Chemo- And diastereoselectivities in the electrochemical reduction of maleimides

Article abstract of DOI:10.1002/cssc.201403184

The electrochemical cathodic reduction of cyclic imides (maleimides) to succinimides can be achieved chemoselectively in the presence of alkene, alkyne, and benzyl groups. The efficiency of the system was demonstrated by using a 3D electrode in a continuous flow reactor. The reduction of 3,4-dimethylmaleimides to the corresponding succinimides proceeds with a 3:2 diastereomeric ratio, which is independent of the nitrogen substituent and electrode surface area. The stereoselectivity of the process was rationalized by using DFT calculations, involving an acid-catalyzed tautomerization of a half-enol occurring through a double hydrogen-transfer mechanism.

Full text of DOI:10.1002/cssc.201403184

DOI: 10.1002/cssc.201403184
Full Papers
Chemo- and Diastereoselectivities in the Electrochemical
Reduction of Maleimides
Kathryn Rix,[a] Geoffrey H. Kelsall,[b] Klaus Hellgardt,[b] and King Kuok (Mimi) Hii*[a]
The electrochemical cathodic reduction of cyclic imides (malei-
mides) to succinimides can be achieved chemoselectively in
the presence of alkene, alkyne, and benzyl groups. The effi-
ciency of the system was demonstrated by using a 3D elec-
trode in a continuous flow reactor. The reduction of 3,4-dime-
thylmaleimides to the corresponding succinimides proceeds
with a 3:2 diastereomeric ratio, which is independent of the ni-
trogen substituent and electrode surface area. The stereoselec-
tivity of the process was rationalized by using DFT calculations,
involving an acid-catalyzed tautomerization of a half-enol oc-
curring through a double hydrogen-transfer mechanism.
Introduction
Metal-catalyzed hydrogenation with H2 is widely employed for
the reduction of C=C bonds to deliver the syn addition of H2
on the surface of a heterogeneous catalyst (e.g., Ni, Pd, Pt, Rh,
Ru).[1] Although the process is 100% atom-efficient, considera-
ble process control is required to overcome mass transport
limitations without engendering explosive hazards associated
with the use of the highly flammable gas in pressurized reac-
tors.[2] Conversely, the anti addition of H2 can be achieved only
by using stoichiometric one-electron reductants to deliver elec-
trons and protons to opposite faces of the alkene in two suc-
cessive steps, typically by dissolving metals in a protic solvent
(e.g., Na/NH3 or Mg/MeOH).[3] Apart from poor atom economy,
such processes can also be difficult to scale up and produce
metal-containing effluents.[4]
electrode potential appropriate to the reduction of a particular
functional group.
Recent advances in flow chemistry and the use of novel
electrode materials and electrochemical reactors have greatly
accelerated the development of electrosynthesis, overcoming
limited productivity associated with mass transport limitations,
as well as competitive reduction of the reaction solvent.[5]
Herein, we describe the deployment of a laboratory-scale flow
electrochemical reactor for the selective cathodic reduction of
cyclic imides 1 to give succinimides 2 (Scheme 1).[6] The pro-
cess affords clean and efficient reduction of the C=C bond
under ambient conditions to generate O2 as the only byprod-
uct.
In comparison, electrochemical synthesis has the potential
to deliver exceptional environmental and safety benefits. The
need for stoichiometric and hazardous reductants is obviated
by using electrons, water, and H+ as reagents, while electrodes
can be regarded as heterogeneous catalysts that can be easily
separated from the products and reused. Energy consumption
is minimized by judicious reactor design and choice of elec-
trode material; this allows reactions to occur under ambient
conditions. Additionally, the chemoselectivity of an electro-
chemical reaction can be achieved, in part, by applying an
[a] Dr. K. Rix, Dr. K. K. Hii
Scheme 1. Electrochemical reduction of maleimides 1.
Department of Chemistry, Imperial College London
Exhibition Road, South Kensington, London SW7 2AZ (UK)
E-mail: mimi.hii@imperial.ac.uk
[b] Prof. G. H. Kelsall, Prof. K. Hellgardt
Department of Chemical Engineering
South Kensington Campus, Imperial College London
London SW7 2AZ (UK)
Results and Discussion
Electrochemical reduction of maleimide (1a) to form succin-
imide (2a) was first reported in 1937 to proceed at a lead or
copper cathode in a solution of sulfuric acid containing ele-
mental nickel particles in suspension.[7] Some four decades
later, the reaction was re-examined by Barradas and co-work-
ers, who reported the selectivity of the process to be highly
pH-dependent: in a basic solution, maleimide undergoes hy-
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201403184.
ꢀ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons
Attribution License, which permits use, distribution and reproduction in
any medium, provided the original work is properly cited.
ChemSusChem 2015, 8, 665 – 671
665 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
drolysis and/or dimerization reactions,[8] whereas the reduction
to succinimide occurs exclusively in acidic media.[9] These ob-
servations were similarly observed in a separate study of N-
substituted maleimide compounds.[10] Although these studies
focused on the mechanistic aspects of the reaction, the syn-
thetic utility of the methodology had not been assessed ade-
quately in terms of its productivity, its current efficiency/charge
yield, as well as chemo- and stereoselectivity. Herein, we ad-
dress these issues by studying the electrochemical reduction
of five maleimide derivatives 1a–e. The corresponding succini-
mides of the methylated derivatives (1d and 1e) are particular-
ly interesting because they are reported to have antifungal ac-
tivity against human opportunistic pathogenic fungi, including
yeasts and dermatophytes.[11]
Table 1. Electrochemical characteristics of maleimide derivatives in an
aqueous solution of H2SO4.[a]
Compound pH Onset reduction potential Transport-limited current
vs. SCE [V]
density[b] [Amꢀ2
]
1a
1b
1c
0
2
4
0
2
4
0
4
ꢀ0.9
ꢀ1.1
ꢀ1.0
ꢀ0.6
ꢀ0.6
ꢀ0.8
ꢀ0.6
ꢀ0.8
130
60
45
110
80
40
80
40
[a] Recorded by using the VC RDE; 10 mm maleimide derivatives. See the
Experimental Section for conditions. Standard reduction potentials are
provided in Table S1 in the Supporting Information. [b] Two-electron re-
ductions established by Levich plots (Figures S1 and S2 in the Supporting
Information).
To eliminate possible metal-catalyzed hydrogenation result-
ing from surface hydride species and toxic metal residues,
carbon-based electrodes, such as vitreous carbon (VC), graph-
ite, and boron-doped diamond (BDD), were utilized during this
work. The use of BDD is particularly attractive because it has
a particularly high overpotential for hydrogen (and oxygen)
evolution, offering a working electrode potential window of
between about ꢀ0.75 and +2.35 V (vs. a standard hydrogen
electrode (SHE)) in a 0.5m aqueous solution of H2SO4.[12] Hence,
loss in efficiency through competitive reduction rates of pro-
tons/water may be minimized, compared with rates on, for ex-
ample, copper or lead.
diminished transport-limited current densities (Table 1); this
suggested smaller diffusion coefficients for these N-substituted
substrates. The effect of nitrogen substitution on the stability
of radical intermediates has been reported previously.[10c] Cru-
cially, no other reduction processes were detected within the
potential range; thus, the chemoselective reduction of the
electron-deficient alkene could be achieved in the presence of
these other unsaturated moieties. In accordance with previous
studies, increasing the pH of the solution decreased the onset
electrode potentials (vs. SCE) for maleimide reduction and de-
creased the current densities. In contrast, no reduction waves
could be recorded for the dimethyl-substituted maleimides 1d
and 1e; the signals were obscured by significant competitive
reduction of H+ to H2.
To identify the potential range over which reduction of 1a
occurred under mass transport control, experiments were per-
formed at a VC rotating disc electrode (RDE). With 10 mm of
the substrate dissolved in 1m H2SO4 (Figure 1), the onset of
a two-electron reduction process was detected at ꢀ0.6 V (vs.
SCE), with mass-transport-limited current densities achieved at
potentials <ꢀ0.9 V (vs. SCE), below which the rate of H2 evolu-
tion increased significantly. At pH 2 and 4, the onset potentials
for the reduction of the C=C bond occurred at slightly more
negative values, and were accompanied by a decrease in cur-
rent densities (Table 1, entries 2 and 3).
Flow reactor
A
commercial parallel-plate reactor (Electrocell microflow
cell[13]) was modified for flow studies (Figure 2). Specifically,
polytetrafluoroethylene (PTFE) gaskets were introduced to pro-
vide chemical inertness, prevent contamination, and to accom-
modate a AgCljAg reference electrode (see Figure 4 below in
Subsequently, the onsets of the reduction of N-allyl (1b) and
N-propargyl (1c) maleimides dissolved in 1m H2SO4 and MeOH
(4:1 v/v) were also determined to be ꢀ0.6 V at pH 0, but with
Figure 1. Cyclic voltammogram of 10 mm 1a in 1m H2SO4 at a VC RDE elec-
trode. Scan rate 25 mVsꢀ1; rotation rate=1000 rpm).
Figure 2. Schematic illustration of the electrochemical reactor operating in
batch recycling mode described herein.
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
666 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
the Experimental Section). The flow reactor was operated in
batch recycling mode with a continuously stirred tank reser-
voir. With the cathode controlled at a fixed electrode potential,
current densities were monitored as a function of time. How-
ever, because maleimide reduction occurred in parallel with
hydrogen evolution (E(H+/H2)ꢁꢀ0.245 V vs. SCE), forming
bubbles that adhered to the electrode, the effective area of
the cathode/solution interface was altered, which caused time-
dependence current densities as bubbles formed and detach-
ed. Hence, observation of the two different processes can be
decoupled by the incorporation of a flow cuvette online to
monitor maleimide formation through UV spectroscopy; a gas
flow counter was installed at the cathode to monitor the com-
petitive formation of H2. Product formation was independently
verified by extracting an aliquot (1 mL) of the catholyte, which
was neutralized to pH 10 with a 1m aqueous solution of NaOH
and extracted with dichloromethane. Following evaporation of
selectivity, at the expense of reduced charge yield (Table 2, en-
tries 3–5). This may be attributed to slower rates of mass trans-
port of these species to the cathode due to their larger molec-
ular/hydrodynamic sizes, and hence, smaller diffusion coeffi-
cients.[15] The nature of the nitrogen substituent appeared to
be particularly important in this context; charge yields of 85
and 45% were recorded for 1b and 1c, respectively. This may
be associated with the lower current densities recorded for
these compounds (Table 1), although the correlation was non-
linear. The introduction of methyl substituents at C-3 and C-4
of 1d led to a further decrease in charge yields to 45%
(Table 2, entry 6), whereas the presence of both C and N sub-
stituents (1e) did not appear to be additive (Table 2, entry 7).
This was mainly attributed to the competitive reduction of pro-
tons at this applied potential. Nevertheless, in both cases, very
high conversions to the corresponding succinimides were ach-
ieved in 5 h.
1
the solvent, the residue was analyzed by H NMR spectroscopy.
Initially, experiments were performed with solutions of
10 mm of the maleimide substrate in 1m H2SO4. This solution
(100 mL) was recycled at a rate of 60 mLminꢀ1 through the
electrochemical reactor, with a potential of ꢀ1.2 V (AgCljAg)
applied to the cathode. Reduction was counterbalanced by ox-
idation of water to generate O2 in the 1m H2SO4 anolyte. By
using a VC cathode, the reduction of 1a to 2a was detected,
with 60% conversion after 5 h, which corresponded to
a charge yield of 62% (Table 2, entry 1). In comparison, by
Electrode geometry and productivity
The volumetric current density, and hence, conversion rate of
the flow reactor was enhanced by increasing the specific sur-
face area of the electrode. This was achieved by inserting an
11 mm thick graphite felt into the PTFE frame in contact with
the BDD cathode. A low-resistance ohmic contact between
a BDD plate feeder electrode and graphite felt enabled the
active cathode area to be extended greatly (from 1.03ꢁ
10ꢀ4 m2 to about 0.2 m2), which should enable the scale up of
reaction rates. However, the local electrode potential is spatial-
ly distributed in 3D electrodes and varies with current densi-
ty.[16]
Table 2. Charge yields for the reduction processes using the flow reactor
(2D electrode).[a]
In this experiment, the concentration depletion of maleimide
was monitored by the disappearance of its UV absorption
band at l=276 nm. By using the 3D electrode, the conversion
of maleimide was extremely rapid, so 95% conversion was ach-
ieved in just 400 s, compared with 60 and 100% conversions
in 5 h for 2D VC and BDD electrodes, respectively. From an ini-
tial maleimide concentration of 0.1m, the reaction was essen-
tially complete within 2400 s, during which time 3645 C of
charge was passed. Because the theoretical charge required for
the two-electron reduction was 1929 C, this corresponded to
a charge yield of about 50%, but would have been greater at
shorter times/lower conversions.
Entry Substrate Cathode t [h] Conversion [%][b] Charge yield [%], Fep[c]
1
2
3
4
5
6
7
1a
1a
1b
1b
1c
1d
1e
VC
BDD
VC
BDD
BDD
BDD
BDD
5
5
5
5
5
5
5
60
99
79
>99
92
88
62
96
79
85
45
45
41
>99
[a] 10 mm of maleimide in 1m H2SO4 (100 mL) was recirculated through
the electrochemical reactor at 60 mLminꢀ1 at room temperature. The ap-
plied potential of ꢀ1.2 V (AgCljAg). [b] Determined by 1H NMR spectros-
copy, following neutralization with an aqueous solution of NaOH, and ex-
traction with CH2Cl2. [c] Ratio of the theoretical charge (Qp) passed for
product formation to the total charge (Q) passed during electrolysis.
Raising the temperature to 508C increased the kinetics, with
complete conversion of the maleimide to succinimide achieved
in 300 s. This may be attributed to an increased diffusion coef-
ficient for maleimide, and hence, increased transport-con-
trolled current densities at the graphite cathode. However, it is
worth noting that the range of potentials in such 3D electro-
des is also sensitive to electrode and electrolyte conductivities,
which are also temperature dependent.
using a BDD electrode,[14] 99% conversion was obtained with
a charge yield of 96% (Table 2, entry 2) within the same time.
This significant enhancement in performance can be attributed
to the higher overpotential for hydrogen evolution at BDD
compared with VC, so inhibiting the competitive reduction of
protons to H2.
The chemoselectivity of the reduction process was deter-
mined subsequently for substrates 1b and 1c, which con-
tained N-allyl and -propargyl substituents, respectively. With
a cathode potential of ꢀ1.2 V (vs. SCE), reduction of the elec-
tron-deficient cyclic olefin was achieved with complete chemo-
Chemoselectivity
Chemoselectivity of the electrochemical reduction of malei-
mide derivatives was reported to be highly dependent upon
pH (Scheme 2).[8,9] At pH values below 7, a two-electron reduc-
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
667 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 4. Reduction of 3,4-dimethylmaleimides under different electrolysis
conditions.[a]
Entry
Compound
Electrode
Conversion[b]
cis/trans[c]
1
2
3
1d
1d
1e
2D BDD[d]
3D graphite[e]
2D BDD[d]
99
94
99
2:3
2:3
2:3
[a] 10 mm substrate in 1m H2SO4/MeOH (4:1 v/v; 100 mL). Potential of
ꢀ1.2 V versus AgCljAg applied. [b] Determined by UV/Vis spectroscopy.
1
[c] Determined by H NMR spectroscopy. [d] 5 h. [e] 30 min.
Scheme 2. The pH-dependent chemoselectivity of the electrochemical re-
duction of 1a.
ence for the trans isomer, that is, the product distribution was
not dependent on either the electrode surface area (Table 4,
entries 1 and 2) or the nitrogen substituent (Table 4, entry 3).
It is generally accepted that the reduction proceeds through
a CE (chemical–electrochemical) mechanism, in which two-
electron transfer and the addition of H+ occur to generate the
transient enol intermediate I (Scheme 2), which then under-
goes tautomerization to form the succinimide. In the case of
3,4-disubstituted substrates (1d and 1e), we postulated that
the process occurred in a stepwise manner, such that the ste-
reoselectivity was determined by the isomerization of a half-
enol intermediate III (Scheme 3).
tion wave was detected; this resulted in the formation of the
corresponding succinimides (2). At higher pH (7–10), stepwise
one-electron reduction occurred, leading to dimerized prod-
ucts 3. At even higher pH values (>10), hydrolysis of the male-
imide becomes a competitive process.
The chemoselectivity of the reduction process towards dif-
ferent unsaturated carbon–carbon bonds was assessed by de-
termining the reduction of several maleimide derivatives that
contained N-allyl (1b), -propargylic (1c), and -benzyl (1e) sub-
stituents (Table 3). In certain cases, the addition of methanol
Table 3. Chemoselective reduction of maleimide derivatives by using
a 3D graphite electrode.[a]
Entry[a] Compound Applied potential [V][b] t [h][c] Conversion [%][d]
1
2
3
4
5
1a
1b
1c
1d
1e
ꢀ1.2
ꢀ1.1
ꢀ1.0
ꢀ1.2
ꢀ1.2
0.1
0.5
0.5
0.5
0.5
95
92
95
94
95
[a] General procedure: 10 mm of the maleimide substrate in 100 mL of
1m H2SO4 or 1m H2SO4/MeOH (4:1 v/v). Flow rate=60 mLminꢀ1. [b] Value
in V versus AgCljAg. [c] Monitored by UV/Vis spectroscopy. [d] Deter-
Scheme 3. Stereodefining step in the reduction of 3,4-disubstituted male-
imides.
1
mined by H NMR spectroscopy.
The enolization of N-methylsuccinimide under pH-neutral
conditions was modeled previously by DFT computations
using the B3LYP/6-31+G** basis set,[17] in which it was pro-
posed that two water molecules were intimately involved in
a triple hydrogen-atom transfer process. Herein, the model
was adopted initially with slight modifications: by using the
WB97XD/6-31G(d,p) basis set to account for dispersion, en-
hanced with a continuum solvation model (CPCM=water).
Two sets of diastereotopic eight-membered cyclic transitions
states were constructed for the half-enol III generated from
substrate 1d, from which the energetic pathways leading to
the cis and trans isomers could be plotted (Figure 3, top
graph). In this model (triple hydrogen transfer), a 4 kJmolꢀ1 ki-
netic bias was observed for the trans isomer accompanied by
a significant activation energy barrier; this does not agree with
our experimental conditions/observations. However, by intro-
ducing a proton to the system, the overall energetic pathway
was transformed greatly. In this case, only two hydrogen
atoms were transferred (Figure 3, bottom graph). The sponta-
neity of the process is supported by a low activation energy
cosolvent was necessary to enable these compounds to dis-
solve in the electrolyte solution. As expected, dimerization of
the maleimide was not detected under acidic conditions.[10c]
More significantly, exclusive reduction of the cyclic C=C bond
was attained at ꢀ1.2 V, whereas the N-substituents remained
unaffected. This is particularly interesting because such orthog-
onal reactivity is difficult, if not impossible, to achieve by cata-
lytic hydrogenation.
Stereoselectivity
The electrochemical reduction of N-ethyl-3,4-dimethylmalei-
mide was described briefly to produce both cis (meso) and
trans (dl) diastereomers in approximately equal amounts at
low pH.[10c] With this in mind, the stereoselectivity of this trans-
formation was examined by comparing the reduction of 3,4-di-
methyl-substituted maleimides 1d and 1e under different
electrolysis conditions (Table 4). In all cases, a mixture of the
diastereomers was obtained in a 2:3 ratio, with a slight prefer-
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
668 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ed maleimides is attributed to a solution process, through an
acid-catalyzed tautomerization of a half-enol; the kinetic profile
is commensurate with a double hydrogen atom transfer pro-
cess. Preliminary theoretical calculations suggested that the se-
lectivity was highly dependent on the nature of the C3 and C4
substituents; this will be studied in our future work.
Experimental Section
General
Commercial reagents were used as received. High-purity water was
prepared by reverse osmosis and deionization. Solvents were dried
over molecular sieve columns in a solvent purification system.
Column chromatography was performed on flash silica gel (Kiesel-
gel 60, 63–200 mm). The pH of solutions was measured by using
a Hanna pH probe.
1H and 13C NMR spectra were obtained by using 400 MHz instru-
ments (1H at 400 MHz and 13C at 100 MHz). Chemical shifts are re-
ported in d (ppm) and referenced to residual protons and 13C sig-
nals in deuterated chloroform. The coupling constants (J) are ex-
pressed in Hertz (Hz). Multiplicities are indicated as singlet (s), dou-
blet (d), triplet (t), and multiplet (m).
IR spectra were recorded for solid or solution samples by using
a PerkinElmer FTIR spectrometer fitted with an attenuated total re-
flectance (ATR) accessory. Mass spectra were recorded on a Micro-
mass Autospec-Q mass spectrometer (EI and CI sources). Melting
points (uncorrected) were determined by using an Electrothermal
Gallenhamp apparatus and a calibrated thermometer (ꢂ28C). GC-
MS results were recorded by using an HP GCD C GC-MS instrument
and a Zebron Wax Plus column (Phenomenex).
Figure 3. Computed energetic pathways for the tautomerization of half-enol
to cis- (*) and trans-succinimide (*) under neutral, triple-hydrogen transfer
(top) and acidic, double-hydrogen transfer (bottom) conditions.
HPLC was conducted by using a HP 1050 HPLC instrument and
a Gemini NX C-18 column (Phenomenex). UV/Vis spectra were re-
corded by using an Agilent 8453 spectrophotometer, HELLMA fiber
optic cables, and flow-through cuvette. Reaction progress was
monitored by the disappearance of the absorbance band of the
substrate at l=276 nm.
barrier and, even more significantly, the formation of the trans
isomer is predicted to be favored by 1.6 kJmolꢀ1; this corre-
sponds to a selectivity of 1.9:1,[18] which may be considered to
be in good agreement with the experimentally observed value.
In theory, lowering the pH of the solution should increase the
selectivity towards the trans isomer, but this is prohibited by
the competitive dimerization process to form 3 (Scheme 2).
To investigate the kinetics of the charge-transfer reactions, a VC
RDE with a PTFE-sheathed 5 mm diameter (1.97ꢁ10ꢀ5 m2; Pine In-
strument Company, USA) was used, and rotated at 3–50 sꢀ1 by
means of an AFMSRCE rotator controller (Pine Research Instru-
ments, USA). A three-compartment glass cell was used for the ro-
tating disc experiments that incorporated a vitreous disc electrode;
a saturated calomel reference electrode (Monocrystally), which was
assumed to have a potential of 0.254 V (SHE); and a platinum flag
counter electrode.
Conclusions
We demonstrated that the reduction of the cyclic C=C bond in
maleimide to form succinimide could be achieved exclusively
in the presence of functional groups that were otherwise sus-
ceptible to competitive reduction under the conditions of cata-
lytic hydrogenation. The reduction potential of these com-
pounds was dependent on the nitrogen substituent, as well as
on substituents at the C=C bond (Table 1).
A glass cell was used for batch electrolysis that incorporated
a Nafion 117 (DuPont Inc.) cation-permeable membrane to sepa-
rate the anolyte and cathode. A VC electrode (0.001 m2) was used
as the cathode, a platinized titanium mesh as the anode, and a sa-
turated calomel electrode (Monocrystally) as the reference elec-
trode; the cathode was polished prior to each experiment. The
anolyte and catholyte volumes were 270 mL; the former was a 1m
aqueous solution of H2SO4 and the latter was 0.01m of the amide
substrate with a 1m aqueous solution of H2SO4 as the supporting
electrolyte. High-purity nitrogen was bubbled through the catho-
lyte to remove dissolved oxygen and to enhance mass transport
rates.
The reactor-scale kinetics could be increased significantly by
using a 3D graphite electrode (Table 3), although the competi-
tive reduction of water was a problem in some of these sys-
tems (particularly 1d and 1e), in which there was a significant
decrease in charge yield. We predict that 3D BDD electrodes
would solve this issue; this will be investigated in our future
work. The stereoselectivity for the reduction of 3,4-disubstitut-
An Autolab PGSTAT302N potentiostat (Metrohm Autolab) instru-
ment was used to apply the required cathode potential and oper-
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
669 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
ated by using Nova software to monitor the resulting currents for
experiments with the RDE and electrochemical reactor, typically at
a range of scan rates (10 to 25 mVsꢀ1).
evaporated and the product was purified by column chromatogra-
phy.
N-Propargylmaleimide (1c):[20] Yellow oil (0.53 g, 13%); Rf =0.4
(hexane/EtOAc, 7:3); IR (ATR): n˜ =3280, 3100, 2961, 1775, 1600,
A commercial electrochemical reactor (Electrocell) was modified to
accommodate a 10 mm thick 46ꢁ96 mm2 PTFE frame, which creat-
ed a cathode compartment with a volume of 18.9 mL and an effec-
tive electrode geometric area of 103 mm2. An additional compart-
ment was bolted onto the side of the reactor and designed to con-
tain a AgCljAg wire reference electrode, which was connected hy-
draulically to the catholyte with a 0.5 mm hole drilled through the
frame, and emerged close to the cathode surface (Figure 4). VC or
1485, 1106, 805, 643 cmꢀ1 1H NMR (400 MHz, CDCl3): d=6.73 (s,
;
2H), 4.21 (d, J=2 Hz, 2H), 5.17–5.1 ppm1 (t, J=2 Hz, 1H); 13C NMR
(100 MHz, CDCl3): d=169.3, 134.5, 76.9, 71.6, 26.7 ppm; MS (EI): m/
z (%): 136 (7) [M]+, 135 (82), 107 (100), 79 (30), 54 (75), 52 (48).
3,4-Dimethylmaleimide (1d): White solid (1.62 g, 83%); Rf =0.4
(CH2Cl2); m.p. 109–1128C (lit.[21] 111–1138C); IR (ATR): n˜ =3242,
1723, 1654, 1453, 1256, 1038, 904, 695, 650 cmꢀ1
;
1H NMR
(400 MHz, CDCl3): d=8.15 (s, 1H), 1.96 ppm (s, 6H); 13C NMR
(100 MHz, CDCl3): d=172.1, 138.3, 8.6 ppm; MS (EI): m/z (%): 126
(12) [M]+, 125 (100), 82 (10), 54 (68), 53 (30).
N-Benzyl-3,4-dimethylmaleimide (1e): Yellow solid (2.23 g, 66%);
Rf =0.4 (hexane/EtOAc, 7:3); m.p. 43–468C (lit.[22] 44–458C); IR
(ATR): n˜ =3026, 1650, 1584, 1452, 1418, 1390, 1035, 963, 706,
1
650 cmꢀ1; H NMR (400 MHz, CDCl3): d=7.25–7.35 (m, 5H), 4.66 (s,
2H), 1.97 ppm (s, 6H); 13C NMR (100 MHz, CDCl3): d=171.8, 137.3,
136.7, 129.3, 128.6, 128.4, 127.7, 41.5, 8.7 ppm; MS (EI): m/z (%):
217 (2) [M]+, 216 (25), 215 (100), 187 (28), 172 (35), 104 (35), 91
(32), 77 (15), 54 (17).
Electrochemistry
The electrochemical cell was operated in batch recycling mode by
using a Masterflex peristaltic pump (Model 07524-40) to supply
flow rates of up to 90 mLminꢀ1 through the catholyte compart-
ment. Both electrode compartments were filled with 1m H2SO4
unless otherwise stated. The glass catholyte and anolyte reservoirs
had a volume of 100 mL each.
Figure 4. Modified catholyte chamber holding the AgjAgCl reference elec-
trode.
BDD were used as the cathode materials, with Ti/Ta2O5ꢀIrO2 as the
anode because of its low overpotential for oxygen evolution and
stability in acidic solutions. Nafion 117 ion-permeable membrane
separated the anode and cathode compartments to prevent malei-
mides or their reduction products reaching the anode, at which
otherwise they would have been oxidized. Hydrogen gas evolution
rates were measured by using a Ritter MGC-1 gas counter.
At time t=0, recirculation of both cell compartments was estab-
lished, the operating potential was applied and the resulting cur-
rent was monitored continuously. Experiments were concluded
after a maximum of 5 h, depending on the progress of the reduc-
tion process. The catholyte was neutralized by the addition of 1m
NaOH, and extracted with dichloromethane. The combined organic
extracts were dried over MgSO4 and the solvent was evaporated.
The residue was analyzed by NMR spectroscopy.
Preparation of 1b by a Mitsunobu reaction[16,19]
PPh3 (7.98 g, 30.6 mmol), maleimide (3.00 g, 31.2 mmol), and diiso-
propyl azodicarboxylate (DIAD; 6.9 g, 34.2 mmol) were added suc-
cessively to a solution of allyl alcohol (1.65 g, 28.4 mmol) in THF
(225 mL). The reaction mixture was left to stir overnight. The sol-
vent was then evaporated and the product was isolated by
column chromatography (5:1 hexane/EtOAc; Rf =0.5) to give 1b as
a yellow solid (1.72 g, 44%). M.p. 41–438C (lit.[19] 40–428C); IR
(ATR): n˜ =3100, 3058, 2985, 1720, 1600, 1475, 1430, 1206, 900, 862,
The NMR spectroscopy data for 2a,[23] 2b,[24] and 2c[25] are consis-
tent with values reported in the literature.
Compounds 2d and 2e were obtained as a mixture of diastereo-
isomers
3,4-Dimethylsuccinimide 2d: cis isomer:[26] 1H NMR (400 MHz,
CDCl3): d=8.52 (brs, 1H), 3.00–2.90 (m, 2H), 1.18 ppm (d, J=
7.0 Hz, 6H); trans isomer:[27] 1H NMR (400 MHz, CDCl3): d=9.35
(brs, 1H), 2.49–2.41 (m, 2H), 1.31 ppm (d, J=7.0 Hz, 6H).
N-Benzyl-3,4-dimethylsuccinimide 2e: cis isomer:[28] 1H NMR
(400 MHz, CDCl3): d=7.34–7.21 (m, 5H), 4.59 (s, 2H), 2.99–2.91 (m,
2H), 1.16 ppm (d, J=7.0 Hz, 6H); trans-isomer:[6b] 1H NMR
(400 MHz, CDCl3): d=7.34–7.21 (m, 5H), 4.59 (s, 2H), 2.49–2.40 (m,
2H), 1.29 ppm (d, J=7.0 Hz, 6H).
1
635 cmꢀ1; H NMR (400 MHz, CDCl3): d=6.71 (s, 2H), 5.82–5.72 (m,
1H), 5.17–5.11 (m, 2H), 4.12–4.06 ppm (m, 2H); 13C NMR (100 MHz,
CDCl3): d=170.3, 134.2, 131.5, 117.6, 39.8 ppm; MS (EI): m/z (%):
138 (7) [M]+, 137 (100), 109 (15), 95 (23), 55 (40), 43 (78).
General procedure for the preparation of maleimide deriva-
tives 1c–e
Propargylamine (1.69 g, 30.6 mmol), ammonium acetate (1.81 g,
23.4 mmol), or benzylamine (2.51 g, 23.4 mmol) were added to a so-
lution of maleic anhydride or dimethylmaleic anhydride (2.0 g) in
acetic acid (50 mL) and the reaction mixture was heated at reflux
for 16 h. After cooling to room temperature, the solvent was
DFT calculations
DFT calculations were performed by using the Gaussian09 software
package.[29] Reactants and products were optimized by using the
WB97XD/6-31G(d,p) level of theory. Transition states were opti-
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
670 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Lund, J. Electroanal. Chem. Interfacial Electrochem. 1986, 202, 299–302;
c) P. H. Zoutendam, P. T. Kissinger, J. Org. Chem. 1979, 44, 758–761.
[11] M. Sortino, A. Postigo, S. Zacchino, Molecules 2013, 18, 5669–5683.
[12] a) J. H. T. Luong, K. B. Male, J. D. Glennon, Analyst 2009, 134, 1965–
1979; b) Y. V. Pelskov, A. Y. Sakharova, M. D. Krotova, L. L. Bouilov, B. V.
Spit syn, J. Electroanal. Chem. 1987, 228, 19–27.
[13] Electrocell A/S, Vennelystvej 1, DK 6880 Tarm, Denmark; see: http://
www.electrocell.com/micro-flow-cell.html.
[14] Procured from ElectroCell (Ref. [13]). The electrode comprised of a con-
ductive diamond layer deposited at high temperature/pressure on
a 2 mm niobium base plate.
mized by using either QST2 or the Berny algorithm, which was fre-
quency checked, followed by intrinsic reaction coordinate (IRC) cal-
culations. Output files have been deposited on a server and are
available through the following links: 1) cis-succinimide (neutral):
http://dx.doi.org/10.6084/m9.figshare.928138; 2) trans-succinimide
(neutral): http://dx.doi.org/10.6084/m9.figshare.928139; 3) cis-succi-
nimide (acid catalyzed): http://hdl.handle.net/10042/132753;
4) trans-succinimide (acid catalyzed): http://hdl.handle.net/10042/
132754
[15] With the reaction operating under mass transport control in a reactor
operating in batch recycling mode, reaction rates would decrease due
to a decreased diffusion coefficient (Di), caused by an increased hydro-
dynamic diameter of a substituted reactant. Hence, at a fixed cathode
hydrogen overpotential/electrode potential, a steady-state current den-
sity due to hydrogen evolution would decrease charge yields/current
efficiencies.
Acknowledgements
We thank the EPSRC for a Doctoral Training Award (to K.R.) and
the Pharmacat Consortium (AstraZeneca, GlaxoSmithKline, and
Pfizer) for additional financial support. K.K.H. is grateful to Prof.
Henry Rzepa (Imperial College London) and Dr. Alexandra Sim-
perler (EPSRC UK National Service for Computational Chemistry
Software, NSCCS) for their help with the theoretical calculations.
[16] For example, see: C. Y. Cheng, G. H. Kelsall, D. Pilone, J. Appl. Electro-
chem. 2005, 35, 1191–1202.
[17] O. Takahashi, K. Kobayashi, A. Oda, Chem. Biodiversity 2010, 7, 1349–
1356.
[18] DDG° =ꢀRTln(trans/cis).
[19] M. A. Walker, Tetrahedron Lett. 1994, 35, 665–668.
[20] a) C. D. Mayer, M. Kehrel, F. Bracher, Org. Prep. Proced. Int. 2008, 40,
574–579; b) M. Dubernet, V. Caubert, J. Guillard, M.-C. Viaud-Massuard,
Tetrahedron 2005, 61, 4585–4593.
Keywords: chemoselectivity
electrochemistry · reaction mechanisms · reduction
·
diastereoselectivity
·
[21] G. B. Gill, G. D. James, K. V. Oates, G. Pattenden, J. Chem. Soc. Perkin
Trans. 1 1993, 2567–2580.
[22] F. Abe, M. Sumoto, Chem. Pharm. Bull. 1991, 39, 1167–1170.
[23] D. W. Manley, R. T. McBurney, P. Miller, R. F. Howe, S. Rhydderch, J. C.
Walton, J. Am. Chem. Soc. 2012, 134, 13580–13583.
[24] B. Alcaide, P. Almendros, J. M. Alonso, Chem. Eur. J. 2006, 12, 2874–
2879.
[25] C. M. Marson, A. Khan, R. A. Porter, J. Org. Chem. 2001, 66, 4771–4775.
[26] J. M. Burns, M. E. Ashley, G. C. Crockett, T. H. Koch, J. Am. Chem. Soc.
1977, 99, 6924–6928.
´
[1] L. Cerveny, Catalytic Hydrogenation, Elsevier, Amsterdam, 1986.
[2] a) G. Eigenberger, U. Wegerle, ACS Symp. Ser. 1982, 196, 133–143; b) M.
Irfan, T. N. Glasnov, C. O. Kappe, ChemSusChem 2011, 4, 300–316.
[3] D. J. Pasto in Comprehensive Organic Synthesis (Eds.: B. M. Trost, I. Flem-
ing), Pergamon, Oxford, 1991, pp. 471–488.
[4] For a discussion on issues associated with the scaleup of dissolving
metal reduction, see: D. J. Joshi, J. W. Sutton, S. Carver, J. P. Blanchard,
Org. Process Res. Dev. 2005, 9, 997.
[5] For recent examples and review, see: a) A. Attour, P. Dirrenberger, S.
Rode, A. Ziogas, M. Matlosz, F. Lapicque, Chem. Eng. Sci. 2011, 66, 480–
489; b) Y. Y. Hosokawa, H. Hakamata, T. Murakami, F. Kusu, Tetrahedron
Lett. 2010, 51, 129–132; c) A. Ziogas, G. Kolb, M. O’Connell, A. Attour, F.
Lapicque, M. Matlosz, S. Rode, J. Appl. Electrochem. 2009, 39, 2297–
2313; d) T. Kashiwagi, F. Amemiya, T. Fuchigami, M. Atobe, J. Flow Chem.
2013, 3, 17–22; e) K. Watts, A. Baker, T. Wirth, J. Flow Chem. 2014, 4, 2–
11.
[6] This can be achieved by Mg/MeOH (dissolving metal), see: a) M. Katoh,
K. Dodo, M. Fujita, M. Sodeoka, Bioorg. Med. Chem. Lett. 2005, 15,
3109–3113; for enantioselective reductions by biocatalysis, see: b) M.
Sortino, S. A. Zacchino, Tetrahedron: Asymmetry 2010, 21, 535–539;
c) M. F. Hegazy, K. Shishido, T. Hirata, Tetrahedron: Asymmetry 2006, 17,
1859–1862.
´
[27] T. Połonski, J. Chem. Soc. Perkin Trans. 1 1988, 629–637.
[28] T. Punniyamurthy, T. Katsuki, Tetrahedron 1999, 55, 9439–9454.
[29] Gaussian 09, Revision D.01, M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.
Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-
nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,
A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro,
M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-
bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-
gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.
Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,
K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,
S. Dapprich, A. D. Daniels, ꢂ. Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-
slowski, D. J. Fox, Gaussian, Inc., Wallingford CT, 2009.
[7] B. Sakurai, Bull. Chem. Soc. Jpn. 1937, 12, 8–11.
[8] a) R. G. Barradas, S. Fletcher, J. D. Porter, Can. J. Chem. 1976, 54, 1400–
1404; b) R. G. Barradas, S. L. Gust, J. D. Porter, Tetrahedron Lett. 1981, 22,
4579–4582.
[9] R. G. Barradas, S. Fletcher, J. D. Porter, J. Electroanal. Chem. 1977, 75,
533–543.
Received: October 12, 2014
[10] a) H. M. Fahmy, B. Nashed, F. A. Khalifa, A. A. Moneim, M. A. Azzem, J.
Electroanal. Chem. Interfacial Electrochem. 1985, 184, 147–160; b) H.
Revised: November 30, 2014
Published online on January 8, 2015
ChemSusChem 2015, 8, 665 – 671
www.chemsuschem.org
671 ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim