Applied Materials Today 18 (2020) 100404

Contents lists available at ScienceDirect

Applied Materials Today

journal homepage: www.elsevier.com/locate/apmt

Discussion

Some thoughts about reporting the electrocatalytic performance of

nanomaterials
Danlei Li, Christopher Batchelor-McAuley, Richard G. Compton ∗
Physical and Theoretical Chemistry Laboratory, Department of Chemistry, University of Oxford, South Parks Road, Oxford OX1 3QZ, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: In reading the ever expanding literature on electrocatalysts, we have become startled by the weakness of
Received 12 April 2019 the electrochemistry often presented, which in some (many?) cases entirely negates the value of the work.
Accepted 24 May 2019 In particular, we have been stimulated to consider the topic of this article by an Editorial (Voiry, Chhowalla
et al., 2018, 12, 9635-9638) in ACS Nano which recently provided ‘guidance’ on the ‘best practices’ for
Keywords: the measuring and reporting the activity of new electrocatalytic materials. From an electrochemical
Characterization of nanomaterials
perspective, at least, contrasting views need to be presented since the suggestions provided are, in places,
Tafel analysis
at odds with conventional wisdom or, more bluntly stated, simply wrong! In the following we do not
Onset potential
Standard potential
seek to provide an alternative set of ‘best practice guidelines’ nor a ‘set of materials characterization
Exchange current density requisites’ – this is likely ultimately an appropriate activity for an IUPAC committee – but rather correct,
Overpotential amplify and develop the discussion provided by the editors of ACS Nano highlighting areas where we
Electrochemical surface area believe additional input is desirable and helpful.
© 2019 Elsevier Ltd. All rights reserved.

In this article we focus on six topics that relate to the recommen- For this reaction the Nernst equation for the electrode potential
dations made by Voiry, Chhowalla et al. in their recent ACS Nano (E) is:
Editorial[1][2]. In each section we start by making a brief state-
RT aB
ment that we believe is correct but different to that made by Voiry, E = EA/B − ln (2)
F aA
Chhowalla et al. This statement is then followed by a more in-depth
discussion and exploration of the issue at hand. where ai is the activity for the ith species and EA/B is the standard
redox potential of the A/B redox couple. This latter value is directly
related to the standard Gibbs energy of the reaction. If we express
1. Standard, formal and equilibrium potentials the solution phase activities of these electroactive solutes in terms
of their activity coefficients on a concentration basis then using
The standard redox potential is not under most conditions equiv- the expression ai =  i ci /c , where c is the standard concentration
alent to the equilibrium potential. The standard potential is the (1 mol dm−3 ) we can write [3]:
equilibrium electrode potential defined at standard conditions (STP) RT B RT cB
with all the electroactive species at unit activity and gases at standard E = EA/B − ln − ln (3)
F A F cA
pressure (assuming ideal behaviour of the gas).
The following definitions are vital to both the Editorial in ACS where  i is the activity coefficient, and the formal (aka conditional)
Nano[1] and the aspects discussed later. To maintain a level of potential (Ef,A/B ) of the A/B couple is then:
clarity we focus on the example of a simple one-electron transfer
process as given by: RT B
Ef,A/B = EA/B − ln (4)
F A

A(sol) + e− = B(sol) (1) First, the formal potential is not a fixed value solely dependent
on the A/B couple but is specific to a given set of experimental
conditions. If the activity coefficients of the electroactive species
change, the formal potential is correspondingly altered. For this
∗ Corresponding author. one-electron case only if  B / A = 1 does the standard potential
E-mail address: [email protected] (R.G. Compton). equal the formal potential. Second, in contrast to the above case

https://doi.org/10.1016/j.apmt.2019.05.011
2352-9407/© 2019 Elsevier Ltd. All rights reserved.
2 D. Li et al. / Applied Materials Today 18 (2020) 100404

where both the reactant and product are solutes, if the reaction summing to zero current. At this potential the current in either the
involves a gas then the standard and formal potentials often dif- anodic or cathodic direction is given by:
fer markedly [4]. This arises due to the fact that most non-polar
gases (such as hydrogen or oxygen) have a low solubility in aque- 0
i0,A/B = FAkA/B cA1−˛ cB˛ (8)
ous solution [5]. Hence, at one bar pressure the gas is commonly
only present in solution at millimolar concentrations; in contrast,
the formal potential is defined at (a possibly hypothetical) [6] unit This so called “exchange current” has little physical significance;
0
the rate of the reaction is controlled by both kA/B and ˛ as expressed
concentration. As such the solution phase dissolved gas concentra-
tion is often three orders of magnitude different between these two by Eq. (6).
definitions. Implicit in the use of Eq. (7) is the assumption that both the
For a simple one-electron transfer process we can define the reduced and oxidized species are present in the bulk solution phase;
equilibrium potential as [7]: as is the case for classical work [11] on the proton/hydrogen redox
couple where both acid and dissolved hydrogen gas are present:
RT cB H+ + e−  1/2H2 .Voiry, Chhowalla et al. emphasize the importance
Eeq,A/B = Ef,A/B − ln (5)
F cA of estimating/measuring the exchange current (i0 ), as was under-
taken classically, to quantify a the activity of catalysts [12]. For the
Even for this simple one-electron transfer reaction, Eq. (5) is explic- H+ /H2 reaction we can usefully define and measure the associated
itly sensitive to the ratio of the oxidized (A) and reduced (B) species. exchange current (i0 ). But, we can see even for the one-electron
If a solution contains one order of magnitude higher concentration example given above, if the concentration of product (cB ) is zero
of the product as compared to the reactant the equilibrium poten- then neither the equilibrium potential (Eq. (5)) nor the exchange
tial will be ∼59.1 mV negative of the formal potential of the system. current (Eq. (8)) is defined!
This may seem a moot point but the differences between these If only the reactant is present in bulk solution we face the
definitions become even more pronounced if an electrode reaction question, how do we measure i0 ? This is a very commonly encoun-
involves multiple steps. For example if the reaction involves the tered situation, for instance when studying the hydrogen evolution
transfer of one proton per electron then the equilibrium potential reaction or the reduction of carbon dioxide where often neither
varies with 59.1 mV pH−1 . Electrocatalytic experiments performed hydrogen nor the products (formate, oxalate, carbon monoxide,
at near neutral pH are often far from being under standard condi- etc.) of the carbon dioxide reduction process are initially present
tions! in the bulk solution phase. Hence, any attempts to provide general
guidelines for quantifying the properties of electrocatalytic materi-
2. How should we quantify electrode-kinetics? als, which are predicated on the measurement of a quantity (i0 ) that
is only relevant to a special situation, is not helpful. Rather for indi-
For a half-cell reaction the idea of the exchange current (i0 ) is only vidual half reactions, A + e− → B or A − e− → B, the measurement of
directly relevant in practical situations where both the electroactive k0 and ˛ (or equally [1 − ˛]) is preferred.
product and reactant are present in the bulk solution phase and the We note that at any given potential the Principle of Microscopic
reaction is to some extent reversible. Consequently, we cannot use i0 Reversibility [13] requires that the anodic and cathodic reactions
as parameter for generally defining electrocatalytic activity. have the same transition state; consequently, for a one-electron
In electrochemistry there are two common formulations of transfer process ˛ + ˇ = 1 when measured at the same potential. How-
the Butler–Volmer equation. First the more general form used by ever, for couples with significant irreversibility the values of ˛ and ˇ
physical electrochemists which is close to that first derived by are commonly evaluated at different potentials for the cathodic and
Erdey-Gruz and Volmer [8]: anodic processes respectively. Since the Principle of Microscopic
     Reversibility only rigorously holds at the same potential, when ˛
0
i = −FAkA/B cA,0 exp
−˛F
(E − Ef,A/B ) − cB,0 exp
ˇF
(E − Ef,A/B ) (6)
and ˇ are measured at different potentials they need not add to
RT RT unity. At other potentials the relative magnitudes of ˛ and ˇ may
vary as the character of the transition state changes, for example
i is the current (A), k0 is the standard (strictly “formal” [9]) elec- as solvation/adsorption or the double layer structure alters with
trochemical rate constant (m s−1 ) and the ci,0 is the concentration the applied potential. Thus generally for experimentally measured
of the ith species at the electrode surface. The second form, origi- transfer coefficients ˛ + ˇ = / 1. Hence, linear extrapolation of the
nating from work by Laidler et al. [10], is regularly employed in the cathodic and anodic branches of a voltammograms to give Eeq and
materials and engineering literature and is given by the expression: i0 will be in error.
Finally, for multi-step reactions the mechanism and hence the
  −˛F    associated reaction product is often potential dependent. Consider
ˇF
i = −i0,A/B exp (E − Eeq,A/B ) − exp (E − Eeq,A/B ) two electrode reactions that occur in parallel, a hypothetical exam-
RT RT
ple might be the reduction of CO2 to either CO or formate. If the
(7)
formation of say, for example, CO on some new catalysts has a
where i0,A/B is the exchange current (A) for the A/B couple. (comparatively) low standard electrochemical rate constant but a
Both expressions assume that ˛ + ˇ = 1. The first major difference high transfer coefficient and the other reaction, in this case leading
between these two equations ((6) and (7)) is the potential against to the formation of formate, has a high standard electrochemical
which they are referenced; in the first the potential is measured rate constant but a low transfer coefficient then the electrochemi-
relative to the formal potential (Eq. (4)) but the second uses the cal reaction product will be potential dependent. At low potentials
equilibrium potential (Eq. (5)). Eq. (7) can be derived from Eq. (6) the reaction with the higher standard electrochemical rate con-
by rearranging the definition of the equilibrium potential (Eq. (5)) to stant will dominate and in this example we will yield formate as
give a definition of the formal potential in terms of the equilibrium a product. Conversely at high overpotentials the relative rates of
potential and by substituting this definition into Eq. (6). Further for these two processes will have switched and the dominant reaction
Eq. (7) the surface concentration terms (ci,0 ) are assumed equal product will be the formation of CO. For multi-step processes the
to their value in bulk. At the equilibrium potential the anodic nature of the electrode reaction mechanism will often change as a
and cathodic currents are equal in size but of opposite direction function of the applied potential!
 D. Li et al. / Applied Materials Today 18 (2020) 100404 3

3. What is an overpotential? Eq. (11) is only exactly correct in the case that the diffusion coeffi-
cients of the reduced and oxidized species are equal [15]. However
In the case where only the electroactive reagent is present in solu- the salient point is that, as a result of conservation of mass at the
tion and under a steady-state mass-transport regime we can use the electrode surface and due to the reactions stoichiometry, the posi-
shift in the voltammetric half-wave potential from that expected for a tion of the reversible voltammetric wave is, as measured by the
reversible process as a measure of the applied overpotential. However, voltammetric half-wave potential, related to the formal potential
precise calculation of this expected reversible half-wave potential for for the reaction but varies as a function of bulk acid concentration
a given set of experimental conditions is not necessarily facile; this is [16]!
especially true for multi-step reactions [4].
The problems outlined in the previous section relating to the 4. What is an onset potential?
exchange current also underlie an issue in the IUPAC definition of
overpotential. IUPAC rigidly define the overpotential in relation to The onset potential is neither a thermodynamically nor kinetically
the equilibrium potential (e = E − Eeq ) [7] but from Eq. (5) if either well-defined parameter; consequently, it is unhelpful when compar-
cA or cB is zero then the overpotential by this definition cannot ing activity of electrocatalysts between laboratories and hence across
be defined! This does not however imply that the thermodynam- different experimental setups. A likely more appropriate method is, as
ics of such a system are also ill-defined, we just need a different is done with ORR catalysts, to report the Faradaic current density at a
definition. Generally, if an electrochemical process is described as given and agreed potential [17].
‘reversible’ under a given mass-transport regime this implies that Although the “onset potential” is a widely reported parameter
the electrode surface concentrations of the reduced and oxidized and at first sight appears to have an obvious definition [18] as some-
species are well described via the Nernst equation (i.e. they are thing akin to the ‘lowest overpotential at which a reaction product
locally at equilibrium). is formed at a given electrode under defined conditions’ or ‘the low-
In the literature it is common in cases where only the reac- est overpotential at which the Faradaic current is observed to be
tant is present in the bulk solution to define the overpotential over and above the measured background’ these statements belie
relative to that of the formal potential or even in some cases the the true complexity of the issue as the following explains. First,
standard potential (s = E − E ). For a simple one-electron transfer consider an irreversible one-electron reduction process for which:
process this definition of overpotential is clear and rational with  
Eq. (6) as the inspiration. For example if we consider the reversible −˛F(E − Ef,A/B )
0
i = FAkA/B exp cA,0 (12)
one-electron oxidation of ferrocene methanol: Fc − e−  Fc+ . Under RT
steady-state mass-transport conditions then the voltammetric
half-wave [14] potential is approximately equal to the formal The cathodic current is predicted to asymptotically approach zero
potential. In cases where the diffusion coefficients of the reduced as E → +∞ and to increase exponentially as E → −∞. Defining where
and oxidized species are equal then the reversible half-wave poten- such an exponential rise ‘starts’ necessarily requires some addi-
tial (E1/2 ) is exactly equal to the formal potential of the system. tional arbitrary definitions. Second, the general interpretation of
For the hydrogen evolution reaction (H+ + e−  1/2H2 ) due to the the onset potential is the potential at which a reductive or oxida-
process involving the formation and breaking of chemical bonds tive Faradaic reaction becomes measurable. Immediately it becomes
the situation is slightly different. Here we briefly consider the clear that such a parameter is essentially a measure of the signal-
case of the hydrogen evolution reaction from an aqueous strong to-noise or signal-to-background ratio of the system and hence
acid solution in the absence of dissolved hydrogen in the bulk does not solely reflect the electrocatalytic properties of the material
phase. The Nernst equation describing the electrode potential can under study but is also influenced by the measurement conditions.
be expressed as: To exemplify this point, Fig. 1(a) shows a simulated one-electron
RT cH+ ,0 irreversible reduction at a rotating disc electrode. The employed
E = Ef,H+ /H + ln (9) simulation numerically calculates the voltammetric profile using
2 F cH 0.5 c 0.5
2,0 a fully implicit finite difference method and makes use of the
where the square-root associated with the concentration of hydro- Hale transform [19,20]. The theoretically predicted Faradaic cur-
gen reflects the stoichiometry of the reaction. Under steady-state rent is plotted in black. Also plotted are some realistic values for
conditions, for example as obtained using a rotating disc electrode, the capacitative current of the electrode. In particular we have
at the voltammetric half-wave potential, i.e. where 50% of the avail- approximated the electrodes capacitance as constant with respect
able protons are converted to hydrogen – and again assuming equal to the applied electrode potential and supposed that the specific
diffusion coefficients for the electroactive species – then the surface capacitance has value that is representative for metals in aqueous
concentrations of the reduced (hydrogen) and oxidized (protons) solutions (20 ␮F cm−2 ) [21].
species will be equal to: To calculate the total electrode area we have multiplied the sim-
ulated geometric electrode (Ageo /m2 ) area by a roughness factor, Rf ,
cH+ ,bulk ≈ 2cH+ ,0 ≈ 4cH2 ,0 at E = E1/2 (10) so that Areal = Ageo × Rf . As discussed later, for heterogeneous elec-
where the subscript zero refers to the surface concentrations of trode surfaces another important quantity is the roughness factor
the species. Eq. (10) expresses that at the voltammetric half-wave which accounts only for the real area of the catalyst (Rf,catalyst ).
potential the surface concentration of the protons (cH+ ,0 ) will be As outlined schematically in Fig. 1(b) in the present example a
half of that in the bulk media (cH+ ,bulk ). Moreover, on the basis of perfectly atomically flat electrode would have a roughness factor
the reactions stoichiometry the reduction of half of the protons to of unity; a very well prepared polycrystalline electrode may be
hydrogen at the electrode surface implies that the surface concen- expected to have a roughness factor of ∼1.5 [17] and a thin-film
tration of hydrogen (cH2 ,0 ) is equal to a quarter of the bulk proton modified RDE will have a widely variable roughness factor in the
concentration. Substitution of equality 10 into Eq. (9) yields an range of 10–1000. Thin-film modified electrodes find routine use
expression for the voltammetric half-wave potential (E1/2 ) as equal in electrocatalyst experiments [17], where a catalyst is mixed with
to: a non-catalytic conductive support such as a form of nano-carbon
and added as a thin-layer across the surface of the electrode. The
RT cH+ ,bulk 0.5 capacitative charging of the carbon support and the catalyst will
E1/2 ≈ Ef,H+ /H + ln (11)
2 F c 0.5 both contribute to the total background charging currents. As can
4 D. Li et al. / Applied Materials Today 18 (2020) 100404

Fig. 1. (a) Comparison of the simulated Faradaic current (black) and capacitative current at a rotating disc electrode (5 mm in diameter) with variable roughness factors
(Rf ), Rf = 1 (red), Rf = 10 (blue), Rf = 50 (yellow) and Rf = 100 (green). Inlay is the zoomed-in version with a potential range from −0.3 V to 0 V. Parameters in the simulation:
scan rate  = 0.02 V s−1 , rotation rate ω = 1600 rpm, concentration c = 1 mM, diffusion coefficient D = 1 × 10−9 m2 s−1 , formal electron transfer rate constant k0 = 1 × 10−7 m s−1 ,
formal potential Ef = 0 V, viscosity = 8.9 × 10−7 m2 s−1 , transfer coefficient ˛ = ˇ = 0.5. The capacitative current Icap = Cdl (20 ␮F cm−2 ) × Rf × Ageo × . (b) Schematic of electrode
surface with different roughness factors. The roughness factor Rf = the real electrode area (Areal )/the geometric area (Ageo ). (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)

be seen in the guidelines provided by Kocha et al. [17] in their exper-

iments for a thin-film modified RDE of diameter 5 mm the total
background currents are of the order of 15 ␮A, this corresponds to
an effective roughness factor of approximately 190.
From Fig. 1(a) we can see that if we define the onset potential in
this model case as when the Faradaic current is expected to be equal
in magnitude to the capacitative contribution then the potential at
which this cross over occurs is extremely sensitive to the capaci-
tance of the electrode. Even in this idealized case the value of the
onset potential varies by more than 200 mV for these realistic elec-
trode capacitances. We further comment that the definition of the
capacitance as being a constant, as is implied in the original editorial
text, and as is used above is an over-simplification [22].

5. What is the appropriate Tafel region of the

current-potential plot of a half-cell reaction in which to
analyze a ‘Tafel-slope’?
Fig. 2. Simulated mass-transport corrected Tafel plots for a rotating disc elec-
It is preferable to define a current range relative to the limiting trode. Solid lines represent the electrode with constant capacitances with variable
roughness factors Rf : Rf = 0 (black), Rf = 1 (red), Rf = 10 (blue), Rf = 50 (yellow)
steady-state value as opposed to defining a suitable potential range and Rf = 100 (green). The capacitative current Icap = Cdl (20 ␮F cm−2 ) × Rf × Ageo × .
for meaningful Tafel analysis giving either a transfer coefficient or, Dashed line stands for the electrode with a variable capacitative current (grey
equivalently, a ‘Tafel-slope’. The analysis provided here indicates that dashed) where the variable capacitative current was calculated from Icap = Cdl
provided a suitable background subtraction to remove the capacita- (20 ␮F cm−2 ) × Rf × Ageo ×  × f(E), where E is the electrode potential. Simulation
parameters: scan rate  = 0.02 V s−1 , rotation rate ω = 1600 rpm, concentration
tive current contribution is first performed then the current in the
c = 1 mM, diffusion coefficient D = 1 × 10−9 m2 s−1 , formal electron transfer rate
range between 10 and 80% of the limiting current is suitable for kinetic constant k0 = 1 × 10−7 m s−1 , formal potential Ef = 0 V, viscosity = 8.9 × 10−7 m2 s−1 ,
analysis once a mass-transport correction has been made. If the latter transfer coefficient ˛ = ˇ = 0.5. Itot = Ifarad + Icap . (For interpretation of the references
correction is neglected only currents below ∼20% of the steady-state to colour in this figure legend, the reader is referred to the web version of this article.)
current are suitable for use.
In a Tafel plot, the log of the current, log10 |i|, is plotted against
the applied potential, E. If we assume the reaction is fully irre- information on alternate methods of presenting Tafel plots and
versible and well described by the Tafel equation (Eq. (12)) then analysis see Ref. [25].
we can see for a one-electron transfer process that such a semi- A linear Tafel plot of log10 |i| vs E requires that, on the basis
log plot will have a slope that is equal to −˛F/(RT × ln 10) and an of Eq. (6), the process is fully irreversible and that the sur-
intercept (at zero overpotential, f ) equal to log10 (FAk0 cA ). Orig- face concentrations of the electroactive species remain constant
inally electrocatalysis experiments were performed using a form throughout the potential range of analysis. In practice the replen-
of current-interrupt technique known as the commutator method ishment of the active species via diffusion is slow and distortions
[23,24]. Here a desired current density was driven to occur on a arise from mass-transport limitations. Consequently, it is neces-
working electrode, the cell was subsequently disconnected and the sary to ‘correct’ for these changes in the redox active species at
rapidly changing cell potential measured against a reference cell. the electrode surface. The rotating disc electrode is to a reasonable
Measurement of the potential of the disconnected cell as a func- approximation [26,27] uniformly accessible. Consequently a plot
tion of the disconnection time allowed the original cell potential to of log10 (1/I − 1/Ilim ) against the applied potential allows for these
be inferred by extrapolation. Hence in reporting data in terms of changes in the surface concentration of the redox active species
current/potential graphs on the y-axis the potential (the measured during the course of the scan to be suitably accounted for. The
variable) was plotted against the current (the controlled variable) need to make a mass-transport correction has long been advocated
[24]. Modern experiments tend to be performed potentiometrically [28,29].
and hence one might expect the axes of the Tafel plot to be swapped, Fig. 2 plots an example mass-transport corrected Tafel plot for
this is virtually never the case; old habits die hard. For more a simulated RDE experiment. In this simulation we have assumed
 D. Li et al. / Applied Materials Today 18 (2020) 100404 5

that the total current can be expressed simply as the sum of both the determination of the magnitude of this value is not necessarily as
Faradaic and non-Faradaic contributions (itot = ifarad + icap ). For the facile as it may seem; this is especially true if the catalytic surface is
solid lines in Fig. 2 the contribution of the capacitance is assumed not uniformly accessible as is the case for porous electrode surfaces.
to be a constant (icap = Cdl × Rf × Ageo × ), alternatively the dotted Related to these values is the (dimensionless) catalyst roughness
line assumes that the capacitance of the electrode varies linearly factor, this value is the surface area of the catalyst per geomet-
as a function of the applied potential, icap = Cdl × Rf × Ageo ×  × f(E), ric area of the electrode. In some contexts the roughness factor
where the function f(E) is a dimensionless scalar that varies lin- may also, vide supra, refer to the total surface area of the electrode
early between 1 and 2 across the simulated voltammetric potential relative to its geometric area.
range. In the absence of a background correction to remove the The measurement of the surface area of the catalysts is not nec-
capacitative current, Fig. 2 shows how the mass-transport cor- essarily straightforward. Note that “best practice” documents in the
rected Tafel plot is distorted by the presence of a capacitative field of ORR tend to only very briefly [17], if at all [29], mention that
“background” current. Capacitative currents are symptomatic of such electrode capacitance measurements (or the measurement of
the used voltammetric technique. This result first highlights the other surface processes such as the under-potential deposition of
importance of background correction to remove the non-Faradaic hydrogen) need to be made using an analogue [35] potentiostat.
contribution from the voltammetric data. It also gives a clear indi- In the case of measuring platinum electrochemical surface areas,
cation as to how sensitive the data will be to the quality of this major errors can be made if staircase voltammetry is used and, of
correction. It is on this basis that Mayrhofer et al. previously pro- course, almost all modern commercial potentiostats provide stair-
posed that the kinetically useful part of an RDE voltammogram is case voltammetry as the default technique, so this can lead to the
the mass-transport corrected current between 10 and 80% of the underestimation of the platinum surface area and will hence cause
steady-state value [29] Kocha et al. [30] were more cautious and an overestimation regarding the material’s specific activity. As an
on the basis of work by Vidal-Iglesias et al. [31], advised that only aside similar issues arise with protein film voltammetry [36]. This
the 10–50% current regime is useable. issue with staircase voltammetry is often covered in potentiostat
manuals [37]; but who reads a manual [38]?
6. Units and electrochemical surface areas Some of the above is opinion and no doubt others in the field will
dispute aspects of it. Neither is the above exhaustively critical of the
The size of the electrochemically available surface area of the cat- editorial written by Voiry, Chhowalla et al. [1]. Some comments will
alyst is an important factor in determining meaningful information be obvious to many in the electrochemical field but we hope that we
regarding its catalytic abilities if electron transfer (not mass-transport) have identified some of the important points for the non-specialist.
is rate determining. Two chemically relevant quantities are the rough-
ness factor (a dimensionless measure of the surface area of the catalyst
Conflict of interest
per geometric area of the electrode) and the electrochemical surface
area (the surface area of the catalyst per gram of catalyst). Measure-
The authors declare no competing financial interest.
ment of the real surface area of the catalyst is often experimentally
challenging and is in some cases sensitive to how the measurement is
made. Acknowledgements
Voiry, Chhowalla et al. [1] present both the electrochemical sur-
face area (ECSA) and the turnover frequency (TOF) as dimensionless D. Li thanks for the China Scholarship Council and the University
parameters! In the guidelines the ECSA is given as the dimension- of Oxford for supporting her DPhil research.
less ratio of the specific double layer capacitance (Cdl ) relative to a
given reference surface specific capacitance, this definition seems References
to follow a previous article in ACS Catalysis [32]. Similarly the TOF
is given as “nproduct /nsite ” [1]; how these two terms (nproduct and [1] D. Voiry, M. Chhowalla, Y. Gogotsi, N.A. Kotov, Y. Li, R.M. Penner, R.E. Schaak,
P.S. Weiss, Best practices for reporting electrocatalytic performance of
nsite ) should be defined and measured is left to the reader. We note nanomaterials, ACS Nano 12 (10) (2018) 9635–9638.
as follows, first, Voiry, Chhowalla et al. have interpreted the ECSA [2] This article was first offered to the editors of ACS Nano as a comment on their
as a roughness factor with a subsequent confusing conflation of recent proposals on the ’Best Practices for Reporting Electrocatalytic
Performance of Nanomaterials’ and was rejected as it is ‘not suitable for the
terminology in the text. Second, we comment that it is essential
interdisciplinary readership of ACS Nano’ [Personal Communication] despite
for a surface area to contain the unit of length squared; similarly a it covering identical material to the editors’ own best practice document.
quantity labelled as a “frequency” has units of reciprocal time. As [3] In Eq. (3) the standard concentration terms have cancelled out, in cases where
the stoichiometry of the reaction is not unity this is not the case.
a consequence of these non-conventional definitions some of the
[4] X. Jiao, C. Batchelor-McAuley, E. Kätelhön, J. Ellison, K. Tschulik, R.G. Compton,
derived expressions provided Voiry, Chhowalla et al. may also be The subtleties of the reversible hydrogen evolution reaction arising from the
usefully reconsidered. nonunity stoichiometry, J. Phys. Chem. C 119 (17) (2015) 9402–9410.
In the literature the definition of the ECSA may vary depending [5] E. Wilhelm, R. Battino, R.J. Wilcock, Low-pressure solubility of gases in liquid
water, Chem. Rev. 77 (2) (1977) 219–262.
on the context in which it is being used; however, following the ORR [6] Note the formal potential is still relevant for understanding the
field the ECSA is probably most usefully defined as the area of the thermodynamics of electrochemical processes involving dissolved gases.
catalyst per gram of material (m2 g−1 catalyst
) [29]. This value in combi- [7] E.R. Cohen, T. Cvitas, J.G. Frey, B. Holström, K. Kuchitsu, R. Marquardt, I. Mills,
F. Pavese, M. Quack, J. Stohner, H.L. Strauss, M. Takami, A.J. Thor, Quantities,
nation with the specific activity of the catalyst (A m−2 catalyst
), [33] the in: Units and Symbols in Physical Chemistry, 3rd Edition, Royal Society of
Chemistry, 2007.
catalytic Faradaic current per catalyst area), yields the mass activity
[8] T. Erdey-Grúz, M. Volmer, Zur theorie der wasserstoff überspannung,
for the material (A g−1
catalyst
). These latter two values vary as a func- Zeitschrift für physikalische Chemie 150 (1) (1930) 203–213.
tion of potential. Consequently to get a relative measure of activity [9] In strict accordance with IUPAC k0 as defined here is the formal
electrochemical rate constant as the used reference potential is the formal not
these values are often reported at a given potential. It is the value for the standard potential see Ref. [7].
the mass activity that most directly translates to the cost of a plat- [10] H. Eyring, S. Glasstone, K.J. Laidler, Application of the theory of absolute
inum based fuel cell device [34]. But in terms of physico-chemical reaction rates to overvoltage, J. Chem. Phys. 7 (11) (1939) 1053–1065.
[11] N. Pentland, J.M. Bockris, E. Sheldon, Hydrogen evolution reaction on copper,
insight the important factor is arguably the specific activity of gold, molybdenum, palladium, rhodium, and iron mechanism and
the catalyst. This value of the specific activity of the catalysts measurement technique under high purity conditions, J. Electrochem. Soc.
gives a measure of the rate of reaction per unit area, accurate 104 (3) (1957) 182–194.
6 D. Li et al. / Applied Materials Today 18 (2020) 100404

[12] S. Trasatti, Work function, electronegativity, and electrochemical behaviour of [24] F.P. Bowden, E.K. Rideal, The electrolytic behaviour of thin films. Part
metals: III. Electrolytic hydrogen evolution in acid solutions, J. Electroanal. I.—Hydrogen, Proc. R. Soc. 120 (784) (1928) 59.
Chem. Interfacial Electrochem. 39 (1) (1972) 163–184. [25] D. Li, C. Lin, C. Batchelor-McAuley, L. Chen, R.G. Compton, Tafel analysis in
[13] W. Thomson, IX.—On the dynamical theory of heat. Part V. Thermo-electric practice, J. Electroanal. Chem. 826 (2018) 117–124.
currents, Trans. R. Soc. Edinb. 21 (Part 1) (1853–1857) 123–171. [26] J. Newman, Current distribution on a rotating disk below the limiting current,
[14] L. Meites, P. Zuman, H.W. Nurnberg, Recommended terms, symbols, and J. Electrochem. Soc. 113 (12) (1966) 1235–1241.
definitions for electroanalytical chemistry (IUPAC Recommendations 1985), [27] W.H. Smyrl, J. Newman, Limiting current on a rotating disk with radial
Pure Appl. Chem. 57 (1985) 1491. diffusion, J. Electrochem. Soc. 118 (7) (1971) 1079–1081.
[15] For the proton/hydrogen redox couple in aqueous solution the diffusion [28] W.J. Albery, Electrode Kinetics, Clarendon Press, 1975.
coefficients of the reduced and oxidised species differ by approximately a [29] K.J.J. Mayrhofer, D. Strmcnik, B.B. Blizanac, V. Stamenkovic, M. Arenz, N.M.
factor of two. This difference in the diffusion coefficients will, for a rotating Markovic, Measurement of oxygen reduction activities via the rotating disc
disc electrode experiment, lead to the half-wave potential being shifted by electrode method: from Pt model surfaces to carbon-supported high surface
approximately −5 mV from that predicted on the basis of Eq. (11). For more area catalysts, Electrochim. Acta 53 (7) (2008) 3181–3188.
information on this additional mass-transport correction to the predicted [30] K. Shinozaki, J.W. Zack, R.M. Richards, B.S. Pivovar, S.S. Kocha, Oxygen
half-wave potential, the reader is directed towards Ref. [4]. In the SI of this reduction reaction measurements on platinum electrocatalysts utilizing
reference the additionally required correction for the half-wave potential is rotating disk electrode technique: I. Impact of impurities, measurement
derived (Eq. (12) of SI in Ref. [4]), simply requiring the definition for the protocols and applied corrections, J. Electrochem. Soc. 162 (10) (2015)
hydrodynamic layer thickness to be substituted into this expression. F1144–F1158.
[16] It is important to recognise that this sensitivity to the bulk acid concentration [31] F.J. Vidal-Iglesias, J. Solla-Gullón, V. Montiel, A. Aldaz, Errors in the use of the
is not the same as that of the so-called ‘Reversible Hydrogen Electrode’. To this Koutecky–Levich plots, Electrochem. Commun. 15 (1) (2012)
end we highlight the square-root present in the natural log term in Eq. (11). 42–45.
[17] S.S. Kocha, K. Shinozaki, J.W. Zack, D.J. Myers, N.N. Kariuki, T. Nowicki, V. [32] E.L. Clark, J. Resasco, A. Landers, J. Lin, L.-T. Chung, A. Walton, C. Hahn, T.F.
Stamenkovic, Y. Kang, D. Li, D. Papageorgopoulos, Best practices and testing Jaramillo, A.T. Bell, Standards and protocols for data acquisition and reporting
protocols for benchmarking ORR activities of fuel cell electrocatalysts using for studies of the electrochemical reduction of carbon dioxide, ACS Catal. 8 (7)
rotating disk electrode, Electrocatalysis 8 (4) (2017) 366–374. (2018) 6560–6570.
[18] A. Maljusch, E. Ventosa, R.A. Rincón, A.S. Bandarenka, W. Schuhmann, [33] Here A is the unit Ampere and not the symbol representing the electrode area
Revealing onset potentials using electrochemical microscopy to assess the which has units of m2 .
catalytic activity of gas-evolving electrodes, Electrochem. Commun. 38 (2014) [34] A. Kongkanand, M.F. Mathias, The priority and challenge of high-power
142–145. performance of low-platinum proton-exchange membrane fuel cells, J. Phys.
[19] R.G. Compton, E. Laborda, K.R. Ward, Understanding Voltammetry: Simulation Chem. Lett. 7 (7) (2016) 1127–1137.
Of Electrode Processes, World Scientific Publishing Company, 2013. [35] C. Batchelor-McAuley, M. Yang, E.M. Hall, R.G. Compton, Correction factors for
[20] J.M. Hale, Transients in convective systems: I. Theory of galvanostatic, and the analysis of voltammetric peak currents measured using staircase
galvanostatic with current reversal transients, at a rotating disc electrode, J. voltammetry, J. Electroanal. Chem. 758, (2015) 1–6.
Electroanal. Chem. (1959) 6 (3) (1963) 187–197. [36] H.A. Heering, M.S. Mondal, F.A. Armstrong, Using the pulsed nature of
[21] B.B. Damaskin, A.N. Frumkin, Potentials of zero charge, interaction of metals staircase cyclic voltammetry to determine interfacial electron-transfer rates
with water and adsorption of organic substances—III. The role of the water of adsorbed species, Anal. Chem. 71 (1) (1999) 174–182.
dipoles in the structure of the dense part of the electric double layer, [37] M. Autolab, Application Note: Comparison between Staircase Cyclic
Electrochim. Acta 19 (4) (1974) 173–176. Voltammetry and Cyclic Voltammetry Linear Scan, https://www.metrohm.
[22] H. Gerischer, R. McIntyre, D. Scherson, W. Storck, Density of the electronic com/en-gb/applications/AN-EC-007.
states of graphite: derivation from differential capacitance measurements, J. [38] A.L. Blackler, R.G. Vesna Popovic, M. Helen Thompson, IG Nobel-Literature
Phys. Chem. 91 (7) (1987) 1930–1935. Prize: For Documenting That Most People Who Use Complicated Products Do
[23] E. Newbery, Overvoltage, J. Chem. Soc. Trans. 105 (1914) 2419–2435. Not Read The Instruction Manual, 2018.