Synergistically enhanced oxygen reduction activity of iron-based nanoshell carbons by copper incorporation

Accepted Manuscript Synergistically enhanced oxygen reduction activity of iron-based nanoshell carbons by copper incorporation Takafumi Ishii, Takuya Maie, Mikiya Hamano, Takeaki Kishimoto, Mayumi Mizushiri, Yasuo Imashiro, Jun-ichi Ozaki PII:

S0008-6223(17)30163-X

DOI:

10.1016/j.carbon.2017.02.035

Reference:

CARBON 11752

To appear in:

Carbon

Received Date: 12 October 2016 Revised Date:

9 February 2017

Accepted Date: 11 February 2017

Please cite this article as: T. Ishii, T. Maie, M. Hamano, T. Kishimoto, M. Mizushiri, Y. Imashiro, J.-i. Ozaki, Synergistically enhanced oxygen reduction activity of iron-based nanoshell carbons by copper incorporation, Carbon (2017), doi: 10.1016/j.carbon.2017.02.035. 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.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Graphical abstract

ACCEPTED MANUSCRIPT

Synergistically enhanced oxygen reduction activity of iron-based

RI PT

nanoshell carbons by copper incorporation

Takafumi Ishii1, Takuya Maie1, Mikiya Hamano1, Takeaki Kishimoto2, Mayumi Mizushiri2,

International Research and Education Center for Element Science, Faculty of Science and

M AN U

1

SC

Yasuo Imashiro2 and Jun-ichi Ozaki*1

Technology, Gunma University, 1-5-1, Tenjin-cho, Kiryu, Gunma 376-8515, Japan Nisshinbo Holdings Inc., 1-2-3 Onodai, Midori-ku, Chiba, Chiba, 267-0056, Japan

AC C

EP

TE D

2

*

Corresponding author. Tel: 81 277-30-1350. E-mail: [email protected] (Jun-ichi Ozaki)

1

ACCEPTED MANUSCRIPT Abstract This study reports the synergistic improvement of the oxygen reduction reaction (ORR) activity of a carbon catalyst prepared from a phenol–formaldehyde resin by simultaneous incorporation

RI PT

of Cu and Fe phthalocyanines. This carbon catalyst exhibited higher ORR activities than those prepared with either Cu or Fe phthalocyanines, with an approx. 28-fold maximum ORR current enhancement observed when Cu/Fe=75:25 was incorporated into the carbon precursor. Surface

SC

and structural characterizations of these catalysts suggested that the incorporated Cu played a

M AN U

role in suppressing the catalytic activity of Fe, forming the reported nanoshell structure, and in keeping more nitrogen atoms attached to carbons. The incorporation of Cu in the carbon catalyst provided a doping effect for nitrogen into the carbon structure and inhibited catalytic graphitization caused by Fe species. These effects resulted in a distinctive carbon structure in

TE D

the nanoshell. These multiple effects of Cu incorporation were considered key factors in the

AC C

EP

enhancement of ORR activity caused by Fe and Cu incorporation.

2

ACCEPTED MANUSCRIPT 1. Introduction The development of a high-performance Pt-free electrocatalyst to accelerate the oxygen reduction reaction (ORR) is crucial for the widespread commercialization of polymer

RI PT

electrolyte fuel cells (PEFCs). Carbon catalysts are some of the most promising candidates for cathode catalysts [1]. Much effort has been directed towards developing a high-performance carbon catalyst by using various transition metals, such as Fe, Co, and Ni [2-7], with the

SC

intention to produce metal–nitrogen complexes on the carbon surfaces.

It is widely accepted that carbon structures vary widely in the presence of transition

M AN U

metals that alter the carbonization processes [8]. Peng et al. reported that the effects of transition metals on these structures, including the nitrogen content, vary depending on the metal used [9]. In our previous study [10], a distinctive structure was observed in carbon derived from a phenol–formaldehyde resin with a transition metal species, which was produced

TE D

by the catalytic action of the incorporated transition metal species altering the carbonization process. The distinctive carbon structure was shell-like, consisting of a few dozen graphene

EP

sheets, and referred to as a nanoshell. These prepared carbons were therefore named nanoshell-containing carbons (NSCCs), of which some had high catalytic activities for the

AC C

ORR. Our previous studies investigated NSCCs produced using one metal, such as Fe or Co [10-14]. Among them, Fe-NSCC, which was derived from a phenol–formaldehyde resin with Fe phthalocyanine, showed excellent ORR activity [14]. The ORR activities of NSCCs were not only influenced by the metal species, but also by the presence of doped nitrogen atoms. In order to maximize the ORR activities of NSCCs, tuning the carbon structure and nitrogen doping levels is critical [14]. Simultaneous incorporation of different transition metals into the preparation would

3

ACCEPTED MANUSCRIPT be a useful method to produce NSCC while controlling the structure and nitrogen content. Although such a cooperative approach would still require further detailed investigation because interactions between doped metals are complicated, it is considered to have potential as a

RI PT

pathway to optimizing and enhancing catalytic activity. Copper has attracted attention as a metallic element that promotes surface catalytic action and deposits graphene on its films or particles under certain deposition conditions [15], whereas it does not cause catalytic

SC

carbonization [8]. The simultaneous incorporation of Fe and Cu into the preparation of NSCC has not yet been examined. In this paper, we report a study conducted to incorporate another

M AN U

additive, Cu, to the abovementioned precursor of Fe-NSCC, to control the carbon structure and nitrogen doping level. We also attempted to elucidate the rules of ORR activity provided by nanoshell and nitrogen atoms in NSCC prepared by the simultaneous incorporation of Fe and

2. Experimental

EP

2.1 Sample preparation

TE D

Cu.

Carbon precursors were obtained by incorporating Fe phthalocyanine (FePc) and Cu

AC C

phthalocyanine (CuPc) into a phenol–formaldehyde resin, where the Cu ratio in the total amount of Cu and Fe varied from 0 to 1, by fixing the total amount of metal species at 5 wt.%. The precursors were carbonized at 800 °C for 1 h in a N2 stream, followed by sequential ball milling and washing with HCl solution. The acid wash was included to remove residual metal species from the surface of the carbon catalysts. Hereafter, carbons prepared using only FePc or CuPc are referred to as 100Fe-C and 100Cu-C, respectively, while carbons prepared using CuPc and FePc together are referred to as xCu-Cs, where x is the mixing ratio (0 < x < 100) of

4

ACCEPTED MANUSCRIPT the two metals. Controls were also prepared by carbonizing the precursors, in which CuPc was replaced with metal-free phthalocyanine (MfPc). The molar amount of MfPc was adjusted according to that of CuPc used in xCu-Cs, and the controls were named as xMf-Cs, where x

RI PT

stands for the value as used in that of xCu-Cs.

2.2 Electrochemical measurements

SC

The catalytic activity of the prepared carbons in the ORR was evaluated by linear sweep voltammetry employing an oxygen saturated 0.5 mol dm–3 H2SO4 aqueous solution. The

M AN U

working electrode was prepared by covering the glass-like carbon disk electrode with a mixture of the prepared carbon and Nafion®, where the amount of loaded carbon sample was kept to 200 µg cm–2. Voltammograms were obtained by sweeping the potential from 1.0 V vs. reversible hydrogen electrode (RHE) to 0 V vs. RHE at 1 mV s–1 by rotating the electrode at 400 to 2500

TE D

rpm. Double layer capacitance was compensated for by subtracting the voltammogram obtained in nitrogen-saturated electrolyte from the obtained ORR voltammogram.

EP

The current observed in rotating electrode voltammetry consisted of kinetic current (ik) and diffusion current (id). The kinetic current, ik, can be separated by applying the

AC C

Koutecky–Levich equation to the obtained voltammograms, given in Figure S1. The plots of 1/i against the square root of rotation speed for the samples prepared here exhibited a linear relationship, as shown in Figure S2. Extrapolating the rotating speed to infinity gave the kinetic current, ik. The number of participating electrons, n, was calculated from the slope of the plots using DO2 = 1.40×10–5 cm2 s–1, υ = 1.00×10–2 cm2 s-1 from ref. [16] and CO2 = 1.20×10–6 mol cm–3 given by our experimental determination using optical oxygen meter (FireStingO2, Pyro science GmbH).

5

ACCEPTED MANUSCRIPT

2.3 Characterization techniques Samples were subjected to transmission electron microscopy (TEM) observation and

RI PT

X-ray diffraction (XRD) analysis to investigate their carbon structures. For TEM observation, the sample was ground in ethanol and sprayed over a Cu microgrid. Sample lattice fringes on the microgrid were observed using a 200-kV transmission electron microscope (JEM-2010,

SC

JEOL Co.). XRD analysis was performed using an X-ray diffractometer (XRD-6100, Cu-Kα, Shimadzu Co.) with Cu Kα radiation. The surface areas and pore structures of the samples were using

nitrogen

adsorption–desorption

measurements

M AN U

evaluated

(BELSORP

mini,

MicrotracBEL Co.) at –196 °C. Surface composition analyses of the elements contained in the samples were performed using X-ray photoelectron spectroscopy (XPS, AXIS NOVA,

TE D

Shimadzu Co.).

3. Results

EP

3.1 Electrochemical properties of carbon catalysts Figure 1a shows the ORR voltammograms of the prepared carbon catalysts obtained

AC C

in an oxygen-saturated 0.5 mol dm–3 H2SO4 aqueous solution. ORR voltammograms defined by kinetic current ik are also shown in Figure 1b. The voltammograms had different onset 0.0

0.0

(a)

-2

75Cu-C

-3

-5.0x10

50Cu-C

100Mf-C 50Mf-C 0.0

-5.0x10

-3.0x10

-1.5x10

-2.0x10

-4.0x10

-3

0.0

0.2

0.4

100Mf-C

-1.0x10

50Mf-C

-2

0.0

-5

-5.0x10

-3

-1.0x10

25Cu-C

k

-2.0x10

50Cu-C

100Fe-C

/ A cm

100Fe-C -3

75Cu-C

-3

-2

25Cu-C

100Cu-C

i

Current density / A cm

-1.0x10

(b)

100Cu-C

-1.5x10

-4

-1.0x10

-4

-1.5x10

-4

0.5

0.6

0.7

0.6

Potential / V vs. RHE

0.8

0.8

0.9

-2.0x10

1.0

-2.0x10

1.0

-5

-2

-2

0.0

0.2

0.4

-4

-4

-4

0.5

0.6

0.7

0.6

0.8

0.8

0.9

1.0

1.0

Potential / V vs. RHE

Figure 1. (a) ORR voltammograms of the prepared carbon catalysts obtained at a rotating 6 speed of 1600 rpm and (b) voltammograms represented in terms of ik.

ACCEPTED MANUSCRIPT potentials and shapes depending on the types of metal species. ORR activity is represented by the onset potential, EO2, as defined by the potential at ik = –10 µA cm–2 (shown in Figure 1b), as used in our previous studies [11, 14]. The EO2 values of the samples are lower than that of Pt/C

RI PT

electrocatalyst (EO2 = ~0.95 V) but to examine their activities is important for understanding catalytic actions occurred on the surface of carbon catalysts. The dependence of EO2 on the additive ratio is presented in Figure 2. Both ends of the plot corresponded to the ORR activities

SC

of 100Fe-C and 100Cu-C or 100Mf-C, respectively. From comparison of the activities of 100Fe-C, 100Cu-C and 100Mf-C, the effectiveness of metal species in introducing ORR

M AN U

activities to carbon materials was in the following order: Fe > Cu > Mf. Carbon catalysts prepared with the simultaneous application of FePc and CuPc, xCu-Cs, resulted in

TE D

enhancement in the ORR activity, which were larger than those observed for xMf-Cs.

xCu-C

AC C

EP

xMf-C

Figure 2. Additive ratio dependence of ORR activity, represented by EO2.

As shown in Figure 1a, the voltammogram shapes for 100Fe-C and 100Cu-C were different. 100Fe-C showed an almost linear increase in ORR current density in the negative direction, even showing a higher EO2 than 100Cu-C. However, in contrast, 100Cu-C showed saturation in the potential region below 0.1 vs. RHE. The xCu-Cs voltammogram shapes were

7

ACCEPTED MANUSCRIPT

(b)

(a)

xCu-C xMf-C

xCu-C

RI PT

xMf-C

number of electrons participating in the ORR.

SC

Figure 3. Additive ratio dependence of (a) ORR activity represented by ik,0.8 and (b) the

M AN U

similar to that of 100Cu-C rather than 100Fe-C, which had onset potentials, EO2s, higher than 100Cu-C, as mentioned above. Saturation of the ORR current indicates that the charge-transfer rate exceeds the diffusion rate in an electrochemical process, while the potential-independent diffusion process determines the whole rate of the process. The absence of saturation in the

TE D

current density of 100Fe-C indicated that the specimen possessed active sites with higher activity than 100Cu-C, but that the number of sites was not enough to overcome the diffusion

EP

rate. The ORR currents of xCu-C and xMf-C showed diffusion controlled trends in the potential region <0.2 V vs. RHE. These trends meant that the simultaneous introduction of FePc and

AC C

CuPc/MfPc produced active sites with higher ORR activities, and in larger numbers than in carbons prepared with single metal species. Figure 3a shows a plot of the kinetic current density at 0.8 V (ik,0.8) against the

additive ratio. The ik,0.8 values of xCu-Cs were higher than those of 100Fe-C and 100Cu-C. In other words, the points for xCu-Cs were located above the line, connecting the points for 100Fe-C and 100 Cu-C. Thus, the simultaneous application of FePc and CuPc to the preparation of carbon catalysts induced synergistic enhancements of ORR activity. The

8

ACCEPTED MANUSCRIPT maximum ik,0.8, which was 28 times of 100Fe-C, was obtained for 75Cu-C. Synergistic enhancement of ORR activity was also observed for xMf-C, as shown in Figure 3a. However, the extent of synergistic enhancement observed for xMf-Cs was smaller than that observed for

RI PT

xCu-Cs. Figure 3b shows the dependence of the number of participating electrons, n, on the additive ratio. The n value of xCu-C was almost 4 which is higher than that of 100Cu-C (n = 3.2) and 100Fe-C (n = 3.0). A similar trend was also observed for xMf-Cs, indicating that the

SC

increase in n was caused by some additive effect other than the introduction of Cu species. Tafel plots are a useful method to understand the detailed mechanisms of the ORR

M AN U

[17, 18]. Figure S3 shows the Tafel plot of ik vs. potential. Tafel slopes were calculated from the linear portions of the plot, and are presented in Figure 4 as a function of the additive ratio. The Tafel slopes of 100Fe-C and 100Cu-C were ~60 mV decade–1 and ~120 mV decade–1,

TE D

respectively, while the same for xCu-Cs were similar to those of 100Fe-C, at ~60 mV decade–1.

AC C

EP

xCu-C

xMf-C

Figure 4. Relationship between Tafel slopes and additive ratio in both xCu-C and xMf-C catalysts.

9

ACCEPTED MANUSCRIPT 3.2 Surface chemical properties of carbon catalysts

SC

xMf-C

RI PT

xCu-C

Figure 5. Additive ratio dependence of N/C surface atomic ratio obtained from XPS

M AN U

measurement.

Figure 5 shows the dependence of the nitrogen-to-carbon (N/C) atomic ratio on the Cu additive ratio. The N/C ratio increased with the Cu ratio, unlike the trend of ORR activity, which showed a maximum at Cu/Fe=75:25, as shown in Figure 3a. The dependence of the

TE D

ORR activities of xMf-Cs is also shown in the plot. In this case, an increasing N/C ratio was also observed with increasing Mf-additive ratio, although the increments were smaller than

EP

those of xCu-Cs. The increase in N/C ratio with increasing additive ratios observed in both cases can be interpreted as being due to the dilution of Fe species, as reported for

AC C

denitrogenation catalysts [19]. The increase in nitrogen content in xCu-C indicated that Cu had weakened the denitrogenation activity of Fe. In other words, Cu spontaneously dopes N atoms into the resultant carbons. This hypothesis was consistent with the previous findings of Peng et al. [9]. They investigated the effects of transition metal incorporation into the structure of carbon materials derived from polyaniline and melamine. They demonstrated that Cu had a positive effect on nitrogen doping in the materials compared with the other transition metals (Mn, Fe, Co, and Ni).

10

ACCEPTED MANUSCRIPT High-resolution N1s XPS spectra of the carbon catalysts indicated the presence of different types of nitrogen species, which were fitted with four peaks corresponding to pyridine (398.5 eV), pyrrole/pyridone (400.5 eV), quaternary (401.5 eV), and oxidized nitrogen (403

RI PT

eV) [20]. To quantitatively investigate the amounts of different types of nitrogen species on the surface, the relative percentages of these types were calculated based on deconvoluted N1s spectra. Table 1 shows the amounts and types of nitrogen groups on the catalyst surfaces. As

SC

has mentioned the total amounts of nitrogen did not obey the ORR trend shown in Figure 3a. Also none of the four types of nitrogen species showed similar trends in ORR activity. Hence,

M AN U

we considered that there is no direct influence of specific nitrogen species on the activity. Metal species can also be an influential factor in determining ORR activity, and there have been a number of discussions of the active sites, including Fe or Co atoms coordinated to surface nitrogen atoms [21, 22]. Table 1 also shows the atomic ratios of metal species, Fe/C,

EP

TE D

and Cu/C, determined by XPS. Although Fe/C and Cu/C were several tenths of the N/C ratio

Nitrogen species / at.% *

AC C

Samples

Metal species / at.%

Pyridine

Pyrrole/Pyridone

Quaternary

Oxide

Fe/C

Cu/C

100Fe-C

0.2

0.2

0.2

0.1

0.05

0.00

25Cu-C

0.4

0.3

0.2

0.1

0.07

0.02

50Cu-C

0.7

0.6

0.3

0.2

0.09

0.00

75Cu-C

1.3

1.2

0.5

0.3

0.09

0.02

100Cu-C

2.0

1.6

0.5

0.5

0.00

0.00

50Mf-C

0.5

0.3

0.3

0.0

0.06

0.00

*Atomic ratio of the specie to carbon atom determined from following formula;
/ × (fraction of the species) × 100.

11

ACCEPTED MANUSCRIPT shown in Figure 5, and scattering of the values was appreciably large. These trends proved the nitrogen atoms and the metal atoms detected by XPS did not participate in the ORR occurring on the surfaces of xCu-Cs. The ORR activity of carbon alloys depended on the

catalysts [14].

SC

3.3 Physical structure of carbon catalysts

RI PT

of metals forming the nanoshell structure and the presence of nitrogen atoms in the carbon

N2 adsorption, XRD analysis, and TEM observations were carried out to investigate

M AN U

the structure of the carbon catalysts. The porosities of the carbon catalysts were assessed by nitrogen adsorption–desorption analysis. The calculated Brunauer–Emmett–Teller (BET) surface areas are listed in Table 2. 25Cu-C had the highest specific surface area among the prepared samples, but no distinct trend was observed between Cu-additive ratio and

TE D

development of porosity. SBET and the pore size distribution did not show any similar dependences to that of the ORR with respect to the additive ratio (see Figure S4 and S5). The

EP

development of porosity evaluated by the N2-adsorption study had no relation to the improvements in ORR activity caused by the incorporation of Cu species.

AC C

Table 2. BET surface area from nitrogen adsorption–desorption analysis. 2

Sample

SBET [m /g]

100Fe-C

315

25Cu-C

363

50Cu-C

344

75Cu-C

282

100Cu-C

303

12

ACCEPTED MANUSCRIPT Figure 6a shows the XRD profiles of the carbon catalysts. The profiles were variable depending on the contents and types of metal species incorporated in the precursor. The coexistence of sharp and broad C (002) peaks at around 2θ=26° were observed only in

RI PT

the profiles of carbon catalysts derived from precursors containing FePc. The intensity of the sharp peak increased with decreasing Cu additive ratio, which was attributed to the relatively developed carbon structure of the nanoshell. These multiple peaks indicated the

SC

inhomogeneity of this material, which consisted of nanoshell carbon and amorphous carbon phases, as discussed in previous papers [14]. Note that the 002 peak position of 100Cu-C was

M AN U

higher than that of 100Mf-C, with even these catalysts exhibiting broad diffraction peaks. This observation indicated that Cu had an ordering effect on the carbon structure during carbonization. Our previous study reported that the degree of nanoshell carbon development could be expressed by a parameter, fsharp, obtained by analyzing the 002 X-ray diffraction of

TE D

(a)

(b) xMf-C xCu-C

AC C

EP

C (002)

(c)

Broad peak Sharp peak

Figure 6. (a) XRD patterns and (b) degree of NS development, fsharp, calculated from deconvolution of the 002 peak shown in (c).

13

ACCEPTED MANUSCRIPT carbons. To investigate the structure quantitatively, the fsharp parameter was employed to express the developmental degree of the nanoshell in the carbon structure, which was obtained by deconvolving the 002 peak as shown in Figure 6c (for more detail, see ref [14]).

RI PT

Figure 6b shows the dependence of fsharp on the Cu additive ratio. The fsharp value decreased with Cu additive ratio due to the decrease in Fe content, which indicated the action of Cu to mitigate the effect of Fe in promoting the nanoshell formation. Figure 6b also presents the

SC

plots for xMf-Cs. The fsharp values of xCu-Cs were lower than those of xMf-Cs with the same Fe ratio. The mitigation effect of Cu when forming graphitic structures was also reported for a

M AN U

Cu-enriched Ni alloy (Ni0.05Cu0.95) used to catalyze carbon nanofiber formation [23]. Development of carbon structure with the increase in Fe content was observed in the Raman spectra (Figure S6), which agrees with the XRD results considering that the development is

AC C

EP

TE D

due to the nano-shell formation.

14

ACCEPTED MANUSCRIPT Figure 7 shows TEM images of 100Fe-C, 100Cu-C, and 50Cu-C. This confirmed the formation of the nanoshell structure in 100Fe-C, while confirming the formation of a

AC C

EP

TE D

M AN U

SC

RI PT

dot-like structure in 100Cu-C characteristic of amorphous carbons. Formation of the

15

ACCEPTED MANUSCRIPT nanoshell was considered not to occur in the presence of Cu, due to its lower activity in catalytic carbonization compared with that of Fe [8]. The TEM images of xCu-Cs showed that

AC C

EP

TE D

M AN U

SC

RI PT

the amount of nanoshell (“B” in Figure 7d) becomes smaller as the Cu/Fe ratio increases and

16

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 7. TEM images of the samples: (a,b) 100Fe-C, (c,d) 50Cu-C, and (e, f) 100Cu-C. The areas denoted “A” and “B” in (d) indicate the typical portions to the newly formed structure and the nanoshell structure, respectively.

xCu-Cs were revealed to have different structures, as shown in the enlarged TEM image (“A” in Figure 7d), than those assigned to nanoshell and amorphous structures found in 100Fe-C and 100Cu-C, respectively. This newly found structure was observed for 50Cu-C and

17

ACCEPTED MANUSCRIPT 75Cu-C.

RI PT

4. Discussion The result in Figure 3a showed that the ORR activity increased with additive ratio up to 75%, and then decreased at additive ratios above 75%. Such a relationship is usually

SC

interpreted as the competition of two opposite effects, i.e., a promoting effect and an opposing inhibiting effect. Both N-doping and nanoshell formation contribute to enhancing the ORR

M AN U

activity of the carbon catalysts. We can tentatively assign the promoting and inhibiting effects on ORR activity to the increase in the N-doping level (Figure 5) and decrease in the degree of nanoshell development, denoted by fsharp (Figure 6b), respectively, due to the following considerations. Many studies on carbon-based ORR catalysts have revealed the promoting

TE D

effect of N-doping [11, 24-30]. We reported the superior ORR activity of N-doped NSCCs over undoped NSCCs, with the former prepared from metal phthalocyanines and the latter

EP

from metal acetylacetonates. These previously reported results elucidated an increase in the nitrogen doping level. The degree of development of the nanoshell structure, fsharp, also

AC C

influenced the ORR activity. In the region of fsharp < 0.2, the ORR activity increased with fsharp. Hence, the decrease in fsharp can be regarded as inhibiting the ORR activity, causing a decrease at additive ratios higher than 75%, as shown in Figure 3a. This study revealed a synergistic enhancement of the ORR activity of carbon catalysts prepared by carbonization of phenol–formaldehyde resin containing both FePc and CuPc, because these catalysts had higher ORR activities than carbon catalysts prepared from single-metal phthalocyanines. Possible explanations for this synergistic enhancement are the

18

ACCEPTED MANUSCRIPT increase in the number of active sites and the formation of new active sites due to the simultaneous application of FePc and CuPc. As described in the Results section, xCu-Cs exhibited higher EO2 values and

RI PT

saturating currents below 0.2 V vs. RHE. This saturation of the ORR current meant a change in the rate determining step (rds). In a sufficiently low potential region (a high overpotential region) charge transfer processes in the ORR are accelerated, obeying the Butler–Volmer

SC

equation and exceeding the rate of diffusion, resulting in the diffusion process controlling the overall rate of ORR. Thus, the saturated ORR current in the lower potential region observed

M AN U

for xCu-Cs indicated that the carbon catalysts should contain sites with higher activities and of sufficient abundance.

The Tafel slope is a parameter to discriminate the rds in a series of electrochemical reactions. Yeager proposed a mechanism for the ORR occurring on carbon surfaces [17],

TE D

claiming that an oxygen molecule was adsorbed on a carbon surface, followed by a single-electron transfer to produce a superoxide anion; this anion then moves from an inactive

EP

site to an active site to undergo further reduction reactions. (1)

O2(ads) + e– → (O2–)

(2)

(O2–) →[O2–]

(3)

AC C

O2 → O2(ads)

In the equations above, the two types of parentheses, () and [], represent inactive and active sites for the adsorption of O2– species on the carbon surface, respectively, step (2) describes the first electron transfer, and step (3) corresponds to the O2– transfer step. A theoretical expression of the Tafel slope, A, is given by the following formula [31, 32]:   = 2.303   + !"# $ 19

ACCEPTED MANUSCRIPT where R, T, γ, ν, r, β, and F are gas constant, the absolute temperature, number of electrons transferred before the rds, stoichiometric number of the reaction, number of electrons transferred in the rds, symmetry factor of the rds, and the Faraday constant, respectively.

RI PT

According to Taylor et al. [33], assuming the ν and β to be 1 and 0.5 in the elementary process of the reaction, a Tafel slope of 120 mV decade–1 indicates that the rds of the ORR was the first electron transfer step as shown in equation (2) where the γ and r are determined as 0 and

SC

1, respectively. While a 60 mV decade–1 Tafel slope indicated that the rds was the anion transfer process, moving O2– from the previous inactive site to the next active site, as shown

M AN U

in equation (3) where the γ and r are 1 and 0, respectively.

Based on these interpretations of the Tafel slopes in ORR, the rds of 100Fe-C and xCu-C was the anion-transfer process between two sites, while the rds of 100Cu-C was the first electron transfer process. Namely, 100Fe-C and xCu-C have active sites with a similar

TE D

catalytic nature. Hence, the synergistic enhancement in xCu-C could be ascribed to an increase in the number of active sites. These carbon catalysts commonly contained nanoshell

EP

carbons. In our previous study, the nanoshell structure have been revealed to accelerate the electron transfer process in the redox reaction of Fe(CN)63–/Fe(CN)64– [34, 35]. Therefore, it

AC C

was natural to assume that the first electron transfer process in ORR could also be accelerated by nanoshell carbons. Thus, we rationalized that the rds of the ORR occurring on 100Fe-C and xCu-C was the process of anion-transfer from the inactive site to active site. On the other hand, 100Cu-C, without nanoshell carbons, had a Tafel slope of 120 mV decade–1, indicating that the rds using this catalyst was the first electron transfer process. The N/C atomic ratio increased with the additive ratio, while fsharp decreased. These two opposing trends explained the synergistic effect that gave a maximum at Cu/Fe=75:25.

20

ACCEPTED MANUSCRIPT The discussion in the previous paragraph indicated two possible rate determining steps, i.e., the first electron transfer step and the anion-transfer step, and that the former step could be accelerated by the presence of nanoshell carbons. Accordingly, the presence of nitrogen atoms

RI PT

should affect the anion-transfer step. Indeed, the above assumption can be supported by the following fact that xMf-Cs, which included nanoshell structures, but lower N/C ratios than xCu-Cs, showed lower ORR activity than xCu-Cs. As mentioned in the results section, there

SC

was no correlation between the type of nitrogen introduced to the carbon catalysts and the ORR activity. Recently, Nakamura et al. stated the importance of pyridinic nitrogen in ORR

M AN U

activity using well-defined surface chemistry [36]. Many researchers have made efforts towards specifying the types of nitrogen responsible for catalyzing ORR. In our case, the absence of a correlation between the types of nitrogen and ORR activity indicated that the presence of nitrogen species did not directly participate in catalyzing the ORR. Finally, we

TE D

concluded that the synergistic increase of ORR activity can be explained by the two opposite factors with the additive ratio, i.e. fsharp and N/C, the former influences the electron transfer

AC C

5. Conclusions

EP

step and the latter the anion transfer step, respectively.

To control the carbon structure and nitrogen doping level of NSCCs, simultaneous

incorporation of Fe and Cu in their preparation was examined. NSCCs prepared with Cu and Fe, xCu-Cs, exhibited higher ORR activity than NSCCs prepared with only Fe or Cu. Their ORR activity did not correspond with the surface concentrations of nitrogen and metal species. The incorporation of Cu in the precursor to NSCC induced two effects: the inhibition of nanoshell formation by diminishing the activity of Fe for catalytic graphitization, and the

21

ACCEPTED MANUSCRIPT promotion of doping nitrogen to the carbon structure. As expected, a synergistic enhancement of ORR activity was observed in the xCu-Cs, which was explained as being due to two kinds of accelerative effect from the first electron transfer step and the anion-transfer step in ORR.

RI PT

In this study, we tentatively assigned the nanoshell structure and the nitrogen species to the factors enhancing the electron transfer and the anion transfer steps, respectively. Further study

SC

should be need in order to clarify the above assumptions.

Acknowledgements

M AN U

We are grateful for the partial financial support from Advanced Low Carbon Technology Research and Development Program (ALCA). A part of this work was supported by NIMS microstructural characterization platform as a program of “Nanotechnology Platform” of the

TE D

MEXT, Japan (proposal A-16-NM-0061).

REFERENCES

Banham D, Ye S, Pei K, Ozaki J, Kishimoto T, Imashiro Y. A review of the stability

EP

[1]

AC C

and durability of non-precious metal catalysts for the oxygen reduction reaction in proton exchange membrane fuel cells. Journal of Power Sources. 2015;285:334-48. [2]

Chung HT, Johnston CM, Artyushkova K, Ferrandon M, Myers DJ, Zelenay P.

Cyanamide-derived non-precious metal catalyst for oxygen reduction. Electrochemistry Communications. 2010;12(12):1792-5. [3] Catalysts

Gupta S, Tryk D, Bae I, Aldred W, Yeager E. Heat-Treated Polyacrylonitrile-Based for

Oxygen

Electroreduction.

Journal

of

Applied

Electrochemistry.

1989;19(1):19-27.

22

ACCEPTED MANUSCRIPT [4]

Jahnke H, Schönborn M, Zimmermann G. Organic dyestuffs as catalysts for fuel

cells. In: Schäfer FP, Gerischer H, Willig F, Meier H, Jahnke H, Schönborn M, et al., eds. Physical and Chemical Applications of Dyestuffs. Berlin, Heidelberg: Springer Berlin

[5]

RI PT

Heidelberg 1976, p. 133-81. Jaouen F, Proietti E, Lefevre M, Chenitz R, Dodelet JP, Wu G, et al. Recent

advances in non-precious metal catalysis for oxygen-reduction reaction in polymer electrolyte

SC

fuel cells. Energy & Environmental Science. 2011;4(1):114-30.

Jasinski R. A New Fuel Cell Cathode Catalyst. Nature. 1964;201(4925):1212-3.

[7]

Lefevre M, Proietti E, Jaouen F, Dodelet JP. Iron-Based Catalysts with Improved

M AN U

[6]

Oxygen Reduction Activity in Polymer Electrolyte Fuel Cells. Science. 2009;324(5923):71-4. [8]

Ōya A, Ōtani S. Catalytic graphitization of carbons by various metals. Carbon.

1979;17(2):131-7.

Peng HL, Liu FF, Liu XJ, Liao SJ, You CH, Tian XL, et al. Effect of Transition

TE D

[9]

Metals on the Structure and Performance of the Doped Carbon Catalysts Derived From

[10]

EP

Polyaniline and Melamine for ORR Application. Acs Catal. 2014;4(10):3797-805. Ozaki J, Tanifuji S, Kimura N, Furuichi A, Oya A. Enhancement of oxygen

AC C

reduction activity by carbonization of furan resin in the presence of phthalocyanines. Carbon. 2006;44(7):1324-6. [11]

Kannari N, Ozaki J. Formation of uniformly and finely dispersed nanoshells by

carbonization of cobalt-coordinated oxine-formaldehyde resin and their electrochemical oxygen reduction activity. Carbon. 2012;50(8):2941-52. [12]

Kobayashi M, Niwa H, Saito M, Harada Y, Oshima M, Ofuchi H, et al. Indirect

contribution of transition metal towards oxygen reduction reaction activity in iron

23

ACCEPTED MANUSCRIPT phthalocyanine-based carbon catalysts for polymer electrolyte fuel cells. Electrochimica Acta. 2012;74:254-9. [13]

Ozaki J, Nozawa K, Yamada K, Uchiyama Y, Yoshimoto Y, Furuichi A, et al.

RI PT

Structures, physicochemical properties and oxygen reduction activities of carbons derived from ferrocene-poly(furfuryl alcohol) mixtures. Journal of Applied Electrochemistry. 2006;36(2):239-47.

Ozaki J, Tanifuji S-i, Furuichi A, Yabutsuka K. Enhancement of oxygen reduction

SC

[14]

activity of nanoshell carbons by introducing nitrogen atoms from metal phthalocyanines.

[15]

Li X, Cai W, An J, Kim S, Nah J, Yang D, et al. Large-Area Synthesis of

High-Quality

and

Uniform

2009;324(5932):1312-4.

Graphene

Films

on

Copper

Foils.

Science.

Wang X, Zhou J, Fu H, Li W, Fan X, Xin G, et al. MOF derived catalysts for

TE D

[16]

M AN U

Electrochimica Acta. 2010;55(6):1864-71.

electrochemical oxygen reduction. Journal of Materials Chemistry A. 2014;2(34):14064-70. Yeager E. Dioxygen electrocatalysis: mechanisms in relation to catalyst structure.

EP

[17]

Journal of Molecular Catalysis. 1986;38(1):5-25. Bockris JOM, Nagy Z. Symmetry factor and transfer coefficient. A source of

AC C

[18]

confusion in electrode kinetics. Journal of Chemical Education. 1973;50(12):839. [19]

Hayashi J-i, Kusakabe K, Morooka S, Nielsen M, Furimsky E. Role of Iron Catalyst

Impregnated by Solvent Swelling Method in Pyrolytic Removal of Coal Nitrogen. Energy & Fuels. 1995;9(6):1028-34. [20]

Raymundo-Piñero E, Cazorla-Amorós D, Linares-Solano A, Find J, Wild U,

Schlögl R. Structural characterization of N-containing activated carbon fibers prepared from a

24

ACCEPTED MANUSCRIPT low softening point petroleum pitch and a melamine resin. Carbon. 2002;40(4):597-608. [21]

Schulenburg H, Stankov S, Schünemann V, Radnik J, Dorbandt I, Fiechter S, et al.

Catalysts

for

the

Oxygen

Reduction

from

Heat-Treated

Iron(III)

RI PT

Tetramethoxyphenylporphyrin Chloride:  Structure and Stability of Active Sites. The Journal of Physical Chemistry B. 2003;107(34):9034-41. [22]

Lefèvre M, Dodelet JP, Bertrand P. Molecular Oxygen Reduction in PEM Fuel

SC

Cells:  Evidence for the Simultaneous Presence of Two Active Sites in Fe-Based Catalysts. The Journal of Physical Chemistry B. 2002;106(34):8705-13.

Podyacheva OY, Shmakov AN, Ismagilov ZR. In situ X-ray diffraction study of the

M AN U

[23]

growth of nitrogen-doped carbon nanofibers by the decomposition of ethylene-ammonia mixtures on a Ni-Cu catalyst. Carbon. 2013;52:486-92. [24]

Lyth SM, Nabae Y, Moriya S, Kuroki S, Kakimoto M, Ozaki J, et al. Carbon nitride

TE D

as a nonprecious catalyst for electrochemical oxygen reduction. Journal of Physical Chemistry C. 2009;113(47):20148-51.

Gong K, Du F, Xia Z, Durstock M, Dai L. Nitrogen-doped carbon nanotube arrays

EP

[25]

with high electrocatalytic activity for oxygen reduction. Science. 2009;323(5915):760-4. Ozaki J, Tanifuji S, Furuichi A, Yabutsuka K. Enhancement of oxygen reduction

AC C

[26]

activity of nanoshell carbons by introducing nitrogen atoms from metal phthalocyanines. Electrochimica Acta. 2010;55(6):1864-71. [27]

Chokai M, Taniguchi M, Moriya S, Matsubayashi K, Shinoda T, Nabae Y, et al.

Preparation

of

carbon

alloy

catalysts

for

polymer

electrolyte

fuel

cells

from

nitrogen-containing rigid-rod polymers. Journal of Power Sources. 2010;195(18):5947-51. [28]

Chen Z, Higgins D, Chen Z. Nitrogen doped carbon nanotubes synthesized from

25

ACCEPTED MANUSCRIPT aliphatic diamines for oxygen reduction reaction. Electrochimica Acta. 2011;56(3):1570-5. [29] and

Geng D, Chen Y, Chen Y, Li Y, Li R, Sun X, et al. High oxygen-reduction activity durability of nitrogen-doped

graphene.

Energy and

Environmental

Science.

[30]

RI PT

2011;4(3):760-4. Choi CH, Park SH, Woo SI. Binary and ternary doping of nitrogen, boron, and

phosphorus into carbon for enhancing electrochemical oxygen reduction activity. ACS Nano.

[31]

SC

2012;6(8):7084-91.

Bockris JO, Nagy Z. Symmetry Factor and Transfer Coefficient - Source of

[32]

M AN U

Confusion in Electrode Kinetics. Journal of Chemical Education. 1973;50(12):839-43. Song C, Zhang J. Electrocatalytic Oxygen Reduction Reaction. In: Zhang J, ed.

PEM Fuel Cell Electrocatalysts and Catalyst Layers: Fundamentals and Applications. London: Springer London 2008, p. 89-134.

Taylor RJ, Humffray AA. Electrochemical Studies on Glassy Carbon Electrodes .2.

TE D

[33]

Oxygen Reduction in Solutions of High Ph ]Ph Greater-Than 10). Journal of Electroanalytical

[34]

EP

Chemistry. 1975;64(1):63-84.

Ozaki J, Mitsui M, Nishiyama Y. Electrochemical Behavior of Iron-Carbon Prepared

AC C

Composites

from

Ferrocene-Poly

(furfuryl

Alcohol).

TANSO.

1994;1994(165):269-74. [35]

Ozaki J, Mitsui M, Nishiyama Y, Cashion JD, Brown LJ. Effects of ferrocene on

production of high performance carbon electrodes from poly(furfuryl alcohol). Chemistry of Materials. 1998;10(11):3386-92. [36]

Guo D, Shibuya R, Akiba C, Saji S, Kondo T, Nakamura J. Active sites of

nitrogen-doped carbon materials for oxygen reduction reaction clarified using model catalysts.

26

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

Science. 2016;351(6271):361-5.

27