Document not found! Please try again

Investigation of Supported Pd-Based Electrocatalysts for the Oxygen

0 downloads 0 Views 2MB Size Report
Nov 25, 2015 - to evaluate their activity, stability, and methanol tolerance properties. ... Keywords: direct methanol fuel cells; oxygen reduction reaction; ... 2.0 V vs. reference hydrogen electrode (RHE) occur [23]. ..... Pd catalysts present similar initial values of half-wave potential (E1/2), .... David Sebastián wrote the paper.
Article

Investigation of Supported Pd-Based Electrocatalysts for the Oxygen Reduction Reaction: Performance, Durability and Methanol Tolerance Carmelo Lo Vecchio, Cinthia Alegre, David Sebastián, Alessandro Stassi, Antonino S. Aricò and Vincenzo Baglio * Received: 12 October 2015; Accepted: 16 November 2015; Published: 25 November 2015 Academic Editor: Federico Bella CNR-ITAE Institute, Via Salita Santa Lucia sopra Contesse, 5, Messina 98126, Italy; [email protected] (C.L.V.); [email protected] (C.A.); [email protected] (D.S.); [email protected] (A.S.); [email protected] (A.S.A.) * Correspondence: [email protected] (V.B.); Tel.: +39-090-624-237

Abstract: Next generation cathode catalysts for direct methanol fuel cells (DMFCs) must have high catalytic activity for the oxygen reduction reaction (ORR), a lower cost than benchmark Pt catalysts, and high stability and high tolerance to permeated methanol. In this study, palladium catalysts supported on titanium suboxides (Pd/Tin O2n´1 ) were prepared by the sulphite complex route. The aim was to improve methanol tolerance and lower the cost associated with the noble metal while enhancing the stability through the use of titanium-based support; 30% Pd/Ketjenblack (Pd/KB) and 30% Pd/Vulcan (Pd/Vul) were also synthesized for comparison, using the same methodology. The catalysts were ex-situ characterized by physico-chemical analysis and investigated for the ORR to evaluate their activity, stability, and methanol tolerance properties. The Pd/KB catalyst showed the highest activity towards the ORR in perchloric acid solution. All Pd-based catalysts showed suitable tolerance to methanol poisoning, leading to higher ORR activity than a benchmark Pt/C catalyst in the presence of low methanol concentration. Among them, the Pd/Tin O2n´1 catalyst showed a very promising stability compared to carbon-supported Pd samples in an accelerated degradation test of 1000 potential cycles. These results indicate good perspectives for the application of Pd/Tin O2n´1 catalysts in DMFC cathodes. Keywords: direct methanol fuel cells; oxygen reduction reaction; methanol tolerance; Pd catalysts; durability tests; titanium suboxides

1. Introduction Fuel cells, fed with H2 or organic fuels, are promising energy conversion devices with low environmental impact [1,2]. DMFCs are very appropriate for portable power applications (such as consumer electronics) where the power requirements are low and a simple compact system with high energy density is required [3,4]. The electrocatalysis of ORR is one of the limiting steps due to the slow kinetics. This requires the use of expensive Pt-based catalysts characterized by suitable activity. Moreover, one of the main drawbacks in DMFC is the methanol crossover from the anode to the cathode side through the membrane. Methanol permeation produces a mixed potential at the cathode, reducing the overall efficiency [5]. Thus, DMFC technology requires the development of a catalyst with lower cost and more suitable activity and selectivity towards ORR than Pt/C [6,7]. To achieve this objective, Pt-free catalysts with better methanol tolerance and proper activity are considered. In this regard, Pd-based electrocatalysts are promising candidates, since they are characterized by a low methanol poisoning [8–10].

Materials 2015, 8, 7997–8008; doi:10.3390/ma8125438

www.mdpi.com/journal/materials

Materials 2015, 8, 7997–8008

Catalyst activity and stability also depend on structural and morphological features of the catalyst support [11]. Durability is an important issue that needs to be addressed in order to achieve a large-scale commercialization of proton exchange membrane fuel cells (PEMFCs) [12], of which DMFCs are a sub-category. In particular, cathode electrocatalyst degradation limits the durability of PEMFCs. The mechanisms causing a decrease of the catalyst (Pt, Pd, etc.) electrochemical surface area Materials 2015, 8, page–page (ECSA) are summarized into the following points: (1) dissolution of metal ions from small particles into the membrane; re-deposition onto (3) and corrosion of carbon support Catalyst(2) activity and stability also larger dependparticles; on structural morphological features of the with loss catalyst support [11]. Durability is an important issue that needs to be addressed in order to achieve of electronic contact or particle coalescence [13]. a large-scale commercialization of proton exchange membrane fuel cells (PEMFCs) [12], of which To improve the electrochemical activity and stability of the supported catalysts, significant work DMFCs are a sub-category. In particular, cathode electrocatalyst degradation limits the durability of has been done to newcausing carbon materials with controlled and tunable nanostructures PEMFCs.synthesize The mechanisms a decrease of the catalyst (Pt, Pd, etc.) electrochemical surface area [14–18]. However, (ECSA) the stability of carbon not appropriate underofsome harsh are summarized into the is following points: (1) dissolution metal ions from operating small particlesconditions intoby thethe membrane; (2) re-deposition ontoelectrolyte larger particles; (3) cell corrosion with are loss prone to experienced cathode in a polymer fuel [19].of carbon These support materials of electronic contact or particle coalescence [13]. corrosion, especially at high electro-chemical potentials or in practical starvation conditions, leading To improve the electrochemical activity and stability of the supported catalysts, significant work to a fast and significant loss of new ECSA over time with during fuel and celltunable operation [20]. With the aim of has been done to synthesize carbon materials controlled nanostructures [14–18]. improvingHowever, the cathode stabilityofincarbon DMFCs, a non-carbonaceous Ti-suboxide material was selected in the stability is not appropriate under some harsh operating conditions the cathode in aelectrocatalysts. polymer electrolyteTi-suboxides fuel cell [19]. are These materials are to materials the presentexperienced work as aby support for Pd recognized asprone robust corrosion, especially at high electro-chemical potentials or in practical starvation conditions, leading in aggressive media and are stable in harsh conditions such as those occurring at the cathode of to a fast and significant loss of ECSA over time during fuel cell operation [20]. With the aim of PEMFCs [21]; moreover, theystability exhibitinhigh electrical conductivity [22]. Ti-suboxides are also improving the cathode DMFCs, a non-carbonaceous Ti-suboxide material was selected in excellent supports for Ir-oxide in for PEM electrolyzersTi-suboxides where electrochemical potentials as high as the present workanodes as a support Pd electrocatalysts. are recognized as robust materials in aggressive media and are stable (RHE) in harshoccur conditions as those occurring at the cathode of 2.0 V vs. reference hydrogen electrode [23].such A comparison of the Pd/Ti O n 2n´1 catalyst PEMFCs [21]; moreover, they exhibit high electrical conductivity [22]. Ti-suboxides are also excellent properties with respect to Pd supported on commonly used carbon blacks (Ketjenblack and Vulcan) supports for Ir-oxide anodes in PEM electrolyzers where electrochemical potentials as high as 2.0 V is provided termshydrogen of ORR electrode activity,(RHE) tolerance to methanol poisoning, resistance to accelerated vs.in reference occur [23]. A comparison of the Pd/Tinand O2n–1 catalyst properties degradation tests. an evaluation of the methanol tolerance properties of isthese Pd catalysts with respectAlso, to Pd supported on commonly used carbon blacks (Ketjenblack and Vulcan) provided in terms of ORR activity, tolerance to methanol poisoning, and resistance to accelerated degradation compared to a commercial Pt/C is presented. tests. Also, an evaluation of the methanol tolerance properties of these Pd catalysts compared to a

Pt/C is presented. 2. Resultscommercial and Discussion 2. Results and Discussion

2.1. XRD Analysis

2.1. XRD Analysis

Figure 1 compares the XRD patterns of Pd catalysts dispersed on different supports and a 1 compares the XRD patterns of Pd that catalysts dispersed on differentPd/Ti supports and a commercial PtFigure catalyst on carbon. It can be seen Pd/Vulcan, Pd/KB, n O2n´1 and Pt/C commercial Pt catalyst on carbon. It can be seen that Pd/Vulcan, Pd/KB, Pd/TinO2n–1 and Pt/C show show the typical face-centered cubic structure (fcc) of Pd and Pt. the typical face-centered cubic structure (fcc) of Pd and Pt. Ti O 4

Ti O 5

Intensity/a.u.

Intensity/a.u.

Ti O 6

Ti O 9

20

25

6.5 nm

20

40

30 2 Θ/ degrees

7 9 11 17

35

40

30% Pd/Vul

7.3 nm

30% Pd/Ti O

4.0 nm

30% Pd/KB

2.8 nm

30% Pt/C

60 80 2 Θ/degrees

n

100

2n-1

120

Figure 1. XRD patterns of 30% Pd/Vul, 30% Pd/KB, 30% Pd/TinO2n–1 and commercial 30% Pt/C.

Figure 1. XRD patterns of 30% Pd/Vul, 30% Pd/KB, 30% Pd/Tin O2n´1 and commercial 30% Pt/C. nO2n–1 catalyst. Inset: Ti-oxide phases for Pd/Ti Inset: Ti-oxide phases for Pd/Tin O2n´1 catalyst. The C (002) graphite carbon reflection occurs at about 2θ = 25°. The typical peaks of Pd appear at 40.1°, 46.6°, 67.8° and 82.1° corresponding to (111), (200), (220) and (311) planes of the fcc structure The Cof(002) graphite carbon reflection occurs at about 2θ = 25˝ . The typical peaks of Pd appear Pd, respectively. Regarding the 30% Pd/Ti nO2n–1, besides Pd reflections, some other peaks appear ˝ ˝ ˝ ˝ at 40.1 , 46.6 , 67.8 and 82.1 corresponding to (111), (200), (220) and (311) planes of the fcc structure 2

of Pd, respectively. Regarding the 30% Pd/Tin O2n´1 , besides Pd reflections, some other peaks appear

7998

Materials 2015, 8, 7997–8008

which are related to Ti-oxide phases such as Ti4 O7 , Ti5 O9 , Ti6 O11 and Ti9 O17 [23]. The main peaks of Ti4 O7 (JCPDS: 18-1402), Ti5 O9 (JCPDS: 11-193), Ti6 O11 (JCPDS: 18-1401) and Ti9 O17 (JCPDS: 18-1405) Materials 2015, 8, page–page are indicated in the inset of Figure 1. The calculated crystallite size (by Scherrer’s equation) of Pd which are and related Ti-oxide phases such as Tiand 4O7, Ti5O9, Ti6O11 and Ti9O17 [23]. The main peaks of is 6.5 nm, 4 nm 7.3tonm in Pd/Vul, Pd/KB Pd/Tin O2n ´ 1 , respectively, and 2.8 nm for Pt in Ti4O7 (JCPDS: 18-1402), Ti5O9 (JCPDS: 11-193), Ti6O11 (JCPDS: 18-1401) and Ti9O17 (JCPDS: 18-1405) Pt/C, referred to the 220 reflection peak at around 2θ = 68˝ . The presence of KB allows us to obtain are indicated in the inset of Figure 1. The calculated crystallite size (by Scherrer’s equation) of Pd 2is ´1 a smaller size Pd-based catalysts due to the higher surface area (950 m ¨ g ) of 6.5 crystallite nm, 4 nm and 7.3for nm the in Pd/Vul, Pd/KB and Pd/Ti nO2n–1, respectively, and 2.8 nm for Pt in Pt/C, 2 ¨ g´1 ) [24]. The largest crystallite size for Pd (7.3 nm) this carbon black compared to Vulcan (250 m referred to the 220 reflection peak at around 2θ = 68°. The presence of KB allows us to obtain a smaller crystallite the Pd-based catalysts by duethe to the higher surface area (950 m2and ·g−1) Teller of this (BET) carbon surface is obtained with size the for support characterized lowest Brunauer, Emmett 2·g−1) [24]. The largest crystallite size for Pd (7.3 nm) is obtained with 2 ´ 1 black compared to Vulcan (250 m area (Tin O2n ´ 1 , 2 m ¨ g ). The commercial Pt/C shows the smallest crystallite size (2.8 nm). the support characterized by the lowest Brunauer, Emmett and Teller (BET) surface area (TinO2n–1, 2·g−1). The commercial Pt/C shows the smallest crystallite size (2.8 nm). 2m 2.2. TEM Analysis

2.2.analysis TEM Analysis TEM (Figure 2) showed a good dispersion for the catalysts supported on carbon blacks (Vulcan andTEM KB)analysis and a(Figure mean2)particle to for thethe crystallite size determined XRD (see showed asize goodsimilar dispersion catalysts supported on carbon by blacks histograms in Figure 2). For the catalyst supported on Ti-suboxides, the Pd nanoparticles are (Vulcan and KB) and a mean particle size similar to the crystallite size determined by XRD (see clearly in Figure thelarge catalyst supported on Ti-suboxides, nanoparticles clearly Ti and visible histograms only at the edges2).ofFor the support particles due to the thePdlow contrast are between visible only at the edges of the large support particles due to the low contrast between Ti and Pd for all Pd species. Nonetheless, a suitable Pd particle distribution on the support is envisaged species. Nonetheless, a suitable Pd particle distribution on the support is envisaged for all catalysts. Thermal gravimetric analysis confirmed a metal concentration (of 30 ˘ 2 wt %) for Pd/C catalysts. Thermal gravimetric analysis confirmed a metal concentration (of 30 ± 2 wt %) for Pd/C electrocatalysts; the same result waswas obtained O2nsample by means EDX analysis. ´1 sample electrocatalysts; the same result obtainedfor forthe the Pd/Ti Pd/TinOn2n–1 by means of EDX of analysis.

(a)

(b)

(c) Figure 2. TEM micrographs for the Pd-based catalysts: (a) 30% Pd/KB; (b) 30% Pd/Vul; (c) 30% Pd/TinO2n–1.

Figure 2. TEM micrographs for the Pd-based catalysts: (a) 30% Pd/KB; (b) 30% Pd/Vul; (c) 30% Pd/Ti O2n´1 . Disk Electrode (RDE) Measurements 2.3.nRotating Figure 3 shows the ORR curves obtained in O2-saturated perchloric acid without methanol and

2.3. Rotating Disk Electrode (RDE)concentrations Measurements with two different methanol (50 mM; 100 mM). These curves allow us to evaluate the

3 2 -saturated perchloric acid without methanol and Figure 3 shows the ORR curves obtained in O with two different methanol concentrations (50 mM; 100 mM). These curves allow us to evaluate the methanol tolerance of the different catalysts. The methanol addition causes a shift of the ORR onset

7999

Materials 2015, 8, 7997–8008 Materials 2015, 8, page–page

methanol the different catalysts. methanol additionThis causes a shift themixed ORR onset towards more tolerance negative of potential values and a The decrease of current. is due toofthe potential towards more negative potential values and a decrease of current.oxidation This is dueatto the the mixed potential caused by the simultaneous oxygen reduction and methanol working electrode causedThis by the simultaneous oxygen reduction and methanol oxidation at the working electrode interface. effect is more remarkable in the commercial 30% Pt/C catalyst showing an evident interface. This effect is more remarkable in the commercial 30% Pt/C catalyst showing an evident methanol oxidation peak during the ORR experiment. The least pronounced shift is noted for the methanol oxidation peak during the ORR experiment. The least pronounced shift is noted for the 30% Pd/KB catalyst with respect to the other Pd-supported catalysts. Such results confirm a better 30% Pd/KB catalyst with respect to the other Pd-supported catalysts. Such results confirm a better tolerance to methanol of the Pd-based tolerance to methanol of the Pd-basedcatalysts. catalysts. 1

1

MeOH 0mM MeOH 50mM MeOH 100mM

0 -1

30% Pd/Vul

2n-1 -2

-2

-2

J / mA·cm

J / mA·cm

-1

30% Pd/Ti O n

-3

-2 -3

-4

-4

-5

-5

-6

MeOH 0mM MeOH 50mM MeOH 100mM

0

-6

0

0.2

0.4

0.6

0.8

1

1.2

0

0.2

0.4

(a)

1

1.2

1

1.2

1 MeOH 0mM MeOH 50mM MeOH 100mM

0

MeOH 0mM MeOH 50mM MeOH 100mM

0

-1

-1 J / mA·cm-2

30% Pt/C E-TEK -2

0.8

(b)

1

J / mA·cm

0.6

Potential vs RHE / V

Potential vs RHE / V

-2 -3

30% Pd/KB -2 -3

-4

-4

-5

-5

-6

-6 0

0.2

0.4

0.6

0.8

1

1.2

Potential vs RHE / V

0

0.2

0.4

0.6

0.8

Potential vs RHE / V

(c)

(d)

Figure 3. Linear sweep voltammetry of oxygen reduction at different MeOH concentrations (0; 50; 100 mM)

Figure 3. Linear sweep voltammetry of oxygen reduction at different MeOH concentrations (0; 50; for 30% Pt/C E-TEK (a); 30% Pd/KB (b); 30% Pd/TinO2n–1 (c) and 30% Pd/Vul (d). Scan rate 5 mV·s−1; 100 mM) for 30% Pt/C E-TEK (a); 30% Pd/KB (b); 30% Pd/Tin O2n´1 (c) and 30% Pd/Vul (d). Scan 1000 rpm; room temperature and 50 μg·cm−2 of Pd or Pt loading. ´1 rate 5 mV¨ s ; 1000 rpm; room temperature and 50 µg¨ cm´2 of Pd or Pt loading.

The commercial 30% Pt/C catalyst shows the highest limiting current (il) when the bare electrolyte (only 0.130% M HClO 4) is catalyst used. Theshows il decreases the addition of methanol. 30% Pd/KB The commercial Pt/C the after highest limiting current (iThe the bare l ) when has an (only excellent and the limiting current is similar to the commercial Pt/C catalyst electrolyte 0.1methanol M HClOtolerance ) is used. The i decreases after the addition of methanol. The 30% 4 l even in the presence of 100 mM methanol. Pd catalysts supported on Ti nO2n–1 and Vulcan both exhibit Pd/KB has an excellent methanol tolerance and the limiting current is similar to the commercial a good tolerance to methanol poisoning, although the best catalytic activity was obtained with KB as Pt/C catalyst even in the presence of 100 mM methanol. Pd catalysts supported on Tin O2n´1 and the support. These results point out the good methanol tolerance of Pd, whereas the support Vulcan both exhibit a good tolerance to methanol poisoning, although the best catalytic activity was influences the level of catalytic activity according to the degree of catalyst dispersion. obtained with KB as theofsupport. These results point out the good methanol tolerance of Pd, whereas A comparison ORR curves for all the above-mentioned catalysts is shown in Figures 4 and 5. the support influences the level of catalytic activity according to the degree of catalyst dispersion. Figure 4a shows oxygen reduction curves obtained in the base electrolyte without methanol. The A comparison ORRhas curves for all the above-mentioned catalysts is shown Figures and 4 and 5. commercial 30%ofPt/C a higher activity than the Pd catalysts in terms of onsetinpotential limiting current. Pd/Tireduction nO2n–1 has an onset potential mV,base which is more negative the Pt/C The Figure 4a shows oxygen curves obtainedofin90the electrolyte withoutthan methanol. −2. The catalyst 30% in thePt/C activation-controlled the the limiting current density is about mA·cm commercial has a higher region, activityand than Pd catalysts in terms of 2.5 onset potential and 30% Pd/Vul and 30% Pd/KB catalysts show a similar behavior in terms of onset potential which is just limiting current. Pd/Tin O2n ´1 has an onset potential of 90 mV, which is more negative than the Pt/C slightly higher than that of Pd/TinOregion, 2n–1. The diffusion-limiting current density of carbon-supported Pd ´2 catalyst in the activation-controlled and the limiting current density is about 2.5 mA¨ cm . catalysts (jd) reaches a value close to the commercial Pt catalyst. The jd is related to the number of The 30% Pd/Vul and 30% Pd/KB catalysts show a similar behavior in terms of onset potential electrons transferred per O2 molecule as reported in the Koutecky-Levich (K-L) equation: which is just slightly higher than that of Pd/Tin O2n ´1 . The diffusion-limiting current density of 4 carbon-supported Pd catalysts (jd ) reaches a value close to the commercial Pt catalyst. The jd is

8000

Materials 2015, 8, 7997–8008

related to the number of electrons transferred per O2 molecule as reported in the Koutecky-Levich (K-L) equation: Materials 2015, 8, page–page 1 1 1 1 1 (1) “ ` “ ` Materials 2015, 8, page–page j jk jd jk 0.62nFC D2{3 v´1{6 ω1{2

O2 1O2 1 1 1 = + = + (1) / ω / 1 1 1 1/ 1 0.62 where j is the overall current density; = + =jK is + the kinetic/ current density; jd is the diffusion (1) / 0.62 current density; n is the number of jelectrons transferred in the/jdωORR; F is thecurrent Faraday constant where j is the overall current density; K is the kinetic current density; is the diffusion density; ´1 ); C (1.26 ˆ 10´3 M) is the saturated concentration of oxygen; D (2.60 ˆ 10−1´5 cm2 ¨ s´1 ) (96485 C¨the mol nwhere is number transferred ORR;current F is the Faraday (96485 C·moldensity; ); Oof O2 current 2 electrons j is the overall current density; jK isin thethe kinetic density; jd isconstant the diffusion ´3 oxygen; 2 ¨ s´1 ) is the kinematic −3 coefficient −5 cm2viscosity −1) is theof −1 is the(1.26 diffusion of oxygen; v (9.00 ˆ 10 cm the × 10 M) is the saturated concentration of (2.60 × 10 ·s diffusion n is the number of electrons transferred in the ORR; F is the Faraday constant (96485 C·mol ); solution, 2·s−1) is the kinematic viscosity−5 of the andrate. w All coefficient oxygen; v400 (9.00 × concentration 10−3 cm and w (100× rpm, 200 rpm, 1000 rpm, 1600 rpm, 2500 rpm) the electrode (1.26 10−3ofM) isrpm, the saturated of oxygen; (2.60 is × 10 cm2·s−1)solution, is rotation the diffusion (100 rpm, 200 rpm, 400 1000 rpm, 1600 rpm, 2500 rpm) is the electrode rotation rate. All the −3 cm 2·s−1 the parameters, tovrpm, the 0.1 HClO solution, are reported in the literature [25]. As shown ) is the kinematic viscosity of the solution, and w coefficient ofreferred oxygen; (9.00 ×M 10 4 ´ 1 ´ 1 { 2 parameters, referred to the 0.1 M HClO 41600 solution, are reported in the literature [25]. As shown in (100 rpm, 200 rpm, 400 rpm, 1000 rpm, rpm, 2500 rpm) is the electrode rotation rate. All the in Figure 4b, the K-L plots (j vs. ω ) of the ORR exhibited good linearity, indicating first-order −1 vs. ω−1/2) of the ORR exhibited good linearity, indicating first-order kinetics Figure 4b, thereferred K-L plots parameters, to(jthe 0.1 M HClO 4 solution, are thePd/KB literature [25]. As shown in kinetics with respect to the reactant concentration. Thereported n valueinfor and Pd/Vul, calculated with respect toK-L theplots reactant concentration. Theexhibited n value for Pd/KB and indicating Pd/Vul, calculated from the −1 vs. −1/2) of the ORR Figure 4b, the (j ω good linearity, first-order kinetics from the slopes in the K-L plots, is 3.8. This indicates that oxygen reduction mainly involves a slopes in the K-L plots, is 3.8.concentration. This indicatesThe thatnoxygen reduction mainly involves a four-electron with respect to the reactant value for Pd/KB and Pd/Vul, calculated from the four-electron transfer topathway to produce water. for ThePd/Ti n value for 2.1 Pd/Ti 2.1 in the same n O2n ´1 isexperimental transfer produce The n value nO2n–1 ismainly in involves the same slopes inpathway the K-L plots, is 3.8. water. This indicates that oxygen reduction a four-electron experimental conditions. This implies a reaction mechanism corresponding to the two-electron conditions. This implies a reaction mechanism corresponding two-electron pathway which transfer pathway to produce water. The n value for Pd/TinOto 2n–1the is 2.1 in the same experimental pathway which involves the reduction of oxygen to hydrogen peroxide. The lower j of Pd/Ti O2n ´1 nto involves the reduction of oxygen to hydrogen peroxide. The lower j d of Pd/Ti n O 2n – 1 is thus related d conditions. This implies a reaction mechanism corresponding to the two-electron pathway which is thus related to reduction its lower number of inlower the ORR. its lower number of electrons transferred in thetransferred ORR. involves the of oxygen toelectrons hydrogen peroxide. The jd of Pd/TinO2n–1 is thus related to

1

its lower number of electrons transferred in the ORR.

(a)

(b)

Figure 4. (a) Comparison (a) of oxygen reduction voltammetric curves(b) for the different Pd/support Figure 4. (a) Comparison of oxygen reduction voltammetric curves for the different Pd/support −1; 1000 rpm; room temperature catalysts and the commercial Pt/C E-TEK catalyst. Scan rate 5 mV·s Figure (a) commercial Comparison Pt/C of oxygen reduction voltammetric curves for the different Pd/support ´1 catalysts and4.the E-TEK catalyst. Scan rate 5 mV¨ s ; 1000 rpm; room temperature −2 and 50 μg·cm Pd or Pt Pt/C loading; (b)catalyst. Koutecky-Levich 30%rpm; Pd/Vul, Pd/KB, ; 1000 room 30% temperature catalysts and theof commercial E-TEK Scan rate 5 plots mV·s−1for ´2 and 50 µg¨ cm of Pd or Pt loading; (b) Koutecky-Levich plots for 30% Pd/Vul, 30% Pd/KB, 30% 30% Pd/Ti nO2n–1−2obtained from the RDE data at 0.7 V vs. RHE. and 50 μg·cm of Pd or Pt loading; (b) Koutecky-Levich plots for 30% Pd/Vul, 30% Pd/KB, Pd/Tin O2n´1 obtained from the RDE data at 0.7 V vs. RHE. 30% Pd/TinO2n–1 obtained from the RDE data at 0.7 V vs. RHE.

Figure 5. Comparison of oxygen reduction curves for the different Pd/support catalysts and the commercial Pt/C E-TEK of catalyst in the presence of 0.1for M MeOH. Scan rate 5 mV·s−1;catalysts 1000 rpm; room Figure 5. Comparison oxygen reduction curves the different Pd/support and the Figure 5. Comparison of oxygen curves for the different Pd/support catalysts and the −2 of Pdreduction or Pt loading. temperature and 50 μg·cm commercial Pt/C E-TEK catalyst in the presence of 0.1 M MeOH. Scan rate 5 mV·s−1; 1000 rpm; room

commercial Pt/C E-TEK catalyst in the presence of 0.1 M MeOH. Scan rate 5 mV¨ s´1 ; 1000 rpm; room −2 of Pd or Pt loading. temperature and 50 μg·cm ´2 temperature µg¨ cm of or Pt reduction loading. reactions for the different catalysts in the presence In Figureand 5, a50 comparison ofPd oxygen of 0.1In MFigure methanol in the bulk solution is shown. Under these thecatalysts Pd/KB catalyst shows 5, a comparison of oxygen reduction reactions forconditions, the different in the presence the most positive onset potential and reaches the largest limiting current density compared to that of Figure 0.1 M methanol in the bulk solution reduction is shown. Under thesefor conditions, the Pd/KB catalyst shows In 5, a comparison of oxygen reactions the different catalysts in the presence obtained with Pt/C. These results show the better tolerance to low methanol concentrations of positiveinonset and reaches the largest limiting current density compared to that of 0.1the Mmost methanol the potential bulk solution is shown. Under these conditions, the Pd/KB catalyst Pd-based catalysts compared to a benchmark Pt catalyst. obtained with Pt/C. These results show the better tolerance to low methanol concentrations of

Pd-based catalysts compared to a benchmark Pt catalyst. 5

8001 5

Materials 2015, 8, 7997–8008

shows the most positive onset potential and reaches the largest limiting current density compared to that obtained with Pt/C. These results show the better tolerance to low methanol concentrations of Pd-based catalysts compared to a benchmark Pt catalyst. KB used as a support material for Pd nanoparticles allows us to achieve a better distribution of Pd particles onto the support. The result is the obtainment of the smallest crystallite size among the Materials 2015, 8, page–page synthesized supported catalysts. These aspects determine the highest ORR activity for the catalyst based on KB KB.used as a support material for Pd nanoparticles allows us to achieve a better distribution of Pd particles onto the support. The result is the obtainment of the smallest crystallite size among the

2.4. Accelerated Tests (ADTs) synthesizedDegradation supported catalysts. These aspects determine the highest ORR activity for the catalyst based on KB.

The main kinetic parameters of Pd electrocatalysts are reported in Table 1. ECSA of Pd-based catalysts, before and after ADT, was determined in deaerated 0.1 M HClO4 using cyclic voltammetry 2.4. Accelerated Degradation Tests (ADTs) (CV) (Figure 6a,b) by integrating the peak related to the Pd-oxide reduction between 0.4 and 0.8 V The main kinetic parameters of Pd electrocatalysts are reported in Table 1. ECSA of Pd-based vs. RHE for Pd/KB and from 0.5 to 0.9 V vs. RHE for Pd/Vul and Pd/Tin O2n ´1 . A charge of catalysts, before and after ADT, was determined in deaerated 0.1 M HClO4 using cyclic voltammetry ´2 for the reduction of a Pd-oxide monolayer was assumed according to the literature [26]. 405 µC cm(Figure (CV) 6a,b) by integrating the peak related to the Pd-oxide reduction between 0.4 and 0.8 V vs. Among Pd/support catalysts, Pd/KB presents highest ECSA ADT.of This due thethe higher RHE for Pd/KB and from 0.5 to 0.9 V vs. RHE forthe Pd/Vul and Pd/Ti nObefore 2n–1. A charge 405is μC cm−2tofor specific surfaceofarea (in terms of BET)was of assumed the carbon support, leading to high and a lower reduction a Pd-oxide monolayer according to the literature [26]. dispersion Among Pd/support catalysts, Pd/KB presents thethe highest ECSA before ADT.the This is dueloss to the Pd particle size (4.0 nm). After ADT, Pd/KB shows largest ofhigher ECSAspecific (71%). surface Pd/Vularea follows (in terms BET) ifofthe theloss carbon support, leading to high lower. dispersion andn O a 2n lower Pd particle the same trendofeven of ECSA (60%) is slightly Pd/Ti the size smallest ´1 presents (4.0 nm). After ADT, Pd/KB showsSuch the largest loss of ECSA (71%). Pd/Vul follows same trend loss of ECSA (20%)the after degradation. differences among carbonaceous and the non-carbonaceous even if the loss of ECSA (60%) is slightly lower. Pd/Ti nO2n–1 presents the smallest loss of ECSA (20%) supports are also evident in the CVs (Figure 6) by comparing the profiles before and after ADT. The after degradation. Such differences among carbonaceous and non-carbonaceous supports are CV of the Tin O2n ´1 -supported Pd catalyst barely changes after ADT, while those related to Vul and also evident in the CVs (Figure 6) by comparing the profiles before and after ADT. The CV of the KB present a significant variation. TinO2n–1-supported Pd catalyst barely changes after ADT, while those related to Vul and KB present a significant variation. Table 1. Kinetic parameters for the Pd/support catalysts. Table 1. Kinetic parameters for the Pd/support catalysts.

Catalyst Catalyst

Before ADT Before ADT Pd/Vul Pd/Vul Pd/KB Pd/KB Pd/Ti n On2n´1 Pd/Ti O2n–1 After ADT After ADT Pd/Vul Pd/Vul Pd/KB Pd/KB Pd/Tin O2n´1 Pd/TinO2n–1

ECSA/m2 ¨ g´1

ECSA/m2·g−1

Tafel Slope I Tafel slope II E /V Tafel slope II Tafel Slope I dec´1 Region/mV¨ Region/mV¨ dec´1 E1/21/2 /V −1 −1 Region/mV·dec

Region/mV·dec

37.8

37.8 39.0 39.0 9.10 9.10

54.8

126.7 126.7 143.0 143.0 126.9 126.9

48.6

133.5 133.5 127.1 127.1 124.6

54.8 63.5 63.5 48.8 48.8

15.2

15.2 11.2 11.2 7.30 7.30

48.6 72.4 72.4 48.6 48.6

0.78

0.78 0.80 0.80 0.79 0.79

0.77

0.77 0.73 0.73 0.79 0.79

124.6

(a)

SA at 0.85 V vs. SA at 0.85 V vs.cm´2 RHE/mA¨ RHE/mA·cm−2

0.022

0.022 0.033 0.033 0.056 0.056

0.054

0.054 0.039 0.039 0.043 0.043

(b)

Figure 6. (a) CV of 30% Pd/Vul, 30% Pd/KB, 30% Pd/TinO2n–1 before ADT; (b) CV of 30% Pd/Vul,

Figure 6. (a) CV of 30% Pd/Vul, 30% Pd/KB, 30% Pd/Tin O2n´1 before ADT; (b) CV of 30% Pd/Vul, 30% Pd/KB, 30% Pd/TinO2n–1 after 1000 potential cycles. 30% Pd/KB, 30% Pd/Tin O2n´1 after 1000 potential cycles.

Figure 7 shows the percentage of the ECSA retention after ADT for the Pd catalysts. This is

Figure 7 shows percentage of ECSA the ECSA ADT and for the Pd catalysts. This is estimated from thethe ratio between the at theretention end of theafter test (EOT) the same value at the beginning theratio test (BOT). In light of the electrochemical results, the differences estimated fromofthe between the ECSA at the end of the test (EOT) and theencountered same valueinat the terms of ECSA decay can be ascribed to the different support surface area and composition, although the contribution of the initial Pd particle size cannot be discarded. The best results in terms of 8002 resistance to corrosion are obtained with the non-carbonaceous support (Ti-suboxides) [27]. 6

Materials 2015, 8, 7997–8008

beginning of the test (BOT). In light of the electrochemical results, the differences encountered in terms of ECSA decay can be ascribed to the different support surface area and composition, although the contribution of the initial Pd particle size cannot be discarded. The best results in terms of resistance Materials to corrosion are obtained with the non-carbonaceous support (Ti-suboxides) [27]. 2015, 8, page–page 100

EOT/BOT

/%

80

ECSA

Materials 2015, 8, page–page

60

40 100

EOT/BOT

/%

20 80 0 60

Pd/KB

Pd/Vul

Pd/Ti O n

2n-1

ECSA

Figure 7. ECSA percentage for the different Pd/support catalysts obtained by the ratio between ECSA

40 Figure 7. ECSA percentage for the different Pd/support catalysts obtained by the ratio between ECSA at EOT and BOT. Scan rate 5 mV·s−1; 1000 rpm; room temperature and 50 μg·cm−2 of Pd loading. at EOT and BOT. Scan rate 5 mV¨ s´1 ; 1000 rpm; room temperature and 50 µg¨ cm´2 of Pd loading. 20

0.74 0.82

E

1/2

vs RHE / V

The specific activity (SA) for the O2 reduction was calculated at 0.85 V vs. RHE after normalizing the kinetic currents by the catalyst loading and the ECSA (Table 1). The SA increases passing The specific activity (SA) for the 0O2 reduction was calculated at 0.85 V vs. RHE after normalizing from carbon-supported Pd catalysts toPd/KB Pd/TinO2n–1Pd/Vul . This appears Pd/Ti O to reflect the typical relationship n 2n-1 the kineticobserved currents bycatalysts the catalyst loading and the ECSAcrystallographic (Table 1). The SA increases for Pt where the occurrence of well-defined orientations in large passing Figure ECSA percentage the different Pd/support catalysts obtained by theparticles ratio between ECSA particles is 7. associated with anforincrease of SA [28,29]. In contrast, small Pd contain a large from carbon-supported Pd catalysts to−1Pd/Ti O . This appears to reflect the typical relationship n 2n ´1 −2 of Pd loading. at EOT and BOT.which Scan rate 5less mV·s ; 1000 rpm; for room temperature and 50 μg·cm number of defects are appropriate a structure-sensitive reaction such as the oxygen observed for Pt catalysts where the occurrence of well-defined crystallographic orientations in large electro-reduction. In addition to these effects, it may be possible that the metal support interaction is The specific activity for the of O2 reduction was calculated at 0.85 V vs. RHE after normalizing particles islarger associated with an(SA) increase SA [28,29]. In contrast, small Pd particles contain a large in carbon-supported samples where there is a significant occurrence of organic functional kinetic which currentsare by less the catalyst loading for and athe ECSA (Table 1). Thereaction SA increases passing number ofthe defects appropriate structure-sensitive such as the oxygen groups on the surface. The metal support interaction may decrease the number of Pd sites available from carbon-supported Pd catalysts to Pd/TinO2n–1. This appears to reflect the typical relationship for the chemisorption of oxygen. electro-reduction. In addition to these effects, it may be possible that the metal support interaction observed for Pt catalysts where the occurrence of well-defined crystallographic orientations in large Pd catalysts present samples similar initial valuesthere of half-wave potential (Eoccurrence 1/2), around 800 mV vs. RHE. is larger inparticles carbon-supported where significant of organic is associated with an increase of SA [28,29].isIna contrast, small Pd particles contain a largefunctional The value of E1/2 for the Pd/TinO2n–1 electrocatalyst does not change after 1000 potential cycles (Figure 8); number of defects which are less appropriate for a structure-sensitive reaction such as the oxygen groups onhowever, the surface. The metal support interaction may decrease the number of Pd sites available it significantly decreases for carbonaceous support-based catalysts, especially for Pd/KB electro-reduction. In addition to these effects, it may be possible that the metal support interaction is for the chemisorption of mV oxygen. with a loss of 70 in E1/2. The values of the Tafel slope are also reported in Table 1, taking into larger in carbon-supported samples where there is a significant occurrence of organic functional account two regions: low current density (region I) and high potential current density (region II). There no vs. RHE. Pd catalysts present similar initial values of half-wave (E1/2 ),ofaround 800ismV groups on the surface. The metal support interaction may decrease the number Pd sites available significant difference in terms of the Tafel slope produced by the support, and no changes have been The valuefor ofthe E1/2 for the Pd/Ti n O2n ´1 electrocatalyst does not change after 1000 potential cycles chemisorption of oxygen. observed for this slope after the ADT. This indicates that the oxygen adsorption mechanism is similar Pd catalysts present similar initial valuesfor of half-wave potentialsupport-based (E1/2), around 800 mV vs. RHE.especially (Figure 8);regardless however, it significantly decreases carbonaceous catalysts, of these variables. Kinetic parameters, reported in Table 1, are in agreement with the value of E1/2 of for 70 the Pd/Ti nO2n–1 electrocatalyst does not change after 1000 potential cycles (Figure 8); for Pd/KBThe with a loss mV in E . The values of the Tafel slope are also reported in Table 1, results previously reported in the 1/2 literature [30,31]. however, it significantly decreases for carbonaceous support-based catalysts, especially for Pd/KB taking intowith account two regions: lowvalues current density (region I) and high current density (region II). a loss of 70 mV in E1/2. The0.82 of the Tafel slope are also reported in Table 1, taking into There is noaccount significant difference in terms of the Tafel slope by the support, and no changes before ADT two regions: low current density (region I) and highproduced current density (region II). There is no after ADT 0.8 0 mV significant difference in terms of the Tafel slope produced by the support, and no changes have been have been observed for this slope after the ADT. This indicates that the oxygen adsorption mechanism observed for of thisthese slope variables. after the ADT. This indicates that the oxygen adsorption mechanism is similar 0.78 Kinetic is similar regardless parameters, - 10 mV reported in Table 1, are in agreement with regardless of these variables. Kinetic parameters, reported in Table 1, are in agreement with the - 70 mV the resultsresults previously reported in the literature [30,31]. 0.76 previously reported in the literature [30,31]. before ADT after ADT

0.72 0.8

vs RHE / V

0.7 0.78 0.68 0.76

0 mV

- 10 mV

- 70 mV Pd/KB

Pd/Vul

Pd/Ti O n

2n-1

E

1/2

Figure 8. E1/2 of Pd/support 0.74 catalysts before and after degradation cycles. Scan rate 5 mV·s−1; 1000 rpm; room temperature and 50 μg·cm−2 of Pd loading. 0.72

Figure 9 shows the behavior0.7of Pd catalysts for the ORR before and after the ADT. The activity of the Pd/TinO2n–1 catalyst is almost unvaried, even after 1000 degradation cycles, indicating a very 0.68

Pd/KB

7

Pd/Vul

Pd/Ti O n

2n-1

Figure 8. E1/2 of Pd/support catalysts before and after degradation cycles. Scan rate 5 mV·s−1; Figure 8. 1000 E1/2rpm; of room Pd/support catalysts before and after degradation cycles. Scan rate 5 temperature and 50 μg·cm−2 of Pd loading. 1000 rpm; room temperature and 50 µg¨ cm´2 of Pd loading.

mV¨ s´1 ;

Figure 9 shows the behavior of Pd catalysts for the ORR before and after the ADT. The activity of the Pd/TinO2n–1 catalyst is almost unvaried, even after 1000 degradation cycles, indicating a very

8003 7

Materials 2015, 8, 7997–8008

Figure 9 shows the behavior of Pd catalysts for the ORR before and after the ADT. The activity of the Pd/Tin O2n ´1 catalyst is almost unvaried, even after 1000 degradation cycles, indicating a very Materials 2015, 8, page–page good stability of the non-carbonaceous support. It can be seen that Pd/KB loses significant activity after 1000 both in kinetic support. and diffusion-limiting regions. densityactivity values for good potential stability ofcycles the non-carbonaceous It can be seen that Pd/KB Current loses significant Pd/Vul, before and after ADT, present a similar trend in terms of onset and kinetic current after 1000 potential cycles both in kinetic and diffusion-limiting regions. Current density values but for the loss of activity is pronounced below 0.8 aVsimilar vs. RHE. aspects canand be kinetic mainlycurrent attributed to the Pd/Vul, before and after ADT, present trendThese in terms of onset but the loss ofbetween activity carbonaceous is pronounced below 0.8 V vs. RHE. These aspectscorrosion. can be mainly attributed to the difference and non-carbonaceous support The carbon blacks easily difference between carbonaceous and non-carbonaceous support corrosion. The carbon blacks easily react with water to produce CO2 in the investigated potential region. In contract, Ti-suboxides show react currents with water to produce CO2 in the lower investigated potentialblacks region.[27] In contract, show corrosion orders of magnitude than carbon and no Ti-suboxides clear electrochemical corrosion currents orders of magnitude lower than carbon blacks [27] and no clear electrochemical corrosion mechanism has yet been individuated in the literature [32,33]. corrosion mechanism has yet been individuated in the literature [32,33]. These results highlight the advantage of using Pd/Tin O2n ´1 in terms of resistance to corrosion These results highlight the advantage of using Pd/TinO2n–1 in terms of resistance to corrosion phenomena. Further work to optimize Pd particle size would lead to high-performing cathode phenomena. Further work to optimize Pd particle size would lead to high-performing cathode catalysts for application in DMFCs. catalysts for application in DMFCs.

(a)

(b)

(c) Figure 9. Oxygen reduction curves for the different Pd/support catalysts before and after degradation

Figure 9. Oxygen reduction curves for the different Pd/support−1catalysts before and after degradation rpm; room temperature and cycles. (a) Pd/KB; (b) Pd/Vul; (c) Pd/TinO2n–1. Scan rate 5 mV·s ; 1000 cycles. (a) Pd/KB; (b) Pd/Vul; (c) Pd/Tin O2n´1 . Scan rate 5 mV¨ s´1 ; 1000 rpm; room temperature 50 μg·cm−2 of Pd loading. and 50 µg¨ cm´2 of Pd loading.

3. Materials and Methods

3. Materials and Methods 3.1. Pd-Based Electrocatalyst Preparation

3.1. Pd-Based Electrocatalyst Preparation

First, 30% Pd supported on Vulcan-XC-72R (from Cabot, Boston, MA, USA), Ketjenblack (from

AkzoNobel, Amsterdam, and Ti(from suboxides (from Atraverda, Wales, UK) First, 30% Pd supportedThe onNetherlands) Vulcan-XC-72R Cabot, Boston, MA, Abertillery, USA), Ketjenblack (from was prepared using the complexand route An appropriate amount of support was AkzoNobel, Amsterdam, Thesulphite Netherlands) Ti [34]. suboxides (from Atraverda, Abertillery, Wales, 8

8004

Materials 2015, 8, 7997–8008

UK) was prepared using the sulphite complex route [34]. An appropriate amount of support was ultrasonically dispersed in water for one hour and then mixed with Pd-sulphite acidic solution. Subsequently, H2 O2 was added to decompose the sulphite complex with the formation of a colloidal suspension (PdOx /support) after pH correction (5.5). Hence, the suspension was filtered, copiously washed with water and dried at 80 ˝ C. The obtained powder was carbothermally reduced at 500 ˝ C in an inert atmosphere (Ar) to form Pd/Vul and Pd/KB. Pd/Tin O2n´1 was formed by the reduction of the dried powder, at room temperature (25 ˝ C), in a 10% H2 –90% Ar atmosphere. 3.2. Physico-Chemical Characterization X-ray diffraction (XRD) patterns for powder catalysts were recorded with a X’Pert 3710 X-Ray (Philips, Eindhoven, The Netherlands) diffractometer using a Cu-Kα source operating at 40 kV and 20 mA. The peak profiles of the (220) reflection for the face-centered cubic (fcc) structure of palladium or platinum phases were obtained by applying the Marquardt algorithm to calculate the crystallite size by the Debye-Sherrer equation. Instrumental broadening was determined by using a standard platinum sample. Transmission electron microscopy (TEM) analysis was made by first dispersing the catalyst powder in isopropyl alcohol. A few drops of these solutions were deposited on carbon film-coated Cu grids and analyzed with a CM12 microscope (Philips, Eindhoven, The Netherlands). An amount of 200 Pd particles was counted to obtain a particle size distribution histogram for the carbon- or Ti-suboxides-supported catalyst. The total metal content in the catalysts was determined by burning the carbon support in a thermal gravimetry experiment up to 950 ˝ C in air (STA analyser, Netzsch, Bayern, Germany). The Pd content in Pd/Tin O2n´1 catalyst was determined by energy dispersive X-ray (EDX) spectroscopy at 25 kV with an XL30 SFEG scanning electron microscope (FEI, Eindhoven, The Netherlands). 3.3. Electrochemical Measurements All the electrochemical measurements of the catalysts were performed at room temperature using a conventional three-electrode cell. The reference electrode was an Hg/Hg2 SO4 saturated with K2 SO4 . The counter electrode was a high-surface-area platinum wire and the working electrode was a glassy carbon (GC) disk of 5 mm where the thin film catalyst was deposited. The 30% Pd/support catalytic inks were prepared by sonicating each catalyst in isopropyl alcohol with 30% of Nafion for 30 min and then added onto a GC disk to obtain a 50 µg¨ cm´2 Pd loading. Before each measurement, the glassy carbon was polished with alumina suspension in an OP-Felt cloth. The commercial 30% Pt/C (from E-TEK) ink was prepared with the same procedure and used for comparison (same metal loading) with respect to supported Pd inks. The base electrolyte was a deaerated 0.1 M HClO4 solution in which methanol (MeOH) was gradually added to investigate the tolerance of the catalyst to the poisoning caused by this organic fuel at 50 mM and 100 mM MeOH concentrations. Conditioning of the catalyst was carried out, in the base electrolyte, by triangular voltage sweeps at a scan rate of 100 mV¨ s´1 between 0.05 and 1 V vs. reference hydrogen electrode (RHE). This procedure consisted of about 50 cycles up to obtaining a constant cyclic voltammogram (CV). Subsequently, three potential cycles were performed from 0.05 to 1 V vs. RHE, at a scan rate of 20 mV¨ s´1 to calculate the ECSA. ORR activity was evaluated by saturating the electrolyte with O2 and carrying out a linear sweep voltammetry from 1.1 to 0.2 V vs. RHE, at a scan rate of 5 mV¨ s´1 . A rotating disk electrode (RDE) was used at different rotations speeds. Accelerated degradation tests (ADTs) were performed by sweeping (1000 cycles) the potential from 0.6 to 1.0 V vs. RHE at a scan rate of 50 mV¨ s´1 . The loss of ECSA and the other kinetic parameters were calculated before and after the stability tests in a deaerated 0.1 M HClO4 solution, whereas the loss of activity was evaluated with a linear sweep voltammetry in an oxygen-saturated electrolyte.

8005

Materials 2015, 8, 7997–8008

The electrochemical potentiostat/galvanostat.

measurements

were

carried

out

with

a

Metrohm

Autolab

4. Conclusions The 30% Pd/Ketjenblack, 30% Pd/Vulcan and 30% Pd/Titanium suboxides have been synthesized and characterized for the oxygen reduction reaction to evaluate their activity in direct methanol fuel cell applications. It appears that Pd/Ketjenblack is the most active electrocatalyst in the presence of methanol in the base electrolyte, thus showing promising characteristics for direct methanol fuel cells. Instead, stability tests of these Pd-based catalysts indicate the best resistance to corrosion for the Ti-suboxide-supported Pd catalyst compared to carbonaceous supports, pointing out their good perspectives in terms of durability. Acknowledgments: Authors acknowledge the financial support of PRIN 2010-11 project “Advanced nanocomposite membranes and innovative electrocatalysts for durable polymer electrolyte membrane fuel cells (NAMED-PEM)”. Author Contributions: Vincenzo Baglio conceived and designed the experiments; Carmelo Lo Vecchio, Cinthia Alegre, Antonino S. Aricò and Alessandro Stassi performed the experiments; Carmelo Lo Vecchio and David Sebastián analyzed the data; Alessandro Stassi, Cinthia Alegre, David Sebastián and Carmelo Lo Vecchio contributed reagents/materials/analysis tools; Carmelo Lo Vecchio, Antonino S. Aricò, Vincenzo Baglio and David Sebastián wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5.

6.

7.

8.

9. 10.

11.

Carrette, L.; Friedrich, K.A.; Stimming, U. Fuel cells: Principles, types, fuels, and applications. Chem. Phys. Chem. 2000, 1, 163–193. Boudghene Stambouli, A.; Traversa, E. Fuel cells, an alternative to standard sources of energy. Renew. Sustain. Energy Rev. 2002, 6, 297–306. [CrossRef] Wee, J.H. A feasibility study of direct methanol fuel cells for laptop computers based on a cost comparison with lithium ion batteries. J. Power Sources 2007, 173, 424–436. [CrossRef] Zhao, T.S.; Chen, R.; Yang, W.W.; Xu, C. Small direct methanol fuel cells with passive supply of reactants. J. Power Sources 2009, 191, 185–202. [CrossRef] Castro Luna, A.M.; Bonesi, A.; Triaca, W.E.; Baglio, V.; Antonucci, V.; Aricò, A.S. Pt-Fe cathode catalysts to improve the oxygen reduction reaction and methanol tolerance in direct methanol fuel cells. J. Solid State Electrochem. 2008, 12, 643–649. [CrossRef] Park, A.R.; Lee, Y.W.; Kwak, D.H.; Roh, B.; Hwang, I.; Park, K.W. Enhanced electrocatalytic activity and stability of PdCo@Pt core-shell nanoparticles for oxygen reduction reaction. J. Appl. Electrochem. 2014, 44, 1219–1223. [CrossRef] Kolary-Zurowska, A.; Zieleniak, A.; Miecznikowski, K.; Baranowska, B.; Lewera, A.; Fiechter, S.; Bogdanoff, P.; Dorbandt, I.; Marassi, R.; Kulesza, P.J. Activation of methanol-tolerant carbon-supported RuSex electrocatalytic nanoparticles towards more efficient oxygen reduction. J. Solid State Electrochem. 2007, 11, 915–921. [CrossRef] Carrera-Cerritos, R.; Baglio, V.; Aricò, A.S.; Ledesma-Garcia, J.; Sgroi, M.F.; Pullini, D.; Pruna, A.J.; Mataix, D.B.; Fuentes-Ramirez, R.; Arriaga, L.G. Improved Pd electrocatalysis for oxygen reduction reaction in direct methanol fuel cell by reduced graphene oxide. Appl. Catal. B Environ. 2014, 144, 544–560. [CrossRef] Shao, M. Palladium based electrocatalysts for hydrogen oxidation and oxygen reduction reactions. J. Power Sources 2011, 196, 2433–2444. [CrossRef] Golikand, A.N.; Lohrasbi, E.; Maragheh, M.G.; Asgari, M. Carbon nano-tube supported Pt-Pd as methanol-resistant oxygen reduction electrocatalyts for enhancing catalytic activity in DMFCs. J. Appl. Electrochem. 2009, 39, 2421–2431. [CrossRef] Limpattayanate, S.; Hunsom, M. Effect of supports on activity and stability of Pt-Pd catalysts for oxygen reduction reaction in proton exchange membrane fuel cells. J. Solid State Electrochem. 2013, 17, 1221–1231. [CrossRef]

8006

Materials 2015, 8, 7997–8008

12. 13.

14. 15.

16. 17.

18.

19.

20.

21.

22. 23.

24.

25. 26. 27.

28. 29. 30.

31.

Zhang, J.; Sasaki, K.; Sutter, E.; Adzic, R.R. Stabilization of platinum oxygen-reduction electrocatalysts using gold clusters. Science 2007, 315, 220–222. [CrossRef] [PubMed] Borup, R.; Meyers, J.; Pivovar, B.; Kim, Y.S.; Mukundan, R.; Garland, N.; Myers, D.; Wilson, M.; Garzon, F.; Wood, D.; et al. Scientific aspects of polymer electrolyte fuel cell durability and degradation. Chem. Rev. 2007, 107, 3904–3951. [CrossRef] [PubMed] Antolini, E. Carbon supports for low-temperature fuel cell catalysts. Appl. Catal. B Environ. 2009, 88, 1–24. [CrossRef] Alegre, C.; Gálvez, M.E.; Moliner, R.; Baglio, V.; Aricò, A.S.; Lázaro, M.J. Towards an optimal synthesis route for the preparation of highly mesoporous carbon xerogel-supported Pt catalysts for the oxygen reduction reaction. Appl. Catal. B Environ. 2014, 147, 947–957. [CrossRef] Pumera, M. Electrochemistry of graphene, graphene oxide and other graphenoids: Review. Electrochem. Commun. 2013, 36, 14–18. [CrossRef] Sebastián, D.; Alegre, C.; Gálvez, M.E.; Moliner, R.; Lázaro, M.J.; Aricò, A.S.; Baglio, V. Towards new generation fuel cell electrocatalysts based on xerogel-nanofiber carbon composites. J. Mater. Chem. A 2014, 2, 13713–13722. [CrossRef] Zeng, J.; Francia, C.; Dumitrescu, M.A.; Monteverde Videla, A.H.A.; Ijeri, V.S.; Specchia, S.; Spinelli, P. Electrochemical performance of Pt-based catalysts supported on different ordered mesoporous carbons (Pt/OMCs) for oxygen reduction reaction. Ind. Eng. Chem. Res. 2012, 51, 7500–7509. [CrossRef] Sebastián, D.; Garcìa Ruiz, A.; Suelves, I.; Moliner, R.; Lázaro, M.J.; Baglio, V.; Stassi, A.; Aricò, A.S. Enhanced oxygen reduction reaction activity and durability of Pt catalysts supported on carbon nanofibers. Appl. Catal. B Environ. 2012, 115–116, 269–275. [CrossRef] Ferreira, P.J.; Shao-Horn, Y.; Morgan, D.; Makharia, S.; Kocha, S.; Gasteiger, H.A. Instability of Pt/C electrocatalysts in proton exchange membrane fuel cells: A mechanistic investigation. J. Electrochem. Soc. 2005, 152, A2256–A2271. [CrossRef] Stassi, A.; Gatto, I.; Baglio, V.; Passalacqua, E.; Aricò, A.S. Oxide-supported PtCo alloy catalyst for intermediate temperature polymer electrolyte fuel cells. Appl. Catal. B Environ. 2013, 142–143, 15–24. [CrossRef] Walsh, F.C.; Wills, R.G.A. The continuing development of Magnéli phase titanium sub-oxides and Ebonexr electrodes. Electrochim. Acta 2010, 55, 6342–6351. [CrossRef] Siracusano, S.; Baglio, V.; D’Urso, C.; Antonucci, V.; Aricò, A.S. Preparation and characterization of titanium suboxides as conductive supports of IrO2 electrocatalysts for application in SPE electrolyzers. Electrochim. Acta 2009, 54, 6292–6299. [CrossRef] Denaro, T.; Baglio, V.; Girolamo, M.; Antonucci, V.; Aricò, A.S.; Matteucci, F.; Ornelas, R. Investigation of low cost carbonaceous materials for application as counter electrode in dye sensitized solar cells. J. Appl. Electrochem. 2009, 39, 2173–2179. [CrossRef] Brocato, S.; Serov, A.; Atanassov, P. pH dependence of catalytic activity for ORR of the non-PGM catalyst derived from heat-treated Fe-phenantroline. Eletrochim. Acta. 2013, 87, 361–365. [CrossRef] Pattabiraman, R. Electrochemical investigations on carbon supported palladium catalysts. Appl. Catal. A Gen. 1997, 153, 9–20. [CrossRef] Siracusano, S.; Stassi, A.; Modica, E.; Baglio, V.; Aricò, A.S. Preparation and characterisation of Ti oxide based catalyst supports for low temperature fuel cells. Int. J. Hydrogen. Energy 2013, 38, 11600–11608. [CrossRef] Aricò, A.S.; Bruce, P.; Scrosati, B.; Tarascon, J.M.; Schalkwijk, W.V. Nanostructured materials for advanced energy conversion and storage devices. Nat. Mater. 2005, 4, 366–377. [CrossRef] [PubMed] Kinoshita, K. Particle size effects for oxygen reduction on highly dispersed platinum in acid electrolytes. J. Electrochem. Soc. 1990, 137, 845–848. [CrossRef] Alexeyeva, N.; Sarapuu, A.; Tammeveski, K.; Vidal-Iglesias, F.J.; Solla-Gullòn, J.; Feliu, J.M. Electroreduction of oxygen on Vulcan carbon supported Pd nanoparticles and Pd-M nanoalloys in acid and alkaline solutions. Electrochim. Acta 2011, 56, 6702–6708. [CrossRef] Pech-Pech, I.E.; Gervasio Dominic, F.; Godìnez-Garcia, A.; Solorza-Feria, O. Nanoparticles of Ag with a Pt and Pd rich surface supported on carbon as new catalyst for the oxygen electroreduction reaction (ORR) in acid electrolytes: Part 1. J. Power Sources 2015, 276, 365–373. [CrossRef]

8007

Materials 2015, 8, 7997–8008

32.

33. 34.

Aricò, A.S.; Stassi, A.; Modica, E.; Ornelas, R.; Gatto, I.; Passalacqua, E.; Antonucci, V. Performance and degradation of high temperature polymer electrolyte fuel cell catalysts. J. Power Sources 2008, 178, 525–536. [CrossRef] Shao, Y.; Liu, J.; Wang, Y.; Lin, Y. Novel catalyst support materials for PEM fuel: Current status and future prospects. J. Mater. Chem. 2009, 19, 46–59. [CrossRef] Aricò, A.S.; Stassi, A.; D’Urso, C.; Sebastiàn, D.; Baglio, V. Synthesis of Pd3 Co1 @Pt/C core-shell catalysts for methanol-tolerant cathodes of direct methanol fuel cells. Chem. Eur. J. 2014, 20, 10679–10684. [CrossRef] [PubMed] © 2015 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).

8008