Nano Energy 26 (2016) 267–275

Contents lists available at ScienceDirect

Nano Energy

http://d2211-28

n CorrE-m

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

Highly active and porous graphene encapsulating carbon nanotubes asa non-precious oxygen reduction electrocatalyst for hydrogen-air fuelcells

Pouyan Zamani, Drew C. Higgins, Fathy M. Hassan, Xiaogang Fu, Ja-Yeon Choi,Md. Ariful Hoque, Gaopeng Jiang, Zhongwei Chen n

Department of Chemical Engineering, University of Waterloo, 200 University Ave. W., Waterloo, Ontario, Canada N2L 3G1

a r t i c l e i n f o

Article history:Received 11 January 2016Received in revised form12 May 2016Accepted 20 May 2016Available online 20 May 2016

Keywords:Oxygen reductionNon-precious catalystFuel cellsPolyanilineCarbon nanotubeIn-situ graphene

x.doi.org/10.1016/j.nanoen.2016.05.03555/& 2016 Elsevier Ltd. All rights reserved.

esponding author.ail address: zhwchen@uwaterloo.ca (Z. Chen).

a b s t r a c t

Heat treated iron-polyaniline-carbon – based non-precious metal catalysts represent a promising class ofmaterial to replace the platinum based ORR catalysts for PEMFC technologies. In the present research, weapply an ammonia treatment to tune the structure and activity of electrocatalysts derived from iron,polyaniline and carbon nanotubes (CNTs). By controlling the NH3 reaction conditions, we were able totune the chemistry of nitrogen incorporation, including concentration and dopant type. The final catalysthad a robust morphology consisting of highly porous 2-D in-situ formed graphene-like structures that,along with the intermixed 1-D CNTs, were decorated with an abundance of nitrogen and iron species.The resultant surface chemistry led to impressive catalyst activity, with a half-wave potential of 0.81 Vobserved through half-cell testing and under H2-air fuel cell testing, a current density of 77 mA cm�2 at0.8 V was achieved, along with a maximum power density of 335 mW cm�2.

& 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Polymer electrolyte membrane fuel cells (PEMFCs) are highlypromising clean energy devices considered as ideal alternatives tothe conventional fossil fuel based technologies used in the auto-motive industry, telecommunications backup and materialshandling [1–3]. Although the target markets exist, technical chal-lenges relating to cost and durability must be addressed [4]. Themajor bottle-neck issues arise from the cathode where the oxygenreduction reaction (ORR) happens [5–10]. Currently the onlytechnologically viable ORR catalysts are platinum-based. This iswhy developing new catalyst materials with higher performance,as well as providing cost drops is of great interest.

Non-precious metal catalysts (NPMCs) are attractive classes ofmaterials to replace the platinum in conventional ORR catalysts.NPMCs are actively under development [6,7,11] and prepared viaseveral synthetic approaches such as applying metal-organic fra-mework precursors [12–14], the sacrificial support method [15–18], hydrothermal or solvothermal synthesis [19,20], pyrolyzing acarbon supported complex [21–25], and polymerization of nitro-gen containing monomers [26–33]. These approaches, using a hightemperature pyrolysis in the presence of iron and/or cobalt

precursors, provide transition metal-nitrogen-carbon complexes(M-N-C) which are so far the most promising class of NPMCs.

In the present study, we apply and optimize an ammoniatreatment to tune the chemistry of a catalyst derived from iron,PANI and carbon nanotubes (CNT). This approach results in im-proved performance, both in half cell and full cell conditions. NH3

contributes by enriching the N-dopant concentration, and bycontrolling the temperature we can tailor the particular identity ofthe nitrogen dopants to maximize catalytic activity. The overallsynthesis process ultimately leads to in-situ graphitization of car-bon, forming catalyst structures that consist of multilayer highlyporous graphene morphologies encapsulating CNTs with anabundance of nitrogen and iron defects. This surface chemistryrenders the catalyst highly active towards the ORR, coupled withthe robustness of the graphene-like morphology.

2. Materials and methods

2.1. Functionalized multi-walled carbon nanotubes

Commercial carbon nanotubes with 30–50 nm diameters and10–20 mm lengths were functionalized first by immersing 10.0 g ofthem in 400 ml of 70% nitric acid. The mixture was then refluxedat 85 °C for 8 h under vigorous agitation. Functionalized CNT(FCNT) were then filtered and washed with DDI water and dried in

P. Zamani et al. / Nano Energy 26 (2016) 267–275268

an oven overnight.

2.2. FeCl3-PANI-FCNT polymerization

3 ml of aniline and 10 g FeCl3 was added into 300 ml HCl atroom temperature while continuously stirring throughout theentire process. Slowly, 5.0 g of (NH4)2 S2O8 (ammonium perox-ydisulfate, APS) was added as an oxidant to polymerize the anilinein the above solution. The mixture was rigorously stirred at roomtemperature for 3 h to allow fully polymerization of the aniline.After 3 h, a suspension of 400 mg FCNTs in DDI water was slowlypipetted into the polyaniline mixture. The final solution was rig-orously stirred for 48 h. After this, the liquid was evaporated andthe solid polymer (FeCl3-PANI-FCNT) was collected.

2.3. Catalyst synthesis

The catalyst precursors were first heat treated at 200 °C for 1 hunder 70 ml min�1 of argon (Ar), using a 30 °C min�1 heating rate.The subsequent heat treatment for the resulting powder was doneat 900 °C for 1 h using the same heating rate and Ar flow rate. Thesolid samples were then leached in 0.5 M sulfuric acid at 80–90 °Cfor 8 h to remove the inactive iron complexes as well as to in-troduce a porous morphology. As it can be seen from the no-menclature outlined in Table S1, after this leaching process, somecatalysts were one step heat treated in ammonia at temperaturesranging from 900 °C to 1000 °C. Others were pyrolyzed at 900 °Cin Ar for 3 h (Fe-P-C_Ar) to further improve the ORR activity and toremove surface functional groups. This is referred to as the con-ventional second heat treatment. After this step, ammonia treat-ment was done at temperature ranges from 800 °C to 1050 °C for15 min and under a 300 ml min�1 ammonia flow rate. Thesesamples are designated Fe-P-C_Ar-NHxxx in which the “xxx” isreplaced with the NH3 temperature (Table S1). A schematic of thecatalyst synthesis is displayed in Fig. 1.

Fig. 1. Schematic of the catalyst synthesis procedure, (a) starting with CNT (b) after functP-C_Ar, and (e) NH3-treated Fe-P-C_Ar-NH900.

2.4. Physical characterization

Transmission electron microscopy (TEM) and scanning electronmicroscopy (SEM) were utilized to investigate the nanostructuredmorphology of the fabricated electrocatalyst materials. The surfacearea of a material is very important in the development of newcatalysts as a higher surface area usually suggests more exposedcatalytic active sites to oxygen in ORR. Brunauer–Emmett–Teller(BET) was utilized to measure the surface area and the pore size ofthe catalyst materials. Raman spectroscopy was applied to studyactive sites in catalysts, as well as examine the effect of heattreatments on the material structures. The surface composition ofcatalysts was analyzed using X-ray photoelectron spectroscopy(XPS). Moreover, energy-dispersive X-ray spectroscopy (EDX) andelectron energy-loss spectroscopy (EELS) were used to analyze theelements and their distribution in the catalyst materials.

2.5. Electrochemical characterizations

Half-cell rotating disk electrode (RDE) testing, a well-estab-lished method to evaluate the electrochemical performance ofcatalyst materials, was used to analyze the ORR taking place on thecatalyst materials, whereby higher onset and half-wave potentialsand increased current densities showed better activity towardORR. In order to simulate the acidic circumstances encounteredduring PEMFC operation at the cathode, a 0.5 M H2SO4 solutionwas used as the electrolyte and saturated with oxygen during ORRtesting. The working electrode was a glassy carbon disk(0.19635 cm2) which was coated uniformly by the catalyst ink toachieve a catalyst loading of 0.6 mg cm�2. A graphite and Ag/AgClelectrode were utilized as the counter and reference electrodes,respectively. All potentials were converted to the reversible hy-drogen electrode (RHE) scale for ease of analysis. In order to ac-tivate the catalyst, cyclic voltammetry (CV) was done under oxy-gen saturated electrolyte and scan rate of 50 mV s�1. During ORRtesting, the potential of the working electrode in oxygen saturatedelectrolyte was changed from ca. 1.0–0.0 V vs. RHE using linearstaircase voltammetry (LSCV) with 30 mV amplitude and 30 s

ionalization to FCNT, (c) polymerized FeCl3-PANI-FCNT composite, (d) pyrolyzed Fe-

P. Zamani et al. / Nano Energy 26 (2016) 267–275 269

period. For the half-cell durability test, the potential of the work-ing electrode was cycled between 1.0 and 0.2 V vs. RHE at50 mV s�1 in N2 saturated electrolyte. After different cycle num-bers, ORR evaluation was performed with the conditions men-tioned above. Rotating ring disk electrode (RRDE) testing was alsoused to measure selectivity of the catalysts towards the four-electron reduction of oxygen. The ring potential was set to 1.2 V vs.RHE to successfully oxidize any peroxide species reaching the ringelectrode surface. For performance under PEMFC operating con-ditions, catalysts were tested at the fuel cell cathode, with acommercial Pt cloth gas-diffusion layer (0.5 mg cm�2 Pt) beingused at the anode. Cathode ink was prepared using a mixture of40 mg catalyst, 480 mg isopropanol, 480 mg DDI water, and440 mg commercial Nafion solution (5 wt%). This way, the overallNafion content in the dry catalyst ink was hold at 35 wt%. A driedNafion 211 membrane (which was previously immersed in 0.5 Mboiling sulfuric acid for 1 h and then in boiling DDI water for 1 h)was placed on the vacuum table. After sonication for 1 h, the inkwas coated on the center of the membrane (5.0 cm2 area) bypainting to reach a cathode catalyst loading of 4.0 mg cm�2.Membrane electrode assembly (MEA) was done as follow: The

Fig. 2. SEM images of (a) FCNT (b) Fe-P-C_Ar and (c) Fe-P-C_Ar-NH900. (d) TEM imagecations. (g) BET surface area analysis and (h) pore size distribution of Fe-P-C_Ar and Fe

catalyst coated membrane (catalyst facing up) was placed on topof the anode and a commercial gas diffusion layer (GDL BC 25) wasplaced on top of the painted catalyst. The MEA was completed byhot pressing at 120 °C for 4 min using a pressure of 600 pounds fora 5 cm2 MEA. Fuel cell testing was carried out by flowing hydrogento the anode at 200 standard cubic centimeters per minutes(sccm) and by flowing air or oxygen to cathode at 200 sccm. Thecell temperature was held at 80 °C and the backpressure was set at20 psig for both the anode and cathode side. Once the cell reachedthe appropriate potential, it was tested with an increment of0.03 V point�1 and 20 s point�1.

3. Results and discussion

The initial phase of the current research was to study the effectof ammonia treatment temperature and heat treatment protocol.In order to optimize the heat treatment conditions, after acidleaching, some catalysts were first pyrolyzed in Ar at 900 °C for 3 hand then heat treated in ammonia at different temperatures. Asbenchmarks, after acid leaching, some samples were heat treated

of Fe-P-C_Ar. (e) and (f) HRTEM images of Fe-P-C_Ar-NH900 at different magnifi--P-C_Ar-NH900.

P. Zamani et al. / Nano Energy 26 (2016) 267–275270

in only ammonia at varying temperatures. Sample nomenclaturein Table S1 indicates sample heating environment and tempera-ture. SEM images (Fig. 2a–c) shows the morphology of the FCNTsused as carbon supports and the resulting catalysts (before andafter ammonia treatment at 900 °C for 15 min). As it can be seenfrom the SEM images (Fig. 2b and c), the distribution of the car-bon–nitrogen precursor (PANI) on the surface of the carbon sup-port (FCNT) is uniform, and it seems that after NH3 treatment alarge amount of porous structures are formed. The TEM image ofFe-P-C_Ar (Fig. 2d) in comparison to Fe-P-C_Ar-NH900 (Fig. 2e)also supports the notion of more porosity after the ammoniatreatment. This was also confirmed from the BET results which aresummarized in Fig. 2g and h. The rapid rise at low relative pres-sures (Fig. 2g) suggests the presence of micro-pores for bothsamples, while at high relative pressure, the lack of saturationindicates the formation of macro-pores that are present in higheramounts for the NH3-treated sample. Both of the Fe-P-C_Ar-NH900 and Fe-P-C_Ar catalysts show remarkably high specificsurface area (higher for Fe-P-C_Ar-NH900) of ca. 1100 m2 g�1 and1050 m2 g�1, respectively. Such high surface area is attractive froma mass transport perspective. This can effectively facilitate reactantaccess to catalytically active ORR sites, thereby improving elec-trode utilization and performance. In Fig. 2h, the pore size dis-tribution clearly indicates the variety in the pore sizes from 2 nm

Fig. 3. (a) Raman spectra of FCNT, Fe-P-C_Ar and Fe-P-C_Ar-NH900, (b) C 1s, (c) N 1s spC_Ar-NHxxx catalysts derived at different NH3 temperatures from 800 °C to 1000 °C.

to 200 nm. The extra macro-pores that were generated duringammonia treatment could decrease the diffusion length of oxygenmolecules which has been reported helpful for improving the ORRperformance [34]. TEM images (Fig. 2d–f) also shows formation ofgraphene like structures (red arrows) after pyrolysis. The in-situformation of graphene structures by using PANI is likely due to thesimilarity between graphene and the aromatic structures of PANI[26,27,29]. The high resolution TEM (HRTEM) image of Fe-P-C_Ar-NH900 indicates that the graphitic structures remained after am-monia treatment (Fig. 2f).

The Raman spectra of samples before and after ammoniatreatment in Fig. 3a shows the asymmetric graphitic carbon re-sonance (G band) peak at ca. 1580 cm�1 and defect carbon (Dband) peak at ca. 1340 cm�1. The lower D/G intensity ratio of thepyrolyzed Fe-P-C_Ar compared to that of the FCNT is probably dueto more graphitic carbon species, which was likely arisen from in-situ formation of graphene structures during the high temperaturepyrolysis of PANI in the presence of the transition metal. However,the lower D/G ratio after ammonia treatment (Fe-P-C_Ar-NH900)was likely due to the lower content of disordered carbon, whichwas due to the faster reaction between amorphous carbons (dis-ordered carbons) with ammonia than the reaction between NH3

and graphitic carbon [10]. XPS C1s spectra of the catalysts (Fig. 3b)shows higher amounts of sp2 bonded carbon (284.85 eV) [35] and

ectra and (d) the corresponding nitrogen content and species distribution of Fe-P-

P. Zamani et al. / Nano Energy 26 (2016) 267–275 271

a lower amount of oxidized carbon species (286.35 eV) [36,37] forthe NH3-treated samples in comparison to Fe-P-C_Ar. For instance,the Fe-P-C_Ar-NH900 has 56.20 at% sp2 carbon and 11.69 at% oxi-dized carbon, compared to 52.02 at% sp2 carbon and 12.35 at%oxidized carbon for Fe-P-C_Ar. In agreement with Raman spectra,XPS indicates a higher degree of graphitization and reducedamount of defect structures for the NH3-treated sample. Thehigher amount of graphitic carbon could be due to an additionalhigh temperature pyrolysis step (under NH3 atmosphere) in thepresence of iron species. However, the amount of sp2 C¼C de-creases with increasing the NH3 temperature from 58.36 at% forFe-P-C_Ar-NH800 down to 53.53 at% of sp2 carbon for Fe-P-C_Ar-NH1000. This finding suggests that the reaction between NH3 andgraphitic carbon is facilitated at higher ammonia treatment tem-peratures, most likely due to the abundance of thermal energy tosurpass the activation energy of the reaction. In Fig. S1a the XPSanalysis also shows the existence of sulfur species mostly in theform of iron sulfide structure at 162.0 eV and carbon sulfide at163.5 and 164.7 eV [38], with these dopant species arising fromthe APS that was used to initiate the polymerization of PANI. XPSN1s spectra of the catalysts derived at different NH3 heat treat-ment are also provided, in addition to the corresponding nitrogencontent column plots in Fig. 3c and d, respectively. It can be seenthat while the ammonia temperature has no significant effect onthe pyrrolic N species quantity (400.2 eV), it does affect theamount of pyridinic (398.6 eV) and graphitic/quaternary nitrogencontents (401.5 eV). On the other hand, the binding energy thatappears at 398.6 eV could also be a contribution of N–Fe boundand it is hard to distinguish between them [28]. From the XPSresults, increasing the NH3 temperature from 800 °C to 1000 °Cresults in higher graphitic/quaternary nitrogen amounts in atomicpercentage. Pyridinic N amount however shows increasing withtemperature and a maximum at 900 °C (pyridinic N content of

Fig. 4. (a) STEM image of Fe-P-C_Ar-NH900 and the corresponding EELS elemental mappN-C (light green, red, blue), and (f) superimposed Fe-N (light green, red) with each pixe

0.96 at%) and then it decreases at higher temperatures. From thetrends of XPS results, when comparing with the sample withoutany ammonia treatment (Fig. S1b), it appears that the NH3 treat-ment results in an increase in nitrogen content, accompanied byan increase in the relative concentration of pyridinic and graphiticnitrogen. While for Fe-P-C_Ar the amount of pyridinic and gra-phitic N were respectively 0.55 and 0.67 at%, they are corre-spondingly as high as 0.6–0.96 and 1.08–1.71 after ammoniatreatment and doping. This trend is consistent with previous re-ports [18,39]. However, as these M-N-C systems are highly het-erogeneous in nature, active site elucidation is very difficult andonly recently have sophisticated in-situ synchrotron and Moss-bauer techniques started to shed light on this important topic [39–41]. With regards to the active site density on these catalysts, theNH3 treatment etches some of the disordered carbon away thatcan allow for exposure of any ORR active sites that were buried.Additionally, the increased surface area and mesopore contentcould mean that the number of active sites accessible to oxygenmolecules is significantly increased.

The scanning transmission electron microscope (STEM) imageshown in Fig. 4a and the corresponding EELS images shown inFig. 4b–d) outline the elemental distribution in this selected areaof the catalyst. Considering that each pixel of the image representsan area of 4.0 nm�4.0 nm, these images confirm that the iron,nitrogen and carbon species are very well dispersed throughoutthe FCNT and in-situ formed graphene structures. The uniformdistribution can particularly be seen when looking at the super-imposed images of Fe-N-C (Fig. 4e) and Fe-N elements (Fig. 4f).Such well-dispersed Fe species suggest that they can potentiallycoordinate with the nitrogen atoms to form Fe-N and Fe-N-Cmoieties, which are believed to be highly active for ORR [10,42].However, sophisticated methods, including in-situ XAS or Moss-bauer spectroscopy are needed to confirm these structures.

ing images for the selected spectrum area of (b) C, (c) N, (d) Fe, (e) superimposed Fe-l represents an area of 4.0 nm �4.0 nm.

P. Zamani et al. / Nano Energy 26 (2016) 267–275272

The EELS spectrum and the element survey of the Fe-P-C_Ar-NH900 sample is displayed in Fig. S2 which shows the location ofeach element peak. As it can be seen from XRD patterns (Fig. S3),the peaks resulting from crystalline phases (2θ¼25.9° and 43.2°)can be predominantly allocated to graphite crystalline structureswhile the peaks appearing at 2θ¼30.8°, 33.5°, 46.8° and 53.0° aredominant formation of FeS species [28] which is in good agree-ment with the XPS S 2p spectra (Fig. S1a). EDX elemental mappingand line scanning were also applied to study the distribution ofelements in the catalyst material in a larger scale. Line scanning

Fig. 5. LSCV results for (a) catalysts derived from different synthesis protocols and (b) Fe(d) durability results for Fe-P-C_Ar-NH900. (e) Half-cell performance comparison of catathe polarization curves are performed in O2 saturated 0.5 M sulfuric acid at 900 rpm rota(squares) under H2-air anode-cathode feeds. Fuel cell tests used 0.5 mg cm�2 Pt at the

EDX in Fig. S4a also supports the well-distributed nature of theseelements along the FCNT and graphene sheets. Moreover, as can beseen in the STEM image of a different area of the same samplewith lower magnification (Fig. S4b) and the corresponding EDXimages (Fig. S4c–f), all of the elements were found to be homo-genously distributed throughout the whole sample area andthrough all of the FCNT and in-situ formed graphene morpholo-gies. The highly mesoporous morphology formed after NH3

treatment, as well as extremely high surface area and the welldispersed Fe-N-C distribution are believed to be highly beneficial

-P-C_Ar-NHxxx catalysts derived at different ammonia treatment temperatures. (c),lysts derived from CNT and carbon black, before and after ammonia treatment. Alltional speed. (f) Fuel cell testing for Fe-P-C_Ar-NH900 (stars) and Fe-P-KJ_Ar-NH900anode.

P. Zamani et al. / Nano Energy 26 (2016) 267–275 273

for the ORR performance [10].From the ORR polarization curves collected at 900 rpm (Fig. 5a

and b), NH3 treatment improved the current densities, onset andhalf-wave (E1/2) potentials significantly. NH3 treatment inducesformation of some types of heteroatom doping that are beneficialfor ORR which could likely affect the catalytic performance bytuning the active sites via generating local heterogeneity in elec-tron density. Moreover, the higher mesoporous and macro porousnanostructures of the NH3 treated catalyst (Fig. 2b–e and h) couldfacilitate the mass transport of oxygen into the catalyst layer andthe removal of water molecules to the electrolyte. The effect ofheat treatment protocol on ORR performance is shown in Fig. 5a.Samples which underwent a second pyrolysis in Ar before NH3

showed better performance compared to those which were di-rectly heat treated in ammonia. The onset potential is howeversimilar, which indicates that the nature of the active sites is likelysimilar. The key to the activity gains following this second pyr-olysis mainly arise due to the acid leach. Any inactive inorganicmetal species are removed through this procedure increasing theporosity and surface exposure of the catalyst. This second heattreatment then likely removes any residues remaining followingthe acid leaching procedure and increases the degree of graphiti-zation. In Fig. 5b it can be seen that the catalyst derived at the NH3

temperature of 900 °C showed the best ORR performance (E1/2 of0.81 V vs. RHE) compared to those synthesized at higher ammoniatemperatures. The obtained half-wave potential places this cata-lyst on par with some of the most active NPMCs reported to date[12,18]. With respect to the XPS results in Fig. 3c and d, the activitytrends show that the pyridinic N content potentially plays animportant contribution in the ORR activity, as the sample with thehighest pyridinic N content (Fe-P-C_Ar-NH900 with pyridinic Ncontent of 0.96 at%) and relatively high amount of graphitic ni-trogen, exhibited the best ORR activity. On the other hand, therewas no clear similarity in trends between ORR activity and in-crease in the graphitic N content. Moreover, the possible con-tribution of sulfur species in ORR was studied and as benchmark,S-doped graphene (SG) was synthesized throughout high tem-perature pyrolysis of a sulfur precursor (benzyl disulfide) andgraphene oxide at 900 °C in Ar. Also N-doped graphene (NG) wasprepared via pyrolysis (similar to SG pyrolysis conditions) ofpolybenzimidazole in the absence of any metal precursor. Whentesting RDE in acid (Fig. S5), the SG shows a very low activitytowards ORR, suggesting that S species do no contribute in ORR inacid. On the other hand, an obvious improvement in ORR activitywas obtained when testing nitrogen doped catalysts, with a re-markable higher performance for Fe-P-C_Ar-NH900 compared toNG. The remarkable higher activity of Fe-P-C_Ar-NH900 catalyst,suggest that unlike S which does not contribute significantly inORR in acid, Fe and N species likely involve in ORR activity.

The reaction between NH3 and carbon based materials includesthe exchange of O2-containing groups with nitrogen-based spe-cies, accompanied with the etching of carbon by radicals formedduring ammonia decomposition at high temperatures and there-fore it is related to the NH3 treatment temperature [43]. When theammonia temperature is above 900 °C, breakdown of the porousstructures that were generated during overall previous pyrolysissteps could likely happen. A similar phenomena was observedpreviously by another group, in which they showed that by in-creasing the NH3 temperature, the average mesoporous size de-creased, pointing out the breaking down of the porous morphol-ogy [44]. Moreover, from the C 1s XPS results in Fig. 3b, it wasobserved that with increasing the NH3 temperature, higheramounts of sp2 graphitic carbon species undergo breakdown. Thebreakdown of graphitic carbon structures and the likeliness ofporous morphologies decomposition could likely be the reasonsfor the lower ORR activity at ammonia treatment temperatures

above 900 °C.From RRDE results for Fe-P-C_Ar-NH900 and Fe-P-C_Ar (Fig.

S6a and b), the number of electrons transferred per reducedoxygen molecule is measured to be 3.9070.05 at relatively highpotentials (0.4–0.8 V vs. RHE), thereby showing a very good se-lectivity towards the four electron reaction. On the other hand,while the peroxide yields for Fe-P-C_Ar in that high potentialrange can be as high as 1075%, it was much lower for Fe-P-C_Ar-NH900 (almost 5%) which indicates that after ammonia treatment,the selectivity to produce H2O during ORR increased meaningfully.The Fe-P-C_Ar-NH900 catalyst also shows relatively good dur-ability behaviour (Fig. 5c and d). After 5000 cycles in N2, a 20 mVand a 45 mV potential loss toward ORR was observed in onset andhalf-wave potentials, respectively. RDE testing for Fe-P-C_Ar-NH900 exhibits improvement compared to the reference samples(Fe-P-KJ_Ar and Fe-P-KJ_Ar-NH900) which were synthesized usingketjen black instead of CNT as carbon support, under the opti-mized condition (similar to the synthesis conditions of Fe-P-C_Arand Fe-P-C_Ar-NH900, respectively), in order to be used asbenchmarks.

In order to evaluate the Fe-P-C_Ar-NH900 catalyst underPEMFC operating conditions, they were integrated as cathodes intoa single cell MEA. Although H2–O2 performance provides a gaugeof intrinsic catalyst layer activity, testing under H2-air conditionsshould be provided as evaluation under practical operative con-ditions. When applying air as the cathode reactant (Fig. 5f), theMEA showed current densities of 77 mA cm�2 at 0.8 V and537 mA cm�2 at 0.6 V, in addition to a maximum power density of335 mW cm�2 with a clear improvement in comparison to thereference sample (Fe-P-KJ_Ar-NH900). From SEM images in Fig.S7a and b, in comparison to Fe-P-C_Ar-NH900, Fe-P-KJ_Ar-NH900has less macro pores which could likely limits diffusion of the feedinto the catalyst bulk towards the active sites and this could be thereason for better performance of Fe-P-C_Ar-NH900 at higher cur-rent densities which is the mass transport control region. To thebest of our knowledge, the obtained power densities and iR-freecurrent densities in an H2-air fuel cell for Fe-P-C_Ar-NH900 are thebest results reported to date in the literatures for a NPMC elec-trode. The electrocatalyst morphology, with its high surface area ofca. 1100 m2 g�1 and variety of pores ranging in size from 2 nm to200 nm, is the likely cause of the high H2-air performance. Thiscatalyst structure is conducive to mass transport through thecatalyst layer, a factor that is increasingly important when usingair as the reactant feed. Clearly the tuned chemical modificationand porous catalyst morphology render Fe-P-C_Ar-NH900 an at-tractive non-precious ORR catalyst for PEMFCs, based on half-celland fuel cell performance evaluations.

4. Conclusion

An NH3 treatment was applied and optimized to tune the ac-tivity and improve the PEMFC performance of ORR catalysts de-rived from iron, PANI and CNT. NH3 contributed by enriching theN-dopant concentration, while careful choice of the temperaturehad an impact on nitrogen dopant identity and ORR activity. Thefinal catalyst consisted of in-situ formed graphene-like structuresthat had well-distributed nitrogen and iron defects. The catalystderived under the optimized condition (F-P-C_Ar-NH900) ex-hibited high catalyst activity, including an E1/2 of 0.81 V vs RHEthrough RDE testing. MEA performance under H2-air conditionsthat are application friendly, current densities of 77 mA cm�2 at0.8 V and 537 mA cm�2 at 0.6 V were achieved. Furthermore, amaximum power density of 335 mW cm�2 at 0.6 V was observed.To the best of our knowledge, this is the best H2-air performanceshown to date for a Pt-free cathode. This catalyst also showed a

P. Zamani et al. / Nano Energy 26 (2016) 267–275274

very good selectivity towards the four electron reaction with thenumber of electrons transferred per reduced oxygen moleculewas calculated to be 3.90 by RRDE. These electrochemical eva-luations indicate that the chemical modification of Fe-PANI-CNTcatalyst by NH3 results in a highly promising Pt-free PEMFC ORRelectrocatalyst.

Acknowledgements

This work was supported by the University of Waterloo and theWaterloo Institute for Nanotechnology. TEM imaging was carriedout at the Canadian Center for Electron Microscopy (CCEM) locatedat McMaster University. This research was conducted as part of theCatalysis Research for Polymer Electrolyte Fuel Cells (CaRPE FC)Network administered from Simon Fraser University and sup-ported by Automotive Partnership Canada (APC) Grant no. APCPJ417858–11 through the Natural Sciences and Engineering ResearchCouncil of Canada (NSERC).

Appendix A. Supporting information

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.nanoen.2016.05.035.

References

[1] Y. Jiao, Y. Zheng, M. Jaroniec, S.Z. Qiao, Origin of the electrocatalytic oxygenreduction activity of graphene-based catalysts: a roadmap to achieve the bestperformance, J. Am. Chem. Soc. 136 (2014) 4394–4403.

[2] R. Wang, D.C. Higgins, D.U. Lee, S. Prabhudev, F.M. Hassan, V. Chabot, et al.,Biomimetic design of monolithic fuel cell electrodes with hierarchical struc-tures, Nano Energy 20 (2016) 57–67.

[3] D.C. Higgins, R. Wang, M.A. Hoque, P. Zamani, S. Abureden, Z. Chen, Mor-phology and composition controlled platinum–cobalt alloy nanowires pre-pared by electrospinning as oxygen reduction catalyst, Nano Energy 10 (2014)135–143.

[4] U.S. Drive, in: U.S. Drive (Ed.), Fuel Cell Technical Team Roadmap, 2013.[5] D. Higgins, P. Zamani, A. Yu, Z. Chen, The application of graphene and its

composites in oxygen reduction electrocatalysis: a perspective and review ofrecent progress, Energy Environ. Sci. (2015).

[6] D.C. Higgins, Z. Chen, Recent progress in non-precious metal catalysts for PEMfuel cell applications, Can. J. Chem. Eng. 91 (2013) 1881–1895.

[7] Z. Chen, D. Higgins, A. Yu, L. Zhang, J. Zhang, A review on non-preciousmetal electrocatalysts for PEM fuel cells, Energy Environ. Sci. 4 (2011)3167–3192.

[8] F. Cheng, J. Chen, Metal–air batteries: from oxygen reduction electrochemistryto cathode catalysts, Chem. Soc. Rev. 41 (2012) 2172–2192.

[9] J. Wu, H. Yang, Platinum-based oxygen reduction electrocatalysts, Acc. Chem.Res. 46 (2013) 1848–1857.

[10] F. Jaouen, E. Proietti, M. Lefèvre, R. Chenitz, J.-P. Dodelet, G. Wu, et al., Recentadvances in non-precious metal catalysis for oxygen-reduction reaction inpolymer electrolyte fuel cells, Energy Environ. Sci. 4 (2011) 114.

[11] H. Chung, G. Wu, D. Higgins, P. Zamani, Z. Chen, P. Zelenay, Heat-treated non-precious metal catalysts for oxygen reduction, Electrochem. N4 Macrocycl.Metal Complexes: Springer (2016) 41–68.

[12] D. Zhao, J.L. Shui, L.R. Grabstanowicz, C. Chen, S.M. Commet, T. Xu, et al.,Highly efficient non-precious metal electrocatalysts prepared from one-potsynthesized zeolitic imidazolate frameworks, Adv. Mater. 26 (2014)1093–1097.

[13] V. Armel, J. Hannauer, F. Jaouen, Effect of ZIF-8 crystal size on the O2 Electro-reduction performance of pyrolyzed Fe–N–C catalysts, Catalysts 5 (2015)1333–1351.

[14] J. Tian, A. Morozan, M.T. Sougrati, M. Lefevre, R. Chenitz, J.P. Dodelet,et al., Optimized synthesis of Fe/N/C cathode catalysts for PEM fuel cells: amatter of iron-ligand coordination strength, Angew. Chem. 52 (2013)6867–6870.

[15] A. Serov, K. Artyushkova, N.I. Andersen, S. Stariha, P. Atanassov, Original me-chanochemical synthesis of non-platinum group metals oxygen reductionreaction catalysts assisted by sacrificial support method, Electrochim. Acta 179(2015) 154–160.

[16] A. Serov, M.H. Robson, M. Smolnik, P. Atanassov, Tri-metallic transition metal–nitrogen–carbon catalysts derived by sacrificial support method synthesis,Electrochim. Acta 109 (2013) 433–439.

[17] U. Tylus, Q. Jia, K. Strickland, N. Ramaswamy, A. Serov, P. Atanassov, et al.,Elucidating oxygen reduction active sites in pyrolyzed metal-nitrogen

coordinated non-precious-metal electrocatalyst systems, J. Phys. Chem. CNanomater. Interfaces 118 (2014) 8999–9008.

[18] A. Serov, K. Artyushkova, P. Atanassov, Fe-N-C oxygen reduction fuel cellcatalyst derived from carbendazim: synthesis, structure, and reactivity, Adv.Energy Mater. (2014) 4, http://dx.doi.org/10.1002/aenm.201301735.

[19] D.C. Higgins, F.M. Hassan, M.H. Seo, J.Y. Choi, M.A. Hoque, D.U. Lee, et al.,Shape-controlled octahedral cobalt disulfide nanoparticles supported on ni-trogen and sulfur-doped graphene/carbon nanotube composites for oxygenreduction in acidic electrolyte, J. Mater. Chem. A 3 (2015) 6340–6350.

[20] X. Fu, J.Y. Choi, P. Zamani, G. Jiang, M.A. Hoque, F.M. Hassan, et al., Co-N de-corated hierarchically porous graphene aerogel efficient oxygen reductionreaction in acid, ACS Appl. Mater. Interfaces 8 (2016) 6488–6495.

[21] J.-Y. Choi, D. Higgins, G. Jiang, R. Hsu, J. Qiao, Z. Chen, Iron-tetracyanobenzenecomplex derived non-precious catalyst for oxygen reduction reaction, Elec-trochim. Acta 162 (2015) 224–229.

[22] J.-Y. Choi, D. Higgins, Z. Chen, Highly durable graphene nanosheet supportediron catalyst for oxygen reduction reaction in PEM fuel cells, J. Electrochem.Soc. 159 (2012) B87.

[23] J. Wu, W. Li, D. Higgins, Z. Chen, Heat-treated nonprecious catalyst using Feand nitrogen-rich 2,3,7,8-Tetra(pyridin-2-y)l)(pyrazino)[2,3-g](quinoxalinecoordinated complex for oxygen reduction reaction in) PEM fuel cells, J. Phys.Chem. C 115 (2011) 18856–18862.

[24] Ja.-Yeon. Choi, Ryan Hsu, Z. Chen, Nanoporous carbon-supported Fe/Co-Nelectrocatalyst for oxygen reduction reaction in PEM fuel cells, ECS Trans. 23(2010) 101–112.

[25] J. Choi, R. Hsu, Z. Chen, Highly active porous carbon-supported nonpreciousmetal- N electrocatalyst for oxygen reduction reaction in PEM fuel cells, J.Phys. Chem. C 114 (2010) 8048–8053.

[26] Gang Wu, Karren L. More, Christina M. Johnston, P. Zelenay, High-performanceelectrocatalysts for oxygen reduction derived from polyaniline, iron, and co-balt, Science 332 (2011) 443–447.

[27] Gang Wu, P. Zelenay, Nanostructured nonprecious metal catalysts for oxygenreduction reaction, Acc. Chem. Res. 46 (2013) 1878–1889.

[28] G. Wu, C.M. Johnston, N.H. Mack, K. Artyushkova, M. Ferrandon, M. Nelson,et al., Synthesis–structure–performance correlation for polyaniline–Me–C non-precious metal cathode catalysts for oxygen reduction in fuel cells, J. Mater.Chem. 21 (2011) 11392.

[29] G. Wu, K.L. More, P. Xu, H.L. Wang, M. Ferrandon, A.J. Kropf, et al., A carbon-nanotube-supported graphene-rich non-precious metal oxygen reductioncatalyst with enhanced performance durability, Chem. Commun. 49 (2013)3291–3293.

[30] H. Peng, Z. Mo, S. Liao, H. Liang, L. Yang, F. Luo, et al., High performance Fe- andN- doped carbon catalyst with graphene structure oxygen reduction, Sci. Rep.(2013) 3.

[31] P. Zamani, D. Higgins, F. Hassan, G. Jiang, J. Wu, S. Abureden, et al., Electrospuniron–polyaniline–polyacrylonitrile derived nanofibers as non–precious oxygenreduction reaction catalysts for pem fuel cells, Electrochim. Acta 139 (2014)111–116.

[32] Drew C. Higgins, Gang Wu, Hoon Chung, Ulises Martinez, Shuguo Ma,Zhongwei Chen, et al., Manganese-based non-precious metal catalyst foroxygen reduction in acidic Media, ECS Trans. 31 (2014) 35–42.

[33] N.R. Sahraie, U.I. Kramm, J. Steinberg, Y. Zhang, A. Thomas, T. Reier, et al.,Quantifying the density and utilization of active sites in non-precious metaloxygen electroreduction catalysts, Nat. Commun. 6 (2015) 8618.

[34] J. Liang, Y. Zheng, J. Chen, J. Liu, D. Hulicova-Jurcakova, M. Jaroniec, et al., FacileOxygen reduction on a three-dimensionally ordered macroporous graphiticC3N4/carbon composite electrocatalyst, Angew. Chem. Int. Ed. 51 (2012)3892–3896.

[35] J. Park, W.C. Mitchel, S. Elhamri, L. Grazulis, J. Hoelscher, K. Mahalingam, et al.,Observation of the intrinsic bandgap behaviour in as-grown epitaxial twistedgraphene, Nat. Commun. 6 (2015) 5677.

[36] X. Huo, J. Liu, B. Wang, H. Zhang, Z. Yang, X. She, et al., A one-step method toproduce graphene–Fe3O4composites and their excellent catalytic activities forthree-component coupling of aldehyde, alkyne and amine, J. Mater. Chem. A 1(2013) 651–656.

[37] K. Zhang, B.T. Ang, L.L. Zhang, X.S. Zhao, J. Wu, Pyrolyzed graphene oxide/resorcinol-formaldehyde resin composites as high-performance super-capacitor electrodes, J. Mater. Chem. 21 (2011) 2663.

[38] F.M. Hassan, R. Batmaz, J. Li, X. Wang, X. Xiao, A. Yu, et al., Evidence of covalentsynergy in silicon-sulfur-graphene yielding highly efficient and long-life li-thium-ion batteries, Nat. Commun. (2015) 6.

[39] U.I. Kramm, J. Herranz, N. Larouche, T.M. Arruda, M. Lefevre, F. Jaouen, et al.,Structure of the catalytic sites in Fe/N/C-catalysts for O2-reduction in PEM fuelcells, Phys. Chem. Chem. Phys.: PCCP 14 (2012) 11673–11688.

[40] A. Serov, K. Artyushkova, E. Niangar, C. Wang, N. Dale, F. Jaouen, et al., Nano-structured non-platinum catalysts for automotive fuel cell application, NanoEnergy 16 (2015) 293–300.

[41] Qingying Jia, Nagappan Ramaswamy, Hasnain Hafiz, Urszula Tylus,Kara Strickland, Gang Wu, et al., Experimental observation of redox-inducedFe-N switching behavior as a determinant role for oxygen reduction activity,ACS Nano 9 (2015) 12496–12505.

[42] U.I. Kramm, I. Herrmann-Geppert, P. Bogdanoff, S. Fiechter, Effect of an am-monia treatment on structure, composition, and oxygen reduction reactionactivity of Fe–N–C catalysts, J. Phys. Chem. C 115 (2011) 23417–23427.

[43] Christian L. Mangun, Kelly R. Benak, James Economy, K.L. Foster, Surfacechemistry, pore sizes and adsorption properties of activated carbon fibers andprecursors treated with ammonia, Carbon 39 (2001) 1809–1820.

[44] X.Q. Wang, J.S. Lee, Q. Zhu, J. Liu, Y. Wang, S. Dai, Ammonia-treated orderedmesoporous carbons as catalytic materials for oxygen reduction reaction,Chem. Mater. 22 (2010) 2178–2180.

P. Zamani et al. / Nano Energy 26 (2016) 267–275 275

Pouyan Zamani received his Master in Polymer En-gineering-Nanotechnology from Amirkabir University(Tehran Polytechnique) in 2010. He is now a Ph.D.student in Chemical Engineering at the University ofWaterloo (since 2012) under the supervision of Dr.Zhongwei Chen, with a focus on the development ofnanostructured based materials as non-precious oxy-gen reduction reaction catalysts for fuel cellapplications.

Dr. Drew Higgins obtained his Ph.D. in Chemical En-gineering from the University of Waterloo in July 2015.His disseration work was under the supervision of Dr.Zhongwei Chen, during which time he spent nearly oneyear as a Visiting Scholar at the Los Alamos NationalLaboratory. He is now an NSERC Banting PostdoctoralFellow at Stanford University, working under the su-pervision of Professor Thomas Jaramillo. His researchinterests include the development of nanostructuredfunctional materials and their integration into cleanenergy electrochemical devices.

Dr. Fathy M. Hassan obtained his Ph.D. in PhysicalChemistry, Cairo University. He held the Japan Societyfor Promotion of Science Award at the National Instituteof Advanced Industrial Science and Technology. He ob-tained a Ph.D. in Chemical Engineering (Nanotechnol-ogy), University of Waterloo. Holder of honored award ofoutstanding achievements and Park and Veva ReillyMedal for Proficiency in Research. A postdoctoral fellowwith research experience in synthetic materials chem-istry and nano-architecture materials for advancedelectrochemical energy storage and conversion. Recentlyhe was awarded NSERC postdoctoral fellow for advanced

materials for next generation lithium-ion batteries.

Dr. Xiaogan Fu is a postdoctoral research fellow at theApplied Nano Materials & Clean Energy Laboratory,Department of Chemical Engineering at the Universityof Waterloo, Canada under Prof. Zhongwei Chen's su-pervision. He obtained his Ph.D. in Materials Chemistryfrom Lanzhou University, China, in 2013. His researchinterests focuses on the design and development ofadvanced materials for energy storage applications.

Ja-Yeon Choi is a Ph.D. student in Chemical Engineer-ing at the University of Waterloo in Dr Zhongwei Chen'sresearch group, and a visiting researcher at BallardPower Systems. He completed his Masters in 2013 onthe application of nanostructured carbon materials andnovel precursors to develop non-precious metal cata-lysts for Proton Exchange Membrane fuel cells. Hiscurrent research interests lie in the development ofadvanced non-precious metal and Pt catalysts withcontrolled nanostructures for fuel cell and batteryapplications.

Dr. Md. Ariful Hoque completed his Ph.D. in ChemicalEngineering at the University of Waterloo, Canada in2016. He obtained his MASc in Chemical Engineeringfrom Yeungnam University, South Korea in 2012 and B.Sc. in Chemical Engineering from Bangladesh Uni-versity of Engineering and Technology (BUET), Bangla-desh in 2009. His research focuses on the developmentof heteroatom-doped nanocarbon materials as plati-num and platinum alloy supports for fuel cellapplications.

Gaopeng Jiang received his master's degree fromDonghua University in 2012 and currently is the Ph.D.candidate on Chemical Engineering (Nanotechnology)in University of Waterloo, Canada. His research interestmainly focuses on the design and development of na-nostructured materials for energy storage and conver-sion devices including fuel cells, batteries and sensors.

Dr. Zhongwei Chen is Canada Research Chair Professorin Advanced Materials for Clean Energy at University ofWaterloo. His research interests are in the developmentof advanced energy materials for metal-air batteries,lithium-ion batteries and fuel cells. He has published1 book, 7 book chapters and more than 150 peer re-viewed journal articles with over 10,000 citations withH-index 44 (GoogleScholar). He is also listed as in-ventor on 15 US/international patents, with several li-censed to companies in USA and Canada. He was re-cipient of the 2016 E. W. R Steacie Memorial Fellowship,which followed shortly upon several other prestigious

honors, including the Ontario Early Researcher Award,

an NSERC Discovery Supplement Award, the Distinguished Performance and theResearch Excellence Awards from the University of Waterloo.