Photocatalysis in an NMR tube: Carbon-carbon coupling ofphenoxyacetic acid with n-substituted maleimides

Article abstract of DOI:10.1016/j.jphotochem.2013.06.002

The semiconductor-sensitised photocatalytic (SPC) carbon-carbon coupling of phenoxyacetic acid (PAA)and N-methylmaleimide (NMI) or maleimide (MI) has recently been reported and is studied in moredetail. Irradiations are performed on a much smaller scale than commonly used, i.e. 1 mL of reac-tion solution in a sol-gel titania coated NMR tube, and NMR is used to monitor the progress of thevarious photoreactions. Use of an NMR tube as the photoreactor allows much faster reaction times(ca. 10 times faster than the 40 mL scale reaction), and helps avoid the more usual need for sample-taking and pre-treatment, before analysis by NMR, HPLC etc. The photochemistries and photocatalyticprocesses associated with the individual reactants (i.e. PAA and NMI or MI) reveal the productionof significant products which are largely absent in the SPC-sensitised coupling reaction. Thus, whenboth PAA and NMI or MI are irradiated together, in the study of the coupling reaction the two mainproducts formed are: 1-methyl-3-(phenoxymethyl)pyrrolidine-2,5-dione (the adduct product), and 2-methyl-3a,4-dihydrochromeno[3,4-c]pyrrole-1,3(2H,9bH)-dione (the cyclic product), which are formedin reasonable yields up to 67% and 15% respectively after only 15 min (I = 3.5 mW cm^-2per photoreac-tor hemisphere). Despite the well-known photochemical feature of NMI to form dimers efficiently, thiswork demonstrates that the background photochemistry and photocatalysis associated with the individ-ual reactants, PAA and NMI, in the photocatalysed coupling reaction are minimal due to the faster SPCkinetics for the coupling reaction, and the UV-filter effect of the sol-gel titania coating. The additionalformation of organic polymeric material in all the photocatalytic processes studied is discussed briefly.

Full text of DOI:10.1016/j.jphotochem.2013.06.002

Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
Contents lists available at SciVerse ScienceDirect
Journal of Photochemistry and Photobiology A:
Chemistry
journal homepage: www.elsevier.com/locate/jphotochem
Photocatalysis in an NMR tube: Carbon–carbon coupling of
phenoxyacetic acid with N-substituted maleimides
Andrew Mills∗, Christopher O’Rourke
School of Chemistry and Chemical Engineering, Queen’s University Belfast, Stranmillis Road, Belfast BT9 5AG, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
The semiconductor-sensitised photocatalytic (SPC) carbon–carbon coupling of phenoxyacetic acid (PAA)
and N-methylmaleimide (NMI) or maleimide (MI) has recently been reported and is studied in more
detail. Irradiations are performed on a much smaller scale than commonly used, i.e. 1 mL of reac-
tion solution in a sol–gel titania coated NMR tube, and NMR is used to monitor the progress of the
various photoreactions. Use of an NMR tube as the photoreactor allows much faster reaction times
(ca. 10 times faster than the 40 mL scale reaction), and helps avoid the more usual need for sample-
taking and pre-treatment, before analysis by NMR, HPLC etc. The photochemistries and photocatalytic
processes associated with the individual reactants (i.e. PAA and NMI or MI) reveal the production
of significant products which are largely absent in the SPC-sensitised coupling reaction. Thus, when
both PAA and NMI or MI are irradiated together, in the study of the coupling reaction the two main
products formed are: 1-methyl-3-(phenoxymethyl)pyrrolidine-2,5-dione (the adduct product), and 2-
methyl-3a,4-dihydrochromeno[3,4-c]pyrrole-1,3(2H,9bH)-dione (the cyclic product), which are formed
in reasonable yields up to 67% and 15% respectively after only 15 min (I = 3.5 mW cm−2 per photoreac-
tor hemisphere). Despite the well-known photochemical feature of NMI to form dimers efficiently, this
work demonstrates that the background photochemistry and photocatalysis associated with the individ-
ual reactants, PAA and NMI, in the photocatalysed coupling reaction are minimal due to the faster SPC
kinetics for the coupling reaction, and the UV-filter effect of the sol–gel titania coating. The additional
formation of organic polymeric material in all the photocatalytic processes studied is discussed briefly.
© 2013 Elsevier B.V. All rights reserved.
Received 25 March 2013
Received in revised form 6 June 2013
Accepted 7 June 2013
Available online xxx
Keywords:
Photocatalysis
Titania
Synthesis
Coupling reaction
NMR
1. Introduction
sensitised photo-Kolbe reaction [8,9], which generates a carbon-
centred radical via the decarboxylation of an unactivated carboxylic
The generally low environmental impact associated with using
semiconductor photocatalysis (SPC) to promote desired organic
reactions makes it an increasingly attractive alternative route in
organic synthesis. SPC has been used for many useful organic tran-
sitions such as: carbon-heteroatom bond formation [1], cyclisation
[2], reduction [3], and oxidation [4], and has been the subject of
numerous reviews [5,6]. The oxidation of benzene is a notable
example of the employment of SPC to produce a useful compound,
i.e. phenol, which is a well-known precursor to a variety of other
organic compounds [5]. Advantages of SPC over conventional syn-
thetic routes are: the high-level control (i.e. the reaction can be
stopped at the flick of a switch), selectivity, and the ability to per-
form reactions at room temperature and pressure.
acid. In the case of acetic acid, this leads to the formation of mainly
methane and carbon dioxide [9] (Eq. (1)), but, most importantly to
this study, also to the formation of acid dimers via the termination
of two carbon-centred radicals (Eq. (2)).
CH3CO2H → CH4 + CO2
(1)
(2)
2CH3CO2H → C2H6 + H2 + 2CO2
Manley et al. [7] recently used a modified version of the photo-
Kolbe reaction to generate the key carbon-centred radical in their
study of the SPC mediated carbon–carbon coupling of different car-
boxylic acids with a range of different alkenes. A typical coupling
reaction studied in this work was that between phenoxyacetic acid
(I) and N-methylmaleimide (II), giving rise to both an adduct (III)
and a chromenedione product (IV) (typically the latter is the major
species) [7]; the reported basic reaction and yields are given in
Scheme 1.
SPC has also been shown to be useful in effecting carbon–carbon
coupling reactions [7] and an early example of a SPC mediated
carbon–carbon coupling reaction is the reported semiconductor
It has been demonstrated recently by Mills et al. [10] that SPC
mediated organic transitions can be performed on a much smaller
scale, i.e. 1 mL, compared with the more typical >75 mL used for
∗
Corresponding author. Tel.: +44 028 9097 4339; fax: +44 028 9097 6524.
E-mail address: andrew.mills@qub.ac.uk (A. Mills).
1010-6030/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jphotochem.2013.06.002

8
A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
O
O
O
N
hv
O
N
O
OH
O
O
O
+
O
+
P25, MeCN
N
O
(IV) 55%
(I) 9 mM
(II) 18 mM
(III) 23%
Scheme 1. SPC C–C coupling of phenoxyacetic acid (9 mM) and N-methylmaleimide (18 mM) using P25 (1 mg mL−1) in a 75 mL reaction solution [7].
clear coating of anatase titania (ca. 1.4 mg cm−2, thickness ∼4 ␮m,
1 mL of reaction solution is exposed to ∼11 cm−2) on the inside of an
otherwise conventional glass NMR tube (500 MHz Wilmad® NMR
tubes 5 mm o.d., length = 20.3 cm). XRD revealed that the annealed
TiO2 film is in its anatase crystal form [12].
most other SPC mediated organic reactions, such as that illustrated
in Scheme 1. In such scaled-down SPC work, usually conducted in an
NMR tube using either a dispersion or coating of the semiconductor,
the kinetics of the overall process are usually much (>10×) faster.
This approach is also very convenient with regard to monitoring
the progress of the SPC reaction since it allows the use of NMR
to monitor the reactions and avoids the need for reaction solution
sampling and filtration of the semiconductor, before analysis (by,
for example, NMR, HPLC or GC) which is typical of most large scale
SPC mediated organic synthesis reactions.
This paper builds on the initial study of Manley et al. [7] and
aims to assess the C–C coupling process of phenoxyacetic acid
with maleimide and N-methylmaleimide in more detail, with the
additional assessment of the homogeneous photochemical and het-
erogeneous photocatalytic processes associated with the individual
components, PAA and NMI or MI. The progress of the reactions is
monitored by NMR, using usually a sol–gel titania coated NMR tube.
2.2. NMR tube irradiations
In situ NMR tube irradiations were carried out using a titania
sol–gel coated NMR tube. A photoreactor, consisting of two lots of
6 × 8 W black light UV lamps (ꢀmax (emission) 365 nm), arranged in
a hemispherical cylinder with an aluminium reflector, were used to
irradiate the contents of the NMR tube. The two photoreactor hemi-
spheres were brought together to surround the NMR tube as shown
in Fig. 1. Unless otherwise stated, the irradiance of each hemisphere
of the photoreactor was 3.5 mW cm−2. For a typical reaction, a 5 mL
stock reaction solution was prepared comprising: the reactant(s)
i.e. 10 mM of the phenoxyacetic acid (PAA) and/or 20 mM of the
N-methylmaleimide (NMI) or maleimide (MI), dissolved in deuter-
ated acetonitrile. This reaction solution was then purged with argon
for 5 min, to remove any dissolved oxygen, prior to pipetting 1 mL
of it into an argon flushed sol–gel coated NMR tube. The cap was
then quickly replaced on the tube and sealed with Parafilm. The
loaded NMR tube was then placed in the overhead stirrer (turned
on its side), which was then spun at 250 rpm. The tube was held at
a slight angle to horizontal in the ‘chuck’ of the overhead stirrer so
that when it was rotated it described a small (ca. 2 cm) circle at its
non-chuck-held end so as to agitate the reaction solution. Once the
NMR tube, filled with reaction solution, was in place, the irradia-
tion was started and, periodically, the tube was removed from the
chuck and placed into the NMR for analysis, before being returned
for further irradiation.
2. Experimental
Unless otherwise stated, all chemicals used in this project were
purchased from Sigma–Aldrich and were used as supplied. The P25
titania was provided by Evonik Degussa Corporation. P25 has a
surface area of ca. 50 m2 g−1 and is 80% anatase and 20% rutile [11].
2.1. Preparation of TiO2 coated Pyrex® NMR tubes
A paste of titania nanoparticles was prepared using the sol–gel
method described by Mills et al. [12]. Briefly, the precursor, tita-
nium(IV) isopropoxide (20 mL), was modified by the addition of
glacial acetic acid (4.65 g) under argon. The titania was then syn-
thesised via the sol–gel process by the addition to the solution of
120 mL of deionised water containing 1.08 g of nitric acid, thus
initiating the hydrolysis reaction, which leads to a condensation
reaction, that produces the hydrous oxide. The colloidal titania
particles were then grown via Ostwald ripening using an auto-
clave (220 ◦C for 720 min). The resulting precipitated particles
were dispersed using an ultrasonic probe and the solution was
rotary evaporated until a weight percent of titania of 10–12% was
achieved. At this stage 50 wt% of polyethylene glycol was added as
a binder to help prevent the formation of cracks in the cast films.
In order to coat an NMR tube with titania, a few drops of the
titania paste were pipetted into the NMR tube and shaken with
the NMR tube’s cap in place until the paste completely covered the
inside of the tube. The cap was then removed and the NMR tube
placed in the chuck of an overhead mechanical stirrer turned on
its side, and rotated horizontal to the floor at 1000 rpm for 5 min
to ensure an even thickness of the paste film on the sides of the
NMR tube. The tube was then placed under vacuum using a rotary
evaporator to remove all the solvent; the use of vacuum ensured
no bubbles were present in the paste, which could lead to poor
adhesion on the sides of the tube and result in flaking when dry. The
tube was then placed in a furnace and heated to 450 ◦C for 90 min
with an initial ramp rate of 10 ◦C min−1. The tube was then left
overnight to cool slowly inside the furnace to create a final robust
2.3. Large scale irradiations
In order to effect comparison with the work of Manley et al.
[7], 40 mL scale irradiations were also performed for the PAA-NMI
carbon–carbon coupling reaction (Scheme 1) using a 50 mL test
tube with either a sol–gel titania coating or a dispersion of tita-
nia. The sol–gel coated test tubes were prepared by first, pipetting
a line of the titania paste down the length of the inside of the tube.
Using a glass rod, the paste was then spread around the inside of
the tube as evenly as possible. The paste was allowed to dry com-
pletely before placing in a furnace for 90 min at 450 ◦C to anneal
(1.2 mg cm−2). The 40 mL reaction solutions were prepared using
HPLC grade acetonitrile, to which 10 mM of PAA acid and 20 mM of
NMI were added. For the titania dispersion work, a 1 mg mL−1 load-
ing of the semiconductor (P25, PC500 (Millennium Chemicals), or
Rutile nanopowder (Sigma–Aldrich (637262-100G))) was used in
the reaction solution, with the dispersion placed in a sonic bath for
5–10 min before being transferred inside the tube. A Suba-Seal®
cap was used to seal the tube, through which the reaction disper-
sion was then purged with argon for 10 min prior to irradiation.
The tubes were irradiated using the same 12 × 8 W BLB cylindri-
cal photoreactor illustrated in Fig. 1(a) but this time being held

A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
9
Fig. 1. (a) Schematic of the NMR tube irradiation setup. (b) Long exposure photograph showing the conical rotation of the NMR tube.
•
vertically, whilst stirred at the bottom by a magnetic flea, rotated
at 400 rpm. Samples were extracted periodically by syringe through
the Suba-Seal® cap and analysed by NMR.
with a second PhOCH2 radical to form the 1,2-diphenoxyethane
(II).
All NMR analyses were performed using a Bruker AVIII 400 Spec-
trometer fitted with a 5 mm PABBO BB-1H/D Z-GRD Z108618/0146
probe and equipped with TopSpin 1.3 software. Unless otherwise
stated, all experiments were performed in deuterated solvents.
When using non-deuterated solvent, (as in the large scale experi-
ments) the NMR was initially locked using an NMR tube containing
CD3CN. Subsequent samples were then scanned manually. This
caused slight broadening of the signals, however, the spectra were
still clear with good separation. Analyses of the data were car-
ried out using ACD/NMR Processor software (version 12.01) by
ACD/Labs.
3.2. Photochemistry and photocatalysis: N-methylmaleimide
(NMI) and maleimide (MI)
Irradiations were also performed with just the NMI (20 mM in
1 mL CD3CN, argon saturated), with and without titania present in
an NMR tube. NMI (ꢀmax = 300 nm) is able to absorb some of the
light provided by the BLB photoreactor, as shown by the UV/vis
absorption spectrum illustrated in Fig. 3 and, therefore, is able to
generate photochemical products. Similar experiments were also
carried out using MI (ꢀmax = 274 nm) which absorbs much less of
the 365 nm light from the photoreactors BLB’s and was also found
to be photochemically active.
Thus, in the absence of a TiO2 photocatalyst, UV irradiation pro-
motes a 2 + 2 cycloaddition of NMI to form the cyclobutane dimer
(III) in an 87% yield after 240 min. This dimerisation process has
been reported in previous studies on the photochemistry of NMI
[14–16]. In contrast, UV irradiation of the same solution (20 mM
NMI in CD3CN) in the presence of a sol–gel titania coating, yielded
the cyclobutane dimer (28%) along with N-methylsuccinimide (IV)
(29%) after 240 min as illustrated in Fig. 4 by the observed changes
in NMR spectra of the reaction solution as a function of irradiation
time.
3. Results and discussion
The names and structures of all the compounds used and formed
in this work, alongside their abbreviations/numerals, are given in
Table 1.
3.1. Photocatalysis: phenoxyacetic acid (PAA)
From the UV/vis absorption spectrum of phenoxyacetic acid
(PAA) (ꢀmax = 270 nm) it is clear that it does not absorb any of the UV
light supplied by the 12 × 8 W BLB photoreactor (ꢀmax = 365 nm).
As a result, it is not surprising to note that irradiation of 1 mL
of 10 mM PAA in CD3CN, in a clean NMR tube, over a 240 min
period showed no reaction/degradation. The introduction of a tita-
nia sol–gel coating in the NMR tube does however give rise (after
90 min of irradiation) to the formation of anisole (I) (45% yield)
and the dimerisation product, 1,2-diphenoxyethane (II) (30% yield),
note, the latter is a sensitiser in thermal paper [13]. The general
reaction scheme along with the change in the NMR spectrum as
the photoreaction proceeds is shown in Fig. 2.
A possible mechanism for the formation of anisole (I) and the
acid dimer (II) is given in Scheme 2. Thus, the PAA is initially oxi-
dised on the surface of the titania by photogenerated holes, h+, to
form a cationic radical, which then subsequen•tly loses a proton
and carbon dioxide to form the neutral PhOCH2 radical. The later
species then either gains a proton and an electron from the surface
of the titania to form the anisole (I), or couples in a termination step
Table 2 lists the yields of the various reactants and products of
the irradiations of PAA with and without titania for both NMI and
also MI.
Fig. 3 shows that the sol–gel coating in the NMR tube absorbs
most of the light before it reaches the reaction solution and so
minimises the homogeneous photochemistry of NMI and MI that
might occur otherwise. In order to demonstrate the UV filter effect
of the titania film, clean NMR tubes were coated on the outside
with the sol–gel titania and used in the UV irradiation of 20 mM
solutions of each maleimide. The results of this work are given in
Table 2 and reveal that the titania (exterior) coating reduces signif-
icantly (e.g. NMI dimer yield reduced from 87% to 21%) the degree
of homogeneous photochemistry that NMI exhibits if a titania film
is present.
Scheme 3 illustrates the homogeneous photochemical mech-
anism in which the cyclobutane dimer is produced upon UV
excitation of NMI, as proposed by Jönsson et al. [14], along with

10
A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
Table 1
The names and structures of all the compounds used and formed in this study, and their corresponding abbreviations/numerals.
Chemical
Structure
Abbreviation
Starting materials
O
O
OH
Phenoxyacetic acid
N-Methylmaleimide
PAA
NMI
N
O
O
O
H
N
O
Maleimide
MI
(I)
Products
O
O
Anisole
O
1,2-Diphenoxyethane
(II)
2,5-Dimethyldihydrocyclobuta[1,2-c:3,4-c’]dipyrrole-1,3,4,6(2H,3aH,3bH,5H)-tetraone (R = CH3)
(III)
O
O
N
O
R
R
O
N
O
Tetrahydrocyclobuta[1,2-c;3,4-c’]dipyrrole-1,3,4,6-tetraone (R = H)
N-methylsuccinimide (R = CH3)
(III)ꢀ
(IV)
R
N
O
O
Succinimide (R = H)
1-Methyl-3-(phenoxymethyl)pyrrolidine-2,5-dione (R = CH3)
(IV)ꢀ
(V)
R
N
O
O
O
3-(Phenoxymethyl)pyrrolidine-2,5-dione (R = H)
(V)ꢀ
2-Methyl-3a,4-dihydrochromeno[3,4-c]pyrrole-1,3(2H,9bH)-dione (R = CH3)
(VI)
O
N
O
R
3a,4-Dihydrochromeno[3,4-c]pyrrole-1,3(2H,9bH)-dione (R = H)
(VI)ꢀ
the photocatalytic pathway in which N-methylsuccinimide (IV) is
the major product. Thus, in Scheme 3, the NMI is excited to its
triplet state by the absorption of UV light, which then reacts with
a neighbouring ground state NMI molecule, attacking the double
bond. The di-radical product then terminates with the formation
of a second C–C bond, forming the cyclobutane ring of the dimer
(III). In contrast, in the SPC mediated production of (IV) from NMI,
an initial hydrogen abstraction reaction is followed by the addi-
tion of a proton and an electron from the surface of the titania to
generate (IV).
3.3. Photochemistry and photocatalysis: phenoxyacetic acid and
N-methylmaleimide
UV irradiation of PAA (20 mM) with NMI (20 mM), in the absence
of titania, not surprisingly gave rise to the 2 + 2 cyclobutane dimer
(III) in a 38% yield after 180 min, with no evidence of coupling
products being formed, despite a 30% consumption of the acid. In
contrast, UV irradiation of the same reaction solution in the pres-
ence of a TiO2 film, generated numerous products, as evident from
the numerous changes in the NMR spectrum shown in Fig. 5, from

A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
11
Fig. 2. Reaction scheme and NMR spectra recorded for the reaction solution (10 mM PAA in 1 mL CD3CN, in a sol–gel coated NMR tube) as a function of irradiation time.
.
+
O
O
+
.
CH2
-
O
O
O
H
hv
OH
OH
-CO2
TiO2
.
O
CH2
PAA
H+
-
e
O
CH3
O
O
(I)
(II)
Scheme 2. Proposed SPC mechanism for the formation of anisole (I) and 1,2-diphenoxyethane (II) from phenoxyacetic acid.
Table 2
Solo irradiations using the 12 × 8 W BLB photoreactor (ꢀmax = 365 nm), of each N-maleimide (20 mM in CD3CN) in an NMR tube with and without titania present. Irradiation
time = 240 min.
N-Maleimide
Semiconductor
None
Starting material
Product yields (%)
Dimer (III)
% Conversion
99
N-Succinimide (IV)
0
87
N
O
O
Sol–gela
97
68
28
21
29
0
None with sol–gel filterb
N-methylmaleimide
None
83
11
0
H
N
O
O
Sol–gela
93
24
6
3
22
0
Maleimide
None with sol–gel filterb
a
Sol–gel titania coated NMR tube, mass = 1.4 0.2 mg cm−2, 1 mL solution is exposed to ∼11 cm2.
b
The sol–gel filter was produced by coating the outside of a clean NMR tube with the titania.

12
A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
3
2
1
0
1
MI
NMI
0.5
Sol-Gel Coated NMR Tube
Standard Pyrex NMR Tube
500 600
0
200
300
400
Wavelength/ nm
Fig. 3. UV/Vis spectra of N-methylmaleimide (NMI), maleimide (MI), a clean Pyrex® NMR tube, and a sol–gel titania coated NMR tube. The broken line represents the emission
profile of the 12 × 8 W BLB photoreactor, recorded using an OceanOptics USB2000+ fibre optic spectrometer.
which the formation of an adduct (V) and cyclic (VI) compounds as
major products is clear. Fortunately, despite the formation of sev-
eral major and minor products in this coupling reaction, the NMR
spectra were well defined and allowed facile monitoring of each
component. Thus, Fig. 6 shows the progress of the concentrations
of the starting materials and all known products as a function of
irradiation time measured by NMR. Note, the reaction is relatively
fast compared to the previous photoreactions reported here, pro-
ducing maximum yields of the main adduct (67%), cyclic (15%), and
N-methylsuccinimide (IV) (19%) products after only 15 min of UV
irradiation.
The general reaction yields for the SPC coupling of PAA with NMI
in CD3CN, using a sol–gel coated NMR tube are given in Scheme 4.
Table 3 lists the NMR-measured yields of the C–C coupling reac-
tions of PAA with either NMI or MI using a sol–gel titania coat
as the SPC mediator in an NMR tube. In both cases, using either
NMI or MI, it is observed that the single component homogeneous
and SPC photoreactions are not significant, since, for example, the
Fig. 4. Reaction scheme and NMR spectra recorded for the reaction solution (20 mM NMI in 1 mL CD3CN, in a sol–gel coated NMR tube) as a function of irradiation time.

A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
13
Scheme 3. Proposed photochemical mechanism for the formation of the cyclobutane dimer as reported by Jönsson et al. [14], and the proposed photocatalytic mechanism
for the formation of N-methylsuccinimide.
Scheme 4. General scheme for the C–C coupling of PAA with NMI, giving rise to the main adduct (V) and cyclic (VI) products, as well as a range of minor side products.
between the adduct and cyclic product yields when using a 1:1 or a
1:2 (acid: N-maleimide) starting material ratio. Using a 1:1 ratio did
however result in much lower yields of the photochemically pro-
duced 2 + 2 cyclobutane dimer, as there was not an excess of the
NMI or MI present to promote the homogeneous photochemistries
of NMI or MI (see Scheme 3).
The proposed mechanism for the SPC C–C coupling reaction
between PAA and the NMI is shown in Scheme 5. The PAA is oxi-
dised by a hole, h+, on the titania to produce a radical cation. This
radical then undergoes decarboxylation to produce a PhOCH2• rad-
ical, which then attacks the double bond of the NMI to produce a
radical adduct. From this point, the radical adduct can either be
reduced by an electron from the titania to form the adduct prod-
uct or cyclise to form eventually the cyclic product via an electron
transfer step, ET, involving another ground state NMI molecule.
The likely additional reactions of the reduced NMI radical product
to form N-methylsuccinimide (IV) are illustrated at the bottom of
Scheme 5. These radicals, as with most organic radicals produced
photochemically or photocatalytically, are also involved in much
less easy to follow polymer generation processes, and this helps
Fig. 5. Adduct and cyclic products formed upon irradiation of phenoxyacetic acid
(20 mM) and N-methylmaleimide (20 mM), in 1 mL CD3CN in a sol–gel titania coated
explain why there is no simple correlation observed between the
yield of the cyclic products and that of the succinimide (vide infra).
NMR tube using the 12 × 8 W BLB photoreactor (ꢀmax = 365 nm).
yields of the cyclobutane dimers (from the homogeneous photo-
chemistry) and the anisole (from the SPC reaction between the
acid and the TiO2 alone) is only 4% or less. The greater rate of the
SPC C–C coupling reaction (15 min) ensures that the single com-
ponent homogeneous and heterogeneous photochemistries, which
are slow (typically 240 min for maximum yields), are significantly
reduced. The filter effect of the titania coating also contributes to
minimising the effect of the homogeneous photochemistries of NMI
and MI. It was also observed that there was no significant difference
3.4. Photocatalysis: phenoxyacetic acid and N-methylmaleimide
(large scale)
In order to better compare the above results to those of Man-
ley et al. [7] with regard to the photocatalysed coupling of PAA
and NMI or MI, sensitised by TiO2, the 1 mL reaction was scaled up
to 40 mL and performed in large 50 mL test tubes (i.d. = 2 cm). One
tube was coated with the sol–gel paste using a similar procedure
Table 3
SPC C–C coupling of PAA with both NMI and MI using a sol–gel titania coated NMR tube, irradiated using the 12 × 8 W BLB photoreactor (ꢀmax = 365 nm).
Starting materials
SC
Adduct product (V)
Cyclic product (VI)
Succinimide (IV)
% Yield
Maleimide Dimer (III)
% Yield
Anisole (I)
% Yield
PAA (mM)
N-Maleimide (mM)
tmax (min)
% Yield
tmax (min)
% Yield
20
20
NMI
MI
20
20
Sol–gela
Sol–gela
15
30
67 (14)
55 (17)
15
60
15 (67)
18 (55)
19
24
1
0
4
4
Notes: tmax = time taken to reach a maximum % yield with respect to the initial concentration of phenoxyacetic acid. For some reactions the two major products did not reach
a maximum at the same time, therefore, the numbers in brackets represent the % conversion of the other product at tmax of the product in question.
a
Sol–gel: mass = 1.4 0.2 mg cm−2, 1 mL solution is exposed to ∼11 cm2.

14
A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
25
20
15
10
5
0
0
5
10
15
20
25
30
Irradiation Time/ min
Fig. 6. Progress of the SPC C–C coupling reaction of 20 mM phenoxyacetic acid (᭹) and 20 mM N-methylmaleimide (ꢁ), showing the evolution of; the adduct product (ꢀ),
the cyclic product (ꢁ), N-methylsuccinimide (ꢂ), and also the minor by-products; 2 + 2 cyclobutane dimer (×), and anisole (ꢃ).
to that used to coat the NMR tubes and the other tube contained
a dispersion of P25 at the same concentration as used by Manley
et al. [7] (i.e. 1 mg mL−1). The larger vessels allowed all reactions
to be mechanically stirred at 400 rpm, thus allowing a good dis-
persion to be maintained. In this system, two other semiconductor
powders were also assessed, namely: PC500 and nano rutile. The
results of the 40 mL scale coupling reactions with PAA (10 mM) and
NMI (20 mM) are given in Table 4 along with the details of each of
the semiconductor used, such as their surface areas, and particle
sizes.
The results of the large scale irradiations given in Table 4 show
that, as found in the NMR tube study, the adduct product is favoured
Scheme 5. Proposed SPC C–C coupling mechanism between PAA and NMI.

A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
15
Table 4
SPC C–C coupling of PAA (10 mM) with NMI (20 mM) in 40 mL of non-deuterated acetonitrile using a range of different semiconductors (1 mg mL−1).
Photocatalyst
Composition
Surface Area (m2 g−1
)
Particle size (nm)
Irradiation time (min)
Product yield %
Adduct (V)
Cyclic (VI)
P25
80% anatase 20% rutile
80% anatase 20% rutile
100% anatase
50
50
–
21
21
53 [17]
5–10 [11]
900
900
300
900
17
23
74
35
30
55
19
18
P25a
Sol–Gelb
PC500
100% anatase
>250 [11]
a
Results obtained by Manley et al. [7] using a 1 mg mL−1 P25 dispersion on a 75 mL scale. Starting material ratio (PAA:NMI) = 1:2.
b
Sol–gel: mass = 1.4 0.2 mg cm−2, 40 mL solution is exposed to ∼74 cm2.
when a sol–gel film is used as a photocatalyst. PC500 (also 100%
and photocatalysis of the single substrates, i.e. acid dimer, anisole,
and the cyclobutene dimer are low (≤4%), and, instead, the main
coupling products are the adduct (V) (67%) and cyclic (VI) (15%)
compound and N-succinimide (IV). The latter is the oxidation prod-
uct derived from the NMI or MI starting product, which acts as an
electron acceptor. Similar products, i.e. large amounts of adduct
and less cyclic product, are observed in scaled up reactions using
a titania paste film or PC500. In contrast, in the same photocat-
alytic reaction performed using P25, the cyclic product is dominant
and with nano rutile powder as the semiconductor sensitiser, no
products are formed.
anatase) powder also generates the adduct (V) as the major prod-
uct and the cyclic product (VI) with a smaller yield. In contrast, and
as reported by Manley et al. [7], P25 appears to favour the pho-
tocatalytic formation of the cyclic product (VI). It is not clear why
P25 favours the generation of the cyclic product, and the sol–gel
paste and PC500 photocatalysts favour the adduct product, nor why
nano rutile powder is found to be ineffective as a photocatalyst for
the coupling reaction. However, such a variation in the product
distribution in semiconductor photocatalysis is not unusual and is
usually attributed to the sensitivity of reaction rates and mecha-
nisms in SPC on a variety of factors including: the crystal phase and
surface morphology of the semiconductor photocatalyst.
With prolonged use the titania photocatalyst (P25 or titania
film) turns brown, due to organic polymer formation, although it
can be restored to its original activity by heating in air at 450 ◦C for
30 min.
3.5. Prolonged irradiation or repeated use: polymer formation
The results of this work show that the use of titania as a pho-
tocatalyst in this coupling reaction is complicated, since it can
photocatalyse different reactions with the individual substrates,
and there is the possibility of homogeneous photochemical pro-
cesses occurring. The underlying radical mechanisms responsible
for product generation lead to organic film formation on the sur-
face of the semiconductor photocatalyst that lowers its activity, but
which can be recovered by a simple heat treatment process. The use
of a titania sol–gel film in an NMR tube to promote various organic
reactions, including, in this case, coupling reactions, appears partic-
ularly effective, as they are very active, reusable, and readily allow
the NMR of the reaction solution to be monitored as a function of
time. This observation suggests that they may be used as a use-
ful tool, before scale-up, to screen various potential SPC mediated
organic reactions.
During the course of all this work, using either the sol–gel
coating or a dispersion of P25 or PC500, it was noted that, with
prolonged irradiation, the titania began to turn brown/yellow in
colour and lose its potency as a photocatalyst. This feature is most
likely due to the formation of insoluble polymeric material adher-
ing to the surface. For example, the discolouration of titania has
been previously reported by Sitkiewitz and Heller [18] in a study of
the SPC oxidation of benzene [18,19] and attributed to polymer film
formation on the surface of the titania. These workers and others
[19] have demonstrated that the brown colour (attributed to poly-
meric material on the surface) could be removed by irradiation of
the stained TiO2 photocatalyst under humid conditions, to CO2 and
CO at the same time returning the titania to its original, active,
white form. In our work, TGA of the brown/yellow sol–gel pro-
duced after prolonged irradiation or repeated use in the PAA/NMI
SPC reaction, showed the presence of 6.6 wt% of adsorbed mate-
rial which was completely removed by heat treating the powder at
450 ◦C (30 min), whilst at the same time returning the titania to its
original photoactivity and colour.
Acknowledgments
The authors would like to thank the EPSRC for funding this work
(Grant EP/I003800/2).
4. Conclusions
The use of a titania photocatalyst to promote the coupling of a
carboxylic acid with an alkene like NMI or MI, appears more compli-
cated than reported previously [7]. In the study of the UV-sensitised
photocatalytic coupling of PAA with NMI or MI, in the absence of
titania, the phenoxyacetic acid on its own does not absorb the UVA
light and so no photochemical products are observed. In the pres-
ence of titania, however, high yields of anisole (45%) and acid dimer
(30%) are produced. In the absence of titania, and as observed by
others, the NMI or MI on their own absorb UV light and generate the
corresponding cyclobutane dimer in reasonable yields (87% or 11%
respectively). In the presence of a titania photocatalyst, UV irradi-
ation of NMI and or MI produces N-succinimide in high yields i.e.
29% and 22% respectively, as well as the dimer (yields 28% and 6%,
respectively).
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.jphotochem.
2013.06.002.
References
[1] K.V. Subba Rao, B. Srinivas, A.R. Prasad, M. Subrahmanyam, Chemical Commu-
nications 16 (2000) 1533–1534.
[2] S. Nishimoto, B. Ohtani, T. Yoshikawa, T. Kagiya, Journal of the American Chem-
ical Society 105 (1983) 7180–7182.
[3] A. Maldotti, L. Andreotti, A. Molinari, S. Tollari, A. Penoni, S. Cenini, Journal of
Photochemistry and Photobiology A 133 (2000) 129–133.
In the coupling of PAA with NMI and MI, the yields of most
of the products associated with the homogeneous photochemistry

16
A. Mills, C. O’Rourke / Journal of Photochemistry and Photobiology A: Chemistry 268 (2013) 7–16
[4] D. Worsley, A. Mills, K. Smith, M.G. Hutchings, Journal of the Chemical Society,
Chemical Communications 11 (1995) 1119–1120.
[12] A. Mills, N. Elliott, G. Hill, D. Fallis, J.R. Durrant, R.L. Willis, Photochemical &
Photobiological Sciences 2 (2003) 591–596.
[5] Y. Shiraishi, T. Hirai, Journal of Photochemistry and Photobiology C 9 (2008)
157–170.
[13] V.J. Smith, D.J. Clark, K.B. Damarell, U.S. Patent 5,071,821.
[14] S. Jönsson, K. Viswanathan, C.E. Hoyle, S.C. Clark, C. Miller, F. Morel, C. Decker,
Nuclear Instruments and Methods in Physics Research Section B 151 (1999)
268–278.
[15] S. Jönsson, P.E. Sundell, M. Shimose, S. Clark, C. Miller, F. Morel, C. Decker, C.E.
Hoyle, Nuclear Instruments and Methods in Physics Research Section B 131
(1997) 276–290.
[16] P. Boule, J. Lemaire, Journal de Chimie Physique et de Physico-Chimie
Biologique 77 (1980) 161–165.
[17] A. Mills, G. Hill, M. Crow, S. Hodgen, Journal of Applied Electrochemistry 35
(2005) 641–653.
[6] G. Palmisano, E. Garcıa-Lopez, G. Marci, V. Loddo, S. Yurdakal, V. Augugliaro, L.
Palmisano, Chemical Communications 46 (2010) 7074–7089.
[7] D.W. Manley, R.T. McBurney, P. Miller, R.F. Howe, S. Rhydderch, J.C. Walton,
Journal of the American Chemical Society 134 (2012) 13580–13583.
[8] B. Kraeutler, A.J. Bard, Journal of the American Ceramic Society 100 (1978)
2239–2240.
[9] B. Kraeutler, C.D. Jaeger, A.J. Bard, Journal of the American Ceramic Society 100
(1978) 4903–4905.
[10] A. Mills, C. O’Rourke, Journal of Photochemistry and Photobiology A 233 (2012)
34–39.
[18] S. Sitkiewitz, A. Heller, New Journal of Chemistry 20 (1996) 233–241.
[19] H. Einaga, S. Futamura, T. Ibusuki, Physical Chemistry Chemical Physics 1 (1999)
4903–4908.
[11] C.S. Uyguner, M. Bekbolet, International Journal of Photoenergy 6 (2004)
73–80.