Electrocatalysts for Hydrogen Oxidation Reaction in Alkaline Electrolytes

26 Pages·2017·8.26 MB·English

Review
CiteThis:ACSCatal.2018,8,6665−6690 pubs.acs.org/acscatalysis
Electrocatalysts for Hydrogen Oxidation Reaction in Alkaline
Electrolytes
Elena S. Davydova,*,† Sanjeev Mukerjee,‡ Fred́ eŕ ic Jaouen,§ and Dario R. Dekel*,†
†
TheWolfsonDepartmentofChemicalEngineeringandtheNancy&StephanGrandTechnionEnergyProgram(GTEP),Technion
− Israel Institute of Technology, Haifa 3200003, Israel
‡
Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts 02115, United States
§Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, UniversitéMontpellier, ENSCM, 34095 Montpellier, France
*
S Supporting Information
s.
e
cl
arti ABSTRACT: In the past 5 years, advances in anion-
d
0:06:50 (UTC). y share publishe Sacdcceoeehlvvnlsieederl(vuoaAeclptEmnirvMeeeecanFertlnCymtsoie)dfwmeaonsbartdrtikachvsnaaelnehnsbcaeeevxdghetiangvanseenhinnoioegworn-apon-tefei-oxnltcinhefheadoatpnfegtaArhefffEeooMrrmmddFeaoaCmbnolsecbrerfcauafunoesrelrrsectneaftultthlelesye-l.
at 2atel of-the-art proton exchange membrane fuel cells. However,
8 m until now, these high AEMFC performances have been
201giti reached with platinum-group metal (PGM)-based anode and
uly 17, ow to le csuatchhocdaetaclaytsatlsyssthso.uInldoinrdtehretonefuarlfiflulttuhreepboetefrneteiaolfoPfGAEMMsFaCnds,,
V on Js on h epvroengrtuesasllyh,asfrbeeeenofachcireitviecdalinratwhemdeavteerloiaplsm.eAntlthoofuPgGhMg-rfreeaet
N UNIoption hcaytdarlyosgtesnfoorxidthaetioonxyrgeeanctiroendu(cHtiOonR)r.eTachteiomnuinchblaoswicermHedOiaR, asicgtniviifitycaonftlPytleinssbaatstiecnmtioedniahcaosmbpeeanredpawidithtothtahteincaatcaildyswisasofitstehlef
TERs for revealed only relatively recently. While several PGM-based composite materials have shown improved HOR activity in basic
HEASdeline mfroemdiaP,GthMeHshOaRvekhinitehtiecrstorermesauinltsedsloinwehrigthhaannondeceeossvaerrypfootrenatniaildse,aslignnoifincpaonlatlryizraebdleuceilnegctrtohdeep.eIrnfoardmdaitniocne,oafttPemGMpts-frtoeemAoEvMeaFwCasy.
a NORTharinggui Tmenheeimrsgbywraobnuaerldrsietbarsebinlaietyem)dashjotaorvebbbeaerfireirnemrolfyvoeerrscttohambeleils.ahrAegdef-utsnocdaolaevmederecnoptamlolyeumnthdeinesrtcsthoaafnlldtehinnisggeto.efTchthhniesolHroegOvyieRwomnpcerecehsteahnnetissomtthhieenrcbutaersrciehcnnmtouleondgdiaiecaraslntdahnuodrfdintlhegseom(feta.hgine.,
ed viorg/s HOR electrocatalysis in basic media and critically discusses the most recent material approaches. Promising future research
wnloadbs.acs. dKiEreYcWtioOnsRiDnSt:hehyddervoegleonpomxeidnattioofntrheeacHtioOnR,anelieocntreoxcchataanlygestms efomrbaralknaelifnueelecleelcl,traolklyatleinseamreedailaso,eloeucttrloinceadta.lyst,platinum-groupmetal,
Dos://pu non-noble
p
htt
e
e
S 1. INTRODUCTION consider three cases toward PGM-free or Pt-free AEMFCs. In
thefirstcase,weconsiderstudieswithPGM-basedanodesand
Anion-exchangemembranefuelcells(AEMFCs) havereceived
PGM-free cathodes. The power density spans between 50 and
increasingattention for more thana decade, mainlybecause of
1500 mW cm−2,12−18 with the BoL performance highly
their potential for resorting to electrocatalysts that are free of
impacted by the oxygen reduction reaction (ORR) activity at
platinum-group metals (PGMs) and, eventually, free of critical
raw materials (CRMs).1,2 Recent achievements in developing thecathode.Inthesecondcase,weconsiderstudieswithPGM-
anion-exchange membranes (AEMs) with high hydroxyl ion free anodes and cathodes. In this case, the BoL power density
conductivity have allowed the fabrication of H -fed AEMFCs spansbetween40and120mWcm−2only,19−21asitisstrongly
2
reaching impressive peak power densities of 1.8 and 0.5 W limited by the very low hydrogen oxidation reaction (HOR)
cm−2 with cathodes fed with oxygen and air, respectively.3 activity of PGM-free catalysts to date. As a third case, we
However, these recent high beginning-of-life (BoL) perform- considerPt-freeAEMFCswithaPd-basedcatalystateitherthe
ances have been obtained with PGM-based catalysts at both anode18,22,23 or the cathode24,25 in combination with a non-
electrodes,usuallywithPt/CatthecathodeandPtoraPt/Ru
alloy at the anode.4−11 Received: February19, 2018
In an effort to logically organize and summarize the BoL Revised: May3, 2018
performance achieved in previously reported studies, we Published: May31, 2018
©2018AmericanChemicalSociety 6665 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
PGM catalyst at the opposite electrode. In this case, power environmentat60°Cwasassumedtobe10−4Acm−2 ,
Pt
densities of up to 500 mW cm−2 have been achieved.22 These i.e., ca. 2 times lower than the specific activity of Pt
data show that while alternatives to PGMs already exist for nanoparticles at 0.9 V when measured under PEMFC
AEMFCcathodes(case1),nocompetitivePGM-freecatalystis conditions at 80 °C (about 2 × 10−4 A cm−2 ).32 This
Pt
available today for AEMFC anodes (cases 2 and 3). assumption is in agreement with the slightly lower ORR
This status can be explained historically, with the quest for activity on Pt single crystals in alkaline versus acidic
PGM-free electrocatalysts for the ORR in acidic media having media.32Theseassumptionsresultinthemassactivityat
greatlyhelpedindevelopingmetal−N−CcatalystsfortheORR 0.9 V of 50 A g −1, in line with experimentally reported
in basic media as well.26−30 The ORR activity of some values in 0.1 MPKt OH.33
pyrolyzed Fe−N−C catalysts is now in fact comparable to, or (b) the Fe−N−C catalyst with the mass activity at 0.9 V of
even higher than, that of Pt-based catalysts in basic media. 10 A g−1 (per total mass of catalyst).33 The Fe−N−C
While the definition of surface-specific activity does not apply cathode loading was assumed to be 2 mg cm−2 (mostly
to Fe−N−C catalysts, it is possible to compare Fe−N−C and carbon, leading to a ca. 50 μm thick cathode layer),
Pt/C on the basis of volume-specific activity.31 In contrast, no which is on the lower end of the range of non-PGM
suchtransferofknowledgehasbeenpossiblefortheHORfrom cathode loadings but leads to a sufficiently thin cathode
acidic to basic media. No PGM-free catalyst has demonstrated layer for proper operation with air. The apparent
reasonable HOR activity in acidic media to date. Attempts to resulting activity of such a Fe−N−C cathode is thus 2
developsuchcatalystshavebeenveryscarcebecauseofthevery × 10−2 A cm−2 at 0.9 V. The Tafel slope was also
geom
fast HOR kinetics on Pt in acidic media, implying that anodes assumed to be 90 mV decade−1 (as for Pt/C, for fair
in proton-exchange membrane fuel cells (PEMFCs) require comparison).
only an ultralow Pt loading to operate efficiently. The main (c) the40%Ag/Ccatalystwiththespecificsurfaceareaof50
advantage of AEMFCs versus PEMFCs, i.e. resorting to PGM- m2 g −1 and i = 10−5 A cm−2 (based on the above
Ag s,0.9V
andCRM-freecatalysts,isthusimpairedatthismomentbythe assumed specific activity for Pt and the one-order-of-
low HOR activity of PGM-free catalysts in basic media. magnitude lower surface-specific activity for Ag vs Pt34)
To illustrate this better, Figure 1 shows calculated polar- and the catalyst loading of 2 mg cm−2 (mass of Ag and
ization curves for the anode and cathode in an AEMFC, C). The Tafel slope was also assumed to be 90 mV
decade−1 (as for Pt/C, for fair comparison).
While not represented in Figure 1, pure Mn oxide and
bimetallic MnCo oxides have also shown high ORR activity in
alkaline media, especially for nanostructured oxide particles
with high surface area. From an analysis of selected literature
and with an oxide content of 30−50 wt % on carbon for an
optimized fuel cell catalyst, the ORR activity at 0.9 V was
estimated to range from ca. 1 to 10 A g −1 for this class of
cathodematerial.35−37Foratotalcathodelcoatadingof2mgcm−2
and a Tafel slope of 90 mV decade−1, the resulting calculated
polarization curves would range from slightly below the Ag/C
oneinFigure1uptotheexactsamecurveascalculatedforthe
Fe−N−C one in Figure 1.
For the anode catalysts, the following were assumed:
(a) the40%Pt/Ccatalystwiththespecificsurfaceareaof50
m 2 g −1 and the exchange current density (i ) of 0.6
Pt Pt 0
mAcm−2,38−41thecatalystloadingof200μgcm−2(mass
ofPtandC),andtheexchangetransfercoefficientof0.5
Figure1.PolarizationcurvesfortheHORandORRcalculatedusing
(for both the anodic and cathodic directions).
the Butler−Volmer equation for the HOR with exchange current
(b) theNi/N-dopedcarbonnanotube(Ni/N-CNT)catalyst
dsuernfsaictye-(sip0e)caifincdtahcetiTviatfyeleaqtua0t.i9onVforvtsheROHRERw(iithmassoarctiivityor, withi0=0.03mAcm−2,42thesurfaceareaof25m2g−1,a
m,0.9V s,0.9V catalyst loading of 1 mg cm−2 , and the exchange
respectively), mimicking the best-in-class PGM-based and PGM-free geom
catalysts. transfer coefficient of 0.5 (for both the anodic and
cathodic directions).
The simulations of anode performance based on electro-
representing best-in-class PGM-based and PGM-free catalysts.
kinetic losses only (Figure 1) illustrate that the slow HOR
The main assumptions made for the calculations include the kinetics is projected to cause significant voltage loss for the
TafellawfortheORRandtheButler−Volmerequationforthe AEMFC at high current density. These calculations are
HOR, a fuel cell temperature of 60 °C, and an AEM consistent with the experimental AEMFC data reported in
characterizedbypH13.Forthecathodecatalysts,thefollowing the literature. For example, Sheng et al.38 reported an anode
were assumed: overpotential (η) of >130 mV at the anode loading of 50 μg
Pt
(a) the 40% Pt/C catalyst with the surface area of 50 m 2 cm−2 and 1.5 A cm−2. In addition, Figure 1 shows that even
Pt
g −1 and i = 10−4 A cm−2 ,31,32 the Tafel slope of with a Pt/C anode and a significant Pt loading of 0.08 mg
Pt s,0.9V Pt Pt
90 mV decade−1,33 and the catalyst loading of 600 μg cm−2, the anode overpotential at 1 A cm−2 is projected to be
cm−2 (mass of Pt and C). The ORR specific activity at non-negligible. This is in contrast to the case of PEMFCs,
0.9 V (i ) of Pt nanoparticles in the AEMFC where the anode behaves almost as a nonpolarizable electrode.
s,0.9V
6666 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
This difference is fundamentally explained by the 2 orders of Pt in acidic electrolytes, the reaction rate in the RDE becomes
magnitude lower HOR activity on Pt (and other PGMs) in limited by hydrogen diffusion in the electrolyte toward the
basic media compared with that in acidic media (Figure 2). electrode, even at very low overpotentials. The measured
polarization curve (see Figure 3A, curve labeled η ) then
diffusion
Figure 2. HER/HOR exchange current densities measured in 0.1 M
NaOH aqueous electrolyte. Data were recorded in a H-saturated
2
electrolyteatambientpressureand40°C.Reprintedwithpermission
fromref 45. Copyright 2014Royal Society of Chemistry.
WhilenoPGM-freecatalysthasshowndecentHORactivityin
acidic media to date, slower kinetics toward the hydrogen
evolution reaction (HER) in basic versus acidic media has also
beenreportedforPGM-freecatalysts,43,44possiblysupportinga
universal effect of pH on the rates of the HER and HOR. In
summary, resorting to ultralow or even zero PGM content at
the anode of an AEMFC with present PGM- or non-PGM Figure 3. (A) HOR/HER polarization curves on polycrystalline
catalysts leads to an unacceptably high anode overpotential. platinum (Pt(pc)) in 0.1 M HClO at 1600 rpm, as measured (solid
4
NovelmaterialsmustbedesignedwhoseintrinsicHORactivity black line, ERDE) and after Ohmic drop correction (dotted red line,
is higher than presently reached with existing materials and EiR‑free). The dashed black curve labeled ηdiffusion represents the
theoretical curve for which there is no kinetic polarization of the
whose surface is not blocked for the HOR catalysis by the
electrode, only H concentration polarization. The inset shows the
formation of oxide at low overpotential. 2
Koutecky−LevichplotobtainedatE=0.1VvsRHEfortheHORat
In order to help foster rational approaches for the differentrotationrates.(B)HOR/HERpolarizationcurvesonPt(pc)
development of highly active PGM-based and PGM-free
in 0.1 M KOH at 1600 rpm.Reprinted with permission from ref 38.
catalysts, this review covers a series of aspects of the HOR Copyright 2010 The Electrochemical Society.
electrocatalysis, including the hypothesized HOR mechanisms
in basic media and the effects of electrocatalyst affinity toward
hydrogenorhydroxylchemisorptionontheHORkinetics.The
only reflects a hydrogen concentration polarization that is not
already developed mono- or bifunctional HOR electrocatalysts
arereviewedinthelightofthisunderstandingandareclassified related to the kinetics (eq 1):
accordingtothemechanismsbywhichtheycatalyzetheHOR. RT ⎛ i ⎞
η = ln⎜1 − ⎟
2. ROTATING DISK ELECTRODE AND ALTERNATIVE diffusion 2F ⎝ ilim⎠ (1)
METHODS TO MEASURE THE HOR ACTIVITY Thus, quantification of the HOR kinetics by the RDE method
The rotating disc electrode (RDE) technique is the most is difficult for very active catalysts.38 This was explained only
common experimental technique applied for the HOR kinetic recently to lead to a broad range of experimentally reported
studies in liquid electrolytes.38,46,47 Applying the well-known values in acidic electrolytes.48,49 This limitation of the RDE
Levich and Koutecky−Levich equations, one can correct the method is currently less important for PGM-based and even
measured I versus E polarization curves for (i) Ohmic drop in more so for non-PGM-based HOR catalysts in alkaline media
theelectrolyte(typically10−20ΩinRDEsetupsdependingon becauseoftheirloweractivity.ThisisshowninFigure3Bfrom
the nature and concentration of the electrolyte, resulting in a the large positive shift of the experimentally measured RDE
maximum Ohmic drop of ca. 10 mV at the diffusion-limited polarizationcurvesrelativetothetheoreticalcurveexpectedfor
current density for an RDE tip geometric area of 0.2 cm2) and H concentration polarization only (eq 1).
2
(ii) the H diffusion limitation and extract the kinetic current In the future, for highly active HOR catalysts in alkaline
2
density values (polarization overpotential). One general electrolytes with i values too high to be measured with the
0
precaution to recall when applying the RDE to the HOR RDE technique, transient techniques such as rapid potentiody-
reaction (a sometimes quasi-reversible reaction, depending on namic scanning50,51 or other steady-state methods that allow
thecatalystandelectrolytepH)istheimpossibilityofassessing highermasstransport(microelectrodes)maybeused.48Forthe
the HOR activity for catalysts whose exchange current density HOR kinetics evaluation on powder catalysts, the cavity
is commensurate with or higher than the diffusion-limited microelectrode technique, which was developed for both
currentdensity,i .Forexample,asaresultofhighi valueson liquid52 and solid electrolyte conditions,53,54 is also applicable.
lim 0
6667 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
Figure4.(A)Structureoftheworkingelectrode,madeupofcatalystdepositedontoagold-coatedporouspolycarbonatetrack-etchmembrane.The
poresarecoatedwiththehydrophobicamorphousfluoropolymertokeeptheporesopenforreactantgas(inthiscaseoxygen)toflowtothecatalyst
from behind. (B) Comparison of the HOR polarization curves measured by the floating electrode and RDE methods using 2.2 μg cm−2 Pt/C
Pt
catalystat 10 mV s−1 at 298 K. Reprintedwith permission fromref 55. Copyright 2013The PCCP Owner Societies.
As an alternative to the RDE method, the floating electrode an ultralow loading, e.g., 3 μg cm−2) while the cathode is
Pt
techniquewasproposedbyZalitisetal.55inordertoinvestigate coatedwithhighloadingofPt/C(e.g.,0.4mg cm−2),defacto
Pt
the HOR electrocatalysts. It combines the high mass transport acting as a reference electrode.57 By the use of this H pump
2
ratecharacteristic ofgas-diffusionelectrodesinfuelcellswitha method with an AEM, the i value of the HOR on Pt/C
0
flat, uniform, and homogeneous catalyst deposition process. interfacedwithananion-exchangeionomer(AEI)wasreported
This approach closely mimicks the operating conditions of for the first time.58 It was found that the HOR activity at the
AEMFCs or PEMFCs (gas-fed electrodes). Uniform Pt layers Pt/C−AEI(TokuyamaA201)interfaceisalmostthesameasin
were produced with loadings down to 0.16 μg cm−2 and a 0.1MKOH,bothactivitiesbeingnearly2ordersofmagnitude
Pt
lowerthanthatmeasuredinacidicelectrolytes,eitherthePEM
thickness of 200 nm on polycarbonate track-etch membrane
filter/Aufloatingelectrodes(Figure4).TheHORrevealedfine electrolyte (Nafion 211) or 0.1 M HClO4.
structure in the diffusion-limited current range and an
3. GENERAL PRINCIPLES OF THE HOR IN BASIC
asymptotic current decay for potentials above 0.36 V due to
MEDIA
theformationofoxidesontheplatinumsurfaceandadsorption
ofanions,whicharenotvisiblewithtechniqueslimitedbymass 3.1.PathwayoftheHORinAlkalineMedia.Itiswidely
transport in aqueous media such as the RDE method. agreed,atleastforinorganicmaterials,thattheHORinalkaline
media proceeds through a combination of the following
Kinetic data on the HOR have also been obtained in single-
elementary steps: (A) dissociative adsorption of molecular
cell fuel cell applying the electrochemical hydrogen pumping
mode(Figure5).56Inthesesystems,hydrogenoxidationoccurs dihydrogen, (B) electron transfer from molecular dihydrogen
to the catalyst, and (C) discharge of the adsorbed hydrogen
attheanode(workingelectrode)andhydrogenevolutionatthe
atom. Hereafter we describe these three steps.
cathode, with protons or hydroxyl ions transferred via the
(A) Dissociative adsorption of molecular dihydrogen, H ,
proton-exchange membrane (PEM) or AEM, respectively. In without electron transfer, which is known as the Tafel (T2)
order to study the HOR kinetics, the anode is coated with the
reaction59 (eq 2):
HOR catalyst under investigation (for PGM-based catalyst, at
H2 + 2* → 2(*−Had) (2)
whereH ismoleculardihydrogeninthevicinityofthecatalyst
2
surface, * is an active site, and H is a chemisorbed hydrogen
ad
atom(oradatom).ThedissociationenergyoftheH molecule
2
is4.52eVataseparationof0.74Å.60Oneoftherequirements
imposed on the catalytic materials is affinity toward molecular
hydrogenchemisorption.Therefore,the*−H bindingenergy
ad
is believed to be a crucial factor controlling the HOR
kinetics.39,61 The surface intermediate of adsorbed hydrogen
onPtmicroelectrodeswasshowntobeindependentofsolution
pH.62Also,aswasrevealedpreviouslyforchalcogenides(RuS ,
2
Co-doped RuS ),63 the ability to dissociate H at room
2 2
temperature is a necessary but not sufficient criterion for the
HOR catalyst.
Figure 5. Representation of a PEM device operating in hydrogen
pumping mode. Humidified hydrogen is fed at the anode, and dry The dissociative adsorption of the diatomic molecule H2, a
nitrogenisfedatthecathode.Thehydrogenmoleculedissociatesinto central step in many industrially important catalytic processes,
protonsandelectronsattheanode,andthelatterrecombineintoH isgenerallyassumedtorequireatleasttwoadjacentandempty
2
at the cathode. atomic adsorption sites (or vacancies).64 The need for
6668 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
ensembles of PGM atoms is experimentally supported by the controllednumberofadjacentactivesites(havinginitiallyhigh
lack of the HOR activity of Pt atoms atomically dispersed as HBE) “diluted” by inactive atoms in order to tune the HBE.
PtS moieties on a sulfur-doped carbon matrix.65 H is known (B) Electron transfer from molecular dihydrogen to the
4 2
tobe achemically reactive gas, which adsorbs dissociatively on catalyst, which is known as the Heyrovsky (H) reaction69 (eq
most transition metal surfaces with heats of chemisorption 3):
between 60 and 120 kJ mol−1. For example, according to the
linear dependence θ − p 0.5,64 hydrogen is considered to H2 + OH− + * → (*−Had) + H2O + e− (3)
H H2 −
adsorb in the atomic state. whereOH isinthevicinityofthecatalystsurface.Thespecies
According to Christmann,66 the preferential adsorption site is argued not to be OH− but most likely OH .18,22,70 If this is
ad
has high coordination. Threefold sites are favored on diagonal, the case, the H reaction could be written as eq 4:
trigonal,andhexagonalsurfaces,whilefourfoldsitesaretherule
for tetragonal surfaces. The latter should therefore adsorb H H2 + (*−OHad) ⇄ (*−Had) + H2O (4)
2
more strongly for a given metal. There is ample evidence that withtheassumptionthatthereactionisprecededbyhydroxide
strong chemisorption of a hydrogen atom requires three to adsorption (eq 5):
seven adjacent metal atoms, coined as the ensemble effect.67
Mitsuietal.68haveshownthatthecreationofactivesitesforH OH− + * ⇄ (*−OHad) + e− (5)
2
dissociation involves the formation of individual vacancies and
Thus, it is supposed that the catalyst surface should be
their subsequent diffusion and aggregation, with the coupling oxophilic, or “hydroxophilic”, in order to provide bifunctional
between these events determining the activity of the catalyst catalysis (Figure 6),22,70,71 which will be discussed in section
surface. Their scanning tunneling microscopy observations of
4.2.Inalkalinemedia,thecatalyticsurfacehastoprovideactive
the transient formation of active sites for the dissociative
sites for the coadsorption of hydrogen atoms and hydroxide
adsorption of H molecules on Pd(111) revealed that two-
2 species, whereas in acid all of the accessible active sites are
vacancy sites are inactive. Two-vacancy sites have low H
2 destined to uniquely chemisorb hydrogen atoms. This could
sticking probability and zero H adsorption events (Table 1).
2 explain, among other reasons, why catalysts exhibit several
times to several orders of magnitude lower electrochemical
Table 1. Active-Site Statistics for Vacancy Clusters for HOR activity in alkaline environments compared with acidic
Pd(111) at 65 K in 2 × 10−7 Torr H2 As Determined by environments (Figure 2).38,39,45,58
Scanning Tunneling Microscopy with a Frame Time of (C) Discharge of the adsorbed hydrogen atom (H ), which
75 s68 is called the Volmer (V) reaction72 (eq 6): ad
vmaecaanncaydcsolursptteironsizteim(eat(osm)s) 2>2×104 3330 4150 5120 (*−Had) + OH− → * + H2O + e− (6)
H stickingcoefficient <0.0001 0.005 0.008 0.008 Nondissociative adsorption of hydrogen to form a hydrogen
2
meanclusterseparationtime(s) 590 710 670 820 moleculeionintermediatewasalsoproposedbyHoriuti73,74as
therate-determiningstep(RDS)forsomesubstratesaccording
to the following reactions (eqs 7 and 8):
AregqguriergeadtefsorofeffithcrieeentorHmodriesshoycdiartoiognen. Avtachanigchieesr acroevethraegreesfooref H2 + * ⇄ H2M → H+2M + e− (7)
2
Had, indirect interactions come into play, which are in most H+M ⇄ MH + H+ (8)
casesrepulsive.Theyleadtoasubstantialloweringoftheinitial 2
metal−hydrogen binding energy (HBE), a phenomenon called The essential feature of the Horiuti mechanism is that the
inducedsurfaceheterogeneity.Consequently,weaklyheld(and hydrogen molecule adsorption step is not accompanied by
chemically reactive) hydrogen species are formed. The hydrogen dissociation and requires only a single-atom catalyst
observations made by the authors68 might be used as an for the reaction. In the Heyrovsky−Volmer (H−V) sequence,
approach to synthesize new electrocatalytic materials with a hydrogen adsorption is dissociative but requires only a single
Figure6.SchematicrepresentationsofthehypothesizedbifunctionalmechanismintheHORelectrocatalysison(A)Pd/C−CeO,(B)PdNi,and
2
(C) Ni(OH)/Pt(111). CeO in (A), Ni in (B), and Ni(OH) in (C) provide the active sites for adsorption of reactive OH , and Pd (or Pt)
2 2 2 ad
providestheactivesitesfordissociativeadsorptionofH andproductionofH ,whichthenreactswithOH .Reprintedwithpermissionfromrefs
2 ad ad
18and 70.Copyright 2016Elsevier and 2013Springer Nature,respectively.
6669 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
site. Only the Tafel−Volmer (T−V) sequence requires dual
g
n
dtarmpdeAfstrorieienliriotlaeffsaueaoteoITnnshcccctensvdos.ththot,ridifricideieoacueoo7swinenarn5goncc.dthtihsphixsiutiaWvmooicerotbdetfhrsimhrnsyaioesbto.eticnnhieabitmchsoTtrhlaeaeeaetneubhdplhmtteiearenbPtshlatirnoreeeswtaotahsasoaeitriteunssyrermtsetueoeydgplgynsrfgerbtatfwioaeponea-omtsnsgclitrfnhuiltheeeeoeleserna.ievraebghoddetsheynhdsueRTtbodtairrhgsh−DoirtrPiesnboaovroVtSsgtefegcur,,sdne7imtiacintah6tnthhHhetbeaeieysgtyvs−tddheemnohfrieVreiofffaeotqros,trueyetgaelgerdse9ebsaahepucen:nssndy5lue-otd1dtaopagsnlfisterrahcooeetHtaetinmahgxhsootoceoeeeneoslxnrturfiioThRuddctdnyetopaeDhiidlatusrPaeSitnrsonlco(thodo,dvtHnewgrcttio.deihhboaH7aVnneeee)6s- ExpectedConsequencesThereof hypothesizedconsequences 1.Dual-siteadsorption. ff2.Noelectrolyteeect. ff3.Nosupportmaterialeect. ff4.Nosizeeect. ff5.Nospillovereect.*−6.TheHbindingenergyisthecrucialfactoradcontrollingtheHOR/HERkinetics.1.Dual-siteadsorption.ff2.Spillovereect. *−3.TheHbindingenergyisthecrucialfactoradcontrollingtheHOR/HERkinetics.*−4.OHadsorptionisthecrucialfactorcontrollitheHOR/HERkinetics.
d1tNdscRhwwisweo0ieyohhsaqncdo:tcsOeei4u1ahasrshrr9detoeeeHea=aa+neglrnnfe−cecgθtdcn=eh1neaisl)tFiica,mrSRiocrwkoasatrgofTan−onenicertHtdsn/sdHhdhipFeeti22teeonr(xh,hktacp1hnganheantc(arieo−sitvgs−astp5fveooen1aTot2eloθrydhllrpfsdy,)ηtaeorchrcg)okaoel−nreeibncpayenu=ecasoffiemnvoattelsaoiaaFrncfeoirblleg0ikludneeifiotpcneenrhθttutrxehtheaoo2soposealpffoy(pftmoe2fttlatlhsahttαeehneottheetcfheidoaeηntee,dlrtH)uRxhowiTmpcRsDenOuieastDSrRrfhage(ifinSamnalPtfos,cdoht-etens(plhlenlelpochoetabPecaapawdRxy)tlstehits).enDdroaatgaA(Sdhikdct7n,iiatestnc0ioadootTe1nrion(trrbmd−1ti(eaceehM09otcVVsαdqee))tf. fitsinBasicSolutionsConrmedExperimentallyand experimentalevidence↔1.TheHORrateisclosetotheHDexchangerateinthegas2251phase.2.TherateofoxidationofpreadsorbedHisatleastanorderofad76magnitudegreaterthantheHORrate.*−3.TheHORissize-independent,whileoxidationofHisad38size-dependent.θawithaslope4.ThereisalinearcorrelationbetweeniandHUPD46of2.625.iisnotpH-dependent.0 fi−391.GoodtofthedatatotheButlerVolmerequation.2.ThereisacorrelationbetweentheHORactivityandtheHBEa39derivedfromtheHpeakpotential.UPD3.IdenticalactivitiesaredeterminedfromtheHORmeasurements45chargetransferresistance.andtheHUPD
values i = 0.233 mA cm−2 , n = 2, and α = 0.43. ys
0 Pt al
The reported data on the HOR kinetics and mechanism for at −e
platinum-based electrocatalysts in basic media in the temper- oc +
ature range of 2−70 °C are summarized in Table S2. ectr HO2
ThissectionhassummarizedthepossibleHORmechanisms El +
*
in alkaline media, considering in detail the Tafel, Heyrovsky, d
e →
Volmer, and Horiuti mechanisms. Experimental evidence for as −
B H
eachmechanismisprovided,whichreveals,asoutlinedinTable t- O
2, that the HOR mechanism is still disputable, even for well- P +
n )
studied Pt catalysts. In the next section, more attention is paid o Had
to the role of dissociative chemisorption (Tafel reaction) of ms *−
hydrogen on the kinetics of the HOR. nis ):(
atHshcclhtaehriohneanenaOanesdt3nddgHHtemdR.drlbct2tieo22Hshiotoec.ikgeero⇄+yaieiaennmlnnTdHn−mes2hreeht2fBtsoOhoietieoiHhtsEcaegutyrreHsleme+,nhel.iHrls−idnocWg+totcEaoghhh⇄adtBemhRealsto2yynatuii/elpnlneeffi2Hyeib−limdaHeasxtetcOitxiirmhepissn2eesaitRcrOsgnieioacnntp(sbmfsr+Ecerbodeabee.aniabweameoyNc2selseetcfeeoerbeir.r−vgtoffvie(aHuipHfeytelltylstre.qyiorutctttudhSroceoo1eeohrrnotc1poloomaelf)ogomgtsaeaeletsraeoonlo,nnyangmtsntesdiuhuiditrocottseereiimhaC,fdlensaylHocsheecerraldteetBisiocacnbesmEil.ructedl)uieslricdaunrfs(tctiamerhdomhechcfseeqaareiece3ntrpcrsrm9ae1eemit,lp((4fos2yirio115toatcefs)h,oo12ri7atan:neea))7srrll Table2.SummaryofProbableHORMecha HORmechanism*→*−dissociation(Tafel):H+22(H)dual-siteH22ad dual-sitedischargeofadsorbedhydrogenatom(Volmer aHdenotesunderpotential-depositedhydrogen.UPD
6670 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
The HBE has a pure thermodynamic definition, sometimes
interpreted as the reaction enthalpy (eq 13 or the simplified
form,eq14)orasthemetal−hydrogenbindingenergy(eq15):
ΔH = − E F − TS° /2 + gRT
H H H (13)
ad UPD 2
where ΔH is the enthalpy of hydrogen adsorption, E is
Had HUPD
the potential (vs the reversible hydrogen electrode (RHE)) of
underpotential deposition (UPD) of hydrogen, S° is the
standard entropy of H adsorption (−130.684 J moHl2−1 K−1),
2
and g is a dimensionless interaction parameter in the frame of
Frumkinadsorption(g<0forattractiveinteractionsandg>0
for repulsive interactions);61
ΔH = − E F − TS° /2
H H H (14)
ad UPD 2
for the Langmuir isotherm of adsorption (with g = 0);
E = −D /2 − E F − TS° /2 Figure 7. Experimentally measured HER/HOR exchange current
M−H H2 HUPD H2 (15) densities (marked as dots) over different metal surfaces in acid
twhheedreissEoMc−iaHtioisntheneemrgeytaol−fthhyedHrogemnobleicnudlieng(4e3n2erkgJymaonld−1D).H612,7i8s slionleutiisotnhse,psilmotpteledkaisneatifcunmcotidoenlopflotthteedcaelicthuelarteadsaHfBuEnc.tTiohneosfotlihdebfrlueee
2 energyforhydrogenadsorptionaccordingtoeq17orasafunctionof
Certain formulas can be obtained, for example eqs 1639 and HBE. Reprinted with permission from ref 79. Copyright 2005 The
17,79 to estimate the i0 value, which is the kinetic parameter at Electrochemical Society.
or near equilibrium conditions (zero overpotential):
⎛ βFE ⎞ estimated from the HUPD peak potential values given the
i = A exp⎜− peak⎟ assumption that H is identical to the HOR intermediate.
0 ⎝ RT ⎠ (16) Though the bindingUPeDnergy values are scattered depending on
the type of material (bulk, pc, facet), the nature of the H
where E is the potential of H desorption;39
peak UPD adsorptionsite(ontop,face-centeredcubic(FCC),monolayer
i = k 1 exp⎛⎜−ΔGHad⎞⎟ v(MaluLe)s,eatrce.),caonrrdeltahteedcowveitrhagethoeftahteomsuicfacneubmybHerasd(oθf),tthheemHBaiEn
0 01 + exp(−ΔGHad) ⎝ kBT ⎠ components, as illustrated in Figure 8. The main components
kBT (17) arerepresentedby18differentelementsofperiod4(Ti,V,Cr,
where k is Boltzmann’s constant and ΔG is the free energy Fe, Co, Ni, and Cu), period 5 (Mo, Ru, Rh, Pd, and Ag), and
B Had period 6 (W, Re, Os, Ir, Pt, and Au) that are potentially
of hydrogen adsorption, which is described by the equation
ΔGHad= HBE + 0.24 eV. A volcano-type correlation between coofnthsiedierrheidghasHc2omstipcokninegntpsroobfathbeiliHtyO(oRreslteiccktrioncgactaoleyffistcsibenect,aui.ese.,
the HBE values and exchange current densities with the
theratioofthenumberofparticlesbeingboundtothenumber
maximum at ΔGHad ∼ 0 eV was observed for the HER and hitting the surface).66
HOR (Figure 7).79 This section has provided the key equations relating
The HBE is a thermodynamic parameter that can be thermodynamic parameters of the hydrogen adsorption/
employed to predict either the HOR or HER kinetics near desorption process such as the HBE, metal−hydrogen binding
equilibriumwithoutdistinctiononthenatureoftheelectrolyte energy, and free energy of hydrogen adsorption. An extensive
(acid or basic medium). However, this does not necessarily number of reported HBE values, along with other thermody-
mean that the HBE values might be used to predict the HOR namicparameters,boththeoreticalandexperimental,havebeen
kinetics far from equilibrium. Moreover, equality of the summarizedforaseriesofcatalysts(elementsofperiods4−6),
theoretically predicted exchange current density values for the taking into account the type of material, nature of adsorption
HORandHERdoesnotnecessarilyimplythesimilarityinthe site,andthecoverageofthesuface.TheHBEvalueshavebeen,
kineticsandmechanismofthereactions.Forinstance,itiswell- within each given period of elements, successfully correlated
known that the HER kinetics on Ni-based electrocatalysts with the atomic numbers of the elements, identifying Ni, Pd,
suffers from deactivation via nickel hydride formation,80 and Pt as potentially highly reactive metallic elements for the
whereas the HOR kinetics is limited by the surface passivation HER/HOR catalysis within the corresponding transition metal
via nickel hydroxide formation.81 Another example supporting groups. Equations to calculate the values of exchange current
this is given by Vilekar et al.82 in a theoretical study to explain density of the HOR, given the Tafel mechanism of the HOR,
the strong asymmetry observed in current versus potential have also been given in this section. Since HBE is still
curvesfortheHERandHORbranchesonPtin0.5MNaOH. considered one of the key HOR descriptors, the summarized
The authors found that the H−V pathway is dominant in the data may temporarily serve as a useful reference database for
regions −0.3 to −0.24 and +0.13 to +0.3 V vs RHE, whereas future studies.
theT−Vmechanismdominatesintheregions−0.1to0Vand 3.3. Effect of Underpotential- and Overpotential-
0 to 20 mV vs RHE.82 DepositedHydrogen.TheHORoccursatpositivepotentials
ThemajorityofHBEvaluesreportedintheliterature(Table with respect to the RHE. Thus, overpotential-deposited
S1)weredeterminedtheroretically,withtheexceptionofsome hydrogen (H ), which is believed to form at negative
OPD
studies39,77,83 where the thermodynamic parameters were potentials and to precede the HER84 (the existence of the
6671 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
Figure8.PeriodicdependenceofselectedtheoreticallycalculatedHBEvaluesfordifferentelements:(A)period4:Ti,V,Cr,Fe,Co,Ni,andCu;
(B)period 5: Mo, Ru,Rh, Pd, and Ag;(C) period 6:W, Re,Os, Ir, Pt,and Au.
speciesisdisputedintheliterature),45isoutofthescopeofthe morepositivethantheRHErevealsthatnotallmaterialshaving
current review. Cyclic voltammetry in the range of potentials high H sticking probability66 are able to host the so-called
2
6672 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
underpotential-deposited hydrogen atoms (H ). As an covered by H . On Pt(100) and Pt(111), the HOR onset
UPD UPD
example of such materials, we can find Re,85,86 Ni,87,88 appeared to correlate with the desorption of H . The
UPD
Fe,89−91 and Ti−Ni.92 Thus, H in aqueous solution is a Pt(111) surface with H coverage higher than 0.5 ML was
UPD UPD
phenomenon characteristic of certain metals,84 such as Pt,38 shown to be almost completely inactive for the HOR. Kinetic
Rh,93 Pd,39 Ru,94 Ir,61,95 and Mo,96 both in acidic and basic analysis according to the empirical relation between H and
UPD
electrolytes. The origin of hydrogen underpotential deposition the number of bare Pt sites available for H dissociation (eq
2
ispurely thermodynamic; onthe other transition metals,H 18),
UPD
formation does not take place in the same or other potential
H = i × (1 − θ )m
regions because ofthe electrosorption ofO-containing species, UPD θHUPD=0 HUPD (18)
namely, oxygen or hydroxide, which is energetically more
favorable. Thus, the effect of HUPD formation can be ruled out rseitveefaolermdtohfatthPet(T1−11V)sienqauleknaclienewistohlumtio=n2f,owllhoewrseatshethiedePatl(d1u1a0l)-
in the case of non-PGM catalysts but cannot be straightfor- surface does not fit the model in the sense that m = 0.5 does
wardly excluded for PGMs. not have a physical meaning. Marković et al.46 hypothesized
AninterestingquestionconcernstherolethatH mayplay
UPD that there are sites for H dissociation on the Pt(110) surface
in the electrochemical oxidation of hydrogen molecules. It is 2
that are not occupied by H , although a full monolayer of
believed that the electrocatalytic oxidation of hydrogen UPD
hydrogen is adsorbed.
moleculesonPtessentiallyinvolvesatomicHinachemisorbed
Strong inhibition of the HOR in the H potential region
state, Pt−H , as a reactive intermediate. There are debates in UPD
ad was also observed for Pt(111), Pt(100), and Pt(110) single-
the literature on the chemical nature of Had,48,50,84 and crystal surfaces in 0.1 M KOH over the temperature range 2−
currently there are two viewpoints. In one, HUPD is considered 60 °C;97 however, it was also proposed that OH coadsorbs
as the intermediate involved in the HOR;84 in the other, it is with H onallthreesurfaces inthe H potentadialregion,98
UPD UPD
believed that another form of adsorbed hydrogen acts as the accordingtothesignificantreactionratesofCOoxidationatE
intDerimffeerdeinatteisaontdheHrmUPsDfaocrtsHaUsPaDbalnodckivnegryindteiffrmereednitatree.4a6c,9ti7vity <be0fo.1reVt.hAesoannseitnodfirtehcet pHrEooRf,wthasedmeatexrimmuinmedcotovebraeg0e.8o2f,H0.U7P8D,
values for the HOR were obtained for Pt(111), Pt(100), and and0.65MLonPt(110),Pt(100),andPt(111),respectively(1
Pt(110)single-crystalsurfacesinbasicmedia(Figure9).46The ML ≡ 1 H per Pt), the rest likely being covered by OH .
UPD ad
HOR was shown to occur on Pt(110) even on a surface fully FortheHOR,theH reactioncanformallybeconsideredas
UPD
the Volmer reaction, which is the RDS. To support this
hypothesis, Durst et al.45 compared the charge transfer
resistance(CTR,R )fortheH reactionwiththeexchange
CT UPD
current density of the HOR/HER, derived from eq 19:
RT i RT 1
i = × = ×
0 F η F R (19)
CT
The exchange current density values for the Volmer reaction
estimatedbyDurstetal.areverysimilartotheHORexchange
current densities of ca. 200 mA cm−2 on Pt/C. In 0.05 M
NaOH, the charge transfer resistance of 13−54 Ω cm−2
obtained on single-crystal Pt surfaces99 corresponds to 0.2−
0.5mAcm−2 whichisveryclosetothevalueof1.0mAcm−2
Pt Pt
obtained for Pt/C in 0.1 M NaOH. One possibility for the
reducedrateoftheVolmerstepinbasewouldbeahigherPt−
Hbondstrength,whichwoulddecreasetheHORrate.Ahigher
Pt−H bond strength is confirmed by the positive shift of the
H peakswithrespecttotheRHEwithincreasingpH.These
UPD
shifts on Pt(553) and Pt(533)100 as well as on Pt(pc),101
amountto−10and−11mVvsRHEperpHunit,respectively
(see other values of d(HBE)/d(pH) in Table S2). The H
UPD
peak shift translated into an HBE difference of 12.5−13.5 kJ
mol−1 from pH 0 to pH 13. If the difference in the HBE is
assumed to be proportional to the difference in the activation
energy (according to the Brønsted−Evans−Polanyi relation-
ship),102 then the difference in the HOR rate between pH 0
and pH 13 for Pt would amount to a factor of 120−200. This
difference is in good agreement with the 210-fold difference in
i values at pH 0 and pH 13 (see Figure 2),45
0,313K
strengthening the hypothesis that the Volmer step is the RDS
for the HOR on Pt.
Figure 9. (A) Plots of the charge associated with the underpotential
Itshouldbementionedthattheconclusionsonthenatureof
depositionofhydrogenandadsorptionofhydroxylspeciesonPt(hkl).
(B) Polarization curves for the HOR on Pt(hkl) in 0.1 M KOH at the RDS of the HOR, even for Pt, are still controversial. For
3600 rpm. Reprinted with permission from ref 46. Copyright 1996 example,ConwayandTilak84indicatedthat“thereseemstobe
RoyalSociety of Chemistry. no dispute regarding the mechanism of the HOR on Pt (and
6673 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690
ACS Catalysis Review
Pt-alloys) being rate-determining dissociative adsorption of Insummary,inthissectiontheeffectofH onthekinetics
OPD
H ”, referring to other studies50,51 to provide rationale for the inthepotentialrangeofhydrogenoxidationhasbeenruledout.
2
statement.Vogeletal.50andRossandStonehart51haveshown The section has discussed the role of H , which is
UPD
that the electrochemical reaction rate parameters are the same experimentally detected only for certain catalytic materials
on smooth Pt, unsupported Pt black, and Pt crystallites such as Pt, Rh, Pd, Ru, Ir, and Mo and, to the best of our
supportedongraphitizedcarbon. ThiscorrelateswiththeH − knowledge, is not detected on Ni-based materials. Thorough
2
D exchange in the gas phase both with and without literature analysis revealed two main contradicting hypotheses
2
chemisorbed carbon monoxide and is dependent only upon on the role of H in the HOR, namely, H as the
UPD UPD
thesurfaceareaoftheplatinumcatalyst.Therefore,theRDSfor intermediateinvolvedintheVolmermechanismversusH as
UPD
hydrogenmoleculeoxidationonPtwasassumedtobethedual- the blocking intermediate, with the second hypothesis being
site dissociative chemisorption of the hydrogen molecule (H more likely.
2
→ 2H+, i.e., the Tafel reaction). In recent work, Zheng et al.39
and Bhowmik et al.77 are inclined to directly relate the HBE 4. HOR ELECTROCATALYSTS
value, which is considered to be linearly dependent on the 4.1. Overview of Different Material Classes. This
H desorptionpotential,totheHORcurrentdensityinbasic
UPD sectionsummarizesthemaintypesoftheHORelectrocatalysts
media.ThismeansthattheVolmerreactionisconsideredtobe for alkaline media that have been investigated hitherto. Figure
thehydrogenoxidationreactionRDS.Sincethesurfacesofreal 11 shows the element-based classification used in the present
catalysts comprise sites with different specific activities (or reviewtodiscussthecharacteristicsandopportunitiesforthese
turnover frequencies), an approach was proposed to correlate different state-of-the-art PGM and non-PGM HOR electro-
the HOR activity and the number of active sites with similar catalysts.
HBE.61 4.1.1. Platinum-Based Catalysts. As evidenced by the
HUPD desorption profiles for Ir/C catalysts were deconvo- activity values presented in Tables S2 and S3, platinum is the
luted into three peaks (likely to be for the (110), (111), and most active catalyst for the HOR in terms of exchange current
(100) facets in the order of increase of the HUPD desorption density (Table S2) or mass activity at low overpotential (η =
peak).103ThepopulationofsiteswiththelowestHBE,namely, 0.05or0.1V)(TableS3).DifferentformsofPt-basedcatalysts
Ir(110) with an HBE value of −0.33 eV, was suggested to be
were investigated to catalyze the HOR in basic electrolytes,
responsiblefortheHOR.AsshownbyShengetal.,83theHOR including Pt(hkl) single-crystal facets,97 bulk/polycrystalline
activityforpolycrystallinePtobtainedusingtheRDEmethodis Pt,49 carbon-supported Pt,39,40,45 Pt−metal alloys,6,105 Pt
found to decrease with the pH. At same time, the HBE adlayers/islands,40 and core−shell structures.41,106
obtained from cyclic voltammograms (CVs) linearly increases
Despite numerous studies on Pt-based materials, the details
with the pH in buffer solutions (pH from 0 to 13) with the
oftheHORmechanismandthenatureoftherate-determining
slopes of −10 meV vs RHE per pH unit for Pt(110) and −8
step are still debated. The controversy arises over the
meV vs RHE per pH unit for Pt(100).
magnitude of the exchange current density i (discussed in
0
The universal correlation between i0 and HBE, which was section 2), the predominant mechanism (T−V, H−V, or dual-
directly derived from the HUPD desorption potential value, pathT−V+H−V),andthenatureandcoverageofthereactive
observed on Pt/C, Ir/C, Pd/C, and Rh/C catalysts (Figure intermediate H (see section 3.3). Conway and Tilak84
10)39 indicates that the HOR on those metals may share the ad
summarized kinetic data for the HOR and concluded that the
HORrateonPtisdeterminedbythedissociativeadsorptionof
H .SincetheydidnotobserveapHeffectontheHORkinetics
2
onPtwhenmeasuredatdifferentpH>4(Figure12),Bagotzky
and Osetrova62 came to the same conclusion.
TheHORrateonasmoothPtsurfacewasshowntobeclose
to the rate of hydrogen dissociation (i.e., the rate of H −D
2 2
exchange) measured in the gas phase (Figure 13).51 These
results indicate that the electrolyte (in the pH range below 4)
and/ortheelectrifiedinterfacedoesnotaffectthenatureofthe
potential energy along the hydrogen−platinum reaction
coordinate. This also leads to the conclusion that the HOR
mechanismoversmoothPtislikelyaslowdual-sitedissociative
adsorption of a hydrogen molecule followed by fast electro-
chemical oxidation of H to H O+.
ad 3
Figure 10. Correlation between the HOR/HER exchange current The HOR kinetics on Pt(pc) in 1 M KOH fits a direct
density on Pt/C, Ir/C, Pd/C, and Rh/C and the HUPD desorption discharge mechanism identical to the overall reaction H2 +
peakpotential. Reprinted fromref 39. 2OH− ⇄ 2H O + 2e−.49 It seems that the step of extracting
2
two electrons directly from H is slow, with an exchange
2
current density of 0.233 mA cm−2. Since essentially identical
samereactionmechanism.Zhengetal.39assumethattheHOR values of the HOR specific exchange current density (0.69 ±
proceedsviatheT−Vpathway,withtheVolmerreactionbeing 0.01 vs 0.57 ± 0.07 mA cm−2; Table S2), activation energy
the RDS, and that the mechanism through which pH affects (28.9 ± 4.3 vs 29.5 ± 4.0 kJ mol−1), and charge transfer
HBE is likely to be metal-independent. However, Santos et coefficient are observed on Pt(pc) and Pt/C, it was
al.104showedthateventhough noH signalisregisteredon hypothesizedthatthereisnosignificanteffectofthePtparticle
UPD
Sn-rich Pt−Sn catalytic systems, they show higher exchange size on the HOR in 0.1 M KOH.38 In contrast, Stonehart and
current density values compared with monometallic Pt. Lundquist75reportedacrystallitesizeeffectfortheoxidationof
6674 DOI:10.1021/acscatal.8b00689
ACSCatal.2018,8,6665−6690