC–F 250, 2JC–F 24, 3JC–F ,5, CF),
1
2
2
141.7 (dd, JC–F 237, JC–F 15, CF), 122.1 (t, JC–F 12, ArC–
2
C¼O), 111.9 (t, JC–F 18, CN3), 70.8 (–CH2–CH2–O–), 70.6
(–CH2–CH2 –O–), 69.7 (–CH2–CH2–O–), 40.4 (NHCO–CH2–).
dF (282 MHz, CDCl3) ꢀ141.08 (dd, J 20, 11, 4F), ꢀ150.73 (dd,
J 23, 11, 4F). m/z (APCI) 627 (100 %, MHþ).
1
(q, JC–F 273, CF3), 40.3 (NHCO–CH2–), 29.7 (NHCO–CH2–
2
CH2–), 28.8 (–CH2–), 28.7 (q, JC–F 40, CN2), 26.8 (–CH2–).
Synthesis of 4-Azido-N-[2-[2-[2-[2-[2-[2-[(4-azido-
2,3,5,6-tetrafluorobenzoyl)amino]ethoxy]ethoxy]
ethoxy]ethoxy]ethoxy]ethyl]-2,3,5,6-
tetrafluorobenzamide 6b
dF (282 MHz, CDCl3) ꢀ65 (s, CF3). m/z (APCI) 555
(100 %, MHþ).
Synthesis of 4-[3-(Trifluoromethyl)diazirin-3-yl]-
N-[12-[4-[3-(trifluoromethyl)diazirin-3-yl]benzoyl]
aminododecyl]benzamide 7b
The title compound (98 mg, 76 %) was obtained as a yellow-
ish oil from 1 (120 mg, 0.36 mmol) and diamine spacer arm 4b
(50.6 mg, 0.18 mmol). nmax/(cmꢀ1) 3299, 2931, 2871, 2125,
1652, 1550, 1488, 1107, 993. lmax/nm (e/Mꢀ1 cmꢀ1) 260
(1630). dH (300 MHz, CDCl3) 3.62–3.53 (m, 24H,
–CH2CH2O–), 7.1 (s, 2H, NHCO). dC (125 MHz, CDCl3)
158.1 (C¼O), 144.4 (ddd, 1JC–F 202, 2JC–F 12, 3JC–F ,5, CF),
The title compound (32 mg, 66 %) was obtained, after chro-
matography on silica gel with hexane/ethyl acetate (2 : 1)
followed by crystallization in a mixture of CHCl3/ hexane, as
a white solid from 2 (51 mg, 0.16 mmol) and diamine spacer arm
3b (15.6 mg, 0.08 mmol). mp 1298C. nmax/cmꢀ1 3311, 2921,
2852, 1633, 1539, 1190, 1147. lmax/nm (e/Mꢀ1 cmꢀ1) 350
(482.5). dH (300 MHz, CDCl3) 1.33–1.27 (m, 16H, –CH2–),
1.61 (m, 4H, NHCO–CH2–CH2–), 3.44 (dt, J 7, 13, 4H, NHCO–
CH2–), 6.13 (m, 2H, NHCO), 7.24 (d, J 8, 4H, ArH), 7.78 (d, J 9,
4H, ArH). dC (125 MHz, CDCl3) 166.6 (C¼O), 136.3 (ArC–
CNN), 132.5 (ArC–C¼O), 127.7 (ArCH), 127.0 (ArCH), 122.3
1
2
2
140.8 (dd, JC–F 249, JC–F 16, CF), 121.8 (t, JC–F 12, ArC–
2
C¼O), 112.4 (t, JC–F 19, CN3), 70.8 (–CH2CH2O–), 70.8
(–CH2CH2O–), 70.7 (–CH2CH2O–), 70.6 (NHCO–CH2–CH2–
), 69.8 (–CH2CH2O–), 40.4 (NHCO–CH2–). dF (282 MHz,
CDCl3) ꢀ141.06 (dd, J 20, 12, 4F), ꢀ150.9 (dd, J 20, 9, 4F).
m/z (APCI) 715 (50 %, MHþ).

Photolinkers for Crosslinking of PEG
201
1
(q, JC–F 273, CF3), 40.6 (NHCO–CH2–), 29.9 (NHCO–CH2–
[3] D. Wilms, M. Scho¨mer, F. Wurm, M. I. Hermanns, C. J. Kirkpatrick,
H. Frey, Macromol. Rapid Commun. 2010, 31, 1811. doi:10.1002/
MARC.201000329
[4] E. A. Phelps, N. O. Enemchukwu, V. F. Fiore, J. C. Sy, N. Murthy,
T. A. Sulchek, T. H. Barker, A. J. Garc´ıa, Adv. Mater. 2012, 24, 64.
doi:10.1002/ADMA.201103574
2
CH2–), 29.7 (–CH2–), 29.5 (–CH2–), 28.7 (q, JC–F 40, CN2),
27.3 (–CH2–). dF (282 MHz, CDCl3) ꢀ65.01 (s, CF3). m/z
(APCI) 625 (100 %, MHþ).
Synthesis of 4-[3-(Trifluoromethyl)diazirin-3-yl]-N-
[2-[2-[2-[2-[[4-[3-(trifluoromethyl)diazirin-3-yl]benzoyl]
amino]ethoxy]ethoxy]ethoxy]ethyl]benzamide 8a
[5] N. Karaca, G. Temel, D. Karaca Balta, M. Aydin, N. Arsu, J. Photoch.
Photobio. A 2010, 209, 1.
[6] N. Arsalani, P. Zare, H. Namazi, Express Polym. Lett. 2009, 3, 429.
doi:10.3144/EXPRESSPOLYMLETT.2009.52
[7] Y. Liu, J. Liu, J. Xu, S. Feng, T. P. Davis, Aust. J. Chem. 2010, 63,
The title compound (33 mg, 57 %) was obtained, after chro-
matography on silica gel with hexane/ethyl acetate (1 : 18), as a
yellowish oil from 2 (61.5 mg, 0.18 mmol) and diamine spacer
arm 4a (18.1 mg, 0.09 mmol). nmax/cmꢀ1 3342, 2923, 2854,
1647, 1550, 1186, 1155. lmax/nm (e/Mꢀ1 cmꢀ1) 350 (490.5). dH
(CDCl3, 300 MHz) 3.63 (m, 16H, –CH2CH2O–), 6.82 (s, 2H,
NHCO), 7.21 (d, J 8, 4H, ArH), 7.80 (d, J 8, 4H, ArH). dC
(125 MHz, CDCl3) 166.6 (C¼O), 135.9 (ArC–CNN), 132.6
1413. doi:10.1071/CH10168
[8] T. Yang, H. Long, M. Malkoch, E. K. Gamstedt, L. Berglund, A. Hult,
J. Polym. Sci. A1 2011, 49, 4044. doi:10.1002/POLA.24847
[9] M. S. Platz, Acc. Chem. Res. 1995, 28, 487. doi:10.1021/
AR00060A004
[10] A. Blencowe, W. Hayes, Soft Matter 2005, 1, 178. doi:10.1039/
B501989C
[11] M. Hashimoto, Y. Hatanaka, Anal. Biochem. 2006, 348, 154.
doi:10.1016/J.AB.2005.10.015
[12] M. Hashimoto, Y. Hatanaka, Eur. J. Org. Chem. 2008, 2513.
doi:10.1002/EJOC.200701069
1
(ArC–C¼O), 127.9 (ArCH), 127.0 (ArCH), 122.3 (q, JC–F
273, CF3), 70.8 (–CH2CH2O–), 70.5 (–CH2CH2O–), 70.1
2
(NHCO–CH2–CH2–), 40.2 (NHCO–CH2–), 28.7 (q, JC–F 40,
CN2). dF (282 MHz, CDCl3) ꢀ65 (s, CF3). m/z (APCI) 617
[13] A. Blencowe, K. Cosstick, W. Hayes, New J. Chem. 2006, 30, 53.
doi:10.1039/B514205A
(100 %, MHþ).
[14] D. Dankbar, G. Gauglitz, Anal. Bioanal. Chem. 2006, 386, 1967.
doi:10.1007/S00216-006-0871-X
10-kDa PEG Photocrosslinking with Compound 6a
[15] L.-H. Liu, M. Yan, Acc. Chem. Res. 2010, 43, 1434. doi:10.1021/
AR100066T
[16] A. Reiser, L. J. Leyshon, L. Johnston, Trans. Faraday Soc. 1971, 67,
Photocrosslinking in Bulk
Under dimmed lighting, in a round-bottomed quartz flask, 6a
(1 mg, 1.5 mmol) and 10-kDa PEG (11 mg, 1.1 mmol) were
introduced in dry dichloromethane (2 mL). After evaporation
of the solvent under vacuum, the mixture was irradiated
(3 ꢁ 8 W, 254 nm, placed at a distance of 10 cm) in a home-
made reactor (rotating 15-mL quartz flask) under an argon
atmosphere for 25 min.
2389. doi:10.1039/TF9716702389
[17] A. Blencowe, C. Blencowe, K. Cosstick, W. Hayes, React. Funct.
Polym. 2008, 68, 868. doi:10.1016/J.REACTFUNCTPOLYM.2007.
12.009
[18] J. F. W. Keana, S. X. Cai, J. Org. Chem. 1990, 55, 3640. doi:10.1021/
JO00298A048
[19] M. Nassal, Liebigs Ann. Chem. 1983, 1510. doi:10.1002/JLAC.
198319830907
[20] E. Leyva, R. Sagredo, E. Moctezuma, J. Fluor. Chem. 2004, 125, 741.
doi:10.1016/J.JFLUCHEM.2003.12.011
[21] G. B. Schuster, M. S. Platz, Adv. Photochem. 1992, 17, 69.
doi:10.1002/9780470133484.CH2
[22] M. Nassal, J. Am. Chem. Soc. 1984, 106, 7540. doi:10.1021/
JA00336A038
[23] S. P. Costas, Macromol. Symp. 2010, 291–292, doi:10.1002/
Photocrosslinking in Solution
Under dimmed lighting, in a round-bottomed quartz flask, 6a
(11 mg, 17.5 mmol) in acetonitrile (2 mL) was mixed with
10-kDa PEG (11 mg, 1.1 mmol) in Milli-Q water (6 mL). The
mixture was irradiated (3 ꢁ 8 W, 254 nm, placed at a distance of
10 cm) under stirring for 90 min. The water/acetonitrile mixture
was evaporated at 508C under reduced pressure to yield an
orange solid.
MASY.V291:1
[24] V. Pourcelle, H. Freichels, F. Stoffelbach, R. Auze´ly-Velty, C. Je´roˆme,
J. Marchand-Brynaert, Biomacromolecules 2009, 10, 966.
doi:10.1021/BM900027R
References
[1] K. Knop, R. Hoogenboom, D. Fischer, U. S. Schubert, Angew. Chem.
Int. Edit. 2010, 49, 6288. doi:10.1002/ANIE.200902672
[2] J. Zhu, Biomaterials 2010, 31, 4639. doi:10.1016/J.BIOMATERIALS.
2010.02.044
[25] L. Renaudie, C. LeNarvor, E. Lepleux, P. Roger, Biomacromolecules
2007, 8, 679. doi:10.1021/BM060961R
[26] J. Moreau, J. Marchand-Brynaert, Eur. J. Org. Chem. 2011, 2011,
1641. doi:10.1002/EJOC.201001666