- Email: [email protected]
A Pt-free catalyst for oxygen reduction reaction based on Fe–N multiwalled carbon nanotube composites
Electrochimica Acta 107 (2013) 126–132
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
A Pt-free catalyst for oxygen reduction reaction based on Fe–N multiwalled carbon nanotube composites Belabbes Merzougui a,∗∗ , Abdouelilah Hachimi a , Akeem Akinpelu a , Saheed Bukola a , Minhua Shao b,∗ a b
Center of Nanotechnology (CENT) & Department of Chemistry, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia UTC Power, South Windsor, CT 06074, USA
a r t i c l e
i n f o
Article history: Received 28 January 2013 Received in revised form 7 June 2013 Accepted 7 June 2013 Available online 17 June 2013 Keywords: Non precious metal catalysts Oxygen reduction reaction Durability Methanol tolerance
a b s t r a c t An iron-based non-precious metal (NPM) catalyst was synthesized by chemically coating a polymer containing nitrogen, such as polyaniline (PANI), on multiwalled carbon nanotubes (MWCNTs), followed by iron complexation and heat treatment in a decomposable source of nitrogen (ammonia) at high temperature. Its excellent activities for oxygen reduction reaction (ORR) in both acidic and alkaline media were confirmed by thin film rotating disk electrode technique. Also, durability test was performed in both media. After 20,000 (0.65–1.0 V) potential cycles, the half-wave potential decayed by only 40 mV in 0.1 M HClO4 , while in 0.1 M KOH it gained 25 mV. In addition, no activity decay was observed by adding methanol in the acidic solution indicating its excellent tolerance to methanol oxidation. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction Cost and durability are considered as the main challenges facing the commercialization in mass scale of polymer electrolyte fuel cells (PEFCs) [1]. A large portion of PEFC systems cost is still associated with cathode catalyst for oxygen reduction reaction (ORR) [2–4]. Nevertheless, great efforts have been so far made to develop advanced ORR catalysts with lower cost and higher activity and durability aiming to replace a substantial amount of the costly Pt-based materials, which are commonly used in today’s fuel cells technologies [3,5,6]. Among them, non-precious metal (NPM) catalysts consisting of transition metals (in particular, Fe and Co)–nitrogen complexes on carbon support prepared by high temperature pyrolysis are the most promising approaches [7–11]. Despite the fact that the nature of the catalytic sites in NPM catalysts are still as a topic of debate, it is generally accepted that ORR activity and durability are largely influenced by the type of the transition metals (M) in the catalyst, the type of C and N precursors used, the ratios of M/N and N/C on the surface, and the synthesis conditions [12]. For example, an excellent ORR activity has been reported by the Dodelet group by mixing a carbon support with iron precursor, and then heating the materials in ammonia as the
source of nitrogen [8]. On the other hand, a similar activity has been achieved by the Zelenay group by polymerizing the aniline (nitrogen precursor) on high surface area carbon support in the presence of iron precursor, and then heating the mixtures in an inert gas [10]. The same group and others explored NPM catalysts supported on multiwalled carbon nanotubes (MWCNTs) [13–16]. To the best of our knowledge, the approach of chemically coating polymer containing nitrogen, such as polyaniline (PANI) on CNTs and heat treatment in a decomposable source of nitrogen, such as ammonia in order to increase the total nitrogen content (aiming to increase the graphitic nitrogen) merits to be studied. Moreover, none of the previous reported works has performed comparative studies of the activity and durability under long cycling in both alkaline and acidic media. Herein, we report an iron-based NPM catalyst through polymerization of aniline on CNTs and subjecting to heat treatment in ammonia at 900 ◦ C. The resulted material shows excellent ORR activity and durability under potential cycling in both acidic and alkaline media. It is important to note that cycling in acidic medium caused a drop in iron content due to dissolution. Yet, the activity of the catalyst did not decay dramatically as one can expect. 2. Experimental
∗ Corresponding author. Tel.: +966 38604378. ∗∗ Corresponding author. E-mail addresses: [email protected] (B. Merzougui), [email protected] (M. Shao). 0013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.electacta.2013.06.016
2.1. Synthesis of Fe-N-C/CNTs Multiwall carbon nanotubes (CNTs) were purchased from Cheap Tubes Incorporation, USA (http://www.cheaptubesinc.com/) and
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
were acid washed in-house before use. The BET area of the acid washed CNTs was in the range of 170 m2 g−1 . Aniline (Aldrich) was distilled before use. Ammonium persulfate {(NH4 )2 S2 O8 } (Aldrich) was used without further purification. MilliQ UV-plus water (Millipore) was used throughout all experiments. In a typical experiment, 5.0 g of aniline was mixed with 50 ml water. Thereafter, 37% HCl was added dropwise with continuous stirring until aniline became completely miscible with water. Then, 1.0 g of CNTs was added into the mixture and sonicated for 0.5 hr in a sonication bath to obtain well dispersed slurry and kept overnight under stirring to impregnate aniline into CNTs surface. As an oxidant agent, 10.0 g of {(NH4 )2 S2 O8 } was dissolved in water and added in dropwise to the mixture and kept overnight under stirring to ensure proper and complete polymerization. The obtained product, named as PANI/CNTs was then filtrated and washed with Di-water for several times until pH became less acidic (pH = 6), then dried under vacuum at 80 ◦ C overnight. In the next step, 1.0 g of the dried PANI/CNTs was dispersed in water in form of slurry using ultrasonication. On the other hand, 0.48 g of FeCl3 ·6H2 O (to give 10 wt.% Fe) was dissolved in water and added to the slurry. The mixture was then kept on a magnetic stirrer overnight to ensure complexation of FeCl3 onto PANI/CNTs. Before heat treatment, two methods were tried to dry the mixture: oven drying and freeze drying. Based on the ORR measurements, freeze drying method was found to lead to a higher activity. Freeze drying might help keep better dispersion or complexation of FeCl3 on PANI/CNTs. Freeze drying was performed with a Freeze Dryer (Labconco) for 3 days. The final product was then heat treated in 10% NH3 /N2 at 900 ◦ C for 1 hr (3 hr for ramping temperature from 25 ◦ C to 900 ◦ C). The resulted catalyst was denoted as Fe-N-C/CNTs. 2.2. Electrochemical measurements For ORR activity measurement, approximately 5 mg of the dried catalyst Fe-N-C/CNTs was dispersed in a mixture of water and isopropanol alcohol (30%, v/v) and 60 l of 1.66 wt.% Nafion® (prepared from 5 wt.%, Aldrich). The mixture was ultrasonicated for 10–20 min to obtain a uniform ink. Then, 20 l of the ink suspension was deposited on the pre-cleaned glassy carbon substrate (5 mm diameter, Pine Instruments) and allowed to dry. For high loading of catalyst, we repeated the operation for several times to achieve the desired loadings. A Ag–AgCl electrode (calibrated and converted to RHE) and Pt mesh were used as a reference and counter electrode, respectively. All the potential reported in this paper were converted to RHE. Before testing ORR activity, each electrode was cycled in nitrogen saturated 0.1 M HClO4 and 0.1 M KOH solutions until a stable CV was obtained (generally for 10–15 cycles between 0 and 1.2 V at 20 mV s−1 ). ORR activity was evaluated in both oxygen saturated acidic (0.1 M HClO4 ) and alkaline (0.1 M KOH) media using a rotating disk electrode (RDE) technique at rotation speed of 400 rpm with a scan rate of 5 mV s−1 . Catalysts durability was also investigated in oxygen saturated 0.1 M HClO4 and 0.1 M KOH under potential cycling between 0.65 and 1.0 V for 20,000 cycles using square-wave signal of 5 s at each potential. ORR activity was evaluated in fresh electrolytes in a separate cell after every 5000 cycles. After 20,000 cycles, the ORR activity was measured again in 0.1 M HClO4 containing 0.1 M methanol to evaluate its methanol tolerance. All the currents were normalized to the geometric area of electrodes. The ring current densities in the RRDE measurements were corrected according to the collection efficiency (30%) of the system. 2.3. Characterization High-end field-emission scanning electron microscope (FESEM, Tescan-Lyra-3) and energy-dispersive X-ray spectroscopy (EDX,
127
Fig. 1. XRD patterns of raw CNTs, PANI/CNTs and Fe-N-C/CNTs (obtained after HT at 900 ◦ C in 10% NH3 /N2 of FeCl3 /PANI/CNTs). The S in FeS and Ni in NiC of Fe-NC/CNTs came from persulfate used for polymerization of aniline and impurity in the raw CNTs (The EDX spectrum showed that Raw-CNTs contain 99 wt.% of carbon and trace of oxygen and nickel.).
Oxford-Xmax) were used to provide morphology and elemental composition information. TEM analysis was performed with FEI ˚ Titan ST operated at 300 kV with a point resolution of 2.0 A. X-ray diffractions (XRD) patterns were collected using a Rigaku ˚ Miniflex II instrument with a monochromator of CuK␣1 (1.5406 A) at 30 kV, 15 mA. The XRD patterns were recorded in the static scanning mode from 5◦ to 60◦ (2) at a detector angular speed of 2◦ min−1 and step size of 0.02◦ . Peak matching was determined with Philips X’Pert plus (V.2) using patterns diffractions files (PDF-2) from ICDD (2011). Raman spectra were taken on an iHR320 Horiba Spectrometer with charge-coupled device (CCD) using monochromatic laser (300 mW, 532 nm), grating of 1200 lines mm−1 and an aluminum substrate. Raman spectroscopy provided further information on the defects in CNTs that may be created during heat treatment in NH3 . BET area measurements were conducted using ASAP-2020 Physisorption Analyzer from Micrometrics. 3. Results and discussion Fig. 1 shows the XRD patterns of as received CNTs, PANI/CNTs, and Fe-N-C/CNTs. The similar intensity of (0 0 2) diffraction peak of graphitic carbon at ∼26◦ before and after polymerization of aniline on CNTs suggests that the polymerization did not cause a noticeable structure change in CNTs. The arising of a broader shoulder at ∼20◦ due to PANI in PANI/CNTs confirms the successful polymerization. However, when the system was subjected to complexation with FeCl3 and followed by heat treatment in ammonia at 900 ◦ C, a composite-like material consisting of Fe–N and Fe–C function groups was formed. The significant decrease in peak intensity at 26◦ also suggests the formation of nitrogen-doped carbon as a result of sacrifice of graphitic structure. A similar XRD pattern for Fe-N-C catalyst was reported by Tsai et al. [17]. The sharp diffraction peak at 45◦ in Fig. 1 may be due to free Fe nanoparticles. The structure changes of CNTs are also supported by the Raman spectroscopic data (Fig. 2). The peaks centered at 1334 cm−1 and 1596 cm−1 are referred to D- and G-band, respectively [18]. The D-band is related to the vibrations of sp3 bonds associated to the bonded carbon atoms of disordered graphite, while the G-band is related to the vibration of all sp2 bonds in well-graphitized structures [19]. The intensity ratio of the D- to G-band (D/G) interprets the degree of disorder that could occur in the carbon structure. The D/G ratios of raw-CNTs and Fe-N-C/CNTs catalyst were estimated to be 0.37 and
128
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
Fig. 2. Raman spectra of as-received CNTs (blue) and Fe-N-C/CNTs catalyst (red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
0.96, respectively. The higher D/G ratio after heat treatment suggests a significant increase in structure defects (though the addition of sp3 carbon structure cannot be totally ruled out). The creation of structure defects can be further supported by the increase in BET area from 170 m2 g−1 for CNTs to 540 m2 g−1 for Fe-N-C/CNTs (though the possible graphitization of PANI could also contribute to the increase of BET area). The defects can be the host for the formation of pyrrolic, pyridinic, and graphitic nitrogen functions, which are believed to be the active sites for oxygen reduction reactions.
Furthermore, as shown in SEM images (Fig. 3b), CNTs were entirely covered by PANI in the PANI/CNTs system as compared to the starting material (Fig. 3a). This may be due to the high ratio of PANI to CNTs used (5.0 g aniline precursors for 1.0 g CNTs) and as a result of high polymerization percentage (based on weight change, polymerization of aniline was almost 100%). After heat treatment in 10% ammonia (Fig. 3c), the system has the morphology of a composite, in accordance with XRD and Raman data. Also, it is noticeable that a significant morphology difference was observed between Fe-N-C/CNTs and PANI/CNTs. Based on these observations, it is assumed that the outer layer of CNTs reacted with PANI moieties under the heat treatment in ammonia forming a kind of carbon composite with increased surface area. TEM images were taken to further analyze the structure of Fe-N-C/CNTs. As shown in Fig. 3d, a carbon composite (graphene-like) structure was observed, consistent with the SEM image. However, a closer inspection (inset of Fig. 3d) revealed remaining of a few layers of multiwall carbon nanotubes. Iron nanoparticles were also observed in Fig. 3d, consistent with the XRD data in Fig. 1. The electrochemical behaviors of the obtained catalyst, Fe-NC/CNTs, were studied using a thin film rotating disk electrode (TFRDE) technique. First, the effect of catalyst loading on ORR activity was checked. When the catalyst loading was 0.4 mg cm−2 or lower, the coverage of the catalysts on the RDE tip was not good and the limiting current was ill-defined in the transport controlled region (Fig. 4a). When the catalyst loading increased to 1.2 mg cm−2 , the limiting current was much better defined. Oxygen reduction may occur via an indirect pathway involving the formation of H2 O2 , the reduction of which depends strongly on catalyst loading (Fig. 4b). It was proposed that the increase in catalyst loading led to more
Fig. 3. SEM images of CNTs (a), PANI/CNTs mixed with FeCl3 (b), Fe-N-C/CNTs (c) and TEM image of Fe-N-C/CNTs (d). The inset of (d) is a high resolution TEM image showing graphitic layers. The scale bar in the inset of (d) is 5 nm.
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
129
Fig. 4. Polarization curves of Fe-N-C/CNTs in O2 saturated 0.1 M HClO4 solution with various catalyst loading at 400 rpm. Scanning rate = 5 mV s−1 . (a) Disk current densities and (b) ring current densities.
active sites in the electrode, and as a result, the formed H2 O2 could not escape the catalyst layer before it got reduced to water [20], as illustrated in Fig. 5. The catalyst loading of 1.2 mg cm−2 was chosen in this study. At this loading the thinknees of the film was estimaed to be 7–8 m by SEM. Prior to ORR activity evaluation, the electrode was cycled in a nitrogen saturated 0.1 M HClO4 solution until a stabilized cyclic voltamogramm (CV) was attained (Fig. 6a, black line). It shows a capacitive envelope between 0 and 1.2 V with a pair of wide redox peaks, which are almost symmetrical at 0.6 V. Such a behavior is similar to that reported elsewhere for these types of catalysts, FeN/C [17]. However, in our system, the redox peaks, which usually assigned to N-Fe(II)/N-Fe(III) are not well pronounced, maybe due to the large capacitive current as a result of high BET area of Fe-NC/CNTs. The increase of BET area during heat treatment may be due to the formation of defects on CNTs, and polyaniline moieties fused on CNTs formed by re-arrangement of aromatic containing nitrogen atoms. Heat treatment may also cause creation of pores due to carbonization of PANI-FeCl3 and degasification, such as evolution of Cl2 , CO2 , CO, and HCl. It is important to mention that cathode material with such a high capacitive current is beneficial to fuel cell operations by acting as a charge storage tank that can reduce cathode voltage excursion during fuel starvation and start-up condition [21]. It is also worth noting that despite the high surface area of Fe-N-C/CNTs (540 m2 g−1 ), the system is very stable. The
Faradic process due to oxidation of catalyst composite appeared negligible up to 1.2 V. Moreover, the CV did not show any significant change after 20,000 cycles between 0.65 and 1.0 V in oxygen saturated 0.1 M HClO4 and 0.1 M KOH solutions. The unchanged capacitive charge clearly indicates the surface resistance of nitrogen doped carbon to corrosion under potential cycling for 20,000 cycles (Fig. 6a, red line). Note that the free iron nanoparticles dissolved quickly in acid. The oxygen reduction activity of Fe-N-C/CNTs was evaluated in an oxygen saturated 0.1 M HClO4 solution with a catalyst loading of 1.2 mg cm−2 . The polarization curves are shown in Fig. 4b. The onset potential of ORR for the fresh Fe-N-C/CNTs is about 0.81 V (Fig. 6b, black line), and is comparable to that of the state-of-theart non-precious metal catalysts in the literatures [12]. In addition to the excellent activity, as compared to that of FeNx-CNT reported eslsewhere [22], the other advantage of this nitrogen doped carbon composite is its stability. The electode was potential cycled between 0.65 and 1.0 V using square-wave signal of 5 s at each potential. ORR activity was evaluated in a fresh 0.1 M HClO4 solution after every 5000 cycles. After 20, 000 cycles (Fig. 6b, red line), the half-wave potential of ORR decreased from 0.666 V for a fresh electode to 0.626 V. By tracking the activity evolution with cycling, it was found that almost all the activity loss occurred during the first 5000 cycles with a degradation rate of 8 V per cycle in half-wave potential. Further activity degradation in the last 15,000 cycles was
Fig. 5. Illustrations of catalyst loading effect on reduction of H2 O2 formed during partial reduction of oxygen in acid medium.
130
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
Fig. 6. Voltammetry curves of Fe-N-C/CNTs in N2 saturated 0.1 M HClO4 (a), 0.1 M KOH (c), 20 mV s−1 and in O2 saturated 0.1 M HClO4 (b), 0.1 M KOH (d), 5 mV s−1 , 400 rpm. Before (black line) and after 20,000 potential cycles (red line) between 0.65 and 1.0 V. Room temperature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
basically very negligible. For such a long cycling the degradation rate of our catalyst is much lower comparing to any other NPM catalysts supported on high surface area carbon. One possible reason for the stability may be due to the stable nitrogen group in Fe-N-C/CNTs. It has been reported elsewhere that there are three types of nitrogen that can be formed in nitrogen-doped carbon. Pyridinic and pyrrolic nitrogen are venerable to protonation in acidic medium and the highly stable nitrogen is the one in quaternary form better known as graphitic nitrogen [23–26]. Therefore, to develop a stable and highly active non-precious catalyst for ORR, it is very important to tailor synthesis method in order to achieve a catalyst with high graphitic nitrogen content. To this extend, coating CNTs with polyaniline and heat treatment in ammonia might be an effective method of increasing both ORR activity and stability by enhancing total nitrogen content and in particular nitrogen in graphitic form. XPS data illustrated in Fig. 7 was used to analyze the nitrogen containing environment of Fe-N-C/CNTs and to show the three types of nitrogen functions mentioned above. The N 1s signal measured by XPS can be deconvoluted into peaks at binding energies, 398.6, 400.5, 401.3, and 402 eV, which correspond to N-pyridinic, N-pyrrolic, N-graphitic, and N-oxide respectively, in agreement with work reported in literature [17,27]. The total atomic percentage of nitrogen was found to be about 2.6%, among it 27% as N-pyridinic (0.7% atom) and 20% as N-graphitic (0.5% atom). Furthermore, the high ORR activity of Fe-N-C/CNTs was also demonstrated in alkaline medium (0.1 M KOH). The onset potential of ORR for the fresh Fe-N-C/CNTs is about 0.94 V/RHE (Fig. 6d, black line), i.e., 130 mV higher than that in acidic solution. The higher ORR activitiy in alkaline solution is due to the formation of stable reaction intermediate HO2 − on the active sites associated with N-Fe(II) and the higher redox potential of N-Fe(II)/N-Fe(III) at high pH [28]. In addition to its high ORR activity, Fe-N-C/CNTs is believed to be more stable than in acidic medium. Effectively, after 20,000 potential cycles (0.65–1.0 V), the half-wave potential shifts by 25 mV to
a positive direction, i.e., the mass activity increases by more than twice instead of decreasing. The increase of the activity may be due to formation of iron oxides. Recently, some studies have demonstrated that cobalt oxides and carbon composites showed excellent ORR activity in alkaline medium [29,30]. Similarly, iron oxides and carbon composites formed after potential cycling are expected to show high ORR activity. EDS analysis after cycling did reveal the presence of Fe in the electrode after potential cycling in alkaline. While in the electrode after cycling in acid, it was difficult to detect any trace of Fe indicating that most Fe was dissolved in acid. The reason for the decrease of limiting current by 15% after cycling in Fig. 6d has not been fully understood.
Fig. 7. XPS spectrum of Fe-N-C/CNTs catalyst showing types of nitrogen functions.
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
131
Innovation Plan (NSTIP). Authors also acknowledge the technical suport provided by the Center of Nanotechnology (CENT). References
Fig. 8. Voltammetry curves of Fe-N-C/CNTs in O2 saturated 0.1 M HClO4 , with (red) and without (black) 0.1 M CH3 OH. 5 mV s−1 , 400 rpm, room temperature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
For direct methanol fuel cells applications (DMFCs), having a catalyst that is active toward ORR and inert for methanol oxidation is a challenge. Due to cross-over of methanol through membrane, Pt based catalysts may not be the right choice. Thus, much attention has been shifted toward non-precious metals and in particular nitrogen-doped carbon. However, almost all the studies were conducted in alkaline medium and on fresh electrode [27,31–34]. In this regard, we conducted the methanol tolerance test in acid medium on the electrode already cycled for 20,000 cycles in oxygen (Fig. 4b). Fig. 8 illustrates that no noticeable change in the ORR polarization curves in the 0.1 M HClO4 solution with and without 0.1 M CH3 OH, suggesting the excellent methanol tolerance of Fe-NC/CNTs, in agreement with the data reported by Y. Li et al. [35]. We deliberately did the test of using a cycled electrode in order to make sure that the degraded catalyt still maintains the methanol tolerance properties. The fresh Fe-N-C/CNTs showed similar excellent methanol tolerance (data not shown). 4. Conclusions In summary, the Fe-N-C/CNTs showed excellent ORR activity, which is comparable to the best NPM catalysts reported in the literatures. The high activity can be attributed to the increased active sites that may be generated by polymerization of PANI and formation of nitrogen doped carbon composite during heat treatment at 900 ◦ C in NH3 . For the first time, a comparative study of catalyst stability for long cycles was evaluated in both acidic and alkaline mediums. The reasons for the stability of the composite, Fe-NC/CNTs, have not been fully understood and may require further study. The methanol tolerance of Fe-N-C/CNTs catalysts in acidic medium before and after cycling were tested and showed no activity loss with methanol in the solution. We believe that Fe-N-C/CNTs as a Pt-free catalyst, developed in this work is a promising low-cost cathode material for DMFCs due to its excellent methanol tolerance as well as for AFCs. Acknowledgements The authors would like to acknowledge the support provided by King Abdulaziz City for Science and Technology (KACST) through the Science & Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) for funding this work through project No. 10-ENE-1375-04 as part of the National Science, Technology and
[1] W. Vielstich, A. Lamm, H.A. Gasteiger, Handbook of Fuel Cells: Fundamentals, Technology, Applications, John Wiley & Sons, Chichester, 2003. [2] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Activity benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs, Applied Catalysis B 56 (2005) 9–35. [3] M.K. Debe, Electrocatalyst approaches and challenges for automotive fuel cells, Nature 486 (2012) 43–51. [4] H.A. Gasteiger, N.M. Markovic, Just a dream-or future reality, Science 324 (2009) 48–49. [5] F.T. Wagner, B. Lakshmanan, M.F. Mathias, Electrochemistry and the Future of the Automobile, The Journal of Physical Chemistry Letters 1 (2010) 2204–2219. [6] M. Shao, Electrocatalysis in Fuel Cells: A Non- and Low-Platinum Approach, Springer, London, 2013. [7] Z. Chen, D. Higgins, A. Yu, L. Zhang, J. Zhang, A review on non-precious metal electrocatalysts for PEM fuel cells, Energy & Environmental Science 4 (2011) 3167–3192. [8] M. Lefevre, E. Proietti, F. Jaouen, J.P. Dodelet, Iron-based catalysts with improved oxygen reduction activity in polymer electrolyte fuel cells, Science 324 (2009) 71–74. [9] E. Proietti, F. Jaouen, M. Lefèvre, N. Larouche, J. Tian, J. Herranz, J.-P. Dodelet, Iron-based cathode catalyst with enhanced power density in polymer electrolyte membrane fuel cells, Nature Communications 2 (2011) 416. [10] G. Wu, K.L. More, C.M. Johnston, P. Zelenay, High-Performance electrocatalysts for oxygen reduction derived from polyaniline, Iron, and Cobalt, Science 322 (2011) 443–447. [11] L. Elbaz, G. Wu, P. Zelenay, Heat-treated non-precious-metal-based catalysts for oxygen reduction, in: M. Shao (Ed.), Electrocatalysis in Fuel Cells: A Nonand Low-Platinum Approach, Springer, London, 2013. [12] J.P. Dodelet, The controversial role of the metal in Fe- or Co-based electrocatalysts for the oxygen reduction reaction in acid medium, in: M. Shao (Ed.), Electrocatalysis in Fuel Cells: A Non- and Low-Platinum Approach, Springer, London, 2013. [13] P. Zelenay, DOE Hydrogen Program Annual Merit Review and Peer Evaluation Meeting, Washington, DC, 2010. [14] T. Schilling, M. Bron, Oxygen reduction at Fe–N-modified multi-walled carbon nanotubes in acidic electrolyte, Electrochimica Acta 53 (2008) 5379–5385. [15] H.R. Byon, J. Suntivich, E.J. Crumlin, Y. Shao-Horn, Fe–N-modified multi-walled carbon nanotubes for oxygen reduction reaction in acid, Physical Chemistry Chemical Physics 13 (2011) 21437–21445. [16] A. Morozan, P. Jegou, B. Jousselme, S. Palacin, Electrochemical performance of annealed cobalt–benzotriazole/CNTs catalysts towards the oxygen reduction reaction, Physical Chemistry Chemical Physics 13 (2011) 21600–21607. [17] C.-W. Tsai, H.M. Chen, R.-S. Liu, K. Asakura, L. Zhang, J. Zhang, M.-Y. Lo, Y.-M. Peng, Carbon incorporated FeN/C electrocatalyst for oxygen reduction enhancement in direct methanol fuel cells: X-ray absorption approach to local structures, Electrochimica Acta 56 (2011) 8734–8738. [18] P. Ayala, A. Gruneis, C. Kramberger, M.H. Rummeli, I.G. Solorzano, J.F.L. Freire, T. Pichler, Effects of the reaction atmosphere composition on the synthesis of single and multiwalled nitrogen-doped nanotubes, The Journal of Chemical Physics 127 (2007) 184709. [19] F. Tuinstra, J.L. Koenig, Raman Spectrum of Graphite, The Journal of Chemical Physics 53 (1970) 1126–1130. [20] A.V. Palenzuela, L. Zhang, L. Wang, P.L. Cabot, E. Brillas, K. Tsay, J. Zhang, Fe–Nx/C electrocatalysts synthesized by pyrolysis of Fe(II)–2,3,5,6-tetra(2pyridyl)pyrazine complex for PEM fuel cell oxygen reduction reaction, Electrochimica Acta 56 (2011) 4744–4752. [21] C.A. Reiser, L. Bregoli, T.W. Patterson, J.S. Yi, J.D. Yang, M.L. Perry, T.D. Jarvi, A Reverse-Current Decay Mechanism for Fuel Cells Batteries, Fuel, Cells, and Energy Conversion, Electrochemical and Solid-State Letters 8 (2005) A273. [22] Th. Palaniselvam, R. Kannan, S. Kurungo, Facile construction of non-precious iron nitride-doped carbon nanofibers as cathode electrocatalysts for proton exchange membrane fuel cells, Chemical Communications 47 (2011) 2910–2912. [23] H.-S. Oh, J.-G. Oh, B. Roh, I. Hwang, H. Kim, Development of highly active and stable non-precious oxygen reduction catalysts for PEM fuel cells using polypyrrole and a chelating agent, Electrochemistry Communications 13 (2011) 879–881. [24] G. Liu, X. Li, P. Ganesan, B.N. Popov, Studies of oxygen reduction reaction active sites and stability of nitrogen-modified carbon composite catalysts for PEM fuel cells, Electrochimica Acta 55 (2010) 2853–2858. [25] G. Liu, X. Li, J.W. Lee, B.N. Popov, A review of the development of nitrogenmodified carbon-based catalysts for oxygen reduction at USC, Catalysis Science & Technology 1 (2011) 207–217. [26] G. Li, G. Liu, B. Popov, Activity and stability of non-precious metal catalysts for oxygen reduction in acid and alkaline electrolytes, Journal of Power Sources 195 (2010) 6373–6378. [27] G. Ma, R. Jia, J. Zhao, Z. Wang, C. Song, S. Jia, Z. Zhu, Nitrogen-Doped Hollow Carbon Nanoparticles with Excellent Oxygen Reduction Performances and Their Electrocatalytic Kinetics, Journal of Physical Chemistry C 115 (2011) 25148–25154.
132
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
[28] N. Ramaswamy, S. Mukerjee, Fundamental Mechanistic Understanding of Electrocatalysis of Oxygen Reduction on Pt and Non-Pt Surfaces: Acid versus Alkaline Media, Advances in Physical Chemistry (2012), http://dx.doi.org/10.1155/2012/491604. [29] Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier, H. Dai, Co3 O4 nanocrystals on graphene as a synergistic catalyst for oxygen reduction reaction, Nature Materials 10 (2011) 780–786. [30] Y. Liang, H. Wang, P. Diao, W. Chang, G. Hong, Y. Li, M. Gong, L. Xie, J. Zhou, J. Wang, T.Z. Regier, F. Wei, H. Dai, Oxygen Reduction Electrocatalyst Based on Strongly Coupled Cobalt Oxide Nanocrystals and Carbon Nanotubes, Journal of the American Chemical Society 134 (2012) 15849–15857. [31] K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Nitrogen-doped carbon nanotube arrays with high electrocatalytic activity for oxygen reduction, Science 323 (2009) 760–764.
[32] B. Piela, T.S. Olson, P. Atanassov, P. Zelenay, Highly methanol-tolerant nonprecious metal cathode catalysts for direct methanol fuel cell, Electrochimica Acta 55 (2010) 7615–7621. [33] L.T. Qu, Y. Liu, J.B. Baek, L. Dai, Nitrogen-doped graphene as efficient metalfree electrocatalyst for oxygen reduction in fuel cells, ACS Nano 4 (2010) 1321–1326. [34] Z. Wang, R. Jia, J. Zheng, J. Zhao, L. Li, J. Song, Z. Zhu, Nitrogen-promoted selfassembly of N-doped carbon nanotubes and their intrinsic catalysis for oxygen reduction in fuel cells, ACS Nano 5 (2011) 1677–1684. [35] Y. Li, W. Zhou, H. Wang, L. Xie, Y. Liang, F. Wei, J.-C. Idrobo, S.J. Pennycook, H. Dai, An oxygen reduction electrocatalyst based on carbon nanotube-graphene complexes, Nature Nanotechnology 7 (2012) 394–400.
Contents lists available at ScienceDirect
Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta
A Pt-free catalyst for oxygen reduction reaction based on Fe–N multiwalled carbon nanotube composites Belabbes Merzougui a,∗∗ , Abdouelilah Hachimi a , Akeem Akinpelu a , Saheed Bukola a , Minhua Shao b,∗ a b
Center of Nanotechnology (CENT) & Department of Chemistry, King Fahd University of Petroleum and Minerals, Dhahran 31261, Saudi Arabia UTC Power, South Windsor, CT 06074, USA
a r t i c l e
i n f o
Article history: Received 28 January 2013 Received in revised form 7 June 2013 Accepted 7 June 2013 Available online 17 June 2013 Keywords: Non precious metal catalysts Oxygen reduction reaction Durability Methanol tolerance
a b s t r a c t An iron-based non-precious metal (NPM) catalyst was synthesized by chemically coating a polymer containing nitrogen, such as polyaniline (PANI), on multiwalled carbon nanotubes (MWCNTs), followed by iron complexation and heat treatment in a decomposable source of nitrogen (ammonia) at high temperature. Its excellent activities for oxygen reduction reaction (ORR) in both acidic and alkaline media were confirmed by thin film rotating disk electrode technique. Also, durability test was performed in both media. After 20,000 (0.65–1.0 V) potential cycles, the half-wave potential decayed by only 40 mV in 0.1 M HClO4 , while in 0.1 M KOH it gained 25 mV. In addition, no activity decay was observed by adding methanol in the acidic solution indicating its excellent tolerance to methanol oxidation. © 2013 Elsevier Ltd. All rights reserved.
1. Introduction Cost and durability are considered as the main challenges facing the commercialization in mass scale of polymer electrolyte fuel cells (PEFCs) [1]. A large portion of PEFC systems cost is still associated with cathode catalyst for oxygen reduction reaction (ORR) [2–4]. Nevertheless, great efforts have been so far made to develop advanced ORR catalysts with lower cost and higher activity and durability aiming to replace a substantial amount of the costly Pt-based materials, which are commonly used in today’s fuel cells technologies [3,5,6]. Among them, non-precious metal (NPM) catalysts consisting of transition metals (in particular, Fe and Co)–nitrogen complexes on carbon support prepared by high temperature pyrolysis are the most promising approaches [7–11]. Despite the fact that the nature of the catalytic sites in NPM catalysts are still as a topic of debate, it is generally accepted that ORR activity and durability are largely influenced by the type of the transition metals (M) in the catalyst, the type of C and N precursors used, the ratios of M/N and N/C on the surface, and the synthesis conditions [12]. For example, an excellent ORR activity has been reported by the Dodelet group by mixing a carbon support with iron precursor, and then heating the materials in ammonia as the
source of nitrogen [8]. On the other hand, a similar activity has been achieved by the Zelenay group by polymerizing the aniline (nitrogen precursor) on high surface area carbon support in the presence of iron precursor, and then heating the mixtures in an inert gas [10]. The same group and others explored NPM catalysts supported on multiwalled carbon nanotubes (MWCNTs) [13–16]. To the best of our knowledge, the approach of chemically coating polymer containing nitrogen, such as polyaniline (PANI) on CNTs and heat treatment in a decomposable source of nitrogen, such as ammonia in order to increase the total nitrogen content (aiming to increase the graphitic nitrogen) merits to be studied. Moreover, none of the previous reported works has performed comparative studies of the activity and durability under long cycling in both alkaline and acidic media. Herein, we report an iron-based NPM catalyst through polymerization of aniline on CNTs and subjecting to heat treatment in ammonia at 900 ◦ C. The resulted material shows excellent ORR activity and durability under potential cycling in both acidic and alkaline media. It is important to note that cycling in acidic medium caused a drop in iron content due to dissolution. Yet, the activity of the catalyst did not decay dramatically as one can expect. 2. Experimental
∗ Corresponding author. Tel.: +966 38604378. ∗∗ Corresponding author. E-mail addresses: [email protected] (B. Merzougui), [email protected] (M. Shao). 0013-4686/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.electacta.2013.06.016
2.1. Synthesis of Fe-N-C/CNTs Multiwall carbon nanotubes (CNTs) were purchased from Cheap Tubes Incorporation, USA (http://www.cheaptubesinc.com/) and
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
were acid washed in-house before use. The BET area of the acid washed CNTs was in the range of 170 m2 g−1 . Aniline (Aldrich) was distilled before use. Ammonium persulfate {(NH4 )2 S2 O8 } (Aldrich) was used without further purification. MilliQ UV-plus water (Millipore) was used throughout all experiments. In a typical experiment, 5.0 g of aniline was mixed with 50 ml water. Thereafter, 37% HCl was added dropwise with continuous stirring until aniline became completely miscible with water. Then, 1.0 g of CNTs was added into the mixture and sonicated for 0.5 hr in a sonication bath to obtain well dispersed slurry and kept overnight under stirring to impregnate aniline into CNTs surface. As an oxidant agent, 10.0 g of {(NH4 )2 S2 O8 } was dissolved in water and added in dropwise to the mixture and kept overnight under stirring to ensure proper and complete polymerization. The obtained product, named as PANI/CNTs was then filtrated and washed with Di-water for several times until pH became less acidic (pH = 6), then dried under vacuum at 80 ◦ C overnight. In the next step, 1.0 g of the dried PANI/CNTs was dispersed in water in form of slurry using ultrasonication. On the other hand, 0.48 g of FeCl3 ·6H2 O (to give 10 wt.% Fe) was dissolved in water and added to the slurry. The mixture was then kept on a magnetic stirrer overnight to ensure complexation of FeCl3 onto PANI/CNTs. Before heat treatment, two methods were tried to dry the mixture: oven drying and freeze drying. Based on the ORR measurements, freeze drying method was found to lead to a higher activity. Freeze drying might help keep better dispersion or complexation of FeCl3 on PANI/CNTs. Freeze drying was performed with a Freeze Dryer (Labconco) for 3 days. The final product was then heat treated in 10% NH3 /N2 at 900 ◦ C for 1 hr (3 hr for ramping temperature from 25 ◦ C to 900 ◦ C). The resulted catalyst was denoted as Fe-N-C/CNTs. 2.2. Electrochemical measurements For ORR activity measurement, approximately 5 mg of the dried catalyst Fe-N-C/CNTs was dispersed in a mixture of water and isopropanol alcohol (30%, v/v) and 60 l of 1.66 wt.% Nafion® (prepared from 5 wt.%, Aldrich). The mixture was ultrasonicated for 10–20 min to obtain a uniform ink. Then, 20 l of the ink suspension was deposited on the pre-cleaned glassy carbon substrate (5 mm diameter, Pine Instruments) and allowed to dry. For high loading of catalyst, we repeated the operation for several times to achieve the desired loadings. A Ag–AgCl electrode (calibrated and converted to RHE) and Pt mesh were used as a reference and counter electrode, respectively. All the potential reported in this paper were converted to RHE. Before testing ORR activity, each electrode was cycled in nitrogen saturated 0.1 M HClO4 and 0.1 M KOH solutions until a stable CV was obtained (generally for 10–15 cycles between 0 and 1.2 V at 20 mV s−1 ). ORR activity was evaluated in both oxygen saturated acidic (0.1 M HClO4 ) and alkaline (0.1 M KOH) media using a rotating disk electrode (RDE) technique at rotation speed of 400 rpm with a scan rate of 5 mV s−1 . Catalysts durability was also investigated in oxygen saturated 0.1 M HClO4 and 0.1 M KOH under potential cycling between 0.65 and 1.0 V for 20,000 cycles using square-wave signal of 5 s at each potential. ORR activity was evaluated in fresh electrolytes in a separate cell after every 5000 cycles. After 20,000 cycles, the ORR activity was measured again in 0.1 M HClO4 containing 0.1 M methanol to evaluate its methanol tolerance. All the currents were normalized to the geometric area of electrodes. The ring current densities in the RRDE measurements were corrected according to the collection efficiency (30%) of the system. 2.3. Characterization High-end field-emission scanning electron microscope (FESEM, Tescan-Lyra-3) and energy-dispersive X-ray spectroscopy (EDX,
127
Fig. 1. XRD patterns of raw CNTs, PANI/CNTs and Fe-N-C/CNTs (obtained after HT at 900 ◦ C in 10% NH3 /N2 of FeCl3 /PANI/CNTs). The S in FeS and Ni in NiC of Fe-NC/CNTs came from persulfate used for polymerization of aniline and impurity in the raw CNTs (The EDX spectrum showed that Raw-CNTs contain 99 wt.% of carbon and trace of oxygen and nickel.).
Oxford-Xmax) were used to provide morphology and elemental composition information. TEM analysis was performed with FEI ˚ Titan ST operated at 300 kV with a point resolution of 2.0 A. X-ray diffractions (XRD) patterns were collected using a Rigaku ˚ Miniflex II instrument with a monochromator of CuK␣1 (1.5406 A) at 30 kV, 15 mA. The XRD patterns were recorded in the static scanning mode from 5◦ to 60◦ (2) at a detector angular speed of 2◦ min−1 and step size of 0.02◦ . Peak matching was determined with Philips X’Pert plus (V.2) using patterns diffractions files (PDF-2) from ICDD (2011). Raman spectra were taken on an iHR320 Horiba Spectrometer with charge-coupled device (CCD) using monochromatic laser (300 mW, 532 nm), grating of 1200 lines mm−1 and an aluminum substrate. Raman spectroscopy provided further information on the defects in CNTs that may be created during heat treatment in NH3 . BET area measurements were conducted using ASAP-2020 Physisorption Analyzer from Micrometrics. 3. Results and discussion Fig. 1 shows the XRD patterns of as received CNTs, PANI/CNTs, and Fe-N-C/CNTs. The similar intensity of (0 0 2) diffraction peak of graphitic carbon at ∼26◦ before and after polymerization of aniline on CNTs suggests that the polymerization did not cause a noticeable structure change in CNTs. The arising of a broader shoulder at ∼20◦ due to PANI in PANI/CNTs confirms the successful polymerization. However, when the system was subjected to complexation with FeCl3 and followed by heat treatment in ammonia at 900 ◦ C, a composite-like material consisting of Fe–N and Fe–C function groups was formed. The significant decrease in peak intensity at 26◦ also suggests the formation of nitrogen-doped carbon as a result of sacrifice of graphitic structure. A similar XRD pattern for Fe-N-C catalyst was reported by Tsai et al. [17]. The sharp diffraction peak at 45◦ in Fig. 1 may be due to free Fe nanoparticles. The structure changes of CNTs are also supported by the Raman spectroscopic data (Fig. 2). The peaks centered at 1334 cm−1 and 1596 cm−1 are referred to D- and G-band, respectively [18]. The D-band is related to the vibrations of sp3 bonds associated to the bonded carbon atoms of disordered graphite, while the G-band is related to the vibration of all sp2 bonds in well-graphitized structures [19]. The intensity ratio of the D- to G-band (D/G) interprets the degree of disorder that could occur in the carbon structure. The D/G ratios of raw-CNTs and Fe-N-C/CNTs catalyst were estimated to be 0.37 and
128
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
Fig. 2. Raman spectra of as-received CNTs (blue) and Fe-N-C/CNTs catalyst (red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
0.96, respectively. The higher D/G ratio after heat treatment suggests a significant increase in structure defects (though the addition of sp3 carbon structure cannot be totally ruled out). The creation of structure defects can be further supported by the increase in BET area from 170 m2 g−1 for CNTs to 540 m2 g−1 for Fe-N-C/CNTs (though the possible graphitization of PANI could also contribute to the increase of BET area). The defects can be the host for the formation of pyrrolic, pyridinic, and graphitic nitrogen functions, which are believed to be the active sites for oxygen reduction reactions.
Furthermore, as shown in SEM images (Fig. 3b), CNTs were entirely covered by PANI in the PANI/CNTs system as compared to the starting material (Fig. 3a). This may be due to the high ratio of PANI to CNTs used (5.0 g aniline precursors for 1.0 g CNTs) and as a result of high polymerization percentage (based on weight change, polymerization of aniline was almost 100%). After heat treatment in 10% ammonia (Fig. 3c), the system has the morphology of a composite, in accordance with XRD and Raman data. Also, it is noticeable that a significant morphology difference was observed between Fe-N-C/CNTs and PANI/CNTs. Based on these observations, it is assumed that the outer layer of CNTs reacted with PANI moieties under the heat treatment in ammonia forming a kind of carbon composite with increased surface area. TEM images were taken to further analyze the structure of Fe-N-C/CNTs. As shown in Fig. 3d, a carbon composite (graphene-like) structure was observed, consistent with the SEM image. However, a closer inspection (inset of Fig. 3d) revealed remaining of a few layers of multiwall carbon nanotubes. Iron nanoparticles were also observed in Fig. 3d, consistent with the XRD data in Fig. 1. The electrochemical behaviors of the obtained catalyst, Fe-NC/CNTs, were studied using a thin film rotating disk electrode (TFRDE) technique. First, the effect of catalyst loading on ORR activity was checked. When the catalyst loading was 0.4 mg cm−2 or lower, the coverage of the catalysts on the RDE tip was not good and the limiting current was ill-defined in the transport controlled region (Fig. 4a). When the catalyst loading increased to 1.2 mg cm−2 , the limiting current was much better defined. Oxygen reduction may occur via an indirect pathway involving the formation of H2 O2 , the reduction of which depends strongly on catalyst loading (Fig. 4b). It was proposed that the increase in catalyst loading led to more
Fig. 3. SEM images of CNTs (a), PANI/CNTs mixed with FeCl3 (b), Fe-N-C/CNTs (c) and TEM image of Fe-N-C/CNTs (d). The inset of (d) is a high resolution TEM image showing graphitic layers. The scale bar in the inset of (d) is 5 nm.
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
129
Fig. 4. Polarization curves of Fe-N-C/CNTs in O2 saturated 0.1 M HClO4 solution with various catalyst loading at 400 rpm. Scanning rate = 5 mV s−1 . (a) Disk current densities and (b) ring current densities.
active sites in the electrode, and as a result, the formed H2 O2 could not escape the catalyst layer before it got reduced to water [20], as illustrated in Fig. 5. The catalyst loading of 1.2 mg cm−2 was chosen in this study. At this loading the thinknees of the film was estimaed to be 7–8 m by SEM. Prior to ORR activity evaluation, the electrode was cycled in a nitrogen saturated 0.1 M HClO4 solution until a stabilized cyclic voltamogramm (CV) was attained (Fig. 6a, black line). It shows a capacitive envelope between 0 and 1.2 V with a pair of wide redox peaks, which are almost symmetrical at 0.6 V. Such a behavior is similar to that reported elsewhere for these types of catalysts, FeN/C [17]. However, in our system, the redox peaks, which usually assigned to N-Fe(II)/N-Fe(III) are not well pronounced, maybe due to the large capacitive current as a result of high BET area of Fe-NC/CNTs. The increase of BET area during heat treatment may be due to the formation of defects on CNTs, and polyaniline moieties fused on CNTs formed by re-arrangement of aromatic containing nitrogen atoms. Heat treatment may also cause creation of pores due to carbonization of PANI-FeCl3 and degasification, such as evolution of Cl2 , CO2 , CO, and HCl. It is important to mention that cathode material with such a high capacitive current is beneficial to fuel cell operations by acting as a charge storage tank that can reduce cathode voltage excursion during fuel starvation and start-up condition [21]. It is also worth noting that despite the high surface area of Fe-N-C/CNTs (540 m2 g−1 ), the system is very stable. The
Faradic process due to oxidation of catalyst composite appeared negligible up to 1.2 V. Moreover, the CV did not show any significant change after 20,000 cycles between 0.65 and 1.0 V in oxygen saturated 0.1 M HClO4 and 0.1 M KOH solutions. The unchanged capacitive charge clearly indicates the surface resistance of nitrogen doped carbon to corrosion under potential cycling for 20,000 cycles (Fig. 6a, red line). Note that the free iron nanoparticles dissolved quickly in acid. The oxygen reduction activity of Fe-N-C/CNTs was evaluated in an oxygen saturated 0.1 M HClO4 solution with a catalyst loading of 1.2 mg cm−2 . The polarization curves are shown in Fig. 4b. The onset potential of ORR for the fresh Fe-N-C/CNTs is about 0.81 V (Fig. 6b, black line), and is comparable to that of the state-of-theart non-precious metal catalysts in the literatures [12]. In addition to the excellent activity, as compared to that of FeNx-CNT reported eslsewhere [22], the other advantage of this nitrogen doped carbon composite is its stability. The electode was potential cycled between 0.65 and 1.0 V using square-wave signal of 5 s at each potential. ORR activity was evaluated in a fresh 0.1 M HClO4 solution after every 5000 cycles. After 20, 000 cycles (Fig. 6b, red line), the half-wave potential of ORR decreased from 0.666 V for a fresh electode to 0.626 V. By tracking the activity evolution with cycling, it was found that almost all the activity loss occurred during the first 5000 cycles with a degradation rate of 8 V per cycle in half-wave potential. Further activity degradation in the last 15,000 cycles was
Fig. 5. Illustrations of catalyst loading effect on reduction of H2 O2 formed during partial reduction of oxygen in acid medium.
130
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
Fig. 6. Voltammetry curves of Fe-N-C/CNTs in N2 saturated 0.1 M HClO4 (a), 0.1 M KOH (c), 20 mV s−1 and in O2 saturated 0.1 M HClO4 (b), 0.1 M KOH (d), 5 mV s−1 , 400 rpm. Before (black line) and after 20,000 potential cycles (red line) between 0.65 and 1.0 V. Room temperature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
basically very negligible. For such a long cycling the degradation rate of our catalyst is much lower comparing to any other NPM catalysts supported on high surface area carbon. One possible reason for the stability may be due to the stable nitrogen group in Fe-N-C/CNTs. It has been reported elsewhere that there are three types of nitrogen that can be formed in nitrogen-doped carbon. Pyridinic and pyrrolic nitrogen are venerable to protonation in acidic medium and the highly stable nitrogen is the one in quaternary form better known as graphitic nitrogen [23–26]. Therefore, to develop a stable and highly active non-precious catalyst for ORR, it is very important to tailor synthesis method in order to achieve a catalyst with high graphitic nitrogen content. To this extend, coating CNTs with polyaniline and heat treatment in ammonia might be an effective method of increasing both ORR activity and stability by enhancing total nitrogen content and in particular nitrogen in graphitic form. XPS data illustrated in Fig. 7 was used to analyze the nitrogen containing environment of Fe-N-C/CNTs and to show the three types of nitrogen functions mentioned above. The N 1s signal measured by XPS can be deconvoluted into peaks at binding energies, 398.6, 400.5, 401.3, and 402 eV, which correspond to N-pyridinic, N-pyrrolic, N-graphitic, and N-oxide respectively, in agreement with work reported in literature [17,27]. The total atomic percentage of nitrogen was found to be about 2.6%, among it 27% as N-pyridinic (0.7% atom) and 20% as N-graphitic (0.5% atom). Furthermore, the high ORR activity of Fe-N-C/CNTs was also demonstrated in alkaline medium (0.1 M KOH). The onset potential of ORR for the fresh Fe-N-C/CNTs is about 0.94 V/RHE (Fig. 6d, black line), i.e., 130 mV higher than that in acidic solution. The higher ORR activitiy in alkaline solution is due to the formation of stable reaction intermediate HO2 − on the active sites associated with N-Fe(II) and the higher redox potential of N-Fe(II)/N-Fe(III) at high pH [28]. In addition to its high ORR activity, Fe-N-C/CNTs is believed to be more stable than in acidic medium. Effectively, after 20,000 potential cycles (0.65–1.0 V), the half-wave potential shifts by 25 mV to
a positive direction, i.e., the mass activity increases by more than twice instead of decreasing. The increase of the activity may be due to formation of iron oxides. Recently, some studies have demonstrated that cobalt oxides and carbon composites showed excellent ORR activity in alkaline medium [29,30]. Similarly, iron oxides and carbon composites formed after potential cycling are expected to show high ORR activity. EDS analysis after cycling did reveal the presence of Fe in the electrode after potential cycling in alkaline. While in the electrode after cycling in acid, it was difficult to detect any trace of Fe indicating that most Fe was dissolved in acid. The reason for the decrease of limiting current by 15% after cycling in Fig. 6d has not been fully understood.
Fig. 7. XPS spectrum of Fe-N-C/CNTs catalyst showing types of nitrogen functions.
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
131
Innovation Plan (NSTIP). Authors also acknowledge the technical suport provided by the Center of Nanotechnology (CENT). References
Fig. 8. Voltammetry curves of Fe-N-C/CNTs in O2 saturated 0.1 M HClO4 , with (red) and without (black) 0.1 M CH3 OH. 5 mV s−1 , 400 rpm, room temperature. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
For direct methanol fuel cells applications (DMFCs), having a catalyst that is active toward ORR and inert for methanol oxidation is a challenge. Due to cross-over of methanol through membrane, Pt based catalysts may not be the right choice. Thus, much attention has been shifted toward non-precious metals and in particular nitrogen-doped carbon. However, almost all the studies were conducted in alkaline medium and on fresh electrode [27,31–34]. In this regard, we conducted the methanol tolerance test in acid medium on the electrode already cycled for 20,000 cycles in oxygen (Fig. 4b). Fig. 8 illustrates that no noticeable change in the ORR polarization curves in the 0.1 M HClO4 solution with and without 0.1 M CH3 OH, suggesting the excellent methanol tolerance of Fe-NC/CNTs, in agreement with the data reported by Y. Li et al. [35]. We deliberately did the test of using a cycled electrode in order to make sure that the degraded catalyt still maintains the methanol tolerance properties. The fresh Fe-N-C/CNTs showed similar excellent methanol tolerance (data not shown). 4. Conclusions In summary, the Fe-N-C/CNTs showed excellent ORR activity, which is comparable to the best NPM catalysts reported in the literatures. The high activity can be attributed to the increased active sites that may be generated by polymerization of PANI and formation of nitrogen doped carbon composite during heat treatment at 900 ◦ C in NH3 . For the first time, a comparative study of catalyst stability for long cycles was evaluated in both acidic and alkaline mediums. The reasons for the stability of the composite, Fe-NC/CNTs, have not been fully understood and may require further study. The methanol tolerance of Fe-N-C/CNTs catalysts in acidic medium before and after cycling were tested and showed no activity loss with methanol in the solution. We believe that Fe-N-C/CNTs as a Pt-free catalyst, developed in this work is a promising low-cost cathode material for DMFCs due to its excellent methanol tolerance as well as for AFCs. Acknowledgements The authors would like to acknowledge the support provided by King Abdulaziz City for Science and Technology (KACST) through the Science & Technology Unit at King Fahd University of Petroleum & Minerals (KFUPM) for funding this work through project No. 10-ENE-1375-04 as part of the National Science, Technology and
[1] W. Vielstich, A. Lamm, H.A. Gasteiger, Handbook of Fuel Cells: Fundamentals, Technology, Applications, John Wiley & Sons, Chichester, 2003. [2] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Activity benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs, Applied Catalysis B 56 (2005) 9–35. [3] M.K. Debe, Electrocatalyst approaches and challenges for automotive fuel cells, Nature 486 (2012) 43–51. [4] H.A. Gasteiger, N.M. Markovic, Just a dream-or future reality, Science 324 (2009) 48–49. [5] F.T. Wagner, B. Lakshmanan, M.F. Mathias, Electrochemistry and the Future of the Automobile, The Journal of Physical Chemistry Letters 1 (2010) 2204–2219. [6] M. Shao, Electrocatalysis in Fuel Cells: A Non- and Low-Platinum Approach, Springer, London, 2013. [7] Z. Chen, D. Higgins, A. Yu, L. Zhang, J. Zhang, A review on non-precious metal electrocatalysts for PEM fuel cells, Energy & Environmental Science 4 (2011) 3167–3192. [8] M. Lefevre, E. Proietti, F. Jaouen, J.P. Dodelet, Iron-based catalysts with improved oxygen reduction activity in polymer electrolyte fuel cells, Science 324 (2009) 71–74. [9] E. Proietti, F. Jaouen, M. Lefèvre, N. Larouche, J. Tian, J. Herranz, J.-P. Dodelet, Iron-based cathode catalyst with enhanced power density in polymer electrolyte membrane fuel cells, Nature Communications 2 (2011) 416. [10] G. Wu, K.L. More, C.M. Johnston, P. Zelenay, High-Performance electrocatalysts for oxygen reduction derived from polyaniline, Iron, and Cobalt, Science 322 (2011) 443–447. [11] L. Elbaz, G. Wu, P. Zelenay, Heat-treated non-precious-metal-based catalysts for oxygen reduction, in: M. Shao (Ed.), Electrocatalysis in Fuel Cells: A Nonand Low-Platinum Approach, Springer, London, 2013. [12] J.P. Dodelet, The controversial role of the metal in Fe- or Co-based electrocatalysts for the oxygen reduction reaction in acid medium, in: M. Shao (Ed.), Electrocatalysis in Fuel Cells: A Non- and Low-Platinum Approach, Springer, London, 2013. [13] P. Zelenay, DOE Hydrogen Program Annual Merit Review and Peer Evaluation Meeting, Washington, DC, 2010. [14] T. Schilling, M. Bron, Oxygen reduction at Fe–N-modified multi-walled carbon nanotubes in acidic electrolyte, Electrochimica Acta 53 (2008) 5379–5385. [15] H.R. Byon, J. Suntivich, E.J. Crumlin, Y. Shao-Horn, Fe–N-modified multi-walled carbon nanotubes for oxygen reduction reaction in acid, Physical Chemistry Chemical Physics 13 (2011) 21437–21445. [16] A. Morozan, P. Jegou, B. Jousselme, S. Palacin, Electrochemical performance of annealed cobalt–benzotriazole/CNTs catalysts towards the oxygen reduction reaction, Physical Chemistry Chemical Physics 13 (2011) 21600–21607. [17] C.-W. Tsai, H.M. Chen, R.-S. Liu, K. Asakura, L. Zhang, J. Zhang, M.-Y. Lo, Y.-M. Peng, Carbon incorporated FeN/C electrocatalyst for oxygen reduction enhancement in direct methanol fuel cells: X-ray absorption approach to local structures, Electrochimica Acta 56 (2011) 8734–8738. [18] P. Ayala, A. Gruneis, C. Kramberger, M.H. Rummeli, I.G. Solorzano, J.F.L. Freire, T. Pichler, Effects of the reaction atmosphere composition on the synthesis of single and multiwalled nitrogen-doped nanotubes, The Journal of Chemical Physics 127 (2007) 184709. [19] F. Tuinstra, J.L. Koenig, Raman Spectrum of Graphite, The Journal of Chemical Physics 53 (1970) 1126–1130. [20] A.V. Palenzuela, L. Zhang, L. Wang, P.L. Cabot, E. Brillas, K. Tsay, J. Zhang, Fe–Nx/C electrocatalysts synthesized by pyrolysis of Fe(II)–2,3,5,6-tetra(2pyridyl)pyrazine complex for PEM fuel cell oxygen reduction reaction, Electrochimica Acta 56 (2011) 4744–4752. [21] C.A. Reiser, L. Bregoli, T.W. Patterson, J.S. Yi, J.D. Yang, M.L. Perry, T.D. Jarvi, A Reverse-Current Decay Mechanism for Fuel Cells Batteries, Fuel, Cells, and Energy Conversion, Electrochemical and Solid-State Letters 8 (2005) A273. [22] Th. Palaniselvam, R. Kannan, S. Kurungo, Facile construction of non-precious iron nitride-doped carbon nanofibers as cathode electrocatalysts for proton exchange membrane fuel cells, Chemical Communications 47 (2011) 2910–2912. [23] H.-S. Oh, J.-G. Oh, B. Roh, I. Hwang, H. Kim, Development of highly active and stable non-precious oxygen reduction catalysts for PEM fuel cells using polypyrrole and a chelating agent, Electrochemistry Communications 13 (2011) 879–881. [24] G. Liu, X. Li, P. Ganesan, B.N. Popov, Studies of oxygen reduction reaction active sites and stability of nitrogen-modified carbon composite catalysts for PEM fuel cells, Electrochimica Acta 55 (2010) 2853–2858. [25] G. Liu, X. Li, J.W. Lee, B.N. Popov, A review of the development of nitrogenmodified carbon-based catalysts for oxygen reduction at USC, Catalysis Science & Technology 1 (2011) 207–217. [26] G. Li, G. Liu, B. Popov, Activity and stability of non-precious metal catalysts for oxygen reduction in acid and alkaline electrolytes, Journal of Power Sources 195 (2010) 6373–6378. [27] G. Ma, R. Jia, J. Zhao, Z. Wang, C. Song, S. Jia, Z. Zhu, Nitrogen-Doped Hollow Carbon Nanoparticles with Excellent Oxygen Reduction Performances and Their Electrocatalytic Kinetics, Journal of Physical Chemistry C 115 (2011) 25148–25154.
132
B. Merzougui et al. / Electrochimica Acta 107 (2013) 126–132
[28] N. Ramaswamy, S. Mukerjee, Fundamental Mechanistic Understanding of Electrocatalysis of Oxygen Reduction on Pt and Non-Pt Surfaces: Acid versus Alkaline Media, Advances in Physical Chemistry (2012), http://dx.doi.org/10.1155/2012/491604. [29] Y. Liang, Y. Li, H. Wang, J. Zhou, J. Wang, T. Regier, H. Dai, Co3 O4 nanocrystals on graphene as a synergistic catalyst for oxygen reduction reaction, Nature Materials 10 (2011) 780–786. [30] Y. Liang, H. Wang, P. Diao, W. Chang, G. Hong, Y. Li, M. Gong, L. Xie, J. Zhou, J. Wang, T.Z. Regier, F. Wei, H. Dai, Oxygen Reduction Electrocatalyst Based on Strongly Coupled Cobalt Oxide Nanocrystals and Carbon Nanotubes, Journal of the American Chemical Society 134 (2012) 15849–15857. [31] K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Nitrogen-doped carbon nanotube arrays with high electrocatalytic activity for oxygen reduction, Science 323 (2009) 760–764.
[32] B. Piela, T.S. Olson, P. Atanassov, P. Zelenay, Highly methanol-tolerant nonprecious metal cathode catalysts for direct methanol fuel cell, Electrochimica Acta 55 (2010) 7615–7621. [33] L.T. Qu, Y. Liu, J.B. Baek, L. Dai, Nitrogen-doped graphene as efficient metalfree electrocatalyst for oxygen reduction in fuel cells, ACS Nano 4 (2010) 1321–1326. [34] Z. Wang, R. Jia, J. Zheng, J. Zhao, L. Li, J. Song, Z. Zhu, Nitrogen-promoted selfassembly of N-doped carbon nanotubes and their intrinsic catalysis for oxygen reduction in fuel cells, ACS Nano 5 (2011) 1677–1684. [35] Y. Li, W. Zhou, H. Wang, L. Xie, Y. Liang, F. Wei, J.-C. Idrobo, S.J. Pennycook, H. Dai, An oxygen reduction electrocatalyst based on carbon nanotube-graphene complexes, Nature Nanotechnology 7 (2012) 394–400.