Electrochemical oxidation behavior of titanium nitride based

G

E

Eu

BC

a

ARRAA

KPCECTEX

1

sdfvcptaPrcwc

1miii

0d

ARTICLE IN PRESSModel

A-16125; No. of Pages 11

Electrochimica Acta xxx (2010) xxx–xxx

Contents lists available at ScienceDirect

Electrochimica Acta

journa l homepage: www.e lsev ier .com/ locate /e lec tac ta

lectrochemical oxidation behavior of titanium nitride based electrocatalystsnder PEM fuel cell conditions

harat Avasarala ∗, Pradeep Haldarollege of Nanoscale Science and Engineering, University at Albany, SUNY, 253 Fuller Road, Albany, NY 12203, USA

r t i c l e i n f o

rticle history:eceived 26 May 2010eceived in revised form 7 July 2010ccepted 9 August 2010vailable online xxx

eywords:roton exchange membrane fuel cells

a b s t r a c t

Titanium nitride (TiN) is attracting attention as a promising material for low temperature proton exchangemembrane fuel cells. With its high electrical conductivity and resistance to oxidation, TiN has a potentialto act as a durable electrocatalyst material. Using electrochemical and spectroscopic techniques, theelectrochemical oxidation properties of TiN nanoparticles (NP) are studied under PEM fuel cell conditionsand compared with conventional carbon black supports. It is observed that TiN NP has a significantly lowerrate of electrochemical oxidation than carbon black due to its inert nature and the presence of a nativeoxide/oxynitride layer on its surface. Depending on the temperature and the acidic media used in the

atalyst supportslectrochemical oxidationorrosionitanium nitridelectrocatalyst

electrochemical conditions, the open circuit potential (OCP) curves shows the overlayer dissolved in theacidic solution leading to the passivation of the exposed nitride surface. It is shown that TiN NP displayspassive behavior under the tested conditions. The XPS characterization further supports the dissolutionargument and shows that the surface becomes passivated with the O–H groups reducing the electricalconductivity of TiN NP. The long-term stability of the Pt/TiN electrocatalysts is tested under PEM fuel

endspropo

PS cell conditions and the trshown to agree with the

. Introduction

In the past decade, fuel cell technology has made significanttrides towards commercialization but much work needs to beone in many aspects of this promising alternative energy sourceor it to compete against the conventional energy sources. Of thearious types of fuel cells, proton exchange membrane (PEM) fuelells have received broad attention due to their low operating tem-erature, low emissions and a quick startup time. But the cost andhe lifetime of a PEM fuel cell system are the major challenges thatre hindering its large-scale commercialization [1,2]. Lifetime of aEM fuel cell is mainly dependent on the durability of its mate-ial components [3]. As the PEM fuel cell operates under harshonditions, high-performance durable materials are required toithstand the degradation caused due to these corrosive operating

onditions.The current fuel cell durability demonstrated in vehicles is

977 h (∼60,000 miles) while the requirement for large-scale com-

Please cite this article in press as: B. Avasarala, P. Haldar, ElectrochemicalPEM fuel cell conditions, Electrochim. Acta (2010), doi:10.1016/j.electacta.

ercialization is 5000 h (∼150,000 miles) [2], which the currentnternal combustion (IC) engine based automotive vehicles can eas-ly accomplish. Apart from the lifetime costs, durability can alsonfluence the reliability of the PEM fuel cell technology against its

∗ Corresponding author.E-mail address: [email protected] (B. Avasarala).

013-4686/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.oi:10.1016/j.electacta.2010.08.035

of the measured electrochemical surface area at different temperatures issed passivation model.

© 2010 Elsevier Ltd. All rights reserved.

counterparts.Investigations have revealed that a considerable part of the

performance loss is due to degradation of the electrocatalyst [4]during extended operation and repeated cycling [5], especiallyfor PEM fuel cells in automotive applications. Currently, the car-bon black supported platinum nanoparticles (Pt/C) remains thestate of the art electrocatalyst for PEM fuel cells. Under the cor-rosive operating conditions, the Pt/C degrades via Pt dissolution, Ptparticle agglomeration and carbon support corrosion mechanismsresulting, primarily, in the loss of electrochemical surface area. Fur-thermore, the catalytic metal, especially Pt, catalyzes the oxidationof carbon [6,7] and the oxidation of carbon black accelerates Pt sin-tering [8]. Vulcan XC-72 carbon black is the most popular catalystsupport currently used in the Pt/C electrocatalysts but its durability,under the oxidizing conditions of a PEM fuel cell, needs significantimprovement [9,10].

An ideal catalyst support material should have corrosion resis-tance properties under strongly oxidizing conditions of PEM fuelcell: high water content, low pH (<1), high temperature (50–90 ◦C),high potentials (>0.9 V) and high oxygen concentration. But car-bon is known to undergo electrochemical oxidation to form surfaceoxides and CO/CO under these conditions [11–13]. Significant

oxidation behavior of titanium nitride based electrocatalysts under2010.08.035

2oxidation of carbon support can be expected to decrease theperformance of a PEM fuel cell [13,14], due to the loss and/oragglomeration of Pt particles caused by the carbon corrosion. Ithas also been reported [14,15] that oxygen containing groups (e.g.,

dx.doi.org/10.1016/j.electacta.2010.08.035


http://www.sciencedirect.com/science/journal/00134686

http://www.elsevier.com/locate/electacta

mailto:[email protected]


ING

E

2 rochim

cbvdwtrotgs

(ecsoshanaamttacaddehpicwrcfts

2

aosaNw[btm

aanhwolt

ARTICLEModel


B. Avasarala, P. Haldar / Elect

arboxyl, carbonyl, hydroxyl, phenol) can be formed on the car-on surface at high temperatures and/or high potentials (>1.0 Vs. RHE at room temperature or >0.8 V vs. RHE at 65 ◦C) whichecrease the conductivity of catalysts and weaken its interactionith the support resulting in an accelerated Pt sintering [8,16,17];

hus drastically affecting the performance of PEM fuel cell. In oureport [9], we showed that carbon black supports have a higher ratef corrosion under potential cycling compared to constant poten-ial conditions and tend to form significant amount of surface oxideroups. These studies indicate towards the vulnerability of carbonupport under prolonged fuel cell operating conditions.

In our recent publication [18], we introduced titanium nitrideTiN) nanoparticles as catalyst supports for synthesizing Pt/TiNlectrocatalyst and showed that it can outperform the Pt/C electro-atalyst in electrochemical surface area and catalytic activity for theame Pt particle size and loading [18]. With its unique propertiesf higher electrical conductivity (relative to carbon) and an out-tanding oxidation and acid corrosion resistance [19–21], TiN has aigh potential for developing durable electrocatalysts. Research islso being pursued on the materials of transition metal carbide anditride family for their application in low temperature fuel cells,s described in a review by Ham and Lee [22]. Recently Musthafand Sampath [23] reported the application of TiN films for directethanol based fuel cells and other groups [24,25] have reported

he application of Ti-based compounds for fuel cells [24,25]. Buthe stability and the long-term durability aspect of the electrocat-lyst material under above room temperature fuel cell operatingonditions is not discussed. As titanium nitride gains attention asn electrocatalyst material, its behavior under PEM fuel cell con-itions remains to be investigated. Considerable research has beenone in the past in electrochemically evaluating TiN thin films inlectrolytes of a wide-pH range. But behavior of TiN vis-à-vis –igh surface area porous material, above RT conditions, cyclic/highotentials, highly acidic conditions (pH < 1) – which are the typ-

cal conditional variables of an electrocatalyst material in a fuelell, has not been reported in the literature. Through this paper,e extend upon our research on the TiN based electrocatalysts and

eport the electrochemical behavior of titanium nitride nanoparti-les by evaluating them under the operating conditions of a PEMuel cell. Along with the bare supports, we also report the elec-rochemical behavior of synthesized Pt/TiN electrocatalyst and itstability under prolonged exposure conditions.

. Titanium nitride

Titanium nitride belongs to the transition metal nitride familynd displays similar formation and physicochemical properties ofther nitrides in the family. Many of its properties are caused by apecial feature of its electronic structure, i.e., transfer of part of s-nd d-valency electrons of the atoms of Ti to the 2p-states of the

thereby depleting the electron population of the d-states of Ti,hich are responsible for its covalent bonding with foreign atoms

26,27]. According to Ham and Lee [22], this deficiency in the d-and occupation near the Fermi level should make the TiN surfaceo have a reduced ability to donate d-electrons to adsorbates; thus

aking it a better electron acceptor.The triple bond between Ti and N and the mixed p and d char-

cteristics of the interaction between Ti 3d and N 2p representsn essential aspect of covalent bonding between titanium anditrogen atoms and is responsible for its inert nature, mechanical


ardness and high melting point [28]. The metallic character arisesith one unpaired electron going into a metal localized sp hybrid

rbital on Ti resulting in the non-zero electron density at Fermievel (EF) [28] contributing to its relatively high electrical conduc-ivity. Oyama [19] reported the electrical conductivity of bulk TiN as

PRESSica Acta xxx (2010) xxx–xxx

4000 S cm−1 and that of C as 1190 S cm−1 while other groups havereported values ranging from 3125 S cm−1 [29] to 33,000 S cm−1

[30] for polycrystalline TiN thin films to ∼55,500 S cm−1 [28 andRefs. within] for TiN single crystal, conductivities that are higherthan that of carbon.

TiN is well known for it oxidation resistance property, an aspectcontributed to the native oxide/oxynitride layer which partiallycovers its surface and is formed due to atmospheric oxidation[28,31–33]. The titanium–oxygen–nitrogen system [34] has beenwell researched and the composition of the native layer has beenattributed to a non-uniformly distributed mixture of oxynitrideand oxide [28,31,32,35–38] components on its surface that pre-vent the surface from further oxidation. The oxide and oxynitridecomponents are also formed during the electrochemical oxidationof titanium nitride under potential scanning conditions [28].

2.1. Oxidation

TiN undergoes a thermodynamically favorable oxidation reac-tion when exposed to air [28,34]:

2TiN + 2O2 → 2TiO2 + N2 ↑ �G◦ = −611.8 kJ mol−1(298 K) (1)

The oxidation process takes place in three stages [35,39]: stage1: oxygen diffuses into the TiN lattice resulting in replacement of Nby O, leading to the formation of titanium oxynitride. Stage 2: thenatural protective films of the phases of TinO2n−1 types are formedon the oxynitride layer during this stage, representing a diffusion-controlled oxidation [40]. Stage 3: slow oxidation of the residualmaterial of TiN coated with TiO2.

A major portion of the replaced nitrogen is released in the inter-stitial positions of surface oxide layers as molecular nitrogen, whichgets released from the surface upon further oxidation [35].

2.2. Electrochemical oxidation of TiN

Many research papers have evaluated the electrochemicaloxidation behavior of titanium nitride films in various media[28,32,36–38], and showed the high stability of TiN against oxi-dation process in a wide-pH range. The electrochemical oxidationresistive properties were attributed to the presence of the nitrogen-enriched surface layer of titanium oxynitride with a large electrondensity that screens the underlying titanium ions and inhibits theoxidation reaction [20,28].

The electrochemical oxidation of TiN follows a similar path ofoxidation process, where: 0.5 V to 0.8–0.9 V: the formation andgrowth of oxide/oxynitride films on the surface of TiN by reaction(2) predominantly leads to the retardation (or passivation) of theseprocesses.

TiN + 2H2O → TiO2 + 1/2N2 + 4H+ + 4e−[28, 41] (2)

TiN = Ti3+ + 1/2N2 + 3e− (3)

∼1.0–1.5 V: oxidation resulting in formation of hydroxides [41],

TiN + 3H2O → Ti(OH)3 + N2 + 3H+ + 3e− (4)

TiN + 3H2O → TiO2·H2O + 1/2N2 + 4H+ + 4e− (5)

Milosev et al. [28] attributed the increase in current in thispotential region to the formation of TiO2 on the surface of TiN ratherthan hydroxide (reaction (4)), but accompanied by a similar lib-eration of N2. The reasoning being that the hydroxides are moresoluble and less protective than the oxides, and can cause the rise


of currents in this region. Above 2.0 V: at these potentials, oxygenevolution takes place with a simultaneous oxidation of titaniumnitride to TiO2.

Through the above discussion in oxidation and electrochemicaloxidation processes, we intend to better address the nature of TiN


IN PRESSG

E

rochimica Acta xxx (2010) xxx–xxx 3

nt

3

3

pUcsi(dapcVa

Faa

ARTICLEModel



anoparticles under PEM fuel cell conditions by fully consideringhe reported electrochemical behavior of TiN in the literature.

. Experimental

.1. Catalyst support materials

The TiN nanoparticles or TiN NP that are used as catalyst sup-orts in this study were commercially purchased (NanoAmor Inc.,SA) and have an average particle size (APS) of 20 nm and a spe-ific surface area (SSA) of 40–55 m2 g−1 [42]. The mean particleize, derived from the transmission electron microscopic (TEM)mage (Fig. 1) and the X-ray diffraction pattern of TiN(1 1 1) peakFig. 2) using Debye–Scherer equation, match closely to the ven-or reported value. The purity ascribed to the TiN was >97% andn energy-dispersive spectroscopy was performed to verify the


resence of impurities, as shown in Fig. A1 of Appendix. For aomparative investigation with carbon, commercially purchasedulcan XC-72R carbon black (Cabot Inc., USA) of a known surfacerea of ∼250 m2 g−1 is used.

ig. 1. (a) Transmission electron micrograph showing the TiN nanoparticles and (b)histogram showing the TiN nanoparticles distribution with average particle size

t 18.9 nm.

Fig. 2. X-ray diffraction pattern of TiN nanoparticles.

3.2. Preparation of catalyst support coated electrodes andelectrochemical tests

For electrochemical comparison of TiN and C, the TiN NP is son-icated in ethanol and an aliquot is deposited on a polished glassycarbon electrode (GCE, 5 mm dia., 0.1962 cm2, Pine Instruments)for a material loading of 0.15 mg cm−2 and the solvent is allowedto evaporate. The TiN NP coated GCE is used as a working electrodein a three-electrode cell set-up and a similar procedure is followedfor depositing carbon black (CB) on polished GCE.

For spectroscopic characterization of TiN, carbon cloth (E-Tek®)is used as a substrate on which TiN NP is coated. The coating isprepared by ultrasonically mixing TiN NP, isopropyl alcohol and0.05 wt% Nafion® for a loading of 3 mg cm−2, with 95 wt% of TiNNP. Electrodes were dried at 80 ◦C for 10 min for the evaporation ofany remaining solvent.

Electrochemical tests were performed in a standard three-electrode cell set-up with a PARSTAT 2273 potentiostat (PrincetonApplied Research, USA). A reversible hydrogen electrode (RHE)(Gaskatel GmBH, Germany) is used as a reference electrode witha Pt mesh acting as a counter electrode. All potentials are reportedwith respect to RHE. Using cyclic voltammetry (CV) techniques,the GCE was cycled between 0 and 1.0 V potentials at 50 mV s−1

in Ar saturated 0.1 M HClO4 (pH < 0.5) electrolyte at 60 ◦C for 24 h.The maximum potential of 1.0 V is chosen as it is near to fuel cellidle conditions. Similarly, the TiN NP coated carbon cloth electrodeis held vertically inside the electrolyte chamber under the sameconditions but cycled between 0 and 1.2 V potentials at 50 mV s−1

for 16 h. Potential cycling is chosen for the electrochemical eval-uation as it is one of the extreme operating conditions in a PEMfuel cell where the electrode is exposed to repeat variations inload due to changing potentials that result in the application ofoxidation/reduction processes on the electrode surface.

3.3. Synthesis of Pt/TiN electrocatalyst

TiN NP is mixed in ethylene glycol (EG) and ultrasonicallytreated before adding H2PtCl6·6H2O drop wise. Sodium hydroxide(NaOH) is added to control the size of Pt nanoparticles by adjustingthe pH of the solution (∼12). The solution is stirred well before heat-


ing at 160 ◦C for 3 h in a N2 atmosphere under refluxing conditions.The solution is stirred overnight at room temperature, after whichit is washed and filtered. The filtrate removed powder is dried inair at 80 ◦C for 4 h.


ING

E

4 rochim

3h

Papl(accu1tua

3

f2osauisl0ismoas

do

4

4

nap(BN

stgl(‘tnNsos

cycling (as the nitride essentially converts to oxide and oxynitridecomponents due to electrochemical oxidation).

In the case of CV curve for TiN NP electrode shown in Fig. 3,there is a wide margin of current density drop after 0.5 h with thelater cycles 8 h, 16 h (not shown in Fig. 3) and 24 h having the same

ARTICLEModel



.4. Preparation of Pt/TiN electrocatalyst coated electrodes andalf-cell performance

For determining the electrochemical surface area (ECSA) oft/TiN, the CV technique is used. The electrocatalyst dispersionsre prepared by ultrasonically mixing the Pt/TiN in ethanol andipetting an aliquot of the dispersion onto the GCE for a cata-

yst loading of 20 mgPt cm−2. After drying, 10 �l of aqueous Nafion5 wt%) solution is pipetted onto the electrode surface in order tottach the catalyst particles onto the electrode surface and also toreate pathways for the H+ ions in the electrolyte to access theatalytic sites. The Pt/TiN deposited GCE is immersed in the Ar sat-rated 0.1 M HClO4 which is at 60 ◦C and scanned between 0 and.2 V at 50 mV s−1. Half-cell performance is measured by the elec-rochemical surface area (ECSA) which is calculated from the areander the hydrogen adsorption and desorption peaks of CV curvesfter double-layer correction.

.5. Surface characterization

X-ray photoelectron spectroscopy (XPS) analysis was per-ormed using a ThermoFisher Theta Probe system which has a-dimensional detector that will allow simultaneous collectionf spectral data from all angles without tilting the sample. Thepectrometer was equipped with a hemispherical analyzer andmonochromator and all XPS data presented here are acquired

sing Al K� X-rays (1486.6 eV) operated at 100 W. Sputter clean-ng was done with a differentially pumped Ar+ sputter gun. Theample to analyzer takeoff angle was 45◦. Survey spectra were col-ected at pass energy (PE) of 125 eV over the binding energy range–1300 eV. High binding energy resolution multiplex data for the

ndividual elements were collected at a PE of 29.55 eV. The XPSpectra were background subtracted using the non-linear, Shirleyethod. The binding energies of the Ti 2p, N 1s, O 1s and C 1s

f the TiN electrodes were calibrated with respect to C 1s peakt 284.6 eV. XPS characterization was performed on the electrodeamples before and after the electrochemical treatments.

A two-theta X-ray scan is performed on the TiN NP using a X-rayiffractometer XDS 2000 (Scintag Inc., USA), where the wavelengthf X-rays from Cu source is 1.54 A.

. Results and discussion

.1. Oxide/oxynitride on TiN nanoparticles

The presence of the oxide/oxynitride components on the TiNanoparticles is confirmed using X-ray diffraction (XRD) technique,s shown in Fig. 2. The strong peaks of TiN seen in the diffractionattern can be indexed to a pure NaCl-type structure (S.G. Fm3m2 2 5)). The lattice constant is ‘a’ = 4.212 A, which is calculated fromragg’s angle of the (1 1 1) diffraction plane of XRD spectra of TiNP (calculations in Appendix).

The oxynitride (TiOxNy) is an intermediate phase and can be con-idered as a solid solution of titanium nitride (TiN) and the cubicitanium oxide (TiO) [43]. Hence, the cubic lattice constant ‘a’ of aiven TiOxNy, according to Vegard’s law relationship [34,44], shouldie between the value of lattice constants of TiN (4.241 A) and TiO4.185 A). The lattice constant calculated for TiN nanoparticles isa’ = 4.212 A is found to lie between these two values thus implyinghe existence of an oxynitride layer (TiOxNy) on the surface of TiN


anoparticles [34,44,45]. A comparison is made between the TiNP and the standard stoichiometric TiN XRD peaks and the discus-

ion is provided in Table A1 of Appendix. Further evidence of thexynitride layer on TiN NP is also shown in the X-ray photoelectronpectra of Ti 2p3/2 in Fig. 6, where the deconvoluted peak between


nitride and oxide peaks is attributed to the oxynitride which has anoxidation state between TiN and TiO2 [28,35].

The native layer on the surface of TiN nanoparticles used inour study is a combination of Stage 1 (TiOxNy phase) and Stage2 (TinO2n−1 phase) of the oxidation process (Section 2.1), with thelatter being predominant while no complete phase separation toTiO2 [28] is seen to occur from the XRD plot in Fig. 2.

Using the TiN NP with partially covered native oxide/oxynitridelayer, the Pt/TiN electrocatalyst was synthesized and its electro-chemical evaluation has been reported in our previous paper [18].The cyclic voltammogram (CV) or the oxygen reduction reaction(ORR) curves for Pt/TiN showed no indication of an ohmic resistancedue to the oxide/oxynitride layer on the surface of TiN nanoparti-cles. In fact, 20 wt% Pt/TiN showed higher electrochemical surfacearea and catalytic activity compared to the conventional 20 wt%BASF® Pt/C [18] for the same Pt loading. If the oxide/oxynitride layerwere to act as a poor electrical conductor, then the ohmic losses(due to the inhibition of electron transportation from the catalystparticle to the support) should cause retardation in the kinetics ofthe catalytic reactions; which was not observed in our tests sug-gesting an alternative possibility. In the later part of this report, anexplanation to this behavior and the role of the oxide/oxynitridelayer on TiN nanoparticles are discussed.

4.2. Electrochemical comparison of CB and TiN NP under PEM fuelcell operating conditions

Fig. 3 shows the CV curves of TiN NP and CB electrodes, where thecurrent densities generated by TiN NP are significantly lower whencompared to those of CB under identical test conditions. The inertnature of TiN and the presence of an oxide/oxynitride layer on itssurface attribute to its initial lower current densities (TiN (0.5 h) vs.C (0.5 h) in Fig. 3). Since the native oxide/oxynitride layer on TiN NPacts as a barrier, the low oxidation current densities observed are,predominantly, a result of the electrochemical oxidation of nitridecomponent on the surface. Based on these two arguments, it canbe reasoned that the composition of oxynitride and oxide compo-nents will increase while that of nitride decreases after the potential


Fig. 3. Cyclic voltammogram (CV) of TiN and CB after potential cycling between 0and 1.0 V (50 mV s−1) in an inert-gas saturated 0.1 M HClO4 at 60 ◦C using a three-electrode cell set-up. Material loading on GCE is 0.15 mg cm−2.


ARTICLE IN PRESSG Model

EA-16125; No. of Pages 11

B. Avasarala, P. Haldar / Electrochimica Acta xxx (2010) xxx–xxx 5

Fa5

mTc

titftdo

vfTssdpatacc

4

ucttuts

((

t

with a decrease of O and N, suggests that the surface is not pre-dominated by the oxide layer after the electrochemical treatment.Further investigation of the elemental peaks can provide informa-tion on the nature of surface groups formed on the surface of TiNand their trends after the electrochemical treatment.

Table 1The atomic concentration ratios of surface elements on TiN electrode: untreated and

ig. 4. Anodic curve of TiN and CB in an oxygen saturated 0.1 M HClO4 at 60 ◦C usingthree-electrode cell set-up. Scan rate is 50 mV s−1 and values are recorded after0 cycles. TiN or CB loading: 0.1 mg cm−2 (normalized to specific surface area).

agnitude of lower oxidation current densities. This shows thatiN NP electrode attains a passivating surface much earlier unlikearbon, which continues to corrode at the same rate.

An anodic corrosion curve, in Fig. A2 of Appendix, shows thathe starting oxidation potential of TiN NP is higher compared to CB,mplying that carbon tends to oxidize at much earlier potentialshan titanium nitride. While the starting potential for CB decreasesrom 70 to 20 mV making it prone to corrosion at much lower poten-ials, the starting potential of TiN NP remains constant for the entireuration indicating that titanium nitride has a higher resistance toxidation.

Considering that the catalyst support is frequently affected byery high potentials on anode side during extreme conditions ofuel starvation in a PEM fuel cell, a comparison is made betweeniN NP and CB at potentials of 1.3 and 1.5 V and the CV curves arehown in Fig. 4. It can be seen that the oxidation of CB increasesteeply compared to TiN NP and so does the surface carbon oxi-ation at ∼0.55 V (hydoquinone/quinone peak). TiN shows highlyassivating behavior, due to its oxide/oxynitride components andlso, possibly due to the titanium dioxide and hydroxide groupshat can be formed at these potentials according to reactions (4)nd (5). To study the surface changes on the TiN NP due to electro-hemical oxidation, XPS characterization is performed on TiN NPoated electrodes.

.3. Spectroscopic analysis of TiN NP electrodes

XPS is a surface analysis technique which can provide a betternderstanding of the changes on the surface due to electrochemi-al oxidation, along with the atomic concentration of elements onhe TiN support surface. As the catalyst support corrosion in lowemperature fuel cells usually occurs at the surface, XPS can be aseful technique in observing these changes. For XPS characteriza-ion and analysis, the TiN electrodes were exposed to two differenturface conditions, (A) and (B).

A) Untreated: TiN nanoparticles (TiN NP), as isB) Treated: TiN NP coated carbon cloth, 16 h of potential cycling


between 0 and 1.2 V at 50 mV s−1 in 60 ◦C argon saturated 0.1 MHClO4 electrolyte.

The XPS survey spectra of untreated and treated TiN NP elec-rodes exhibited the characteristic Ti 2p, O 1s and N 1s peaks at

Fig. 5. Relative atomic % concentrations of surface elements on TiN electrode:untreated and treated, obtained using XPS (excluding C and F).

the corresponding binding energies of 528.2, 456.5 and 396.2 eV,respectively, in accordance with values reported in literature[28,31,33,35,37,38,46]. Based on the elemental peaks in the surveyspectra and the sensitivity factors of individual elements, the rela-tive atomic concentration on the surface of untreated and treatedTiN electrodes are plotted in Fig. 5.

The initial oxygen composition on the ‘untreated’ TiN NP elec-trode (in Fig. 5) is attributed to the native oxide/oxynitride layer onits surface; other surface groups such as TiN–O*, TiN–OH, adsorbedH2O may also be contributing to oxygen composition, althoughminimally. The increase in the composition of Ti is a result of theelectrochemical oxidation of TiN into Ti ions (through reaction (3)).Similarly, the decrease of N can be attributed to the decrease innitride composition due to its oxidation. From the electrochemicalreactions, the oxygen composition should increase as an outcomeof the oxidation of TiN due to potential cycling. But, as seen in Fig. 5,a slight decrease of oxygen composition is observed for the ‘treated’TiN NP electrode.

The atomic concentration ratios of the elements on ‘treated’ TiNelectrodes, drawn from their respective peak intensities are pre-sented in Table 1. The O/Ti ratio of the TiN electrodes decreasesafter the electrochemical treatment due to the relative increase ofTi while the increase of O/N ratio indicates towards the relativedecrease of N over O. This suggests that the electrochemical oxi-dation does not necessarily replace one-to-one of N with O. Basedon these arguments, the decrease of O/TiN ratios from 0.39 to 0.35,suggests a relative increase in TiN composition after the electro-chemical treatment, the increase attributed to the increase in Ticomposition.

The significant increase in Ti/N ratio suggests that the Ndecreases while Ti increases simultaneously, presumably from thesame compound. Since the oxidation of TiN inevitably results innitrogen as a by-product in the form of N2, NO2, NO, etc. [37] onecan see the decrease in nitrogen composition. The increase of Ti,


treated.

Sample O/Ti O/N O/TiN Ti/N

Untreated 0.60 1.15 0.39 1.91Treated 0.50 1.22 0.35 2.42


ARTICLE ING Model


6 B. Avasarala, P. Haldar / Electrochim

Table 2Assignments to the component peaks of the high-resolution XPS of the ‘untreated’and ‘treated’ TiN NP electrodes.

Peak Component Binding energy (eV) Assignment

Untreated Treated

Ti2p3/2

Ia 455.6 455.6 TiNIIa 456.9 457.2 TiOxNy

IIIa 459.1 459.0 TiO2

A – 460.7 Ti(OH)22+

Ti2p1/2

Ib 461.7 – TiNIIb 463.4 462.1 TiOxNy

IIIb 465.0 464.4 TiO2

B – 466.2 Ti(OH)22+

O1s

IV 530.9 530.5 TiO2

V – 531.2 O–H (Ti(OH)2)2+)VI 532.5 533.2 Adsorbed O2/H2OVII 533.9 534.0 TiOxNy

VIII – 535.5 NO

o1a

4

ehapo

F

N1s

IX 396.2 396.5 O–N396 ev (TiOxNy)X 397.2 397.3 TiNXI 399.1 399.4 O–N399 ev (TiOxNy)

The most probable assignments to the origin of the componentsf high-resolution X-ray photoelectron spectra of Ti 2p, O 1s and Ns are presented in Table 2 along with their peak assignments thatre based on the reported literature [28,31,33,35,37,38,46].

.3.1. Ti 2pThe deconvoluted Ti 2p peaks of ‘untreated’ and ‘treated’ TiN

lectrodes are shown in Fig. 6(a) and (b), respectively, where the


igh-resolution Ti 2p spectra shows the spin orbit doublet Ti 2p3/2nd Ti 2p1/2 peaks. The peak positions and their assignments areresented in Table 2. The presence of titanium nitride (Ia) and thatf oxynitride (IIa) and oxide (IIIa) components of Ti 2p3/2 formed

ig. 6. Deconvoluted Ti 2p XPS spectra of: (a) untreated; (b) treated TiN NP electrode.


due to atmospheric oxidation can be seen in Fig. 6(a) whose bind-ing energies, shown in Table 2, match well with the literature[28,31,35,38,46]. Using XPS technique, which is strictly a surfacecharacterization technique, the oxide peak at 459.1 eV is assignedto TiO2 (IIIa) but a distinct phase of the same is not observed in theXRD plot in Fig. 2.

Comparing the spectra of Ti 2p in Fig. 6(a) and (b), one canobserve the changes in peak areas and heights after the electro-chemical treatment. The binding energies of Ti 2p3/2 and 2p1/2 of‘untreated’ electrode are assigned to TiN (Ia) and TiN (Ib), respec-tively, with 455.6 and 461.7 eV as their respective peak positions.After the electrochemical treatment, both the binding energies of Ti2p3/2 and 2p1/2 shifted to 460.7 eV (A) and 466.2 eV (B), respectively,as shown in Fig. 6(b). These characteristic peaks of ‘A’ and ‘B’ arerelated to the tetravalent titanium ion, Ti(OH)2

2+, where the Ti ionpossibly oxidized to the states of higher valence during potentialcycling following reaction (6) [37]:

TiN + 3H2O = Ti(OH)22+ + NO + 4H+ + 6e− (6)

For Ti in acidic media, the Pourbaix diagram predicts that thethermodynamically stable states of Ti ions under the condition ofanodic polarization have 3+ and 4+ valence, with the dissociatedtitanium, most likely, as Ti3+, Ti(OH)2+ and Ti(OH)2

2+ [36,37]. Con-sidering the XPS peak positions of ‘A’ and ‘B’ of Ti 2p in Fig. 6(b) andthe tetravalence of Ti ion according to the Pourbaix diagram of Ti inacidic media, it is reasoned that the Ti ion could be taking the formof Ti(OH)2

2+ with a 4+ oxidation state, after the electrochemicaltreatment.

The areas under the deconvoluted peaks of Ti 2p3/2 are pre-sented in Table 3 where one can see a significant presence of oxideand oxynitride components on the surface of the untreated TiNelectrode, suggesting that the process of atmospheric oxidationhas been initiated to certain extent, supporting the earlier state-ments. Due to their inert nature, the existing oxide/oxynitride layerpresent on the surface of TiN NP does not oxidize any further. Itshould be noted that the overlayer was never transformed to apure titanium oxide even after 2 years of air exposure, but ratherremained a mixture of oxynitride and oxide. This implies that theoxidation mainly occurs on the nitride part of the surface on TiNnanoparticles resulting in the decrease of its composition, with acorresponding increase in the oxide and oxynitride components[28] by the same magnitude.

In Table 3, it can be seen that the nitride component, TiN (Ia),decreases by 17.2% after the electrochemical treatment, with a10.4% increase in the relative composition of TiOxNy. This increaseagrees well with the reported literature [28], where the elec-trochemical oxidation of TiN forms TiOxNy compound on thesurface. The remaining percentage of decrease in nitride should beattributed to the increase of oxide composition. But the increase inthe oxynitride composition is not accompanied by a correspond-


ing increase in the oxide component. In fact, the oxide component,as shown in Table 3, significantly decreases by a relative compo-sition of 46.7%. The area ratio of nitride (Ia) and oxide (IIIa) peaks,(ATiN/ATiO2), in Ti 2p3/2 (Fig. 6) increases from 0.51 to 0.95 sug-gesting the significant decrease of oxide after the electrochemical

Table 3Approximate functional group distribution of deconvoluted Ti 2p3/2 for “untreated”and “treated” TiN electrode (relative composition, %).

Functional groups Untreated (1) Treated (2) Change in relativecomposition %,(1)–(2)

TiN 31.5 14.3 17.2↓TiOxNy 6.8 17.2 10.4↑TiO2 61.5 14.8 46.7↓Ti(OH)2

2+ – 53.7 –


IN PRESSG

E

rochimica Acta xxx (2010) xxx–xxx 7

tmaTitoooefte

4

‘ti[egoetoatdc(Nib

4

stabTaafi[

(

dtXd

mdoo(tnt

w

ARTICLEModel



reatment. The oxide on TiN NP surface dissolves in the acidicedia exposing the nitride surface which further reacts with the

cidic media to form Ti–OH compound according to reaction (8).he above proposed mechanism explains the significant increasen Ti(OH)2

2+, the magnitude (53.7%) of which can be equaled tohe dissolution of oxide component (↓46.7%) and the dissolutionf oxide (↓6.8%) which is newly formed due to the electrochemicalxidation of nitride component. Although an elaborate discussionn this topic is provided later in Section 4.4, it is emphasized thatlectrochemical treatment under cycling potentials will lead to theormation of oxide/oxynitride layer [28]. But, unlike in buffer solu-ions (pH ∼ 7), perchloric acid solution tends to dissolve the layerxposing the nitride surface underneath.

.3.2. O 1sThe deconvoluted O 1s XPS spectra for the ‘untreated’ and

treated’ TiN electrode are shown in Fig. 7(a) and (b), respec-ively. The dominant O 1s component (IV) for ‘untreated’ electroden Fig. 7(a) is located at 530.2 eV and is assigned to TiO228,31,33,35,38,46,47] but the dominant O 1s peak for the ‘treated’lectrode has shifted to 531.2 eV, which is characteristic to O–Hroup [28,35,37,46,48]. The significant increase of O–H peak (V)f O 1s in Fig. 7, after the electrochemical treatment, supports thearlier argument about the formation of Ti–OH groups indicated byhe peaks ‘A’ and ‘B’ of Ti 2p in Fig. 6(b) and the Pourbaix diagramf Ti in acidic media. The remaining components of the O 1s peaksre labeled in Table 3 and illustrated in Fig. 7. The dissolution ofitanium oxide in acidic media can also be seen from the significantrop of the oxide peak (IV) at 530.9 eV. Similar to the oxynitrideomponent in Ti 2p and N 1s spectra (Fig. A3, Appendix), the TiOxNy

VII) peak in O 1s is seen to increase for the ‘treated’ TiN electrode.O (VIII) is a possible by-product formed during the electrochem-

cal oxidation of TiN [36] along with NO2 and N2 and its peak cane seen near 535.0 eV [49].

.4. Dissolution of oxide/oxynitride layer

It is well known [50,51] that the oxides of titanium, due to theirtrong basic character, dissolve in acidic media to form colorlessitanium (IV or 4+) oxysalts that decompose in water. In perchloriccid media such as the one used in our case, Ti (IV) species cane present as (Ti(OH)2)2+ or TiO2

+ that is derived from the salts ofiOCl42− anions [51]. To confirm the oxide dissolution in perchloriccid, a few ml of H2O2 was added TiN NP dispersed in 0.1 M HClO4nd the solution turned to intense orange color. This is due to theact that colorless hydrated Ti(IV) ions in the solution get convertedn to orange hydrated (Ti(O2)OH)+ ions according to reaction (7)50] thus confirming the presence of dissolved Ti(IV) species:

Ti(OH)3)+ + H2O2 → (Ti(O2)OH)+ + 2H2O (7)

The immediate change in color of the solution suggests the rapidissolution of titanium oxides in the acidic media. Further evidenceo the dissolution of oxide component is shown in the analysis ofPS spectra of Ti 2p and O 1s in Figs. 6 and 7, where one can see theecrease of oxide composition.

Unlike the oxides which were quick to dissolve in the acidicedia, it can be reasoned that the oxynitride takes a longer time to

issolve due to its nitrogen composition. Fig. 8 shows the changes ofpen circuit potential (OCP) of TiN NP with immersion time at vari-us temperatures of perchloric acid. OCP or the corrosion potentialEcorr) measures the equilibrium potential above which the oxida-


ion is initiated on the surface. A stable OCP indicates the stableature of the electrode while the surface reactions cause the OCPo fluctuate.

The OCP (at 60, 70, 80 ◦C) of TiN NP in Fig. 8 shifted rapidly down-ards during the initial stage of immersion. The downward shift of

Fig. 7. Deconvoluted O 1s XPS spectra of: (a) untreated; (b) treated TiN NP electrode.

OCP towards the active potentials is due to the self-activation ofTiN and is an indication of a dissolution of the oxynitride on thesurface of TiN [37,52], which increasingly exposes the nitride sur-face directly to acidic solution. The anodic chemical reaction thatcan be related to TiN at these OCP conditions can be expressed as[37]:


TiN + H2O → Ti(OH)2+ + 1/2N2 + H+ + 3e− (8)

The potential-pH (or Pourbaix) diagram of Ti-H2O predicts thatthe possible thermodynamically stable state of Ti ions under theOCP conditions can be 3+ valence; and the reaction (8) indicates that


ARTICLE ING Model


8 B. Avasarala, P. Haldar / Electrochim

Fi

ttfiidii

sustaCuaiOsottlbsN

2lwpIao

dattdsi

ig. 8. Shifts in open circuit potential (OCP) with the immersion time (4 h) for TiNn 0.1 M HClO4 solution saturated with argon.

his dissociated titanium, Ti(III), is most likely Ti(OH)2+ [36,37]. Ashe temperature of the perchloric acid solution in Fig. 8 increasesrom 60 to 80 ◦C, the hydrogen ion strength in the acid solutionncreases thus causing a faster dissolution of oxynitride layer dur-ng the initial stage of immersion. The steeper slope and a lowerissolution time with an increase in temperature in Fig. 8 are an

ndication of faster rate of oxynitride dissolution due to increasedonic strength of the acid solution.

The rapid decrease of OCP, for the curves in Fig. 8, in the initialtage of immersion is related to the formation of the corrosion prod-ct of Ti(OH)2+ (according to reaction (8)) on the exposed nitrideurface, with its nature of positive charge [37]. The formation ofhese corrosion products on the TiN surface leads to a subsequentdsorption of other negative ions existing in the solution such aslO4

−. As a result, an adsorption configuration is likely to be set-p where the surface area of TiN NP will be covered with suchnions, naturally prohibiting the dissolution of TiN and thus mak-ng the electrode passive and attain more noble potentials. TheCP curves of 60, 70 and 80◦C in Fig. 8 drop steeply in the initial

tages of immersion after which it gradually plateaus for the restf the immersion, in the passive region of the plot. This increaseowards nobler potentials is likely the result of subsequent adsorp-ion of anions in the solution by the positively charged Ti(OH)2+

ayer on the nanoparticles surface. This adsorption model, proposedy Chyou et al. [37] for TiN thin films in varying concentrations ofulfuric acid solutions, explains the electrochemical behavior of TiNP under 0.1 M perchloric acid at different temperatures.

The minimal decline of OCP curve at room temperature (RT,3 ◦C) in Fig. 8 indicates that the rate of dissolution of the oxynitride

ayer on the surface of TiN NP is very low at the given conditions;hich imply that TiN NP will take a significantly longer time toassivate under the RT conditions of 0.1 M HClO4 acid electrolyte.

t should be noted that the variation in the starting OCP potentialt various temperatures is most likely due to the varying thicknessf the oxynitride layer on the surface of TiN NP.

Although the OCP plot shows the decrease of oxynitride layerue to its dissolution, Table 3 and the Ti 2p and O 1s spectra shown increase of TiOxNy composition for the ‘treated’ TiN NP elec-


rode. Unlike potential cycling, no external potential is imposed onhe TiN NP electrode surface during the OCP conditions. Hence, theissolution of TiOxNy in the initial stage and a passivation of nitrideurface in the later stage are observed. Potential cycling involvesmposing oxidizing potentials on the TiN NP electrode surface and


leads to the formation of oxide and oxynitride components [26].Thus the oxide/oxynitride components on the surface of TiN NPundergo dissolution in the acidic media and also formation due tooxidizing potentials during the potential cycling treatment. As aresult of the two parallel phenomena, one can see a net increase ofTiOxNy component on TiN nanoparticles.

4.5. Electrical conductivity of oxide/oxynitride components onTiN NP surface

An aspect of concern for TiN nanoparticles is the change inits electrical conductivity due to the presence of the oxide andoxynitride components on the surface. Since the surface is notpredominated with TiO2, the oxide component largely exists asa sub-stoichiometric composition (TinO2n−1) of Magneli phase,which has electrically conductive properties [24]. With an increasein surface oxygen composition due to electrochemical oxidationat high potentials, the surface layers can become titanium dioxidebut get easily dissolved in the acidic conditions of fuel cell withoutacting as an insulator.

The oxynitride component of the native layer, TiOxNy, has arelatively slower rate of dissolution compared to oxide and con-sists of lower percentage of oxygen composition than the oxidecomponent. TiOxNy has an electrically conductive nature due to itsNaCl crystal structure which is similar to TiN and also a significantpresence of nitrogen in its lattice [34]. It has been reported [53]that titanium oxynitride thin films with low oxygen contents (<40at.%) are partly metallic suggesting that the oxynitride layer on thesurface of TiN nanoparticles may have relatively higher electricalconductivity. To further support the statement, the elemental com-position of oxygen, from oxide and oxynitride components, on the‘untreated’ TiN nanoparticles surface was measured by XPS and wasfound to be ∼35% atomic concentration. This means that the oxyni-tride component will have a much lower composition of oxygenthan the total value. It is reasonable to speculate that this con-ductive nature of oxynitride component on the surface of TiN NPresulted in the minimal ohmic resistance (or poor retardation ofkinetics) from the catalyst supports during the initial cycles of CVand ORR tests of Pt/TiN presented in our previous report [18].

4.6. Stability of Pt/TiN electrocatalyst

As the temperature affects the active or passive behavior of theTiN NP, the durability of the Pt/TiN electrocatalyst (synthesizedusing TiN NP) will also be invariably affected in a similar man-ner. The synthesized Pt/TiN electrocatalyst has an average platinumparticle size of 2.4 nm as measured by TEM (Fig. 9). To test the dura-bility of the electrocatalyst, Pt/TiN was subjected to accelerateddurability test (ADT) where the electrocatalyst coated electrodeundergoes repeated potential cycling for a prolonged period of timeunder simulated PEM fuel cell conditions. The durability perfor-mance in this case was monitored by the no. of active Pt sites on theelectrocatalyst i.e., the electrochemical surface area (ECSA), whichwas measured at intervals of 100 cycles; the plot of ECSA of Pt/TiNvs. no. of cycles at different temperatures is shown in Fig. 10.

The ECSA of Pt/TiN at 60 ◦C in Fig. 10 has an initial high valueof 70 m2 g−1

Pt but falls to near zero within few hours of operationand continues to plateau near zero for the remaining duration. Thistrend can be explained by the OCP curve of TiN NP at 60 ◦C (in Fig. 8)where the dissolution of oxynitride layer takes place in the initialstages and later passivation occurs due to the formation of Ti(OH)


layer on its surface. XPS characterization also revealed a similarhydroxide layer on the surface of TiN NP after repeated potentialcycling, which is similar to ADT. This hydroxide layer prevents theTiN NP supports (of Pt/TiN electrocatalyst) to conduct electronsfrom Pt catalyst particles resulting in near zero ECSA as shown


Please cite this article in press as: B. Avasarala, P. Haldar, Electrochemical oxidation behavior of titanium nitride based electrocatalysts underPEM fuel cell conditions, Electrochim. Acta (2010), doi:10.1016/j.electacta.2010.08.035



B. Avasarala, P. Haldar / Electrochimica Acta xxx (2010) xxx–xxx 9

Fig. 9. (a and b) Transmission electron micrograph showing the Pt particles on TiN catalyst supports; (c) histogram showing the Pt nanoparticles distribution with meanparticle size at 2.4 nm.

Fig. 10. Decrease of ECSA of 20 wt% Pt/TiN after potential cycling between 0 and1.3 V (50 mV s−1) in N2 saturated 0.1 M HClO4 for 1200 cycles or 16 h. Pt loading:20 �g cm−2.

Fig. 11. CV curves of 20 wt% Pt/TiN showing the 100 and 1200 cycles during thepotential cycling between 0 and 1.3 V (50 mV s−1) in N2 saturated 0.1 M HClO4 at70 ◦C. Pt loading: 20 �g cm−2.




10 B. Avasarala, P. Haldar / Electrochimica Acta xxx (2010) xxx–xxx

rious

itbih

isa(Pu

hdPcacdccbp

iticpaods

satsicdcd

t

solve in acidic solution depending on the media and temperature.The exposed nitride surface reacts with the solution and using XPS

Fig. 12. Scheme showing the surface condition of TiN nanoparticles at va

n Fig. 10. Passivation of TiN surface due to hydroxide layer andhe resulting poor conductivity of TiN particles has been reportedy Qiu and Gao [21] who measured that the electrical conductiv-

ty of TiN powder decreases by an order of magnitude due to theydrolyzed layer on the surface.

As mentioned earlier, the OCP curves in Fig. 8 show that anncrease in temperature leads to faster passivation of the TiN NPurface. Hence, one can see a faster rate of decline of ECSA for testt 70 ◦C and significantly lower rate of decrease of ECSA at 23 ◦CRT) in Fig. 10. This shows that the decreasing trends of ECSA oft/TiN can be explained by the adsorption model of TiN describedsing OCP curves.

It may be argued that the decline of ECSA for the Pt/TiN atigher temperatures (60 and 70 ◦C) of acidic media may also beue to the complete loss of Pt catalyst and/or a sharp increase int particle size due to agglomeration. But an increase in Pt parti-le size does not amount to a corresponding rapid decline of ECSAs seen in Fig. 10 and the presence of Pt (on the Pt/TiN coatedarbon cloth electrode) was still seen after the ADT using an energy-ispersive spectrometer. Also, an increased Pt particle shows lowerurrent densities but retains the typical shape of the platinum CVurve unlike the ‘cyc 1100’ curve in Fig. 11 that shows negligi-le (hydrogen adsorption–desorption) HAD peaks with a completeassivation suggesting the poor conduction of support.

It can be observed that the passivation of TiN surface (plateau-ng of OCP in Fig. 8 at 60 ◦C) is attained in ∼1 h whereashe passivation of TiN surface of Pt/TiN (plateauing of ECSAn Fig. 10 at 60 ◦C) is attained after ∼4 h (or 300 potentialycles). The longer time taken by the TiN NP to passivate underotential cycling conditions, compared to the OCP conditions,grees well with the earlier proposed mechanism of simultane-us dissolution and formation of oxide/oxynitride componentsuring potential cycling. The illustration of the process can beeen in Fig. 12.

Though the above results of Pt/TiN electrocatalyst may seem touggest only the strong dependence of temperature on its stability,closer analysis indicates that factors such as concentration and

ype of acid media can also influence the TiN surface. It is empha-ized that the passivation behavior of TiN NP shown in this studys specific to the temperature and concentration conditions of per-hloric acid electrolyte. By changing the operating window to aifferent set of parameters of—acidic media, temperature and con-


entration, crystallinity of TiN powder [54], an active phase or aelayed passivation of TiN nanoparticles can be observed.

The passivation behavior of TiN NP under the prolonged elec-rochemical conditions may not be a limitation for its application

stages during the exposure to PEM fuel cell conditions (acid electrolyte).

as a catalyst support in PEM fuel cells. Unlike carbon black, TiNnanoparticles do not predominantly oxidize to gaseous oxides thatcan lead to a significant mass loss but forms a passivating surfacethat prevents further oxidation. If the chemical stability of the TiNnanoparticles towards water in acidic media can be modified toreduce surface hydrolysis, TiN NP can become a durable catalystsupport that can increase the lifetime of PEM fuel cells. Zhang andBinner [55] proposed a coating of stearic acid on nitride powderas a stabilizing agent against hydrolysis while Müller et al. [56]described the protection of nanoscale non-oxidic particles fromoxygen uptake by coating with N2-containing surfactants. Thougha carbon-based coating can provide a hydrophobic surface, there isa possibility that it may undergo electrochemical oxidation similarto carbon supports under PEM fuel cell conditions. Alternatively,identifying an active phase of TiN, in a set of conditions that fall inthe operating window of a PEM fuel cell, can result in an increaseddurability of the Pt/TiN electrocatalyst.

Though carbon black supports cost less than TiN NP, an increaseddurability obtained from the novel TiN NP supports can reduce thelifetime costs of a PEM fuel cell system by enabling the system tofunction till its intended life span. Further research is being pursuedto effectively use the oxidation resistance properties of titaniumnitride nanoparticles for increasing the durability of electrocata-lysts in PEM fuel cells. The research also opens different avenueswhere TiN supports can be used in combination with other catalystsor mixed with other conventional supports. But it is emphasizedthat attention must be paid to the stability of the nitride supportover prolonged exposure under PEM fuel cell conditions as it pro-vides the true understanding of the durability of the electrocatalyst.

5. Conclusion

The oxidation currents of TiN are significantly lower comparedto that of CB at operating potentials 1.0 and 1.2 V of a PEM fuelcell thus indicating its corrosion resistance properties. XRD and XPScharacterization of TiN NP revealed an existing oxide and oxynitridelayer on the surface, which prevents further diffusion of oxygen intothe bulk TiN. It is reported that the nitride component decreases andoxide/oxynitride components increase on the TiN surface duringthe potential cycling but the oxide/oxynitride layer tends to dis-


and OCP data, it is shown to form Ti(IV) hydroxide ions that pas-sivates the surface by adsorbing oppositely charged anions in thesolution. The durability of Pt/TiN electrocatalyst also agrees wellwith the adsorption model of TiN NP.


ING

E

rochim

A

tDo

A

t

R

[[

[[

[

[

[[

[[

[

[

[[[

[[[

[

[[[[

[

[[

[

[[[[

[

[

[

[[

[

[[[[

[

[[

ARTICLEModel



cknowledgements

The authors gratefully acknowledge Richard Moore for his assis-ance with XPS characterization of the TiN electrode samples andr. Thomas Murray for transmission electron microscopy imagesf TiN and Pt/TiN powders.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.electacta.2010.08.035.

eferences

[1] R. Borup, et al., Chem. Rev. 107 (2007) 3904–3951.[2] S. Satyapal, Hydrogen program overview, 2009 US DOE hydrogen program and

vehicle technologies program, in: Annual Merit Review and Peer EvaluationMeeting, Virginia (US), May 18, 2009.

[3] G. Hinds, Performance and durability of PEM fuel cells: a review, NPL ReportDEPC-MPE 002, September, 2004.

[4] T. Tada, T. Kikinzoku, in: H. Vielstich, A. Gasteiger, Lamm (Eds.), Handbookof Fuel Cells—Fundamentals, Technology and Applications, vol. 3, Wiley, NewYork, 2003, p. 481.

[5] G. Gasteiger, W. Gu, R. Makharia, M. Mathias, B. Sompalli, in: W. Vielstich, H.Gasteiger, A. Lamm (Eds.), Handbook of Fuel Cells—Fundamentals, Technologyand Applications, vol. 3, Wiley, New York, 2003, p. 593.

[6] Z. Siroma, K. Ishii, K. Yasuda, Y. Miyazaki, M. Inaba, A. Tasaka, Electrochem.Commun. 7 (2005) 1153–1156.

[7] L. Li, Y.C. Xing, J. Electrochem. Soc. 153 (2006) A1823–A1828.[8] Y.Y. Shao, G.P. Yin, Y.Z. Gao, P.F. Shi, J. Electrochem. Soc. 153 (2006)

A1093–A1097.[9] B. Avasarala, R. Moore, P. Haldar, Electrochim. Acta (2010),

doi:10.1016/j.electacta.2010.03.056.10] Y Shao, G. Yin, Y. Gao, J. Power Sources 171 (2007) 558–566.11] K. Kinoshita, Carbon: Electrochemical and Physicochemical Properties, first ed.,

Wiley, New York, 1988.12] J. Willsau, J. Heitbaum, J. Electroanal. Chem. 161 (1984) 93–101.13] K.H. Kangasniemi, D.A. Condit, T.D. Jarvic, J. Electrochem. Soc. 151 (2004)

E125–E132.14] S.D. Knights, K.M. Colbow, J. St-Pierre, D.P. Wilkinson, J. Power Sources 127

(2004) 127–134.15] L.M. Roen, C.H. Paik, T.D. Jarvic, Electrochem. Solid-State Lett. 7 (2004) A19–

A22.16] Y.Y. Shao, G.P. Yin, J. Zhang, Y.Z. Gao, Electrochim. Acta 51 (2006) 5853–5857.17] F. Coloma, A. Sepulvedaescribano, F. Rodriguezreinoso, J. Catal. 154 (1995)

299–305.


18] B. Avasarala, T. Murray, W. Li, P. Haldar, J. Mater. Chem. 19 (2009) 1803–1805.19] S.T. Oyama, The Chemistry of Transition Metal Carbides and Nitrides, first ed.,

Springer, 1996.20] A.D. Verkhoturov, M.A. Kuzenkova, N.V. Lebukhova, I.A. Podchernyaeva, Soviet

Powder Metall. Met. Ceram. 27 (1988) 162–166.21] Y. Qui, L. Gao, J. Phys. Chem. B 109 (2005) 19732–19740.

[

[[

PRESSica Acta xxx (2010) xxx–xxx 11

22] D.J. Ham, J.S. Lee, Energies 2 (2009) 873–899.23] O.T. Musthafa, S. Sampath, Chem. Commun. 1 (2008) 67–69.24] T. Ioroi, Z. Siroma, N. Fujiwara, S. Yamazaki, K. Yasuda, Electrochem. Commun.

7 (2005) 183–188.25] K.-W. Park, K.-S. Seol, Electrochemistry Communications 9 (2007) 2256–2260.26] G.V. Samsonov, I.F. Pryadko, L.F. Pryadko, Moskva Nauka (1976) 339–401.27] A.I. Ivanovskii, V.A. Gubanov, O.P. Schveikin, E.Z. Kurmaev, J. Less-Common Met.

78 (1981) 1–19.28] I. Milosev, H.H. Strehblow, B. Navinsek, M. Metikos-Hukovic, Surf. Interface

Anal. 23 (1995) 529–539.29] R. Kroger, M. Eizenberg, C. Marcadal, L. Chen, J. Appl. Phys. 91 (2002) 5149–5154.30] V. Lingwal, N.S. Panwar, J. Appl. Phys. 97 (2005) 104902–104910.31] N.C. Saha, H.G. Tompkins, J. Appl. Phys. 72 (1992) 3072–3079.32] Y. Massiani, A. Medjahed, P. Gravier, L. Argeme, L. Fedrizzi, Thin Solid Films 191

(1990) 305–316.33] C. Ernsberger, J. Nickerson, T. Smith, A.E. Miller, D. Banks, J. Vac. Sci. Technol. A

4 (1986) 2784–2788.34] G. Hett, M. Green, I. Parkin, J. Am. Chem. Soc. 129 (2007) 15541–15548.35] F. Esaka, K. Furuya, H. Shimada, M. Imanura, N. Matsubayashi, H. Sato, A. Nishi-

jima, A. Kawana, H. Ichimura, T. Kikuchi, J. Vac. Sci. Technol. A 15 (1997)2521–2528.

36] J.C. Francois, Y. Massiani, P. Gravier, J. Grimblot, L. Gengembre, Thin Solid Films223 (1993) 223–229.

37] S.D. Chyou, H.C. Shih, T.T. Chen, Corros. Sci. 35 (1993) 337–347.38] A. Vesel, M. Mozetic, J. Kovac, A. Zalar, Appl. Surf. Sci. 253 (2006) 2941–2946.39] G.N. Komratov, Powder Metall. Met. Ceram. 36 (1997) 9–10.40] R. Juza, in: H.J. Emeléus, A.G. Sharpe (Eds.), Advances in Inorganic Chemistry

and Radiochemistry, vol. 9, Academic Press, 1966, p. 81.41] N.D. Tomashov, T.V. Chukalovskaya, I.L. Medova, F. Egorov, Prot. Met. 21 (1985)

682–688.42] Nanostructured and Amorphous Materials, Inc.: Titanium Nitride

http://www.nanoamor.com/inc/sdetail/3691 (accessed May 2, 2010).43] C.H De Novion, B. Beuneu, T. Priem, N. Lorenzelli, A. Finel, in: R. Freer (Ed.),

The Physics and Chemistry of Carbides, Nitrides, and Borides (NATO ASI Series.Series. E, No. 185), Kluwer Academic Publishers, Dordrecht Amsterdam, 1990,p. 329.

44] X. Yang, C. Li, B. Yang, W. Wang, Y. Qian, Chem. Phys. Lett. 383 (2004) 502–506.45] Powder Diffraction File, Card Nos. 38 – 1420 (TiN) and 77 – 2170 (TiO), Joint

Committee on Powder Diffraction Standards (JCPDS), Swarthmore, PA, 1955.46] Y.L. Jeyachandran, K. Narayandass, D. Mangalaraj, S. Areva, A. Mielczarski,

Mater. Sci. Eng. A 445–446 (2007) 223–236.47] H.C. Mana, Z.D. Cuia, X.J. Yang, Appl. Surf. Sci. 199 (2002) 293–302.48] P. Prieto, R.E. Kirby, J. Vac. Sci. Technol. A 13 (1995) 2819–2826.49] M.H. Matloob, M.W. Roberts, J. Chem. Soc. Faraday Trans. 1 (1977) 1393–1405.50] N. Wiberg, A.F. Holleman, E. Wiberg (Eds.), Holleman-Wiberg’s Inorganic Chem-

istry, thirty-fourth ed., Academic Press, New York, 2001.51] R. Sarkar, General & Inorganic Chemistry (Part II), first ed., Books and Allied (P)

Ltd, Kolkata, 2001.52] J. Krysa, R. Mraz, I. Rousar, Mater. Chem. Phys. 48 (1997) 64–67.53] J.–M. Chappeı, N. Martin, J. Lintymer, F. Sthal, G. Terwagne, J. Takadoum, Appl.


Surf. Sci. 253 (2007) 5312–5316.54] A. Delblanc Bauer, M. Herranen, H. Ljungcrantz, J.O. Carlsson, J.E. Sundgren,

Surf. Coat. Technol. 91 (1997) 208–214.55] Y. Zhang, G.P. Binner, Int. J. Inorg. Mater. 1 (1999) 219–227.56] R. Muller, M. Knapp, K. Heckmann, M. Ruthendorf, G. Boden, Langmuir 20 (2004)

2598–2606.


http://dx.doi.org/10.1016/j.electacta.2010.08.035

http://www.nanoamor.com/inc/sdetail/3691

Documents

Electrochemical oxidation behavior of titanium nitride based