Mechanistic aspects of aldehyde and imine electro-reduction in a liquid-liquid carbon nanofiber membrane microreactor

Article abstract of DOI:10.1016/j.tetlet.2012.04.092

A simple and electrolyte-free ion-transfer electrosynthesis micro-reactor system (volume 100 μL, up to 10 mg batches) for processes at liquid-liquid interfaces is developed and demonstrated for the reduction of aldehydes and imines. These cathodic reactions occur at an amphiphilic carbon nanofiber membrane accompanied by proton cation transfer from an aqueous phase into an organic phase.

Full text of DOI:10.1016/j.tetlet.2012.04.092

Tetrahedron Letters 53 (2012) 3357–3360
Contents lists available at SciVerse ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Mechanistic aspects of aldehyde and imine electro-reduction in a
liquid–liquid carbon nanofiber membrane microreactor
⇑
John D. Watkins , James E. Taylor, Steven D. Bull, Frank Marken
Department of Chemistry, University of Bath, Claverton Down, Bath BA2 7AY, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A simple and electrolyte-free ion-transfer electrosynthesis micro-reactor system (volume 100 lL, up to
Received 20 February 2012
Revised 31 March 2012
Accepted 19 April 2012
Available online 27 April 2012
10 mg batches) for processes at liquid–liquid interfaces is developed and demonstrated for the reduction
of aldehydes and imines. These cathodic reactions occur at an amphiphilic carbon nanofiber membrane
accompanied by proton cation transfer from an aqueous phase into an organic phase.
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Hydrogenation
Electro-reduction
Electrosynthesis
Carbon nanofiber membrane
Microreactor
As a methodology in organic synthesis, electrosynthesis1 offers
a variety of benefits over other procedures. Reactions can be con-
trolled with a potential bias, or conducted under mild conditions
with strong acids or bases generated in situ.2 Much attention has
been paid to single phase electrosynthetic reactions where a con-
siderable concentration of electrolyte is required in the organic
media. Water is an ideal electrochemical medium due to its high
dielectric constant, but is often dismissed as a potential solvent
due to its incompatibility with many organic processes. However,
electrochemical processes occurring in organic media in simulta-
neous contact to an aqueous electrolyte phase and an electrode
(a ‘triple phase boundary’) provide an alternative approach to car-
rying out preparative electrosynthesis.3 It has been found that re-
dox processes can only occur at the contact line of both liquid
phases and electrode and new approaches have been investigated
to increase the size of the triple phase boundary reaction zone. Re-
cent studies have extended the methodology to include a range of
water immiscible carrier phases,4 dual phase flow channels,5–7
acoustic emulsions,8,9 high shear force emulsions,10 and most re-
cently an amphiphilic carbon nanofiber membrane-based microre-
actor.11 Another benefit of the triple phase boundary synthetic
approach is similar to the biphasic approach used by Amemiya
et al.12,13 for biphasic systems. In the triple phase boundary process
however, no electrolyte is required in the product phase leading to
cleaner syntheses with no need for an electrolyte separation step.
Traditionally hydrogenation reactions have been carried out
using molecular hydrogen and a suitable catalyst, or via transfer
hydrogenation reactions.14 Examples of the electro-reduction of
unsaturated alkenes to their corresponding alkanes using a medi-
ated approach15 and single phase reactions have recently been re-
viewed.16 The electro-reduction of aldehydes and ketones is also of
considerable interest due to the possibility of using different
experimental conditions and reactor design to control product dis-
tribution. The radical mechanism involved in these processes leads
17
primarily to coupled diol products as found by Bian and Bai
although competing reaction pathways that lead to alcohol and/
or condensation products also took place. Kise et al. reported the
electro-reductive acylation of aromatic ketones using acylimidaz-
oles and chlorotrimethylsilane to form
a-trimethylsiloxy ketones
and esters.18 They have also demonstrated the inter- and intramo-
lecular electro-reductive cyclization of dicarbonyl compounds to
form trans-1,2-diols.19 Atobe et al. have since shown that the use
of acoustic emulsification can be used to change the diasteromeric
selectivity of aldehyde reduction products.20
An alternative electrochemical strategy to the direct reduction
of carbonyl compounds followed by the abstraction of hydrogen
is the electrogeneration of hydrogen at an electrode surface prior
to its addition to the carbonyl compound. The challenge in achiev-
ing good yields using this approach is to find an electrode material
that is suitable for the efficient evolution of hydrogen that is also
capable of catalyzing hydrogenation of the carbonyl group. Santana
et al. reported the efficient hydrogenation of a range of alkene and
carbonyl substrates using conditions that were optimized for
hydrogen evolution at a nickel deposit on iron electrodes.21
⇑
Corresponding author at present address: Frick Chemistry Laboratory, Princeton
University, Princeton, NJ 08544, USA.
E-mail address: jdw4@princeton.edu (J.D. Watkins).
0040-4039/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.tetlet.2012.04.092

3358
J. D. Watkins et al. / Tetrahedron Letters 53 (2012) 3357–3360
In this study the bulk electrolysis of aromatic aldehydes and imi-
nes is reported in an amphiphilic carbon nanofiber microreactor.22
Figure 1A shows a photograph of the 100 L reactor in a glass cap-
(3–5B) products (Table 1, entries 3–5). Pleasingly, the reduction
also worked well in the presence of a chloroaryl-substituent, with
no additional side products observed from competing chloro group
reduction (Table 1, entry 6). However, the reduction potential re-
quired for para-methoxybenzaldehyde was found to be too nega-
tive to allow reduction to occur (Table 1, entry 7). It was also
found that para-nitro and para-cyano substituents were incompat-
ible with these electro-reductive conditions, affording a complex
mixture of products.
Finally, the effect of the concentration of the inert supporting
electrolyte in the aqueous phase was investigated. Increasing the
concentration of Na2SO4 to 2 M was found to give a pinacol (A)
to alcohol (B) ratio of 1.0:1.7, whilst decreasing the concentration
down to 0.1 M Na2SO4 resulted in a reversal in product ratio
(1.4:1.0) in favor of the pinacol product.
The electro-reduction of imines to amines or vicinal diamines in
organic/aqueous biphasic media would provide a useful synthetic
tool for the controlled reduction of hydrolytically stable imines.
Therefore, electro-reductions of a small series of para-trifluoro-
methyl-benzalimines25 were carried out, the results of which are
presented in Table 1 (entries 9–11). It was found that the benzyl-
imine underwent a selective 2-electron reduction to afford the
corresponding amine, although only in 11% conversion after 2 h
(Table 1, entry 9). The overall conversion could be improved to
around 50% by increasing the reaction time to 4 h (Table 1, entry
10). A faster reaction was observed for the reduction of the isopro-
pyl-imine derivative, with complete conversion into products
observed within 4 h. However, this reduction was less selective,
with the amine and diamine reduction products being formed in
a 1:1 ratio (Table 1, entry 11).
The ratio of products produced in the aldehyde reduction reac-
tions (pinacol (A) versus alcohol (B) products) was surprisingly
independent of the composition and pH of the aqueous phase used.
The reaction conditions within the carbon nanofiber membrane are
likely to be dominated by the nature of the ‘triple phase boundary’
zone.26 Based on the known solubilities of the organic reactants,
the reaction should occur predominantly in the organic phase.
However, in the region close to the interface, fast transport of pro-
tons is possible and the formation of the 2-electron product is
more likely (see ‘monomer zone’ in Fig. 2). The ‘proton availability’
reflects the ability of protons to transfer from the aqueous phase
into a thin zone within the organic liquid during the reduction.
At distances further away from the triple phase boundary reaction
zone, electrochemical reduction at the carbon nanofiber electrode
is still possible, but the formation of a 1-electron dimer can com-
pete and dominate (see ‘dimer zone’ in Fig. 2).
l
illary with the carbon nanofiber membrane held in place by sili-
cone. Once immersed into the aqueous phase (see Fig. 1B) an
organic liquid is added and a stable interface forms for triple phase
boundary electrolysis. The inclusion of a high speed micro-propel-
ler ensures effective mixing and transport within the organic phase.
In the triple phase boundary mechanism the redox active spe-
cies is only reduced at the boundary between solid electrode, aque-
ous electrolyte, and organic carrier phases with the concerted
insertion of one proton per electron to maintain charge neutrality
(Fig. 1C). Initially, the electro-reduction reaction of benzaldehyde
derivatives was undertaken under triple phase boundary condi-
tions. Voltammetry experiments (not shown) were conducted for
160 mM benzaldehyde in acetonitrile in contact with a 1 M Na2-
SO4 + 0.1 M PBS pH7 aqueous phase and suggest that the onset of
reduction was at ca. À1.0 V versus SCE with a broad peak at ca.
À1.9 V versus SCE,23 meaning that a preparative electrolysis was
conducted at À1.9 V versus SCE.24 When the products of the benz-
aldehyde reduction were analyzed by 1H NMR spectroscopy, it was
found that although the conversion from starting material was
essentially 100%, two distinct products had been formed (see Ta-
ble 1). These were assigned as benzhydrol (1B, 2-electron product)
and the meso-pinacol (1A, 1-electron product) as reported previ-
ously by Bian and Bai.17 When changing the benzaldehyde concen-
tration from 320 to 80 to 40 mM the product ratio changed from
1.0:1.8 to 1.0:1.4 to 1.0 to 1.0 respectively, indicative of more
dimerization occurring at higher concentrations. In an attempt to
control the pinacol-to-alcohol ratio the pH of the aqueous phase
was varied, however even in 1 M Na2SO4 + 1 M H3PO4 the pina-
col-to-alcohol ratio only reached 1.4:1.0. Therefore, the composi-
tion and pH of the aqueous phase appear to have a surprisingly
insignificant effect on the product ratio of this electro-reductive
reaction. A similar experiment using toluene as solvent instead of
acetonitrile gave very similar results in terms of yield and product
distribution (see Table 1).
Further investigations were then carried out to determine the
functional group compatibility of this electro-reduction reaction
using various aromatic aldehydes as substrates (Table 1). The
reduction of para-trifluoromethylbenzaldehyde proceeded well,
giving complete conversion in favor of the 2-electron reduction
product (2B) and with a lower reduction potential being required
(Table 1, entry 2). Reduction of para-, meta-, and ortho-tolualde-
hyde required more negative potentials to proceed giving slightly
reduced conversions in favor of their corresponding alcohol
Figure 1. (A) Photograph of the tubular microreactor. (B) Schematic drawing of the microreactor capillary with organic phase and micro-propeller immersed with the
amphiphilic carbon nanofiber membrane separating the aqueous electrolyte phase. (C) Schematic drawing of the overall interfacial process during reduction in the organic
phase.

J. D. Watkins et al. / Tetrahedron Letters 53 (2012) 3357–3360
3359
Table 1
Summary data for the reduction of substituted benzaldehydes (160 mM) in a liquid–liquid carbon membrane microreactor in an organic phase of 100
phase of 1 M Na2SO4 + 0.1 M PBS pH 7
l
L MeCN and an aqueous
X
(A)
XH
R
Low proton availability
-
Far from triple phase boundary
R
R
2 H+
XH
X
X
aq: 1 M Na2SO4
+ 0.1 M PBS pH 7
XH
(B)
R
R
2 h
High proton availability
-
1 e-
1 e-, 2 H+
Near to triple phase boundary
R
X = O, NBz, NiPr
Entry
1
Reactant
Conditions
MeCN/2 h
Triple phase reduction potential versus SCE
Ratioa A:B
1.0:1.8
Conversionb
100%
O
À1.90 V
O
2
3
MeCN/2 h
MeCN/2 h
À1.80 V
À2.10 V
1.0:1.7
1.0:3.1
100%
66%
F3C
O
O
4
MeCN/2 h
À2.10 V
1.0:1.8
94%
O
O
5
6
MeCN/2 h
MeCN/2 h
À2.10 V
À2.10 V
1.0:1.7
1.0:1.8
70%
99%
Cl
O
7
8
MeCN/2 h
À2.10 V
À1.85 V
—
Negligible
100%
MeO
O
Toluene/2 h
1.0:1.6
N
Ph
Ph
9
MeCN/2 h
MeCN/4 h
À1.70 V
À1.70 V
Amine only
Amine only
11%
50%
F3C
N
10
F3C
N
11
MeCN/4 h
À1.85 V
1.0:1.0
100%
F3C
The initial reduction current was found to decay rapidly to a limiting current of 1–1.5 mA giving ca. 7–11 C of charge flow in a 2 h electrolysis. This value represents an excess
charge of 2 or 3 times compared to the required current for the 2 electron reduction of 160 mM of aldehyde or imine in 100 lL of solvent.
a
Product distribution calculated from integration of resonances arising from diol (2H) and alcohol (2H) at ca. 4.75 and 4.6 ppm in their 1H NMR spectra, respectively for
aldehydes and from integration of resonances arising from diamine (2H) and amine (2H) at ca. 4.80 and 4.65 ppm in their 1H NMR spectra, respectively for imines.
b
Conversion calculated from comparison of integration of the aldehyde peak (1H) at ca. 10 ppm and the new singlet peaks for diol (2H) and alcohol (2H) at ca. 4.75 and
4.6 ppm in their 1H NMR spectra, respectively for aldehydes or from comparison of integration of the imine peak (1H) at ca. 8.4 ppm and the new singlet peaks for diamine
(2H) and amine (2H) at ca. 4.80 and 4.65 ppm in their 1H NMR spectra, respectively for imines.
Based on this model, it appears that the pH of the aqueous
phase is less important than the mobility of protons through the
organic phase and this can explain the experimental observations.
It was also found that the use of the organic soluble proton donor
trifluoroacetic acid, did not change the product yields appreciably.
In future we believe that further study of the conditions at the car-
bon nanofiber-supported liquid–liquid interface will be important
for a wider range of electro-synthetic processes to be possible.

3360
J. D. Watkins et al. / Tetrahedron Letters 53 (2012) 3357–3360
11. Watkins, J. D.; Ahn, S. D.; Taylor, J. E.; Bulman-Page, P. C.; Bull, S. D.; Marken, F.
Electrochim. Acta 2011, 56, 6764–6770.
12. Amemiya, F.; Fuse, K.; Fuchigami, T.; Atobe, M. Chem. Commun. 2010, 46(16),
2730.
13. Amemiya, F.; Matsumoto, H.; Fuse, K.; Kashiwagi, T.; Kuroda, C.; Fuchigami, T.;
Atobe, M. Org. Biomol. Chem. 2011, 9, 4256.
14. Noyori, R.; Hashiguchi, S. Acc. Chem. Res. 1997, 30, 97.
15. da Silva, A. P.; Mota, S. D. C.; Bieber, L. W.; Navarro, M. Tetrahedron 2006, 62,
5435.
16. Popp, F. D.; Schultz, H. P. Chem. Rev. 1962, 62, 19.
17. Bian, Y. J.; Bai, D. S. J. Indian Chem., Sect. B 1890, 2007, 46.
18. Kise, N.; Agui, S.; Morimoto, S.; Ueda, N. J. Org. Chem. 2005, 70, 9407.
19. Kise, N.; Shiozawa, Y.; Ueda, N. Tetrahedron 2007, 63, 5415.
20. Atobe, M.; Tonoi, T.; Nonaka, T. Electrochem. Commun. 1999, 1, 593.
21. Santana, D. S.; Melo, G. O.; Lima, M. V. F.; Daniel, J. R. R.; Areias, M. C. C.;
Navarro, M. J. Electroanal. Chem. 2004, 569, 71.
22. Procedure for creation of the carbon membrane microreactor: First a glass tube of
internal diameter 3.5 mm and external diameter 5 mm was cut to be 50 mm in
length. A strip, ca. 3 mm wide of graphite foil was cut to be 70 mm in length
and attached to the inside of the glass tube by a chemically resistant silicon
based adhesive and the ends of the foil folded around the top and bottom of the
glass tube. A contact was formed at one end of the tube by the adhesion of a
copper wire by silver epoxy. At the other end a carbon membrane disc was cut
to 5 mm diameter and placed in contact with the foil. Silicon adhesive was
used to seal the carbon to the tube with a circle of carbon ca. 3.5 mm exposed
on both the inside and outside of the glass tube. Upon curing the silicon
adhesive contracted and forced the carbon membrane and foil into a physical
electrical contact suitable for electrochemical study.
Figure 2. Schematic drawing of the extended triple phase boundary reaction zone
and the reaction zone with a change in proton availability causing the formation of
a mixture of monomer and dimer reduction products.
Particularly interesting will be the use of pulse-electrolysis where
the diffusion time of reagents within the membrane electrode can
be controlled to increase the yield of 2-electron product versus 1-
electron product.
In conclusion, a microreactor system for hydrogenation and
electro-reduction reactions has been developed based on a carbon
nanofiber membrane that separates an aqueous electrolyte and an
immiscible organic phase (toluene or acetonitrile). For aldehyde
reduction and imine reduction processes good conversions and
interesting selectivity effects have been observed. Although not
isolated in this case, the use of this reactor for preparative reac-
tions can yield up to 5 mg of products to be produced. This as-
sumes a maximum concentration of 250 mM redox material can
be effectively electrolyzed, as found in our previous study.11 There
is however potential for scale up through the use of continuous
flow methods. The novel electro-reductive method described
should be easily adapted to other related synthetic challenges.
23. Voltammetric experiments were performed using a microAutolab III system
(Ecochemie, Netherlands) in staircase voltammetry mode, with the step
potential maintained at approximately 1 mV. The counter and reference
electrodes were platinum gauze and KCl-saturated calomel (SCE,
Radiometer), respectively. The working electrode was
carbon nanofiber membrane disc (‘Bucky paper’, Nanolabs US, with low
resistivity (ꢀ0.1 cm) containing relatively low impurity levels (Fe 0.36, Si
a 4 mm diameter
X
0.31, Al 0.23, Na 0.32, S 0.23 at %) mounted with Ambersil silicone (Silicoset
151) on a glass capillary of 3.5 mm inner diameter and 5 mm outer diameter.
The electrical contact was made with a 1 mm stripe of pyrolytic graphite film
(Goodfellow, UK) inside of the glass capillary. Solutions were deaerated with
argon (Pureshield, BOC). The pH was measured with a glass electrode (3505 pH
meter, Jenway). All experiments were conducted at a temperature of 22 2 °C.
24. General procedure for electro-reduction reactions: In order to perform a synthetic
reduction in the microreactor the electrode was first placed ca. 20 mm deep in
an aqueous solution containing electrolyte (Fig. 1). After visual inspection for
Acknowledgements
leaking of aqueous phase into the tube, 100 lL of organic phase containing
redox probe was added to the inside of the glass tube. The electrode was left to
equilibrate for several minutes to confirm no leaking of either phase though
the carbon membrane. A cyclic voltammogram was performed to ascertain the
J.D.W. gratefully acknowledges the EPSRC for funding support
and J. E. T. thanks the University of Bath for the award of a Ph.D.
Research Fellowship.
suitability of the electrode for synthesis and
a suitable potential for
potentiostatic reduction was chosen. Resistance variations were seen to
change the voltammetric shape slightly between electrodes. The overhead
stirrer was inserted into the glass tube to a depth of ca. 1 mm from the surface
of the electrode and was engaged at a rotation speed of ca. 300 Hz while the
potential was held at a value corresponding to a suitable reductive current (1–
10 mA). After the synthesis was complete the overhead stirrer was removed
and the internal organic solution separated by pipette as well as three
washings of the inside of the glass tube by fresh acetonitrile. The electrode was
then immersed in clean MeCN followed by 1 M H3PO4 followed by distilled
water for about an hour each before reuse. The aqueous phase can be recycled
immediately.
References and notes
1. Utley, J. Chem. Soc. Rev. 1997, 26, 157.
2. Moeller, K. D. Tetrahedron 2000, 56, 9527.
3. Marken, F.; Webster, R. D.; Bull, S. D.; Davies, S. G. J. Electroanal. Chem. 1997,
437, 209.
4. Scholz, F.; Komorsky-Lovric, S.; Lovric, M. Electrochem. Commun. 2000, 2, 112.
5. MacDonald, S. M.; Watkins, J. D.; Gu, Y.; Yunus, K.; Fisher, A. C.; Shul, G.; Opallo,
M.; Marken, F. Electrochem. Commun. 2007, 9, 2105.
25. General procedure for the synthesis of imines: Both imines were produced by the
6. Watkins, J. D.; MacDonald, S. M.; Fordred, P. S.; Bull, S. D.; Gu, Y. F.; Yunus, K.;
Fisher, A. C.; Bulman-Page, P. C.; Marken, F. Electrochim. Acta 2009, 54, 6908.
7. MacDonald, S. M.; Watkins, J. D.; Bull, S. D.; Davies, I. R.; Gu, Y.; Yunus, K.;
Fisher, A. C.; Page, P. C. B.; Chan, Y.; Elliott, C.; Marken, F. J. Phys. Org. Chem.
2009, 22, 52.
8. Atobe, M.; Ikari, S.; Nakabayashi, K.; Amemiya, F.; Fuchigami, T. Langmuir 2010,
26, 9111.
9. Watkins, J. D.; Bull, S. D.; Marken, F. J. Phys. Chem. C 2009, 113, 15629.
10. Watkins, J. D.; Amemiya, F.; Atobe, M.; Bulman-Page, P. C.; Marken, F.
Electrochim. Acta 2010, 55, 8808.
same reaction process. To
a solution of 4-trifluoromethylbenzaldehyde
(3.8 mmol) in anhydrous MeCN (20 ml) was added (3.8 mmol) benzylamine
under nitrogen and the mixture was stirred at room temperature for 24 h. In
some cases a white precipitate formed due to the acid base pairing of oxidized
aldehyde and amine but was removed by gravity filtration. The MeCN was
removed by rotary evaporation and the residue analyzed by 1H and 13C NMR
spectroscopy showing the imine product with no aldehyde being present. The
sample was used without further purification.
26. Marken, F.; Watkins, J. D.; Collins, A. M. Phys. Chem. Chem. Phys. 2011, 13,
10036.