Cesium and Rubidium Salts of Keggin-type

SCHOOL OF ADVANCED STUDIES Doctorate course in Chemical Sciences

PhD thesis Cesium and rubidium salts of Keggin-type heteropolyacids as stable meso-microporous matrix for anode catalyst for H2/O2 Proton Exchange Membrane Fuel Cell, Direct Methanol Fuel Cell and Direct Ethanol Fuel Cell

Cycle XXI Scientific-sector CHIM/01

PhD Candidate

Tutor

Artur śurowski

Professor Roberto Marassi Professor Paweł J. Kulesza

2005/2006 – 2007/2008

Table of contents Overview of the dissertation ……………………………………………………………4 Introduction 1. Fuel cell……………………………………………………..………………………7 1.1 History………………………...…………………………………………7 1.2 Basic principles, types and efficiency………………………………...…8 2. Anode electrocatalysis in PEMFC…………………………………………..…...…14 2.1 Hydrogen oxidation…………………………………………………….14 2.2 Methanol oxidation……………………………………………………..16 2.3 Ethanol oxidation…………………………………………………….....20 3. Heteropoly compounds……………………………………….............…….............24 3.1 Structure of heteropoly compounds (heteropolyacids)…………………24 3.2 Properties of heteropolyacids…………………………………………..27 3.3 Salts of the heteropoly compounds…………………………………..…30 4.

Experimental techniques……..………………………………..………………..34 4.1 Cyclic voltammetry……………………………………………….…….34 4.1.1

Rotating disk voltammetry…………………………………………37

4.2 Chronoamperometry……………………………………………………39 4.3 Staircase voltammetry………………………………………………….40 4.4 Electrochemical impedance spectroscopy……………………………...41 4.5 Transmission electron microscopy ……………...……………………..43 4.6 Scanning electron microscopy ……..…………………………………..45 4.7 Infrared spectroscopy………………………………………………….46 Experimental part 5. Chemical reagents and measuring equipment…………………..………..………...48 6. Preparation and characterization of Keggin-type matrix………………..………….51 6.1 Preparation of Keggin-type heteropolyacid salts…………..…………...51 6.2 IR characterization of Keggin-type matrix……………………………..51 6.3 SEM characterization of Keggin-type matrix…………………………..54 6.4 Cyclic voltammetry characterization of Keggin-type matrix…………..57 6.5 Conclusions…………………………………………………………….61 7.

Hydrogen oxidation reaction (HOR) on the catalytic layers containing Cs2.5PW12 matrix…………………………………………………....62 1

7.1

Preparation of the catalytic layers…………………………….……..62

7.1.1

Mixing method……………………………………………………..62

7.1.2

Electrochemical method……………………………………………63

7.2 Electrochemical measurements on the Pt/C modified with Cs2.5PW12 system prepared by mixing method……………………….65 7.2.1

Cyclic voltammetry (CV) and rotating disc voltammetry (RDE) measurements……………………………………………….65

7.2.2

Test of the Cs2.5H0.5PW12O40-containing anode’s catalyst in working single PEM fuel cell…………………………...74

7.3 Electrochemical measurements on the Pt/C modified with Cs2.5H0.5PW12O40 system prepared by electrochemical method..…….78 7.3.1

Cyclic voltammetry (CV) study….……………………….………..78

7.3.2

CO electrooxidation on the catalytic layer containing Cs2.5H0.5PW12O40 as a matrix……………………………………….80

7.3.3

HRTEM characterization…………………………………………..81

7.3.4

Rotating disc voltammetry (RDE) measurements…………………83

7.4

Comparison of catalytic layers containing Cs2.5PW12 matrix prepared by mixing and electrochemical methods……………88

8. Methanol oxidation reaction (MOR) on the catalytic layers containing Keggin-type heteropolyacid salts as a matrix……………..…….……...90 8.1 Preparation of the catalytic layers……………………………………...90 8.2 Electrochemical measurements at the Pt40%/C modified with Cs2.5-HPAs matrix………………………………...……91 8.2.1 Cyclic voltammetry (CV) study……………………………………...91 8.2.2 Staircase voltammetry (SV) measurements…………………………..95 8.2.3 Chronoamperometry (CA) measurements……………………………97 8.2.4 Electrochemical impedance spectroscopy for methanol electrooxidation ………………………………………..99 8.2.5 CO stripping voltammetry study at Pt/C-Cs2.5HPAs…….………….105 8.2.6 Electrochemical stability of investigated materials containing Cs2.5-HPAs matrix………………………………………107

2

8.3 Electrochemical measurements at system containing Rb2.5-HPAs matrix…………………………….………...108 8.3.1 Cyclic voltammetry (CV) measurements…………………………...108 8.3.2 Staircase voltammetry (SV) measurements…………………………112 8.3.3 Chronoamperometry (CA) measurements…………………………..114 8.3.4 Electrochemical impedance spectroscopy measurements…………..116 8.3.5 CO stripping voltammetry study at Pt/C modified with Rb2.5-HPAs………...………………………………...118 8.3.6 Electrochemical stability of investigated materials containing Rb2.5-HPAs……………………………………………...120 8.4 Summary and conclusion……………………………………………..122 9. Ethanol oxidation reaction (EOR) on the Pt40%/Vulcan XC-72 carbon modified with Cs2.5-HPAs………………………………………………………...127 9.1 Cyclic voltammetry (CV) measurements……………………….…….127 9.2 Staircase voltammetry (SV) measurements…………………………..130 9.3 Chronoamperometry (CA) measurements……………………………132 9.4 Electrochemical impedance spectroscopy for ethanol electrooxidation…………………………………………...134 9.5 Electrochemical stability of investigated materials containing Cs2.5-HPAs………………………………………………...136 Conclusions…………………………………………………………………………...138 Bibliography…………………………………………………………………………..141

3

Overview of the dissertation In various applications, fuel cells are widely recognized as very attractive devices to obtain directly electric energy from the electrochemical combustion of chemical products. Low temperature fuel cells, which generally utilize proton electrolyte membranes, seem to be also of utility for a large of power applications. However, the final choice of the fuel is still difficult and depends greatly on the field of application. Utilization of pure hydrogen or hydrogen-rich gases, rather than alcohols, as fuels in Polymer Electrolyte Membrane Fuel Cell (PEMFC) leads to higher electric efficiencies. Due to the problems related to the hydrogen storage, the hydrogen-oxygen PEMFC is the best choice for stationary applications. Direct alcohols (methanol and ethanol) fuel cells based on solid polymer electrolytes are widely proposed for portable and mobile applications due to relatively low prices of the methanol and ethanol fuels and easiness of storage. The major impediment in the development of fuel cells operating on hydrogen, methanol, and ethanol is deactivation of the anode electrocatalyst by trace level of CO. Thus, the most active Pt electrocatalyst for the oxidation of hydrogen and alcohols (e.g. methanol, and ethanol) is deactivated by strong adsorption of carbon monoxide, and leads to decreased performance of fuel cell. Therefore, a new inexpensive, stable electrocatalyst must be developed, which is tolerant to high levels of CO (particularly, for direct alcohol fuel cell) or that could preferably utilize CO. There are currently two state-of-the-art methods which increase CO tolerance of the fuel cell anode. One method is to use a Pt-alloy catalyst (e. g. PtRu, PtSn, PtMo, PtW).[1-6] The other method is to use the zeolite as matrix, and limit the preferential formation of CO clusters on platinum by the steric constraints imposed by the zeolites framework, followed by facile oxidation to CO2 by interaction with the surface or bridged hydroxyls of the zeolites. Furthermore, the porous nature of the zeolite support material provides relatively improved gas permeability and minimizes the disadvantage associated with restricted gas diffusion in the electrode. The ideal support would also an enhanced electrochemically active surface area by dispersion of the metal catalyst.[7, 8] Keggin-type heteropolyacids show appreciable acid catalytic properties[9-14] that are of practical importance[15] as illustrated in numerous recent reviews.[13, 15-18] The fact that heteropolyanions undergo spontaneous adsorption (from aqueous solution) on

4

various substrates[32,33] provides a simple tool for modification of electrode surfaces[1922]

Among other important issues related to electrocatalysis are their ability to undergo

fast reversible multi-electron transfers and super-acid properties resulting in the increased availability and mobility of protons at the electrocatalytic interfaces. Consequently, polyacids were considered for fuel cell research.[23-29] With respect to the oxygen reduction, the adsorbed heteropolyacid (particularly H3PW12O40) nanostructures do not seem to block access of reactant molecules[21] to catalytic Pt. Further, their presence at the interface shifts formation of the inhibiting Pt-oxo (PtOH or PtO) species towards more positive potentials thus increasing the potential range where catalytic metallic (Pt0) sites exist. It is reasonable to expect that, by analogy to the activating role of WO3 and partially reduced hydrogen tungsten oxide bronzes during elect oxidation of methanol, the related heteropolyblue tungstates should also enhance reactivity of Pt during oxidation of organic fuels.[26,

30]

Stanis and co-workers[27] reported that the

addition of adsorbed HPAs can improve the performance of Pt anodes in a fuel cell under CO-poisoned conditions. Also Farell et al. postulated that HPAs can act as cocatalysts with platinum for methanol electrooxidation.[28] Among limitations in practical applications of heteropolyacids is their very good solubility in aqueous solutions including acids as well as their ability to undergo desorption during long-term operation. Thus it is necessary to stabilize heteropolyacid layers at electrocatalytic interfaces without loosing their activating properties. An interesting alternative arises from the possibility of formation of acidic salts of heteropolytungstates or molybdates by partial exchange of protons with large cations (such as Cs+, Rb+, NH4+ or K+) in the parent heteropolyacid. Consequently, a watersoluble polyacid of low surface area (<5 m2g-1) is transferred into a water-insoluble acid salt precipitate characterized by the surface area exceeding 100 m2g-1.[31, 32] Contrary to zeolites exchanged with alkali metals[14, 33], heteropolyacid salts remain strong acidity. The resulting materials have occurred to be efficient solid shape-selective acid catalysts for a variety of organic liquid-phase reactions[18] that include hydrogenations and oxidations. The pore size and acidity of heteropolyacid salt can be controlled by the cation content. For example, when the Cs content, x, is initially increased in CsxH3xPW12O40

from 0 to 2, the number of surface protons decreases but, later, it significantly

increases when x changes from 2 to 3 to show the highest surface acidity at x = 2.5.[14] In the present work, we consider the salts of Keggin-type heteropolyacids containing 2.5 moles of Cs+ and Rb+ cations in 1 mole of the heteropoly salt. The 5

system with such Cs or Rb-content is micro-mesoporous, and it is characterized by very good stability and, while being insoluble in water, it exhibits high acidity (proton availability and mobility). The aim of this work is to study the applicability of the Cs and Rb salts of Keggin type heteropolyacids as a stable meso-microporous matrix for anode catalyst for H2/O2 Proton Exchange Membrane Fuel Cell, Direct Methanol Fuel Cell and Direct Ethanol Fuel Cell. The experimental part is divided into four chapters (from 6 to 9). In the Chapter 6, we present characterization of cesium, rubidium and ammonium salts of Keggin types heteropolyacids by infrared spectroscopy (IR), scanning electron microscopy (SEM), and cyclic voltammetry (CV) measurements. In Chapter 7, we illustrate incorporation and activation of Pt centers in the conductive high surface-area zeolite-type robust, Cs2.5H0.5PW12O40 matrices through application of mixing and electrochemical methods. To evaluate electrocatalytic activity towards the hydrogen oxidation of the investigated electrode material, we have performed the diagnostic rotating disc electrode voltammetric measurements. In a case of the catalytic layer prepared by corrosion of Pt counter electrode (electrochemical method), CO-stripping and HRTEM measurements have been performed to comment the electrochemically active area of catalyst, dimensions of Pt particles and platinum loading. In Chapter 8, we report on the performance of electrocatalysts (prepared by mixing method) towards methanol oxidation in acidic media. The Cs and Rb salts of H3PW12O40, H3PMo12O40, H4SiW12O40, H4SiMo12O40 were used as zeolite matrix for Pt40%/Vulcan XC-72 carbon nanoparticles (Pt/C). The techniques of cyclic voltammetry, staircase voltammetry, chronoamperometry, electrochemical impedance spectroscopy, CO stripping voltammetry were applied to compare the Pt-base electrocatalyst activity and stability to methanol oxidation. In Chapter 9, the system composed of Pt40%/Vulcan XC-72 carbon modified with Cs2.5-HPAs matrix (prepared by mixing method) has been examined with respect to ethanol electrooxidation by several different electrochemical techniques. Comparison has been made to commercial Pt/C.

6

1. Fuel cell

1.1 History The principle of the fuel cell was discovered by the German scientist, who in 1839 published paper about current production in reaction between hydrogen and oxygen. Based on this work, in 1839, British physicist and lawyer, Sir William R. Grove develops the first working fuel cell. By connecting a hydrogen anode and an oxygen cathode, he produced an electric current with the experimental set-up shown in Fig. 1.

Fig.1. Four-cell version of Grove’s gas battery.[34]

He also, in 1844/1845, presented the first fuel cell power generator which consisted 10 cells connected in series and was supplied with hydrogen from corrosion of zinc in acid.

Fig. 2. Grove’s fuel cell power generator.[34]

In 1905 Wilhelm Ostwald and Walter H. Nerst presented a general theory of the fuel cells. Due to easily accessible and large amounts of oil and the invention of the

7

combustion engine by Carl Friedrich Benz and Gottlieb Daimler fuel cells were forgotten until the middle of the 20th. Because of US Apollo space programme in the 1960’s fuel cells exhibited their first renaissance. Gemini 5 was the first space shuttle using a polymer membrane fuel cell instead of battery. In 1969 alkaline fuel cells were used in the Apollo missions and supplied the electric power when the USA landed on the moon. The first oil crisis in 1973 led to the second renaissance of the fuel cells. At this time interest for large power plants based on high-temperature fuel cells considerably increased. In 1970 professor Karl Kordesch (early fuel cell pioneer) form University of Graz, Austria, and his co-workers developed an alkaline fuel cell motorbike and a car. The focus on increasing pollution level over the last two to three decades has forced world society to search for cleaner energy technology, and thus fuel cells have experienced an exponential increase in attention.

1.2 Basic principles, types and efficiency A fuel cell is electrochemical device that continuously converts chemical energy into electric energy (and some heat) for as long as fuel oxidant is supplied. In the exchange membrane fuel cell (PEMFC), hydrogen (fuel) is oxidised to protons and electrons at the anode. Protons migrate through the membrane electrolyte to the cathode. As the membrane is an electric insulator, electrons are forced to flow in an external electric circuit. At the cathode, oxygen (oxidant) reacts with protons to produce water, which is the only waste product from a hydrogen-operated PEMFC. A schematic representation of a fuel cell with the reactant/product gases and the ion conduction flow directions through the cell is shown in Figure 3. Anode reaction: H2 →2H+ + 2e-

(1)

Cathode reaction: ½O2 + 2H+ + 2e- →H2O

(2)

Total cell reaction: H2 + ½O2 →H2O

(3)

8

Fig.3. Schematic of an individual fuel cell.[35]

From hydrogen and oxygen we obtain water, heat and power. There are other fuels, electrolytes and charge-transferring ions for the other fuel cell types - but the principle is the same. When an external resistance, commonly referred to as a “load”, is applied to the cell, non-equilibrium exists and a net current flows through the load. The net rate of an electrochemical reaction is proportional to the current density which is defined as the current of the electrochemical system divided by the active area devices. The cell voltage becomes smaller as the net reaction rate increase because of irreversible losses.[36] A representative polarization curve for hydrogen-oxygen PEMFC is illustrates in Fig. 4. The voltage-current density relationship for a given fuel cell (geometry, catalyst/electrode characteristic, and electrolyte/membrane properties) and operating conditions (concentration, flow rate, pressure, temperature) is a function of the kinetic, ohmic, and mass transfer resistance. Fuel crossover and internal current losses result from the flow of fuel and electric current in the electrolyte. The electrolyte should only transport ions, however a certain fuel and electron flow will always occur. Although the fuel loss and internal currents are small, they are the main reason for the real open circuit voltage (OCV) being lower than the theoretical one (Erev).

9

Fig. 4. A typical performance curve for a solid polymer fuel cell showing the relative effects of electrode activation (kinetic losses), ohmic resistance and mass transport losses.[37]

Activation losses are caused by the slowness of the reactions taking place on the electrode surface. The voltage decreases somewhat due to the electrochemical reaction kinetics. The ohmic losses result from resistance to the flow of ions in the electrolyte and electrons through the cell hardware and various interconnections. The corresponding voltage drop is essentially proportional to current density, hence the term "ohmic losses". Mass transport losses result from the decrease in reactant concentration at the surface of the electrodes as fuel is used. At maximum (limiting) current, the concentration at the catalyst surface is practically zero, as the reactants are consumed as soon as they are supplied to the surface. In the thesis the investigation was principally concerned with the activation losses of low temperature fuel cells.

Fuel cells can be classified according to their operating temperature, electrolyte and the corresponding conductive ions (Tab. 1).

10

Alkaline Fuel Cell (AFC)

Proton Phosphoric Exchange Acid Fuel Cell Membrane Fuel (PAFC) Cell (PEMFC)

Molten Carbonate Fuel Fell (MCFC)

Solid Oxide Fuel Cell (SOFC)

Operating temperature

70-220°C

Up to 120°C

130-220°C

600-800°C

700-1000°C

Electrolyte

Potassium hydroxide (KOH)

Polymer membrane

Concentrated phosphoric acid

Melted Li/K carbonate

Solid oxide ceramic

Fuels

Pure hydrogen

Hydrogen, natural gas

Hydrogen, natural gas

Hydrogen, natural gas

Power range realised Application

Up to 12 kW

Hydrogen (+reformate), methanol, ethanol Up to 250 kW

Up to 1 MW

Up to 2 MW

Up to 10 MW

Portable, Small power mobile, APU, plants, APU, CHP CHP APU - Auxiliary Power Unit, CHP - Combined Heat and Power

Power plants

Power plants, APU, CHP

Space, submarine

Table. 1. Typical properties of the different fuel cell types.[37]

To the low-temperature group of fuel cells we can include Polymer Electrolyte Membrane Fuel Cells (PEMFC), Direct Methanol Fuel Cells (DMFC), Direct Ethanol Fuel Cells, Phosphoric Acid Fuel Cells (PAFC) and Alkaline Fuel Cells (AFC). Solid Oxide Fuel Cells (SOFC) and Molten Carbonate Fuel Cells (MCFC) are hightemperature fuel cells.

The most important disadvantage of fuel cell at the present time is the same for all types – the cost. However, the are varied advantages[38]:


Efficiency. Fuel cell is generally more efficient than combustion engines whether piston or turbine based. They are small system which can be just as efficient as large ones. This is very important in the case of the small local power generating system needed foe combined heat and power system. Fuel cells are not limited due to the Carnot theorem therefore, they are more efficient in extracting energy from a fuel than conventional power plant. Waste heat from some cells can also be harnessed, boosting system efficiency still further. Typical efficiency of different fuel cells is placed below, in Table 2.

11

Type

Efficiency

Solid Oxide

45 – 65%

Molten Carbonate

50%

Phosphoric Acid

40%

Alkaline

50 – 60%

Direct Methanol

40%

Proton Exchange

40%

Membrane (PEM) Table. 2. Typical efficiency of fuel cells.[37]


Simplicity. The essential of a fuel cell are very simple, with few if any moving parts. This can lead to highly reliable and long-lasting system.


Low emissions. The by-product of the main fuel cell reaction, when hydrogen is the fuel, is pure water, which means a fuel cell can be essentially ‘zero emission’. This is their main advantage when use in vehicles, as there is a requirement to reduce vehicle emissions, and even eliminated them within cities.

Water Vapor g / mile

CO2 g /mile

CO g /mile

NOx g /mile

Hydro Carbons g /mile

176.90

415.49

20.9

1.39

2.80

Gasoline ICE Light Truck 1

N/a

521.63

27.7

1.81

3.51

Methanol FC 2

113.40

68.04

0.016

0.0025

0.0034

Hydrogen FC 2

113.40

0.00

0.00

0.00

0.00

Engine Type

Gasoline ICE Passenger Car 1

1 2000 U.S. EPA Average Annual Emission for Passenger Cars and Light Trucks 2 Calculations from Desert Research Institute 1mile = 1.6 km

Table. 3. Systems emissions.[39]

12


Silence. Fuel cell are very quiet, even those with extensive extra fuel processing equipment. This is very important in both portable power applications and for local power generation in combined heat and power schemes. The advantages of fuel cell impact particularly strongly on combined heat and power system (for both large and small scald applications), and on mobile power systems, especially for vehicles and electronic equipments such as portable computers, mobile telephones, and military communications equipment. These areas are the major fields in which fuel cells are being used. A key point is the wide range of applications of a fuel cell power, from systems of a few watts up to megawatts.

Table. 4. Chart to summarized the application of fuel cells of different types and different applications.[38]

Among the applicable low-temperature acid-type systems, polymer electrolyte membrane fuel cells (PEMFCs), direct methanol fuel cells (DMFCs) and direct ethanol fuel cells (DEFCs) are probably the most promising devices, and they are subjects of interest in many laboratories worldwide. While in the case of PEMFCs, hydrogen is utilized; the methanol or ethanol fuel is oxidized in the anodic compartment of DMFCs or DEFCs, respectively. To improve power densities of PEMFCs, DMFCs, and DEFs there is a need to develop new electrocatalyst to inhibit the poisoning and significantly increase the rate of electrooxidation. This study has been concerned with the catalysts of the three low temperature fuel cells, the PEMFC, the DMFC and the DEFC.

13

2. Anode electrocatalysis in PEMFC

2.1 Hydrogen oxidation The hydrogen oxidation reaction (HOR) is very important reaction which, serve as a model for electrocatalytic reaction, and has significant practical utility in some kinds of electrochemical sensors and fuel cells.[40-57] In acid media the catalysis of the reaction is usually promoted by the platinum catalyst dispersed as fine particles on high surface area carbon supports.[44,58,59] One important class of such devices are those assembled using polymer electrolyte (PE) and polymer impregnated gas diffusion electrodes, most commonly employing Nafion® (E.I. DuPoint, U.S.A.) membranes and ionomers. The electrochemical oxidation of dihydrogen: H2 → 2H+ + 2e-

(E0 = 0.00 V vs. RHE)

(4)

on noble metal surfaces such as Pt and Pd is very facile.[44] Other metals also show high activity for H2 electrooxidation, but in acidic electrolytes, noble metals show the greatest stability towards corrosion or passivation. The accepted mechanism of hydrogen electro-oxidation on Pt in acidic electrolytes is formed by a primary chemical: H2 + 2Pt → 2Pt-Hads

(Tafel reaction)

(5)

and / or electrochemical: H2 + Pt → Pt-Hads + H+ + e- (Heyrovsky reaction)

(6)

adsorption steps, followed by a discharge path of the adsorbed hydrogen atom given by MH → M + H+ + e-

(Volmer reaction)

(7)

For a given electrode material and the electrolyte, the rates of reactions (5) and (6) may be quite different, and the mechanism may be formed preponderantly by the Tafel/Volmer steps or alternatively by the Heyrovsky-Volmer steps. On Pt electrode, a Tafel-Volmer mechanism, with dissociative adsorption of dihydrogen (Tafel reaction) being the rate determining step (rds), has been proposed for this reaction[34,44,45,51] in acidic electrolytes. 14

While pure H2 is the ideal choice of fuel for the PEMFC, economical sources of pure H2 are not readily available. Therefore, currently the most practical source of H2 is the catalytic processing of hydrocarbons. Hydrogen produced by the stream reforming or partial oxidation of hydrocarbon fuels (gasoline, diesel, methane, alcohols) contains impurities such as CO (1 - 3%) and CO2 (19 – 25%). It is well known that CO binds strongly Pt sites and reduces the number its catalytic centers available for H2 adsorption and oxidation. Although the electrochemical oxidation of CO: CO + H2O → CO2 + 2H+ + 2e-

(E0 = -0.1 V vs. RHE)

(8)

is thermodynamically favorable, in practice a large overpotential is required on pure Pt surfaces before oxidation occurs. For example, on dispersed Pt catalysts, the onset of CO oxidation is no observed until 0.5 V at 80 0C. Therefore, in the potential region where anodes need to operate (i.e. 0 - 0.1 V), CO is an inert adsorbate. The degree of CO poisoning of Pt catalysts is very dependent on both temperature and CO concentration.[34] CO2 poisoning on pure Pt catalysts is modest when compared to the effect of CO, especially when the large differences in relative concentrations in reformate are considered. The poisoning effect comes from two possible mechanisms: 1) H2 + CO2 → H2O + CO

(9)

2) CO2 + 2Pt-Hads → Pt-CO + H2O + Pt

(10)

Both are forms of the “reverse water-gas shift” reaction with (9) being the familiar gas-phase reaction, and (10) the electrochemical equivalent. In both cases, the product is CO, which has the same effect as fuel stream CO. The most elegant way to overcome anode poisoning is through the development of CO- and CO2- tolerant electrocatalysts. Much effort has been spent modifying Pt with others metals to improve CO tolerance. Niedrach et al.[60], in early 1960s, found that the addition of Ru, Rh and Ir to Pt in the form of unsupported mixed metal powders (blacks) get substantial tolerance, compared to Pt black alone, to the presence of CO in a fuel stream at 85 0C in 2.5 M H2SO4 electrolyte. Moreover, they found that mixing Pt blacks with metal oxides, such as CoMoO4, MoO2 and WO3, also improved CO tolerance over Pt alone.[60] In particular, the oxide-containing electrodes showed

15

remarkable performances with pure CO as a fuel, with reported activities approaching those with pure H2 even at low potentials.[34] In the recent years, a number of workers have reported the superior CO and CO2 tolerance of carbon-supported PtRu catalysts when compared to Pt-only catalysts.[1,61-67]

2.2 Methanol oxidation

Introduction The electro- oxidation of methanol is very important from practical point of view and has been studied for more than three decades.[1,28,61-67,68-97] Methanol is used as a liquid fuel in Direct Methanol Fuel Cell (DMFC). Among the different possible alcohols, methanol is the most promising organic fuel because its use as a fuel has several advantages in comparison to hydrogen: high solubility in aqueous electrolytes, liquid fuel available at low cost, easily handled, transported and stored, high theoretical density of energy (6 kWh/kg) comparable to that of gasoline (10-11 kWh/kg).[68,69,97]

Fig. 5. Sketch of a DMFC illustrating proton, water and methanol permeation across the PEM and related characterization methods.[98]

The sketch of the DMFC is shown in Fig. 5.[98] The DMFC consist of an anode at which methanol is electro-oxidized to CO2 and a cathode where oxygen, generally from air, is reduced to form water. Both electrodes, usually formulated with platinum or Pt-based catalysts, are separated by a proton-conducting electrolyte. The direct

16

methanol oxidation involves the transfer of six electrons to the electrode for complete oxidation to carbon dioxide. In an acidic medium, this reaction can be written as follows:
Anode reaction:

CH3OH + H2O → CO2 + 6H+ + 6e-

(11)


Cathode reaction: 3

/2 O2 + 6H+ + 6e- → 3 H2O

(12)


Overall reaction:

CH3OH + 3/2 O2 → CO2 + 2 H2O

(13)

The free energy associated with the overall reaction at 25 oC and 1 atm and the electromotive force are[77,99]: ∆G = -686 kJ mol-1CH3OH; ∆E = 1.18 V vs. SHE Methanol oxidation reaction Iwastia[84] reported that the thermodynamic potential for methanol oxidation to CO2, is very close to the equilibrium potential for hydrogen (E0=0.02 V), Eq. 11. However, compared with hydrogen oxidation, this reaction is by several orders of magnitude slower and requires high loading Pt-based electrocatalyst (≈ 2 mg cm-2).[77] In fact, the ideal anodic reaction is not completely reached as methanol is mainly decomposed into CO, which can further be oxidized into CO2. The formations of CO and CO2 are assumed occurring according to the dual path mechanism in the oxidation, one leading to CO (12) and another to CO2 (11)[100]: CH3OH → CO+ 4H+ + 4e-

(14)

Other CO-like species are also formed during the adsorption phase, COHads, HCOads, HCOOads, and the principle by-products are formaldehyde and formic acid.[77,93, ,101,102] The generally accepted mechanism of methanol electro-oxidation pathway is considered as follows[83,85,103]: Pt + CH3OH → Pt(CH3OH)ads

(15)

Pt(CH3OH)ads → Pt(CH4-nO) + n H+ + n e-

(16)

M + H2O → M-OH + H+ + e-

(17)

Pt-CO + M-OH → CO2 + H+ + e+

(18) -

Pt-CO + H2O → CO2 + 2 H + 2 e

(19)

17

where, M is a metal, e.g. Pt, Ru, W. The complexity of the methanol oxidation reaction on involving different path of reaction is shown in Fig. 6.

Fig. 6. Detailed reaction mechanism of the oxidation of methanol on a platinum electrode.[97]

Since the complete oxidation of methanol to CO2 involves the transfer of 6 electrons to the electrode, the overall reaction mechanism (Eqs. 15 – 19) involves several steps including dehydrogenation, chemisorption of methanolic residues, rearrangement of adsorbed residues, chemisorption of oxygenated species (preferentially on the alloying element) and surface reaction between CO and OH to give CO2. There is a reasonable consensus regarding the general mechanisms at different potentials on Pt surfaces. It is known that at low potentials, less than 0.4 V vs. RHE, the rate determining step for Pt and PtRu electrode is the methanol dissociative adsorption expressed by Eqs. (15) and (16).[82,84] The most important potential regime for a fuel cell is ca. 0.4 - 0.45 V vs. RHE and the consensus agreement is that at this potential the rate-determining process is the oxidation of CO.[34,82] Clean Pt is a good catalyst for methanol oxidation and initially show very high activity for methanol oxidation, but these very rapidly decay in current on the formation of strongly bound intermediates.[62,77,104-106] These intermediates (mainly COads) are only removed on going to high overpotentials where they are oxidized. At potentials below 0.45 V, the surface of Pt becomes poisoned with a near-monolayer coverage of CO, and

18

further adsorption of water or methanol cannot occur. Hence, the methanol oxidation rate drops to an insignificant level.[34] The development of advanced Pt based catalysts has focused on the addition of a secondary component (e.g., Ru, Sn, W) that is able to provide an adsorption site capable of forming OHads species at low potentials adjacent to poisoned Pt sites. The adsorption site is also less effective at adsorbing methanol itself. The OHads is then able to react with the bound CO to produce CO2 and free sites for further methanol adsorption. Superior catalytic activities have been reported for Pt based alloys, such as Pt-Ru, PtMo, Pt-Sn, Pt-Os, Pt-Ru-Os.[61-66,106-108] This enhancement effect has been explained by models such as the “bifunctional mechanism”[64,109,110] and/or the “electronic effect”[65,111,112], which indicates a promotional effect of the alloyed metal on Pt. For “promoters” such as Ru, stable methanol oxidation currents occur at significantly lower potentials (<0.25 V) to Pt[34], indicating the Ru is capable of the formation of OHads without itself being poisoned by CO.[62,84] Ru + H2O → (OH)ads + H+ + e-

(20)

(CO)ads + (OH)ads → CO2 + H+ + e-

(21)

Until now the most successful results have been achieved through the alloying of Sn and Ru with Pt. It has been shown that these alloys give rise electrocatalysis which strongly promote the oxidation of both methanol and CO. The promotion of CO oxidation reaction on Pt-Sn catalysts appears to be mainly due to a modification of the electronic environment around Pt-sites.[113-115] It becomes clear that Pt-Ru catalysts are more effective for methanol oxidation since the reaction desires the electrocatalyst to be operated in a potential regime where labile-bonded oxygen should be present on the surface. In this situation, the supply of active oxygen to the surface is of paramount importance since this apparently would facilitate the oxidation of adsorbed methanolic residues to CO2. The presence of strongly-bonded oxygen species on Sn-sites in the PtSn system limits the oxidation of methanol to CO2.[113-115] Out of all tested catalysts, for the practical applications in DMFC the most active and stable catalysts are these based on platinum ruthenium alloys.[1,61,63,64,66,67,112]

19

2.3 Ethanol oxidation

Introduction Direct ethanol fuel cells (DEFCs) have spurred more and more in the recent years due to intrinsic advantages such as its low toxicity, renewability, and its easy production in great quantity by the fermentation from sugar-containing raw materials.[97,116] Furthermore, the high theoretical mass energy density (about 8.00 kWh/kg)[4,97] provides it with a potential candidate fuel for polymer electrolyte membrane fuel cells. Therefore, ethanol electro-oxidation reaction is extensively studied by numerous of researchers.[2-6,79,91,92,97,116-139]

Fig. 7. Schematic principle of a direct ethanol/oxygen fuel cell (DEFC).[117]

The parts of which a direct ethanol fuel cell (DEFC) constitutes and its working principle are shown in Fig. 7. At the anode (negative pole of the cell) the electrooxidation of ethanol aqueous solutions takes place as follows[97,117,118]: CH3CH2OH + 3 H2O → 2 CO2 + 12 H+ + 12 e-

(22)

E10 = 0.085 V vs. SHE whereas the cathode (positive pole) undergoes the electro-reduction of oxygen, i.e.:

O2 + 4 H+ + 4 e- → 2 H2O

(23)

E20 = 1.229 V vs. SHE

20

where Ei0 are the standard electrode potentials versus the standard hydrogen electrode (SHE). This corresponds to the overall combustion reaction of ethanol in oxygen: CH3CH2OH + O2 → 2 CO2 + 3 H2O

(24)

This gives a standard electromotive force (Eemf) equal to 1.145 V. The complete oxidation of ethanol involves 12 electrons per methanol molecule (Eq. 22). and the cleavage of the C-C bond. In view of the direct electro-oxidation of ethanol in the fuel cells, the materials that could facilitate ethanol complete oxidation and shift the onset oxidation potentials to lower values are the most interest.[119]

Ethanol oxidation reaction The complete oxidation of ethanol to CO2 is the central challenge in the electrocatalysis of this alcohol. In the recent years, several spectroscopic and electrochemical studies have been taken to evaluate the mechanism of ethanol electrooxidation on Pt-based catalysts.[2,79,92,97,120-122] A generally accepted reaction sequence comprises of the following steps (where M is an active site)[2]: M + CH3–CH2OH → M–CHOH–CH3 + H+ + e-

(25)

M–CHOH–CH3 → M–CHO–CH3 + H+ + e-

(26)

+

-

M–CHO–CH3 → M–CO–CH3 H + e

(27)

M–CO–CH3 + M → M–CO + M–CH3

(28)

M + H2O → M–OH + H+ + e-

(29)

M–CO + M–OH → 2M + CO2 + H+ + e-

(30)

Ethanol can also react with adsorbed hydroxyl species directly to produce acetate via a four-electron oxidation pathway[2,122]: C2H5OH + M–OH → M–CH3COO + 4H+ + 4e-

(31)

An efficient electrocatalyst should facilitate each of the processes of dehydrogenation (Eqs. (25)–(27)), C–C bond cleavage (Eq. (28)) and COads oxidation (Eq. (30)) for the complete conversion of ethanol to CO2 to take place. In addition, water activation (Eq. (29)) at low electrode potential is important for the subsequent COads oxidation step.

21

Platinum is recognized to be the most active material for ethanol oxidation, however, it should be noted that the self-inhibition happens in the case of Pt alone, especially in the steady state operation mode. Furthermore, in order to increase the fuel utilization and the fuel cell efficiency, it is crucial to break C-C bond and provoke its complete oxidation into carbon dioxide. Therefore, in order to improve the electrocatalytic activity of platinum to ethanol electro-oxidation, platinum was often modified by a second or a third additive like PtRu[2,3], PtRh[123], PtMo[124], PtSn[4,5], PtW[6], PtRuMeOx (Me = W, Mo, V ect.).[91,125-130] Fig. 8 presents the effect of different additives to Pt’s activity to ethanol electrooxidation in a single DEFC. As one can distinguish that all the additives can promote more or less the platinum’s electrocatalytic activity towards ethanol oxidation.[119]

Fig. 8. The effect of the different additives on the Pt’s activity toward ethanol electrooxidation in single direct ethanol PEMFCs under the same operation conditions. Tcell = 90 0C. Anode: PtM/C, 1.3 mg Pt/cm2, Cethanol = 1.0 mol/L, flow rate: 1.0 mL/min. Cathode: Pt/C (20%, Johnson Matthey Corp.), 1.0 mg/ cm2, PO2 = 2.0 atm. Electrolyte: Nafion1®-115 membrane.[119]

From the attempts of Lamy et al.[131] to identify the suitable electrocatalysts for ethanol oxidation, it was concluded that Sn can lead to encouraging results especially at lower potential values, three times increase in maximum power density. They also found that when the Sn atomic ratio in PtSn electrocatalyst is in the range of 10–20%, it shows desirable results. Moreover, they found that the addition of tin to Pt promotes the oxidation of ethanol to acetic acid at lower potentials and proposed the mechanism of ethanol electrooxidation over Pt and PtSn catalysts.[119] Both the bifunctional mechanism and ligand effect have been proposed to be involved in the ethanol electrooxidation over PtSn catalyst.[79] 22

At the end of this chapter we have to note that even if PtSn/C catalysts exhibit higher electrocatalytic activity to ethanol oxidation, the majority of the oxidation products are still the species containing C–C bond, which will have an obviously negative effect on the fuel cell performance.[4] It is crucial and necessary to develop a novel catalyst or add a third element to modify the PtSn/C and PtRu/C to present higher specific activity of dehydrogenation, C–O and C–C bond cleavage during the ethanol oxidation process.

23

3. Heteropoly compounds Transition metal polyoxometalates (POMs) are well-defined clusters with an enormous variation in size, metal-oxygen framework topology, composition, and functions. The preparation of POMs is based on the programmed self-assembly of metal oxide building blocks, which result in discrete, structurally uniform, nanoscopic clusters.[140] Although the first polyoxometalates [PMo12O403-] was reported in the 1826 by Berzelius[16] they have wide drawn attention since 1933 when it was possible to characterized a believing variety of POM structure and the first X-ray crystal structure analysis of the Keggin anion, [PW12O403-], appeared in the literature.[141] Most of the chemical and physical properties of POMs have been growing interest for variety applications, particularly in medicine, homogeneous and heterogeneous catalysis.[18,142] There are two generic families of polyoxometallates[19]: (1) the isopoly compounds (isopolyanions or isopolyoxometalates) contain only d0 metal cations and oxide anions (2) the heteropoly compounds (called heteropolyanions, heteropolyoxometalates, or heteropolyacids, when contain in the structure H+, H3O+, H5O2+) contain one or more p-, d- or f-block “heteroatom” in addition the other ions. As heteropolyanions are more numerous and their structural and electronic properties are easier to modify synthetically than those of the isopolyanions, the former is a field of increasing importance (particularly, in acid catalysis).

3.1 Structure of heteropoly compounds (heteropolyacids) Heteropolyacids

(HPAs)

are

complex

proton

acids

that

incorporate

polyoxometallates anions (heteropolyanions) having metal-oxygen octahedral as basic structure units.[18] Their general formula may be presents as [XxMmOy]q- (x < m) for heteropoly anions. M is usually Mo or W and to a lesser extent V, Nb, or Ta. The heteroatom X, can be one of 64 elements that belong to a various groups of the periodic table except the noble gases.[143]

24

Heteropolyacids can be classified into four different groups according to their molecular architectures[144,145]: (1) Keggin (e. g. H3[PW12O40]), (2) Wells-Dawson (e. g. H7[P2Mo17VO62]), (3) Finke-Droege (e. g. Na16[Cu4(H2O)2(P2W15O56)2]) and (4) Pope-Jeannin-Preyssler (e. g. (NH4)14[NaP5W30O110]). Among the various structural classes, the heteropolyacids comprised of the Keggin structure (represented by the formula Xn+M12O4n-8), and in particular the compounds containing molybdenum and tungsten, are the most often studied due to their acids and redox properties, stability at elevated temperatures, availability and relative ease of synthesise what is the most important in catalysis.[16] In 1864 Marignac observed two isomeric forms of the H4SiMo12O40, now know as the α- and βisomers.[146] In 1933 Keggin reported an the structure the α-isomer of [PW12O40]3-.[147] The α-Keggin anions are know with a wide range of heteroatoms (for example, X = Al(III), Si(IV), P(V), Fe(III), Co(III), Cu (I) and Cu (II) for M=Mo and W. β-isomers are much less common and the structure of β-[SiW12O40]4-, was first determinated by crystallography in 1973 by Yamamura and Sasaki.[146] Figure 9 illustrates two views of the α-Keggin structure. The Keggin anions has a diameter of ca. 1.2 nm and is composed of a central tetrahedron XO4 surrounded by 12 ege- and corner–sharing metal-oxygen octahedra MO6. The octahedra are arranged in four M3O13 groups, one of which is shown in Fig. 1A. Each group is formed by three octahedra sharing edges and having a common oxygen atom which is also shared with the central tetrahedron XO4 to give molecular symmetry. The metal atoms M occupy the centres of distorted octahedra with one terminal M-Ot bond. There are four types of atoms in the Keggin anion: 12 terminal M=Ot, twelve edge-bridging angular M-Oc-M shared by the octahedra within a M3O13 group, 12 corner-bridging quasi-linear M-Oe-M connecting two different M3O13 groups, and 4 internal X-Op-M.[148] The β-isomer structure may be considered as derived from the α-structure by rotation of one M3O13 group by 60o about its threefold axis.

25

Op X

A

B

Fig. 9. The α-Keggin (primary) structure shown (A) as a combination of [M3O13] groups and (B) as individual bonds showing the distorted octahedral geometry around each metal.[28,146]

The examples of the Keggin type heteropolyacids:


H3PW12O40 (PW12) 12-phosphotungstic acid
H3PMo12O40 (PMo12) 12-phosphomolybdic acid
H4SiW12O40 (SiW12) 12-silicotungstic acid
H4SiMo12O40 (SiMo12) 12-silicomolybdic acid Generally, heteropolycompound (heteropoly acids and their salts) form ionic crystals composed of heteropolyanions (primary structure), countercanions (H+, H3O+, H5O2+), hydratation water, and other molecules (Fig. 10). This three – dimensional arrangement is the secondary structure (pseudoliquid).[149] A relatively stable form of hydrated HPA contains six water molecules per Keggin unit (KU). The hexahydrate has a body-canter cubic structure with Keggin units at the lattice points and H5O2+ bridges along the face. Each of 12 terminal oxygen atoms of the Keggin unit is bond to hydrogen atom of an H5O2+ bridge. The secondary structure of HPAs depends on the amount of hydration water and heteropolyanion and can change from flexible to solid. Fro example 12tungstophosphoric acid (H3PW12O40) exhibits different packing arrangements as the hydration water is lost.[148] This water can be easily removed on heating, whereby the acid strength is increased due to the dehydration of protons. This a reversible process accompanied by chancing the volume of crystal cell. Unlike the rigid network structure of zeolites, in HPA crystal the Keggin anions are quiet mobile. Not only water but also

26

a variety of polar organic molecules can enter and leave HPA crystal. Such structural flexibility is important when using HPA as a heterogeneous catalyst.

Fig. 10. Schematic diagram illustrating the structure of the hexahydrate H3PW12O40·6H2O as two interpenetrating simple cubic structures. The H+ of an H5O2+ species coordinated to the body centered KU is located at the midpoint of an edge of the conventional cubic cell. The KUs are shown in polyhedral representation.[150]

The tertiary structure is the structure of heteropolyacids as assembled. The size of particles, pore structure, surface area, and distribution of protons in the particles are the elements of the tertiary structure.[14]

3.2 Properties of heteropolyacids

Heteropolyacids (HPAs), also know as polyoxometalates (POMs), are early transition metal oxygen anion clusters that exhibit a wide range of molecular size, compositions and architectures.[16,151] A most attractive attribute of heteropoly compounds is the size dependence of their physicochemical properties. Most notable is the size-dependent tendency of the metaloxygen framework to accommodate excess electrons.[140,152] Heteropoly compounds are commonly strong oxidizing agents, which are readily reversibly reduced by addition of various specific numbers of electrons depending upon the pH and the potential employed. The ability of heteropoly compounds to accept electrons under alteration of the optical properties can be use for the construction of functional electrooptical materials.[153,154] The reduction products, which typically retain the general structures of their oxidized parents, are characteristically deep blue in colour and comprise a vary large group of complexes known as the “poly blue”, or “heteropoly blue” if the

27

framework includes heteroatoms.[155] Heteropoly blues corresponds to class II systems in the Robin and Day classification of mixed-valence compounds. The added (“blue”) electrons are delocalized over certain atoms or regions of the structure. The electron delocalization is viewed as operating through two mechanisms[156]:
a thermally activated electron hopping processes from one addendum (e. g. Mo or W) atom to the next,
a ground-state delocalization presumably involving π-bonding through bridging oxygen atoms from reduced metal atom to its neighbours. The heteropoly blue provide important potentialities as specialized reducing agent, with a wide range of controllable reduction potentials. Heteropoly blue are use as electroreduction catalyst. By fixing the potential, one fixes the heteropoly blue species involved and thus controls the number of electrons per reduction event.[147] At the lattice energies of heteropoly compounds are low, and so are the salvation energies of heteropoly anions, the solubility of heteropoly compounds largely depends on the salvation energy of the cation. Heteropolyacids (HPAs) are extremely soluble in water and oxygen-containing organic solvents such as lower alcohols, ethers, ketones, ect. On the other hand, they are insoluble in nanpolar solvents such as benzene.[148] The Keggin anions may be consider a nanoscale, isolated particle within an aqueous matrix, which allows classification of Keggin type heteropolyacids as a strongly acidic particle hydrate.[149,157] In aqueous solution H3PW12O40 (PW12), H4SiW12O40 (SiW12) and H3PMo12O40 (PMo12) are strong fully dissociated acids, and there are stronger then the usual mineral acids such as H2SO4, HCl, HNO3, ect. (Table 5).[18]

Table 5. Dissociation Constants of Heteropolyacids in Acetone at 25oC.[18]

The acid strength of crystalline heteropolyacids decrease in the series PW12 > SiW12 > PMo12 > SiMo12 which is the same to that in solution (Table 5). The tungsten acids are markedly stronger than molybdenum ones. Although, the effect of the central atom is no

28

as great as that of the addenda atoms, phosphorus-based heteropolyacids are slightly more acidic than silicon based heteropolyacids. A strong polarization of the outer electrons of the surface oxygen towards the Mo6+ ions gives a strong Brønsted acidity to all surface oxygen, particularly the terminal oxygen (Ot). As a result, the equilibrium reaction is shifted strongly the right to create a high concentration of protons in hydrogen-bonded aqueous matrix with which the anions are located (Equation 32).[149]

(32)

However, when the HPA has been fully dehydrated, the location of the protons is not as easily defined. The terminal (M=Ot) and bridging (M-O-M) atoms of oxygen are the most probable positions for the protons. The calculation were proved that the most energetically favourable site for the acid proton is a bridging oxygen atom.[158] The heteropolyacids are widely use as acid, redox, and bifunctional catalysts in homogenous and heterogeneous system because of their high solubility in polar solvents and fairly high thermal stability in the solid sate.[14,15,18] The thermal stability of hydrogen forms of heteropolyacids change as follows: H3PW12O40 > H3PMo12O40 > H4SiMo12O40 > H4SiW12O40. The ability of heteropolyanions to undergo spontaneous adsorption from aqueous solution on various electrode substrates[159,160] provides a simple tool for the modification of the surface.[19-23] Further, their ability to exhibit fast reversible multielectron transfers and super-acid properties resulting in the increased availability and mobility of protons at the electrocatalytic interface. Heteropolyacids have been known as good matrix for traces of platinum reactive towards reduction of oxygen[19-21]. Because of interaction between heteropolyanions and Pt surface by mainly corner oxygen (from heteropolyanions) only a few percent of interfacial reactive platinum atoms would be blocked to the access of oxygen molecules.[21] Another issue concerns ability of the adsorbed heteropolyanions (particularly H3PW12O40) to shift of voltammetric peaks referring to the formation of Pt-oxo (PtOH or PtO) species towards more positive potential and increase of the potential range where platinum is not

29

covered with platinum oxide adsorption and activation of oxygen molecules during electrocatalysis may be facilitate.[21] Heteropolyanions due to their exceptionally high solid state protonic conductivity were considering for fuel cell research.[23-25,27,28,161] Stanis and co – workers have presented that the addition of adsorbed heteropolyacids can improved the performance of Pt anodes in a fuel cell under CO – poisoned conditions.[27] Though, heteropolyacids have been identified to be good candidates as materials for catalytic applications, the insolubility in water and increasing the surface area of heteropoly compound are necessary. For this purpose, the mine approach has been followed in the literature: the direct preparation of acid porous and water-insoluble salts.

3.3 Salts of the heteropoly compounds Keggin type heteropolyacid salts, described by formula M1xHy-xM2M312O40, where M1 is Cs+, Rb+, M2 is P or Si, M3 is W or Mo, x is 2.5 and y is 3 or 4 if M2 is P or Si, respectively, are produced by partially exchanging protons in the parent heteropolyacids. They are efficient solid acid catalyst for a variety of organic reaction, particularly for liquid-phase reaction.[14,18,162] Salts of heteropoly compounds can be classified into two groups[163]: (1) group A, the small cation group (like Na+, Cu2+), which posses:
low surface area (1 – 15 m2 g-1),
high solubility in water,
absorption capability of polar or basic molecules in the solid bulk, (2) group B, the large cations group (like Cs+, Rb+), which is:
with high surface area (50 – 200 m2 g-1),
insoluble in water,
unable to adsorb molecules. Partial exchange of heteropolyacids with large cations, such as Cs+, Rb+, NH4+, K+, ect. change a water-soluble acid with low surface area (<5 m2g-1) into a water-insoluble acid salt precipitates with surface area exceeding 100 m2g-1.[31,32] Further, strong acidity remains on acids salts what is opposite to alkali-exchanged zeolites.[14,33]

30

Primay structure

Secondary structure

W3O13

The crystal structure of Cs salts are the same as

P

the H+ (H2O)n, Cs2+, Cu2+, K+, Mg2+, ect.

cubic

of

H3PW12O40·6H202,

with

cations at the sites of H+ (H2O)2

sites,

called

second structure, which corresponds Tertiary structure

to

the

of

the

micocrystallites. Aggregates

microcrystallites called

the

are tertiary

structures corresponding. [14, 164]

Fig. 11. Hierarchical structure of Keggin type heteropolyacids salts.[164,165]

Okuhara has presented the change of the surface area with the extent of Cs substitution for proton in H3PW12O40 (Fig. 12).[162] The pore size of heteropolyacids salts can be precisely controlled by the cation content. The surface are decrease when the Cs content, x in CsxH3-xPW12O40 increase from 0 to 2, and then the surface are increase when x change from 2 (1m2 g-1) to 3 (156 m2 g-1). The author has estimated the acid amount on the surface (called surface acidity) from the surface area and the formal concentration of proton attached to the polyanion. Figure 12 illustrates that the surface acidity decreases at first with the content of Cs, but sharply increases when x exceeds 2. The maximum appeared at x = 2.5. The Cs2.5H0.5PW12O40 and H3PW12O40 have similar acid strength.[162,166]

31

Fig. 12. Changes in the surface area and surface acidity of CsxH3−xPW12O40 as a function of the Cs content.[162]

Okuhara and co – workers have estimated from surface area the particles size of the heteropoly aids and its salts.[160] The particles size were 2000 Å for H3PW12O40 and 5000-10000 Å for Cs1 (x = 1) and Cs2 (x = 2), respectively. Conversely, particle sizes of Cs2.5 and Cs3 were 60-70 Å. Morphologies of the CsxH3−xPW12O40 were examined using scanning electron microscopy (Fig. 13). It was apparent that the surface of Cs2 is smooth (Fig. 13A). The same observations were made for Cs1. On the contrary, Cs2.5 is composed of fine particles with the size about 100 Å and the surface are rough (Fig. 13B). As well as the primary spherical particles, pores between particles can be seen for Cs2.5. The authors obtained the similar SEM image for Cs3 to that of Cs2.5.

Fig. 13. SEM images of (A) Cs2HPW12O40 and (B) Cs2.5H0.5PW12O40.[160]

It is well know that the pore-width of zeolites can be controlled by the kind of cation and the pore-width increased as the cation size decreased (Cs > Rb > K).[167] The salts of Rb and K gave change in the surface area (Fig. 14) similar to that of Cs salts 32

(Fig. 12).[162] A marked increase of the surface area when the Rb content increases was observed at x = 1.8, while it was detected at 2.1 for the Cs salts. In the case of K salts, the change of the surface are with the increase of the content of K was rather loose.

Fig. 14. Surface areas of RbxH3−xPW12O40 (A) and KxH3−xPW12O40 (B) as a function of cation content.[162]

Heteropoly salts are frequently more stable than the parent acid. The relative stabilities, however, depended on the counteraction.[148] The thermal stability change generally in the order of Ba2+, Co2+ < Cu2+, Ni2+ < H+, Cd2+ < Ca2+, Mn2+ < Mg2+ < La3+,Ce3+ < NH4+ < K+, Tl+, Cs+.[14] The acidic cesium salt Cs2.5H0.5PW12O40 is more stable than H3PW12O40. No decomposition of the salt was observed at 500 0C. Due to the presence of meso-micropore in the Keggin type heteropolyacids salts (particularly,

where

x

=

2.5),

it

is

possible

to

introduce

the

platinum

nanoparticles.[162,164] The size and dispersion of nanoparticles of Pt can be control by quantity of platinum in the structure of heteropolyacids salt. Okuhatra and Nakato reported that the presence of Pt 0.5 wt% in the Cs2.5H0.5PW12O40 did not influence the pore width of these Cs salts, and the size of Pt was probably less than 10 Å. The Keggin type heteropolyacids salts (particularly, when cation content is equal 2.5, e.g. Cs2.5H0.5PW12O40) are evidently a vary promising materials as matrix for catalytic centres (for example Pt nanoparticles) for fuel cell research.

33

4. Experimental techniques

4.1 Cyclic voltammetry

Cyclic voltammerty (CV) is often the first experimental performed in an electroanalytical study. In particular, it offers a rapid location of redox potentials of the electroactive species, and convenient evaluation of the effect of media upon the redox process.[168] Cycling voltammetry consists of scanning linearly the potential of a stationary working electrode immersed in a quiescent solution (Fig. 15) and measuring the resulting current.[168-170]

Fig. 15. Potential – time excitation signal in cyclic voltammetric experiment.[168]

In CV a constant-surface electrode (platinum, gold, glassy carbon, hanging mercury drop electrode) are used as working electrodes. During the change of the potential (when the oxidized form, O) from E1 to E2 (in negative direction), a cathodic current begins to increase, until a peak is reached (Fig. 16). After traversing the potential region in which the reduction process take place, of the potential sweep is reversed. During the reverse scan, R molecules (generated in the forward half cycle, and accumulate near the surface) are deoxidised back to O and the anodic peak results.[168]

34

Fig. 16. Typical cyclic voltammogram for a reversible O + ne- ↔ R redox process.[168]

In cyclic voltammetry for the reversible systems the position of the peaks on the potential axis (Ep) is related to the formal potential of the redox process. The formal potential for a reversible couple is centered between the anodic peak potentials (Ep,a) and the cathodic peak potentials (Ep,c)[168]:

Eo =

E p,a + E p,c

(33)

2

The separation between the peak potentials (for reversible couple) is given by

∆E p = E p,a − E p,c =

0.059 V n

(34)

Thus, the peak separation ca be used to determinate the numbers (n) of electrons transferred, as a criterion for a Nernstian behaviour. Accordingly, a fast one-electron process exhibits a ∆Ep of about 59 mV. Both the cathodic and anodic peak potentials are independent of the scan rate. For irreversible process, the individual peaks are reduced in size, widely separated and depended of the scan rate.

35

The peak current (ip) for a reversible electrode process is presented by the Randles-Sevcik equation[168,169]: ip = 2.72 x 105n3/2D1/2AV1/2c0

(35)

It follows from Eq. 35 that the peak current depends on the concentration of the depolarized in the bulk of the solution (c0), the diffusion coefficient of the substance being reduced or oxidized (D), the area of the electrode surface (A), and the number of electrons taking parts in the elementary electrode process (n). Furthermore, the current increase with the increasing polarisation rate (V). The liner dependence of the peak current on the concentration of the reacting substance makes this method useful in quantitative analysis. The peak dependent on the square root of the scan rate, and the liner dependents means that the value of current is control by diffusion of electroactive substance to the electrode surface. Different behaviour is observed when the reagent or product of an electrode reaction is adsorbed strongly or weakly on the electrode. The separation between the peak potentials is smaller than expected for solution phase process. The ideal Nernstian behaviour of surface-confined nonreacting species is manifested by symmetrical cyclic voltammetric peaks (∆Ep = 0), and a peak half-width of 90.6/n mV (Fig. 17).

Fig. 17. Ideal cyclic voltammetric behaviour for a surface layer on an electrode. The surface coverage, Γ, can be obtained from the area under the peak.[168]

36

The peak current is directly proportional the surface coverage (Γ) and potential scan rate[168]: ip =

n 2 F 2 ΓAν 4 RT

(36)

The surface coverage, Γ, can be calculate from the peak area (i.e., the quantity of charge consumed during the reduction or adsorption of the adsorbed layer):

Q = nFAΓ

(37)

This can be used for calculation the area occupied by the adsorbed molecule and hence to predict its orientation of the surface. In practices, the ideal behaviour is approached for a relatively slow scan rate, and for adsorb layer that show non intermolecular interaction and fast electron transfer.[168] The cyclic voltammetry method was used in the thesis for initial electrochemical studies of Keggin type heteropolyacids salts and examination of proposed catalysts towards oxidation of hydrogen, methanol and ethanol.

4.1.1 Rotating disk voltammetry The rotating disk electrode (RDE) is vertically mounted in the shaft of the synchronous controllable-speed motor and rotated with the constant angular velocity (ω) about an axis perpendicular to the plan disc surface (Fig. 18).[168]

Fig. 18. Rotating disc electrode.[168]

37

The primary advantages gained by the utilization of electrochemical techniques based on rotated electrodes is the precise, quantitative control

of mass transport to the

electrode through forced convection included by electrode rotation.[171] This is possible because the motion of a RDE drags a layer of fluid near the disc surface along with it as rotates. At the same time the liquid layers is subjected to centrifugal forces that cause to move rapidly away from the rational axis of the electrode describing an S-shape path as it does so. As a consequence of fluid motion parallel to the disk surface, new liquid is drawn to the disc along the path that is parallel to the rotational axis of the electrode. According to the simple Nernst diffusion layer concept, a thin layer of stagnant solution is present at the electrode surface, within which the concentration of the electroactive species that is undergoing oxidation or reduction varies linearly from its value in the bulk solution to a new value at the electrode surface. The current, i, is given approximately by the expression[171]:

i=

nFAD( c o − c )

δ

(38)

where A is electrode area, δ is the thickness of diffusion layer, F is the Faraday constant, D is the diffusion coefficient, n is the number of electrons taking parts in the elementary electrode process, co and c are the concentration of the electroactive species in the bulk solution and at the electrode surface, respectively. For RDE in a solution with the kinematic viscosity, ν, Levich presented that the diffusion layer thickness is dependent on the inverse square root of the angular velocity ω of the rotating electrode: δ = 1.61D1/3ν1/6ω-1/2

(39)

The limiting current il (for a reversible system) is thus proportional to the square root of the angular velocity, as described by Levich equation[169]: il = 0.62nFAD2/3ω1/2ν-1/6co

(40)

An increase in ω from 400 to 2500rpm thus results in a twofold increase of the signal. A deviation from linearity of a plot of il vs. ω-1/2 suggests some kinetic limitations. 38

Since 1960 the rotating disc technique has been employed in the study of the kinetics of electrode process, as well as of that of chemical reactions taking place at the electrode surface.[169] In the thesis rotating disc voltammetry was applied to study the kinetics of hydrogen oxidation reaction (Chapter 7.2.1 and 7.3.4).

4.2 Chronoamperometry Chronoamperometry[168] involves stepping the potential of the working electrode from a value at which no faradic reaction occurs to a potential at which the surface area concentration of the electroactive species is effectively zero (Fig. 19a). A stationary working electrode and unstirred solution are used. The resulting current – time dependence is monitored. As mass transport under these conditions is solely by diffusion, the current – time curve reflects the change in the concentration gradient in the vicinity of the surface. This involves a gradual expansion of the diffusion layer associated with the depletion of the reactants, and hence decreased slope of the concentration profile as time progress (Fig. 19b). The current corresponding to the transformation of Ox in to Red, for linear diffusion of reactants to the electrode, change with the time according to the Cottrell equation[169]:

1/ 2 0 nFDOx AcOx il = π 1/ 2 t 1/ 2

(41)

where, iL is the limiting current, i.e. the maximum current that can be obtained under the given conditions. Its value depends on the bulk depolarized concentration (C0ox), the diffusion coefficient (Dox), the electrolysis time (t), and the electrode surface area (A). F is the Faraday constant, and n is the number of electrons exchanged between one ion or molecule of a reactant an electrode. According to equation 41 the current of chronoamperometric electrocatalysis tends to zero when the time tends to infinity. This is due to the progressive decrease of the reactant concentration in the region close to the surface, and change the current with time corresponds to the curve illustrates in Fig. 19c.

39

Fig. 19. Chronoamperometric experiments: (a) potential time waveform, (b) change of concentration profiles with time, (c) the resulting current – time response.[168]

In the thesis the chronoamperometry method was used to evaluate the reactivity of our electrocatalytic system for the methanol (Chapter 8.2.3, 8.3.3) and ethanol (Chapter 9.3) oxidation.

4.3 Staircase voltammetry Staircase voltammetry[168,172] has been proposed as a useful tool for rejecting background charging current. The potential – time waveform involves successive potential step (∆E) of c.a. 10 mV height and about 50 ms duration (tp) (Fig. 20). The current is sampled at the end of each step, where the charging current has decayed to a negligible value. This method of polarization of the electrode enable the double layer to be charged, making it possible to discriminate capacitive current component, if the current is measured at a sufficiently long time after application of the pulse. Such a method of polarisation and current sampling should give current vs. potential curves that are similar to those recorded in linear scan voltammetry.[169]

40

Fig. 20. Potential – time waveform used in staircase voltammetry.[168]

In the thesis staircase voltammetry method was applied to get insight into the system’s reactivity towards oxidation of methanol and ethanol (Chapter 8.2.2, 8.3.2, 9.2).

4.4 Electrochemical impedance spectroscopy

Electrochemical impedance spectroscopy is an experimental technique that involves imposing a small sinusoidal (AC) voltage or current signal of knows amplitude and frequency - the perturbation – to an electrochemical cell and monitoring the AC amplitude and phase response of the cell. The AC perturbation is typically applied over a wide range of frequencies, from 10 kHz or greater to less than 1 Hz, hence the name impedance spectroscopy. The ratio and phase-relation of the AC voltage and current signal response is the complex impedance, Z (jω). The results of an impedance spectroscopy experiment is a rich data set from which many properties of the electrochemical cell may be extracted via application of physically – responsible equivalent circuit models.[36,173] Properties of the electrochemical system commonly evaluated using impedance spectroscopy include ohmic (bulk) resistance, electrode properties such as charge transfer resistance and double layer capacitance, and transport (diffusion) effects. An electrochemical impedance experiments is based on monitoring the AC response of an electrochemical cell that results from imposing a small AC signal (Fig. 21). The impedance is the ratio of the AC voltage and current output. 41

Fig. 21. Schematic of an electrochemical impedance spectroscopy.[36]

A

sinusoidal

current

signal

of

amplitude

IAC

(amps)

and

frequency

ω

(radiations/seconds) can be defined:

I(ω) = IAC·sin (ω·t)

(42)

where t is time (s). The output AC voltage signal from the electrochemical cell can be defined:

V (ω) = VAC·sin[(ω·t)-Ө]

(43)

where VAC is the amplitude for the output voltage signal (volts) and Ө is the phase angle (radians). The phase angle is the difference in the phase of a sinusoidal voltage and current signals. In the case of AC signal, the “resistance” of circuit of the electrochemical device, which is not purely resistive, will be function of the frequency of oscillation of the input signal. Ohm’s Law for the AC case is expressed:

Z(jω) = V(jω) / I(jω)

(44)

where Z (jω) is the complex impedance (Ω) and j is imaginary operator, j ≡ √-1

(45)

Equation (44) indicates that impedance is a complex value. That is, it can take on both real and imaginary components. Note that the imaginary component of the impedance is a real measurable quantity: “j” is for bookkeeping purposes and allows description of the out-of-phase component of the impedance. The complex relationship of impedance is implicit so Z(jω) is normally written as Z(ω). Although one can think of impedance as “resistance” to current, it is more general than that because it takes into account the phase difference between voltage and current. Equation (44) also indicated that 42

impedance depends on the frequency at which it is measured. Z can change as the frequency of the AC signal changes. Frequency in cycles per second, f (Hz = 1/s), is obtained through the relation: ω = 2π·f

(46)

Equation (44) can be written in complex notation:

Z = Z´ + Z´´

(47)

where, Z´ = Re(Z) = |Z|cosӨ

real (in-phase) component of impedance

(48)

Z´´ = Im(Z) = |Z|sinӨ

imaginary (out-of-phase) component of impedance

(49)

Z = ( Z ′) 2 + ( Z ′′) 2

magnitude of impedance

(50)

and, Ө = tan-1 (Z´´/Z´)

(51)

In the impedance spectroscopy experiment, the frequency of the AC perturbation is swept over a range, from ~ 10 kHz to less than 1 Hz and the impedance is evaluated as a function of frequency to evaluate the properties of the electrochemical system under investigation. Usually the real (Z´) and imaginary (Z´´) parts of the impedance are plotted as a function of frequency. In the thesis the AC-Impedance was used to analyzed reaction mechanism of methanol and ethanol electrooxidation on Pt40%/Vulcan XC-72 carbon modified with Keggin type heteropolyacids salts (Chapter 8.2.4, 8.3.4, and 9.4).

4.5 Transmission electron microscopy

The fathers of electron microscopy were Knoll and Ruska (1931), and the first commercial TEM was built in 1939 by Siemens. Since then, the theory and the instrumentation have developed and modern TEM’s have become a fundamental tool for material science.[174]

43

A TEM microscope is composed by an optical column, operated under high vacuum that enclose[175]:


Illumination system. It takes the electrons from the gun and transfers them to the specimen giving either a broad beam or a focused beam. In the ray-diagram, the parts above the specimen belong to illumination system.
The objective lens and stage. This combination is the heart of TEM.
The TEM imaging system. Physically, it includes the intermediate lens and projector lens.

Fig. 22. Schamatic of a Transmition Electron Microscope.[175]

The diffraction pattern and image are formed at the back focus plane and image plane of the objective lens. If we take the back focus plane as the objective plane of the intermediate lens and projector lens, we will obtain the diffraction pattern on the screen. It is said that the TEM works in diffraction mode. If we take the image plane of the objective lens as the objective plane of the intermediate lens and projector lens, we will form image on the screen. It is the image mode. TEM is applied to analysis of electrochemical power sources, since novel electrode materials for electrochemical energy storage devices are currently synthesized with constantly smaller dimensions, down to nanosize level. This is very important in catalysts used for fuel cells, since smaller particles have higher specific surface and catalytic activity. Control of morphology, granulometry and microstructure at nanometric level is, thus, very important in predicting or explaining the performances obtained by the different electrodes or preparation methodologies. In the thesis TEM was used to investigate size of Pt nanoparticles obtained by corrosion of platinum counter electrode (Chapter 7.3.3).

44

4.6 Scanning electron microscopy

In Scanning Electron Microscopy (SEM), a high energy (up to 50 keV), very thin electron beam is finely focused over a sample and swept in a raster across the surface. The electron beam/sample interactions cause the emission of different signals that are collected by specific detectors and converted into an image of the sampled area and viewed or recorded on a cathode ray tube (CRT). The generated signals include secondary electrons, backscattered and Auger electrons, photons of various energies and characteristic X-rays. The signals of greatest interest for surface topography are secondary and backscattered electrons.[174] The basic components of the SEM microscope are presented in Fig. 23.

Fig. 23. Schematic of operation a typical Scanning Electron Microscope.[174]

The SEM permits the observation of materials in macro and submicron ranges. When SEM is used in conjunction with EDS (Energy Dispersive X-ray Spectrometer) the analyst can perform an elemental analysis on microscopic sections of the material or contaminants that may be present.

45

In the thesis the scanning electron microscopy was used to investigate morphology of the Keggin types heteropolyacids salts (Chapter 6.3).

4.7 Infrared spectroscopy Infrared spectroscopy (IR) involves examination of the twisting bending, rotating and vibrational motions of atoms in a molecule.[170] Molecules contain bonds of specific spatial orientation energy. These bonds are seldom completely rigid, and when is supplied, they may band, distort or stretch. A vary approximate model compares the vibration to that of a harmonic oscillator, such as an ideal spring. If the spring has a force constant, k, and masses mA and mB at the end, than the theoretical vibration frequency ν is given by[176]: ν = (1/2π)√(k/µ)

(52)

where µ = mA · mB / (mA + mB) is called the reduced mass. Each type of molecular vibration is characterized by a vibrational quantum number, υ. For a simple stretching vibration, there is a series of levels whose energy is given by approximately by:

E = hν0·(υ + ½)

(53)

This means there is a set of levels spaced energy by hν0 or in wavenumber by ν 0. The selection rule for an ideal harmonic oscillator allows transition where ∆υ = ± 1, giving a single, fundamental vibrational absorption peak. However, when the bonds are stretched they weaken, so better model takes this into account, and the molecules are treated as anharmonic oscillators. Thus, where high energies are involved, larger energy transition may occur, where ∆υ = +2, +3, ect. giving the first overtone at wavenumber approximately double that the fundamental, and so on. The electric field of incident radiation interacts with the molecular dipole. When the frequency of the radiation (~1013 Hz) resonates with a molecular vibration, absorption can occur, particularly if excitation of that vibration has an effect on the

46

molecular dipole moment. The energy changes involved in exciting vibrational modes in this way correspond to the infrared spectral region. A full infrared spectrum consists of bands (group frequencies), assignable to particular moieties (e.g. -CH2-, -CH3, C=O), in characteristic frequency regions that are relatively independent of the other groups in the molecule. Since infrared spectroscopy probes molecular vibrations that involve changes in the dipole moment, the vibrations of polar molecular bonds generally correspond to strong infrared bands. Most IR spectra are recorded by measuring the absorption of the incident radiation as a function of wavelength (wavenumber) in the range of 2.5 - 25 µm (4000 400cm-1) by solid, liquid or gaseous samples. Both qualitative and quantitative information can be obtained from vibrational spectra. The use of IR spectrometry for qualitative measurements is extensive and wide ranging, and for this purpose transmission spectra are conventionally recorded as function of wavenumber. In order to make quantitative measurements, it is necessary to convert the transmittance readings to absorbance, A, the relation between the two being[176]:

A = log (100 / T%)

(54)

In the thesis infrared spectroscopy was used as an established tool for the structural characterization of heteropolyacids salts (Chapter 6.2).

47

5. Chemical reagents and measuring equipment

A) Chemical reagents using in the measurements:

Phosphotungstic acid hydrate, H3PW12O40·nH2O (PW12), Sigma-Aldrich Phosphomolybdic acid hydrate, H3PMo12O40 nH2O (PMo12), Sigma-Aldrich Tungstosilicic acid hydrate, H4SiW12O40·nH2O (SiW12), Sigma-Aldrich Silicomolybdic acid hydrate, H4SiMo12O40·nH2O (SiMo12), Sigma-Aldrich Cesium nitrate, CsNO3, Aldrich Rubidium nitrate, RbNO3, Aldrich Ammonium Chloride, NH4Cl, Carlo Ebra Potassium chloride, KCl, Aldrich Sulphuric acid 99,999%, H2SO4, Aldrich Methanol 99, 9%, CH3OH, J. T. Baker Ethanol 99, 9%, C2H5OH, J. T. Baker Nafion solution 5% wt, Ion Power, Inc Pt10% on Vulcan XC-72 (Pt/C), E-tek Pt40% on Vulcan XC-72 (Pt/C), E-tek Vulcan XC-72, E-tek High purity argon, carbon monoxide, hydrogen, oxygen gas Ultra-pure water (Millipore Milli-Q)

B) Equipments used during performed experiments:
IR spectra in the range 1800 to 500 cm-1 were recorded with a Perkin Elmer System 2000 FT-IR instrument.
The morphology of platinum particles was monitored using a JEM2100F electron microscope (TEM) operating at 200 kV.
The morphology of heteropolyacid salts was monitored using a JOEL Model JSM-5400 scanning electron microscope (SEM).
Electrochemical measurements were carried out by CHI 660B electrochemical working station (CH Instruments Inc.).

48


RDE voltammetric measurements were done using a variable speed rotator, Pine Instruments, USA.
For the electrochemical impedance studies, the data were collected by CHI 660B with a frequency sweep of 0.05 Hz to 100 KHz and ac sine wave amplitude of 5 mV.
The PEMFC test stand consists of fuel cell hardware, backpressure regulators, water traps, a humidifier and mass flow controller. The fuel cell hardware (Fuel Cell Technologies Inc.) is a single cell with area 4.84 cm2 and single-serpentine flow fields.

C) Electrodes applied during the measurements:
Glassy carbon electrode (as working electrode), GC (area 0.071cm2), CH Instruments, Inc.
Rotating disc (glassy carbon) electrode (as working electrode), RDE (area 0.1256cm2), Pine Researcher Instrumentation.
A saturated calomel electrode (as references electrode), SCE, Amel Electrochemistry
Platinum flag (as counter electrode, area ca. 1cm2)
Nafion Membrane 115 (as solid electrolyte in PEM cell), Ion Power, Inc.
Low-temperature ELAT® GDL microporous layer on woven web, tin configuration (as diffusion layer in electrode for fuel cell), E-tek

D) Experimental conditions:


The measurements were carried out in a three-electrode cell with a platinum flag as a counterelectrode and a saturated calomel electrode serving as the reference electrode, placed in a separated compartment and connected to the main cell by a Luggin capillary.
CO stripping voltammetry study: prior to electrochemical measurements 0.5 mol dm-3 sulphuric acid solution was purged with argon for 30 minutes. Subsequently, five consecutive CV (scan rate 20 mV s-1) were performed in the potential range 0.025 – 1.125 V vs. RHE. For CO

49

stripping measurements, pure CO was bubbled into the electrolyte for 10 minutes and then its adsorption on the electrode was driven under potential control at 0.1 V vs. RHE for 4 minutes. The electrolyte was purged for 35 minutes with argon, keeping electrode potential at open circuit potential (OCP) to eliminate CO reversibly adsorbed on the surface. Five cyclic voltammetry (scan rate 20 mV s-1) were recorded from 0.025 to 1.125 V versus RHE. The first anodic sweep from 0.025 to 1.125 V vs. RHE was performed to electro-oxidize the irreversibly adsorbed CO and the subsequently voltammetries in order to verify the completeness of the CO oxidation.
At the beginning of experiments, each catalytic layer was cycled continuously (san rate, 50 mV s-1) through the potential region from 0 to 1.05 V vs. RHE until a steady state voltammogram was obtained
Before experiment working electrode was polished (on a cloth) with Al2O3 water suspension, particle size 5 - 0.05 µm.
All the electrochemical experiments were carried out at 22 ±2 0C (except PEMFC measurements).
All potentials in these studies were reported here versus RHE.

50

6. Preparation and characterization of Keggin-type matrix This thesis is focused on applying salts of Keggin-type heteropolyacids as a material for the new stable, water-insoluble matrix for the films containing Pt nanoparticles. From several possible salts of Keggin-type heteropolyacids those containing 2.5 moles of Cs+, Rb+ and NH4+ cations in 1 mole of the heteropolyacids salt have been chosen. Advantages of these salts are the highest surface acidity, high ionic conductivity, water insolubility and high surface area characterized by the presence of meso and micro pores.

6.1 Preparation of Keggin-type heteropolyacid salts

Keggin-type heteropolyacids salts were prepared by adding stoichiometric amounts of 0.25 mol dm-3 aqueous solution of CsNO3, RbNO3 or NH4Cl, to the desired volume of 0.1 mol dm-3 aqueous solution of desired heteropolyacid[177-180], using a molar ratio M/Y = 2.5, (where M = Cs+, Rb+ or NH4+ and Y = P or Si). The suspension was then stirred for 24 h. The precipitates was washed four times with ultra pure water, separated from the liquid phase by centrifugation and freeze dried (40 - 75 0C).

6.2 IR characterization of Keggin-type matrix IR

spectra

of

H3PW12O40

(HPW),

Cs2.5H0.5PW12O40

(Cs2.5PW)

and

Rb2.5H0.5PW12O40 (Rb2.5PW) are shown in Fig. 24A. Four bands at 700-1100 cm-1 region corresponding to Keggin unit structural vibrations[181,182] are observed for all samples what suggesting that the framework of the primary Keggin structure remained unaltered. The origin of the Keggin anion vibration bands is as follows: at 1077 cm-1 is from the νas (P-Oa) vibration; 971 cm-1 is due to the terminal νas (W=Od) vibration; at 883 and 756 cm-1 should be assigned to νas (W-Ob-W) and νas (W-Oc-W), respectively. Weaker absorptions appearing at 592 and 523 cm-1 are due to bending vibrations of the type δ (Oa-P-Oa) and νs (W-O-W), respectively.[183] The absorption at c.a. 1697 cm-1 is indicative of the presence of the protonated water clusters, probably proton-type H5O2+, and it is assigned to δ (H2O) vibration.[178,184] The proton substitution with Cs+ or Rb+ ion to form M2.5H0.5PW12O40 (M = Cs+, Rb+) causes a decrease of the intensity of the δ

51

(H2O) peak at 1697 cm-1. Absorption band detected at 1620 cm-1 for cesium and rubidium salt of 12-phosphotungstic acid can be attributed to the presence of neutral water.[178] Fig. 24B shows the IR spectra of 12-phosphomolybdic acid, H3PMo12O40 (HPMo) as well as its cesium and rubidium salts. There is no apparent difference between the three spectra. All showed bands at 1060, 961, 875, 760 cm-1 that are in agreement with those reported in the literature.[22,185-187] A shift in the frequency in the oxygen Mo-Ob-Mo stretching mode from 875 cm-1 (for HPMo) to 865 cm-1 (for Cs2.5H0.5PMo12O40 or Rb2.5H0.5PMo12O40) is probably related to the organization and in particular to the presence of positively charged cation in the material. The absorption at 1610 cm-1 is related to the presence of neutral water.[178] The IR spectra recorded for Fig. 24C shows the 12-silicotungstic acid, H3SiW12O40 (HSiW) and its cesium and rubidium salts, while Fig. 24D shows the IR spectra of 12-silicomolybdic acid, H3SiMo12O40 (HSiMo) and their cesium and rubidium salts. The band around 1700 cm-1 (Fig. 24C) for HSiW is due to oxonium ions (H3O+) or more likely to dioxonium ions (H5O2+).[188,189] As in the case of HPW, proton substitution with Cs+ or Rb+ causes a decrease of the intensity of the peak at 1700 cm-1. The absorption band at c.a. 1624 cm-1 is indicative of the presence of “neutral” water. From the spectrum of HSiW and their cesium and rubidium salts in the 700-1100 cm-1 region, we can obtain information on the bonds between Keggin units.[190] The bands at ca. 730 and 871 cm-1 were assigned to the stretching of tungsten-oxygen-tungsten chains, the former to W-Oc-W, and the latter to W-Ob-W. The Oc oxygen atom is common for two [WO6] octahedra in [W3O10] subunits, joined by Ob atoms. The band at ca. 905 cm-1 was assigned to Si-Oa stretching (there are four Oa atoms connected to the central Si atom). The absorption at 970 cm-1 is related to the W=Od bond. The assignments of the bond at ca. 1020 cm-1 is not known. In the spectra of HSiMo (Fig. 24D) and its cesium and rubidium salts, we can see four bands at 700-1100 cm-1 region corresponding to Keggin unit structural vibrations. This means that, also in this case, the structure of the salts still retains the basic framework of the Keggin structure.

52

A Rb 2.5 H 0.5PMo 12 O 40

R b 2 .5 H 0 .5 P W 1 2 O 4 0 883

1620 1077 883 592 975 760

865 591

523

760

C s 2 .5 H 0 .5 P W 1 2 O 4 0

1060

592

975

523

961

Transmittance

1077

Transmittance

B

1610

1620

760

1610

Cs 2.5 H 0.5PMo 12 O 40 1060

865 591

961

760

1697 1610 1060

1077 883 971

H 3P W 12O 40

1800

1600

1400

1200

1000

W avenum ber / cm

875

H 3 PMo 12 O 40

756

592 523

800

600

590

961 1800

1600

-1

1400

1200

760

1000

W avenum ber / cm

800

C

D

1020

162 4

1614

R b 2 .5 H 1.5 S iW 12 O 40

R b 2.5 H 1.5 S iM o 12 O 40

97 0 8 71 91 0

959 851

735

900

1020

162 4

Transmittance

Transmittance

600

-1

C s 2.5 H 1.5 S iW 12 O 40 97 0

87 1

740

1612

C s 2.5 H 1.5 S iM o 1 2 O 40 959

910

900

735

851 740

102 1 1700 964

1610

9 70 9 05 8 71

H 4 S iW 12 O 40

H 4 S iM o 12 O 40 727

851 901

180 0

1600

1400

1 200

1000

W avenum ber / cm

800

600

1800

1600

-1

1400

1200

1000

W avenum ber / cm

740 800

600

-1

Fig. 24. IR spectra of the: A) H3PW12O40 and their cesium and rubidium salts (Cs2.5H0.5PW12O40, Rb2.5H0.5PW12O40), B) H3PMo12O40 and their cesium and rubidium salts (Cs2.5H0.5PMo12O40, Rb2.5H0.5PMo12O40), C) H4SiW12O40 and their cesium and rubidium salts (Cs2.5H1.5SiW12O40, Rb2.5H1.5SiW12O40)

and

D)

H4SiMo12O40

and

their

cesium

and

rubidium

salts

(Cs2.5H1.5SiMo12O40, Rb2.5H1.5SiMo12O40) recorded at ambient temperature.

53

We can conclude from the results that the primary Keggin structure remains unaltered even when the protons form parental heteropolyacid are substituted by the cesium or rubidium cations. The peak at ca. 1700 cm-1 in the HPW and HSiW IR spectra is due to the presence of dioxonium ions (H5O2+) and their relative intensity decay is observed upon proton substitution by Cs+ or Rb+ cation. Such pattern is observed here because the hexahydrate structures isomorphous with H3PW12O40·6H2O also exist in other dodecaheteropolyacids, like tungstosilicic acids in which only three of four protons are forming dioxonium ions. Thus their proper formula is (H5O2+)3(HSiW12O403-) as described in the literature.[190]

6.3 SEM characterization of Keggin-type matrix Scanning electron micrographs (SEM) of pure 12-phosphotungstic acid, 12phosphomolybdic acid, 12-silicotungstic acid, 12-silicomolybdic acid and their cesium and rubidium salts are shown in Fig. 25, Fig. 26, Fig. 27 and Fig. 28, respectively. The SEM image of pure phosphotungstic acid (HPW) reveals the presence of a (Fig. 25A) mixture of small (micro size) crystals together with few larger crystals.[191,192] The SEM micrograph of Cs2.5H0.5PW12O40 (Fig. 25B) shows that the bulk salt consists of agglomerates of smaller nanocrystals. The nanosize spherical aggregates of bulk Rb2.5H0.5PW12O40 (Fig. 25C) appear to be bigger than those of Cs2.5H0.5PW12O40. Fig. 26A shows the SEM micrograph of the pure phosphomolybdic acid (HPMo) in which arrays of uniformly small crystals are observed. The SEM micrograph of pure HPMo suggests a more uniform crystal texture as compared to that of pure HPW. The presence of Keggin structure is not so prominent in HPMo as compared to HPW because molybdenum (Mo) atom is of smaller size, Cs2.5H0.5PMo12O40, when compared to tungsten (W). Both salts consist of aggregates composed of small spherical nanocrystals grains that appear to be smaller for Cs2.5PMo12 than for the analogous rubidium salt.

54

Fig. 25. SEM micrographs of

Fig. 26. SEM micrographs of

A) H3PW12O40 (3000x magnification),

A) H3PMo12O40 (3000x magnification),

B) Cs2.5H0.5PW12O40 (8000x magnification),

B) Cs2.5H0.5PMo12O40 (8000x magnification),

C) Rb2.5H0.5PW12O40 (8000x magnification).

C) Rb2.5H0.5PMo12O40 (8000x magnification).

55

Fig. 27. SEM micrographs of

Fig. 28. SEM micrographs of

A) H4SiW12O40 (3000x magnification),

A) H4SiMo12O40 (3000x magnification),

B) Cs2.5H1.5SiW12O40 (8000x magnification),

B) Cs2.5H1.5SiMo12O40 (8000x magnification),

C) Rb2.5H1.5SiW12O40 (8000x magnification).

C) Rb2.5H1.5SiMo12O40 (8000x magnification).

The SEM image of pure silicotungstic acid (Fig. 27A) and silicomolybdic acid (Fig. 28A) shows sizeable agglomerates containing mixture of micro size crystals. In contrast, the cesium and rubidium salts of HSiW and HSiMo are composed of fine

56

particles. Smaller particles are obtained for Rb+ than for Cs+ salts of HSiMo and HSiW (Fig. 27, 28), what is opposite to cesium and rubidium salts of HPW and HPMo (Fig. 25, 26). In conclusion, we have to note that the surface state and microstructure of Cs+ and Rb+ salts are in strict contrast with heteropolyacids. The relatively smallest particles were obtained for cesium salt of tungsten heteropolyacid (Fig. 25B).

6.4 Cyclic voltammetry characterization of Keggin-type matrix

In order to prepare inks for cyclic voltammetric characterization, 0.02 g of cesium, rubidium and ammonium salts of 12-phosphotungstic acid, (HPW), 12phosphomolybdic acid, (HPMo), 12-silicotungstic acid, (HSiW) or 12-silicomolybdic 3 acid, (HSiMo) (described by the general formula M1x H y−x M 2 M 12 O 40 , where M1 is Cs+,

Rb+ or NH4+ and M2 is P or Si, M3 is W or Mo and x is 2.5 and y is 3 or 4 if M3 is P or Si, respectively) was mixed with Nafion (5% alcoholic solution) and ethanol (>99.9%) as a solvent and left for 12 hours on the magnetic stirrer. Next 3 µl of the suspension was dropped on the glassy carbon electrode surface (GC) and left for 30 minutes to dry. Cyclic voltammetric response of investigated heteropolyacid salts deposits on glassy carbon electrode recorded in argon saturated 0.5 mol dm-3 H2SO4 (scan rate: 50 mV s-1) are presented in Fig. 29. Fig.

29a

shows

typical

cyclic

voltammogram

for

Keggin-type

12-

phosphotungstic acid. The two couples of peaks at potentials -27 mV and 240 mV correspond to two redox reactions and can be described in terms of two consecutive reversible one-electron processes[20,21,193]: VI

3-

-

+

V

VI

PW12 O40 + ne + nH ↔ HnPWn W12-n O40

3-

(55)

where n is equal to 1 or 2. Voltammograms recorded for cesium (Fig. 29b), rubidium (Fig. 29c) and ammonium (Fig. 29d) salt of tungsten heteropolyacid are slightly different from those obtained for H3PW12O40. In the cyclic voltammogram of Cs2.5H0.5PW12O40 and Rb2.5H0.5PW12O40 instead of two couples of peaks only one peak at circa 68 mV can be seen in the potential range from –0.2 to 0.85 V. Thus the peaks characteristic of salts are shifted towards lower potentials in comparison to the original tungsten heteropolyacid and that only the first reduction can be observed in the potential 57

window examined. A different behavior is observed for ammonium salt of HPW, ((NH4)2.5H0.5PW12O40). In the Fig.29d we see two couples of peaks at 23 mV and 235 mV. This voltammogram is similar to that of pure HPW and may indicate that the salt is partially soluble in aqueous solution of H2SO4 thus releasing the acidic anion. Cyclic voltammogram of 12-phosphomolybdic acid (HPMo) is presented in the Fig. 29e. In this case we can see three couples of peaks at 280, 445 and 640 mV. These peaks correspond to the three consecutive two-electron processes that can be described by the following reaction[21,22,187]: VI

3-

-

+

V

VI

PMo12 O40 + ne + nH ↔ HnPMon W12-n O40

3

(56)

where n is equal to 2, 4 or 6. Fig. 29f-h illustrates the cesium (Cs2.5PMo12), rubidium (Rb2.5PMo12) and ammonium ((NH4)2.5PMo12) salt of HPMo. In all these voltammograms we observe three couples of peaks like in parental 12phosphomolybdic acid, but only for Cs2.5PMo12 the peaks are shifted towards less positive potentials values (from 640 to 570 mV, from 445 to 400 mV and from 280 to 210 mV). The cyclic voltammogram of (NH4)2.5H0.5PMo12O40 is very similar to that obtained for unmodified H3PMo12O40. A typical cyclic voltammogram for Keggin-type silica heteropolyacid, H4SiW12O40 (HSiW) is shown in Fig. 29i. The two couples of peaks that are present in the cyclic voltammogram at potentials -170 mV and 42 mV correspond to two redox reaction and can be described in terms of two consecutive reversible one-electron processes[20,193]: VI

4-

-

+

V

VI

SiW12 O40 + ne + nH ↔ HnSiWn W12-n O40

4-

(57)

where n is equal to 1 or 2. In Fig. 29j we can see that for Cs2.5SiW peaks are shifted towards less positive potential values (from 42 to –25 mV and from –170 to –220 mV).

58

0.6

0.04

0.05

0.3

0.00

-0.05

-0.3

a

-0.04

e

-0.6 0.03

0.2

-2

-0.2

b

0.2 0.0

-0.03

-0.2

-0.06

0.1

i

-0.10

0.00

0.0

j / mA cm

0.00

0.0

f

0.1

-0.4 0.2

0.0

0.0

-0.1

-0.2

j

0.0

c

-0.1

g

-0.2

0.2

0.6 0.3

0.0

k

-0.4 0.05 0.00

0.0

-0.05

-0.3

d

-0.2

h

-0.6

-0.2

0.0

0.2

0.4

0.6

0.2

0.8

E / V vs. RHE

0.4

0.6

0.8

l

-0.10 -0.2 0.0 0.2 0.4 0.6 0.8

E / V vs. RHE

E / V vs. RHE

0.3

Fig.29. Cyclic voltammetric response of 0.0

(a) H3PW12O40, (b) Cs2.5H0.5PW12O40,

-0.3

j / mA cm

-2

(c) Rb2.5H0.5PW12O40, (d) (NH4)2.5H0.5PW12O40,

em

-0.6 0.2

(g) Rb2.5H0.5PMo12O40, (h) (NH4)2.5H0.5PMo12O40,

0.0

(i) H4SiW12O40, (j) Cs2.5H1.5SiW12O40, (k) Rb2.5H1.5SiW12O40, (l) (NH4)4SiW12O40,

-0.2

fn

-0.4 0.2

0.4

0.6

0.8

-2

(m) H4SiMo12O40, (n) Rb2.5H1.5SiMo12O40, (o) Cs2.5H0.5SiMo12O40 deposits on glassy carbon electrode.

E / V vs. RHE j / mA cm

(e) H3PMo12O40, (f) Cs2.5H0.5PMo12O40,

Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.

0.02

Scan rate: 50 mV s-1. 0.00 -0.02

go

-0.04 0.3

0.4

0.5

0.6

0.7

0.8

E / V vs. RHE

59

This behavior is not observed for rubidium (Rb2.5SiW12) and ammonium ((NH4)4SiW12) salts of 12-silicotungstic acid. Fig. 29l shows, instead of cyclic voltammogram of (NH4)2.5H1.5SiW12O40 ((NH4)2.5SiW12), voltammetric response of (NH4)4SiW12O40. Having in mind the fact that ammonium salt of 12-silicotungstic acid containing 2.5 moles of NH4+ is completely soluble in water, it cannot be used as a matrix material. Fig. 29m shows a typical cyclic voltammogram of Keggin-type 12silicomolybdic acid. The three couples of peaks at potentials 270 mV, 450 mV and 550 mV correspond to three consecutive redox processes each involving two-electron[20,194]: VI

4-

-

+

V

VI

SiMo12 O40 + ne + nH ↔ HnSiMon W12-n O40

4-

(58)

where n is equal to 2, 4 or 6. The cyclic voltammetric responses of Rb2.5H1.5SiMo12O40 (Rb2.5SiMo12) and Cs2.5H1.5SiMo12O40 (Cs2.5SiMo12) are presented in Fig.29n and Fig.29, respectively. Three peaks located at potential similar to those of HSiMo are seen for Rb2.5SiMo. In the potential range studied, only two couples of peaks can be observed for Cs2.5SiMo12. The cyclic voltammetric response of ammonium salt of HSiMo is not presented because (NH4)2.5H1.5SiMo12O40 is completely soluble in water. Full substitution of all the protons in the 12-silicomolybdic acid to produce (NH4)4SiW12O40 also results into a completely soluble product. Some of the voltammograms recorded for investigated materials deposited on the glassy carbon electrodes shows similar behavior to those voltammograms obtained for their analogues in the solution. It is due to partial solubility in solution. It can be concluded, judging from CV measurements, that the most promising materials for the anode matrix in fuel cells are cesium salts of Keggin-type heteropolyacids (HPA). It seems that rubidium salt of HPA could also be of practical importance. Ammonium salts of Keggin-type heteropolyacids are certainly less attractive because they are more soluble in comparison to cesium and rubidium salts.

60

6.5 CONCLUSIONS 1) The results obtained using Infrared Spectroscopy technique (IR) demonstrate that the primary Keggin structure remains unaltered when the protons existing in parental heteropolyacid (HPW, HPMo, HSiW, HSiMo) are substituted with cesium or rubidium cations (Cs2.5PW12, Rb2.5PW12, Cs2.5PMo12, Rb2.5PMo12, Cs2.5SiW12, Rb2.5SiW12, Cs2.5SiMo12, Rb2.5SiMo12). The peak at c.a. 1700 cm-1 in the HPW and HSiW IR spectra due to dioxonium ions (H5O2+) decreases in intensity with increasing the number of protons substituted by Cs+ and Rb+ cation. 2) Images obtained using Scanning Electron Microscope (SEM) clearly show that cesium and rubidium substituted Keggin-type heteropolyacids posses higher surface area than theirs parental heteropolyacids. Consequently formation of small spherical crystallites where proton in HPA is substituted by cesium or rubidium cations is feasible. The smallest nanoparticles were obtained when protons in Keggin-type 12phosphotungstic acid were substituted by cesium cations (to produce Cs2.5H0.5PW12O40). 3) The voltammograms indicate that the most promising material for the anode matrix in fuel cells are salt of Keggin-type heteropolyacids containing cesium and rubidium cations. 4) Results obtained here show that, relatively high solubility in the water, ammonium salts of HPAs are less attractive than their cesium or rubidium salt analogues. This observation is of practical importance when it comes to preparation of catalytic layers for fuel cells. 5) On the basis of our results described in this chapter, the salts of Keggin-type heteropolyacids containing 2.5 moles of cesium and rubidium cations in 1 mole of the heteropolyacids salts (Cs2.5PW12, Rb2.5PW12, Cs2.5PMo12, Rb2.5PMo12, Cs2.5SiW12, Rb2.5SiW12, Cs2.5SiMo12, Rb2.5SiMo12) have been chosen for future research.

61

7. Hydrogen oxidation reaction (HOR) on the catalytic layers containing Cs2.5PW12 matrix This chapter is devoted to the development of a catalyst consisting of Pt nanoparticles supported on insoluble salts of Keggin-type heteropolyacids to be used for hydrogen oxidation reaction (HOR). Two methods of preparation of catalytic layers are described through mixing and the electrochemical method. The best systems in each preparation method contained a new type of matrix as demonstrated and compared to unmodified commercial carbon supported platinum catalyst using rotating disc electrode (RDE) voltammetry. High resolution transmission electron microscopy (HRTEM) has been used to study the dispersion platinum nanoparticles deposited on the matrix by using electrochemical method. CO stripping voltammetry has also been applied to estimate the active surface of Pt existing in the catalytic layer (prepared by electrochemical method).

7.1. Preparation of the catalytic layers

7.1.1. Mixing method The steps followed for the preparation of composite catalytic layers by mixing is shown schematically in Fig. 30. A suspension in ethanol of Pt10%/Vulcan XC-72 carbon and heteropolyacid salt (Cs2.5H0.5PW12O40) was mixed with know amount of Nafion (5% alcoholic solution) and stirred in a close vial for 12 h. The mass ratio between Pt10%/Vulcan XC-72 : heteropolyacids salts : Nafion was 1:2:1.1, respectively. A portion of the resulting ink was then dropped using micropipette on to the surface of glassy carbon electrode (RDE electrode) in such way to obtain Pt loading equal 30 µg cm-2. The resulting catalytic layer was left to dry for 30 min. at room temperature. For comparison, the heteropolyacid salt-free ink of Pt10%/Vulcan XC-72 carbon and Nafion was also prepared.

62

Ethanol Nafion

Pt 10%/C – Cs2.5PW12 Mixing for 24h

Cs2.5H0.5PW12O Pt on Vulcan Nafion

Glassy carbon

Pt 10%/C – Cs2.5PW12 with Nafion

Fig. 30. Schematic of the preparation of a composite catalytic layer by mixing.

7.1.2. Electrochemical method To prepare the matrix for electrochemical plating the Cs2.5H0.5PW12O40 was mixed with Vulcan XC-72 carbon, Nafion solution (5% of aliphatic alcohols) and ethanol. After 12 hours of mixing (under magnetic stirring) desired amount of the suspension was dropped on the glassy carbon rotating disc electrode (RDE) surface and left for 30 minutes to dry. The best results were obtained when the catalytic ink was prepared by mixing Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and Nafion in the following proportion 2:1:1.1. A standard three-electrode cell (described previously in the chapter 5) was used to prepare films and to perform other electrochemical measurements. A platinum flag served as a counter electrode; a saturated calomel electrode (SCE), placed in a separated compartment and connected to the main cell through a Luggin capillary, was used as reference electrode. Preparation of catalytic layer Pt-Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and Nafion was accomplished by cycling (2500 cycles) the electrode between -0.05 V

63

to 1.05 V vs. RHE at scan rate 50 mV s-1 in a solution of 0.5 M of H2SO4 containing 5 10-3 M KCl. During cycling the counter electrode can reach very anodic potentials causing partial oxidation of the Pt flag. The Pt ions may eventually deposit on the -

working electrode. The presence of ions Cl in the solution helps the process, and decrease of the standard oxidation potential Pt/Pt(II), E0 = 1.2 V vs. RHE, by formation of complexes[195]: 2-

2-

PtCl6 + 2e = PtCl4 + 2Cl 2-

PtCl4 + 2e = Pt + 4Cl

-

-

0.68 V vs. RHE

(59)

0.73 V vs. RHE

(60)

After deposition of platinum within the matrix, the electrode was washed with water and subjected to cycling in 0.5 M H2SO4 in potential range from 0 V to 1.05 V (vs. RHE) to remove Cl- from the catalytic film. The steps followed for the electrode preparation are shown schematically in Fig. 31.

Counter electrode Pt flag

Anodic dissolution of Pt counter electrode

Working Electrode (GC)

Matrix (Cs2.5H0.5PW12O40)

Fig. 31. Scheme of the preparation catalytic layer by electrochemical method.

64

7.2 Electrochemical measurements on the Pt/C modified with Cs2.5PW12 system prepared by mixing method 7.2.1 Cyclic voltammetry (CV) and rotating disc voltammetry (RDE) measurements

Fig. 32 shows the cyclic voltammetric responses (recorded in argon-saturated 0.5 M H2SO4 ) of a glassy carbon electrode modified with Nafion-treated (---) unmodified Pt10%/Vulcan XC-72 carbon (Pt10%/C) and (—) Pt10%/Vulcan XC-72 carbon modified with cesium salt of 12-phosphotungstic acid (Cs2.5H0.5PW12O40). The platinum loading in both cases was 30 µg cm-2.

4

j / mA cm

-2

2

0

-2

-4

-6 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 32. Cyclic voltammetric response of Nafion-containing films (deposited on rotating disc glassy carbon electrode) of (—) Pt10%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, and (---) unmodified Pt10%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 M H2SO4. Scan rate, 50 mV s-1.

It is apparent from Fig. 32 that reduction of Cs2.5H0.5PW12O40 (in the range of potentials from 0 to 0.3 V) tends to overlap with the so-called hydrogen adsorption peaks that typically exist on clean bare platinum. The electroactivity of Cs2.5H0.5PW12O40 in the hydrogen adsorption/desorption region of Pt (at potential lower than 0.35 V) is clearly evident in Fig. 32. Further, following CV response for Cs2.5PW12 with Pt/C, the

65

voltammetric peak (at potential higher than 0.7 V) referring to the formation of Pt-oxo (PtO or PtOH) species is shifted towards more positive potentials. Increase of the potential range, where platinum is not covered with platinum oxide, may facilitate adsorption and activation of hydrogen molecules during electrocatalysis.[21] The hydrogen electrooxidation reaction on Pt in acid solution is one of the fastest known electrochemical reactions. To evaluate electrocatalytic activity towards the hydrogen oxidation, we have performed the diagnostic rotating disc electrode (RDE) voltammetric measurements using a glassy carbon disc electrode modified with Nafiontreated (Fig. 33A) Pt10%/Vulcan XC–72 carbon modified with Cs2.5H0.5PW12O40 in comparison to (Fig. 33B) unmodified Pt10%/Vulcan XC–72 carbon. The RDE measurements (Fig. 33) for the hydrogen oxidation have been performed at different rotation rates (ranging from 900 to 2500 rpm). The anodic potential was limited to 100 mV, because the diffusion limiting current for both catalytic layer reach a constant value at ca. 70 mV, and the coupled hydrogen adsorption-electron transfer steps of the reaction are only observed at potentials below 50 mV. The examination procedure was the same for both systems. It becomes obvious that current density obtained for catalytic layer modified with cesium salt of tungsten heteropolyacid (Cs2.5H0.5PW12O40) is higher (Fig. 33A), when compared to the corresponding current density of the bare commercial Pt10%/Vulcan XC-72 carbon nanoparticles (Fig. 33B). The above results may imply that modification of Pt10%/Vulcan XC-72 carbon with Cs2.5H0.5PW12O40 matrix results in increasing dispersion of catalytic centers (higher exposing of Pt particles) on which proceed hydrogen oxidation reaction (HOR). It can be also explained by the different diffusion speed into the catalytic layer, what was described in the literature.[46]

66

3.0

ω / rpm

A

2500

j / mA cm-2

2.5 2.0 900

1.5 1.0 0.5 0.0 0.00

0.03

0.06

0.09

0.12

0.15

E / V vs. RHE

3.0

B

ω / rpm

2.5

j / mA cm

-2

2500

2.0 900

1.5

1.0

0.5

0.0 0.00

0.03

0.06

0.09

0.12

0.15

E / V vs. RHE

Fig. 33. RDE voltammograms recorded in the hydrogen saturated 0.5 M H2SO4 solution at different rotation rates (900, 1100, 1300, 1500, 1700, 1900, 2100, 2300, 2500 rpm) using a glassy carbon disc electrode covered with Nafion-treated (Fig. 33A) Pt10%/Vulcan XC–72 carbon modified with Cs2.5H0.5PW12O40 and (Fig. 33B) unmodified Pt10%/Vulcan XC–72 carbon. Scan rate, 10 mV s-1. The loading of Pt 30 µg cm-2.

The above effects become clearer when we compare them at certain rotation rate for modified and unmodified systems (Fig. 34). Moreover, based on the literature data[27,28] we can expect that HPAs can act as redox mediators for the electrochemical

67

oxidation of CO. This could be advantageous knowing that Pt can easily poison with CO.

3.0

Cs2.5H0.5PW12O40 with 10%Pt/Vulcan and Nafion 2 LPt = 30 µg/cm

2.5

j / mA cm-2

10%Pt/Vulcan and Nafion

2.0

2

LPt = 30 µg/cm

1.5 1.0 0.5 0.0 0.00

0.03

0.06

0.09

0.12

0.15

E / V vs. RHE Fig. 34. Comparison of RDE voltammograms recorded in the hydrogen saturated 0.5 mol dm-3 H2SO4 solution using rotating disc glassy carbon electrode covered with films of Nafioncontaining inks of (red line) Pt10%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, and (black line) unmodified Pt10%/Vulcan XC-72 carbon Rotation rate, ω = 2500 rpm. Scan rate, 10 mV s-1.

Fig. 33 and Fig. 34 confirm that modification of commercial Pt10%/Vulcan XC72 by Cs2.5H0.5PW12O40 matrix increases dispersion of platinum centers which are in contact with hydrogen. The dependence of the RDE limiting currents versus the square root of rotation rates (Fig. 35) shows linearity behavior, i.e. ideal behavior characteristic of a system limited solely by convective diffusion of hydrogen in solution.

68

3.0 on afi N +

j / mA cm-2

2.5

O 40 12 W P

H 0.5

.5

2.0 C %/ 10 t P

1.5

) 72 CX (

s2 +C

C %/ 10 t P

) 72 CX (

on afi N +

1.0

0.5

0.0 0

10

20

30 1/2

40

50

1/2

ω / (rpm)

Fig. 35. Levich plots j vs. ω1/2 prepared using the data of Fig. 33 for (red symbols) Cs2.5PW12modified GC-supported Pt10%/Vulcan XC-72 carbon nanoparticles. Currents were measured at 0.1 V. For comparison the same plot is provided for (black symbols) GC-supported Pt10%/Vulcan XC-72 carbon nanoparticles of the same loading (30 µg cm-2).

Assuming laminar flow and the mass transport rate, the diffusion limited current density is mathematically described by the Levich equation, as a function of the rotational frequency of the RDE, ω (in radians per seconds)[196,197]: 2/3 -1/6

ίd = 0.62nFD

ν

1/2

c0ω

= Bc0ω

1/2

(61)

where, D is the diffusivity of hydrogen in 0.5 M H2SO4 (D298

K

-5

2

= 3.7 x 10 cm /s,

estimated from the product of H2 diffusivity at infinite dilution and the ratio of the dynamic viscosities of the electrolyte and pure water), n is the number of electrons in the H2 oxidation reaction (i.e., n = 2), ν is the kinematics viscosity of the electrolyte -2

2

(ν298 K = 1.07 x 10 cm /s), c0 is the solubility of H2 in 0.5 M H2SO4 (c298 K = 7.14 x 10

-3

M). In 0.5 M H2SO4, the calculated value of Levich constant, B, and the solubility, c0, -2

-2

i.e., Bc0, at 298 K is equal to 6.54 x 10 (mA cm )rpm

-1/2

.

It have to be point out that in the presence of Nafion® film, mass-transport limitations through the film lowered the limiting current at each rotation speed. That’s why for the

69

Nafion®-coated RDE we have additional term including hydrogen diffusion into the layer[46]: ifilm = nFADfcf / δf

(62) 2

where, Df is the diffusion coefficient of H2 in the recast Nafion® (cm s-1), A is a geometric area of an electrode area (cm2), cf is a solubility of H2 in recast Nafion® (mol cm-3), δf is a nafion film thickness. The thickness of the Nafion® film on the investigated catalytic layers was equal 0.23 µm. Maruyama and coworkers[46] reported absence of the mass transfer effect when the thickness of Nafion® film varied in the range 1 to 13 µm. This means that in our case the influence of the ifilm is negligible and this element can be omission. According to the Eq. 61, a plot of the current density at constant potential versus ω1/2 (Fig. 35) ought to result in a straight line with the slope defined by Bc0. The regression, made for ours new type of system, yielded a slope of Bc0 = 6.07 · 10-2 (mA cm-2) rpm-1/2 -2

-2

which is in fairly good agreement with above calculated value (6.54 x 10 (mA cm ) rpm

-1/2

) (-8%) considering the uncertainties associated with the evaluation of the

diffusivity of hydrogen in the electrolyte. Incidentally, the above values apply to the entire potential range between 0.1 and 0.3 V. For Nafion-containing Pt10%/Vulcan XC72 carbon (Cs2.5PW12-free system showed as a comparison) the regression yielded a slope of Bc0 = 6 · 10-2 (mA cm-2) rpm-1/2. Fig. 36, presents mass transport corrected Tafel diagrams for the investigated catalytic layers. Mass transport correction for rotation disc electrode is described as follows

[54]

:

ik = id i / id – i

(63)

where, i is the experimentally obtained current, id refers to the measured diffusionlimited current, and ik the mass-transport free kinetic current.. The RDE polarization data analyzed here allow obtaining information on the kinetics of the HOR. On polycrystalline platinum, oxidation of hydrogen involving two electrons in acidic media can proceed according to two different pathways:

by chemical adsorption step (Tafel-Volmer mechanism) H2 + 2M → 2MH

(Tafel reaction)

(5) 70

or electrochemical adsorption step (Heyrovsky-Volmer mechanism) H2 + M → MH + H+ + e-

(Heyrovsky reaction)

(6)

followed by the hydrogen atom discharge path given by MH ↔ M + H+ + e-

(Volmer reaction)

(7)

The rates of the reactions (5) and (6) can change depending on the electrode material and electrolyte.[44,51] Taking attention that in our case influence of the ifilm is negligible the equations have been derived for analysis of the RDE voltammograms for the cases of reversible or irreversible electrochemical reaction. The kinetic equations are[45,49,52,198]:

E = E10' −

j − 2.303RT log L nF  jL

j  

(64)

E = E20' −

 j ⋅j 2.303RT log L  αnF  jL − j 

(65)

where, for the reversible (Eq. 64) and irreversible (Eq. 65) cases, respectively. In the 0’

0’

equations E1 and E2 are current independent constants, j is the current density, jL is the diffusion limiting current density and α is the change transfer coefficient.

71

0.0

a

900 rpm 1700 rpm 2100 rpm

log[(jL-j)/jL]

log[(jL-j)/jL]

0.0

-0.5

c

900 rpm 1700 rpm 2100 rpm

-0.5

-1.0

d

b

log[(jLx j)/(jL - j)]

log[(jLx j)/(jL - j)]

-1.0

1

0

900 rpm 1700 rpm 2100 rpm

1

0 900 rpm 1700 rpm 2100 rpm

-1 0.00

0.01

0.02

0.03

0.04

E / V vs. RHE

0.00

0.01

0.02

0.03

0.04

E / V vs. NHE

Fig. 36. Mass transport corrected Tafel diagrams [E ~ log(jL – j / jL) and E ~ log(jL · j / jL - j)] for HOR on catalytic layer prepared from Nafion-containing Cs2.5H0.5PW12O40 with Pt10% on Vulcan XC-72carbon, (a) equation (64) and (b) equation (65) and as a comparison catalytic layer made from Nafion-containing Pt10% on Vulcan XC-72 carbon, (c) equation (64) and (d)equation (65) in 0.5 M H2SO4. Rotation rates: (■) 900 rpm, (ο) 1700, (∆) 2500 rpm. Data taken from Fig. 33.

If either condition is satisfied, according to equations (64) and (65), the plots log(jL – j / jL) vs. E should be linear with slopes independent of the rotation rates and equal to nF/2.303 RT or αnF/2.303 RT.[169,197] As we can see in Fig. 36, for all the catalytic layers the plots are linear only if a reversible kinetics is assumed. The reverse values of the slopes (the Tafel slope E ~ log(jL · j / jL - j)), presented in Table 6, are very close to the theoretical value obtained -1

for platinum 29 mV dec at temperature 293 K. However, there is no agreement with the prediction coming from equation (65). In this case we should obtain slope equal 58 -1

mV dec from mass transport corrected Tafel diagram. For potentials below 0.01 V, there is no linear relationship for the plot E ~ log(jL · j / jL - j), however, if we consider -1

slope values for E > 0.01 V, we obtain values on the level 28 - 31 mV dec what is -1

more then two times lower in comparison to the predicted value of 58 mV dec . From these results it may be concluded that the HOR takes places via the Tafel-Volmer mechanism with the atom-atom recombination step (Tafel) as a rate determining step

72

(rds) at both layers. On this type of electrode, in strong acidic media, a Tafel-Volmer mechanism, with Tafel as the rds, has also been proposed by Mello and co-workers.[44] Exchange currents (I0) was calculated from the slope of the linear polarization response and corrected for diffusion by using the following equation[52]:

∆E RT  1 1   +  = ∆I nF  I 0 I L 

(66)

where: n = 2 and IL is the limiting current. Values obtained from this equation are presented in Table 6.

breversible (mV dec-1)

birreversible (mV dec-1)

Pt10% on Vulcan XC-72 carbon

32

28

Cs2.5H0.5PW12O40 with Pt10% on Vulcan XC-72 carbon

32

31

32

35

Composition of catalytic layer

Babić at all[52]. Pt20% on Vulcan XC-27 carbon

Theoretical breversible (mV dec-1)

Theoretical birreversible (mV dec-1)

29 (Tafel-Volmer) 118 (HeyrovskyVolmer)

58 (Tafel-Volmer, HeyrovskyVolmer)

I0 (mA)

0.63

0.83

0.55

Table 6. Theoretical and experimental Tafel parameters, mechanism and corresponding values of exchange currents for HOR in acid medium at 25 0C for the catalytic layer made from: Nafion-containing Cs2.5H0.5PW12O40 with Pt10%/Vulcan XC-72 carbon and for Cs2.5H0.5PW12O40 – free system (as a comparison).

In conclusion, we should note that, in both systems, hydrogen oxidation reaction follows the Tafel-Volmer mechanism with the atom-atom recombination step (Tafel) as rds. Calculated values of Bc0 for investigated materials are close to the theoretical value (6.54 x 10-2 mA cm-2 rpm-1/2). Nevertheless the value of diffusion limiting current for ours system is 12% higher than that for unmodified one. Moreover exchange current calculated for Cs2.5PW12-containing catalytic layer is 32% higher than that for the system containing unmodified commercial available Pt/Vulcan XC-72 carbon nanoparticles. The value of I0 obtained for unmodified system (Pt/C) is in a good agreement with that reported by Babić at all.[52] All these issues clearly show that our

73

catalytic layer containing Pt10%/Vulcan XC-72 carbon together with a matrix obtained from cesium salt of 12-phosphotungstic acid (Cs2.5H0.2PW12O40) possesses many advantages in comparison with Cs2.5PW12-free system.

7.2.2 Test of the Cs2.5H0.5PW12O40-containing anode catalyst in working single PEM fuel cell The mixing method was used to prepare inks to be utilized as anode catalyst in a working fuel cell. Cs2.5H0.2PW12O40, Pt10%/Vulcan XC-72 and Vulcan XC-72 were mixed in 99% ethanol and stirred in a close container for 24 hours. 200 µl of Nafion solution were then added to the homogeneous suspension and stirring was continued for 12 hours. The relative amounts of the different component are listed in Table 7. Name of the component Cs2.5H0.2PW12O40 Pt10% on Vulcan XC-72 carbon Vulcan XC-72 carbon Nafion

Weight in grams 0.0073 0.0059 0.0067 0.0087

Table 7. Amounts of components used for preparation anode for PMEFC.

The ink was then spread homogeneously onto the surface of a commercial gas diffusion layer (low-temperature ELAT® GDL microporous layer on woven web, tin configuration), purchased from E-TEK, (5x2.5 cm) previously weighted using a brush. The resulting GDL with catalytic layer (CL) was left to dry in an oven for 30 min. at 80 0

C and weighted again to determine the amount of catalytic mixture in terms of Pt

content per cm2. Pt loading was 43 µg cm-2. The same procedure was used to prepare a Cs2.5H0.2PW12O40-free catalytic layer using commercial Pt10%/Vulcan XC-72 carbon from E-TEK. The Pt loading for this electrode was 50 µg cm-2. 2.2 x 2.2 cm portions of the two layers were then cut to obtain the electrode to be used for preparing a MEA. Membrane electrode assemblies (MEA) are built by hot pressing two of the above electrodes at the opposite sides of a Nafion membrane (thick, 127 µm). Nafion® Membrane 115 purchased from Ion Power Inc., (dimension) 7.5x7.5 cm, was used as a proton conducting membrane. Two types of MEA have been built and tested: MEA1 was prepared using the composite electrode at one side and the commercial catalyst on the other side; MEA2, to

74

be used for comparison, had the commercial catalyst on both sides. The preparation conditions are listed in Table 8 in the case of MEA1. Anode electrode Cathode electrode Cs2.5H0.2PW12O40 Pt10%/C Pt10%/C Vulcan XC-72 Vulcan XC-72

Membrane Temperature Pressure Time / 0C / bar / min. Nafion® 125 15 3 115

Table 8. Components and conditions for the hot pressing MEA preparation.

The MEA have been tested using a 5 cm2 single cell, from Fuel Cell Technologies, operated equipped with single serpentine flow channel as gas feeds, showed in Fig. 37. The cells were operated using pure H2 and O2.

H2 part

A

B serpentine flow channel gas intlet

gas outlet

O2 part bipolar plate

Fig. 37. Picture of the (A) single PEM fuel cell and (B) one side of PEMFC with bipolar plate and single serpentine flow channel

Fig. 38 and Fig. 39 show the polarization curves and the power curves for both MEA under the conditions specified in Table. 9.

75

1,0

MEA1

Cell voltage / V

0,8

MEA2

0,6

0,4

0,2 2

2

H2 -- LPt = 50 µg/cm Pt // Pt LPt = 50 µg/cm -- O2 2

2

H2 -- LPt = 43 µg/cm Pt + Cs2.5H0.5PW12O40 // Pt LPt = 50 µg/cm -- O2

0,0 0

200

400

600

800 -2

Current density / mA cm

Fig. 38. Current-voltage PEMFC’s single cell performances of (red symbols) Cs2.5H0.5PW12O40containing and (black symbols) Cs2.5H0.5PW12O40-free anode: 70 0C of cell temperature, 70 0C for both humidifiers, 3 bars of operating pressure.

Power Density / mW cm-2

400

300

200

100 2

2

H2 -- LPt = 50 µg/cm Pt // Pt LPt = 50 µg/cm -- O2 2

2

H2 -- LPt = 43 µg/cm Pt + Cs2.5H0.5PW12O40 // Pt LPt = 50 µg/cm -- O2

0 0

200

400

600

800

-2

Current density / mA cm

Fig. 39. PEMFC’s single cell performances of (red symbols) Cs2.5H0.5PW12O40-containing and (black symbols) Cs2.5H0.5PW12O40-free anode: 70 0C of cell temperature, 70 0C for both humidifiers, 3 bars of operating pressure.

76

Catalysts and their loading (µg cm-2) Anode Cathode Cs2.5H0.2PW12O40 Pt10%/C Vulcan XC-72 Pt10%/C 50 µg cm-2 Vulcan XC-72 -2 43 µg cm Pt10%/C Pt10%/C Vulcan XC-72 Vulcan XC-72 50 µg cm-2 50 µg cm-2

Feeding conditions TCell (0C) Anode Cathode H2, O2, 70 50 ccm 50 ccm 3 bars 3 bars H2, 50 ccm 3 bars

O2, 50 ccm 3 bars

70

TAnode TCathode (0C) (0C)

Electrolyte

70

70

Nafion® 115

70

70

Nafion® 115

Table 9. Composition of electrodes, their loadings and operation conditions of single working PEMFC.

The polarization curves were obtained by applying 50 mA steps and recording the cell potential after 60 sec. As it may be seen better performances are obtained (either in terms of potential and power) when using MEA1. In this type of fuel cells the performances are limited by the cathode where oxygen reduction (a very sluggish electrochemical reaction) occurs. As the cathode catalyst is the same in both cases, the better performances of MEA1 have to be ascribed to the anode side that contains a catalyst layer modified with Cs2.5H0.5PW12O40.

Our results suggest that modification of commercial Pt10%/Vulcan XC-72 carbon with cesium salt of 12-phosphotungstic acid (Cs2.5H0.2PW12O40) leads to better utilization of catalytic centers as apparent from Fig. 38 and Fig. 39. Similar results were obtained by Stanis et al. where Pt/C was modified with H3PW12O40 heteropolyacid was used as an anode in PEM fuel cell operating on pure hydrogen as a fuel.[27] In conclusion, we should also remark that results obtained for our new system by using single working PEM fuel cell are in agreement with previous results received by using electrochemical methods, particularly cyclic voltammetry (CV) and rotating disc voltammetry (RDE) methods.

77

7.3 Electrochemical measurements on the Pt/C modified with Cs2.5H0.5PW12O40 system prepared by electrochemical method 7.3.1 Cyclic voltammetry study Corrosion of Pt was accomplished by applying sufficiently positive potential to -3 platinum flag counter electrode in 0.5 mol dm H2SO4 containing ions Cl (as was

described previously in chapter 7.1.2) on composite catalytic layers and layers prepared using the commercial catalyst alone. During negative potential scans applied to the working electrode electrodeposition take places onto catalytic layers existing on the surface of glassy carbon RDE electrode. The voltammograms before and after electrochemical deposition were recorded in the 0.5 mol dm-3 H2SO4 without Cl- (Fig. 40 and 41). Before recording voltammetric curve of the catalytic layer after electrodeposition of platinum, the modified electrode was washed with water and subjected to cycling in 0.5 M H2SO4 in potential range from 0 V to 1.05 V (vs. RHE) to remove Cl- from the catalytic film. 4

Pt as counter electrode after Pt corrosion before Pt corrosion

2

Carbon paper as counter electrode

0.4 -2

-2 j / mA cm

j / mA cm

-2

0

-4

after cycling beafore cycling

0.2 0.0 -0.2 -0.4

-6

0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE

-8 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE

Fig. 40. Cyclic voltammetric responses of Cs2.5H0.5PW12O40 with Vulcan before (dash line) and after (solid line) Pt corrosion from platinum counter electrode; Inset: Cs2.5PW12O40 with Vulcan before (dash line) and after (solid line) cycling where carbon paper was used as counter electrode; Electrolyte: argon saturated 0.5 M H2SO4. Scan rate, 50 mV s-1.

78

Fig. 40 shows the final cyclic voltammetric response of the composite catalytic layer before and after Pt continuous cycling together with the initial cyclic voltammograms. The final cyclic voltammogram shows the classic signature of hydrogen adsorption/desorption on Pt. The insert in Fig. 40 shows the cyclic voltammograms obtained under the same conditions with the same electrode when carbon paper instead of Pt was used as counter electrode. Apart from a slight increase of the currents in the hydrogen region, due to the reduction of the cesium salts (chapter 6.4), the shape of the curve is practically unaltered after prolonged cycling. The obvious conclusion that may be drawn is that the high current increase when the counter electrode is Pt is due to deposition of Pt on the composite electrode because of corrosion of the counter electrode. For a comparison the same procedure has been applied to Nafion-containing catalytic layer made from Vulcan XC-72 carbon (Cs2.5PW12-free system). Fig. 41 shows the cyclic voltammetric response of Vulcan before and after Pt corrosion. The results demonstrate that the procedure works also on Vulcan alone.

6

Vulcan before Pt corrosion Vulcan after Pt corrosion

4

j / mA cm

-2

2 0 -2 -4 -6 -8 -10 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 41. Cyclic voltammetric responses of Vulcan before (dash line) and after (solid line) Pt corrosion from platinum counter electrode; Electrolyte: argon saturated 0.5 M H2SO4; Scan rate, 50 mV s-1.

Again the development of the cyclic voltammetric curves with increasing number of scan reveals the progressive grow of the characteristic hydrogen adsorption/desorption waves close to 0 V that indicate presence of Pt on the electrode. 79

In order to obtain further evidences, beside the signature of hydrogen adsorption/desorption, of the effectiveness of the proposed method to dope the composite electrode (and Vulcan) with Pt nanoparticles by corrosion of a Pt counter electrode and to have an idea of the electrochemically active area, dimensions and Pt loading that may be obtained, alternative methods to detect Pt have been used. These include both CO stripping and TEM. The electrochemically active area at Pt electrodes is usually measured using the hydrogen adsorption/desorption peaks. In this case this method can not be applied because the Cs salts is electro-active in the same potential region and, hence, the integrated areas under the peaks contain the contributions due to the reduction of the compound itself.

7.3.2 CO electrooxidation on catalytic layer containing Cs2.5H0.5PW12O40 as a matrix It is well known in the literature that CO adsorbs on Pt poisoning it and cancelling any catalytic activity for hydrogen oxidation. Adsorbed CO can be stripped from the Pt surface at potentials in the range 0.8 - 1 V vs. RHE. The charge corresponding to one monolayer of adsorbed CO is equal to 0.484 mC cm-2.[199-201] Hence, by measuring the charge under the stripping peak, after subtraction of the background, the electrochemically active area can be obtained using the formula:

EAS = QCO / 0.484

(67)

where QCO is the measured charge.

Fig. 42 shows several voltammograms obtained at 20 mV/s using the composite electrode under different conditions. The black continuous line is relative to the Pt containing composite electrode as prepared, while the dashed black line shows the voltammogram obtained after CO adsorption. The red line corresponds to the voltammogram of the composite electrode before Pt deposition and the dashed blue line to that obtained with the same electrode after CO adsorption. The last two voltammograms demonstrate that there is no absorption of CO before Pt corrosion and, hence, that no errors are introduced due to CO adsorption in the salt.

80

The peak relative to hydrogen desorption present in the pristine electrode after activation in the hydrogen region is completely absent after CO absorption. This testifies that the entire Pt surface is completely blocked by CO. The stripping of CO at about 0.95 V is very well defined and sharp. The computed charge under the peak is equal to 0.2207 mC. Using Eq. 67 this translates into an electrochemically active area of 0.456 cm2.

Cs2.5PW12 -containing system after Pt corrosion

3

CO-adsorption on Cs2.5PW12 - containing system after Pt corrosion Cs2.5PW12 -containing Pt-free system

j / mA cm

-2

CO-adsorption on Cs2.5PW12 - containing Pt-free system

2

CO

1

0

-1 0.0

0.2

0.4

0.6

0.8

1.0

1.2

E / V vs. RHE Fig. 42. Base voltammetry (—) and CO-stripping (---) on a Nafion-containing Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded platinum and for a comparison base voltammetry (—) and CO-stripping (---) at a Pt-free Nafion-containing Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon electrode. Electrolyte, argon saturated 0.5 mol dm-3 H2SO4. Scan rate, 20 mV s-1.

7.3.3 HRTEM characterization

Fig. 43a and b show HRTEM images of Pt nanoparticles obtained using powders from the composite and the pure Vulcan electrodes after electrochemical Pt deposition. The images show that the nanoparticles of electrodeposited Pt on catalytic layer containing cesium salt of tungsten heteropolyacid are largely spherical in shape with a diameter of about 2 nm (Fig. 43a), while for the catalytic layer containing only Vulcan XC-72 carbon the platinum particles have diameter about 4 nm (Fig. 43b). A likely explanation of the different particle size may be that in the case of the Cs2.5H0.5PW12O40 containing composite Pt deposition occurs inside the nanochannels of 81

the porous tertiary structure of the salts that posses characteristic diameters while in the case of pure Vulcan the particles grow on the surface where are more prone to grow.

(a)

(b)

Fig. 43. HRTEM image of nafion-containing catalytic layer made from: (a) Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded Pt and (b) Vulcan XC-72 carbon with corroded Pt for a comparison.

In the case of the composite catalytic layer where the surface area is known one may attempt to compute the Pt loading by assuming uniform diameters and the absence of bigger aggregates average. Knowing dimensions of Pt nanoparticles and assuming theirs spherical shape we are able to calculate platinum loading into the catalytic layer by using the following equation[52,202]: S = (6 x 103) / ρd

(68)

where S is specific surface area (m2 g-1), d is the mean particle size in nm (from HRTEM) and ρ is the density of Pt metal (21.4 g cm-3). Assuming that all nanoparticles have dimension ca. 2 nm as well as the absence of aggregated nanoparticles, we obtain the surface area for one nanoparticle equal to 140.18 m2 g-1. Knowing EAS, electrochemically active area (0.456 cm2) received from CO stripping measurements (assuming 100% coverage) and S for one particle we are able to calculate loading of platinum into the catalytic layer by applying following equation:

82

LPt = (EAS/ S) / A

(69)

where, A is the area of the RDE electrode (A = 0.1256 cm2). According to this equation the calculated Pt loading was 2.6 µg cm-2. This approximate calculation was made to comment on the order of magnitude of platinum. The amounts of Pt are indeed varying very low what means that we are able to decrease quantity of this material without decreasing the systems performance towards hydrogen oxidation reaction (HOR).

7.3.4 Rotating disc voltammetry measurements The kinetic data of the HOR occurring at the catalytic layers prepared by electrochemical method has been done by applying the same type of analysis (using the same procedure) as for the systems prepared by mixing method (chapter 7.2.1). Fig. 44 presents the hydrogen oxidation polarization curves recorded at several rotation speeds (1300 rpm ≤ ω ≤ 2500 rpm) in 0.5 mol dm-3 H2SO4 on the following catalytic layers: (A) Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon, Nafion and corroded platinum and (B) Vulcan XC-72 carbon with Nafion and corroded platinum (for comparison). The anodic potential limit was to 100 mV because the limiting currents reach a constant value at ca. 80 mV in the case of both catalytic layers. Also the coupled hydrogen adsorption-electron transfer steps of the reaction are only observed at potentials below 50 mV. Current density obtained for catalytic layer containing Cs2.5H0.5PW12O40 matrix is higher than that obtained for catalytic layer containing only Vulcan XC-72 carbon.

83

ω / rpm

j / mA cm-2

2.5

A

2500

2.0 1300

1.5

1.0

0.5

0.0 0.00

0.02

0.04

0.06

0.08

0.10

0.12

E / V vs. RHE

j / mA cm-2

ω / rpm

B

2.5

2500

2.0 1300

1.5

1.0

0.5

0.0 0.00

0.02

0.04

0.06

0.08

0.10

0.12

E / V vs. RHE Fig. 44. Hydrogen oxidation RDE voltammograms for several rotation speeds obtained in 0.5 M H2SO4 under H2 atmosphere on Nafion-containing catalytic layer made from (A) Cs2.5H0.5PW12O40 with Vulcan XC-72 Carbon and corroded platinum and (B) Vulcan XC-72 Carbon with corroded platinum as a comparison, at different speed rotation, 1300 rpm ≤ ω ≤ 2500 rpm. Scan rate, 2 mV s-1.

84

Comparison of hydrogen oxidation RDE curves at 2500 rpm recorded for the system containing Cs2.5PW12 matrix with corroded platinum and this unmodified one with corroded platinum is illustrated in Fig. 45. Higher current density obtained for the modified material could be explained by formation of smaller Pt nanoparticles not only on the surfaces but also possibly inside the channels in micro-meso porous structure of the matrix.

3.0 Cs2.5H0.5PW12O40 + Vulcan XC-72 carbon after corrosion of Pt flag

j / mA cm-2

2.5 2.0

Vulcan XC-72 carbon after corrosion of Pt flag

1.5 1.0 0.5 0.0 0.00

0.02

0.04

0.06

0.08

0.10

0.12

E / V vs. RHE Fig. 45. Comparison of hydrogen oxidation RDE voltammograms at Nafion-coated catalytic layers made from: (red line) Cs2.5H0.5PW12O40 with Vulcan XC-72 Carbon and corroded platinum and (black line) Vulcan XC-72 Carbon with corroded platinum in 0.5 M H2SO4 under H2 atmosphere. Rotation rate, ω = 2500 rpm; scan rate, 2 mV s-1.

The dependence of the respective RDE limiting currents versus the square root of rotation rates (Fig. 46) shows linearity behavior. Assuming laminar flow, the mass transport rate and the limiting diffusional current density can be described mathematically by the Levich equation, as a function of the rotational frequency of the RDE, ω (in radians per seconds) (Equation 61).

85

2.5

j / mA cm

-2

2.0

1.5

1.0

+ O 40 12 W P5

Cs 2

lca Vu

a on arb c 72 CX n

-72 XC n a lc Vu

H 0.

.5

0.5

lag tf P f no sio o r or rc fte

ca

c er aft n rbo

orr

ag t fl fP o on osi

0.0 0

10

20 1/2

ω

30

40

50

1/2

/ (rpm)

Fig. 46. Levich plots j vs. ω1/2 prepared using the data of Fig.44 for (red symbols) Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded platinum. Currents were measured at 0.1 V. For comparison the same plot is provided for (black symbols) unmodified GCsupported Vulcan XC-72 carbon with corroded platinum.

The Bc0 calculated from the slope of Levich plot (Fig. 46) is equal 5.1 · 10-2 (mA cm-2) -3 rpm-1/2 and is about 20% lower than the theoretical value of Bc0 in 0.5 mol dm H2SO4,

at 25 0C (6.54 · 10-2 (mA cm-2) rpm-1/2). In order to overcome limitation of the technique employed, steady-state experiments were also carried out. They are complementary to the RDE results. Steady state data were used to prepare Tafel plots for the hydrogen oxidation reaction occurring at the investigated electrode material and for Nafion-containing Cs2.5PW12-free system as a reference. The Tafel diagrams for Nafion-containing catalytic layer made from Cs2.5PW12 matrix with Vulcan XC-72 carbon and corroded platinum are presented in [

,

Fig. 47. As a comparison, the same graphs were made for unmodified system. 44,45 50,51] Like it was described previously, on polycrystalline platinum two electron HOR in acidic media can proceed by chemical adsorption step (Tafel) or electrochemical adsorption step (Heyrovsky), followed by the adsorbed hydrogen atom discharge step (Volmer), as described in Eqs. (5) – (7).

86

log[(jL-j)/jL]

(c)

(a)

0.0

-0.8

-1.6

(d)

(b) log[(jLx j)/(jL - j)]

0.8

0.0

-0.8

0.00

0.02

0.04

0.00

0.02

0.04

E / V vs. RHE

Fig. 47. Mass transport corrected Tafel diagrams [E ~ log(jL – j / jL) and E ~ log(jL · j / jL - j)] for HOR on catalytic layer prepared from Nafion-containing Cs2.5H0.5PW12O40 with Vulcan XC72 carbon and corroded platinum, (a) equation (64) and (b) equation (65) and as a comparison catalytic layer made from Nafion-containing Vulcan XC-72 carbon with corroded platinum, (c) equation (64) and (d) equation (65) in 0.5 M H2SO4. Rotation rates: (■) 1500 rpm, (ο) 2500 rpm. Data taken from Fig. 44.

The values of the Tafel slopes for reversible reaction, presented in Table 10, are very close to theoretical value obtained for platinum 29 mV dec-1 at temperature 295 K. Based on the experimental data reversible nature of the electrochemical hydrogen oxidation reaction at both investigated catalytic layers is electrochemically reversible seems experimentally supported. The HOR on investigated systems takes places via the Tafel-Volmer mechanism with the atom-atom recombination step (Tafel) as a rate determining step (rds).[44] Exchange currents (I0) was calculated from the slope of the linear polarization response and corrected for diffusion by using previously showed and described equation (66). Values obtained from this equation are presented in Table 10. In conclusion, we should admit that the results obtained for catalytic layers prepared by electrochemical method studied using CV and RDE method, clearly shows that modification of the system by using insoluble cesium salt of 12-phosphotungstic acid matrix provide to better performance towards hydrogen oxidation reaction (HOR). 87

The exchange current (I0) calculated for Cs2.5PW12-containig material is almost twice of the value obtained for unmodified one. breversible (mV dec-1)

Composition of catalytic layer Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded Pt Vulcan XC-72 carbon with corroded Pt

birreversible (mV dec-1)

Theoretical breversible (mV dec-1)

Theoretical birreversible (mV dec-1)

I0 (m A) 0.98

31

28

31

28

29 (TafelVolmer) 118 (HeyrovskyVolmer)

58 (Tafel-Volmer, HeyrovskyVolmer) 0.5

Table 10. Theoretical and experimental Tafel parameters, mechanism and corresponding values of exchange currents for HOR in acid medium at 22 0C for the Nafion-containing catalytic layer made from: Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon

and corroded Pt and for

Cs2.5H0.5PW12O40 – free system (as a comparison).

7.4 Comparison of catalytic layers containing Cs2.5PW12 matrix prepared by mixing and electrochemical methods Table 11 presents summary of results obtained for Cs2.5H0.5PW12O40-containing systems prepared by mixing method and by electrochemical method.

Composition of catalytic layer Pt10% on Vulcan XC72 carbon

breversible (mV dec-1) 32

Theoretical breversible (mV dec-1)

I0 (mA) 0.63

29 (TafelVolmer)

Cs2.5H0.5PW12O40 with Pt10% on Vulcan XC72 carbon

32

Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded Pt

31

118 (HeyrovskyVolmer)

0.83

0.98

Table 11. Summary table containing mechanistic parameters and corresponding values of exchange currents for HOR in acid medium at 22 0C for the Nafion-containing catalytic layer made from: Cs2.5H0.5PW12O40 with 10% Pt/C, Cs2.5H0.5PW12O40 with Vulcan XC-72 carbon and corroded Pt and 10% Pt/C (for comparison).

88

The mechanism of the hydrogen oxidation reaction is the same for all investigated electrode materials. The value of exchange currents (I0) calculated for the system containing Cs2.5PW12-matrix is higher than that obtained for Cs2.5PW12-free catalyst. Furthermore, the catalyst prepared by electrochemical deposition appears to be more active, towards hydrogen oxidation than that prepared by mixing method. Catalytic layer, prepared by mixing method, containing Cs2.5H0.5PW12O40 matrix was tested as an anode in single working PEM fuel cell operating on pure hydrogen as a fuel. Advantage of this method is its simplicity. Results obtained from this measurements confirm that the modification of the commercial available Pt/C by using Cs2.5H0.5PW12O40 matrix provide to better performance of the system towards HOR what is in agreement with RDE measurements. It is important to underscore that results described in this chapter being the part of the Italian and international patent.[203,204]

89

8. Methanol oxidation reaction (MOR) on the catalytic layers containing Keggin-type heteropolyacid salts as a matrix Methanol oxidation reaction is a very important reaction from a practical point of view. This small organic molecule has attracted considerable attention due to the development of direct liquid fuel cells that require highly reactive fuels with high energy density. However, the formation of strongly adsorbed intermediates species such as (CO)ads on the Pt catalyst, which is the most active metal for methanol oxidation, results in high oxidation overpotentials, usually far from the thermodynamic limit. Therefore this chapter is devoted to optimize of Pt-based electrocatalysts, by applying new type of matrices with high porosity, strong acidic properties and high ionic conductivity. New systems have been characterized with respect to their electrochemical properties ( voltammetry, CO stripping) and their electrocatalytic activity for methanol oxidation. The stability of catalytic properties will be also discussed. The chapter is divided into two parts related to Keggin-type heteropolyacid salts (Cs+ or Rb+)

the cation present into the

used as a matrix for commercial

Pt/Vulcan XC-72 carbon. Methanol oxidation reaction on the unmodified catalytic layer (Nafion-containing Pt/Vulcan XC-72 carbon) is shown for a comparison.

8.1 Preparation of the catalytic layers The electrocatalyst layer was prepared by mixing via the following procedure. Calculated amounts of commercial Pt40%/Vulcan XC-72 carbon and of cesium or rubidium salt of Keggin-type heteropolyacid matrix were mixed in 1:1 ethanol - water mixture (proportion 1:1) and

stirred for 24 h. A measured volume of Nafion (5%

alcoholic solution) was then added and the stirring continued for additional 12 h to obtain a homogenous dispersion. The mass ratio between Pt40%/Vulcan XC-72 : heteropolyacids salts : pure Nafion solution was 1:1.5:0.3, respectively. The resulting ink was dropped with a micropipette the surface of a glassy carbon disk electrode ( 0.071 cm2 surface area) . The thin film

catalytic layer was dried at room temperature

for 20 minutes. For comparison, heteropolyacid salts-free ink of Pt40%/Vulcan XC-72 carbon and Nafion was also prepared. The platinum loading in each system is equal to 100 µg cm-2. 90

8.2 Electrochemical measurements at the Pt40%/C modified with Cs2.5-HPAs matrix

8.2.1 Cyclic voltammetry (CV) study

Fig. 48 shows the cyclic voltammetric responses (recorded in argon-saturated 0.5 mol dm-3 H2SO4) of the glassy carbon electrode modified with Nafion-treated pure Pt40%/Vulcan XC-72 carbon (Pt40%/C) and Pt40%/Vulcan XC-72 carbon modified with matrix made from cesium salts of Keggin-type heteropolyacids: Cs2.5H0.5PW12O40, Cs2.5H0.5PMo12O40, Cs2.5H1.5SiW12O40 and Cs2.5H1.5SiMo12O40. The platinum loading in all investigated electrocatalytic films was equal to 100 µg cm-2. The voltammetric peaks appearing at about 0.46 V (Fig. 48, light-blue line) and 0.43 V (Fig. 48, green line) should be attributed to the reduction of Cs2.5H1.5SiMo12O40 and Cs2.5H0.5PMo12O40, respectively, while Cs2.5H1.5SiW12O40 (Fig. 48, dark-blue line) is electroactive at potential lower ca. 0.1 V. Cs2.5H0.5PW12O40 becomes electroactive (Fig. 48, red line) at the more negative potentials and its reduction overlaps with the so-called hydrogen adsorption peaks that typically exist on clean bare platinum at potentials lower than 0.35 V. The presence of cesium salts of Keggin-type heteropolyacids (particularly Cs2.5PMo12 and Cs2.5SiMo12) causes a shift of the voltammetric peaks (at potential higher than 0.8 V) relative to the formation of Pt-oxo (PtO or PtOH) species towards more positive potentials.

91

0.6

j / mA cm

-2

0.3

0.0

-0.3

-0.6

-0.9 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 48. Cyclic voltammetric responses of Nafion-containing films (deposited on glassy carbon electrode) of (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4. Scan rate, 10 mV s-1. Temperature, 240C.

To investigate the influence of the matrix in the catalytic

behavior towards

methanol oxidation reaction (MOR), cyclic voltammetric measurements (Fig. 49) were carried out on four kinds of systems (Pt40%/C with Cs2.5PW12, Pt40%/C with Cs2.5PMo12, Pt40%/C with Cs2.5SiW12 and Pt40%/C with Cs2.5SiMo12) and on commercial unmodified electrocatalyst (Pt40%/C) as a reference. Because the slow kinetics, methanol oxidation reaction occurs at high oxidation overpotentials, far from the thermodynamic limit (E0 = 0.02 V).[84] In Fig. 49, for all investigated materials we observe two characteristic irreversible current peaks (A and B) during the electrooxidation of methanol. The peak obtained in the forward scan (peak A) at around 0.85 V is typically attributed to the methanol electrooxidation and the backward peak (peak B) at ca. 0.73 V is known to be due to the oxidation reaction on the Pt of residual intermediate species such as CH2OH, CH2O,HCOOH and CO.[205,206]

92

35

B

30

A

j / mA cm

-2

25 20 15 10 5 0 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 49. Cyclic voltammetric responses of Nafion-containing films (deposited on glassy carbon electrode) of (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Scan rate, 10 mV s-1. Temperature, 240C.

Moreover, Fig. 49 clearly show that modification of commercial Pt40%/C by cesium HPA salts matrix increases noticeably the electrocatalytic activity of the system for MOR. The highest current densities were obtained for the catalytic layers containing Cs2.5H0.5PMo12O40 and Cs2.5H1.5SiW12O40 as a matrix. It is also important to note that methanol oxidation reaction starts at more negative potentials for all modified systems in comparison to unmodified catalytic layer containing commercial Pt40%/C. This indicates that the presence of the matrix can increase the kinetics of methanol oxidation by lowering the onset potential for this reaction. The reaction mechanism was investigated on the basis of the Tafel plot analyses. Fig. 50 shows the Tafel plots for the methanol oxidation obtained from the cyclic voltammograms at the scan rate 5 mV s-1 for the systems modified by cesium salts of Keggin-type heteropolyacids and, as a comparison, for the unmodified catalytic layer.

93

-1

106 mV dec

1

-1

95 mV dec-1 95 mV dec

log j / mA cm-2

region 1 0

-1

93 mV dec

88 mV dec-1

-1

region 2 -2

-3 0,3

0,4

0,5

0,6

0,7

E / V vs. RHE Fig.50.Tafel plots of methanol oxidation at the Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Temperature, 240C.

The reverse of the slopes (RT/nF) of the Tafel lines (1 / b, expressed in mV dec1

) change at ca. 0.45 V, indicating the presence of two different reaction mechanisms.

The Tafel slope in region 1 is strongly dependent on the concentration of methanol as well as the reaction time[85] and this does not allow to draw conclusions on the kinetic equation in region 1. Since the Tafel slope in region 2 is unchanged with the concentration of methanol and reaction time a kinetic parameters at a given reaction time can be determined from this region.[85] The main slope, in the region 2, for unmodified 40% Pt/C is 88 mV decade-1. The

Tafel

slopes

for

the

catalytic

layers

containing

Cs2.5H0.5PW12O40,

Cs2.5H0.5PMo12O40, Cs2.5H1.5SiW12O40 and Cs2.5H1.5SiMo12O40 matrix were estimated, in the same manner as for the pure Pt40%/Vulcan XC-72 carbon system, to be 93, 95, 106 and 95 mV decade-1, respectively. Those results are in good agreement with those reported by Inada and co-workers.[85] They obtained Tafel slope of the methanol oxidation at the Pt electrode to be 96 ± 10 mV decade-1 in the potential range of 0.55 – 0.7 V vs. RHE at room temperature[85], which well agrees with the result of Fig. 50. Thus, it is postulated that the number of electrons involved in the electrode reaction n is 94

ca. 1 (n ≈ 1) at 240C. The methanol electrooxidation pathway is considered as follows[85]: CH3OH ↔ (CHnO)ads + (4 – n) H+ + (4 – n) e-

(70)

H2O ↔ (OH)ads + H+ + e-

(71)

(CHnO)ads + (OH)ads → CO2 + n (H)ads

(72)

(H)ads → H+ + e-

(73)

where methanol is oxidized via the strongly adsorbed species (CHnO)ads. If (CHnO)ads is (COH)ads, step (72) is written as (COH)ads + (OH)ads → (CO)ads + H2O

(74)

Reaction described in equation (70) (methanol dissociative adsorption) occurs around 0.1-0.3 V vs. RHE at Pt[84,85] and it is not include to the rate-determining process (Tafel slop is observed at 0.45 – 0.66 V). Therefore, the rate determining step is step showed in Eq. (71) or (72); although Eq. (71) is reported to be rate determining process only by several workers.[83,84,207]

8.2.2. Staircase voltammetry (SV) measurements

For a better insight of the system reactivity towards methanol oxidation reaction staircase voltammetry has been used. Fig. 51 shows

dependencies of staircase

voltammetric responses (step period of 50 s recorded every 25 mV) of methanol oxidation on catalytic layers containing Nafion treated Pt40%/Vulcan XC-72 carbon and cesium salts of Keggin-type heteropolyacids (Cs2.5PW12, Cs2.5PMo12, Cs2.5SiW12, Cs2.5SiMo12) as a matrix and on commercial unmodified electrocatalyst (Pt40%/Vulcan XC-72 carbon) for a comparison. For all materials containing Keggin-type cesium salts as a matrix a significant increase of electrocatalytic currents was observed. This can be rationalized in terms of the relative ability, potential mutual interactions, and the existence of sufficient numbers of Pt centers for efficient oxidation of methanol. Presence of Cs2.5H0.5PMo12O40 and Cs2.5H1.5SiW12O40 salts in the system results in some increase of methanol electrocatalytic currents (compare curves for unmodified and modified systems in Fig. 51).

95

25

j / mA cm

-2

20

15

10

5

0 0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 51. Staircase voltammetric current densities for the methanol (0.5 mol dm-3) oxidation recorded every 25 mV (between 0.25 and 1.07 V) following application of 50-s potential steps at the

Nafion-containing

layer

of

(-●-)Pt40%/Vulcan

XC-72

carbon

modified

with

Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4..Temperature, 240C.

Fig. 52 presents dependencies of staircase voltammetric responses (step period of 50 s recorded every 25 mV) of methanol oxidation on the same layers as in Fig. 51 but in a shorter potential range (from 0.43 to 0.65 V). An important issue of the data of Fig. 52 is that the methanol oxidation currents densities tend to appear at less positive potentials than on bare Pt. The best performances are shown by a catalytic layer containing Cs2.5H0.5PMo12O40 salt as a matrix. The methanol oxidation reaction starts at potential c.a. 20 mV less positive than for the unmodified system containing Nafiontreated Pt40%/Vulcan XC-72 carbon. This enhancement effect and also the increased current densities, may originate from the fact that PMo12 (in Cs2.5PMo12), added to platinum, increase the CO tolerance of Pt-based system.

96

8

j / mA cm

-2

6

4

2

0 0.45

0.50

0.55

0.60

0.65

E / V vs. RHE Fig. 52. Staircase voltammetric current densities for the methanol (0.5 mol dm-3) oxidation recorded every 25 mV (between 0.43 and 0.65 V) following application of 50-s potential steps at the

Nafion-containing

layer

of

(—)Pt40%/Vulcan

XC-72

carbon

modified

with

Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.. Temperature, 240C.

8.2.3 Chronoamperometry (CA) measurements

To further evaluate the reactivity of our electrocatalytic systems towards the methanol oxidation, current-time measurements at different constant potentials (0.47 or 0.52 V) were performed (Fig. 53). As expected, the largest currents densities are observed at the more positive applied potential (Fig. 53B). As observed in Fig. 53, when investigated materials are polarized at constant potential (0.47 V or 0.52 V) in methanol solutions, the current density decays continuously indicating a pronounced loss in activity. The current density reaches almost stationary state after 800 seconds (Fig. 53A) and 600 seconds (Fig. 53B). The factor causing the decay of current density is apparently, a blockage of the surface by some organic residue, which is slowly formed and can only be oxidized at high anodic potentials.[84] The impregnation of cesium salts

97

of Keggin-type heteropolyacids matrix with commercial Pt40%/Vulcan XC-72 carbon results in an increase of the methanol electrooxidation current densities.

A

j / mA cm-2

0.3

0.2

0.1

0.0 0

200

400

600

800

1000

t/s

B j / mA cm-2

0.6

0.4

0.2

0.0 0

200

400

600

800

1000

t/s Fig. 53. Chronoamperometric curves recorded for the methanol oxidation at the Nafioncontaining layer of (—)Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon upon application of (A) 0.47 V and (B) 0.52 V. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.. Temperature, 240C.

98

At both studied constant potentials, catalytic layers containing Cs2.5H0.5PW12O40 and Cs2.5H1.5SiMo12O40 shows only small difference in current densities in comparison to unmodified platinum (at 0.47V current decreased to the same level with that of pure Pt). The highest activity in both potentials (0.47 V and 0.52 V) is displayed by the system containing Cs2.5H0.5PMo12O40 salts as a matrix. We have to state that we not have produced a practically more active catalyst here. The system’s activity here is due to the higher number of the Pt active site which the methanol molecules reached to CO2 per second, per surface site, but not to the mass of platinum used.

8.2.4 Electrochemical impedance spectroscopy for methanol electrooxidation In order to further compare the activity of methanol electrooxidation on a different electrochemically polarized catalysts and to investigate the reaction mechanism, electrochemical impedance spectroscopy (EIS) was carried out at different potentials. Fig. 54 shows Nyquist plots of methanol electrooxidation for different electrochemically polarized catalysts modified by cesium salts of Keggin-type heteropolyacid matrix (Pt40%/C-Cs2.5HPA) and for unmodified Pt40%/Vulcan XC-72 carbon catalyst as a reference, at potentials of 0.50, 0.65 and 0.75 V. The EIS results indicate that the methanol electrooxidation on Nafion-containing Pt40%/Vulcan XC-72 carbon with cesium salts of HPAs (Cs2.5PW12, Cs2.5PMo12, Cs2.5SiW12, Cs2.5SiMo12) matrix at various potentials shows different impedance behaviors. In order to analyze reaction mechanism of methanol electrooxidation on Pt/C-Cs2.5HPAs catalysts, a simple two-step model for methanol electrooxidation can be assumed[209]: I1 CH 3OH → COads + 4 H + + 4e −

(75)

I2 COads + H 2O → CO2 + 2 H + + 2e −

(76)

I1 is the rate leading to the adsorbed surface intermediates COads. I2 is the rate for oxidation of COads. IF is the Faradaic current and stand for net rate of charge transfer. In this assuming, we just consider only one intermediate COads in methanol electrooxidation.

99

I 1 = k1cm (1 − θ CO ) I 2 = k 2θ COθ OH ,

(77)

I F = I1 + I 2

(78)

where θCO and θOH are the fractional surface coverage of COads and OH, respectively. For simplifying the analysis, the variation of θOH is assumed to have little effect on impedance behavior of methanol electrooxidation, defined as Θ=

dθ F = ( I1 − I 2 ) = K ( I1 − I 2 ) dt qCO

(79)

where qco is the quantity of charge needed to complete a full coverage of CO on the electrocatalyst.[209] According to the kinetic theory derived by Harrington and Conway[218] and Cao[219] for reactions involving intermediate adsorbate, in the electrode process of methanol electrooxidation, the Faradaic current depends on the electrode potential E and one other state variable θCO varying with E and affecting the Faradaic current.[209] So the Faradaic admittance of methanol electrooxidation is[209] YF =

1 B + Rct a + jω

(80)

where Rct = (∂E/∂IF)SS is charge transfer resistance of the electrode reaction and is the only circuit element that has a simple physical meaning describing how fast the rate of charge transfer during methanol electrooxidation changes with changing electrode potential when the surface coverage of the intermediate is held constant. The subscript “ss” denotes steady state. Rct also can be defined as Rct = limω→0 Re{Zf} where Re{Zf} is the real component in complex plots (Nyquist plots, Fig. 54), ω the circular frequency

ω = 2πf. So Rct can be obtained directly from Nyquist plots.[209] In Eq. 80, a = −(∂Θ/∂θ)SS > 0 a is the always positive for a stable steady process and with a dimension of s−1, defined as B = mb where m = ∂IF/∂θ, b = (∂Θ/∂E)SS = dθ/dE, and B with dimensions of Ω−1 cm−2 s−1. The impedance Nyquist plots can be classified according to the sign of B value. When B > 0, Eq. 80 can be rewritten as follows[209]: YF =

1 + Rct

1 1 1 = + a + jω Rct R0 + jωL B B

(81)

100

The dimension of R0 is Ω cm2 and of L is cm2. Thus there will be an inductive component involved in Faradaic impedance. The impedance Nyquist plot will be a capacitive arc in the high frequency range and an inductive arc in the low frequency range. When B < 0, in this case: YF =

B 1 − Rct a + jω

(82)

The Faradaic impedance can be rewritten as

Rct2 B Ra 1 ZF = = Rct + = Rct + YF a − Rct B + jω 1 + jωRa C a

(83)

with Ra = Rct 2|B|/a − Rct|B|, Ca = 1/Rct2|B|. The dimension of Ra is Ω cm2 and of Ca, F cm−2. In this group there are still two cases, which can be classified.[209] When a − Rct|B| > 0, then Ra > 0, two capacitive arcs will be displayed on the first quadrant of Nyquist plot. When a − Rct|B| < 0, then in this case Ra < 0, and the capacitive arc in low frequency range will enter into the second quadrant of Nyquist plot.[209] According to above analysis, for methanol electrooxidation on Pt/C-Cs2.5HPAs catalysts, the impedance parameters from Eqs. 77 and 78 can be deduced and are as follows[209]:

∂[K ( I 1 − I 2 )] αF  ∂Θ  [k1cm − (k1cm + k 2θ OH )θ CO ] b= =K  = ∂E RT  ∂E  SS

(84)

∂( I1 − I 2 )  ∂I  m= F  = = k 2θ OH − k1cm ∂θ  ∂θ  SS

(85)

1 When methanol electrooxidation at low potential range (0.5 V), assumes reaction (1), methanol dehydrogenation is rate-determining step, then k1 « k2. So, when k2θOH > k1cm, According to Eqs. 84 and 85, thus b < 0, and m > 0, namely B < 0. From Eq. 83, Nyquist plots of EIS at 0.4V (Fig. 54A) should show capacitive behaviors. Moreover, in this case, a − Rct|B| > 0, so two overlapped capacitive semicircles should exist in Nyquist plot and signify a reaction with one adsorbed intermediate.[209]

101

6000

A

-Zimag / ohm

5000 4000

Fig. 54. Nyquist plots recorded in Ar saturated

3000

0.5 mol dm-3 CH3OH + 2000

0.5 mol dm-3 H2SO4

1000

aqueous solution for Nafion-containing

0 0

1000

2000

3000

4000

(-●-) Pt40%/Vulcan XC-72

5000

carbon modified with

Zreal / ohm

-Zimag / ohm

500

Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72

B

400

carbon modified with

300

Cs2.5H0.5PMo12O40,

200

(-▼-) Pt40%/Vulcan XC-72

100

carbon modified with Cs2.5H1.5SiW12O40,

0 -100

(-♦-) Pt40%/Vulcan XC-72

-200

carbon modified with

-300

Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan

-400 0

200

400

600

800

1000

Zreal / ohm

different potentials: (a) 0.50 V; (b) 0.65 V; (c) 0.75 V.

300 200

XC-72 carbon catalysts at

C

The frequency range is from 0.05Hz to 100kHz.

-Zimag / ohm

100

Each electrode contained 100 µg 0

cm-2 of the platinum.

-100 -200 -300 -400 -500

-400

-300

-200

-100

0

100

Zreal / ohm

102

The analyses from EIS indicate that, at low potential region , reaction (I1) might be ratedetermining step, and means the oxidation of COads with the OHads is fast.[208,209] 2. When methanol electrooxidation at intermediate potential range (0.65 V), with an increase of potential, the rate of reaction (75) is increasing, but is not enough to exceed the rate of reaction (76) obviously. So in this case the rate-determining step of methanol electrooxidation is in transition region[209,210]. Thus maybe has following relationship[209]: k2 θOH θCO + k1 cm θCO > k1 cm > k2 θOH From Eqs. 84 and 85, b < 0, and m < 0 can be deduced and then B = mb > 0. According to Eq. 81 an inductive arc in low frequency range will be exhibited in Nyquist plot. The inductive behavior from theoretical analysis is also observed in our experiments shown in Fig. 54B. In general, the condition of occurrence of an inductive behavior in Nyquist plot is mb > 0. This means, if the variation of the electrode potential causes a variation of the Faradaic current density not only through its effect on the strength of the electric field in the double layer but also through its effect on another variable X and the both effects act in the same direction, an inductive component will be involved in the Faradaic impedance.[219] According to impedance parameters, b and m, inductive behavior in methanol electrooxidation reveals the COads coverage decreases with increasing potential (b < 0), and the decreasing COads lead to an increase of Faradaic current (m < 0). A reasonable explanation is with potential increasing, the large amounts of OHads are formed on Cs2.5HPAs (particularly tungsten containing HPAs) sites and react with COads and decrease its coverage. Meanwhile, the decreasing surface coverage of COads will contribute to adsorption of methanol on Pt and enhance the Faradaic current. So at the intermediate potential range, with an increase of potential from 0.5 to 0.65 V, the transition from capacitive behavior to inductive behavior indicates that the rate-determining step maybe is changing.[209,217] The similar inductive behavior in EIS at about 0.65V was also observed by Wang and co-workers.[217] 3. When methanol electrooxidation at high potential range (0.75 V), in this case, reaction (76) can be assumed as rate-determining step.[209,217] Thus, k1 » k2 and then k2θOH < k1cm. According to Eqs. 84 and 85, b > 0 and m < 0 can be deduced, then B < 0. Moreover, in this case a − Rct |B| < 0. The capacitive arc at low frequency in Nyquist plot will flips to the second quadrant with the real component of the impedance

103

becoming negative. The theoretical analyses agree well with the experimental results shown in Fig. 54C. This means that resistance Ra becomes negative resulting from passivation of electrode surface.[221] Melnick et al[220] indicated that the passivation of the Pt electrode during methanol electrooxidation is probably due to the reversible formation of oxide species. Meanwhile, due to reaction (76) is rate-determining step, the oxidation of COads with OH is very slow, so the passivation at higher potentials can be explained by the formation of a large amount of COads and OH on surface of Pt/CCs2.5HPAs catalyst. Therefore, adsorption of methanol on Pt sites is inhibited due to an increase of coverage of COads and OH on Pt sites and the electrooxidation rate is almost no obvious increase.[209] From the EIS analyses, reaction (76) as rate–determining step at high potential range can well explain experimental results.[209,217] Impedance electrochemically

spectroscopy polarized

of

the

Pt/C-Cs2.5HPAs

electrooxidation catalysts

of

indicates

methanol that

on

cathodic

polarization leads to an enhancement for methanol oxidation. The reaction proceeds with one adsorbed intermediate of appreciable surface concentration and lifetime. The different impedance behaviors in three different potential regions reveal that the mechanism and rate-determining step in methanol electrooxidation vary with potentials. At low potential range (0.50 V), methanol dehydrogenation is rate-determining step while at high potential range (0.75 V), the oxidation and removal of COads became ratedetermining step. Meanwhile, at intermediate potential region(0.65 V), the ratedetermining step in methanol electrooxidation is maybe in transition range.[209,217] In spite of the different impedance behaviors at different potentials, the diameters of the primary semicircle of the Nafion-containing Pt40%/C with Cs2.5PMo12 and Pt40%/C with Cs2.5SiW12 catalysts are always smaller that of the Nafion-containing Pt40%/C with Cs2.5PW12, Pt40%/C with Cs2.5SiMo12 and unmodified Pt40%/C systems. This behavior means that the lowers charge transfer resistances for the electrode reaction can be obtained on the Pt40%/Vulcan XC-72 carbon modified by Cs2.5H0.5PMo12O40 or Cs2.5H1.5SiW12O40 matrix.

104

8.2.5 CO stripping voltammetry study at Pt/C-Cs2.5HPAs Fig. 55 shows cyclic voltammograms obtained with catalytic layers modified by cesium salts of Keggin-type heteropolyacids (Cs2.5PW12, Cs2.5PMo12, Cs2.5SiW12 and Cs2.5SiMo12) and with unmodified system from pure Pt40%/Vulcan XC-72 carbon with a CO adsorbed ad-layer. An insert is put into the figure to emphasize the difference between the potential of CO oxidation in the first anodic cycle for all investigated systems. It is clear that the first anodic cycle is different from the second anodic cycle in that a peak between 0.632 and 0.89 V exist in the first anodic cycle in each voltammogram. The peak between 0.632 and 0.89 V is the oxidation of adsorbed CO. For all the layers CO-adsorbed on the Pt surfaces is oxidized already in the first anodic cycle, which is confirmed by the absence of peak in the second anodic cycle at potentials between 0.632 and 0.89 V. Moreover, for the CO oxidation peaks, important information such as onset and peak potential for CO oxidation is summarized in Table 12. Catalyst Pt/C Pt/C-Cs2.5PW12 Pt/C-Cs2.5PMo12 Pt/C-Cs2.5SiW12 Pt/C-Cs2.5SiMo12

Onset potential Peak potential (V) (V) 0.743 0.813 0.665 0.792 0.692 0.817 0.665 0.784 0.632 0.733

Table 12. Onset and peak potentials for oxidation of adsorbed CO with the four modified systems and for pure Pt/C as a reference.

Among the Pt/C, Pt/C-Cs2.5PW12, Pt/C-Cs2.5PMo12, Pt/C-Cs2.5SiW12 and Pt/CCs2.5SiMo12 catalysts, the Pt/C-Cs2.5SiMo12 system has the lowest onset and peak potentials for CO oxidation, while the Pt/C has the highest ones. CO starts to be oxidized with the Pt/C-Cs2.5SiMo12 electrode at a potential of 0.632 V which is 111 mV lower than with the Pt/C electrode. The addition of others cesium salts of Keggin-type heteropolyacids (Cs2.5PW12, Cs2.5PMo12, Cs2.5SiW12) also results in shifting of the onset potential of CO oxidation in to a lower values (Table 12). These results suggest that the HPA cesium salts behave as a redox mediator for the CO oxidation reaction with Pt present in the system. The proposed mechanism proceeds via the following reaction[27]: 105

HPA + CO + H2O → CO2 + H2HPA

(86)

H2HPA → HPA + 2H+ + 2e-

(87)

This mechanism is considered particularly for the molybdenum containing heteropolyacids, in which oxidation potential is shifted to the more positive values of potentials

in

comparison

to

the

tungsten

containing

heteropolyacids.[19,21]

5 j / mA cm

-2

4

j / mA cm

-2

4

2

3 0 0,6

2

0,7

0,8

E / V vs. RHE

1 0 -1 0,0

0,2

0,4

0,6

0,8

1,0

1,2

E / V vs. RHE

Fig. 55. Cyclic voltammograms in the potential range 0.025 to 1.125 V vs. RHE on the Nafioncontaining layer of (—)Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon with (first anodic cycle) and without (second anodic cycle) a CO adsorbed ad-layer. Electrolyte, argon saturated 0.5 mol dm-3 H2SO4. Scan rate, 20 mV s-1. Temperature 24 0C.

The catalytic enhancement of the layer with cesium salt of tungsten containing heteropolyacids (Cs2.5PW12, Cs2.5SiW12) as a matrix may be due to the synergistic effect between Pt and PW12 (or SiW12). Assuming that this enhancement effect may originate from the fact that tungsten containing heteropolycompound is analogous to the parent tungsten oxide[74,211], tungsten units may provide additional -OH groups or radicals

106

capable of facilitating oxidation of passivating intermediates (COads)[26,74,222] on Pt. This reaction can be described as follows[222]: CO + OH → CO2 + H+ + ePt

(88)

WO3

Alternatively, introduction of PW12 (or SiW12) may induce morphological differences and lead to better Pt catalyst utilization.[74]

8.2.6 Electrochemical stability of investigated materials containing Cs2.5-HPAs matrix Electrochemical stability, that is a very important factor in the application to practical fuel cells, was tested using long-term CV (Fig. 56) for methanol oxidation reaction under argon flow. Current densities where obtained from the forward peak of the last cycle (9th cycle) of methanol oxidation recorded in argon saturated 0.5 mol dm-3 H2SO4 containing 0.5 mol dm-3 CH3OH solution at 50 mV s-1. The gap between measurements was 20 min. As Fig. 56 shows, the electrochemical stability of the system containing Cs2.5H1.5SiW12O40 and Cs2.5H1.5SiMo12O40 as a matrix are fairly good, indicating that the catalytic layers are stable in these conditions. Moreover, also Cs2.5H0.5PW12O40 and Cs2.5H0.5PMo12O40 –containing systems are quite stable during long-term stability test. It is important to note that for all modified systems the current densities of MOR are higher than for unmodified catalytic material containing Nafion treated Pt40%/Vulcan XC-72 carbon. The reduction of currents densities during long-term stability test may result not only from accumulation of poisonous species (such as COads) on the surface of Pt particles but also because of methanol consumption during the successive scans and change of the surface structure of the Pt particles. Nevertheless, the best results (the highest current densities) during long-term

stability test

were

obtained

for

the

catalytic

layers

containing

Cs2.5H1.5SiW12O40 and Cs2.5H0.5PMo12O40 salts as a matrix.

107

30

jp / mA cm

-2

25

20

15

10

5

0 0

200

400

600

800

t / min Fig. 56. Long-term stability test of the Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon electrodes in argon saturated 0.5 mol dm-3 H2SO4 containing 0.5 mol dm-3 CH3OH solution at 50 mV s-1. Temperature, 240C. Same Pt loadings mounted in all cases (LPt = 100 µg cm-2).

8.3 Electrochemical measurements at system containing Rb2.5-HPAs matrix At the beginning of this subsection we have to mark that the interest of the rubidium salts of Keggin-type heteropolyacids were much lower than theirs cesium equivalents. Nevertheless they seem to be very attractive in point of view theirs future application as a matrix in anode materials for the alcohol fuel cells.

8.3.1. Cyclic voltammetry (CV) measurements Fig. 57 illustrates cyclic voltammetric responses of a glassy carbon electrode modified with Nafion-treated pure Pt40%/Vulcan XC-72 carbon (Pt40%/C) and Pt40%/Vulcan XC-72 carbon modified with a matrix made from rubidium salts of

108

Keggin-type heteropolyacids (Rb2.5H0.5PW12O40, Rb2.5H0.5PMo12O40, Rb2.5H1.5SiW12O40 and Rb2.5H1.5SiMo12O40) recorded in argon-saturated 0.5 mol dm-3 H2SO4 at room temperature. The platinum loading in all investigated materials was equal to 100µgcm-2.

0.6

j / mA cm

-2

0.3

0.0

-0.3

-0.6

-0.9 0.0

0.2

0.4

0.6

0.8

1.0

1.2

E / V vs. RHE Fig. 57. Cyclic voltammetric response of Nafion-containing films (deposited on glassy carbon electrode) of (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4. Scan rate, 10 mV s-1. Temperature, 240C.

The most positive potential of anodic peak (0.46 V, Fig. 57 light-blue) is due to the reduction of Rb2.5H1.5SiMo12O40 salt. The voltammetric peak appearing at about 0.43 V (Fig. 57, green line) should be attributed to the reduction of Rb2.5H0.5PMo12O40 salt, while Rb2.5H1.5SiW12O40 (Fig. 57, dark-blue line) and Rb2.5H0.5PW12O40 (Fig. 57, red line) salts becomes electroactive at a less positive potential of ca. 0.35 V. The presence of rubidium salts of Keggin-type heteropolyacids (particularly Rb2.5PMo12) leads to shift of the voltammetric peaks referring to the formation of Pt-oxo (PtO or PtOH) species toward more positive potentials. Cyclic voltammetric responses of the

109

systems containing rubidium salts of HPAs matrix are very similar to these obtained for their cesium equivalents. To investigate the influence of the matrix in the catalytic layer on the behavior towards methanol oxidation reaction (MOR) cyclic voltammetric measurements (Fig. 58) were carried out with four kinds of systems containing Nafion treated Pt40%/Vulcan XC-72 carbon modified by rubidium salts of Keggin-type HPAs (Rb2.5PW12, Rb2.5PMo12, Rb2.5SiW12 and Rb2.5SiMo12) matrix and a commercial unmodified electrocatalyst (Pt40%/C) as a reference. In the Fig. 58 for all investigated materials we observe two characteristic irreversible peaks (A and B) during the methanol electrooxidation. Forward scan peak A, at potential ca. 0.85 V, is the oxidation current density that involves the formation of intermediates.[205,206] As it was mention previously, it is certain that (CO)ads is a poison to platinum catalyst for further oxidation of methanol. The backward peak B at ca. 0.73 V, correspond to the oxidation of the intermediates created during forward cycle.

B

j / mA cm

-2

40

A

30

20

10

0 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 58. Cyclic voltammetric response of Nafion-containing films (deposited on glassy carbon electrode) of (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Scan rate, 10 mV s-1. Temperature, 240C.

110

Moreover, Fig. 58 clearly shows that modification of commercial Pt40%/C by rubidium HPA salts matrix increases noticeably electrocatalytic activity of the system for MOR, what can be attributed to the better Pt catalyst utilization. The highest current density was obtained for the catalytic layer containing Rb2.5H1.5SiW12O40 as a matrix. Nevertheless the current densities obtained for materials containing Rb2.5H0.5PMo12O40 and Rb2.5H1.5SiMo12O40 are only 7% lower from this obtained from the layer containing Rb2.5SiW12. Furthermore the onset potential of methanol oxidation on all modified catalytic layers is shifted to less positive values of potential in comparison to unmodified catalytic layer containing commercial Pt40%/C. This indicates that addition of the new type of matrix can facilitate oxidation of methanol and produce passivating intermediates during this process. The reaction mechanism was investigated on the basis of the Tafel plot analyses. The Tafel plots for the methanol oxidation obtained by cyclic voltammograms at the scan rate 5 mV s-1 at the systems modified by rubidium salts of Keggin-type heteropolyacids and for the unmodified catalytic layer made from Pt40%/Vulcan XC-72 carbon are showed in Fig. 59. The main slope, in the region 2 (1/b, expressed in mV dec.-1), for unmodified Pt40%/C is 88 mV decade-1. The other Tafel slopes for the catalytic layers containing Rb2.5H0.5PW12O40, Rb2.5H0.5PMo12O40, Rb2.5H1.5SiW12O40 and Rb2.5H1.5SiMo12O40 matrix were estimated, in the same manner as for the pure Pt40%/Vulcan XC-72 carbon systems, to be 106, 102, 106 and 113 mV decade-1, respectively. These results are in good agreement with those reported by Inada and co-workers.[85] Small deviations are observed for the system containing Nafion treated Pt40%/C with Rb2.5H1.5SiMo12O40 matrix. For the slope in the shown range, it is postulated that the number of electrons involved in the electrode reaction n is c.a. 1 (n ≈ 1) at 240C[82] and that the rate determining process is the step in which CO from Pt surfaces is oxidized by the –OH groups to CO2.[85] In spite of this Eq. (71) showed previously is reported to be rate determining process only by several workers.[83,84,207]

111

-1

-1 106 mV dec 113 mV dec -1 102 mV dec -1 106 mV dec

log j / mA cm-2

1

0

region 1

-1

88 mV dec

region 2

-1

-2

-3 0,3

0,4

0,5

0,6

0,7

E / V vs. RHE Fig. 59.Tafel plots of methanol oxidation at the Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Temperature, 240C.

8.3.2 Staircase voltammetry (SV) measurements Reactivity of the Pt/C modified with Rb2.5-HPAs towards methanol oxidation was also examined by using staircase voltammetry method. Fig. 60 present dependencies of staircase voltammetric responses (step period of 50 s recorded every 25 mV) of methanol oxidation on catalytic layers containing Nafion treated Pt40%/Vulcan XC-72 carbon and rubidium salts of Keggin-type heteropolyacids (Rb2.5PW12, Rb2.5PMo12,

Rb2.5SiW12,

Rb2.5SiMo12)

matrix

and

unmodified

electrocatalyst

(Pt40%/Vulcan XC-72 carbon) as a reference. For all materials containing Keggin-type rubidium salts matrix a significant increase of electrocatalytic currents was observed. This can be rationalized in terms of the relative ability, potential mutual interactions, and the existence of sufficient numbers of Pt centers for efficient oxidation of methanol. The highest current densities for

112

methanol electrooxidation were obtained for catalytic layers containing rubidium salts of 12-silicotungstic and 12-silicomolybdic acid as a matrix.

25

j / mA cm

-2

20

15

10

5

0 0.4

0.6

0.8

1.0

E / V vs. RHE Fig. 60. Staircase voltammetric current densities for the methanol (0.5 mol dm-3) oxidation recorded every 25 mV (between 0.25 and 1.07 V) following application of 50-s potential steps at the

Nafion-containing

layer

of

(-●-)Pt40%/Vulcan

XC-72

carbon

modified

with

Rb2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4..Temperature, 240C.

Fig. 61 present dependencies of staircase voltammetric responses of methanol oxidation on the same layers as in Fig. 60 but in a shorter potential range (from 0.43 to 0.65 V). An important issue of the data of Fig.60 is that the methanol oxidation currents densities tend to appear at less positive potentials than on bare Pt, the same as for Cs+ salts

of

HPAs.

The

best

performance

shown

catalytic

layer

containing

Rb2.5H1.5SiW12O40 salt as a matrix. The onset potential of methanol oxidation is ca. 25 mV less positive than for the unmodified system containing Nafion-treated Pt40%/Vulcan XC-72 carbon. However, it is important to underline that the onset potentials for the catalytic layers modified by the residue rubidium salts of 12phosphotungstic acid, 12-phosphomolybdic acid and 12-silicomolybdic acid are also shifted to the less positive values of potential. This enhancement effect, as well as the 113

increasing of current densities, may originate from the fact that rubidium salts of Keggin-type heteropolyacids matrix added to platinum increase the CO tolerance of Ptbased system.

j / mA cm

-2

8

6

4

2

0 0.45

0.50

0.55

0.60

0.65

E / V vs. RHE Fig. 61. Staircase voltammetric current densities for the methanol (0.5 mol dm-3) oxidation recorded every 25 mV (between 0.43 and 0.65 V) following application of 50-s potential steps at the

Nafion-containing

layer

of

(—)Pt40%/Vulcan

XC-72

carbon

modified

with

Rb2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40,

(—) Pt40%/Vulcan XC-72

carbon modified with Rb2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.. Temperature, 240C.

8.3.3 Chronoamperometry (CA) measurements To further evaluate the reactivity of our electrocatalytic systems for the methanol oxidation, current-time measurements at different constant potentials (0.47 or 0.52 V) were performed (Fig. 62). As observed in Fig. 62, when investigated materials are polarized at constant potential (0.47 V or 0.52 V) in methanol solutions, the current density decays continuously indicating a pronounced loss in activity. The current density reaches almost stationary state after 600 seconds (Fig. 62A) and 400 seconds (Fig. 62B). As it was mentioned before the factor causing the decay of current density is

114

apparently, a blockage of the surface by some organic residue, which is slowly formed during methanol oxidation process.[84] In Fig. 62B, for the Rb2.5SiW12 containing catalytic layer the initial current decay is greater than for the others catalysts. This behavior can be addressed to the greater extent of the surface coverage with partially oxidized intermediates, than for other layers.

A

j / mA cm-2

0.3

0.2

0.1

0.0 0

200

400

600

800

1000

t/s 1.0

B j / mA cm-2

0.8

0.6

0.4

0.2

0.0 0

200

400

600

800

1000

t/s Fig. 62. Chronoamperometric curves recorded for the methanol oxidation at the Nafioncontaining layer of (—)Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon upon application of (A) 0.47 V and (B) 0.52 V. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.. Temperature, 240C.

115

All catalytic layers modified with rubidium salts of Keggin-type heteropolyacids matrix exhibit higher methanol electrooxidation current densities (at constant potential 0.47 or 0.52 V) during all experiment time in comparison to unmodified system (Pt40%/C). The highest activities in both constant potentials are displayed by systems containing Rb2.5H0.5PMo12O40, Rb2.5H1.5SiW12O40 and Rb2.5H1.5SiMo12O40 salts as a matrix. This can be explained by the increasing Pt active surface area, which can be reached by methanol molecules and increment the CO tolerance of Pt-based system.

8.3.4 Electrochemical impedance spectroscopy measurements

In order to further compare the activity of the catalysts containing rubidium salts of HPA`s matrix for methanol oxidation, electrochemical impedance spectroscopy (EIS) were carried out at different potentials in 0.5 mol dm-3 H2SO4 and CH3OH aqueous solutions. The Nyquist plots for the corresponding Pt40%/C-Rb2.5HPAs systems and for unmodified one, as a reference, recorded at 0.50, 0.65 and 0.75 V are showed in Fig. 63. It is clearly seen that EIS of the methanol electrooxidation on investigated electrode materials at various potentials shows different impedance behavior. The polarization resistance can be measured from diameter of the primary semicircle. As it was in the case of Cs2.5-HPAs modified catalytic layers, at low potentials range of methanol electrooxidation Pt/C-Rb2.5HPA electrode interface is dominated by adsorption and electrical double layer at low potential range.[208] The Nyquist plots of electrochemical impedance spectroscopy should show capacitive behaviors, which could be seen in Fig. 63A. At intermediate potentials range (Fig. 63B) the rate-determining step of MOR is in transition region. The inductive behaviors observed in our experiments shown in Fig. 63B are in agreement with the literature data.[209,210] When EIS are performed at high potential range (Fig. 63C), the oxidation and removal of COads become rate-determining step. The capacitive arc at low frequency in Nyquist plot reverses to the second and third quadrants, with the real component of the impedance becoming negative[209], which is shown in Fig. 63C.

116

A

4000

-Zimag / ohm

3000

Fig. 63. Nyquist plots recorded in Ar saturated

2000

0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4

1000

aqueous solution for 0

Nafion-containing 0

1000

2000

3000

4000

5000

(-●-) Pt40%/Vulcan XC-72

Zreal / ohm

carbon modified with Rb2.5H0.5PW12O40,

400

(-▲-) Pt40%/Vulcan XC-72

B

-Zimag / ohm

carbon modified with Rb2.5H0.5PMo12O40,

200

(-▼-) Pt40%/Vulcan XC-72 carbon modified with 0

Rb2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 -200

carbon modified with Rb2.5H1.5SiMo12O40 and

-400 0

200

400

600

800

1000

Zreal / ohm

(-■-) unmodified Pt40%/Vulcan XC-72 carbon catalysts at different potentials: (a) 0.50 V;

200

(b) 0.65 V; (c) 0.75 V.

C

The frequency range is from

-Zimag / ohm

100

0.05Hz to 100kHz. Each electrode contained 100 µg

0

cm-2 of the platinum. -100

-200

-300

-400 -400

-300

-200

-100

0

Zreal / ohm

117

The electrochemical impedance spectroscopy of methanol electrooxidation presented in Fig. 63 shows that the mechanism and rate-determining step vary with potentials. At low, intermediate and high potentials range the rate-determining steps of methanol electrooxidation are methanol dehydrogenation, transition range and oxidation and removal of COads, respectively. Moreover the diameters of the primary semicircle of the Nafion-containing Pt40%/C with Rb2.5SiW12 and Pt40%/C with Rb2.5SiMo12 catalysts are always smaller than that of the Nafion-containing Pt40%/C with Rb2.5PW12, Pt40%/C with Rb2.5PMo12 and the unmodified Pt40%/C systems in all electrochemical impedance spectroscopy (EIS) measurements, performed at different potentials This behavior can be explained by the lower charge transfer resistance of the electrode reaction on the Pt40%/Vulcan XC-72 carbon modified by Rb2.5SiW12 and Rb2.5SiMo12 matrix.

8.3.5 CO stripping voltammetry study at Pt/C modified with Rb2.5-HPAs The cyclic voltammograms obtained on Nafion containing catalytic layers modified by rubidium salts of Keggin-type heteropolyacids (Rb2.5PW12, Rb2.5PMo12, Rb2.5SiW12 and Rb2.5SiMo12) and on unmodified system (Pt40%/Vulcan XC-72 carbon), as a comparison, with a CO adsorbed ad-layer are shown in Fig. 64. An insert is put into the figure to emphasize the difference between the potential of CO oxidation in the first anodic cycle for all investigated systems. For all the layers, CO-adsorbed on the Pt surfaces, is oxidized already in the first anodic cycles, as confirmed by the absence of peak in the second anodic cycles at potentials between 0.63 and 1 V. Moreover, for the CO oxidation peaks in the first anodic cycles onset and peak potentials for CO oxidation are summarized in Table 13.

118

5 4

j / mA cm

-2

4

j / mA cm

-2

3 2

2

0 0.6

0.8

E / V vs. RHE

1 0 -1 -2 0.0

0.2

0.4

0.6

0.8

1.0

1.2

E / V vs. RHE Fig. 64. Cyclic voltammograms in the potential range 0.025 to 1.125 V vs. RHE on the Nafioncontaining layer of (—)Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon with (first anodic cycle) and without (second anodic cycle) a CO adsorbed ad-layer. Electrolyte, argon saturated 0.5 mol dm-3 H2SO4. Scan rate, 20 mV s-1. Temperature 24 0C.

Catalyst

Onset potential Peak potential (V)

(V)

Pt/C

0.743

0.813

Pt/C-Rb2.5PW12

0.679

0.831

Pt/C-Rb2.5PMo12

0.656

0.790

Pt/C-Rb2.5SiW12

0.656

0.780

Pt/C-Rb2.5SiMo12

0.632

0.750

Table 13. Onset and peak potentials for oxidation of adsorbed CO with the four modified systems and for pure Pt/C as a reference.

From the Table 13 we can easily calculate that the catalytic enhancement toward CO oxidation is on the order of 64 – 110 mV in comparison to the unmodified Pt40%/Vulcan XC-72 carbon electrode. The lowest onset potential of oxidation of CO119

adsorbed on the Pt surfaces was recorded for the system containing Rb2.5SiMo12 matrix. This shift to less positive values of potential is in well agreement with the onset potential obtained for catalytic layer containing Cs2.5SiMo12 salt as a matrix. These results suggest that the salts containing SiMo12 heteropolycompound can provide the best protection of the Pt surface against poisoning by CO, moreover they can help oxidize strongly adsorbed CO, e.g. by weakening the Pt-CO bond. For the others catalytic layers modified by rubidium salts of Keggin-type heteropolyacids matrix (Rb2.5PW12, Rb2.5PMo12, Rb2.5SiW12) we also obtain lower value of the onset potential for CO oxidation (Table 13). This shift of potential on the modified catalyst could be attributed to the presence of oxygenated species on Rb2.5-HPAs at lower potentials compared to platinum. According to the bifunctional mechanism[212], these oxygenating species allow the oxidation of CO into CO2 at lower potentials. These results, obtained in this subchapter, suggest that the presence of POMs anions (attend in the rubidium salts of Keggin-type heteropolyacids) might facilitate the electrooxidation of intermediate species such as the CO that adsorbed on the Pt catalyst surfaces[27], leading to suppression of the poisoning effect on Pt catalysts by CO or COlike intermediates. The catalytic enhancement of the layer containing Rb2.5SiW12 matrix may be due to the synergistic effect between Pt and SiW12. This behavior can be explained by assuming that tungsten units may provide additional -OH groups or radicals capable of facilitating oxidation of passivating intermediates (COads) on Pt (like in the parent WO3).[74,211] Tungsten containing salts may also induce morphological differences and lead to better Pt catalyst utilization.

8.3.6 Electrochemical stability of investigated materials containing Rb2.5-HPAs matrix The long-term stability of Nafion treated Pt40%/Vulcan XC-72 carbon modified by Rb2.5-HPAs matrix systems in 0.5 mol dm-3 H2SO4 containing 0.5 mol dm-3 CH3OH aqueous solution has been investigated by cyclic voltammetry and the corresponding results are shown in Fig. 65. Long-term stability test for unmodified system (Pt40%/C) is presented as a reference.

120

30

jp / mA cm

-2

25

20

15

10

5

0

0

200

400

600

800

t / min

Fig. 65. Long-term stability test of the Nafion-containing (-●-)Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Rb2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon electrodes in argon saturated 0.5 mol dm-3 H2SO4 containing 0.5 mol dm-3 CH3OH solution at 50 mV s-1. Temperature, 240C. Same Pt loadings mounted in all cases (LPt = 100 µg cm-2).

The current densities were read from the forward anodic peak of the last cycle (9th cycle). The gap between measurements was 20 min. The most stable performances during long-term test are shown by catalytic layers containing Rb2.5H0.5PW12O40 and Rb2.5H0.5PMo12O40, in which drop of the current densities was equal 23% and 27% respectively. However, the best results (the highest current densities) during long-term stability tests were obtained for the catalytic layers containing Rb2.5H1.5SiW12O40 and Rb2.5H1.5SiMo12O40 salts as a matrix. We should also mention that, during all stability tests, the current densities of methanol electrooxidation on modified systems were higher that that for Nafion treated Pt40%/Vulcan XC-72 carbon electrode. The reduction of currents densities during long-term stability tests may result not only from accumulation of poisonous specious (such as COads) on the surface of Pt particles but also because of methanol consumption during the successive scans and change of the surface structure of the Pt particles.

121

8.4. Summary and conclusions 1) The values of peak current densities of methanol electrooxidation obtained from cyclic voltammetry measurements, as a function of electrode composition are presented in Fig. 66. Our result clearly shows that the modification of commercial Pt catalyst by Cs+ or Rb+ salts of HPAs increase electrocatalytic activity of the system towards methanol electrooxidation. Based on this results we can conclude that the presence of Cs2.5-HPAs or Rb2.5-HPAs matrix into the catalytic layer may facilitate the electrooxidation of intermediate species such as the CO that is adsorbed on the Pt catalyst surfaces leading to suppression of the poisoning effect on Pt catalysts by CO or CO-like intermediates.

jp / mA cm

-2

30

20

10

0 12

o Si M 2.

/C 0% Pt 4

40

%

/C

-R

-R

b

2.

b

5

5

5 2.

b

12

Si W

12

PM

o

12

16

-R /C 0% Pt 4

Pt

b -R /C

0%

12

2. 5

PW

/C 0%

12

Pt 4

2.

5

/C

-C

s

2. 5

0%

Pt 4

PW

12

o

8

PM

12

s

Pt 4

Pt 4

0%

/C

-C /C

0%

-C

s

5 2.

s -C

Pt 4

/C % 40 Pt

2. 5

Si W

12

4

Si M

o

0

Fig. 66. Peak current densities obtained from forward scan of methanol electrooxidation on the investigated materials recorded in argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Data taken from Fig.49 and 58.

2) The kinetic data of the methanol electrooxidation on investigated catalytic materials obtained from the Tafel plot analyses shows that, in all cases, the rate determining step is the step in which CO from Pt surfaces is oxidized by the –OH groups to CO2.

122

Catalyst

Tafel slope b / mV dec

Catalyst

Tafel slope

-1

b / mV dec-1

Pt/C

88

Pt/C + Rb2.5PW12

106

Pt/C + Cs2.5PW12

93

Pt/C + Rb2.5PMo12

102

Pt/C + Cs2.5PMo12

95

Pt/C + Rb2.5SiW12

106

Pt/C + Cs2.5SiW12

106

Pt/C + Rb2.5SiMo12

113

Pt/C + Cs2.5SiMo12

95

Table 14. Tafel slope values for all investigated catalytic materials.

3) The results received from staircase voltammetry (Fig. 67), which were used to better insight into the systems reactivity, clearly shows that the onset potential of the methanol oxidation reaction is lower for almost all modified materials. Only in the case of Cs2.5PW12 and Cs2.5SiMo12 modified layers onset potential is not change in comparison to the Pt/C electrode. These results are very important in a point of view of theirs practical application as an anode materials for the fuel cells.

Eonset / V vs. RHE

0,56

0,54

0,52

0,50 12

M

o

12

W

Si

Si -R

2.

b

5

5 2.

-R

/C

/C Pt

40

%

% 40 Pt

% 40 Pt

12

b

5 2.

b -R /C

/C %

16

PM

o

12

PW 5

-R

b

2.

% 40 40 Pt

% 40 Pt

12

/C

12

Pt

5

s -C /C

2.

PM

PW

12

o

8

5

s -C %

40 Pt

% 40 Pt

2.

Si 5

/C

-C

2.

s

Si 5

/C

s

2.

-C /C % 40 Pt

12

W

12

4

M

o

0

Fig. 67. Onset potential of methanol electrooxidation on the investigated materials recorded in argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Data taken from Fig. 52 and 61.

Based on the results obtained by Samant et al.[7] we can conclude that the superior oxidation kinetics in presence of zeolites matrix could be due to the preferential formation of CO clusters on platinum that are limited by the steric constraints imposed

123

by the zeolites framework, followed by facile oxidation to CO2 by interaction with the surface or bridged hydroxyls of the zeolites species.

4) Chronoamperometry tests on catalytic layers (Fig. 68) confirm earlier obtained results that modification of commercial Pt/C with ours zeolite matrix increase catalytic activity of the system towards methanol oxidation. The best performance after 1000 seconds of polarization at constant potentials (0.47 and 0.52 V) is obtained with the system modified with Rb2.5SiW12 matrix. Nevertheless, the layers containing Cs2.5PMo12, Cs2.5SiW12, Rb2.5PW12, Rb2.5PmO12 and Rb2.5SiMo12 salts also exhibit higher current densities than the unmodified Pt/C electrode at both studied constant potentials.

0 ,0 6

A 0 ,0 4

j / mA cm

-2

0 ,0 2

0 ,0 0

B 0 ,3 0

0 ,1 5

o iM

12

12 2.

b

5

S

S 5 2.

-R /C

/C P

t4

0%

0% t4 P

16

iW

o M

b

P 5

-R

b

2.

-R /C 0% t4

P

12

12

W 5

b /C

0% t4 P

2.

-R

t4 P

12

P

/C

12

0%

P 5 2.

s

P 5

-C /C

0% t4 P

8

W

o

12

4

M

12

s

S 5 2.

/C

P

t4

0%

/C 0% t4

-C

s

S 5

-C

s

2.

-C /C P

0% t4 P

2.

iW

o iM

0

12

0 ,0 0

Fig. 68. Current densities of methanol electrooxidation as a function of electrode composition. The data were obtained from chronoamperometric curves (Fig. 53 and 62) after 1000 seconds of electrodes polarization at (A) 0.47 V and (B) 0.52 V in argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. T = 24 0C

124

5) The CO stripping voltammetry is a very important method, because CO is one of the main poisoning intermediate products during methanol electrooxidation. Therefore the studies on increasing resistance of the system towards this strongly adsorbed intermediate species were made. The results presented in Fig. 69, clearly shows that the onset potential of the CO oxidation is shifted to the lower value of potentials for the systems modified with cesium and rubidium zeolite matrix. Our results confirm that the system modification increase tolerance towards CO. The PMo12 or SiMo12 containing materials behave as a redox mediator for the CO oxidation reaction with Pt present in the system, while the PW12 and SiW12 materials may provide additional hydroxide groups which facilitate oxidation of passivating intermediates (COads) on Pt.

Eonset / V vs. RHE

0,75

0,70

0,65

0,60

12

o M b

-R

2. 5

Si

Si 2. 5

b -R

/C

/C Pt

40

%

% 40 Pt

16

12

W

12

o PM 2. 5

b -R

Pt

40 %

/C

/C % 40

Pt

12

12

PW 2. 5

-R

b

40 %

/C

12

Pt

PW s

-C /C %

40 Pt

2. 5

o

8

PM s

-C /C

Pt

40

%

/C % 40 Pt

2. 5

Si -C

s

2. 5

Si s

2. 5

-C /C % 40 Pt

12

W

12

4

M

o

0

12

0,55

Fig. 69. Onset potential of methanol electrooxidation on the investigated materials recorded in argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Data taken from Fig.55 and 64.

6) The results of the cyclic voltammetry long-term stability test on investigated electrode materials are placed in Fig. 70. The peak current densities obtained at the forward scan were read at 600 minutes of the test.

125

jp / mA cm

-2

20

15

10

5 12

o Si M 5

-R b

2.

Si W

/C

b

Pt 40 %

/C -R 40 % Pt

2. 5

5 2.

b /C -R

12

16

12

PM o

12

PW

12

5 2.

b -R

Pt 40 %

12

/C /C Pt 40 %

5 2.

s /C -C

Pt 40 %

PW

12

o

8

PM 5

Pt 40 %

-C s /C

2.

5 2.

Pt 40 %

-C

s

5 2.

s

Pt 40 % /C

/C -C 40 % Pt

12

Si W

12

4

Si M

o

0

Fig. 70. Peak current densities obtained at 600 minute of long-term stability test of methanol electrooxidation on the investigated materials recorded in argon saturated 0.5 mol dm-3 CH3OH + 0.5 mol dm-3 H2SO4. Data taken from Fig. 56 and 65.

For all the modified catalytic layers the values of the peak current densities of methanol electrooxidation were higher in comparison to the platinum electrode. The highest value of jp was obtained for the Rb2.5SiMo12 containing material. However, the most stable during long-term cyclic voltammetry stability test appears to be the catalytic layers modified with Rb2.5PW12 and Rb2.5PMo12 matrix.

126

9. Ethanol oxidation reaction (EOR) on the Pt40%/Vulcan XC-72 carbon modified with Cs2.5-HPAs Among the small organic molecules, ethanol, being a renewable biofuel, is a promising candidate for a fuel cell. Compared with methanol, ethanol has many merits: less toxic, better stability, lower permeability across proton exchange membrane. Ethanol is safer than other hydrocarbons, and the energy density of ethanol is higher than that of methanol (1325.31 and 702.32 kJ mol-1 for ethanol and methanol, respectively). Nevertheless, the process of the oxidation of ethanol is more difficult than that of methanol with is necessity to break the C-C bond to obtain its complete oxidation. However, the formation of strongly adsorb intermediates species such as (CO)ads on the Pt catalyst, which is usually considered as the best single metal to adsorb organic molecules and break intermolecular bond, reflect in high oxidation overpotentials. Therefore the aim of the present chapter is to investigate the influence of the modification of Pt nanoparticles for the electrooxidation of ethanol. Studies of the electrooxidation of ethanol in acidic media on the new systems prepared by mixing method were carried out using different electrochemical techniques such as cyclic voltammetry, staircase voltammetry, chronoamperometry, A.C. Impedance and Tafel plots. The stability of catalytic properties will be also discussed. Cyclic voltammograms of the systems modified by Keggin-type heteropolyacid salts as a matrix recorded in argon saturated 0.5 mol dm-3 H2SO4 as well as CO stripping of these layers will be not presented because of the great similarity to the voltammograms obtained previously in chapter 8.2.1 and 8.2.5, respectively.

9.1 Cyclic voltammetry (CV) measurements In order to compare the electrocatalytic capabilities of all four electrodes studied (and Pt40%/C electrode as a reference), all the steady cyclic voltammograms of the electrooxidation of ethanol in argon saturated 0.5 mol dm-3 C2H5OH in 0.5 mol dm-3 H2SO4 aqueous solution are presented in Fig. 71. In the forward scan, ethanol oxidation produced a prominent symmetric peak (A) around 0.89 V. In the reverse scan, an anodic peak current density (B) is detected at around 0.79 V. This peak is due to the oxidation

127

of all adsorbed carbonaceous species (e.g. Pt-OCH2CH3, Pt-CHOH-CH3, (Pt)2=COHCH3, Pt-COCH3 and Pt-C≡O).

A 25

B

j / mA cm

-2

20

15

10

5

0 0.0

0.2

0.4

0.6

0.8

1.0

E / V vs. RHE

Fig. 71. Cyclic voltammetric response of Nafion-containing films (deposited on glassy carbon electrode) of (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 C2H5OH + 0.5 mol dm-3 H2SO4. Scan rate, 10 mV s-1. Temperature, 240C.

It is noticeable in Fig. 71 that the Nafion-treated Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40 matrix shows the highest positive scan peak current density for ethanol electro-oxidation among the catalysts investigated here. It is quite apparent from the figure that backward anodic current density peak (jb) is greater for an electrode where forward anodic current density peak (jf) is grater. Moreover, the ratio of the forward anodic peak current density (jf) to the reverse anodic peak current density (jb), jf / jb, can be used to describe the catalyst tolerance to carbonaceous species accumulation. These results as well as peak currents densities and corresponding peak potentials of the forward and backward scan are listed in Table 15. Low jf / jb ratio indicates poor oxidation of ethanol to carbon dioxide during the anodic scan, and extensive accumulation of carbonaceous residues on the catalyst surface. High jf / jb ratio shows the converse case.

128

Forward peak Catalysts

potential /

jf / mA cm

Backward peak -2

jb /

potential / V vs. mA cm

V vs. RHE

jf/jb -2

RHE

Pt40%/C

0.894

16.60

0.796

14.77

1.12

Pt40%/C-Cs2.5H0.5PW12O40

0.897

21.62

0.800

18.44

1.17

Pt40%/C-Cs2.5H0.5PMo12O40

0.897

25.43

0.804

22.09

1.15

Pt40%/C-Cs2.5H1.5SiW12O40

0.898

20.30

0.802

19.14

1.06

Pt40%/C-Cs2.5H1.5SiMo12O40

0.898

19.31

0.798

17.00

1.14

Table 15. CV results of the investigated catalysts (24 0C).

It

is

clearly

shown

that

the

layers

containing

Cs2.5H0.5PW12O40,

Cs2.5H0.5PMo12O40 and Cs2.5H1.5SiMo12O40 salts as a matrix have higher ability to oxidize ethanol to CO2 (particularly Pt-(CO)ads to CO2 ) in comparison to the platinum electrode (higher value of jf / jb).

1.0 -1

log j / mA cm

-2

162 mV dec

0.5

0.0

-0.5

-1.0 0.4

0.5

0.6

0.7

E / V vs. RHE Fig. 72. Tafel plots of ethanol oxidation at the Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Temperature, 240C.

129

Moreover, Fig. 71 clearly show that modification of commercial Pt40%/C by cesium HPA salts matrix increases noticeably electrocatalytic activity of the system for EOR (higher current densities were obtained for all modified catalytic layers). This enhanced activity could be attributed to the increased surface area of the catalytic sites created by high dispersion of Pt. A very convenient way of comparing the performance of different electrode materials for an ethanol electro-oxidation process is the use of steady-state polarization curves and the corresponding Tafel plots. This innovative procedure allows a clear and pictorial view of two important parameters, namely the starting potential and the current density value for the systems under investigation. Thus, Fig. 72 shows the Tafel plots carried out in the potentiostatic mode for ethanol oxidation process at the Nafioncontaining (red symbols) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (green symbols) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (dark blue) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (light blue) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (black symbols) unmodified Pt40%/Vulcan XC-72 carbon as a reference. For the oxidation of ethanol on the investigated Pt - Cs-HPAs composite electrodes and for the unmodified Pt/C system the b (slope) value is ca. 162 mV dec-1. However, the existing mechanism[124] is extremely complex and not allow for the establishment of the rite determining step (rds) for theoretical calculations of the Tafel slope value (b). It is clear that the current density related to the active surface area (at constant value of potential) increases with platinum dispersion by shifting the Tafel plot cathodically: e.g. at 0.5 V the current density for modified layers (particularly for the systems containing Cs2.5PW12 and Cs2.5SiW12 matrix) is much higher that this for pure system (Pt40%/C). Nevertheless, the same value of b shows that there is no change in the reaction mechanism, particularly the rate determining step. These results suggest a weaker poisoning for the electrodes modified by Cs2.5-HPAs matrix.

9.2 Staircase voltammetry (SV) measurements Fig. 73 present dependencies of staircase voltammetric responses (step period of 50 s recorded every 25 mV) of ethanol oxidation on catalytic layers containing Nafion treated Pt40%/C and cesium salts of Keggin-type heteropolyacids, particularly12phosphotungstic

acid

(Cs2.5H0.5PW12O40),

12-phosphomolybdic

acid 130

(Cs2.5H0.5PMo12O40), 12-silicotungstic acid (Cs2.5H1.5SiW12O40) and 12-silicomolybdic acid (Cs2.5H1.5SiMo12O40), respectively, as a matrix and commercial unmodified electrocatalyst

(Pt40%/Vulcan

XC-72

carbon)

as

a

reference.

Presence

of

Cs2.5H0.5PMo12O40 matrix in the system results in some increase of ethanol electrocatalytic currents (compare curves for unmodified and modified systems in Fig. 73). 12

j / mA cm

-2

10 8 6 4 2 0 0.4

0.6

0.8

1.0

E / V vs. RHE Fig.73. Staircase voltammetric current densities for the ethanol (0.5 mol dm-3) oxidation recorded every 25 mV (between 0.25 and 1.07 V) following application of 50-s potential steps at the Nafion-containing layer of (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4..Temperature, 240C.

It is important to note that electrocatalytic activities (shown as the current densities at corresponding peaks) for all modified electrode materials are higher than that of Cs2.5HPAs - free system. To express more clearly, the mass activity (MA, mA mg-1), defined by peak current density per unit of catalyst loading, was used to evaluate the electrocatalytic activity of the system for ethanol electrooxidation. The MA was calculated according to the following equation[213]: MA = jp/md x 103

(89)

131

where jp (mA cm-2) is the peak current density and the md (µg cm-2) is the loading mass of Pt. The obtained values are presented in Table 16.

jp /

LPt /

MA /

mA cm-2

mg cm-2

mA mg-1

Pt40%/C

8.770

100

87.70

Pt40%/C-Cs2.5H0.5PW12O40

10.13

100

101.3

Pt40%/C-Cs2.5H0.5PMo12O40

12.02

100

120.2

Pt40%/C-Cs2.5H1.5SiW12O40

10.73

100

107.3

Pt40%/C-Cs2.5H1.5SiMo12O40

9.570

100

95.70

Catalysts

Table 16. The mass activity of the investigated catalysts (24 0C).

The highest MA value was obtained for catalytic layer modified by Cs2.5H0.5PMo12O40 zeolite type matrix (Table 16). The mass activity value for this system is almost 40% higher in comparison to platinum electrode (Pt40%/C) while for the other investigated electrode materials the MA value was from 9 to 22 % higher than for unmodified system. This can be explained by the more exposed nanoparticles of Pt in the zeolite matrix (Cs2.5-HPAs), which result in increasing specific surface and also effect in higher electrocatalytic activity of Pt/C-zeolite (Cs2.5-HPAs) catalyst. Moreover, zeolite material with special porous structure provides relatively high permeability and good micromedia for ethanol oxidation.[7] We should to point out that for Pt/C - Cs2.5HPAs catalyst, Pt act as the main catalyst for catalyzing the dehydrogenation of ethanol during the oxidation reaction and the oxygen containing species can be provided by the framework oxygen sites or surface hydroxyls of the zeolite particles. What’s more these oxygen-containing species strongly react with CO-like intermediate species on Pt surface to release the active sites for further ethanol oxidation.[7]

9.3 Chronoamperometry (CA) measurements Chronoamperometric experiments were carried out to observe the stability and possible poisoning of the catalysts under short-time continuous operation. Fig. 74 shows the current-time curves recorded for the several Pt containing electrode systems in argon saturated 0.5 mol dm-3 C2H5OH + 0.5 mol dm-3 H2SO4 solutions at a fixed potential of 0.47 V versus RHE. The initial current density decay can be addressed to a 132

great extent to the increasing surface coverage with partially oxidized intermediates.[214] Meanwhile, that initial decay was much less pronounced in the case of ethanol oxidation for Pt40%/C-Cs2.5PW12 composite suggesting less poisoning of the electrode surface than for the other electrode materials.

0.6

j / mA cm-2

0.5

0.4

0.3

0.2 0

200

400

600

800

1000

E / V vs. RHE Fig. 74. Chronoamperometric curves recorded for the ethanol oxidation at the Nafioncontaining layer of (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (—) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (—) unmodified Pt40%/Vulcan XC-72 carbon upon application of (A) 0.47 V. Electrolyte: argon saturated 0.5 mol dm-3 H2SO4.. Temperature, 240C.

Most of the current densities present a quasi-stationary behavior after 400 seconds of polarization. This slow decay of the current densities is due to the poisoning of the material during the advance process.[215] However, it is important to note that the current density for the layer containing Cs2.5PMo12 zeolite matrix after initial decay start to increases with time what suggest that the presence of PMo12 into the system help to cleaning Pt surface from CO-like intermediate species.

133

0,30 0,29

j / mA cm

-2

0,28 0,27 0,26 0,25 0,24 0,23

40

40

O 12

o Si M H

2. 5

2. 5

s

Pt 40 %

/C

-C

s

-C /C Pt 40 %

1. 5

Si W 1. 5

H

10

12

O

40

12

5 0.

H 2. 5

Pt 40 %

/C

-C

s

-C /C Pt 40 %

8

PM

o

12

s

2. 5

H

0. 5

PW

/C Pt 40 %

6

O

4

O

2

40

0,22

Fig. 75. Current densities of ethanol electrooxidation as a function of electrode composition. The data were obtained from chronoamperometric curves after 1000 seconds of electrodes polarization at 0.47 V in argon saturated 0.5 mol dm-3 C2H5OH + 0.5 mol dm-3 H2SO4. Temperature, 24 0C.

Values of the current density for ethanol oxidation measured after 1000 second at 0.47 V are plotted against the electrode composition in Fig. 75. It can be observed that electrodes with Cs2.5H1.5SiW12O40 and Cs2.5H1.5SiMo12O40 matrix exhibit lower activity than unmodified system. Layer containing cesium salt of 12-phosphotungstic acid after 1000 seconds of oxidation of ethanol at 0.47 V shows the same current densities like unmodified Pt40%/C electrode. The highest current density for EOR was found for Pt40%/Vulcan XC-72 carbon containing Cs2.5PMo12 salt as a matrix, which is in good agreement with CV and SV results.

9.4. Electrochemical impedance spectroscopy for ethanol electrooxidation In order to understand the characteristics of the electrocatalysis reaction on modified electrodes, additional analysis of the ac impedance behaviour was carried out. The electrochemical impedance spectra of Pt modified electrodes recorded in argon saturated 0.5 mol dm-3 H2SO4 + 0.5 mol dm-3 C2H5OH at 0.75 V are shown in Fig. 76.

134

-Zimag / ohm

900

600

300

0

-300 0

500

1000

1500

2000

Zreal / ohm Fig. 76. Nyquist plots recorded in Ar saturated 0.5 mol dm-3 C2H5OH + 0.5 mol dm-3 H2SO4 aqueous solution for Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40,(-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon catalysts at 0.75 V. The frequency range is from 0.05Hz to 100kHz. Each electrode contained 100 µg cm-2 of the platinum.

The diameters of the semicircles (Fig. 76.) decreased in order Pt40%/C > Pt40%/C + Cs2.5SiMo12 > Pt40%/C + Cs2.5SiMo12 > Pt40%/C + Cs2.5SiMo12 > Pt40%/C + Cs2.5PMo12 > Pt40%/C + Cs2.5PW12. This behavior can be related to the charge transfer resistances, which decrease with increasing effective surface area for the charge transfer reaction. A significantly different x-axis intercept were observed for the Pt40%/C and Pt40%/C-Cs2.5SiMo12. This behavior is due to the higher film resistivity for these samples in comparison to the others investigated materials. The impedance data from Fig. 76 suggest inductive behavior of the spectrum, which has been well studied by several researchers.[208,216] They showed that an inductive loop appeared if the reactions step of an intermediate species is the rate determining. It is considered that CO forms on the electrocatalyst as a strongly adsorbed intermediate and that the electro-oxidation of COads to CO2 is a rate determining step.

135

9.5 Electrochemical stability of investigated materials containing Cs2.5-HPAs Electrochemical stability of the catalytic layers containing Cs2.5-HPAs zeolite matrix during long-term ethanol oxidation was shown in Fig. 77.

30

jp / mA cm

-2

25

20

15

10

5

0 0

200

400

600

800

t / min Fig. 77. Long-term stability test of the Nafion-containing (-●-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PW12O40, (-▲-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H0.5PMo12O40, (-▼-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiW12O40, (-♦-) Pt40%/Vulcan XC-72 carbon modified with Cs2.5H1.5SiMo12O40 and (-■-) unmodified Pt40%/Vulcan XC-72 carbon electrodes in argon saturated 0.5 mol dm-3 H2SO4 containing 0.5 mol dm-3 C2H5OH solution at 50 mV s-1. Temperature, 240C. Same Pt loadings mounted in all cases (LPt = 100 µg cm-2).

As a reference stability test for unmodified Nafion treated Pt40%/Vulcan XC-72 carbon electrode is also presented. It is evident that the worst performance shows system containing Cs2.5SiMo12 as a matrix. The final current density for this electrode material at the end of the test was 32% lower than the onset potential. The drop of the current densities during long-term stability test for ethanol electro-oxidation on the investigated layers is presented in Table 17.

136

Catalysts

Current density drop

Pt40%/C

22.3 %

Pt40%/C-Cs2.5H0.5PW12O40

13.5 %

Pt40%/C-Cs2.5H0.5PMo12O40

22.8 %

Pt40%/C-Cs2.5H1.5SiW12O40

29.7 %

Pt40%/C-Cs2.5H1.5SiMo12O40

32.4 %

Table. 17. Current density drops during long-term stability test.

The data obtained from the stability test indicate that the best performance towards ethanol oxidation in the long term shows catalytic layer modified by the Cs2.5PW12. However, the highest current densities values during all experiment were obtained for the Nafion-containing Pt40%/Vulcan XC-72 carbon modified with Cs2.5PMo12 matrix.

137

Conclusions

This thesis is focused on the development of stable, proton highly-conductive, meso-microporous matrices for Pt that can be used to increase the CO tolerance of PEMFC, DMFC, and DEFC anodes. In the first experimental part of this thesis (Chapter 6), the IR spectra clearly show that the primary Keggin structure remains unaltered even when the protons form parental heteropolyacid are substituted by the cesium or rubidium cations. This is important from the point of view of the systems CO tolerance that shall be attributed to the presents of tungstate and molybdenum units in the structure. The SEM results presented here clearly suggest that morphology of heteropolyacids changes dramatically when the protons in heteropolyacid are substituted by cesium or rubidium cations. This substitution supports the existence of small spherical particles. The cyclic voltammograms recorded for investigated materials deposited on the glassy carbon electrodes show that some of the heteropolyacid salts exhibit behavior similar to that characteristic of their analogues in solution. This phenomenon can be explained in terms of the partial solubility in solution, particularly ammonium salts. On the basis of the results obtained in Chapter 6, the salts of Keggin-type heteropolyacids containing 2.5 moles of cesium and rubidium cations in 1 mole of the heteropolyacid salt (such as Cs2.5H0.5PW12O40,

Rb2.5H0.5PW12O40,

Cs2.5H0.5PMo12O40,

Rb2.5H0.5PMo12O40,

Cs2.5H1.5SiW12O40, Rb2.5H1.5SiW12O40, Cs2.5H1.5SiMo12O40, Rb2.5H1.5SiMo12O40) can be applied as matrices for Pt catalysts. The next part of the thesis (Chapter 7) is devoted to the hydrogen oxidation reaction on the Pt/Vulcan catalysts modified with Cs2.5H0.5PW12O40 prepared by two methods (mixing and electrochemical deposition). We demonstrate that incorporation and activation of catalytic Pt centers in these conductive high-surface-area zeolite-type robust matrices is feasible. Immobilization of platinum nanoparticles within the micro-mesoporous hybrid material was achieved through electrochemical deposition of platinum by corrosion of Pt counter electrode. HRTEM investigation shown that the particles have spherical sizes and their diameters range 1 – 2 nm and 3 - 4 nm for Cs2.5PW12 modified and free system, respectively. The difference in particle size may be due to fact that Pt deposition, on the Cs2.5H0.5PW12O40 containing composite, occurs inside the nanochannels of the porous tertiary structure of

138

the salts that posses characteristic diameters, while in the case of pure Vulcan the particles grow on the surface where are more prone to grow. The electrochemically active area obtained from CO stripping voltammetry combined with the results from HRTEM allows estimating the loading of Pt, ca. 2.6 µg cm-1. Comparison was made to the electrocatalytic system produced by simple mixing of Vulcan XC-72 supported platinum (10 wt %) nanoparticles with the cesium heteropolytungstate salt. In both cases, regardless the method of preparation and the nature of immobilization of Pt sites, clear enhancement of the platinum activity were observed during hydrogen oxidation. The next chapter of the thesis (Chapter 8) deals with the applications of the cesium and rubidium salts of the Keggin-type heteropolyacids as matrices for Pt/Vulcan during methanol oxidation reaction. The results obtained with cyclic voltammetry and staircase voltammetry methods clearly show that the modification of commercial Pt/C with cesium and rubidium salts of HPAs significantly increases the electrocatalytic currents (compare to unmodified Pt/C), and shifts the onset potential of the methanol oxidation reaction to the less positive values. This suggests that the presence of Cs2.5HPAs or Rb2.5-HPAs matrix into the catalytic layer may facilitate the electrooxidation of intermediate species such as the CO that is adsorbed on the Pt catalyst surfaces leading to suppression of the poisoning effect on Pt catalysts by CO or CO-like intermediates. The results are also consistent with the data obtained from CO stripping voltammetry, where the modification of Pt catalyst by the Cs2.5-HPAs or Rb2.5-HPAs results in shifting onset potential of CO oxidation to the less positive values. An important issue is that, among all examined catalysts, the Pt/C-Cs2.5SiMo12 and Pt/CRb2.5SiMo12 shows the lowest onset and peak potentials for CO oxidation, while the Pt/C exhibits the highest parameters, respectively. Long-term stability test confirm earlier obtained results that modification of commercial Pt/C with zeolite matrix increase catalytic activity of the system towards methanol electro-oxidation. The highest value of the jp (peak current densities of methanol electrooxidation) was obtained for the Rb2.5SiMo12 containing catalytic layer. However, the most stable during long-term cyclic voltammetry stability test is the system modified with Rb2.5PW12 and Rb2.5PMo12 matrices. The results presented in chapter 9 clearly show that the Pt/C modified with Cs2.5HPAs catalysts studied display electrocatalytic activity with respect to ethanol oxidation 139

as evidence by the voltammetric and chronoamperometric measurements. The data obtained from the stability test indicate that the best performance towards ethanol oxidation (in the long term experiment) is exhibited by the catalytic layer modified by the Cs2.5PW12. However, the highest current densities values during all experiment were obtained for the Pt40%/Vulcan XC-72 carbon modified by Cs2.5PMo12 system. The results presented in the thesis suggest that cesium and rubidium heteropolyacids salts should be considered as good matrices for anodes in the low temperature fuel cells (PEMFC, DMFC, DEFC). It is reasonable to expect that the amounts of precious Pt (or Ru) can be somewhat diminished during practical applications. We should to admit that the all experiments presented in the thesis were reproducible within 5-7% in different measurements. The results presented in Chapter 7 of the thesis are the part of the Italian and international patent.[203,204]

140

Bibliography [1] C. Arbizzani, S. Beninati, E. Manferrari, F. Soavi, M. Mastragostino, J. Power Sources 161 (2006) 826. [2] S.S. Gupta, J. Datta, J.Electroanal. Chem. 594 (2006) 65. [3] M.S. Volkmar, I. Remo, P. Elena, G. Sergio, J. Pchys. Chem. 100 (1996) 17901. [4] S.Q. Song, W.J. Zhou, Z.H. Zhou, L.H. Jiang, G.Q. Sun, Q. Xin, V. Leontidis, S. Kontou, P. Tsiakaras, Int. J. Hydrogen Energy 30 (2005) 995. [5] W.J. Zhou, B. Zhou, W.Z. Li, Z.H. Zhou, S.Q. Song, G.Q. Sun, Q. Xin, S. Douvartzides,M. Goula, P. Tsiakaras, J. Power Sources 126 (2004) 16. [6] E. Passalacqua, F. Lufrano, G. Squadrito, A. Patti, L. Giorgi, J. New. Mat. Electrochem. Syst. 3 (2000) 131. [7] P. V. Samant, J., B. Fernandes, J. Power Sources 125 (2004) 172 [8] H. Pang, J. Chen, L. Yang, B. Liu, X. Zhong, X. Wei, J. Solid State Elechtrochem. 12 (2008) 237. [9] M. T. Pope, A. Muller, Angew. Chem. Int. Ed. Engl. 30 (1991) 34. [10] I. V. Kozhevnikov, Catal. Rev. – Sci. Eng. 37 (1995) 311. [11] C. L. Hill, C. M. Prosse-McCartha, Coord. Chem. Rev. 143 (1995) 407. [12] A. Corma, Chem. Rev. 95 (1995) 559. [13] T. Okuhara, N. Mizuno, M. Misono, M. Adv. Catal. 41 (1996) 113. [14] N. Mizuno, M. Misono, M. Chem. Rev. 98 (1998) 199. [15] M. Sadakane, E. Steckham, Chem. Rev. 98 (1998) 219. [16] C.L. Hill, Chem. Rev. 98 (1998) 1. [17] D.E. Katsoulis, Chem. Rev. 98 (1998) 359. [18] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171. [19] R. Wlodarczyk, M. Chojak, K. Miecznikowski, A. Kolary, P. J. Kulesza, R. Marassi, J. Power Sources 159 (2006) 802. [20] R. Wlodarczyk, A. Kolary-Zurowska, R. Marassi, M. Chojak, P. J. Kulesza, Electrochim. Acta 52 (2007) 3958. [21] M. Chojak, A. Kolary-Zurowska, R. Wlodarczyk, K. Miecznikowski, K. Karnicka, B. Palys, R. Marassi, P.J. Kulesza, Electrochim. Acta 52 (2007) 5574. [22] K. Karnicka, M. Chojak, K. Miecznikowski, M. Skunik, B. Baranowska, A. Kolary, A. Piranska, B. Palys, L. Adamczyk, P.J. Kulesza, Bioelectrochemistry 66 (2005) 79. [23] N. Giordano, P. Staiti, S. Hocevar, A.S. Arico, Electrochim. Acta 41 (1996) 397. [24] Y.S. Kim, F. Wang, M. Hickner, T.A. Zawodzinski, J.E. McGrath, J. Membr. Science 212 (2003) 263. [25] V. Ramani, H.R. Kunz, J.M. Fenton, J. Membr. Science 232 (2004) 31. [26] T. Moiyalagan, B. Viswanathan, J. Power Sources 175 (2008) 789. [27] R.J. Stanis, M.-C. Kuo, J.A. Turner, A.M. Herring, J. Electrochem. Soc. 155 (2008) B155. [28] J.R. Ferrell, M.-C. Kuo, J.A. Turner, A.M. Herring, Electrochim. Acta 53 (2008) 4927. [29] R.J. Stanis, M.-C. Kuo, J.A. Turner, A.M. Herring, Electrochim. Acta 53 (2008) 8277. [30] J. Shim. C.-R. Lee, H.-K. Lee, J.-S. Lee, E.J. Cairns, J. Power Sources, 102 (2001) 172. [31] S. Soled, S. Miseo, G. Mc Vicker, W.E. Gates, A. Gutierrez, J. Paes, Catal. Today 36 (1997) 441. [32] S. Soled, S. Miseo, G. McVicker, W.E. Gates, A. Gutierrez, J. Paes, Chem. Eng. J. 64 (1996) 247. [33] M. Misono, Mater. Chem. Phys. 17 (1987) 103. [34] G. Hoogers, Fuel cell technology handbook, 2003 [35] hydrogencommerc.com [36] K.R. Cooper, V. Ramani, J.M. Fenton, H.R. Kunz, Experimental Methods and data analyses for polymer electrolyte fuel cell, 2005. [37] www.fuelcell.no [38] J. Larminie, A. Dicks, Fuel Cell System Explained, 2003. [39] www.fuelcells.org [40] S. Shuldiner, J. Electrochem. Soc. 107 (1960) 452. [41] V. S. Bagotzky, N. V. Osetrova, Electroanal. Chem. 43 (1973) 233. [42] W. Vogel, J. Lundquist, P. Ross, P. Stonehart, Electrochim. Acta 20 (1975) 79. [43] F.T.A. Vork, L.J.J. Janssen, E. Barendecht, Electrochim. Acta 31 (1986) 1569. [44] R.M.Q. Mello, E.A. Ticianelli, Electrochim. Acta 42 (1997) 1031. [45] N.M. Marković, B.N. Grugur, P.N. Ross, J. Phys. Chem. B 101 (1997) 5405. [46] J. Maruyama, M. Inaba, K. Katakura, Z. Ogumi, Z. Takehara, J. Electroanal. Chem. 447 (1998) 201.

141

[47] M.J. Croissant, T. Napporn, J.-M. Léger, C. Lamy, Electrochem. Acta 43 (1998) 2447. [48] T.J. Schmidt, H.A. Gasteiger, G.D. Stäb, P.M. Urban, D.M. Kolb, R.J. Behm, J. Electrochem. Soc. 145 (1998) 2354. [49] J. Jiang, A. Kucernak, J. Electroanal. Chem. 567 (2004) 123. [50] P.M. Quaino, M.R.G. de Chialvo, A.C. Chialvo, Phys. Chem. Chem. Phys. 6 (2004) 4450. [51] A.F. Innocente, A.C.D. Ângelo, J. Power Sources 162 (2006) 151. [52] B.M. Babić, L.M. Vračar, V. Radmilović, N.V. Krstajić, Electrochim. Acta 51 (2006) 3820. [53] A. śurowski, M. Łukaszewski, A. Czerwiński, Electrochem. Acta 51 (2006) 3112. [54] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Applied Catalysis B: Environmental 56 (2005) 9. [55] M. Łukaszewski, A. śurowski, M. Grdeń, A. Czerwiński, Electrochem. Commun. 9 (2007) 671. [56] M. Łukaszewski, A. śurowski, A. Czerwiński, J. Power Sources 185 (2008) 1598. [57] A. śurowski, M. Łukaszewski, A. Czerwiński, Electrochimica Acta 53 (2008) 7812 [58] V. A. Paganin, E. A. Ticianelli, E.R. Gonzales, J. Applied Electrochem. 26 (1996) 297. [59] T. Frelink, W. Visscher, J.A.R. van Veen, J. Electroanal. Chem. 382 (1995) 65. [60] L.W. Niedrach, I.B. Weinstock, Electrochem. Technol. 3 (1965) 270. [61] J.H. Choi, K.-W. Park, B.-K. Kwon, Y.-E. Sung, J. Electrochem. Soc. 150 (2003) A973. [62] A.L. Ocampo, M. Miranda-Hernández, J. Morgado, J. A. Montoya, P.J. Sebastian, J. Power Sources 160 (2006) 915. [63] A. Hamnett, Catal. Today 38 (1997) 445. [64] S. Wasmus, A. Kuver, J. Electroanal. Chem. 461 (1999) 14. [65] D. Lee, S. Hwang, I. Lee, J. Power Sources 145 (2005) 147. [66] M. Mastragostino, A. Missiroli, F. Soavi, J. Electrochem. Soc. 151 (2004) A1919. [67] C. He, H.R. Kunz, J.M. Fenton, J. Electrochem. Soc. 144 (1997) 970. [68] A.S. Aricó, P. Cretí, P.L. Antonucci, J. Cho, H. Kim, V. Antonucci, Electrochim. Acta 43 (1998) 3719. [69] D. Chu, R. Jiang, Solid State Ionics 148 (2002) 591. [70] P.J. Kulesza, B. Grzybowska, M.A. Malik, M. Chojak, K. Miecznikowski, J. Electroanal. Chem. 512 (2001) 110. [71] C. Arbizzani, M. Biso, E. Manferrari, M. Mastragostino, J. Power Sources 178 (2008) 584. [72] C. Arbizzani, S. Beninati, F. Soavi, A. Varzi, M. Mastragostino, J. Power Sources 185 (2008) 615. [73] J.M. Macak, P.J. Barczuk, H. Tsuchiya, M.Z. Nowakowska, A. Ghicov, M. Chojak, S. Bauer, S. Virtanen, P.J. Kulesza, P. Schmuki, Electrochem. Communications 7 (2005) 1417. [74] M. Chojak, M. Mascetti, R. Wlodarczyk, R. Marassi, K. Karnicka, K. Miecznilowski, P.J. Kulesza, J. Solid State Electrochem. 8 (2004) 854. [75] F. Sen, G. Gökağaç, J. Phys. Chem. C 111 (2007) 5715. [76] L.C. Ordóñez, P. Roquero, P.J. Sebastian, J. Ramírez, Catalysis Today 107-108 (2005) 46. [77] A.S. Aricò, S. Srinivasan, V. Antonucci, Fuel Cells 1 (2001) 133. [78] J. Prabhuram, R. Manoharan, J. Power Sources 74 (1998) 54. [79] F. Viger, C. Coutanceau, F. Hahn, E.M. Belgsir, C. Lamy, J. Electroanal. Chem. 563 (2004) 81. [80] D. Yuan, S. Tan,Y. Liu, J. Zeng, F. Hu, X. Wang, P. Sen, Carbon 46 (2008) 531. [81] T. Schultz, S. Zhou, K. Sundmacher, Chem. Eng. Technol. 24 (2001) 12. [82] M. Umeda, H. Sugii, I. Uchida, J. Power Sources 179 (2008) 489. [83] M. Umeda, M. Kokubo, M. Mohamedi, I. Uchida, Electrochim. Acta 48 (2003)1367. [84] T. Iwastia, Electrochim. Acta 47 (2002) 3663. [85] R. Inada, K. Shimazu, H. Kita, J. Electroanal. Chem. 277 (1990) 315. [86] M. Shen, S. Roy, K. Scott, J. App. Electrocham. 35 (2005) 1103. [87] G. Girishkumar, T.D. Hall, K. Vinodgopal, P.V. Kamat, J. Phys. Chem. B 110 (2006) 107. [88] Y.T. Kim,T. Mitani, J. Catal. 238 (2006) 394. [89] M.H. Seo, S.M. Choi, H.J. Kim, J.H. Kim, B.K. Cho, W.B. Kim, J. Power Sources 179 (2008) 81. [90] Z. He, J. Chen, D. Liu, H. Tong, W. Deng, Y. Kuang, Materials Chem. Phys. 85 (2004) 396. [91] Z. Jusys, T.J. Schmidt, L. Dubau, K. Lasch, L. Jörissen, J. Garche, R.J. Behm, J. Power Sources 105 (2002) 297. [92] J.M. Léger, S. Rousseau, C. Coutanceau, F. Hahn, C. Lamy,Electrochim. Acta 50 (2005) 5118. [93] W.S. Li, L.P. Tian, Q.M. Huang, H. Li, H.Y. Chen, X.P. Lian, J. Power Sources 104 (2002) 281. [94] P. Waszczuk, J. Solla-Gullón, H.S. Kim, Y.Y. Tong, V. Montiel, A. Aldaz, A. Wieckowski, J. Catal. 203 (2001) 1. [95] Y. Morimoto, E.B. Yeager, J. Electroanal. Chem. 444 (1998) 95.

142

[96] C. Lamy, J.-M. Léger, S. Srinivasan, Direct methanol fuel cells-from a 20th century electrochemists dream to a 21st century emerging tehnology, in: J.O.M. Bockris, B.E. Conway (Eds.), Modern Aspects of Electrochemistry, Vol. 34, Plenum Press, New York, 2000 (Chapter 3), pp.53-117. [97] C. Lamy, A. Lima, V. LeRhun, F. Delime, C. Coutanceau, J.-M. Léger, J. Power Sources 105 (2002) 283. [98] V.S. Silva, J. Schirmer, R. Reissner, B. Ruffmann, H. Silva, A. Mendes, L. M. Madeira, S. P. Nunes, J. Power Sources 140 (2005) 41. [99] J.O`M. Bockris, S. Srinivasan, Fuel Cells: Their Electrochemistry, McGraw – Hill Cook Company, New York, 1969. [100] E. Herro, W. Chrzanowski, A.Wieckowski, J. Phys. Chem. 99 (1995) 10423. [101] E. Antolini, J.R.C. Salgado, E.R. Gonzalez, Appl. Catal. B 63 (2006) 137. [102] K. Sundmacher, T. Schultz, S. Zhou, K. Schott, M. Ginkel, E.D.Gilles, Chem. Eng. Sci. 56 (2001) 333. [103] L.D. Burke, J.K. Casey, Electrochim. Acta 37 (1992) 1817. [104] C. Lamy, E.M. Belgsir, J.-M. Léger, J. Appl. Electrochem. 31 (2001) 799 [105]. R. Parsons, T. Vandernoot, J. Electroanal. Chem. 257 (1988) 9. [106] K.L. Ley, R.Liu,C. Pu, Q.Fan, N. Leyarovska, C. Segre, E.S. Smotkin, J. Electrochem. Soc. 144 (1997) 1543. [107] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairnes, J. Phys. Chem. 98 (1994) 617. [108] B.Gurau, R. Viswanathan, R.Liu, T.J. Lafrenz, K.L. Ley, E.S.Smotkin, E. Reddington, A. Sapienza, B.C. Chan, T.E. Mallouk, S. Sarangapani, J. Phys. Chem. B 102 (1998) 9997. [109] M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 267. [110] K.A. Friedrich, K.P. Geyzers, U. Linke, U. Stimming, J. Stumper, J. Electroanal. Chem. 402 (1996) 123. [111] Y. Tong, H.S. Kim, P.K. Babu, P. Waszczuk, A. Wieckowski, E. Oldfield, J. Am. Chem. Soc. 124 (2002) 468. [112] T. Freelnik, W. Visscher, J.A.R. van Veen, Langmuir 12 (1996) 3702. [113] J. Mc Breen and S. Mukerjee, J. Electrochem. Soc. 142 (1995) 3399. [114] A.S. Aricò, V. Antonucci, N. Giordano, A.K. Shukla, Barman, D.D. Sarma, J. Power Sources 50 (1994) 295. [115] N. Markovic, A. Widelov, P.N. Ross, O.R. Monteiro, I.G. Brown, Catal. Lett. 43 (1997) 161. [116] A.S. Arico, P. Cretí, P.L. Antonucci, V. Antonucci, Electrochem. Solid State Lett. 1 (1998) 66. [117] F. Vigier, S. Rousseau, C. Coutanceau, J.-M. Leger, C. Lamy, Topics in Catalysis 40 (2006) 111. [118] W.J. Zhou, S.Q. Song, W.Z. Li, Z.H. Zhou, G.Q. Sun, Q.Xin, S. Douvartzides, P. Tsiakaras, J. Power Sources 140 (2005) 50. [119] S. Song, P. Tsiakaras, App. Catal. B: Environmental 63 (2006) 187. [120] T. Iwanista, B. Rasch, E. Cattaneo, W. Vielstich, Electrochim. Acta 34 (1989) 317. [121] V.M. Schmidt, R. Ianniello, E. Pastor, S. Gonzalez, J. Pchys. Chem. 100 (1996) 17901. [122] M.H. Shao, R.R. Adzic, Electrochim. Acta 50 (2005) 2415. [123] J.P.I. De Souza, S.L. Queiroz, K. Bergmaski, E.R. Gonzalez, F.C. Nart, J. Phys. Chem. B 106 (2002) 9825. [124] A.O. Neto, M.J. Giz, J. Perez, E.A. Ticianelli, E.R. Gonzalez, J Electrochem. Soc. 149 (2002) A272. [125] A.C.C. Tseung, K.Y. Chen, Catal. Today 38 (1997) 439. [126] P.K. Shen, K.Y. Chen, A.C.C. Tseung, J. Electrochem. Soc. 142 (1995) L85. [127] K.Y. Chen, P.K. Shen, A.C.C. Tseung, J. Electrochem. Soc. 142(1995) L185. [128] K.Y. Chen, Z. Sun, A.C.C. Tseung, Electrochem. Solid-State Lett. 3 (2000) 10. [129] C. Roth, M. Goetz, H. Fuess, J. Appl. Electrochem. 31 (2001) 793. [130] W. Xu, C. Liu, W. Xing, T. Lu, Electrochem. Commun. 9 (2007) 180. [131] C. Lamy, S. Rousseau, E.M. Belgsir, C. Coutanceau, J.M. Léger, Electrochim. Acta 49 (2004) 3901. [132] A.O. Neto, E.G. Franco, E. Arico, M. Linardi, E.R. Gonzalez, J. European Ceramic Society 23 (2003) 2987. [133] X. H. Xia, H.-D. Liess, T. Iwastia, J. Electroanal. Chem. 437 (1997) 233. [134] R. Alcala, J.W. Shabaker, G. W. Huber, M. A. Sanchez-Castillo, J. A. Dumestic, J. Phys. Chem. B 109 (2005) 2075. [135] E. Antolini, J. Power Sources 170 (2007) 1. [136] L. Colmenares, H. Wang, Z. Jusys, L. Jiang, S. Yan, G.Q. Sun, R. J. Behm, Electrochim. Acta 52 (2006) 221 [137] Z. Liu, L. Hong, J. App. Electrochem. 37 (2007) 505.

143

[138] R. Alcalá, M. Mavrikakis, J.A. Dumestic, J. Catal. 218 (2003) 178. [139] V. Rao, C. Cremers, U. Stimming, L. Cao, S. Sun, S.Yan, G. Sun, Q. Xin, J. Electroanal. Soc. 154 (2007) B1138. [140] S. Liu, D.G. Kurth, B. Bredenkötter, D. Volkmer, J. Am. Chem. Soc. 124 (2002) 12279. [141] S. Liu, D. Volkmer, D.G. Kurth, Journal of Cluster Science 14 (2003) 405. [142] J.T. Rhule, C.L. Hill, D.A. Judd, Chem. Rev. 98 (1998) 327. [143] E. Papaconstantinou, Chem. Rev. 18 (1989) 1. [144] M.S. Kaba, I.K. Song, D.C. Duncan, C.L. Hill, M.A. Barteau, Inorg. Chem. 37 (1998) 398. [145] D.R. Park, H. Kim, J.C. Jung, S.H. Lee, I.K. Song, Catayst Communicaton 9 (2008) 293. [146] A.J. Bridgeman, Chemical Physics 287 (2003) 55. [147] L. C. W. Baker, D. C. Glick, Chem. Rev. 98 (1998) 3. [148] I. Kozhevnikov, “Catalysts for fine chemical synthesis. Catalyst by polyoxometallates” 2000. [149] A. Yamada, J.B. Goodenough, J. Electrochem. Soc. 145 (1998) 737. [150] M.J. Janik, R.J. Davis, M. Neurock, J. Phys. Chem. B 108 (2004) 12292. [151] I.K. Song, M.A. Barteau, J. Mol. Catal. A: Chemical 182 (2002) 185. [152] M.T. Pope, Prog. Inorg. Chem. 39 (1991) 181. [153] S. Liu, D.G. Kurth, H. Möhwald, D. Volkmer, Adv. Mater. 14 (2002) 225. [154] I. Moriguchi, J.H. Fendler, Chem. Mater. 10 (1998) 2205. [155] M. Kozik, C.F. Hammer, L.C. W. Baker, J. Am. Chem. Soc. 108 (1986) 2748. [156] M. Kozik, N. Casan-Pastor, C. F. Hammer, L. C. W. Baker, J. Am. Chem. Soc. 110 (1988) 7697. [157] W.A. England, M.G. Cross, A. Hamnett, P.J. Wiseman, J.B. Goodenough, Solid State Ionics, 1 (1980) 231. [158] B.B. Bardin, S.V. Bordawekar, M. Neurock, R.J. Davis, J. Phys. Chem. B 102 (1998) 10817. [159] C. Rong, F.C. Anson, Anal. Chem. 66 (1994) 3124. [160] T. Okuhara, H. Watanabe, T. Nishimura, K. Inumaru, M. Misono, Chem. Mater. 12 (2000) 2230. [161] W.B. Kim, T. Voilt, G.J. Rodriguez-Rivera, J.A. Dumesic, Science 305 (2004) 1280. [162] T. Okuhara, Applied Catalysis A: General 256 (2003) 213. [163] G. Koyano, K. Ueno, M. Misono, Applied Catalysis A: General 181 (1999) 267. [164] T. Okuhara, T. Nakato, Catal. Surv. from Japan 2 (1998) 31. [165] Y. Kamiya, T. Okuhara, M. Misono, A. Miyaji, K. Tsuji, T. Nakajo, Catal. Surv. Asia 12 (2008) 101. [166] T. Okuhara, T. Nishimura, M. Misono, in: Proceedings of the 11th International Congress on Catalysis, Stud. Surf. Sci. Catal., 101 (1996) 559. [167] M. Yoshimune, Y. Yoshinaga, T. Okuhara, Micropor. Mesopor. Mater. 51 (2002) 165. [168] J. Wang, Analytical Electrochemistry, second edition, 2001. [169] Z. Galus, Fundamentals of electrochemical analysis, second edition, 1994. [170] H.H. Willard, L.L. Merritt, J.A. Dean, F.A. Settle, Instrumental methods of analysis, 7th edition, 1988. [171] R. Varma, J.R. Selman, Techniques for characterization of electrodes and electrochemical process, 1991. [172] Z. Stojek, Electroanalytical Methods. Guide to experiments and Applications, F. Scholz-Editor, Springer, Heidelberg 2002, p.102. [173] E. Barsoukov, J.R. Macdonald, Impedance spectroscopy theory, experimental, and Applications, second edition, 2005. [174] R. Marassi, F. Nobili, chapter in Encyclopaedia of electrochemical power sources, 2008. [175] Y. Ding, Fundamental Theory of Transmission Electronic Microscopy. www.nanoscience.gatech.edu [176] D. Kealey, P.J. Haines, Analytical Chemistry, 2002. [177] N. Essayem, Y.Y. Tong, H. Jobic, J.C. Vedrine, Applied Catal. A. General., 194-195 (2000) 109. [178] N. Essayem, A. Holmqvist, P.Y. Gayraud, J.C. Vedrine, Y. Ben Taarit, J. Catal., 197 (2001) 273. [179] N. Essayem, Y. Ben Taarit, P.Y. Gayraud, G. Sapaly, C. Naccache, J. Catal., 204 (2001) 157. [180] N. Essayem, G. Coudurier, J.C. Vedrine, D. Habeamacher, J. Catal., 183 (1999) 199. [181] C. Rocchiccioli-Deltcheff, R. Thouvenot, R. Franck, Spectrochim. Acta A 32 (1976) 587. [182] Q.J. Shan, J. Gong, L.Y. Qu, J. Coord. Chem. 55 (2002) 319. [183] L.R. Pizzio, M.N. Blanco, App. Catal. A General 255 (2003) 265. [184] N.K. Mizuno, K.Y. Yoneda, M. Misono, J. Catal. 83 (1983) 384 [185] M. Fourier, C. Rocchiccioli-Deltchef, L.P. Kazansky, Chem. Phys. Lett. 223 (1994) 297. [186] M. Lira-Cantú, P. Gómez-Romero, Chem. Mater. 10 (1998) 698.

144

[187] M. Skunik, B. Baranowska, D. Fattakhova, K. Miecznikowski, M. Chojak, A. Kuhn, P.J. Kulesza, J. Solid State Electrochem. 10 (2006) 168. [188] A. Bielanski, J. Datka, B. Gil, A. Malecka-Lubanska, A. Micek-Ilnicka, Cat. Lett. 57 (1999) 61. [189] G.M. Brown, M.R. Noe-Spirlet, W.R. Busing, H.A. Levy, Acta. Cryst. B 33 (1977) 1038. [190] A. Bielanski, A. Lubanska, J. Pozniczek, A. Micek-Ilnicka, App. Catal. A: General 256 (2003) 153. [191] C. Danumah, S. Vaudreuil, L. Bonneviot, M. Bousmina, S. Giasson, S. Kaliaguine, Microporous Mesoporous Mater. 44-45 (2001) 241. [192] M.I. Ahmad, S.M.J. Zaidi, S.U. Rahman, S. Ahmed, Microporous Mesoporous Mater 91 (2006) 296. [193] D. Ingersoll, P.J. Kulesza, L.R. Faulkner, J. Electrochem. Soc. 141 (1994) 140. [194] A. Lewera, M. Chojak, K. Miecznkiowski, P.J. Kulesza, Electroanalysis 17 (2005) 1471. [195] W.L. Latimer, The Oxidation States of the Elements and Their Potentials in Aqueous Solutions, 2nd Edition, Prentice-Hall, INC. Englewood Cliffs, N. J, 1952 (Chapter12). [196] H.A. Gasteiger, N.M. Marković, P.N. Ross Jr., J. Phys. Chem. 99 (1995) 8290. [197] A.J. Bard, L. Faulkner, Electrochemical Metchods, Fundamentals and Applications, John Wiley & Sons INC., New York, 2001 (Chapter 3, chapter 9). [198] Z.D. Wei, S.H. Chan, L.L. Li, H.F. Cai, Z.T. Xia, C.X. Sun, Electrochim. Acta 50 (2005) 2279. [199] A. Pozio, M. De Francesco, A. Cemmi, F. Cardellini, L. Giorgi, J. Power Sources 105 (2002) 13. [200] M. Ciureanu, H. Wang, J. Electrochem. Soc. 146 (1999) 4031. [201] M.J. Weaver, S.C. Chang, L.W.H. Leung, X. Jiang, M. Rubel, M. Szklarczyk, D. Zurawski, A. Wieckowski, J. Electroanal. Chem. 327 (1992) 247. [202] F. Gloaguen, J.M. Léger, C. Lamy, J. Appl. Electrochem. 27 (1997) 1052. [203] B. Scrosati, R. Marassi, A. Kolary, A. Zurowski, P.J. Kulesza, S. Dsoke “Catalizzatori a basso contenuto di Platino per celle a combustibile” RM 2007A000228, April 2007. [204] B. Scrosati, R. Marassi, A. Kolary, A. Zurowski, P.J. Kulesza, S. Dsoke, “ Catalysts having low Pt content for fuel cells”, PCT/IB2008/051462, April 2008. [205] V.E. Kazarinov, G.Y. Tysyachnava, V.N. Andreev, J. Electroanal. Chem. 65 (1975) 391. [206] G.R. Mundy, R.J. Potter, P.A. Christensen, A. Hammett, J. Electroanal. Chem. 27 (1990) 257. [207] Y. Zhu, H. Uchida, T. Yajima, M. Watanabe, Langmuir 17 (2001) 146. [208] R.E. Melnick, G.T.R. Palmore, J. Phys. Chem. B 105 (2001) 1012. [209] G. Wu, L. Li, B.Q. Xu, Electrochim. Acta 50 (2004) 1. [210] W. Sugimoto,K.Aoyama, T. Kawaguchi, Y. Murakami, Y. Takasu, J. Electroanal. Chem. 576 (2005) 215. [211] P.K. Shen, A.C.C. Tseung, J. Electrochem. Soc. 141 (1994) 3082. [212] M. Watanabe, S. Motoo, J. Electroanal. Chem. 60 (1975) 275. [213] H. Tang, J.H. Chen, L.H. Nie, D.Y. Liu, W. Deng, Y.F. Kuang, S.Z. Yao, J. Colloid Interface Sci. 269 (2004) 26 [214] T. Iwasita, E. Pastor, Electrochim. Acta 39 (1994) 531. [215] H. Hoster, T. Iwasita, H. Baumgartner, W. Vielstich, J. Electrochem. Soc. 148 (2001) A496. [216] J. Otomo, X. Li, T. Kobayashi, C. Wen, H. Nagamoto, H. Takahashi, J. Electroanal. Chem. 573 (2004) 99. [217] J.J. Wang, G.P. Yin, J. Zhang, Z.B. Wang, Y.Z. Gao, Electrochim. Acta 52 (2007) 7042. [218] D.A. Harrington, B.E. Conway, Electrochim. Acta 32 (1987) 1703. [219] C.N. Cao, Electrochim. Acta 35 (1990) 831. [220] R.E. Melnick, G.T.R. Palmore, J. Phys. Chem. B 105 (2001) 9449. [221] V.S. Bagotzky, Y.B. Vassilyew, Electrochim. Acta 12 (1967) 1323. [222] Z. Hou, B. Yi, H. Yu, Z. Lin, H. Zang, J. Power Sources 123 (2003) 116.

145