REVIEWS

A practical guide to electrosynthesis

Matthew C. Leech and Kevin Lam    ✉
Abstract | Organic electrosynthesis is an old and rich discipline. By exploiting the cheapest and
greenest source of electrons, electricity itself, electrolysis has been shown to be a powerful
method to perform redox reactions under mild, safe and green conditions. The field is in the
midst of a renaissance and there is little doubt that it will become one of the classic methods
to activate small organic molecules in the very near future. Nevertheless, electrosynthesis
can be rather daunting for a beginner. In this Review, we will guide synthetic chemists
through their first organic and organometallic electrosyntheses by reviewing the essential
aspects of the field and by sharing practical tips. We will also cover the fundamentals of
electro­analytical techniques, such as cyclic voltammetry, since they are powerful methods to
investigate mechanisms. Finally, these concepts will be examined in practice through three
case studies.

On March 20, 1800, Count Alessandro Volta sent a let­ Basic electrosynthetic concepts
ter to the then President of The Royal Society, Joseph We will discuss here the basic concepts of electrosynthe­
Banks, including a first description of a battery. It sis, focusing mostly on practical aspects. We have delib­
would not be exaggerating to say that it shocked the erately kept the discussion of theoretical concepts that
scientific world of the time; the first source of direct underlie the practical to a minimum. There are several
current was born. It did not take long for two chemists textbooks and reviews dedicated to the topic4–6.
to find a synthetic use for this discovery. A little under
two months later, on May 2, 1800, Sir Anthony Carlisle Current. The easiest way to understand, intuitively,
and William Nicholson performed the first electrolysis the concepts of current and potential difference is by
of water (to produce hydrogen and oxygen). Following analogy to a river.
this discovery, the electrolyses of many inorganic The potential difference can be compared with
compounds were performed, with varying degrees the elevation of a waterfall. The river’s water loses
of success. It was not until the early 1830s, however, potential energy when it falls and, in the same way,
that Michael Faraday became interested in the electro­ an electron loses energy when moving from a high
lysis of organic compounds. In 1834, he published potential point to a low potential point within an elec­
the following observation: “The electrolysis of a solu­ trical circuit. The electrical current is similar to the flow
tion of sodium acetate gives rise to the formation of a of a river. A high current means a rapid flow of elec­
gas (ethane)”1. The very first example of what is now trons in the circuit — a high number of electrons per
known as the Kolbe decarboxylation had just been per­ unit time (Eq. 1).
formed and the field of organic electrosynthesis was
dQ
born2. Though the field proved to be highly popular i(t ) = (1)
at that time, by the middle of the twentieth century, dt
the general interest in organic electrochemistry had During an electrolysis, oxidation occurs at the pos­
begun to fade. This might be partly attributed to a itively charged anode, while reduction occurs at the
lack of understanding of the processes — the field has negatively charged cathode. Per International Union
been regarded as something of a black box — but also a of Pure and Applied Chemistry (IUPAC) convention,
high entry barrier, with the need to build sophisticated reductions lead to negative currents, while oxidations
equipment. In addition, there was little concern at the lead to positive currents. A source of potential confu­
School of Science, time for the development of green or more sustain­ sion for novices here is that US chemists use the oppo­
The University of Greenwich,
Chatham Maritime, UK.
able processes. Recent technological developments in site convention, so one must take note of an author’s
✉e-mail: k.lam@ electroanalytical methods and the greater availability affiliation before attempting to understand their
greenwich.ac.uk of standardized electrosynthetic equipment has led discussion — especially if comparing work from two
https://doi.org/10.1038/ to a renaissance of the discipline and a resurgence of or more different labs (note, the USA uses the opposite
s41570-022-00372-y interest from the synthetic community3. convention).

NAture Reviews | CHEMiStry volume 6 | April 2022 | 275

0123456789();:
Reviews

Potential. A redox couple is defined by an equilibrium Kinetic aspects. Just like conventional reactions, electron
between an oxidized species (O) and a reduced species transfers can be analysed in terms of kinetics. The rela­
(R) — with the chemical equation (Eq. 2) balanced by a tionship between the rate (intensity of electric current)
number of electrons. and the potential depends on both the rate of the charge
O + ne − ⇌ R (2) transfer reaction and the rate of mass transport. As with
thermal reactions, the limiting step will impose its rate
The potential (E) of a redox couple (O/R) is a ther­ on the electrochemical reaction. Unlike conventional
modynamic value that reflects the oxidizing or reducing reactions, however, the temperature is not the only fac­
power of a given species and is given by the Nernst equa­ tor influencing the kinetics; in electrochemical reactions,
tion, which describes the electrochemical potential of the electrode potential is the main influential kinetic
the couple at equilibrium (Eq. 3), where R is the ideal gas factor (Butler–Volmer model).
constant, T the temperature, F the Faraday constant, ared The overpotential is an essential kinetic compo­
the activity in reductant and aox is the activity in oxidant: nent. Consider an ideal electrolytic cell, with no resis­
RT a tance, containing an acidic aqueous solution. In theory,
E = E0 − ln red (3) immersing an electrode in the solution and adjusting its
nF aox
potential to a slightly negative value should be sufficient
The absolute potential cannot be measured directly to generate hydrogen gas. In reality, however, under such
and is, thus, always given as a relative value with respect conditions, no hydrogen evolution occurs. A potential
to a reference electrode, the potential of which is set more negative than that predicted by thermodynam­
by convention. Colloquially, the concepts of ‘potential’ ics is required before hydrogen bubbles are noticeably
and ‘difference of potential’ are used interchangeably, formed on the surface of the electrode. The difference
although they are, strictly speaking, different. By con­ between this potential and the theoretical potential is
vention, the standard hydrogen electrode (SHE) has called the overpotential (η). It relates to the additional
a potential of 0.0 V. Electrons (from an electrode or a energy needed to compensate for the slow charge trans­
redox reagent) will flow from the lowest potential to the fers. To state it differently, the need for overpotential is
highest one. Thus, a species with a negative redox poten­ the result of the kinetics involved in the electronic trans­
tial is one that is more difficult to reduce than a proton. fer as well as from the ohmic drop (resistance). Applying
To draw an analogy between electrochemical potential a surplus of potential can accelerate these different
and a concept perhaps more familiar to synthetic chem­ equilibria.
ists, an electrode with a slightly negative potential could
be compared with a mild reducing agent, while an elec­ Electrolysis potential. Having reviewed the various
trode with a highly negative potential could be compared factors that influence the electrolysis, it is now possible
with a strong reducing agent. to establish the equation that links these factors to the
However, the SHE is rarely used in laboratories applied potential of an electrolysis (Eq. 4), in which Eanode
because of the safety issues posed by working with and Ecathode are the potentials, determined by the Nernst
hydrogen gas and, thus, is complex and time-consuming equation, of the redox couples found at the anode and
to install. For studies conducted in aqueous solution, the cathode, respectively; R is the resistance of the solution, i
use of an Ag/AgCl or calomel reference electrode is often the current and Ση the sum of the moduli of the cathodic
preferred to the use of the SHE7. and anode overpotentials (the cathodic overpotentials
Meanwhile, in organic solvents, the use of commercial are always negative).
Ag/AgCl or calomel electrodes is frequently impractical
— most are aqueous references and water may contam­ ΔEelec = E anode − Ecathode + Ri + ∑ η (4)
inate the organic electrolytic solution. In addition, the
high electrical resistance of organic solvents is problem­ Cyclic voltammetry
atic. Most commercial reference electrodes have only Here, we describe the main technique used in electro­
small, sintered glass dividers to enable ion mobility, analytical chemistry, in particular, its use in the elucidation
which dramatically increases the resistance of the system. of electroorganic mechanisms.
One solution to these issues is the use of a quasi-
reference electrode, for example, a silver wire coated Principle. The equipment needed to run cyclic voltam­
with AgCl (easily prepared by the electrolysis of 0.1 M metry experiments consists of a potentiostat connected
aqueous HCl with a Ag anode). Such a quasi-reference to an electrochemical cell (Fig. 1). Convection effects are
can be directly immersed into the solution or separated detrimental to electrochemical measurements and, thus,
by a highly porous sintered glass divider. The potential must be avoided. The cell must, therefore, be placed on
of such a quasi-reference is unknown, and the IUPAC a vibration-free surface and the temperature carefully
recommendation is to add ferrocene (Fc) as an internal controlled with a thermostat.
standard at the end of a set of experiments and to express
the measured potentials versus that of the Fc+/Fc couple. The basic setup consists of the following:
It is essential to appreciate the extremely narrow 1. A working electrode: the potential ramp is applied to
action range of the electrochemical potential. During an this electrode and the current recorded.
Potentiostat
An electronic instrument that
electrosynthesis, knowing the composition of the layer in 2. A reference electrode with a constant potential:
controls the voltage difference the immediate vicinity of the electrode is more important the potential applied to the working electrode is
between two electrodes. than the composition of the bulk solution. expressed in relation to this reference electrode.

276 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();:
 Reviews

3. A counter electrode: the potential of this electrode is Reference

electrode
adjusted using the potentiostat, such that the poten­ Counter electrode
Inert gas
tial difference measured between the reference elec­ inlet
trode and the working electrode matches the desired
applied potential: ΔEapplied = Ework(t) − Eref. The counter Working
electrode
electrode is also essential to close the electrochemical
circuit.

Cyclic voltammetry cells are standard three-electrode

setups. From a practical perspective, one selects a cell that
allows the three electrodes to be immersed in a minimum
of solution in order to minimize the amount of chemical
needed to run the experiments. As a rule of thumb, the
three electrodes should be less than 1 cm apart to mini­
mize the impedance of the system, especially if, for fear of
possible contamination, one decides to separate the coun­
ter and reference electrodes from the bulk of the solution,
using a porous sintered glass filter. Typically, the cell is
filled with 0.1 M solution of support electrolyte. Oxygen
is then removed and excluded from the system by sparg­
ing the solution with argon or nitrogen gas. The analyte
is then added in order to produce a 1-mM solution. It is
crucial to have 10–100 times more supporting electrolyte
than analyte. With a lower ratio, the analyte becomes a Electrochemical cell
current carrier, its resulting migration leading to ohmic
drop and a biased response. Organic species tend to Fig. 1 | An electrochemical cell used for cyclic voltam-
adsorb rapidly onto electrodes — particularly carbon metry. A standard three-electrode setup comprising
electrodes — therefore, it is advisable to polish the elec­ working, counter and reference electrodes, immersed in
trode between measurements to avoid a loss of sensi­ a solution of supporting electrolyte (0.1 M) and analyte
(1 mM). An inert gas inlet provides a means to sparge the
tivity and an altered signal. This is easily achieved by
system with dinitrogen or argon.
immersing the electrode in an aqueous suspension of
either nanometric alumina or diamond. Commercially
available, disposable, screen-printed three-electrode set­ peak (Epa and Ipa). Assuming the oxidized form of the
ups are more expensive but avoid the need for polishing analyte is stable and does not undergo any further chem­
because a fresh setup is used for each experiment. Cyclic ical or electrochemical reaction, it will be reduced back
voltammetry relies on diffusion, therefore, it is impor­ during the back scan and an opposite current, Ipc, will
tant to make sure that the solution is ‘quiet’. Indeed, the be recorded. The system is then said to be chemically
solution should only be stirred between, rather than dur­ reversible. In a chemically reversible system, Ipc must
ing, cyclic voltammetry scans. With this setup in place, be equal to the anodic Ipa. The number of exchanged
a linearly increasing or decreasing potential (E) ramp is electrons can also be determined by measuring the dif­
applied on the working electrode with a precise sweeping ference of potential between the cathodic peak and the
rate ν (V s−1) and the current (I) response of the system is anodic peak using Eq. 5, in which R is the ideal gas con­
recorded (I versus E; Fig. 2). stant, T the temperature, F the Faraday constant, n the
From a practical perspective, the potential is number of electrons, Epc the cathodic process potential
increased at a defined scan rate from an initial potential and Epa is the anodic process potential
(Ei) to a final potential (Ef ). Until the oxidation poten­
tial (E) of the analyte is reached, the recorded current is RT 56.5 mV
∣E pc − E pa∣ = 2.2 ≈ at 25 °C (5)
merely a capacitive current. However, once the critical nF n
oxidation potential is reached, the analyte is oxidized and
a positive current is recorded. The local electrolysis leads If the electrolysis products are unstable and undergo
to a rapid depletion of the analyte from the area near a rapid chemical reaction on the cyclic voltamme­
the electrode and a decrease of current is observed until try experiment timescale, no opposite current will be
a plateau is reached (diffusion current). At this point, recorded during the back scan. The system is then said
in the absence of stirring, the electrolysis rate is con­ to be chemically irreversible. Equations to describe such
Supporting electrolyte
trolled by the diffusion of the analyte from the bulk of systems are more complex and will not be discussed here.
A chemical species that is not
electroactive in the range of the solution to the electrode. The potential sweep is then For more complex systems, digital simulation packages
the applied potentials being reversed and, depending on the system under investiga­ can help beginners (as well as more experienced users)
studied, which is added to tion, a reductive current might be observed during the to extract the thermodynamic and kinetic information
a solvent medium in order back scan, giving rise to the classic ‘duck-shaped’ cyclic from a single voltammogram.
to increase its conductivity,
ideally without affecting the
voltammetry curve. The chemical reversibility of a system depends on
electrochemical behaviour Figure 2e shows a typical example of a cyclic voltam­ the scanning rate at which the analysis is performed.
of the analyte. mogram. The analyte is oxidized and shows an anodic A system that appears to be irreversible at a low scan

NAture Reviews | CHEMiStry volume 6 | April 2022 | 277

0123456789();:
Reviews

a b e
Epa

E I/A 0
Ipa

Scan
Time E/V direction
c I/A
d E/V
Ipc

E I/A 0

Epc
Time E/V

Fig. 2 | Principles of cyclic voltammetry. a | Initial sweep with potential increased linearly up to a set point. b | Observed
current response of the initial sweep shown in part a versus potential. c | Reverse sweep with potential returning to initial
level linearly (at the same rate as the initial sweep over time). d | Observed current response of the reverse sweep versus
potential. e | Typical cyclic voltammogram with important parameters annotated. Epa, anodic peak potential; Epc, cathodic
peak potential; Ipa, anodic peak current; Ipc, cathodic peak current.

rate might become reversible, or at least partially revers­ In other words, by using coulometry, one is able to
ible, at a higher scan rate. Therefore, by recording cyclic determine the number of electrons transferred per mol­
voltammograms at different scan rates, it is very often ecule of substrate. To do this, a voltammetry is carried
possible to study the different chemical kinetics that out beforehand in order to determine the reduction or
accompany the electron transfers. oxidation potential of the compound under investiga­
The concept of ‘reversibility’ often refers specifically tion. Once this value is known, the working electrode
to ‘chemical reversibility’ (i.e. the stability of the electro­ is set to this potential and, unlike in voltammetry or
lysed product). However, the concept of ‘electrochemical chronoamperometry experiments, the solution is vig­
reversibility’ also exists. An electrochemically irrevers­ orously stirred and electrolysed until the current value
ible system implies a slow electron transfer, while fast has fallen to 0.1% of its initial value. During this opera­
electron transfers lead to an ‘electrochemically reversible’ tion, the number of coulombs is recorded by integrating
system (also called Nernstian systems). the current value with respect to time (Eq.  6), where
i is the current in A, t the time in s and Q is the total
Chronoamperometry and chronocoulometry. Studying charge in C.
electrochemical changes with respect to time can pro­ tfin
vide valuable information about the kinetics of different Q= ∫ i(t )dt (6)
t0
electrochemical phenomena. The basic principle and
equipment needed to perform chronoamperometry Electronic integrators are now available, which, when
and chronocoulometry are substantially similar to those placed in series in the electrolysis circuit, directly measure
for cyclic voltammetry. During a chronoamperometric the number of charges consumed during electrolysis.
experiment, however, an initial applied potential is cho­ A word of caution is necessary here. It is frequently
sen so that no redox reaction occurs and is then quickly observed that the number of exchanged electrons meas­
pulsed to a value where the redox species is active, and ured by cyclic voltammetry or chronoamperometry dif­
the resulting current versus time plot is recorded. The fers from that found by coulometry. The difference is the
mathematical analysis of these currents makes it possible result of the timescales of the different analyses. While
to extract valuable information about the kinetics of the voltammetries and chronoamperometries are transient
different phenomena that occur during the electrochem­ methods performed in less than a second, coulometry is
ical process and is especially useful to study adsorption of performed over a time range from a few minutes to sev­
the analyte on the electrode surface8. Although chron­ eral hours. It is not uncommon for side chemical reac­
oamperometric methods offer some advantages over tions to take place over such a long time period and, if
cyclic voltammetry, it comes at the cost of a loss of signa­ the products are electroactive at the applied potential,
ture in ‘potentials’, since discreet potentials are selected they will have an impact on the final charge count.
during chronoamperometry, while a potential range is
scanned during cyclic voltammetry analysis, so the two Electrochemical parameter measurement. Here, we
Adsorption
The adhesion of a chemical
methods are often combined. briefly explain how the previously discussed analyti­
substance (known as the Coulometry is the measurement of the number cal methods can be used to extract important electro­
adsorbate) onto a surface. of charges involved in an electrochemical reaction. chemical parameters. In order to determine the redox

278 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();:
 Reviews

potential of a redox couple, voltammetry is the method parameters of the system, for instance, the kinetic rate
of choice. It does not provide the value of the standard constant of a chemical reaction that occurs following the
potential, which is essential for theoretical calculations, initial electron transfer. For simple systems, such as an
but, fortunately, it is not required for the purposes of electron transfer followed by a unique chemical reac­
electrosynthesis. For an electrochemically reversi­ tion, one can rely on the use of Nicholson and Shain’s
ble system, one can measure, from a voltammogram, semiempirical method11. However, for more complex
the half-wave potential (E1/2) — the potential at which the systems and in order to obtain more accurate data, the
current is half that at the apex of the cathodic or anodic use of a software package such as digital simulation
peak. In the case of an electrochemically non-reversible is often necessary to extract the kinetic and thermo­
system, there is only one peak, and either the cathodic or dynamic parameters from a chronoamperogram or
the anodic peak potential is reported (Epc or Epa), keeping cyclic voltammogram12.
in mind that this value depends on the parameters of the
cyclic voltammetry system, such as the scan rate and Electrosynthesis
the concentration of the analyte. In literature descriptions of electrosynthesis, technical
The next question is: ‘how many electrons have terms such as H-cell, standard cell or quasi-reference cell
been exchanged?’ This question may seem trivial, but are frequently used without proper definition and with­
it is challenging to determine the number of electrons out describing the construction of the apparatus. Our
exchanged for an unknown system. aim here is to provide a practical guide such that anyone
As we have seen earlier, in the case of a reversible can attempt to apply organic electrolysis13.
system, it is easy to determine the number of exchanged
electrons by cyclic voltammetry, whereas for a chemi­ Constant potential versus constant current. Electrolysis
cally non-reversible system, the same task is more com­ can be conducted under two different regimes: with
plex. Coulometry also allows for the determination of constant current or with constant/controlled potential.
exchanged electrons. However, if the oxidized or reduced To explain the fundamental difference between the two,
analyte generates a new electroactive species at the elec­ consider the general reaction shown in Eq. 7.
trolysis potential, then this will impact the final electron e− e− (7)
count as further reactions occur. Thus, coulometry meas­ A→B→C
ures the number of electrons exchanged during the global Constant-current electrolysis requires only a simple
electrolysis and not necessarily the number of electrons source of direct current (i.e. a battery) to be effective. At
exchanged in a specific process. Both the Compton9 and the beginning of the electrolysis, A is available in suffi­
Amatore10 groups have developed advanced methods cient quantity to ensure that the entire current is carried
to measure the number of exchanged electrons using by its electrolysis alone. Over time, however, the con­
voltammetric and other transient methods. centration of species A gradually decreases until there is
Finally, in order to propose a plausible reaction insufficient quantity for it to provide the entire current
mechanism, it is often interesting to know the kinetic on its own. At this point, the working electrode’s poten­
tial will drift until electrolysis of species B begins (assum­
ing it is electroactive) and, at this point, selectivity issues
+ – may arise. For example, stopping the reaction at B when
it can be further reduced to C would be challenging,
even though some A remains.
By contrast, constant-potential electrolysis makes
high selectivity possible by keeping the working elec­
Ammeter A trode’s potential constant during the electrolysis. The
Divided cell drawbacks of this method are that a longer electrolysis
+ DC power
– supply time is required and it may be necessary to use a more
sophisticated electrochemical setup.
Cell
The cell. The construction of an electrolysis cell is rela­
Constant-current tively simple and is within the reach of any competent
electrolysis schematic glass-blower.
Depending on the reaction to be performed, it is pos­
sible to use a divided or undivided cell (Fig. 3). In the sim­
Non-divided cell plest of configurations, in an undivided cell, molecules
are free to move to the cathode or to the anode. This is
not a problem if the reactant in question can only be
reduced or oxidized and the product is not electroac­
tive. However, in many cases, such a cell configuration
Fig. 3 | Simple electrochemical reaction setup. A simple circuit diagram shows how a
DC power supply is connected in series with a reaction cell and an ammeter to perform is not appropriate, for example, in the event that the ini­
electrolysis at constant current. Such a setup can be used with a non-divided cell if the tial product may undergo a further reduction or oxida­
reactant can be either only oxidized or only reduced and if the product is not electro­ tion, or if the initial electrolysis product may react with
active. In other cases, a divided cell is used and the electrolyte solutions surrounding the the counter electrode or one of the products generated
cathode and the anode are separated using a diaphragm. at the counter electrode. In this case, a divided cell,

NAture Reviews | CHEMiStry volume 6 | April 2022 | 279

0123456789();:
Reviews

Table 1 | Advantages and disadvantages of different electrode materials activity may facilitate the reduction of a radical anion
by protonating it. It can also prevent a proton elimi­
Material Pro Con nation reaction from an oxidation-generated radical
Mercury (cathode) Large H2 overpotential (can access Highly toxic (can form cation.
very negative potentials) organomercurial 2. The electrochemical window. This is the range of
compounds)
Renewable surface potentials accessible to an electrochemical system. The
Now rarely used latter depends on several factors, such as temperature,
Lead (cathode) Large H2 overpotential Easily passivated by solvent, supporting electrolyte or electrode.
organic compounds 3. The  dielectric constant. This has a direct influence
(needs cleaning during on the value of the ohmic drop (resistance) because
the electrolysis)
it controls the dissociation of the electrolytes. It is,
Silver (plated) Excellent for the reduction of High cost therefore, wise to use solvents with a strong dielectric
(cathode) organic halides constant.
Nickel (anode) Forms the strongly oxidizing agent High cost
NiOOH in basic conditions Protic acidic solvents such as sulfuric, fluorosulfonic,
Magnesium, Avoids the need for a large excess Corrosion of the trifluoroacetic and acetic acids have been used histor­
aluminium, iron, zinc of supporting electrolyte electrode during the ically owing to their ability to dissolve both organic
(sacrificial anodes) Avoids the need for a divided cell
electrolysis and inorganic compounds. However, their strong cor­
rosivity and their incompatibility with certain organic
Generates Lewis acids in situ
compounds means that they have fallen into disuse.
Carbon (cathode and Highly versatile Can be brittle Neutral protic solvents are very often preferred to
anode) the previous category. Indeed, alcoholic solvents, such
Large H2 overpotential (as a cathode) Difficult to clean
Low cost as methanol and ethanol, have been found to be excel­
Wide varieties available (graphite,
lent solvents to carry out oxidations in undivided cells,
glassy carbon, pyrolytic carbon etc.) with the reduction of the alcohol solvent occurring at
the cathode. High-molecular-weight alcohols are typi­
Easily chemically modified
cally unsuitable as electrolytic solvents, owing to their
Platinum (cathode Highly versatile Low H2 overpotential poor conductivity. Hexafluoroisopropanol, however,
and anode) (as a cathode)
High stability has recently been shown to be an ideal solvent for many
Easy to clean High cost electrosynthetic transformations16. Interestingly, water
can be used as solvent for electrolysis, unless it is incom­
patible with the expected reaction mechanism. The main
equipped with a diaphragm, should be used to separate problem with water is, nevertheless, that it is a poor sol­
the  catholyte from the  anolyte. The shape of the latter vent for many organic species. A practical trick consists
type of cell often resembles the letter H and, as a result, of using a ‘hydrotropic’ supporting electrolyte, such as
Diaphragm it is generally referred to as a ‘H-cell.’ tetrabutylammonium tosylate, to increase the solubility
Alternatively a membrane or
of organic substances17,18. Another solution is to run an
frit, a semipermeable material
that allows the flow of ions The electrodes. The electrode material undeniably plays emulsion-type electrolysis by adding a phase-transfer
between the anolyte and the a key role in electrolysis and the importance of choosing catalyst to the cell19–21.
catholyte compartments in a the right material has been reviewed elsewhere14,15. Basic amine-based solvents, such as ammonia,
divided cell without mixing the The construction of an electrode remains a trivial methyl amine or ethylene diamine, have been used
two solutions.
step that consists of cutting a piece of metal or carbon mostly to perform reductions and to study highly basic
Catholyte and drilling into its upper part in order to attach a cop­ species, such as dianions4. Another useful property
The electrolyte in the presence per wire that will be connected to a current generator. specific to these solvents is their ability to generate sol­
of the cathode in an For safety reasons, it is recommended that users insu­ vated electrons at the cathode that can be put to use in
electrochemical cell.
late copper–electrode connections with Teflon. This reactions similar to the well-known Birch reduction22.
Anolyte is especially important if the electrode is used as the Aprotic polar solvents are the ideal candidates for
The electrolyte in the presence anode, since the slightest copper surface exposed to electrosynthesis when trace amounts of protons would
of the anode in an the solution will oxidize until the connections dissolve. be detrimental to the desired reaction. In addition, it is
electrochemical cell. An alternative is to use wires made from a nickel-chrome often easier to study the mechanism of electrochemi­
Dielectric constant
alloy. These are more resistant to corrosion and can be cal reactions in aprotic solvents, since mechanisms can
A measure of the polarity of an connected to the electrodes using an epoxy resin. The vary as a function of the pH (ref.23). However, one must
organic solvent and its ability different cathode and anode materials are summarized keep in mind that these solvents are usually difficult to
to insulate charge. in Table 1. dry and may contain residual water, alongside other
nucleophilic impurities. These can be easily removed
Ohmic drop
Also known as IR drop, a The solvent. Electrolysis can only occur at the interface by passing the solvent through a column of activated
potential drop caused by between the electrode and a conducting solvent. The alumina immediately prior to the electrolysis.
the inherent resistance of the nature of the solvent will, therefore, play an impor­ Acetonitrile is one of the most popular aprotic sol­
solvent, which can cause tant role in electrolysis and several factors must be vents, whether for reductions or oxidations, owing
shifts in peak potential, reduce
observed currents and increase
considered. to its large electrochemical window. In addition, the
the separation between anodic 1. The proton activity. Related to the ‘acidity’ of the conductivity of the electrolyte solutions in acetonitrile
and cathodic peaks. solvent in question. A solvent with a strong proton is improved. The solvent is water-soluble and is easily

280 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();:
 Reviews

removed during the workup. Interestingly, acetoni­ are much more stable than perchlorates and more
trile, unlike N,N-dimethylformamide (DMF), is a poor difficult to oxidize than hexafluorophosphates.
hydrogen atom donor in radical processes. Tetrabutylammonium triflate is also much more sol­
DMF is an excellent choice for organic electrolysis, uble in organic solvents than perchlorate or the cor­
since it can dissolve a wide range of organic and inor­ responding tetrafluoroborate. Finally, this quaternary
ganic compounds. The major drawback with this solvent ammonium salt is easily prepared, in one step, from
is that it slowly breaks down into dimethylamine and tributylamine and butyl triflate. Tosylates are so-called
formic acid. N-methylpyrrolidine is a more stable alter­ ‘hydrotropic’ salts. Once dissolved in water (up to
native to DMF but is more difficult to remove at the end 1 M), their hydrophobic properties facilitate the disso­
of the reaction. lution of organic species. The effect is even more pro­
Dimethyl sulfoxide is equally an excellent solvent nounced when partnered with a tetraalkylammonium
that dissolved many inorganic and organic compounds. cation.
It is relatively resistant to oxidation but is easily cathod­ Lithium cations can be reduced into metallic lithium
ically decomposed. Dimethyl sulfoxide is not always when a platinum electrode is used. The presence of traces
chemically inert and can take an active role during of water can then lead to the passivation of the electrode
electrolysis when it is reduced into dimethylsulfide or through the formation of an insulating layer of lithium
methylsulfinylmethylide (dimsyl) anion. hydroxide. Sodium salts are rarely soluble in organic
Less conductive solvents such as nitromethane, ethers solvents and, thus, with the exception of sodium per­
and dichloromethane have been used to carry out anodic chlorate, are rarely used. Tetraalkylammoniums are the
oxidations, even though a large amount of supporting most commonly employed cations in organic electrosyn­
electrolyte had to be used24. Ethers are poor oxidation thesis. The halides of these cations are sometimes used,
solvents because they are themselves easily oxidized but tetrafluoroborate and hexafluorophosphate anions
into their corresponding oxonium, while dichlo­ are usually preferred due to their large anodic window.
romethane releases chloride ions as a result of cathodic The presence of a strong base in the medium tends to
decomposition. degrade tetraalkylammonium salt into trialkylamine via
Ionic liquids have been introduced into organic a Hofmann-type degradation.
electrochemistry with mixed success25,26. Indeed, these
salts, which are liquids at room temperature, have large Case studies
electrochemical windows. However, they are very resist­ Here, we use three case studies that illustrate three key
ant and viscous. One way to alleviate this problem is concepts associated with electrosynthesis, the impact of
to add a small amount of DMF to the medium27. This the nature of the electrodes, the influence of the support­
solvent seems to disturb the stacking of ions and facil­ ing electrolyte and the importance of the electroanalyti­
itates their mobility, leading to a significant increase in cal studies, to elucidate a redox mechanism and develop
conductivity. a new reactivity.

The supporting electrolyte. The passage of current Importance of the electrode material. Here, we will
through a solution depends directly on the mobility describe how the nature of the working electrode could
of ions found in the solution. The choice of electrolyte dramatically change the outcome of an electrosynthetic
depends on its solubility, its dissociation in the solvent reaction, in this case, an anodic decarboxylation of
and its chemical compatibility with the reaction one carboxylic acids.
wishes to perform. If the electrolysis is being carried out In contrast to most organic electrosyntheses, the elec­
in a divided cell, it is essential that both the catholyte trochemical behaviour of carboxylic acids has been thor­
and the anolyte contain broadly similar concentrations oughly studied. The Kolbe reaction or anodic oxidation
of supporting electrolyte. Otherwise, the solvent may of carboxylates2 was among the first reported organic
migrate from one compartment to the other during the electrosyntheses1, if not the first, with many articles
electrolysis. published on the topic28–32.
In this section, we will broadly discuss the main ani­ In the Kolbe reaction, a carboxylate anion RCO2− is
ons and cations used in supporting electrolytes. More oxidized at the anode to produce the corresponding
in-depth information, such as the solvent windows for carboxyl radical (RCO2•). When R is an alkyl group,
various solvent and supporting electrolyte combinations, a decarboxylation occurs very rapidly and leads to the
can be found elsewhere24. formation of an alkyl radical R•. The outcome of the reac­
Perchlorates are very difficult to oxidize and, there­ tion is slightly different when R is aromatic. Although an
fore, offer an excellent oxidation window. While lith­ aroyloxy radical is still produced, its decarboxylation is
ium or sodium perchlorates are relatively stable at much slower than for the aliphatic analogues33. Typically,
room temperature, even in presence of organic com­ the reaction is carried out on an alcoholic solution of the
pound, tetrabutylammonium perchlorates should carboxylic acid that has been partially neutralized (usu­
always be vacuum-dried without ever being heated. ally up to 25%) and the electrolysis is performed in an
Tetrafluoroborates offer a slightly wider anodic window undivided cell using platinum electrodes. Interestingly,
than that of perchlorates and, importantly, are much cyclic voltammetry experiments show that the solvent
more stable. Hexafluorophosphates are also highly stable is actually easier to oxidize than the carboxylate! One
and even more resistant to oxidation than tetrafluorob­ might then wonder how the reaction could still pro­
orates or perchlorates. Trifluoromethanesulfonates ceed. The key to understanding this is that, during an

NAture Reviews | CHEMiStry volume 6 | April 2022 | 281

0123456789();:
Reviews

electrolysis, the composition of the electrode and the addition product is obtained (Fig. 4c). The methodol­
layer of solution surrounding it very often matters more ogy was extensively used to rapidly access complex
than the composition of the bulk of the solution. Indeed, carbocyclic and heterocyclic compounds through a
the use of high current densities (usually higher than Kolbe–intramolecular cyclization tandem reaction29,45–47.
100 mA cm−2) during the electrolysis helps to ensure As noted above, when traditional Pt electrodes are
the adsorption of the negatively charged carboxylates substituted for graphite electrodes, the fate of the anodic
on the positive anode such that solvent molecules are oxidation is very different. In the Hofer–Moest variation
excluded. Such high current densities also lead to a high of the Kolbe reaction, the radical (R•), initially produced
concentration of alkyl radicals produced by the anodic via the usual Kolbe decarboxylation, is further oxidized
decarboxylation of the carboxylic acids and favours their to produce the corresponding carbocation (R+)28,29,48
combination on or near the electrode’s surface34. Finally, (Fig. 4d). Although other factors such as the temperature
another key element of the reaction is the use of ‘smooth’ and the nature of the solvent–electrolyte system have
anode materials, such as precious metals (Pt, Au, Ir etc.) been found to favour such multielectron oxidation, the
or glassy carbon electrodes. As we will see, electrodes nature of the anode remains the most important.
made from porous materials such as graphite induce a The exact reason for this dramatic change in mechanism
very different result. remains poorly understood. One plausible hypothesis
The Kolbe electrolysis is a highly versatile method to would be that organic molecules tend to adsorb more
generate radicals from simple, unactivated and readily strongly on graphite than on platinum, thus, allow­
available carboxylic acids. The method has been found ing enough time for further oxidation to the cation
to possess a wide range of applications, some of which to occur.
are summarized in Fig. 4. This simple modification of the Kolbe electrolysis
One of the most common uses of the Kolbe reaction has been used to perform classic cationic reactions.
is the preparation of homodimers29,35–38 (Fig. 4a). Kolbe The most common use of the Hofer–Moest reaction
electrolysis has also been employed to perform unsym­ is depicted in Fig.  4d . A carbocation is generated
metrical radical sp3–sp3 couplings (Fig. 4b). To achieve through the decarboxylation of a carboxylic acid, and
this, a carboxylic acid (R1COOH) is electrolysed in the the generated carbocation is trapped with a nucleophile.
presence of an excess of a co-acid (R2COOH) in order Carboxylic acids49, alcohols50, water51 and nitriles52 have
to obtain the cross-coupling product (R1–R2)39. This been found to be suitable nucleophilic partners. The
method has been intensively used for the synthesis of methodology has been shown to be particularly useful
pheromone derivatives, among other applications29,40–44. for the modification of carbohydrates53. The Hofer–
Unfortunately, the formation of the dimer of the Moest reaction has also been applied to other classical
co-acid (R2–R2) is unavoidable. Therefore, the method carbocation transformations, such as eliminations54 or
is restricted to co-acids that lead to the formation of rearrangements55.
volatile or easily removable dimers. The structure of the carboxylic acid itself also has a
When a carboxylate is submitted to the Kolbe dramatic impact on the outcome of the anodic oxida­
electrolysis in the presence of an alkene, the radical tion. Oxidation of allylic, benzylic and tertiary carbox­
ylic acids results in overoxidation of the alkyl radical to
form the corresponding stabilized carbocation inde­
One electron per molecule pendent of the electrode material employed. Anodic oxi­
a O O
– 2e–
dation of secondary carboxylic acids generates a mixture
+ MeO2C CO2Me
– 2CO 2 R1 R1 of carbocations and radicals.
R1 O R1 O
Besides the advances in the field discussed here, more
recent applications of the Hofer–Moest reaction include
b O O
+
– 2e– the mild formylation of amines56, as well as its very
R1 R2 R1
R1 O R2 O
– 2CO 2
O
promising use in the synthesis of biofuels57–60.
O

From mechanism to a new reaction. Here, we will discuss

c O R1 R1
how the use of electroanalytical methods helped to eluci­
O Y
– 2e– date the mechanism of disulfide oxidation and how this
+ HO Y – 2CO 2
R1 O N ultimately led to the development of new methodology61.
X X COR2 The reader will have to keep in mind that the US
current–potential convention will be used for the cyclic
Two electrons per molecule voltammetry analysis (negative oxidative current,
d MeO positive reduction current).
O NuH
– 2e– The oxidation of disulfides (R1SSR2) is an important
R+ R1 Nu HO
– 2CO 2 HO OH
R1 O
OH biological process and has, unsurprisingly, attracted a lot
of attention. For a long time, the monoelectronic oxida­
Fig. 4 | Applications of the Kolbe reaction. a | Homocoupling of two carboxylates. tion of disulfides was thought to lead to the rupture of
b | Heterocoupling of two carboxylates. c | Formation of substituted carbocycles and het- the S–S-bond and to the formation of a thionium (RS+)
erocycles. d | Hofer–Moest decarboxylation and subsequent trapping with a nucleophile. and of a radical (RS•), even though the sulfur-centred
R1 and R2, aliphatic chains. Examples taken from refs36,44,45,53, with the formed bond(s) radical was never observed nor trapped. The rad­
highlighted in red. ical cation of the disulfide was also thought to be too

282 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();:
 Reviews

a b

F
F
S S
S S
F F
F
S
S

F
Initial compound IJ
IJ

IC IC
10 µA
Reducible
Reducible 10 µA
product
product

–0.3 –0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 1.7 –0.7 –0.2 0.3 0.8 1.3 1.8
E (volt vs Cp2Fe0/+) E (volt vs Cp2Fe0/+)

c S R
– S R R S R
S R -e S R R S S
R S R S 2
S S R
S R R S
R

Fig. 5 | Experimental cyclic voltammograms and proposed mechanism for the oxidation of (p-FC6H4S)2. a | Cyclic
voltammetry of (p-FC6H4S)2 in CH2Cl2/0.05 M [Bu4N][PF6] (dashed lines) and CH2Cl2/0.05 M [Bu4N][B(C6F5)4] (solid lines),
scan rate = 0.2 V s−1. In the presence of PF6−, the oxidation is irreversible; the radical cation (red circles) is, thus, understood
to have reacted further, as soon as it is formed. In the presence of B(C6F5)4−, however, the radical cation is stable and the
oxidation is seen to be reversible. b | Cyclic voltammetry of (p-FC6H4S)2 in CH2Cl2/0.05 M [Bu4N][B(C6F5)4] before (solid lines)
and after (dashed lines) bulk electrolysis, scan rate = 0.2 V s−1. After bulk electrolysis, no disulfide cation remains but a new
product has been formed that is itself irreversibly reduced at 0.25 V versus Fc+/0. This new product can be oxidized back
to the starting disulfide. c | Proposed new mechanism for disulfide oxidation. Figure adapted with permission from ref.61,
American Chemical Society.

unstable to be observed by cyclic voltammetry. Indeed, Fc+/0 (dashed line, Fig. 5b). In this same reductive scan,
a cyclic voltammogram of para-fluorobenzenedisulfide the only observed reduction is that at Epc = 0.25 V versus
(p-FC 6H 4S) 2 (dashed lines) in a dichloromethane Fc+/0 (a reduction of the product previously observed).
(DCM) solution containing [NBu4][PF6] as supporting Interestingly, this reduction is also chemically irrevers­
electrolyte shows a chemically irreversible oxidation at ible and, on the back scan, a peak at around 1.21 V ver­
Epa = 1.29 V versus Fc+/0 (Fig. 5a). However, the reductive sus Fc+/0 indicates regeneration of the starting disulfide.
return scan shows a reduction at Epc = 0.25 V versus Fc+/0. Coulometry shows that 2/3e− are consumed per mole of
Although the PF6− anion is very often considered unre­ disulfide in this reduction. When the same experiment is
active, it remains quite coordinating in the presence of performed in DCM in the presence of [NBu4][B(C6F5)4],
a strong electrophilic cation, such as a disulfide radical a solution of the disulfide radical cation can be obtained
cation. When PF6− was substituted by a weakly coordi­ and no other chemical reaction is observed. The solu­
nating anion such as B(C6F5)4−, a very different oxidative tion can be reduced to reform the original disulfide. In
behaviour is observed. The oxidation of the disulfide is this case, the number of exchanged electrons per mole of
now chemically reversible (E1/2 = 1.21 V versus Fc+/0) and disulfide is close to one. This confirms that, in the pres­
no reduction is observed on the back scan. ence of a weakly coordinating anion, the radical cation
These two simple experiments immediately begin to of the disulfide is quite stable and only traces of the new
shed light on the mechanism of disulfide oxidation. In reducible product are formed. Additional cyclic voltam­
the presence of PF6−, a coordinating anion, a chemically metry experiments, using different concentrations of
irreversible oxidation is observed, indicating that, as disulfides and experimenting at different temperatures
soon as it is formed, the disulfide radical cation under­ in DCM/[NBu4][B(C6F5)4], show that higher concentra­
goes a rapid follow-up chemical reaction that leads to the tions of disulfide and lower temperatures both favour the
formation of a new electroactive compound that could formation of the reducible product.
be reduced at Epc = 0.25 V versus Fc+/0. In the presence We now have more clues about the possible nature of
of B(C6F5)4−, however, the radical cation is surprisingly the reducible product. Indeed, the fact that only 2/3e− are
stable at room temperature. consumed during the anodic oxidation of the disulfide
The next step is to perform bulk electrolysis and in DCM/[NBu4][PF6] suggests that one neutral disulfide
coulometry of this oxidation. After bulk electrolysis in has combined with two disulfide radical cations to form
DCM containing [NBu4][PF6] at 1.38 V versus Fc+/0, no a new compound that can be reduced back into the start­
disulfide radical cation remains. Its total consumption ing disulfide. Combined with the knowledge that radical
is confirmed by the absence of current at 1.21 V versus cations dimerize easily and that the reducible product

NAture Reviews | CHEMiStry volume 6 | April 2022 | 283

0123456789();:
Reviews

formation is favoured at high concentration of disulfide a. The formation of the disulfide radical cation (as seen
and low temperatures, we are able to propose a plausible by the oxidation wave at 1.21 V versus Fc+/0 on the
mechanism for the reaction: cyclic voltammetry experiments).
b. The dimerization of two radical cations (suggested
by cyclic voltammetry experiments at various
a
concentrations and temperatures).
–e –
Fe Fe c. The formed dimer then reacts with one molecule of
neutral disulfide (as suggested by the 2/3e− per mol­
ecule of disulfide measured by coulometry) to form a
2+
compound that regenerates the initial disulfide when
4(–e – ) reduced.
oxidation
Ru Hg Ru 4 Ru 4 Ru
Hg electrode
Taken together, this suggested that the identity of the
I new reducible compound was a trisulfide cation. Indeed,
–H + 2 × (+2e –) the dimer of the disulfide radical cation would be highly
+2e
electrophilic and would be attacked by a nucleophilic
+ 2+ 2+
disulfide rapidly upon its formation. The reduction of
the trisulfide cation would also be in good agreement
2 M H + Ru 2 Ru Ru with the formation of the initial disulfide.
Ru The proposed mechanism (Fig. 5c) was supported by
IV II chemical oxidation of the disulfide with SbCl5, which
III Favoured at low T was found to result in the trisulfide cation. Cyclic
Favoured at high T voltammetry of this trisulfide cation confirmed its
reduction at 1.21 V versus Fc+/0 to reform the disulfide.
b c
1 Ultimately, a methodology for the addition of
1 disulfides across double bonds was developed based on
this mechanistic investigation61.

Importance of the supporting electrolyte. The choice

Conc. ia ia of an appropriate supporting electrolyte is frequently
Current ic 298 K ic 233 K overlooked in the development of electrosynthetic
1 2+
1 + reactions, as it is often believed to be present solely to
1+ improve conductivity of the reaction medium and
to have no influence on the reactions performed. This
–0.5 0 0.5 1.0 –0.5 0 0.5 1.0 is not always the case, however, as exemplified by the
Volt vs FeCp2 Volt vs FeCp2 electrochemical studies conducted by Geiger and col­
leagues on ruthenocene — the second-row analogue of
d ferrocene62 (Fig. 6).
III III Ferrocene and the ferrocenium ion exhibit almost
ideally reversible redox behaviour in non-aqueous
III media, so much so that this redox couple has been
25 °C designated by the IUPAC as the internal reference of
choice for electrochemical studies. The electrochemical
behaviour of ruthenocene is far more complex.
It was initially reported that the anodic oxidation of
ruthenocene using a mercury electrode was a reversible
0 °C one-electron process. However, isolation of the oxidized
product revealed a Hg(I) adduct, Hg[(RuCp2)2]+, instead
II I of the expected ruthenocenium. Oxidation (either
IV chemical or electrochemical) of ruthenocene (RuII) was
known to produce RuIV complexes, although this had
–45 °C been attributed to disproportionation of ruthenocenium
(RuIII).
Instead, it was suggested that the oxidation of ruthe­
ppm 5.5 5.0 4.5 –7.0 –7.5 nocene was a two-electron process when conducted in
media containing traditional supporting electrolytes (i.e.
Fig. 6 | Proposed reaction pathways after electrochemical oxidation of rutheno-
[BF4]−, [PF6]−, [ClO4]−). However, this assertion could not
cene. a | General scheme for the oxidation of ferrocene versus ruthenocene. b | Cyclic
voltammetry plot showing ruthenocene → ruthenocenium formation at room tempera- be reconciled with the reversible one-electron process
ture. c | Lower temperature cyclic voltammetry plot showing formation of dimer after observed when oxidation was conducted in molten salts
oxidation. d | Stacked NMR spectra (rather than overlay, which may be confusing) at dif- nor with the fact that the decamethylruthenocenium
ferent temperatures, showing how peaks/compounds appear and disappear with varying ion was not observed to disproportionate under similar
temperature. Figure adapted with permission from ref.62, American Chemical Society. conditions. Thus, it was postulated that the supporting

284 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();:
 Reviews

electrolyte, specifically, its anionic component, was at high temperatures, as evidenced by the broadening
responsible for promoting the two-electron oxidation and eventual disappearance of the Cp-H resonance at
as a result of nucleophilic attack at the in situ-generated 5.38 ppm at higher temperatures. Similarly, the appear­
RuIII centre. A simple (reversible) one-electron oxida­ ance of a pair of triplets at 5.53 and 5.02 ppm in addition
tion could be observed when oxidation was performed to a singlet at 5.14 ppm, which integrate to 2:2:5, respec­
in solutions containing weakly coordinating anions at tively, is indicative of the formation of the dinuclear
room temperature (i.e. BArF24; Fig. 6). C–H activated species. The generation of [Cp2RuH]+
When electrochemical studies were carried out was evident through the presence of a hydride reso­
in low-donor solvents with supporting electrolytes nance between −7.1 and −7.3 ppm. The expected Cp-H
comprising weakly coordinating anions (e.g. TFAB; resonance was highly temperature sensitive, appearing
[B(C6F5)4]−, tetrakis(perfluoroaryl)borate anion), the between 1.8 (298 K) and 3.7 ppm (228 K). The formation
voltammetric response observed was consistent with of these complexes is summarized in Fig. 6.
reversible formation of the desired ruthenocenium ion When small quantities of [NBu4][PF6] were added to
(E½ = 0.4 V versus Fc). Slower scan rates gave little indica­ solutions of ruthenocene containing either TFAB-based
tion of dimerization, with near-reversible behaviour con­ or BArF24-based supporting electrolytes, the reversibility
sistently observed. However, at high scan rates or at low of the ruthenocene oxidation was immediately changed.
temperatures, the reversibility of the voltammogram was When ten equivalents of [NBu4][PF6], were added,
reduced, with a cathodic wave observed at ca. −0.06 V the reversibility of the oxidation had been completely
(versus Fc at 243 K; ν = 0.2 V s−1), which was attrib­ quenched, with a shift of the oxidative process to more
uted to formation of the bis(ruthenocenium) dication negative potential. This is consistent with either increased
by analogy to the corresponding osmium system. ion pairing of the ruthenocenium with the [PF6]− or a
After exchanging the TFAB counterion for BArF24, coupled follow-up reaction, such as the formation of
the reversible one-electron process corresponding to the a metal-fluoride or metal-F-PF5 Ru(IV) species.
ruthenocene/ruthenocenium redox couple appeared at a
more positive potential (E½ = 0.57 V versus Fc). Moreover, Conclusions
two additional cathodic product peaks were observed, In this article, we have introduced organic electrosynthe­
with their relative amounts dependent upon the cyclic sis, its history and the essential aspects of the method. An
voltammetry scan rate and temperature. An additional overview of the theoretical and experimental aspects has
consequence arising from the use of BArF24 counte­ been discussed, alongside practical advice for those seek­
rions is that the electrogenerated species possessed ing to undertake their first electrosyntheses. Key param­
a much greater solubility when compared with their eters such as cell design, electrode material, solvent and
TFAB counterparts, thus, facilitating characterization supporting electrolyte choice have been reviewed in
using NMR spectroscopy. order to demystify some of the most overlooked factors
Reducing the supporting electrolyte concentrations by those taking their first steps into the field. In com­
enabled NMR shimming of samples of ruthenocene that bination with organic and organometallic electrosyn­
had been electrolysed at low temperatures in CD2Cl2 at thesis case studies, we have shown the dramatic effects
a concentration appropriate for NMR studies (~10 mM that these parameters can have, including their ability
in the metal complex). In addition to one Ru(IV) side to dramatically alter the fate of an electrosynthetic
product, it was possible to identify three different Ru(III) reaction.
species, which coexist in a temperature-sensitive equilib­
rium. Formation of the metal–metal dimer is disfavoured Published online 21 March 2022

1. Faraday, M. Experimental researches in electricity. 10. Amatore, C. et al. Absolute determination of 19. Eberson, L. & Helgée, B. Studies on electrolytic
Philos. Trans. R. Soc. Lond. 122, 125–162 (1832). electron consumption in transient or steady state substitution reactions. IX. Anodic cyanation of
2. Kolbe, H. Untersuchungen über die Elektrolyse electrochemical techniques. J. Electroanal. Chem. aromatic ethers and amines in emulsions with the
organischer Verbindungen. Justus Liebigs Ann. Chem. 288, 45–63 (1990). aid of phase transfer agents. Acta Chem. Scand. B
69, 257–294 (1849). 11. Nicholson, R. S. & Shain, I. Theory of stationary 29, 451–456 (1975).
3. Yan, M., Kawamata, Y. & Baran, P. S. Synthetic organic electrode polarography: single scan and cyclic 20. Eberson, L. & Helgée, B. Studies on eectrolytic
electrochemical methods since 2000: on the verge methods applied to reversible, irreversible, and substitution reactions. XII. Synthesis of 4-alkoxy-4′-
of a renaissance. Chem. Rev. 117, 13230–13319 kinetic systems. Anal. Chem. 36, 706–723 (1964). cyanobiphenyls — a class of liquid crystals — via
(2017). 12. Britz, D. & Strutwolf, J. Digital Simulation in anodic cyanation of 4,4′-dialkoxybiphenyls in
4. Hammerich, O. & Speiser, B. Organic Electrochemistry: Electrochemistry (Springer, 2018). emulsion systems. Acta Chem. Scand. B 31, 813–817
Revised and Expanded (CRC Press, 2015). 13. Leech, M. C., Garcia, A. D., Petti, A., Dobbs, A. P. (1977).
5. Bard, A. J. & Faulkner, L. R. Electrochemical Methods: & Lam, K. Organic electrosynthesis: from academia 21. Eberson, L. & Helgée, B. Studies on electrolytic
Fundamentals and Applications (Wiley, 2000). to industry. React. Chem. Eng. 5, 977–990 (2020). substitution reactions. XIII. Anodic acyloxylation of
6. Elgrishi, N. et al. A practical beginner’s guide to 14. Heard, D. M. & Lennox, A. J. J. Electrode materials aromatic substrates in emulsion systems with the aid
cyclic voltammetry. J. Chem. Educ. 95, 197–206 in modern organic electrochemistry. Angew. Chem. of phase transfer agents. Acta Chem. Scand. B 32,
(2018). Int. Ed. 59, 18866–18884 (2020). 157–161 (1978).
7. Sawyer, D. T., Sobkowiak, A. & Roberts, J. L. 15. Pletcher, D. & Walsh, F. C. Industrial Electrochemistry 22. Teherani, T., Itaya, K. & Bard, A. J. An electrochemical
Electrochemistry for Chemists (Wiley, 1995). (Springer, 1993). study of solvated electrons in liquid ammonia. Nouv. J.
8. Kissinger, P. & Heineman, W. R. Laboratory 16. Colomer, I., Chamberlain, A. E. R., Haughey, M. B. Chim. 2, 481–487 (1978).
Techniques in Electroanalytical Chemistry, Revised & Donohoe, T. J. Hexafluoroisopropanol as a 23. Liu, T. et al. New insights into the effect of pH on
and Expanded 2nd edn (Taylor & Francis, 1996). highly versatile solvent. Nat. Rev. Chem. 1, 0088 the mechanism of ofloxacin electrochemical detection
9. Paddon, C. A., Silvester, D. S., Bhatti, F. L., (2017). in aqueous solution. Phys. Chem. Chem. Phys. 21,
Donohoe, T. J. & Compton, R. G. Coulometry on the 17. McKee, R. H. & Gerapostolou, B. G. Electrolytic 16282–16287 (2019).
voltammetric timescale: Microdisk potential-step reduction of nitro compounds in concentrated 24. Izutsu, K. Electrochemistry in Nonaqueous Solutions
chronoamperometry in aprotic solvents reliably aqueous salt solutions. Trans. Electrochem. Soc. 68, (Wiley, 2011).
measures the number of electrons transferred in an 329 (1935). 25. Kathiresan, M. & Velayutham, D. Ionic liquids as
electrode process simultaneously with the diffusion 18. McKee, R. H. & Brockman, C. J. A new method for an electrolyte for the electro synthesis of organic
coefficients of the electroactive spec. Electroanalysis electro-organic reductions. Trans. Electrochem. Soc. compounds. Chem. Commun. 51, 17499–17516
19, 11–22 (2007). 62, 203 (1932). (2015).

NAture Reviews | CHEMiStry volume 6 | April 2022 | 285

0123456789();:
Reviews

26. Torriero, A. A. J. Electrochemistry in Ionic Liquids pure methyl hydrogen (R)-3-methylglutarate. 55. Shono, T., Hayashi, J., Omoto, H. & Matsumura, Y.
(Springer, 2015). Tetrahedron 41, 627–633 (1985). The migratory aptitude in the anodic oxidation
27. Comminges, C., Barhdadi, R., Laurent, M. & 42. Jensen-Korte, U. & Schäfer, H. -J. Pheromone, 7. of β-hydroxycarboxylic acids, and a new synthesis of
Troupel, M. Determination of viscosity, ionic Kolbe-Synthese v­on 2­9-­te­rt­-B­ut­yl­di­me­th­yl­si­ly­lo­xy­-3, di-muscone. Tetrahedron Lett. 18, 2667–2670
conductivity, and diffusion coefficients in some binary 11-dimethyl-1-nonacosen, einer Schlüsselverbindung (1977).
systems: ionic liquids + molecular solvents. J. Chem. zur Darstellung eines optisch aktiven Sexuallockstoffes 56. Lin, D. Z. & Huang, J. M. Electrochemical
Eng. Data 51, 680–685 (2006). der Deutschen Hausschabe. Liebigs Ann. Chem. N-formylation of amines via decarboxylation of
28. Schäfer, H. J. Recent contributions of Kolbe 1982, 1532–1542 (1982). glyoxylic acid. Org. Lett. 20, 2112–2115 (2018).
electrolysis to organic synthesis. Top. Curr. Chem. 43. Seebach, D. Preparation of enantiomerically pure 57. Schäfer, H. J., Harenbrock, M., Klocke, E., Plate, M.
152, 91–151 (1990). compounds employing anodic oxidations of carboxylic & Weiper-Idelmann, A. Electrolysis for the benign
29. Tanaka, H., Kuroboshi, M. & Torii, S. in Organic acids – A late review of research done in the 1980ies. conversion of renewable feedstocks. Pure Appl. Chem.
Electrochemistry 5th edn 1267–1307 (CRC Press, Helv. Chim. Acta 102, e1900072 (2019). 79, 2047–2057 (2007).
2015). 44. Brecht-forster, A., Fitremann, J. & Renaud, P. 58. Dos Santos, T. R., Harnisch, F., Nilges, P. &
30. Moeller, K. D. Synthetic applications of anodic Synthesis of (±)-nephromopsinic, (−)-phaseolinic, Schröder, U. Electrochemistry for biofuel generation:
electrochemistry. Tetrahedron 56, 9527–9554 and (−)-dihydropertusaric acids. Helv. Chim. Acta 85, transformation of fatty acids and triglycerides
(2000). 3965–3974 (2002). to diesel-like olefin/ether mixtures and olefins.
31. Schäfer, H. J. Recent synthetic applications of the 45. Becking, L. & Schäfer, H. J. Pyrrolidines by ChemSusChem 8, 886–893 (2015).
Kolbe electrolysis. Chem. Phys. Lipids 24, 321–333 intramolecular addition of Kolbe radicals generated 59. Meyers, J. et al. Electrochemical conversion of a
(1979). from β-allylaminoalkanoates. Tetrahedron Lett. 29, bio-derivable hydroxy acid to a drop-in oxygenate
32. Wiebe, A. et al. Electrifying organic synthesis. 2797–2800 (1988). diesel fuel. Energy Environ. Sci. 12, 2406–2411
Angew. Chem. Int. Ed. 57, 5594–5619 (2018). 46. Lebreux, F., Buzzo, F. & Marko, I. E. Studies in (2019).
33. Hayrapetyan, D., Shkepu, V., Seilkhanov, O. T., the oxidation of carboxylic acids: new twists for 60. Holzhäuser, F. J. et al. Electrochemical cross-coupling
Zhanabil, Z. & Lam, K. Electrochemical synthesis of an old reaction. Synthesis of various cyclic systems of biogenic di-acids for sustainable fuel production.
phthalides via anodic activation of aromatic carboxylic and substituted orthoesters. ECS Trans. 13, 1 (2008). Green Chem. 21, 2334–2344 (2019).
acids. Chem. Commun. 53, 8451–8454 (2017). 47. Lebreux, F., Buzzo, F. & Markó, I. E. Synthesis of 61. Lam, K. & Geiger, W. E. Anodic oxidation of disulfides:
34. Dickinson, T. & Wynne-Jones, W. F. K. Mechanism five- and six-membered-ring compounds by detection and reactions of disulfide radical cations.
of Kolbe’s electrosynthesis. Part 3. — Theoretical environmentally friendly radical cyclizations using J. Org. Chem. 78, 8020–8027 (2013).
discussion. Trans. Faraday Soc. 58, 400–404 (1962). Kolbe electrolysis. Synlett 2008, 2815–2820 (2008). 62. Swarts, J. C., Nafady, A., Roudebush, J. H.,
35. Schäfer, H. J. Anodic and cathodic CC-bond formation. 48. Hofer, H. & Moest, M. Ueber die Bildung von Trupia, S. & Geiger, W. E. One-electron oxidation
Angew. Chem. Int. Ed. Engl. 20, 911–934 (1981). Alkoholen bei der Elektrolyse fettsaurer Salze. of ruthenocene: reactions of the ruthenocenium ion
36. Nuding, G., Vögtle, F., Danielmeier, K. & Steckhan, E. Justus Liebigs Ann. Chem. 323, 284–323 (1902). in gentle electrolyte media. Inorg. Chem. 48,
Rodlike molecules by Kolbe electrolysis. Synthesis 49. Yoshikawa, M., Wang, H. K., Tosirisuk, V. & Kitagawa, I. 2156–2165 (2009).
1996, 71–76 (1996). Chemical modification of oleanene-oligoglycosides by
37. Seebach, D. & Renaud, P. Chirale Synthesebausteine means of anodic oxidation. Chem. Pharm. Bull. 30, Acknowledgements
durch Kolbe-Elektrolyse enantiomerenreiner 3057–3060 (1982). The authors are grateful to the Engineering and Physical
β-Hydroxy-carbonsäurederivate. (R)- und (S)-Methyl- 50. Kitagawa, I., Kamigauchi, T., Ohmori, H. & Yoshikawa, M. Sciences Research Council (grant EP/s017097/1 awarded to
sowie (R)-Trifluormethyl-γ-butyrolactone und - Saponin and Sapogenol. XXIX. Selective cleavage of K.L. and M.C.L.) and the University of Greenwich for their
δ-valerolactone. Helv. Chim. Acta 68, 2342–2349 the glucuronide linkage in oligoglycosides by anodic financial support.
(1985). oxidation. Chem. Pharm. Bull. 28, 3078–3086 (1980).
38. Stang, C. & Harnisch, F. The dilemma of supporting 51. Pergola, F., Nucci, L., Pezzatini, G., Wei, H. & Author contributions
electrolytes for electroorganic synthesis: a case study Guidelli, R. Direct electro-oxidation of d-gluconic acid The authors contributed equally to all aspects of the article.
on Kolbe electrolysis. ChemSusChem 9, 50–60 (2016). to d-arabinose. Electrochim. Acta 39, 1415–1417
39. Matzeit, A. et al. Radical tandem cyclizations by (1994). Competing interests
anodic decarboxylation of carboxylic acids. Synthesis 52. Thomas, H. G. & Katzer, E. Acylale durch anodische The authors declare no competing interests.
11, 1432–1444 (1995). oxydation von α-alkoxy-carbonsäuren. Tetrahedron
40. Bestmann, H. J. et al. Pheromone, 57. Lett. 15, 887–888 (1974). Peer review information
Synthese Methylen-unterbrochener Lepidopteren- 53. Stapley, J. A. & BeMiller, J. N. The Hofer–Moest Nature Reviews Chemistry thanks R. Brown and the other,
Polyenpheromone und Strukturanaloger. decarboxylation of d-glucuronic acid and anonymous, reviewer(s) for their contribution to the peer
Acetylensynthese, Wittig-Reaktion und Kolbe- d-glucuronosides. Carbohydr. Res. 342, 610–613 review of this work.
Elektrolyse. Liebigs Ann. Chem. 1987, 417–422 (2007).
(1987). 54. Torii, S., Okamoto, T. & Tanaka, H. Electrolytic Publisher’s note
41. Rossi, R., Carpita, A. & Chini, M. Synthesis of the two decarboxylation reactions. I. Electrosynthesis of Springer Nature remains neutral with regard to jurisdictional
enantiomers of the sex pheromone of Diabrotica γ-substituted butyrolactones and γ-substituted claims in published maps and institutional affiliations.
Undecimpunct at a Howardi and of chiral precursors α,β-butenolides from γ-substituted paraconic acids.
of other pheromones starting from enantiomerically J. Org. Chem. 39, 2486–2488 (1974). © Springer Nature Limited 2022

286 | April 2022 | volume 6 www.nature.com/natrevchem

0123456789();: