doi.org/10.26434/chemrxiv.7125107.v1

Flow Battery Electroanalysis 2: Influence of Surface Pretreatment onFe(III/II) Redox Chemistry at Carbon ElectrodesTejal Sawant, James McKone

Submitted date: 24/09/2018 • Posted date: 25/09/2018Licence: CC BY-NC-ND 4.0Citation information: Sawant, Tejal; McKone, James (2018): Flow Battery Electroanalysis 2: Influence ofSurface Pretreatment on Fe(III/II) Redox Chemistry at Carbon Electrodes. ChemRxiv. Preprint.

Redox flow batteries are attractive for large-scale electrochemical energy storage, but sluggish electrontransfer kinetics often limit their overall energy conversion efficiencies. In an effort to improve ourunderstanding of these kinetic limitations in transition metal based flow batteries, we used rotating-diskelectrode voltammetry to characterize the electron-transfer rates of the Fe3+/2+ redox couple at glassy carbonelectrodes whose surfaces were modified using several pre-treatment protocols. We found that surfaceactivation by electrochemical cycling in H2SO4(aq) electrolyte resulted in the fastest electron-transfer kinetics:j0 = 0:90 mA/cm2 in an electrolyte containing 10 mM total Fe. By contrast, electrodes that were chemicallytreated to either remove or promote surface oxidation yielded rates that were at least an order of magnitudeslower: j0 = 0:07 and 0:08 mA/cm2, respectively. By correlating these findings with X-ray photoelectronspectroscopy data, we conclude that Fe3+/2+ redox chemistry is catalyzed by carbonyl groups whose surfaceconcentrations are increased by electrochemical activation.

File list (2)

download fileview on ChemRxivFlow Battery 2-Main.pdf (1.13 MiB)

download fileview on ChemRxivFlow Battery 2- SI.pdf (1.44 MiB)

Flow Battery Electroanalysis 2: Influence of

Surface Pretreatment on Fe(III/II) Redox

Chemistry at Carbon Electrodes

Tejal V. Sawant and James R. McKone∗

Department of Chemical and Petroleum Engineering, Swanson School of Engineering, University

of Pittsburgh, Pittsburgh, PA 15261, USA

E-mail: jmckone@pitt.edu

Abstract

Redox flow batteries are attractive for large-scale electrochemical energy storage, but slug-

gish electron transfer kinetics often limit their overall energy conversion efficiencies. In an

effort to improve our understanding of these kinetic limitations in transition metal based flow

batteries, we used rotating-disk electrode voltammetry to characterize the electron-transfer

rates of the Fe3+/2+ redox couple at glassy carbon electrodes whose surfaces were modified

using several pre-treatment protocols. We found that surface activation by electrochemical cy-

cling in H2SO4(aq) electrolyte resulted in the fastest electron-transfer kinetics: j0 = 0.9±0.07

mA/cm2 in an electrolyte containing 10 mM total Fe. By contrast, electrodes that were chem-

ically treated to either remove or promote surface oxidation yielded rates that were at least an

order of magnitude slower: j0 = 0.07± 0.01 and 0.08± 0.04 mA/cm2, respectively. By cor-

relating these findings with X-ray photoelectron spectroscopy data, we conclude that Fe3+/2+

redox chemistry is catalyzed by carbonyl groups whose surface concentrations are increased

by electrochemical activation.

1

Keywords

Redox flow battery, electroanalysis, kinetics, glassy carbon, rotating disk electrode voltammetry,

surface treatment, cleaning

Introduction

Renewables like solar and wind power are highly desirable energy sources for their environmental

sustainability but problematic because they are spatially and temporally intermittent. As a result,

there is a clear need for cost effective technologies that can efficiently store large quantities of

renewable electricity.1–4 Although conventional (solid phase) secondary batteries can be deployed

for this type of energy storage, they are considerably more developed for mobile applications in

which high energy and power density are paramount.5 By contrast, redox flow batteries (RFBs)

are less technologically mature, but they have garnered significant attention for their potential to

store electricity on a large scale.

RFBs have been under investigation for decades, but academic and commercial interest

has grown substantially over the last several years.6–15 Early work focused on aqueous transition-

metal compounds—particular Fe and Cr aquo complexes—as the active components of liquid-

phase battery electrolytes.16–18 The aqueous all-vanadium flow battery has also been studied ex-

tensively and remains popular due to its use of a single set of interconvertible vanadium aquo

complexes as both the negative and positive electrolytes.19–28 More recently, there has emerged

significant interest in non-aqueous RFB chemistries, many of which are also based on transition

metal complexes as the redox-active components of the electrolytes.29–34

Because RFBs are anticipated to be useful in large-scale installations, they require stack

components—electrodes, separators, bipolar plates, cell housing, and plumbing—that are chem-

ically robust and low in cost. To this end, carbon is the material of choice for the positive and

negative electrodes in practical RFBs.35–38 Generally graphitic carbons are processed into cloths,

2

foams, or felts to produce a porous electrode with high surface area over which redox reactions

can occur.39–42 These materials are broadly similar in chemical composition to glassy carbon (GC)

electrodes, which are commonly employed for electroanalytical studies.43–47

In the context of electroanalysis using carbon electrodes, a variety of preparation proto-

cols are commonly used to clean and/or activate their surfaces toward electrochemistry.48 These

procedures often include polishing steps using aqueous alumina slurries, followed by additional

treatments intended to remove impurities or modify the chemical functionality of the carbon sur-

face.49,50 For example, freshly polished carbon electrodes are often activated by electrochemi-

cal cycling in aqueous acidic electrolytes over a potential range that slightly exceeds the elec-

trode/solvent stability window.44,51 These treatments are understood to help remove impurities

and generate a partially oxidized carbon surface. Chemical oxidants can also be used to gener-

ate surface oxygen species.25,48,52,53 By contrast, thermal treatments in vacuum or under reducing

atmospheres have been used to remove surface oxygen functional groups.24,54–57 McCreery, et al.

reported a straightforward treatment involving incubation of carbon electrodes in organic solvents

that had been pre-purified with activated carbon.45,58,59 This method was also found to decrease the

amount of oxidized carbon on the electrode surface (albeit less effectively than heat treatments),

which was attributed to the removal of residual impurities that remain even after polishing.

A variety of carbon surface preparation protocols have been found to improve the appar-

ent rates of electron transfer to aqueous transition metal RFB redox couples, but the mechanistic

basis of these improvements remains unclear. Several groups have argued that catalytic enhance-

ment results from the presence of surface oxygen functionalities broadly.44,60,61 By contrast, some

have suggested that the primary reason for enhanced kinetics is the presence of a greater ratio of re-

active edge sites to unreactive basal planes.62,63 Recent work on vanadium electrolytes in particular

has suggested that both oxidative and reductive surface treatments can have a significant effect on

electron-transfer rates.64–66 Thus, ambiguity in the kinetics of carbon electrodes in RFB applica-

tions closely mirrors controversies in the electroanalytical literature regarding the catalytic proper-

ties of nominally well-defined model carbon surfaces.67–70 Resolving these ambiguities associated

3

with carbon surface chemistry and electron-transfer rates will greatly improve our understanding

of carbon-based electrocatalysis and help drive improved performance in functional RFBs.

We are working to understand, and ultimately eliminate, kinetic limitations in RFBs by

leveraging insights and methods from electroanalytical chemistry to investigate electron-transfer in

model flow battery electrolytes. The present study extends on prior results in which we character-

ized the kinetics of aqueous Fe3+/2+ as a prototypical RFB positive electrolyte at Pt and Au surfaces

using rotating disk electrode (RDE) voltammetry.71 We have now executed a series of experiments

examining the electron transfer kinetics of the Fe3+/2+ redox couple at GC electrodes as a function of

surface preparation. Using an electrolyte containing 10 mM total Fe in HCl(aq) solution, we found

exchange current densities of 0.07 ±0.01 mA/cm2, 0.9 ±0.07 mA/cm2, and 0.08 ±0.04 mA/cm2

at GC electrodes that were pretreated by incubation in pre-purified isopropanol, electrochemical

activation, and chemical oxidation with H2O2, respectively. These results were surprising because

the electrochemical and H2O2 treatments both increased the extent of carbon surface oxidation, but

only electrochemical activation improved the electron-transfer kinetics. Moreover, X-ray photo-

electron spectroscopy (XPS) and Raman analysis showed only modest differences in the apparent

extent of surface oxidation between all three treatments. These results lead us to conclude that

the catalytic activity of carbon toward aqueous Fe3+/2+ redox chemistry depends on specific sur-

face chemical functionality rather than defects or oxidized sites generally. Based on additional

scrutiny of XPS results, we conclude that the relatively higher proportion of carbonyl groups on

electrochemically treated GC surfaces are responsible for catalyzing Fe3+/2+ electrochemistry.

Experimental

FeCl2 (tetrahydrate salt, 98%), FeCl3 (hexahydrate salt, 97%), hydrochloric acid (Certified ACS

plus), sulfuric acid (trace metal grade), isopropanol (Certified ACS grade), hydrogen peroxide (30

wt%, stabilized with sodium stannate) and Ag/AgCl (gel-type reference electrodes with 3 N NaCl

fill solution) were obtained from Fisher scientific and were used as obtained. Deionized water with

4

resistivity of ≥18.2 MΩ·cm was obtained using a Millipore Milli-Q Advantage A10 system and

used for preparation of all solutions. Activated carbon (425–850 µm particle size) was obtained

from Alfa Aesar. Two water-based detergents, Citranox and Alconox, were obtained from W.W.

Grainger Inc. 5, 1 and 0.05 µm alumina powders and 8 inch polishing micropads were obtained

from Pace Technologies. Graphite electrodes were obtained from Electron Microscopy Sciences;

they were of spectroscopic grade with a porosity of 16.5 % and a diameter of 1/4 in. Zero grade

N2(g) (99.998%) and Ultra-high purity H2(g) (99.999%) were obtained from Matheson.

The electrochemical apparatus was a Pine MSR electrode rotator, equipped with a 5mm

Change-Disk electrode assembly. A Gamry Interface 1000E potentiostat was used for all elec-

trochemical measurements. Sonication was performed in a Branson M1800 ultrasonic cleaner, in

which electrodes were first placed in 20 mL vials filled with deionized water that were in turn

placed in the sonicator bath. The electrochemical cell comprised of a 100 mL glass chamber with

a tight Teflon cap bearing holes to introduce the electrodes and a gas purge tube.

Raman microscopy was perfomed on a Renishaw InVia Raman microscope. Spectra

were collected using 633 nm laser light with 1800 l/mm grating through a 20× objective lens at

50% laser power. Three accumulations of the Raman spectra were taken over the range 1000–1800

cm−1 for 20 seconds each, and each electrode was tested at 2 different locations on the sample.

X-ray photoelectron spectroscopy (XPS) was performed on a Thermo Scientific ES-

CALAB 250Xi. Each electrode was tested at 3 different points on the sample. Survey spectra

were first collected to identify atomic concentrations of surface carbon and oxygen. The C 1s

and O 1s peaks were found at 284 eV and 532 eV and the relative sensitivity factors used were

1 and 2.93, respectively. High resolution scans of the C 1s region were then collected and peak

deconvolution was performed to extract ratios of functional groups present in the sample.72–74 For

this deconvolution, XPS data were fit to graphitic groups (285 eV), alcoholic groups (286 eV),

carbonyl groups (287 eV) and carboxylic groups (288.6 eV) after background subtraction using a

built-in protocol in the instrument software. The peak positions were allowed to vary over ±0.1

5

eV and the full width half max (FWHM) of the peaks was constrained between 1.3-1.9 eV.

For electrochemistry measurements, we followed the protocol that we recently reported

for Pt and Au RDE voltammetry of Fe-based RFB electrolytes.71 Briefly, all glassware was cleaned

by boiling in aqueous detergents (Citranox followed by Alconox), and the cell was assembled with

an electrolyte consisting of 5mM FeCl2 and 5mM FeCl3 in 0.5 M HCl. The working electrode GC

was prepared first by polishing with 5, 1, and 0.05 µm alumina slurries. Then the GC surfaces

were further treated using one of three different procedures. The first carbon treatment comprised

of solvent cleaning using isopropanol that had been pre-purified with activated carbon (hereafter

referred to as AC/IPA).45 A 1:3 vol/vol of activated carbon and isopropanol was first prepared by

stirring for 5 minutes followed by sonicating for 5 minutes. This slurry was then allowed to stand

for at least 30 minutes prior to use. The active surface of a freshly polished glassy carbon electrode

was submersed in this mixture for 10 minutes followed by a copious water rinse and sonicating in

water for 30 seconds. The electrode was then introduced into the cell before it dried. The second

carbon treatment involved first completing the AC/IPA cleaning step, followed by electrochemical

activation. The electrode was cycled in 0.5 M H2SO4(aq) solution from -0.25 to 1.5 V versus

Ag/AgCl for a total of 100 cycles at 200 mV/s and then from -0.25 to 1.7 V versus Ag/AgCl for 20

cycles at 200 mV/s. It was then rinsed and introduced into the Fe3+/2+ electrolyte before it dried.

The third carbon treatment involved again cleaning with AC/IPA, followed by chemical oxidation

by hydrogen peroxide. After solvent cleaning, the electrode surface was introduced into a vial

containing 5 ml of 30% hydrogen peroxide and left for 10 minutes. It was then thoroughly rinsed

with water and introduced into the test cell before it dried.

RDE measurements and analysis also followed the methods we have reported previ-

ously.71 While we found that it was essential to clean the surface of Pt and Au electrodes between

data collection at each rotation rate, this was not found to be necessary for GC electrodes. From the

RDE data, diffusivity was extracted using Koutecky-Levich (KL) analysis and exchange current

density was found by fitting the transport-free polarization data to a version of the Butler-Volmer

equation in which the symmetry factors, αox and αred, were allowed to vary arbitrarily between 0

6

and 1.

Results and Discussion

As an initial indication of the relative rates of electron transfer, Figure 1(a) presents cyclic voltam-

metry (CV) data for glassy carbon electrodes after each surface treatment at a rotation rate of 0

rpm and a scan rate of 200 mV/s. The peak-to-peak separation values, which vary inversely with

reaction kinetics, were considerably smaller for electrochemically-activated GC (140mV) as com-

pared to electrodes with AC/IPA (420mV) or H2O2 (510mV) treatments. Thus, it is immediately

apparent that the kinetics of Fe3+/2+ were considerably faster at electrochemically activated GC

as compared to the other treatments. This result is in good agreement with the popular use of

electrochemical cycling methods to activate carbon electrodes for electroanalysis and RFB elec-

trocatalysis.43,64 Figure 1(b) presents peak-to-peak separations for all GC electrodes over a range

of scan rates from 10–3000 mV/s. We made further use of these data to estimate reaction kinetics

using the method reported by Nicholson.75 This analysis showed that the electron-transfer rate at

electrochemically treated GC was ∼15 times faster than AC/IPA treated GC, and ∼30 times faster

than H2O2 treated GC. Complete details are included in the Supporting Information.

Figure 2 shows RDE current density vs. potential (j–E) data for each electrode type

over a range of rotation rates from 100 to 2500 rpm. The equilibrium potential was found to be

0.45 V versus Ag/AgCl (0.68 V vs. NHE), which agrees reasonably well with the reported value

of 0.7 V vs. NHE for 1 M HCl supporting electrolyte.76 Oxidative and reductive current densi-

ties increased monotonically in magnitude with increasing rotation rate at the outer limits of the

measured potential range. Electrochemically activated GC electrodes exhibited well-differentiated

current densities at low overpotentials, whereas AC/IPA and H2O2 treated electrodes both exhib-

ited overlapping j–E response over a range of at least 100 mV about the equilibrium potential,

which is indicative of predominantly kinetic limitations. Moreover, larger potential ranges were

required to achieve transport-limited behavior for the AC/IPA and H2O2 treated electrodes; in fact,

7

Figure 1: (a) Current density versus potential data representing peak to peak separations at a scanrate of 200mV/s at 0 rpm for AC/IPA treated GC, electrochemically treated GC and H2O2 treatedGC, (b) Peak to peak separation as a function of scan rate over the range from 10-3000 mV/s in5mM FeCl2 and 5 mM FeCl3 in 0.5 M HCl at 0 rpm.

Figure 2: RDE current density versus potential data for (a)AC/IPA treated GC, (b)Electrochemically treated GC and (c) H2O2 treated GC in 5mM FeCl2 and 5 mM FeCl3 in 0.5 MHCl.

8

steady-state limiting currents were not observed even over a 1.2 V scan range for H2O2 treated GC.

Figure 3 depicts KL data for each GC treatment, comprising plots of inverse current

vs. inverse square root of rotation rate for overpotentials from 40 to 180 mV in the positive and

negative directions. These data exhibit linear trends in which the intercepts decrease monotonically

and the slopes of the best fit lines approach a nearly constant value as overpotential increases. Thus,

the slopes of the lines above 100 mV overpotential were used to extract diffusivity values.

The KL data for Fe2+ oxidation at AC/IPA treated GC exhibited greater variability, partic-

ularly at relatively low overpotentials. Upon further examination, we found that the variation was

consistent with a monotonic increase in the electron-transfer rate over the course of the measure-

ment. We attribute this variability to changes in electrode surface composition during electrochem-

ical measurements, which are qualitatively similar to the changes that occur during electrochemical

activation.

Figure 4 presents the natural log of kinetic current density versus overpotential (η) along

with the associated Butler-Volmer fits. The results from these fits were found to be in good agree-

ment with the experimental data if the α values were constrained between 0 and 1 but were not

required to sum to 1. Table 1 summarizes the transport and kinetics properties of the Fe3+/2+ redox

couple (10 mM total Fe concentration) for each carbon surface preparation. Uncertainty values are

reported as 1 standard deviation from mean of 5 replicates. Diffusivity values ranged from 2–4

x 10−6 cm2/s and were indistinguishable within two standard deviations. The exchange current

densities were found to be 0.07 ±0.01 mA/cm2 for AC/IPA treated GC, 0.9 ±0.07 mA/cm2 for

Electrochemically treated GC, and 0.08 ±0.04 mA/cm2 for H2O2 treated GC. All α values were

found to be between 0.25 and 0.45 and were also indistinguishable within two standard deviations.

Our initial expectation was that any surface treatment that oxidized GC surfaces would

result in increased electron-transfer rates.60,61,77,78 We did observe substantial increases in interfa-

cial capacitance, along with the emergence of pseudocapactive redox features, in GC electrodes

9

Figure 3: Koutecky-Levich analysis representing inverse current versus inverse square root ofrotation rate for (a) iron oxidation at AC/IPA treated GC, (b) iron reduction at AC/IPA treated GC,(c) iron oxidation at Electrochemically treated GC, (d) iron reduction at Electrochemically treatedGC, (e) iron oxidation at H2O2 treated GC and (f) iron reduction at H2O2 treated GC in 5mM FeCl2and 5 mM FeCl3

10

Figure 4: Butler-Volmer fit for iron oxidation and reduction at (a) AC/IPA treated GC, (b) Electro-chemically treated GC and (c) H2O2 treated GC in 5mM FeCl2 and 5 mM FeCl3 in 0.5 M HCl

Table 1: Transport and kinetics properties of iron oxidation and reduction at GC electrodes. Stan-dard deviations are reported based on n=5 set of experiments.

Electrode DiffusivityDox (cm2/s)

DiffusivityDred (cm2/s)

Exchange current densityj0 (mA/cm2) αox αred

AC/IPAtreated GC (n=5)

2.2 x 10−6

(±0.4 x 10−6)3.7 x 10−6

(±0.4 x 10−6)0.07

(±0.01)0.33

(±0.09)0.31

(±0.05)Electrochemicallytreated GC (n=5)

3.8 x 10−6

(±0.5 x 10−6)2 x 10−6

(±0.5 x 10−6)0.90

(±0.07)0.35

(±0.07)0.44

(±0.09)H2O2

treated GC (n=5)2.0 x 10−6

(±0.6 x 10−6)2.5 x 10−6

(±1 x 10−6)0.08

(±0.04)0.26

(±0.06)0.36

(±0.01)

11

after the electrochemical and H2O2 treatments (see Supporting Information). These observations

are consistent with substantial surface oxidation and/or increased electrode surface area. Thus,

we anticipated enhanced electron-transfer kinetics in both cases relative to the AC/IPA treatment.

Instead, the electrochemical treatment resulted in faster rates while the H2O2 treatment gave elec-

tron transfer kinetics that were essentially indistinguishable from AC/IPA cleaning alone. These

unexpected results motivated us to further explore electrode surface chemistry using Raman spec-

troscopy and XPS.

Figure 5 presents Raman microscopy data collected immediately after completion of

each surface preparation. Two clear peaks are resolved, which are attributable to the presence of

graphitic carbon at ∼1590 cm−1 (G band) and structural defects at ∼1330 cm−1 (D band).79–82

Some reports also describe an additional peak that appears as a shoulder on the G band at ∼1610

cm−1, which is also indicative of defective carbon.53 Rather than a shoulder, we observed a broad

G peak with substantial intensity at 1610 cm−1. The intensity ratio of D to G bands (ID/IG) for

the AC/IPA and H2O2 treated GC was ≈ 2.1 while that for the electrochemically treated GC was

1.7. This intensity ratio of D to G band has been reported to increase when the graphite crystallite

size decreases, and a larger ID/IG ratio has also been found to correlate with increase in edge

plane density and enhanced electron transfer kinetics.51,62,63 These data are all indicative of highly

defective graphitic carbon, which is consistent with the fact that glassy carbon is comprised of

essentially amorphous, randomly oriented graphite nanocrystallites. Thus, the high density of

crystallographic defects and edge planes in the structure of glassy carbon provides ample sites for

surface oxidation. Nevertheless, these Raman data do not provide a clear indication of the relative

extent of oxidation resulting from each electrode treatment.

To better quantify and compare the extent of surface oxidation, we performed XPS mea-

surements on carbon electrodes after each treatment. Figure 6 presents a region of the XPS survey

scan data for carbon electrodes based on their surface treatments. Surface oxidation was estimated

based on O/C ratios (after accounting for relative intensity factors), which were 0.13 for AC/IPA

cleaned GC, 0.15 for electrochemically treated GC, and 0.17 for hydrogen peroxide treated GC.

12

Figure 5: Raman data depicting D band at 1330 cm−1 and G band at 1590 cm−1 for AC/IPA treated,electrochemically treated and H2O2 treated GC electrodes. The intensities are normalized to themaxmimum intensity of G band for each electrode type.

These data show, then, that indeed the electrochemical and H2O2 treatments resulted in increased

surface oxidation, although the difference between all three treatments was small. More impor-

tantly, it is clear that the degree of oxidation of carbon alone does not correlate well with the rel-

ative kinetics of electron transfer for Fe3+/2+. Instead, we conclude that specific surface functional

groups are responsible for catalysis in this system.

Figure 6: XPS data presenting intensity as a function of binding energy(eV) for AC/IPA treated,Electrochemically treated and H2O2 treated GC electrodes representing carbon and oxygen peaks.

13

We collected high-resolution XPS data to elucidate the relative amounts of various types

of oxidized surface species, under the hypothesis that one or more functionality is responsible for

catalyzing Fe3+/2+ redox chemistry. Figure 7(a) presents a representative XPS peak deconvolution

performed on the C 1s spectrum for H2O2 treated GC. These data were fit to find the relative ratios

of graphitic groups (285 eV), alcoholic groups (286 eV), carbonyl groups (287 eV) and carboxylic

groups (288.6 eV). Figure 7(b) collects the corresponding results from all three surface treatments

as a bar chart depicting fractional ratios of each type of functional group. Graphitic carbon com-

prised the primary component in each case, which is unsurprising in view of the fact that XPS

measurements probe several nm of sample depth, whereas only surface sites are susceptible to ox-

idation. If we treat the distribution of functional groups in the AC/IPA treated carbon as a baseline

corresponding to a “clean” and relatively inactive carbon surface, the electrochemical treatment

was found to increase the relative proportion of carbonyl and carboxylic groups, while decreasing

the proportion of alcohol groups (likely via further oxidation of alcohols to carbonyls or carboxly-

ates). By contrast, the H2O2 treated electrodes slightly increased the fraction of carboxylic groups,

greatly increased the fraction of alcoholic groups, and did not significantly alter the fraction of

carbonyl groups.

The XPS data clearly indicate that carbonyls are the only oxidized surface functionality

whose coverage increases in electrochemically activated GC, but not in H2O2 treated GC. Thus,

we conclude that carbonyl groups are primarily responsible for catalyzing electron transfer from

carbon surfaces to the Fe3+/2+ redox couple. This agrees with several prior studies in which the

catalytic activity of carbon electrodes toward transition metal aquo complexes was also found to

correlate with carbonyl coverage.45,46,57 One possible mechanistic rationale for this enhancement

involves the formation of bridging complexes mediated by carbonyls. This picture is attractive for

its similarity to the comparatively well-established catalysis of transition metal redox chemistry by

forming bridging complexes with adsorbed anions at noble metal electrodes.83–86

Based on further analogies to solution-phase coordination chemistry, a single carbonyl

group is less likely to bind Fe centers than multiple adjacent carbonyl species resembling, e.g.,

14

Figure 7: (a) Representative XPS deconvolution data for H2O2 treated GC fitted after backgroundsubtraction to graphitic groups (285 eV), alcoholic groups (286 eV), carbonyl groups (287 eV) andcarboxylic groups (288.6 eV). The residual curve at the top indicates error on peak fitting to thehigh resolution carbon peak, (b) Bar chart of various functional groups present in AC/IPA treated,electrochemically treated and H2O2 treated GC electrodes obtained from carbon peak deconvolu-tion using XPS.

para-quinones (or hydroquinones) or 1,3-diones. This may help explain why electron-transfer rates

are faster in electrochemically activated GC relative to AC/IPA or H2O2 treated GC, even though

each electrode surface contains carbonyl groups. As total carbonyl surface coverage increases, the

statistical likelihood of near-adjacent carbonyl groups also increases. We speculate that the cat-

alytic mechanism may therefore involve self-catalysis by Fe species that are weakly coordinated

to the carbon surface by two or more carbonyl groups. Surface coordination could then facili-

tate rapid electron transfer from the electrode to surface-bound Fe, followed by a rate-determining

self-exchange type electron-transfer between surface-bound and solution phase Fe species. This

mechanism is analogous to prior observations of self-catalysis by quinones bound to carbon sur-

faces through pi stacking.59 It also bears resemblance to more recent observations of redox-silent

but catalytically competent coordination complexes that were covalently bound to graphite surfaces

through well-defined molecular linkers.87,88 Further improvements in our ability to control the spe-

cific surface chemistry of model carbon electrodes would be advantageous for probing this and

other mechanistic hypothesis, and would in turn support the design of functionalization strategies

15

that could be used to minimize overpotential losses in functional RFBs.

Conclusions

We have used quiescent and hydrodynamic voltammetry to assess the rate of electron-transfer

between glassy carbon and RFB-mimicking Fe3+/2+ electrolytes after three different surface pre-

treatments. We found that electrochemical activation and chemical treatments with concentrated

H2O2 both increased the degree of carbon surface oxidation, but only the former resulted in en-

hanced kinetics. XPS data showed that electrochemical activation increased the proportion of

carbonyl groups on the surface, whereas H2O2 did not, leading us to conclude that carbonyls are

primarily responsible for catalyzing Fe3+/2+ redox chemistry on carbon. Nonetheless, the specific

molecular structure and areal density of active catalytic sites remain unclear.

These results underscore the remarkable complexity underlying the kinetics of flow bat-

teries, even for well-studied coordination complexes involving only a single electron transfer. Fur-

ther work is warranted to elucidate the precise surface chemistry responsible for catalysis in Fe-

based electrolytes and to determine whether the associated mechanism is general to all transition

metal aquo compounds of interest for RFB applications. These insights will help motivate im-

proved surface functionalization strategies that can be applied to technologically relevant carbon

electrodes in the interest of minimizing energy efficiency losses attributable to kinetics in practical

flow batteries.

Acknowledgements

We gratefully acknowledge the Swanson School of Engineering at the University of Pittsburgh for

support of this work via startup funds for the McKone Laboratory.

16

References

(1) Noack, J.; Roznyatovskaya, N.; Herr, T.; Fischer, P. The Chemistry of Redox Flow Batteries.

Angew. Chem., Int. Ed. 2015, 54, 9776–9809.

(2) Tokuda, N.; Kumamoto, T.; Shigematsu, T.; Deguchi, H.; Ito, T.; Yoshikawa, N.; Hara, T.

Development of a Redox Flow Battery System. SEI Tech Rev 1998, 8894.

(3) Leung, P.; Li, X.; Ponce de Leon, C.; Berlouis, L.; Low, C. T. J.; Walsh, F. C. Progress in

Redox Flow Batteries, Remaining Challenges and their Applications in Energy Storage. RSC

Adv. 2012, 2, 10125–10156.

(4) Nguyen, T.; Savinell, R. F. Flow Batteries. The Electrochemical Society Interface 2010, 19,

54–56.

(5) Tarascon, J.-M.; Armand, M. Issues and Challenges facing Rechargeable Lithium Batteries.

Nature 2001, 414, 359.

(6) Weber, A. Z.; Mench, M. M.; Meyers, J. P.; Ross, P. N.; Gostick, J. T.; Liu, Q. Redox Flow

Batteries: A Review. Journal of Applied Electrochemistry 2011, 41, 1137–1164.

(7) Bartolozzi, M. Development of Redox Flow Batteries. A Historical Bibliography. J. Power

Sources 1989, 27, 219–234.

(8) Pan, F.; Wang, Q. Redox Species of Redox Flow Batteries: A Review. Molecules 2015, 20,

20499 – 20517.

(9) Wei, W.; Qingtao, L.; Bin, L.; Xiaoliang, W.; Liyu, L.; Zhenguo, Y. Recent Progress in Redox

Flow Battery Research and Development. Adv. Functi. Mater. 2012, 23, 970–986.

(10) Yang, Z.; Zhang, J.; Kintner-Meyer, M. C.; Lu, X.; Choi, D.; Lemmon, J. P.; Liu, J. Electro-

chemical Energy Storage for Green Grid. Chem. Rev. 2011, 111, 3577–3613.

17

(11) Skyllas-Kazacos, M.; Chakrabarti, M. H.; Hajimolana, S. a.; Mjalli, F. S.; Saleem, M.

Progress in Flow Battery Research and Development. J. Electrochem. Soc. 2011, 158, R55.

(12) de Leon, C. P.; Frıas-Ferrer, A.; Gonzalez-Garcıa, J.; Szanto, D. A.; Walsh, F. C. Redox Flow

Cells for Energy Conversion. J. Power Sources 2006, 160, 716 – 732.

(13) Sauer, D. U.; Wenzl, H. Comparison of Different Approaches for Lifetime Prediction of

Electrochemical Systems –Using Lead-Acid Batteries as Example. J. Power Sources 2008,

176, 534–546.

(14) Soloveichik, G. L. Flow Batteries: Current Status and Trends. Chemical Reviews 2015, 115,

11533–11558.

(15) Darling, R. M.; Gallagher, K. G.; Kowalski, J. A.; Ha, S.; Brushett, F. R. Pathways to Low-

Cost Electrochemical Energy Storage: A Comparison of Aqueous and Nonaqueous Flow

Batteries. Energy Environ. Sci. 2014, 7, 3459–3477.

(16) Hagedorn, N.; Thaller, L. Redox Storage Systems for Solar Applications. NASA TM 81464,

National Aeronautics and Space Administration, U S Dept. of Energy 1980,

(17) Gahn, R. F.; Hagedorn, N. H. Single Cell Performance Studies on the Fe / Cr Redox Energy

Storage System Using Mixed Reactant Solutions at Elevated Temperature Conservation and

Renewable Energy Division of Energy Storage Systems. NASA TM-83385, National Aero-

nautics and Space Administration, U S Dept. of Energy 1983,

(18) Thaller, L. H. Electrically Rechargeable Redox Flow Cell. United States Patent 3,996,064

1976,

(19) Kaneko, H.; Nozaki, K.; Wada, Y.; Aoki, T.; Negishi, A.; Kamimoto, M. Vanadium Redox

Reactions and Carbon Electrodes for Vanadium Redox Flow Battery. Electrochim. Acta 1991,

36, 1191–1196.

18

(20) Maria, S.; George, K.; Grace, P.; Hugh, V. Recent Advances with UNSW Vanadium Based

Redox Flow Batteries. Int. J. Energy Res. 2010, 34, 182–189.

(21) Skyllas-Kazacos, M. Novel Vanadium Chloride/ Polyhalide Redox Flow Battery. J. Power

Sources 2003, 124, 299–302.

(22) Rychcik, M.; Skyllas-Kazacos, M. Characteristics of a New All-Vanadium Redox Flow Bat-

tery. J. Power Sources 1988, 22, 59–67.

(23) Rychcik, M.; Skyllas-Kazacos, M. Evaluation of Electrode Materials for Vanadium Redox

Cell. J. Power Sources 1987, 19, 45–54.

(24) Sun, B.; Skyllas-Kazacos, M. Modification of Graphite Electrode Materials for Vanadium

Redox Flow Battery Application I. Thermal Treatment. Electrochim. Acta 1992, 37, 1253–

1260.

(25) Sun, B.; Skyllas-Kazacos, M. Chemical Modification of Graphite Electrode Materials for

Vanadium Redox Flow Battery Application-Part II. Acid Treatments. Electrochim. Acta 1992,

37, 2459–2465.

(26) Aaron, D. S.; Liu, Q.; Tang, Z.; Grim, G. M.; Papandrew, A. B.; Turhan, A.; Zawodzin-

ski, T. A.; Mench, M. M. Dramatic Performance Gains in Vanadium Redox Flow Batteries

through Modified Cell Architecture. J. Power Sources 2012, 206, 450–453.

(27) Houser, J.; Clement, J.; Pezeshki, A.; Mench, M. M. Influence of Architecture and Material

Properties on Vanadium Redox Flow Battery Performance. J. Power Sources 2016, 302, 369–

377.

(28) Zhong, S.; Skyllas-Kazacos, M. Electrochemical Behaviour of Vanadium(V)/Vanadium(IV)

Redox Couple at Graphite Electrodes. J. Power Sources 1992, 39, 1–9.

(29) Ding, Y.; Zhao, Y.; Li, Y.; Goodenough, J. B.; Yu, G. A High-Performance All-Metallocene-

Based, Non-Aqueous Redox Flow Battery. Energy Environ. Sci. 2017, 10, 491–497.

19

(30) Gong, K.; Fang, Q.; Gu, S.; Li, S. F. Y.; Yan, Y. Nonaqueous Redox-Flow Batteries: Organic

Solvents, Supporting Electrolytes, and Redox Pairs. Energy Environ. Sci. 2015, 8, 3515–

3530.

(31) Milshtein, J. D.; Barton, J. L.; Carney, T. J.; Kowalski, J. A.; Darling, R. M.; Brushett, F. R.

Towards Low Resistance Nonaqueous Redox Flow Batteries. J. Electrochem. Soc. 2017, 164,

A2487–A2499.

(32) Suttil, J. A.; Kucharyson, J. F.; Escalante-Garcia, I. L.; Cabrera, P. J.; James, B. R.;

Savinell, R. F.; Sanford, M. S.; Thompson, L. T. Metal Acetylacetonate Complexes for High

Energy Density Non-Aqueous Redox Flow Batteries. J. Mater. Chem. A 2015, 3, 7929–7938.

(33) Sevov, C. S.; Brooner, R. E. M.; Chenard, E.; Assary, R. S.; Moore, J. S.; Rodrıguez-

Lopez, J.; Sanford, M. S. Evolutionary Design of Low Molecular Weight Organic Anolyte

Materials for Applications in Nonaqueous Redox Flow Batteries. J. Am. Chem. Soc 2015,

137, 14465–14472.

(34) Xiaoliang, W.; Wu, X.; Jinhua, H.; Lu, Z.; Eric, W.; Chad, L.; Vijayakumar, M.; Hender-

son, W. A.; Tianbiao, L.; Lelia, C.; Bin, L.; Vincent, S.; Wei, W. Radical Compatibility

with Nonaqueous Electrolytes and Its Impact on an All Organic Redox Flow Battery. Angew.

Chem., Int. Ed. 2015, 54, 8684–8687.

(35) McCreery, R. L. Advanced Carbon Electrode Materials for Molecular Electrochemistry.

Chem. Rev. 2008, 108, 2646–2687.

(36) Yue, L.; Li, W.; Sun, F.; Zhao, L.; Xing, L. Highly Hydroxylated Carbon Fibres as Electrode

Materials of All-Vanadium Redox Flow Battery. Carbon 2010, 48, 3079–3090.

(37) Zhao, P.; Zhang, H.; Zhou, H.; Yi, B. Nickel Foam and Carbon Felt Applications for Sodium

Polysulfide/Bromine Redox Flow Battery Electrodes. Electrochim. Acta 2005, 51, 1091–

1098.

20

(38) Chakrabarti, M. H.; Brandon, N. P.; Hajimolana, S. A.; Tariq, F.; Yufit, V.; Hashim, M. A.;

Hussain, M. A.; Low, C. T. J.; Aravind, P. V. Application of Carbon Materials in Redox Flow

Batteries. J. Power Sources 2014, 253, 150–166.

(39) Kazacos, M.; SkyllasKazacos, M. Performance Characteristics of Carbon Plastic Electrodes

in the AllVanadium Redox Cell. J. Electrochem. Soc. 1989, 136, 2759–2760.

(40) Zhong, S.; Kazacos, M.; Burford, R. P.; Skyllas-Kazacos, M. Fabrication and Activation

Studies of Conducting Plastic Composite Electrodes for Redox Cells. J. Power Sources 1991,

36, 29–43.

(41) Qian, P.; Zhang, H.; Chen, J.; Wen, Y.; Luo, Q.; Liu, Z.; You, D.; Yi, B. A Novel Electrode-

Bipolar Plate Assembly for Vanadium Redox Flow Battery Applications. J. Power Sources

2008, 175, 613–620.

(42) Yazici, M. S.; Krassowski, D.; Prakash, J. Flexible Graphite as Battery Anode and Current

Collector. J. Power Sources 2005, 141, 171–176.

(43) Engstrom, R. C.; Strasser, V. A. Characterization of Electrochemically Pretreated Glassy

Carbon Electrodes. Anal. Chem. 1984, 56, 136–141.

(44) McDermott, C. A. A.; Kneten, K. R.; Mccreery, R. L. Electron Transfer Kinetics of Aquated

Fe(+3/+2), Eu(+3/+2), and V(+3/+2) at Carbon Electrodes: Inner Sphere Catalysis by Surface

Oxides. J. Electrochem. Soc. 1993, 140, 2593–2599.

(45) Ranganathan, S.; Kuo, T. C.; McCreery, R. L. Facile Preparation of Active Glassy Carbon

Electrodes with Activated Carbon and Organic Solvents. Anal. Chem. 1999, 71, 3574–3580.

(46) Chen, P.; Fryling, M. A.; McCreery, R. L. Electron Transfer Kinetics at Modified Carbon

Electrode Surfaces: The Role of Specific Surface Sites. Anal. Chem. 1995, 67, 3115–3122.

(47) Stulıkova, M.; Vydra, F. Voltammetry with Disk Electrodes and its Analytical Application:

21

IV. The Voltammetry of Iron(III) at the Glassy Carbon Rotating Disk Electrode in Acid Me-

dia. J. Electroanal. Chem. Interfacial Electrochem. 1972, 38, 349–357.

(48) Kinoshita, K. Carbon : electrochemical and physicochemical properties; Wiley: New York,

1988; p 316.

(49) R.L. McCreery, In Electroanalytical Chemistry, Vol. 17; Bard, A. J., Ed.; 1991; pp 221–374.

(50) R.L. McCreery,; K.K. Cline, Laboratory Techniques in Electroanalytical Chemistry; P.T.

Kissinger and W.R. Heineman: Dekker, NY, 1996; pp 293–332.

(51) Bowling, R.; Packard, R. T.; McCreery, R. L. Mechanism of Electrochemical Activation of

Carbon Electrodes: Role of Graphite Lattice Defects. Langmuir 1989, 5, 683–688.

(52) Eifert, L.; Banerjee, R.; Jusys, Z.; Zeis, R. Characterization of Carbon Felt Electrodes for

Vanadium Redox Flow Batteries: Impact of Treatment Methods. J. Electrochem. Soc. 2018,

165, A2577–A2586.

(53) Yi, Y.; Weinberg, G.; Prenzel, M.; Greiner, M.; Heumann, S.; Becker, S.; Schlogl, R. Elec-

trochemical Corrosion of a Glassy Carbon Electrode. Catal. Today 2017, 295, 32–40.

(54) Rice, R. J.; Pontikos, N. M.; McCreery, R. L. Quantitative Correlations of Heterogeneous

Electron-Transfer Kinetics with Surface Properties of Glassy Carbon Electrodes. J. Am.

Chem. Soc 1990, 112, 4617–4622.

(55) Rice, R.; Allred, C.; McCreery, R. Fast Heterogeneous Electron Transfer Rates for Glassy

Carbon Electrodes without Polishing or Activation Procedures. J. Electroanal. Chem. Inter-

facial Electrochem. 1989, 263, 163–169.

(56) Poon, M.; McCreery, R. L. In Situ Laser Activation of Glassy Carbon Electrodes. Anal. Chem

1986, 58, 2745–2750.

(57) Kuo, T.-C.; McCreery, R. L. Surface Chemistry and Electron-Transfer Kinetics of Hydrogen-

Modified Glassy Carbon Electrodes. Anal. Chem. 1999, 71, 1553–1560.

22

(58) Ranganathan, S.; McCreery, R. L. Electroanalytical Performance of Carbon Films with Near-

Atomic Flatness. Anal. Chem. 2001, 73, 893–900.

(59) DuVall, S. H.; McCreery, R. L. Self-Catalysis by Catechols and Quinones during Heteroge-

neous Electron Transfer at Carbon Electrodes. J. Am. Chem. Soc 2000, 122, 6759–6764.

(60) He, Z.; Jiang, Y.; Meng, W.; Jiang, F.; Zhou, H.; Li, Y.; Zhu, J.; Wang, L.; Dai, L. HF/H2O2

treated Graphite Felt as the Positive Electrode for Vanadium Redox Flow Battery. Appl. Surf.

Sci. 2017, 423, 111–118.

(61) Zhang, W.; Xi, J.; Li, Z.; Zhou, H.; Liu, L.; Wu, Z.; Qiu, X. Electrochemical Activation of

Graphite Felt Electrode for VO2+/VO2+ Redox Couple Application. Electrochim. Acta 2013,

89, 429–435.

(62) Park, M.; Jeon, I.-Y.; Ryu, J.; Baek, J.-B.; Cho, J. Exploration of the Effective Location of

Surface Oxygen Defects in Graphene-Based Electrocatalysts for All-Vanadium Redox-Flow

Batteries. Adv. Energy Mater. 2014, 5, 1401550.

(63) Pour, N.; Kwabi, D. G.; Carney, T.; Darling, R. M.; Perry, M. L.; Shao-Horn, Y. Influence

of Edge- and Basal-Plane Sites on the Vanadium Redox Kinetics for Flow Batteries. J. Phys.

Chem. C 2015, 119, 5311–5318.

(64) Miller, M. A.; Bourke, A.; Quill, N.; Wainright, J. S.; Lynch, R. P.; Buckley, D. N.;

Savinell, R. F. Kinetic Study of Electrochemical Treatment of Carbon Fiber Microelectrodes

Leading to In Situ Enhancement of Vanadium Flow Battery Efficiency. J. Electrochem. Soc.

2016, 163, A2095–A2102.

(65) Bourke, A.; Miller, M. A.; Lynch, R. P.; Gao, X.; Landon, J.; Wainright, J. S.; Savinell, R. F.;

Buckley, D. N. Electrode Kinetics of Vanadium Flow Batteries: Contrasting Responses of

V(II)-V(III) and V(IV)-V(V) to Electrochemical Pretreatment of Carbon. J. Electrochem.

Soc. 2016, 163, A5097–A5105.

23

(66) Bourke, A.; Miller, M. A.; Lynch, R. P.; Wainright, J. S.; Savinell, R. F.; Buckley, D. N. Ef-

fect of Cathodic and Anodic Treatments of Carbon on the Electrode Kinetics of V(IV)/V(V)

Oxidation-Reduction. J. Electrochem. Soc. 2015, 162, A1547–A1555.

(67) Dumitrescu, I.; Unwin, P. R.; Macpherson, J. V. Electrochemistry at Carbon Nanotubes: Per-

spective and Issues. Chem. Commun. 2009, 6886–6901.

(68) Nugent, J. M.; Santhanam, K. S. V.; Rubio, A.; Ajayan, P. M. Fast Electron Transfer Kinetics

on Multiwalled Carbon Nanotube Microbundle Electrodes. Nano Lett. 2001, 1, 87–91.

(69) Chen, P.; McCreery, R. L. Control of Electron Transfer Kinetics at Glassy Carbon Electrodes

by Specific Surface Modification. Anal. Chem 1996, 68, 3958–3965.

(70) Banks, C. E.; Davies, T. J.; Wildgoose, G. G.; Compton, R. G. Electrocatalysis at Graphite

and Carbon Nanotube Modified Electrodes: Edge-Plane Sites and Tube Ends are the Reactive

Sites. Chem. Commun. 2005, 829–841.

(71) Sawant, T. V.; McKone, J. R. Flow Battery Electroanalysis: Hydrodynamic Voltammetry of

Aqueous Fe(III/II) Redox Couples at Polycrystalline Pt and Au. ACS Applied Energy Mate-

rials 2018, A–K.

(72) Cabaniss, G. E.; Diamantis, A. A.; Murphy, W. R.; Linton, R. W.; Meyer, T. J. Electrocatalysis

of Proton-Coupled Electron-Transfer Reactions at Glassy Carbon Electrodes. J. Am. Chem.

Soc 1985, 107, 1845–1853.

(73) Levi, G.; Senneca, O.; Causa, M.; Salatino, P.; Lacovig, P.; Lizzit, S. Probing the Chemical

Nature of Surface Oxides during Coal Char Oxidation by High-Resolution XPS. Carbon

2015, 90, 181–196.

(74) Proctor, A.; Sherwood, P. M. A. X-ray Photoelectron Spectroscopic Studies of Carbon Fibre

Surfaces–II: The Effect of Electrochemical Treatment. Carbon 1983, 21, 53–59.

24

(75) Nicholson, R. S. Theory and Application of Cyclic Voltammetry for Measurement of Elec-

trode Reaction Kinetics. Anal. Chem 1965, 37, 1351–1355.

(76) J Bard, A.; R. Faulkner, L. Electrochemical Methods : Fundamentals and Applications, 2nd

ed.; John Wiley and Sons, Inc., 1980; Vol. 18.

(77) Taylor, R. J.; Humffray, A. A. Electrochemical Studies on Glassy Carbon Electrodes: II.

Oxygen Reduction in Solutions of High pH (pH¿10). J. Electroanal. Chem. Interfacial Elec-

trochem. 1975, 64, 63–84.

(78) Zhang, G.; Sun, S.; Yang, D.; Dodelet, J.-P.; Sacher, E. The Surface Analytical Charac-

terization of Carbon Fibers Functionalized by H2SO4/HNO3 treatment. Carbon 2008, 46,

196–205.

(79) Chu, P. K.; Li, L. Characterization of Amorphous and Nanocrystalline Carbon Films. Mater.

Chem. Phys. 2006, 96, 253–277.

(80) Wang, Y.; Alsmeyer, D. C.; McCreery, R. L. Raman Spectroscopy of Carbon Materials:

Structural Basis of Observed Spectra. Chemistry of Materials 1990, 2, 557–563.

(81) Cancado, L. G.; Jorio, A.; Ferreira, E. H. M.; Stavale, F.; Achete, C. A.; Capaz, R. B.;

Moutinho, M. V. O.; Lombardo, A.; Kulmala, T. S.; Ferrari, A. C. Quantifying Defects in

Graphene via Raman Spectroscopy at Different Excitation Energies. Nano Lett. 2011, 11,

3190–3196.

(82) Jin, J.; Fu, X.; Liu, Q.; Liu, Y.; Wei, Z.; Niu, K.; Zhang, J. Identifying the Active Site

in Nitrogen-Doped Graphene for the VO(2+)/VO2(+) Redox Reaction. ACS Nano 2013, 7,

4764–4773.

(83) Weber, J.; Samec, Z.; Marecek, V. The Effect of Anion Adsorption on the Kinetics of the

Fe3+/Fe2+ Reaction on Pt and Au Electrodes in HClO4. J. Electroanal. Chem 1978, 89,

271–288.

25

(84) Junsuke, S. Hydrodynamic Voltammetry with the Convection Electrode. IV. The Measure-

ments of the Kinetic Parameters of the Electrode Reaction. Bull. Chem. Soc. Jpn. 1970, 43,

755–758.

(85) Hung, N. G.; Nagy, Z. Kinetics of the Ferrous/Ferric Electrode Reaction in the Absence of

Chloride Catalysis. J. Electrochem. Soc. 1987, 134, 2215–2220.

(86) Angell, D. H.; Dickinson, T. The Kinetics of the Ferrous/Ferric and Ferro/Ferricyanide Reac-

tions at Platinum and Gold Electrodes. Part I. Kinetics at Bare-Metal Surfaces. J. Electroanal.

Chem. 1972, 35, 55–72.

(87) Oh, S.; Gallagher, J. R.; Miller, J. T.; Surendranath, Y. Graphite-Conjugated Rhenium Cata-

lysts for Carbon Dioxide Reduction. J. Am. Chem. Soc 2016, 138, 1820–1823.

(88) Jackson, M. N.; Oh, S.; Kaminsky, C. J.; Chu, S. B.; Zhang, G.; Miller, J. T.; Surendranath, Y.

Strong Electronic Coupling of Molecular Sites to Graphitic Electrodes via Pyrazine Conju-

gation. J. Am. Chem. Soc 2018, 140, 1004–1010.

26

download fileview on ChemRxivFlow Battery 2-Main.pdf (1.13 MiB)

Supporting Information

Flow Battery Electroanalysis 2: Influence of

Surface Pretreatment on Fe(III/II) Redox

Chemistry at Carbon Electrodes

Tejal V. Sawant and James R. McKone∗

Department of Chemical and Petroleum Engineering, Swanson School of Engineering, University

of Pittsburgh, Pittsburgh, PA 15261, USA

E-mail: jmckone@pitt.edu

Tabulated Fe3+/2+ kinetics data from literature

Table S1 present reported kinetics data for Fe3+/2+ redox chemistry at carbon electrodes. The rate

constants, all of which were collected using RDE voltammetry, varied over nearly three orders of

magnitude. Reported α values are close to 0.5 in most cases, while some reports did not explicitly

include α values.

Peak current vs. scan rate data

We performed cyclic voltammetry over a series of scan rates ranging from 10 to 200 mV/s (Fig-

ure S1) and collected values of anodic and cathodic peak currents. Figure S2 plots peak current

versus square root of scan rate, which show linear behavior as expected from a diffusional process.

S1

Table S1: Kinetics of Fe3+/2+ as reported in literature

Electrode Rate constantk0 (cm/s) α

Measuringtechnique

Supportingelectrolyte [Fe] Ref.

GC 4.4 x 10−3 0.5 RDE 0.1 M HCl 1

GC 1 x 10−3 0.5 RDE 0.3 M HCl 1

GC 2.9 x 10−3 0.5 RDE 0.1 M HNO31

GC >10−2 0.57 RDE 0.1-1 M HClO41

GC 1.3 x 10−3 0.5 RDE 0.1 M H2SO41

Carbon paste 5.4 x 10−5 0.63 RDE 1 M H2SO42

Carbon paste 8.5 x 10−5 RDE 0.1 M HCl 2

GC 2.3 x 10−3 RDE 0.2 M HClO4 5 mM 3

GC 2.3 x 10−4 RDE 1 M NaCl 0.1 M 4

Pyrolytic graphite 5.2 x 10−4 RDE 4 M HCl 0.1 M 5

Figure S1: Cyclic voltammograms of (a) AC/IPA treated GC, (b) electrochemically treated GC and(c) H2O2 treated GC electrodes depicting scan rate dependence in an electrolyte containing 5mMFeCl2 and 5mM FeCl3 in 0.5 M HCl(aq).

Figure S2: Peak current density vs. square root of scan rate for (a) AC/IPA treated GC, (b) electro-chemically treated GC and (c) H2O2 treated GC electrodes in an electrolyte containing 5mM FeCl2and 5mM FeCl3in 0.5 M HCl(aq).

S2

Kinetics estimates from CV Data

To estimate electron-transfer kinetics directly from CV data, we followed the method reported by

Nicholson, who showed that the relationship between ∆Ep and scan rate can be used to extract

k0.6 We first extracted the following mathematical relationship between ∆Ep and an empirical

parameter Ψ from Nicholson’s report:

∆Ep · n = 0.051 + 0.0345ψ−0.671 (1)

Nicholson showed that Ψ can be related to the apparent electron-transfer rate constant k0,app via

the following equation:

ψ = ν−12k0,app

[πnFD

RT

]− 12

(2)

such that a plot of Ψ vs ν−0.5 should give a line with y intercept that is near zero and a slope that is

proportional to k0. Figure S3 shows the corresponding plots for Fe3+/2+ voltammetric data at carbon

electrodes prepared by the three different methods over an extended range of scan rates from 10–

3000 mV/s. The data for AC/IPA and electrochemically treated carbon both gave excellent linear

Figure S3: Plots of ψ versus inverse square root of scan rate along with least-squares fits over arange of scan rates from 10–3000 mV/s for (a) AC/IPA treated GC, (b) Electrochemically treatedGC and (c) H2O2 treated GC

fits, whereas the H2O2 treated carbon gave a considerably poorer fit, which is consistent with the

greater relative uncertainty we found in the RDE results for these surfaces.

Table S2 collects apparent k0 values (k0,app) extracted from the slopes of these data along

S3

with the corresponding j0 values for a 5 mM reactant concentration under the assumption that

the diffusivity was 4 x 10−6 cm2/s and the relationship between k0,app and j0 obeys the following

equation:

j0 = nFCk0,app (3)

These j0 values agree reasonably well with those reported from RDE results in the main text.

However, because these results relied on a mathematical fit to Nicholson’s empirical relation, we

recommend that they be treated only as an order-of-magnitude estimate of reaction kinetics.

Table S2: j0 and k0,app obtained from analysis of voltammetry data

Electrodek0,app(cm/s)

j0(mA/cm2)

AC/IPA treated GC 2.4 x 10−4 0.12Electrochemically treated GC 3.7 x 10−3 1.8

H2O2 treated GC 1.2 x 10−4 0.06

Results of conventional Butler-Volmer fits with contrained α values

Figure S4 reports conventional Butler-Volmer fits of kinetics data obtained by constraining the fit

to αox + αred = 1. These results show a clear deviation in the fits for all 3 electrode surface

treatments, which were resolved by allowing the symmetry factors to vary arbitrarily between 0

and 1 as shown in the main text. Such a deviation is consistent with a reaction mechanism involving

both electrochemical and chemical elementary steps.

Residual surface roughness

As a part of our experimental procedure, we polished the surface of GC electrodes before em-

ploying various activation protocols. Although polishing is an important step in removing adventi-

tious species from GC surfaces, we were unable to consistently eliminate polishing damage in the

form of sub-micron scratches, as shown in Figure S5. This residual surface structure complicates

S4

Figure S4: Transport free polarization data for (a) AC/IPA treated GC, (b)Electrochemically treatedGC and (c) H2O2 treated GC electrodes in RFB electrolyte containing 5mM FeCl2 and 5mM FeCl3in 0.5 M HCl(aq). The black lines correspond to fits using the Butler-Volmer equation where thesymmetry factors were constrained to αox + αred = 1.

quantitative analysis of interfacial electron transfer rates due to the increase in electrochemically

accessible surface area per unit geometric area. However, even GC electrodes with completely

smooth surfaces are likely to exhibit considerable site heterogeneity arising from the high density

of crystallographic defects, which makes the determination of the “intrinsic” kinetics of these sur-

faces difficult. In fact, we argue that this roughness may in fact be advantageous for the study of

RFB electrokinetics, in that it is likely to expose a diversity of interfacial active sites as would be

expected to exist in structured carbon materials that are used in functional RFBs.

Figure S5: Scanning electron micrographs of (a) AC/IPA treated GC, (b) Electrochemically treatedGC and (c) H2O2 treated GC

S5

Effect of surface treatment on interfacial capacitance

Figure S6 overlays CV data of GC electrodes in 0.5 M H2SO4 solution over the potential range

from -0.2 to 0.8 V vs Ag/AgCl. In the absence of a redox couple, the data in this range are repre-

Figure S6: Current density versus potential data for AC/IPA treated, electrochemically treated, andH2O2 treated GC electrodes in 0.5 M H2SO4(Aq) solution.

sentative of the interfacial capacitance and/or surface redox chemistry of the electrodes. Whereas

prior work involving similar solvent treatments resulted in interfacial capacitance values on the

order of 50–100 µF/cm2,7 we found AC/IPA treated electrodes gave values on the order of 400

µF/cm2. We attribute this difference to residual surface roughness from alumina polishing, as dis-

cussed in the previous section. The apparent interfacial capacitance also increased markedly with

electrochemical and H2O2 treatments (by factors of 3 and 5, respectively). Moreover, reversible

voltammetric features emerged between 0.2 and 0.5 V vs Ag/AgCl, which can be attributed to

pseudocapacitive redox chemistry of oxidized surface species. These data collectively agree with

XPS results showing that electrochemical and H2O2 treatments both oxidize GC surfaces, where

the latter treatment introduces somewhat more surface oxygen functionality in spite of the lack of

increase in Fe3+/2+ electron transfer kinetics.

S6

References

(1) Stulıkova, M.; Vydra, F. Voltammetry with Disk Electrodes and its Analytical Application: IV.

The Voltammetry of Iron(III) at the Glassy Carbon Rotating Disk Electrode in Acid Media. J.

Electroanal. Chem. Interfacial Electrochem. 1972, 38, 349–357.

(2) Galus, Z.; Adams, R. N. The Investigation of the Kinetics of Moderately Rapid Electrode

Reactions using Rotating Disk Electrodes. J. Phys. Chem 1963, 67, 866–871.

(3) McDermott, C. A.; Kneten, K. R.; McCreery, R. L. Electron Transfer Kinetics of Aquated Fe

(+3/+2), Eu(+3/+2), and V(+3/+2) at Carbon Electrodes: Inner Sphere Catalysis by Surface

Oxides. J. Electrochem. Soc. 1993, 140, 2593–2599.

(4) Hawthorne, K. L.; Wainright, J. S.; Savinell, R. F. Studies of Iron-Ligand Complexes for an

All-Iron Flow Battery Application. J. Electrochem. Soc. 2014, 161, A1662–A1671.

(5) Ateya, B. G.; Austin, L. G. The Kinetics of Fe(2+)/ FeCl(2+)/ HCI (aq) on Pyrolytic Graphite

Electrodes. J. Electrochem. Soc. 1973, 120, 1216–1219.

(6) Nicholson, R. S. Theory and Application of Cyclic Voltammetry for Measurement of Electrode

Reaction Kinetics. Anal. Chem 1965, 37, 1351–1355.

(7) Ranganathan, S.; Kuo, T. C.; McCreery, R. L. Facile Preparation of Active Glassy Carbon

Electrodes with Activated Carbon and Organic Solvents. Anal. Chem 1999, 71, 3574–3580.

S7

download fileview on ChemRxivFlow Battery 2- SI.pdf (1.44 MiB)