spinel catalysts in aqueous solutions

Accepted Manuscript Title: Oxygen reduction at the surface of polymer/carbon and polymer/carbon/spinel catalysts in aqueous solutions Author: V.G. Khomenko K.V. Lykhnytskyi V.Z. BarsukovISE Member PII: DOI: Reference:

S0013-4686(13)00654-3 http://dx.doi.org/doi:10.1016/j.electacta.2013.04.019 EA 20319

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

1-9-2012 27-1-2013 5-4-2013

Please cite this article as: V.G. Khomenko, K.V. Lykhnytskyi, V.Z. Barsukov, Oxygen reduction at the surface of polymer/carbon and polymer/carbon/spinel catalysts in aqueous solutions, Electrochimica Acta (2013), http://dx.doi.org/10.1016/j.electacta.2013.04.019 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Oxygen reduction at the surface of polymer/carbon and polymer/carbon/spinel catalysts in aqueous solutions V.G. Khomenko1, K.V. Lykhnytskyi and V.Z. Barsukov*1 Kiev National University of Technologies and Design, 2 Nemirovich-Danchenko str.,

ip t

Kiev, 01011, Ukraine; E-mail: [email protected]; [email protected]

cr

Abstract

us

The mechanism and kinetics of oxygen reduction were studied at such composite catalysts as PANI/C, PPy/C, transition metal oxides/C and PPy/C/transition metal oxides by the method

an

of rotating disc electrode. It was shown that for the PANI/C and PPy/C catalysts there are almost no diffusion limits and the sluggish stage of the reaction is chemo sorption of oxygen

M

at the polymer surface followed by desorption and formation of hydrogen peroxide (a twoelectron mechanism). In contrast, for the inorganic composites transition metal oxides/C the

approximately 3.0.

ed

limiting stage is oxygen diffusion, whereas the calculated effective number of electrons is

pt

In organic/inorganic composite catalysts of the PPy/C/transition metal oxides type oxygen

Ac ce

reduction proceeds by a mixed diffusion-adsorption mechanism to form water (a four-electron mechanism). The calculated effective number of electrons for such available catalysts is 3.8 in alkaline electrolytes, which, by efficiency, approaches the ideal value 4.0 for noble metals. Keywords: ORR, composite catalysts, conducting polymers, transition metal oxides, spinels.

*

Corresponding author. Tel.: +380675044565, Fax: +380 44 284 8266.

E-mail addresses: [email protected]; [email protected] 1

ISE Member 1

Page 1 of 44

1. Introduction Oxygen is the cheapest, lightest, most efficient and available oxidizing agent, which is widespread in nature. It is a paradox that in electrochemical power-engineering other solid and liquid oxidizers are conventionally used in energy conversion devices instead of oxygen,

ip t

though these oxidizers are inferior to oxygen in all the above criteria. To change the situation radically and make wide use of fuel cells and high-efficiency metal-air batteries, development

cr

of inexpensive, efficient and available catalysts of oxygen reduction is of primary importance

us

as these can substitute for existing catalysts based on silver, platinum and the other precious metals [1]. Research into the mechanisms and kinetics of oxygen reduction reaction (ORR) at

an

some promising polymer/carbon and polymer/carbon/oxide composites is essential for the development of such catalysts.

M

In [2-4] a noticeable catalytic activity of films of polyaniline (PANI), polypyrrole (PPy), polythiophen (PTh) and some other electronically conducting polymers (ECPs) toward the

ed

oxygen reduction was reported. In particular, PANI and PPy showed the highest catalytic

toward ORR at all.

pt

activity, whereas poly(3,4-ethylenedioxythiophene) (PEDOT) did not reveal any activity

Ac ce

In order to explain the mechanism of catalytic activity of ECPs, a quantum-chemical modeling of ECPs and adsorption complexes of “ECP-oxygen” has been performed [4]. In principle, each carbon atom of ECP can be an adsorption site. Nevertheless, a dissolved molecular oxygen can be adsorbed on the ECP surface only when both oxygen atoms form bonds to surface carbon atoms, i.e. a so called “bridge model” of adsorption has been realized. Furthermore, an optimization of structure for adsorption complex on criterion of potential energy minimum gives possibility to determine the most likely adsorption sites for each type of ECP, as well as the bond orders and the bond lengths. For example, the calculations show that the bond orders in chemisorbed oxygen molecules at PANI decrease by a third, and the 2

Page 2 of 44

bond length increases by more than 20% in comparison with that in a free oxygen molecule. Thus, chemisorbed oxygen molecules have a fairly high degree of activation and can be readily reduced at the PANI surface. A similar mechanism takes place on the active carbon atoms of PPy, PT, PMeT. Thus, quantum-chemical modeling of adsorption complexes of

ip t

oxygen with the above polymers presented in [4] allowed a hypothesis that such catalytic activity is bound up with the electronic structure of ECPs and the possibility of chemisorption

cr

of oxygen molecules dissolved in the electrolyte at the ECP surface. According to the model

us

reported the absence of catalytic effect at some ECPs is due to the structure of the polymers, as, for example, chemisorption of an oxygen molecule at the PEDOT surface does not occur

an

at all because of steric difficulties.

In [2-4] an assumption was made that the electrochemical stage of the reaction proceeds by a

M

2-electron mechanism, which, depending on the pH of solution, can be represented by the following well-known equations:

ed

in acidic solution О2 + 2H+ + 2e- = H2O2;

pt

in alkaline solution:

E0 = 0.682 V

E0 = – 0.065 V

(2).

Ac ce

О2 + H2O + 2e- = HO2- + OH-;

(1),

The presence of hydrogen peroxide in the electrode layer was verified by experiment, namely the sample of electrolyte taken from this layer changed its color on interacting with the KI indicator solution [4].

Thus, today there is much evidence for the two-electron ORR mechanism at PANI and PPy catalysts. At the same time, the nature of the limiting stage of electrode reaction for such processes is still a question that needs investigating. To answer the question, in this work a detailed study of ORR at PANI and PPy was carried out by the method of rotating disc electrode [5, 6]. A 3

Page 3 of 44

study like this usually allows one not only to determine the limiting stage of a reaction, but also to calculate the effective number of electrons involved in the electrode process and therefore to gain additional arguments for the ORR mechanism at a catalyst. Besides, of a great theoretical and practical interest is studying the possibility of developing

ip t

polymer composite catalysts, which would be available and would allow the effective number of electrons to be as close to four as possible, as it is the case with the Pt catalysts, for

cr

example. Some composite catalysts based on PPy, although combined with such precious

us

metals as Pt-Co [7] or Pd [8], show rather high efficiency.

In our work, it was assumed that some transition metal oxides, which are capable of

an

catalyzing decomposition of hydrogen peroxide to water, can promote a deeper oxygen reduction at the ECP/carbon/transition metal oxide composite catalyst. In this case, the ORR

M

proceeding by the 4-electron mechanism can be expected:

O2 + 2 H2О + 4e- = 4 ОH-

О2 + 4H+ + 4e- = 2 H2O;

pt

in acidic solution:

E0 = 0.401 V

(3),

E0 = 1.229 V

(4).

ed

in alkaline solution

Ac ce

Development of such composite polymer catalysts can enable one not only to increase considerably the efficiency of ORR, but also to get rid of hydrogen peroxide, which create numerous problems with corrosion, in an electrochemical device. This paper aims at a detailed study of the above questions.

2. Experimental 2.1. Synthesis of composites 2.1.1. Electrochemical synthesis of ECPs

4

Page 4 of 44

To electrochemically synthesize ECP films at Pt or carbon materials the following electrolytes were used: −

PANI was deposited in 1 М acid solution, containing 0.1 М aniline;

−

PPy was produced in 0.5 М КСl solution, containing 0.3 М pyrol;

ip t

Before synthesis, all the electrolytes were blown with argon to remove gases. The synthesis of ECPs was carried out by gradually building up the film under repeated cyclic

cr

potential sweeps at sweep rates of 10-100 mV/s. The electrodes with ECP films were

us

thoroughly washed with a solvent which was used in the synthesis and acetone and then were

an

air- and vacuum dried.

2.1.2. Chemical synthesis of ECPs

M

In many cases, PANI and PPy were synthesized chemically as well, which allowed homogeneous polymer layers well-keyed to the Pt surface, with a thickness of about 1 µm, the

ed

layers being produced chemically quicker than electrochemically. PPy was produced as a black powder by oxidation of the monomer in the presence of a

pt

catalyst. FеС13 solution was used as an oxidizer [9, 10]. The Fе(III)/monomer concentration

Ac ce

ratio was chosen according to [11] and amounted to 2.4, the PPy yield in this case is close to 100%. To achieve high conductivity the PPy polymerization was carried out at 0 – 5 °С [12]. In polymerization by this way the PPy is doped with Сl– ions during the synthesis.

The

polymerization reaction can be expressed as follows [9]: nC4H5N + (2 + y)n FeCl3  [(C4H3N)n+ ny Cl–] + (2 + y) FeCl2 + 2n HCl

(5),

where у is the oxidation (PPy doping) number. PANI was synthesized by oxidation of of aniline salt in acid solution, K2Cr2O7 being used as an oxidizer. The aniline salt was produced by dissolving aniline in hydrochloric acid. The synthesis was carried out at 0 - 5°С [13-16]. 5

Page 5 of 44

Carbon materials (graphite powder, carbon black and others) were introduced into the reaction mixture prior to polymerization in order that the polymer is deposited on the carbon surface immediately in the process of synthesis. The reaction mixture was thoroughly mechanically stirred.

ip t

2.1.3. Synthesis of transition metal oxides on the surface of carbon materials Chemical-purity metal (Ni, Co) nitrates were used as precursors for the synthesis. To produce

cr

the oxides thermal or sol-gel method [17, 18] were usually used.

us

In the thermal method, the oxides were produced by decomposition of nitrates at 300оС. To produce the composites with graphite-carbon materials, a solution of nitrates was prepared in The

an

the distilled water-ethanol mixture (50:50), where the carbon material was placed.

suspension produced was evaporated under constant stirring, and then the product was placed

M

into an oven at 300оС for 1 hour.

The sol-gel method allowed an oxide material with a more developed surface and better

ed

coherence with the carbon substrate [19]. To produce the oxide, a nitrate solution of the corresponding metal was prepared, to which NH4HCO3 solution was added. The precipitate of

pt

metal carbonates produced was filtered with the Buckner funnel, washed with distilled water

Ac ce

and dried. The powder produced was dissolved in propionic acid. To produce the composites, graphite-carbon material was added. The suspension was evaporated under constant stirring until gel was formed. The material produced was placed into a crucible and kept in an oven at 280оС for 1 hour.

2.1.4. Synthesis of ECP composites with carbon and transition metal oxides The synthesis of a three-component composite was carried out after the synthesis of transition metal oxides at the carbon surface according to 2.1.3. The ECP was introduced to the oxide composite either mechanically (by finely milling with zirconium oxide grinding balls and

6

Page 6 of 44

stirring the mixture with ultrasound) or by introducing the oxide composite into the reaction mixture prior to polymerization (as described in 2.1.2).

2.2. Microscopy and elementary analysis

ip t

To determine the microstructure and composition of the composite surface, the РЭММА2000 scanning electron microscope (SEM) with Energy Dispersive X-Ray Analysis (EDX)

cr

was used. Each individual particle of catalysts has been separately probed with the focused

us

electron beam. The resulting EDX spectra allow identifying the different elemental composition from each particle probed. This information was used to determine the nature of

an

catalyst.

Besides, the crystal structure of catalysts were determined using an X ray diffractometer

M

(XRD, DRON-1, Russia) with Cu Kα radiation (λ=0.1514178 nm). The diffractograms were

(º)·min-1.

ed

obtained in a scan range of 2θ from 10º to 90º with a step width of 0.04 º and scan rate of 5

pt

2.3. Electrochemical measurements

Ac ce

To determine kinetic parameters of oxygen reduction and to calculate the effective number of electrons involved in the reaction, a glassy carbon rotating disc electrode with a diameter of 3 mm was used as a working one [5, 6]. The electrochemical measurements were taken in a three-electrode cell using a silver chloride reference electrode and a glasscarbon electrode. Composite catalyst was dispersed in 15 wt.% Nafion solution (LQ-1015) from Ion Power Inc. (USA) and the resultant suspension was agitated in an ultrasonic bath for 10 min to make catalyst ink. The weight percentage of Nafion binder in the ink was in the range 15-30 wt%. The suspension was coated on the glassy carbon electrodes, which were air-dried for 30 min to evaporate the solvent. The weight of the ink after drying is controlled at 2 mg∙cm-2. 7

Page 7 of 44

Cyclic voltammetric curves or cathode polarization curves were taken at rotation speeds from 0 to 2500 rpm in solutions saturated with oxygen or argon (for comparison). Potentiodynamic curves were taken at a potential sweep rate of 2 mV/s in the air and argon atmosphere. Voltammetric curves were taken in the -1.0 − +0.4 V potential range with the

ip t

standard hydrogen electrode as a reference one. Electrochemical investigations were performed using the VMP3 automatic multi-channel

us

MSTAT potentiostat MSTAT from Arbin Instruments, USA.

cr

potentiostat/galvanostat from Princeton Applied Research, UK, as well as with the 32-channel

an

3. Results and discussion 3.1. Morphology of PANI and PPy films

M

3.1.1. Electrochemical synthesis

Thin PANI films at a Pt electrode were produced by subsequent electrochemical oxidation of

ed

0.1М monomer solution in 1М solution of hydrochloric acid. A typical cyclic voltammogram (CVA) for the PANI electrosynthesis is presented in Fig.1a (first cycles of electrosynthesis)

pt

and Fig.1b (continued formation of PANI’s film).

Ac ce

An absolute advantage of this method is the possibility of producing electrodes with a controlled thickness of the polymer. The thickness of PANI depends on the number of cycles. Fig. 2 shows the morphology of PANI produced in the first (a) and the second (b) cycles of CVA. Microscopic studies prove that an electrochemically deposited ECP consists of two layers – a thick film layer (a), adjacent to the electrode surface, and a more friable one (b), which is formed at the polymer surface, that is on the thick layer. The friable layer is composed of subprimary structures such as fibrils and globules. The structure of these depends on the nature of polymer, synthesis conditions and electrolyte composition. Our studies show that PANI mainly forms the fibril structure, whereas PPy forms the globular 8

Page 8 of 44

structure. With increasing the thickness of the ECP film, a decrease in the polymerization rate at active centers is observed, the active centers being the fibrils and globules. This results in the formation of friable agglomerates with a thickness of 10-50 μm. These agglomerates are easily removed from the electrode surface, forming a fine-dispersion suspension. That is why

ip t

it appears impossible to produce films with a thickness of more than a few dozens of microns

cr

by this method.

us

3.1.2. Chemical synthesis

ECPs are chemically synthesized by polymerization of a monomer under the action of an

an

oxidizing agent by the techniques presented in 2.1.2. Fig. 3 shows microphotographs of chemically synthesized PANI (a) and PPy (b), which, as in the case of electrochemical

M

synthesis, have fibril and globular structure, respectively. It is necessary that such conditions of the chemical synthesis should be created in which the ECP is the main product of the

ed

reaction. Our studies show that, in this case, the chemical and electrochemical syntheses give

pt

identical polymers, as their electrochemical behavior is almost the same.

Ac ce

3.2. Oxygen reduction at ECPs

To determine the mechanism and kinetics of oxygen reduction, thin PANI and PPy films were chemically applied to a glassy carbon disc electrode. Cathode polarization curves were taken on these films at different rotation speeds. Results are given in Fig. 4 in the “current vs square root of rotation speed” coordinates, as it is conventional when using the method of rotating disc electrode. The current density of oxygen reduction appeared not to be dependent on the rotation speed of the disc electrode for the PANI and PPy films (Fig. 4).

9

Page 9 of 44

This is a quite unusual result, which indicates that the ORR rate on the PANI and PPy films does not depend on the rate of oxygen diffusion, at least at the polymer-electrolyte interface. Consequently, the sluggish stage of the process has a pronounced kinetic nature, but not a diffusion one.

ip t

From this a conclusion can be drawn that the kinetics of oxygen reduction is most likely to be limited by the rate of formation of the PANI−О2 and PPy−О2 adsorption complexes, which is

cr

also proved by the quantum-chemical calculations reported in [4].

us

Thus the mechanism of the electrocatalytic oxygen reduction at the surface of PPy/C and PANI/C composites can be described by the following scheme shown in Fig. 5.

an

According to this scheme, the molecules of oxygen dissolved in the electrolyte chemisorb on the surface of electroconducting polymers, noticeable weakening and extension of the bonds

M

between oxygen atoms take place up to breaking of the double bond. The interaction between the oxygen activated in this way and hydroxonium ions in acidic solution becomes easier and

ed

results in the formation of hydrogen peroxide.

Thus, the 2-electron mechanism of ORR on the films of PANI and PPy catalysts, which

pt

proceeds via the formation of bridge-type adsorption complexes on the ECP surface, finds

Ac ce

additional proof in this paper due to the employment of the method of rotating disc electrode.

3.3. Investigation and development of ORR at composite catalysts 3.3.1. The bases for the development of a composite catalyst Another important problem is devoted to development of an inexpensive and efficient polymer composite catalyst, which could allow ORR by the 4-electron mechanism. Our main hypothesis consists in the following. It is well-known that some oxides of transition metals (like NiO2, CoO2, Fe2O3, MnO2, etc.) are quite good catalysts of H2O2 to H2O conversion. 10

Page 10 of 44

Let us assume that one of these oxides (or even a few of them) is introduced to a polymer composite material, then the hydrogen peroxide formed in the 2-electron process could be expected to immediately decompose to water, which could allow the 4-electron reaction by the following scheme (Fig. 6).

ip t

The maximum possible number of electrons for the ORR in such a three-component (ECP/carbon/transition metal oxide) composite is likely to reach 4. Actually, n being close to

cr

4 will depend on how complete the decomposition of hydrogen peroxide to water is and can

us

vary from 3 to 4.

To bring the calculated n value as much as possible to 4, which is characteristic of the Pt

an

metals, the three components should be most properly selected: (1) the polymer;

ed

(3) the transition metal oxide(s).

M

(2) the carbon;

3.3.2.The selection of the polymer component

pt

As was reported in [4], PANI demonstrates the highest catalytic activity towards ORR of all

Ac ce

the ECPs, PPy being the second most catalytically active one. That is why, in terms of the selection of the polymer catalyst for acid and neutral electrolytes, it is PANI that can be recommended as the polymer component (in combination with thermally expanded graphite, for example). Such a catalyst can be rather efficient in low-power fields, because, as it was shown above and in [2-4], it provides only the 2-electron ORR. If one tries to provide the 4electron ORR with the help of polymer catalysts along with transition metal oxides, which are stable only in alkaline solution, PPy is undoubtedly the best choice as it shows stable work in alkaline solution as well.

11

Page 11 of 44

3.3.3. The selection of the carbon materials As the most expedient carbon material for the composite catalyst developed, such a comparatively new material as graphitized carbon black was chosen (the trademark PureBlack®) from Superior Graphite Co. (Chicago, IL, USA) and Columbian Chemicals Co.

ip t

(Marietta, GA, USA). This nano-structured material combines an advantage of carbon black (a sufficiently high area ca 40-60 m2∙g-1) and that of graphite (high conductivity) [19].

cr

PureBlack® could be produced from ungraphitized precursors (different types of carbon

us

black), which proved to be graphitizable, i.e. they were partially converting into graphite-like structures upon heat treatment at temperatures in the range of 1500–3000 ºC. Besides graphitization, as a result of treatment at so high temperatures one can receive a super pure

an

carbon material with a carbon content ca 99.95%. In particular, the elemental mineral impurities in the ash of PureBlack® which were measured with emission spectroscopy are as

M

following: As, Co, Cr, Cu, Mo, Ni, Pb, Sb, Sn, V ≤ 1 ppm; Fe ≤ 3 ppm; Si, Al ≤ 5 ppm; Ca ≤ 10-12 ppm. Besides, the material has very low to zero volatile content (external oxygen, sulfur, etc., groups) and very low equilibrium moisture pickup (20 ppm level) [19].

ed

On the other hand, our studies show that this material has high resistance to corrosion

pt

compared to other carbon materials, especially at high temperature. In particular, Fig. 7 presents mass loss curves for acetylene black, activated carbon and

Ac ce

graphitized PureBlack® under heating in the air in identical conditions. It can be easily concluded from Fig. 7 that the graphitized PureBlack® demonstrates a minimum mass loss in heating in the atmospheric conditions up to 600 °С (about 10-12%), whereas the carbon black loses almost 70% of its mass, and the activated carbon loses almost 90% in the same conditions. Such resistance to corrosion is important when using hightemperature methods of synthesis of composite catalysts on the surface of carbon, in particular, the thermal decomposition of salts or sol-gel method. CVAs for an electrode based on PureBlack® in 1М NaOH solution saturated with argon are shown in Fig. 8. It should be noted that the absence of pronounced peaks on the CVAs proves 12

Page 12 of 44

that there are no oxygen-containing surface groups (of the quinone/hydroquinone type), which indicates that the material is quite inert and can be used as carbon substrate in the composite. These CVAs, especially at high potential sweep rates, allow a conclusion that the electrode based on the graphitized black has fairly high double layer capacity, which is typical of

ip t

materials with high specific surface.

cr

3.3.4. The investigation of the catalytic activity of graphitized black and its composites with

us

transition metal oxides

To evaluate the catalytic activity of PureBlack® and its composites with transition metal

an

oxides towards the ORR, these materials were tested using the rotating disk electrode (RDE) technique. CVAs taken for PureBlack® in 1M NaOH solution saturated with oxygen at

M

different rotation speeds are given in Fig. 9a.

Results of the investigation show that graphitized carbon black itself possesses a certain

ed

catalytic activity towards the ORR.

The PureBlack®/nickel oxide and PureBlack®/cobalt oxide composites were synthesized by

pt

the techniques described in 2.1.3. and investigated using the rotating disc electrode (Fig. 9b,

Ac ce

9с). Moreover, the investigation showed that the catalytic activity of mixed nickel and cobalt oxides appeared to be considerably higher than that the individual activities of these simple oxides. For this reason, the PureBlack® – NiСo2O4 spinel composite was also synthesized by the sol-gel method. The ratio of materials was calculated from the masses of components used (polypyrrol and the NiCo2O4/PureBlack composite) prior to mechanical-chemical stirring. The chemical composition of the oxide/ PureBlack composites was controlled with the help of EDX spectra (2.2). Electrochemical results of the study of such a composite are presented in Fig. 9d.

13

Page 13 of 44

The reproducibility of the systems and the stability of the interface are very good. For example, taking the CVAs given in Fig. 9a – 9d has been repeated for hundreds cycles and there has not been observed any noticeable change in the curves. To treat and compare the potentiometric curves, the data from Fig. 9 (a-d) were replotted as

ip t

Koutecky-Levich plots (Fig. 10). Linear dependence is observed in the coordinates given. Calculations were done by the Koutecky-Levich equation:

1  1  1 i iLev ik

cr

(6)

us

Here ILev is the Levich current and Ik is the kinetic current and are defined by eqs 7 and 8. The diffusion current dependences on ω1/2 for all the materials studied in this section are in

an

rather good accord (taking into account the background current different from 0) with linear

M

dependences derived from the well-known Levich equation [5, 6]

(7)

ik  nF  S  С  k  

(8),

ed

iLev  0.62  nF  S  D 2 3  1 6  C   0,5

where n is the number of electrons, F is the Faraday constant, D is the diffusion coefficient, ν

pt

is the viscosity, C is the oxygen concentration, S is the electrode area,  is the electrode

Ac ce

rotation speed (rad/s), Ψ is the surface coverage of the oxygen, k is the rate constant. Proceeding from the slope of the line the effective number of electrons for any material used as an ORR catalyst can be easily calculated. Table 1 presents the corresponding empirical equation of the dependence shown in Fig. 10 for every catalyst under study as well as the calculated effective number of electrons involved in the ORR.

14

Page 14 of 44

Table 1. Calculated kinetic ORR parameters for different catalysts The effective number

The kinetic equation for diffusion

of electrons, n

current [mA-1]

PureBlack®

2.29

1/I = – 25,44/√ω – 1.86

5% СoOx/ PureBlack®

2.79

1/I = – 19,29/√ω – 1.67

5% NiOx/ PureBlack®

3.41

1/I = – 16,34/√ω – 1.39

5% NiСo2O4/ PureBlack®

3.72

1/I = – 17,66/√ω – 0.91

us

cr

ip t

Catalyst

The data given show that the ORR on graphitized black (as well as on the other carbon and

an

graphite materials) proceeds by the two-electron mechanism.

The activity of inorganic oxide catalysts synthesized on PureBlack® increases from cobalt

M

oxide to nickel oxide. The effective number of electrons for the oxide catalysts is, on average, about 3. The NiСo2O4 spinel on the same carbon substrate demonstrates n = 3.74 and,

ed

therefore, appears to be much more active than each of the oxide catalysts.

pt

3.3.5. Selection of an optimal method for producing the composite.

Ac ce

Figure 11 shows voltammetric curves for the NiCo2O4/carbon material composites produced by different methods. Such composites can be produced by simple thermal decomposition of nitrates of the corresponding metals taken in a certain ratio (curve 1), sol-gel method with using acetic acid (curve 2) and sol-gel method with using propionic acid (curve 3). From comparison of these curves it can be easily concluded that the sol-gel method with using propionic acid (curve 3) is an optimal one, since it is the method of synthesis that enables the highest catalytic activity in the whole possible potential range. Fig.12a presents XRD diffractograms for NiCo2O4, produced by different methods of synthesis. The crystal structure of the nickel-cobalt spinel was confirmed by the presence of 15

Page 15 of 44

corresponding reflexes [21] in the diffractogram for the material produced by thermal decomposition of corresponding nickel and cobalt salts. It should be noted that there is a decrease in the intensity of reflexes (Fig. 12b) for the material produced by the sol-gel method, which is indicative of a considerable decrease in the size of particles. That is why the

ip t

method is preferable when producing a high-activity catalyst. 3.3.6. Optimization of the composition of two-component NiCo2O4/PureBlack® catalyst

cr

In order to select an optimal composition of the two-component catalyst, catalyst specimens

us

with different contents of the basic component (NiCo2O4) were synthesized.

Data given in Figure 13 enable the conclusion that the optimal content of the basic component

M

an

is about 5%. This conclusion is in accord with the data on similar catalysts reported in [11].

3.3.7. Optimization of the composition of three-component PPy/NiCo2O4/PureBlack®

ed

catalyst

pt

Comparative data in Figure 14 allow the conclusion that introducing a conductive polymer into the composite noticeably decreases the electrode polarization.

Ac ce

Introducing PPy into the composite can be carried out either by a direct chemical synthesis of the polymer on the oxide/carbon material composite surface or by a mechanical mixing. The latter was chosen for the further use, since the synthesis of PPy involves oxidizing agents, which can affect the oxide material. The effects of polypyrrol on the shape of cathode curves of oxygen reduction are shown in Figure 15. It is seen from the Figure that the highest reduction current is observed at a polypyrrol content of 15%.

16

Page 16 of 44

For the 15% PPy /5%NiCo2O4/PureBlack® the dependence Koutecky-Levich plots was determined (Figure 16). It is seen from Fig. 16 that at relatively small electrode polarizations (0.10-0.15 V) there is a bend on the lines. This gives evidence for a mixed mechanism of the process, which is limited

ip t

by both the oxygen diffusion stage and the rate of formation of the oxygen-polymer adsorption complex. Thus, in a mixed organic-inorganic composite catalyst at small

cr

polarizations there are limiting stages, characteristic of both the organic and inorganic

us

catalyst.

With increasing the electrode polarization to 0.4 – 0.5V the bend on the lines disappears. This

an

indicates that at high polarizations the limiting stage of the process is oxygen diffusion. From the slope of the lines (Fig. 16) the calculated effective number of electrons involved in

M

the oxygen electroreduction was found out to be n = 3.80 at both lower and higher polarizations. This number of electrons approaches the idealized value n = 4, characteristic of

4. Conclusion

pt

ed

the Pt catalyst of oxygen reduction [20].

Ac ce

Thus, in this work the mechanism of oxygen reduction on different composite catalysts was considered. In particular, it was shown that there are no diffusion limitations in the oxygen reduction on the PANI/C and PPy/C catalysts in acid solution. The assumption made previously that the sluggish stage of the ORR in similar systems is oxygen chemo sorption on the polymer surface followed by desorption and formation of hydrogen peroxide (the 2electron mechanism) was experimentally verified. The limiting stage and main kinetic parameters were determined for the transition metal oxides/C composites. The process in such systems was shown to be limited by the stage of oxygen diffusion, and the average effective number of electrons is about 3.0. 17

Page 17 of 44

It was established that the oxygen reduction on the PPy/С/transition metal oxides composites proceeds by a mixed diffusion-adsorption mechanism to form water (the 4-electron mechanism). Compositions were developed for inexpensive composite catalysts in alkaline

ip t

electrolytes, which, by their efficiency, are close to the materials containing precious metals.

Acknowledgements

cr

This work was supported by the Ukrainian Ministry of Education and Sciences in the

us

framework of a bilateral cooperation project between Ukraine and Greece. The authors would like to acknowledge Professor Georgios Kokkinidis and Ass. Professor Sotirios Sotiropoulos

an

from Aristotle University of Thessaloniki for the valuable discussions and interest for these

M

investigations.

References

ed

[1] D. Linden, T. B. Reddy, Handbook of Batteries, 3rd ed., McGraw Hill, New York, 2001. [2] V.Z. Barsukov, S.V. Chivikov, Electrochim. Acta 41 (1996) 1773.

pt

[3] V.Z. Barsukov, V.G. Khomenko, S.V. Chivikov, et al. Electrochim. Acta 46 (2001) 4083.

Ac ce

[4] V.G. Khomenko, V.Z. Barsukov, A.S. Katashinskii, Electrochim. Acta 50 (2005) 1675. [5] K.J. Vetter, Electrochemical kinetics; theoretical aspects, Academic Press, New York, 1967. [6] F.P. Miller, A.F. Vandome, J. McBrewster, Levich Equation, Dr. Mueller e.K., VDM Verlag, 2010. [7] W.M. Martínez, T. T. Thompson, M. A. Smit, International Journal of Electrochemical Science 5 (2010) 931. [8] C. Jeyabharathi, P. Venkateshkumar, J. Mathiyarasu, K. L. N. Phani, J. Electrochem. Soc. 157 ( 2010) B1740. 18

Page 18 of 44

[9] S.P.Armes, Synthetic Metals 20 (1987) 365. [10] A.Pron, Z.Kucharski, C.Budrowski, et al., J. Chem. Phys. 83 (1985) 5923. [11] T. Skotheim, Handbook of Conducting Polymers, M. Dekker, New York, 1998. [12] S.Rapi, V.Bocchi, G.P.Gardini, Synthetic Metals 24 (1988) 217.

ip t

[13] M.Vilchik, M. Zigon, M.Zupan, A. Sebenik, Acta Chimica Slovenica 45 (1998) 173. [14] J.Stejskal, Pure Appl.Chem. 74 (2002) 857.

cr

[15] S. Bhadraa, D. Khastgir, N.K. Singhaa, J.H.Leeb, Progress in Polymer Science 34(2009)

us

783. [16] A. Malinauskas, Polymer 42 (2001) 3957.

Applications. H. Kozuka (Ed.) Springer, 1, 2005.

an

[17] Handbook of Sol-Gel Science and Technology. Processing, Characterization and

M

[18] M. Crisan, A. Jitianu, D. Crisan, at al., Journal of Optoelectronics and Advanced Materials 2 (2000) 339.

ed

[19] I.V. Barsukov, M.A. Gallego, J. E. Doninger, J. Power Sources. 153 (2006) 288.

(2005) 209.

pt

[20] J. Vondrak, B. Klapštˇe, J. Velicka, J. of New Materials for Electrochemical Systems 8

Ac ce

[21] Mao-Cheng Liu, Ling-Bin Kong, Chao Lu, Xiao-Ming Li, Yong-Chun Luo, Long Kang Electrochem. Soc. 159 (2012) A1262.

19

Page 19 of 44

Captions to figures Fig.1. Subsequent electrochemical synthesis of PANI at the Pt electrode in 1 М solution of sulfuric acid at the potential sweep rate of 50 mV/s (a – 1-3 cycles; b – 4 -15 cycles)

CVA cycles in 1 М H2SO4 solution at a potential sweep rate of 50 mV/s

ip t

Fig. 2. Morphology of subprimary PANI structures produced in the first (a) and second (b)

Fig. 3. Microphotographs of PANI (a) and PPy (b) chemically synthesized by the techniques

cr

in 2.1.2

us

Fig. 4. Current vs ω1/2 (rps) for the PPy/C (a) and PANI/C (b) composites in 1M H2SO4 Fig. 5. The scheme of ORR at the surface of PPy/C and PANI/C composites

an

Fig. 6. The expected scheme of ORR at the surface of ECP/C composites after introducing some transition metal oxide MetOx to such composites

M

Fig. 7. Relative mass loss for acetylene black (1), activated carbon (2) and graphitized

ed

PureBlack® (3) specimens under heating in the air from 100 to 600 °С Fig. 8. Cyclic voltammetric curves for the graphitized PureBlack® at different potential

pt

sweep rates 1 - 100, 2 - 40, 3 - 10 mV/s in 1M NaOH solution saturated with argon Fig. 9. Cyclic voltammetric curves in 1M NaOH solution saturated with argon (Ar) and with

Ac ce

oxygen at different rotation speeds of the disc electrode (rpm) on the following materials: (a) PureBlack®; (b) “PureBlack®-NiOx”; (c) “PureBlack®-CoOx”; (d) “PureBlack®-NiCo2O4” Fig. 10. Koutecky-Levich plots for oxygen reduction for the following materials при потенциале -0.5 V: 1 − PureBlack®; 2 − “PureBlack®-CoOx” composite ; 3 − “PureBlack®NiOx” composite ; 4 − “PureBlack®-NiCo2O4” composite Fig. 11. Cathode curves of the 5% NiCo2O4/PureBlack® composite produced by thermal decomposition (1), sol-gel method with using acetic acid (2); sol-gel method with using propionic acid (3) in 1 M NaOH under the scan rate 10 mV/s

20

Page 20 of 44

Fig. 12. XRD pattern of NiCo2O4/PureBlack® synthesized by thermal decomposition of graphite (а) and sol-gel method using propionic acid (b). Fig. 13. Cathode curves of the NiCo2O4/PureBlack® composites in 1 M NaOH at different content of NiCo2O4.

ip t

Fig. 14. Cathode curves of the PPy/NiCo2O4/PureBlack® (1), NiCo2O4/PureBlack® (2), PPy/PureBlack (3) composites in 1 M NaOH under the scan rate 10 mV/s.

cr

Figure 15. Cathode curves of the PPy/5%NiCo2O4/ PureBlack® composite in 1 M NaOH at

us

different PPy content

Fig. 16. Koutecky-Levich plots of oxygen reduction for the 15% PPy / 5% NiCo2O4 /

Ac ce

pt

ed

M

an

PureBlack® composite catalyst at different potentials in 1 М NaOH solution

21

Page 21 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig 1a

Page 22 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig 1b

Page 23 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 2a

Page 24 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 2b

Page 25 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 3a

Page 26 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 3b

Page 27 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig4a.tif

Page 28 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig4b.tif

Page 29 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 5

Page 30 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 6

Page 31 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 7

Page 32 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig. 8

Page 33 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig9a_New.tif

Page 34 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig9b_New.tif

Page 35 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig9?.tif

Page 36 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig9d_New.tif

Page 37 of 44

Ac

ce

pt

ed

M

an

us

cr

i

#10.tif

Page 38 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig-11.tif

Page 39 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig-12.tif

Page 40 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig-13.tif

Page 41 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig-14.tif

Page 42 of 44

Ac

ce

pt

ed

M

an

us

cr

i

fig-15.tif

Page 43 of 44

Ac

ce

pt

ed

M

an

us

cr

i

Fig_16.tif

Page 44 of 44