Organic Process Research & Development
Article
to be used in organic electrochemistry applications.13 This
electrode material has a high surface-to-volume ratio, and the
three-dimensional porosity promotes turbulent mixing and
efficient mass transport. A recent study by Weber and co-
workers analyzed the flow behavior through GF electrodes and
showed that this material serves as a static mixer that supports
plug-flow behavior.13b
RESULTS AND DISCUSSION
■
Previous studies have shown that ACT is a very effective
electrocatalytic mediator for oxidation of primary alcohols to
carboxylic acids.6a,9 Chronoamperometry studies revealed that
ACT exhibits turnover frequencies (TOFs) of 400−1900 h−1
for primary aliphatic alcohols. Much slower rates are observed
under bulk electrolysis conditions. The reaction rates observed
when conducting ACT-mediated oxidation of LEV-CH2OH
under typical batch electrolysis conditions with reticulated
vitreous carbon (RVC)6a and graphite felt (GF) (see Table 1,
The GF anode was paired with a stainless-steel plate cathode
(10 cm2), and a PTFE static mixing spacer was included in the
the Supporting Information for details of the reactor
construction). The cell was tested with and without a
sulfonated fluoropolymer membrane between the anode and
cathode compartments. Finally, a Ag/AgCl reference electrode
was interfaced with the inlet tube of the anodic chamber to
allow the anode potential to be monitored while performing
flow electrolysis at constant current. Prior to each electrolysis
experiment, the flow was initiated and a voltammogram was
recorded in order to measure the current at 0.7 V (i.e., just
beyond ACT/ACT+ potential: E1/2(ACT) = 0.63 V vs Ag/
AgCl). This current value was used to conduct the ensuing
constant current electrolysis. The anode potential was stable
throughout the electrolytic oxidation of LEV-CH2OH (de-
tailed below) until substrate was consumed, and the reaction
was terminated when the anode potential rose to 1.0 V.
Representative electrolysis data obtained under different
conditions are provided in Table 1. The reaction time, product
yield, and faradaic efficiency were determined, together with
the extent of epimerization, defined according to the % ee
retained in the product [i.e., (ee of LEV-CO2H/ee of LEV-
CH2OH) * 100]. An initial batch experiment was conducted as
a benchmark, using the same electrode materials used for the
details). This reaction exhibits a comparatively long reaction
time (11 h) and a notable drop in ee, from 95.2% ee in the
alcohol to 87.5% ee in the LEV-CO2H product (92% ee ret)
(Table 1, entry 1). Flow conditions were initially tested using
an undivided flow cell (UDFC), which avoids the need for a
membrane and uses a single reservoir for the reaction solution.
Flow electrolysis under conditions otherwise identical to the
batch reaction (pH 9, 0.5 M Na2CO3/NaHCO3 electrolyte, rt,
10 cm2 GF electrode) led to a much faster rate, reaching
completion in only 1.2 h, and exhibited a significant
improvement in enantiomeric retention (97% ee ret, entry
2). Lowering the pH led to a slight improvement in
enantiomeric retention (98% ee ret at pH 8) but a significantly
lower rate, resulting in only 50% yield of product over the same
1.2 h time period (entry 3). A divided flow cell (DFC) was also
tested by incorporating a Nafion 324 membrane between the
anode and cathode compartments and using separate feed
solutions for each electrode. A preliminary experiment at pH 8
revealed very poor reaction performance (entry 4). This
outcome was traced to acidification of the anodic reaction
solution during electrolysis (a solution pH of 6−7 was measure
after 0.6 h). It was possible to maintain the pH by use of a pH
controller that titrated 1 M NaOH solution into solution
(Figure 2). Optimization of this approach (see Supporting
Information for further details) led to an excellent yield (92%)
within 0.6 h and near-perfect enantiomeric retention (entry 5).
Analysis of the reaction time course provided valuable
insights into the reaction performance during the DFC
electrolysis conditions (cf. Table 1, entry 5). The anode
potential was monitored relative to the Ag/AgCl reference
Table 1. Comparison of Batch and Flow Reaction
Conditions
c
d
e
Time
(h)
Yield
(%)
FE
ee ret
(%)
a
b
Entry
Format
pH
(%)
1
2
Stirred batch
Flow,
undivided
9.0
9.0
11
1.2
92
91
89
81
92
97
3
4
Flow,
undivided
8.0
1.2
50
44
98
Flow, divided
8.0
0.6
23
92
N.D.
N.D.
>99
f
5
Flow, divided
9.0
0.6
83
a
Conditions: 0.1 M (5 mmol) LEV-CH2OH, 5 mol % ACT, 50 mL of
carbonate buffer electrolyte (0.5 M; pH adjusted by varying the ratio
of Na2CO3/NaHCO3), flow rate = 50 mL min−1, rt. Variable constant
currents, set according to the current recorded at 0.7 V vs Ag/AgCl
during CV measurement at scan rate of 50 mV/s, corresponded to the
following values: batch, Iapp = 50 mA; undivided flow, Iapp = 500 mA;
divided flow, Iapp = 1000 mA. electrode dimensions = 10 cm2 × 0.5
b
cm. pH controlled by using different ratios of NaHCO3 and
c
Na2CO3. 1H NMR yields with DMSO as the internal standard.
d
e
Faradaic Efficiency. % ee ret = (ee of LEV-CO2H/ee of LEV-
f
CH2OH)*100; enantiomeric excess (ee) determined by HPLC. The
pH of the anode solution maintained by titration with 1.0 M NaOH
using a pH controller.
entry 1) electrodes correspond to an “effective TOF” of only
0.6−7 h−1, drawing attention to the influence of mass transport
on the reaction rate. Postulating that the longer reaction times
imposed by the batch reaction format contributes to the loss of
enantiomeric purity during the oxidation of LEV-CH2OH, we
initiated an effort to explore flow electrolysis conditions.
Single-pass flow conditions have been reported previously for
TEMPO-mediated alcohol oxidation;10 however, this approach
requires slow flow rates to ensure complete conversion in a
single pass and negates many of the beneficial mass transport
contribution of flow-based electrolysis.
As an alternative, we considered a recirculating flow process
that would be compatible with high flow rates. A commercially
available parallel-plate electrolysis cell (Micro Flow Cell from
ElectroCell11) was selected for these studies, owing to their
modular construction and availability of larger-scale reactor
formats. Three-dimensional (3D) porous graphite felt (GF; 10
cm2 geometric surface area with a 0.5 cm thickness) was
integrated in the anode compartment. GF has been commonly
used in redox flow batteries due to its wide operating potential
range, good electrical conductivity, high specific surface area,
and chemical and mechanical stability12 and recently has begun
B
Org. Process Res. Dev. XXXX, XXX, XXX−XXX