N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O2 Reduction to H2O2

Article abstract of DOI:10.1016/j.chempr.2017.10.013

Electrochemical oxygen reduction (ORR) is a challenging approach for the sustainable production of hydrogen peroxide (H₂O₂) and is also a reaction of relevance in fuel-cell applications. Here, we propose an outstanding metal-free electrocatalyst for the unexpectedly selective ORR to H₂O₂, consisting of graphitized N-doped single-wall carbon nanohorns (CNHs). The catalyst can operate at acidic pH to a faradic efficiency as high as 98%, but it also shows excellent performance at either physiological or alkaline pH. Moreover, the very positive onset potentials observed at all pH values investigated (+0.40 V, +0.53 V, and +0.71 V at pH 1.0, 7.4, and 13.0, respectively), good stability, and excellent reproducibility make this material a benchmark catalyst for ORR to H₂O₂. The outstanding activity arises from a combination of several factors, such as CNH-dependent facilitation of electron delivery, suitable porosity, and a favorable distribution of the types of N atoms. The versatility of hydrogen peroxide as a chemical for a wide range of applications justifies its heavy industrial production. Hydrogen peroxide is currently produced through the anthraquinone process, which is an energy-intensive process relying on the use of a precious metal (palladium). Therefore, there is huge interest in finding alternative low-cost and low-energy synthetic schemes. Electrocatalytic processes exploiting the reduction of molecular O₂ are undoubtedly of high appeal, particularly if they can rely on electrocatalysts based on metal-free materials. The electrocatalytic reduction of O₂ is also of high interest from an energy perspective for fuel-cell development. Hence, the development of an electrocatalyst able to selectively reduce O₂ to H₂O₂ potentially has a double utility, provided that the catalytic reaction occurs at low applied overpotentials and the material possesses long-term stability and high current efficiency. Fornasiero and colleagues describe an alternative strategy for producing hydrogen peroxide through a more sustainable method than the current synthetic approaches. The strategy relies on the use of electrocatalysis, made possible by the use of a catalyst with high efficiency and selectivity toward H₂O₂ formation. The prepared material is particularly appealing because it does not contain any metal, implying a greener and cheaper synthetic scheme.

Full text of DOI:10.1016/j.chempr.2017.10.013

Article
N-Doped Graphitized Carbon Nanohorns as a
Forefront Electrocatalyst in Highly Selective O2
Reduction to H O2
2
Daniel Iglesias, Angela Giuliani,
Michele Melchionna, ...,
Francesco Vizza, Maurizio Prato,
Paolo Fornasiero
melchionnam@units.it (M.M.)
prato@units.it (M.P.)
pfornasiero@units.it (P.F.)
HIGHLIGHTS
A metal-free electrocatalyst for
2
the reduction of O is prepared
The catalysts displays high
selectivity toward the formation of
2 2
H O
The catalyst’s selectivity remains
excellent over a wide pH range
Fornasiero and colleagues describe an alternative strategy for producing
hydrogen peroxide through a more sustainable method than the current synthetic
approaches. The strategy relies on the use of electrocatalysis, made possible by
2 2
the use of a catalyst with high efficiency and selectivity toward H O formation. The
prepared material is particularly appealing because it does not contain any metal,
implying a greener and cheaper synthetic scheme.
Iglesias et al., Chem 4, 1–18
January 11, 2018 ª 2017 Elsevier Inc.
https://doi.org/10.1016/j.chempr.2017.10.013

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
2
2 2
Reduction to H O , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
Article
N-Doped Graphitized Carbon Nanohorns
as a Forefront Electrocatalyst
in Highly Selective O Reduction to H2O2
2
1
1
1,
1
2
Daniel Iglesias, Angela Giuliani, Michele Melchionna, * Silvia Marchesan, Alejandro Criado,
Lucia Nasi, Manuela Bevilacqua, Claudio Tavagnacco, Francesco Vizza, Maurizio Prato,1,2,6,*
3
4
1
4
and Paolo Fornasiero1,5,7,
*
SUMMARY
The Bigger Picture
Electrochemical oxygen reduction (ORR) is a challenging approach for the
sustainable production of hydrogen peroxide (H ) and is also a reaction of
relevance in fuel-cell applications. Here, we propose an outstanding metal-
free electrocatalyst for the unexpectedly selective ORR to H , consisting of
The versatility of hydrogen
peroxide as a chemical for a wide
range of applications justifies its
heavy industrial production.
Hydrogen peroxide is currently
produced through the
2
O
2
2
O
2
graphitized N-doped single-wall carbon nanohorns (CNHs). The catalyst can
operate at acidic pH to a faradic efficiency as high as 98%, but it also
shows excellent performance at either physiological or alkaline pH. Moreover,
the very positive onset potentials observed at all pH values investigated
anthraquinone process, which is
an energy-intensive process
relying on the use of a precious
metal (palladium). Therefore,
there is huge interest in finding
alternative low-cost and low-
energy synthetic schemes.
Electrocatalytic processes
(
+0.40 V, +0.53 V, and +0.71 V at pH 1.0, 7.4, and 13.0, respectively), good
stability, and excellent reproducibility make this material a benchmark catalyst
for ORR to H . The outstanding activity arises from a combination of several
2
O
2
factors, such as CNH-dependent facilitation of electron delivery, suitable
porosity, and a favorable distribution of the types of N atoms.
exploiting the reduction of
INTRODUCTION
2
molecular O are undoubtedly of
The importance of hydrogen peroxide (H
2
O
2
) is witnessed by its very large annual
6
high appeal, particularly if they
can rely on electrocatalysts based
on metal-free materials.
production, approaching 4 3 10 tons worldwide for many end uses. The current in-
1
dustrial manufacturing of H
2 2
O relies on the anthraquinone process, which is asso-
2
,3
ciated with low sustainability. On the other hand, plasma-assisted or metal-cata-
4
–6
lyzed
explosive reaction. A very promising approach is the utilization of a fuel-cell device
to simultaneously provide electricity and selectively carry out on-site O electrore-
at the cathode. Efficient catalysts based on noble metals
have been accurately designed to perform ORR at convenient electric potentials
2 2 2 2
direct formation of H O from H and O is possible, but it is a potentially
The electrocatalytic reduction of
2
O is also of high interest from an
2
7
energy perspective for fuel-cell
development. Hence, the
2 2
duction (ORR) to H O
8
,9
development of an electrocatalyst
and with moderate to good faradic efficiency (FE). An immediate outlook for in-
dustrial convenience points toward the use of systems not relying on expensive
precious metals, thus strongly encouraging the development of metal-free electro-
catalysts. Carbon-based materials have established themselves in recent years as
able to selectively reduce O
2
to
2 2
H O
potentially has a double
utility, provided that the catalytic
reaction occurs at low applied
overpotentials and the material
possesses long-term stability and
high current efficiency.
1
0
front runners in energy applications and in particular to replace precious metal-
1
1
based electrocatalytic conversions. A definite advantage for electrocatalytic
ORR is to perform N doping of carbon material because appropriately introduced
1
2
N atoms are efficient catalytically active sites for the reaction. For example,
N-doped carbon nanotube (CNT) arrays have been reported as very competitive cat-
1
3
alysts for ORR. However, the traditional N-doped carbon-based catalysts are
known to favor the four-electron pathway for reduction of O , generating water as
2
1
4–16
the main product.
Chem 4, 1–18, January 11, 2018 ª 2017 Elsevier Inc.
1

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Here, we report the synthesis of a catalytic system based on graphitized single-
1
7
wall carbon nanohorns (CNHs) with a high degree of N doping (g-N-CNHs),
accessed via coating and controlled annealing of polydopamine (PDA), that ex-
2 2 2
hibits an outstanding performance for the electroreduction of O to H O , over-
coming the latest precious metal- and carbon-based catalysts. ORR by this
metal-free carbon catalyst proceeds at very advantageous positive potentials in
a wide pH range (1–13) with a FE as high as 98%. Moreover, the catalyst sustains
a very stable electric current for at least 24 hr for the full range of pH values
tested. The very high performance, stability, low cost, and reproducibility
of the catalyst make this material a state-of-the-art catalyst for O
2
electro-
reduction to H , inspiring a possible alternative and viable pathway for the
2
O
2
sustainable, large-scale synthesis of H
applications.
2 2
O or for developments in fuel-cell
RESULTS
The g-N-CNH nanomaterial was synthesized through a simple multi-step proto-
col (Figure 1, top). The CNHs were first oxidized (ox-CNHs) in order to introduce
COOH groups essential for anchoring the N-graphite precursor. The ox-CNHs
were then combined with dopamine hydrochloride, which polymerized on the
surface of the CNHs in aqueous buffer (ox-CNHs@PDA). Finally, the material
ꢀ
was treated thermally at 700 C under an argon atmosphere to graphitize the
1
8
PDA while preserving the structure of the CNHs. In order to evaluate the
effect of thermal treatment on the structure, composition, and performance of
the calcined material, several other temperatures were used to calcine the
ꢀ
ꢀ
ꢀ
ꢀ
sample: 600 C, 800 C, and 900 C. The sample treated at 600 C did not show
complete graphitization of the PDA shell, and the material turned out to be
ꢀ
an insulator. The materials obtained with a pyrolysis temperature of 800 C and
ꢀ
9
7
00 C have the same type of structure as the g-N-CNH catalyst treated at
ꢀ
00 C as can be seen from the transmission electron microscopy (TEM) micro-
graphs (Figure S2) but with compositional differences that were critical to the
performance.
The morphology of the material at different stages of the synthesis was investi-
gated by TEM (Figure 1, bottom; Figure S1). No significant structural changes
were observed after oxidation of the CNHs, whereas after coating with the PDA
polymer, the carbon shell was evident. Finally, annealing produced a core-shell
structure, with a thick layer of N-doped graphitized carbon enveloping the
CNHs. TEM analysis of the final g-N-CNH material confirmed that the CNHs core
structure was not evidently altered by the thermal treatment. Energy dispersive
X-ray spectroscopy (EDX; Figure 2) confirmed the presence of the three elements
C, N, and O, with C uniformly distributed in the nanocomposite, and N and O ap-
pearing enriched and co-localized in the shell structure, as expected for the graph-
itized PDA coating.
1Department of Chemical and Pharmaceutical
Sciences, Consorzio Interuniversitario Nazionale
per la Scienza e Tecnologia dei Materiali,
University of Trieste, Via L. Giorgieri 1, 34127
Trieste, Italy
2Carbon Bionanotechnology Group, CIC
biomaGUNE, Parque Technol o` gico de San
Sebasti a` n, Paseo Miram o` n, 182, 20014 San
Sebasti a` n, Guip u` zcoa, Spain
3CNR-IMEM Institute, Parco area delle Scienze
37/A, 43124 Parma, Italy
4ICCOM-CNR, Via Madonna del Piano 10, Sesto
Fiorentino, 50019 Florence, Italy
Thermogravimetric analysis (TGA) was conducted under both N
2
and air (Figures 3A
5ICCOM-CNR Trieste Associate Unit, University
of Trieste, Via L. Giorgieri 1, 34127 Trieste, Italy
and 3B). Under N , the introduction of COOH groups made it possible to accurately
2
estimate a 7% weight loss in the ox-CNH material (in comparison with pristine
6Ikerbasque, Basque Foundation for Science,
CNHs). TGA under air showed that the graphitized materials had a lowering of the
48013 Bilbao, Spain
ꢀ
combustion temperature (400 C) with respect to both p-CNHs and ox-CNH; the
7Lead Contact
sharp combustion step also indicated higher homogeneity of g-N-CNHs than of
ox-CNHs@PDA. At all stages of the synthetic sequence there was no appreciable
residue, suggesting that metals were not present in detectable amounts. This aspect
*
Correspondence: melchionnam@units.it (M.M.),
prato@units.it (M.P.), pfornasiero@units.it (P.F.)
https://doi.org/10.1016/j.chempr.2017.10.013
2
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 1. Scheme of the Synthesis and Relative TEM Analysis
(
(
Top) Scheme of the synthetic protocol to obtain g-N-CNHs.
Bottom) TEM images of the nanostructured material at each stage of the synthesis.
was further confirmed by inductively coupled plasma optical emission spectrometry
ICP-OES) of the final sample, where no metal above 0.0005 wt % was found. In
(
agreement with TGA and ICP-OES, full X-ray photoelectron spectroscopy (XPS)
screening of the materials confirmed that no metals were detected. XPS also pro-
vided important information on the atomic hybridization of the three constituent el-
ements C, O, and N (Table 1 and Figure 3D). XPS of p-CNHs showed solely peaks
attributed to the C=C aromatic pattern in the C1s binding energy (BE) range, with
no obvious peaks relative to O or N atoms, whereas the oxidation treatment intro-
duced only a low level of O atoms (Table 1 and Figure S4) and no signal relative
to N atoms. After PDA coating, one signal in the N1s range was observed at BE
(
400.0 eV) consistent with a C–N type of bond present in the dopamine monomer.
In excellent agreement with the theoretical N/O ratio of a dopamine molecule, we
found an N:O ratio of 1:2 in the ox-CNHs@PDA material (after subtracting the
5
.49% contribution for the O atoms as a result of the first oxidation step). Annealing
of the material produced a significant change, with the appearance of two peaks in
the N1s BE range, which were assigned to an N–C bond (400.3 eV), possibly as a
result of a pyrrolic N, and a C=N bond due to pyridinic units (398.9 eV). Importantly,
no peak relative to quaternary or oxidized N atoms was found in the g-N-CNH sam-
1
9
ple, whereas in other reported materials, pyrrolic or pyridinic N atoms have been
2
0
shown to have low stability after high T treatments. Moreover, the amount of O
atoms is still considerable, in line with the EDX data, although much lower than
the fresh sample ox-CNHs@PDA. We compared these data with those of materials
ꢀ
ꢀ
obtained by increasing the pyrolysis temperature (800 C and 900 C), named g-N-
CNHs800 and g-N-CNHs900 in Table 1 and Figure S5. It can be seen that the total
content of N progressively decreases by increasing the calcination temperature, in
2
1,22
accordance with previous reports.
However, the nature of the N atoms (pyrrolic
and pyridinic) and their ratio did not vary with the increased temperature. This is of
central importance as the activity toward ORR is correlated with the N atoms, which
Chem 4, 1–18, January 11, 2018
3

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 2. Further Microscopy Investigation Confirming Composition and Structure
EDX mapping of the three elements C, N, and O and HAADF-STEM of ox-CNHs@PDA (top) and g-N-CNHs (bottom).
was lower in our experiments with g-N-CNHs800 and g-N-CNHs900 (see below), indi-
ꢀ
cating that treatment at 700 C is the best.
In the Raman spectra (Figure 3C), we found the typical signature of nanocarbons,
ꢂ1
ꢂ1
with the characteristic D band at ꢁ1,350 cm , G band at ꢁ1,588 cm , and the
ꢂ1
2
D band at ꢁ2,700 cm . Raman analysis confirmed the oxidation of p-CNHs,
because there was a definite increase in the intensity ratio I
D G
/I , a semiquantitative
parameter often used for assessing the number of defects related to the partial
2
3
disruption of the p-extended framework. We also noted that after coating with
the PDA polymer, the D, G, and 2D peaks appeared broad, and graphitization
partially restored the original sharpness.
Analysis of the textural properties of the final material g-N-CNHs by N
2
physisorp-
tion at the temperature of liquid nitrogen revealed a high Langmuir surface area of
2
4
81 m /g. The material presented considerable microporosity, with a type I
2
4
isotherm according to IUPAC recommendations (Figure S7) that presented no
3
hysteresis and with a microporous volume of 0.15 cm /g. The external surface
2
area calculated from the t-plot was very low (26 m /g) and therefore the greatest
contribution to the total surface area comes from the microporous surface area
2
(
455 m /g), as calculated from the difference between the total surface area and
2
the external surface area. In comparison, N physisorption analysis of the ox-
CNHs (Figure S7) indicated that this material is also extensively microporous,
2
but with a higher Langmuir area (928 m /g), as a consequence of the tip opening
process caused by the oxidation treatment, which makes the internal channels of
2
5
CNHs available. Hence, the covering with the graphite layer results, as expected,
in blocking of the internal space of CNHs and thus a drop in surface area. The
3
microporous pore volume was also higher in this precursor material (0.18 cm /g)
and was reduced after covering with PDA and graphitization. Overall, however,
4
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 3. Physical Characterization of the CNH-Based Materials
(
(
(
(
2
A) TGA under N flow.
B) TGA under air flow.
C) Raman of the samples at each synthetic stage.
D) XPS analysis of g-N-CNHs in the C1s, O1s, and N1s BE ranges.
the area of the final material remained high enough to guarantee efficient adsorp-
tion of the O molecule.
2
In order to gain more insight into the mechanistic aspects in terms of ORR selec-
tivity, we synthesized several reference materials. As a first set of reference mate-
rials, other carbon nanostructures such as CNTs, carbon nanocones (CNCs), and
graphene oxide (GO) were covered with PDA and annealed under the same condi-
tions. These samples were named g-N-CNTs, g-N-CNCs, and g-N-GO, respec-
tively. The Raman spectra did not differ significantly from those obtained for the
CNH-based material; they showed a narrowing of the D band after the thermal
treatment as a sign of graphitization of the polymeric PDA layer (Figure S9). TGA
profiles also closely resembled those obtained for the CNH-based materials, such
that the graphitized samples showed a lower combustion temperature than the
bare carbon nanostructures (Figure S10). Inspection of the morphology via TEM re-
vealed some differences: with the CNTs, coverage with PDA turned out to be very
homogeneous (like that observed with CNHs), whereas with CNCs and GO, amor-
phous free-standing carbon aggregates were present to a large extent (Figure S11).
Analysis of the textural properties of the three reference materials showed that they
all were predominantly microporous (type I isotherm, see Figure S8), but the
measured microporous volume was in all cases lower than that of g-N-CNHs. More-
over, both BET (Brunauer, Emmett, and Teller) and Langmuir surface areas were
lower (see Table S2).
We also synthesized a second set of reference materials in order to evaluate whether
a change in the synthetic protocol would affect the type distribution of the N atoms
and in turn the catalytic performance. Hence, in place of PDA, we used two different
Chem 4, 1–18, January 11, 2018
5

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Table 1. XPS Data for All Materials at Each Stage of the Synthetic Protocol
Sample
Core
Atomic %
99.10
0.90
–
p-CNHs
C
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
1s
O
N
Ox-CNHs
C
94.51
5.49
–
O
N
Ox-CNHs@PDA
g-N-CNHs
C
73.08
19.45
7.47
90.63
3.80
5.56
93.05
2.68
4.27
93.05
4.37
2.57
O
N
C
O
N
g-N-CNHs800
g-N-CNHs900
C
O
N
C1s
O
1s
N
1s
The atomic percentage of the three constituent atoms for the four materials was calculated by XPS
analysis.
precursors, polypyrrole and polyaniline, and the final catalysts were named
CNHs@Pyrr-700 and CNHs@PANI-700, respectively (the number 700 indicates
the annealing temperature). The materials were characterized by Raman spectros-
copy, TGA, TEM, and XPS. From the Raman spectra, similar effects to the PDA-based
materials were observed. TEM inspection revealed that indeed coverage with poly-
pyrrole produced a less defined core-shell structure than g-N-CNHs after annealing,
whereas for CNHs@PANI-700, the difference in morphology was even more stark
(
Figure S14). XPS analysis confirmed that the annealed materials present only two
types of N atoms, identified as N pyrrolic and N pyridinic (Figures S6A and S6B).
However, in contrast to the g-N-CNHs, the Npyridinic/Npyrrolic ratio changed (1.13
for g-N-CNHs, 0.85 for CNHs@pyrr-700, and 0.61 for CNHs@PANI-700), as well
as the atomic percentage of the total N atoms (5.56 for g-N-CNHs, 6.54 for
CNHs@pyrr-700, and 6.73 for CNHs@PANI-700) (Table S1).
The electrochemical behavior of g-N-CNHs at the different synthetic stages was first
evaluated by cyclic voltammetry (CV; Figure 4A) between ꢂ0.34 and +1.65 V versus
reversible hydrogen electrode (RHE) in standard conditions. CV was performed in
ꢂ1
phosphate buffer (pH 7.4) with a scan rate of 0.050 V s under N
2
. The capacitive
current of ox-CNHs@PDA was lower than that of ox-CNHs or g-N-CNHs. The PDA
layer embedding the carbon nanostructure scaffold therefore acts as an electrical
ꢀ
insulator, hindering the mobility of the charge carrier. Graphitization at 700 C re-
stores the electrical conductivity of the material. The ox-CNH CV trace is character-
ized by a very broad oxidation peak centered at +0.82 V as a result of the redox pro-
cess involving oxygen containing functional groups formed during the oxidation
treatment. g-N-CNHs exhibited a reversible electrochemical process characterized
by E1/2 = (Eap + Ecp)/2 = +0.60 V (where Eap and Ecp are the anodic and cathodic peak
6
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 4. Analysis of the Electrochemical Properties through CV Experiments
A) CVs obtained on glassy carbon electrode (GCE) modified with ox-CNHs (black line), ox-
CNHs@PDA (red line), and g-N-CNHs (green line) in 0.10 M PBS (pH 7.4) under N
(
2
.
(
(
B–D) Multicycle (up to 20 cycles) CVs obtained on GCE modified with ox-CNHs (B), ox-CNHs@PDA
ꢂ1
C), and g-N-CNHs (D) in 0.10 M PBS (pH 7.40) under N
2
. Scan rate, 0.05 V s
.
potentials, respectively). In g-N-CNHs, this redox process can be ascribed to the
presence of different nitrogen functionalities on the graphitic layer. A closer look
into the redox behavior of the three materials was carried out by consecutive CV cy-
cles in the range +0.5 V to +1.5 V (Figures 4B–4D). A slight change of CV profile was
observed after the first cycle for ox-CNHs and g-N-CNHs, probably because of
wetness issues, and successive cycles remained similar (Figures 4A and 4C). In
contrast, the change after the first cycle for ox-CNHs@PDA was much more evident.
In particular, a clear redox process was observed in the first CV, related to the
2
6,27
oxidation and reduction of dopamine and dopaminequinone, respectively,
indi-
2
8
cating the presence of non-covalently adsorbed dopamine domains. The redox
process was greatly diminished in five successive cycles and then disappeared in
subsequent cycles, a sign that all the reacting supramolecular domains have been
oxidized and polymerized.
The electrocatalytic activity toward O
2 2 2
reduction to H O was first examined by CV
under a pure O flow for all three materials (ox-CNHs, ox-CNHs@PDA, and g-N-
2
CNHs). For ox-CNHs and ox-CNHs@PDA, the activation proceeds at much more
negative potentials and with lower current than with the final material (g-N-CNHs).
More in-depth analysis was therefore performed only on g-N-CNHs. Figure 5 shows
the electrocatalytic activity of g-N-CNHs at three different pH values (1.0, 7.4, and
1
2
3.0) after subtraction of the background current measured under N . At all three
pH values, clear activation was observed, although it occurred at three different po-
tential values. The peak current density at alkaline pH was slightly lower than that at
acidic or neutral pH, but on the other hand, the most favorable potential was ob-
tained by operating at pH 13.0, with an onset potential of +0.71 V (Table 2).
Chem 4, 1–18, January 11, 2018
7

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 5. Electrochemical Analysis of the O
2
Reduction through Cyclic Voltammetry
Background-subtracted CVs obtained on GCE modified with g-N-CNHs at pH 1.0 in 0.1 M H
2
SO
4
(
(
red line), at pH 7.4 in 0.10 M phosphate buffer solution (green line), and at pH 13.0 in 0.1 M NaOH
ꢂ1
.
blue line) under pure O
2
. Scan rate, 0.05 V s
However, also at pH 1.0 and 7.4, the onset potentials are very positive, confirming
the flexibility of the catalyst to work under different pH regimes and with excellent
efficiency.
Figures S15 and S16 show the CVs measured for g-N-CNHs800 and g-N-CNHs900
and the relative cathodic current profiles after subtraction of the capacitive currents.
The current densities were lower than those of g-N-CNHs at all three pHs as a
consequence of the lower levels of N atoms. However, the onset potentials in
both g-N-CNHs800 and g-N-CNHs900 were slightly more positive, presumably
because of the increased level of graphitization of the carbon shell by annealing
the PDA layer at higher temperatures, which favors the reduction process. Hence,
the optimum thermal treatment is related to the final utilization of the material:
higher annealing temperature decreases the energetic barrier for the ORR but is
2 2
detrimental for the selectivity toward H O .
We performed rotating disk electrode (RDE) experiments (Figure S17) to confirm the
reduction product identity in each medium and obtain the number of electrons
involved in the reduction process. The linearity of the Koutecky-Levich (K-L) plots
(
Figure S18) implies a first-order reaction toward dissolved O
2
at each set pH. The
values of n calculated were 2.6 G 0.2, 2.0 G 0.1, and 3.1 G 0.2 at pH 1.0, 7.4,
2
9
and 13.0, respectively. In order to confirm the selectivity of the process, we per-
formed rotating ring disk electrode (RRDE) experiments under the same pH condi-
tions. The number of electrons n calculated (Equation 1) from the ratio of disk and
ring currents were 2.4 and 3.2 at pH 1.0 and 13.0, respectively (Tables 2 and S3).
The values are consistent with the RDE experiments.
n = 4 ꢂ 2ðI
Hence, at lower pH, where the proton concentration is higher, the reduction of O
leads mainly to H , involving the required two electrons, whereas the higher
number of electrons involved at alkaline pH indicates that the process is less selec-
tive toward H production.
R
=NÞ
(Equation 1)
2
2
O
2
2
O
2
8
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Table 2. Values of the Various Electrocatalytic Parameters at Each Applied pH
ꢂ2
ꢂ1
ꢂ
pH
ORRonset (V)
0.40
E
peak (V)
j
peak (mA cm
)
j
norm (mA mg
)
Ne (RRDE)
1
7
1
.0
0.18
0.40
0.59
ꢂ3.6
8.7
5.5
4.3
2.4
a
.4
0.53
ꢂ2.3
2.1
3.0
0.71
ꢂ1.79
3.2
The potential values reported in the table refer to the RHE reference electrode. The current density
values are normalized by the electroactive surface area obtained with hydroquinone as redox probe
(
see Supplemental Information). Further normalization by the amount of catalyst is also reported.
a
The RRDE experiments performed in neutral media were affected by instability problems of the ring cur-
ꢂ
rent attributed to phosphate surface adsorption; the Ne value was derived from Koutecky-Levich plots
obtained with the same experimental setup.
On the basis of the CV data, bulk electrolysis experiments were then performed in
2 2
each medium in order to quantify the as-formed H O product (Figures 6A–6C).
Three different applied potentials were explored to optimize conditions. In partic-
ular, each fixed potential at each pH was chosen on the basis of the values observed
for the onset potential, calculated by the tangent method, and the peak potential in
the CV. In all three cases, higher FEs were obtained by working at more positive po-
tentials. The highest FEs were obtained by working at pH 1.0 at a potential of +0.30 V
versus RHE. FEs obtained at pH 7.4 and 13.0 were also very good, and such FEs were
achieved operating at very positive controlled potentials.
By considering the set potential and the FEs obtained, the present catalyst outper-
forms the most recently reported state-of-the-art catalysts, working particularly well
under acidic conditions (see Table S4 and Figure S20). As a comparison, all the refer-
ence materials were tested as catalysts under acidic conditions (conditions of
maximum FE of g-N-CNHs). The samples based on different nanocarbons, g-N-
CNTs, g-N-CNCs, and g-N-GO showed much reduced FEs toward H
2 2
O of 20%,
3
9%, and 25%, respectively. The FE for the polypyrrole- and polyaniline-derived cat-
alysts was 6% and 65%, respectively.
2 2
In many reported cases, the long-term stability of electrocatalysts for H O evolution
2 2
is severely compromised as the accumulation of H O in solution leads to etching of
the catalyst. Other factors can also be involved, such as the reaction conditions,
including the pH, the working potentials, and the oxygen radicals, which can evolve
by other parallel O
2
reduction reactions or species deriving from the decomposition
itself. In order to evaluate the long-term performance of
reactions involving H
2
O
2
g-N-CNHs, long-term bulk electrolysis experiments were carried out (Figures 6D
and 6E). The potential was set at the value where the FE% was highest in each
medium: +0.30 V at pH 1.0, +0.45 V at pH 7.4, and +0.65 V at pH 13.0.
In the first 6 hr, the FEs at pH 1.0 and 13.0 remained almost constant (Figure 6D),
such that deactivation took place only after the sixth hour. By working at pH 7.4,
the FE over time decreased much more steeply. Overall, however, the 24-hr-long ex-
periments gave a stable current at all pH values. Thus, the drop in FE% is not related
to the stability of the material but is most likely a result of the further reduction of
H
2
O
2
to H
by electrochemical impedance spectroscopy [EIS]) or simply decomposition of
via chemical processes, calling for a future assembly of continuous-flow sys-
tems that can store the formed H . However, to prove our hypothesis, we used
the modified electrode several times with new electrolyte, and the FE in the initial
hr remained similar. We compared the FEs for the materials calcined at higher tem-
perature (Figure S19), finding that FEs decreased, although they were still good, as a
2
O over long experimental times (favored at alkaline pH, as shown later
2 2
H O
2
O
2
6
Chem 4, 1–18, January 11, 2018
9

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 6. Study of the Variation of FE in Relation to pH and Applied Potential and Evaluation of the Stability over Long-Term Electrolysis Experiments
A–C) FE% of H produced after 1 hr of electrolysis at three different fixed potentials (for each pH value) using a g-N-CNH modified electrode at pH
.0 (A), pH 7.4 (B), pH 13.0 (C) at three different potential values for each medium.
(
2 2
O
1
(
(
D) Plots of FE percentage versus time.
E) A stability test at pH 1.0, 7.4, and 13.0 over 24 hr was carried out with a three-electrode cell comprising a working electrode made of g-N-CNH-
modified glassy carbon, a counter electrode made of Pt net separated from the bulk solution by a porous membrane, and a reference Ag/AgCl
electrode.
consequence of the lower content of N atoms, thus confirming the strict correlation
between activity and N atoms regardless of pH.
2 2
We calculated the H O production rates by g-N-CNHs at each pH after 1, 3, 6, 9,
and 24 hr (Table S1). The rates were normalized by the amount of catalyst and
area of the electrode. Production rates after 1 hr were very high (54, 42, and
ꢂ1 ꢂ1
ꢂ2
7
3 mmol g
h
cm at pH 1.0, 7.4, and 13.0, respectively) although there was a
decline over time for successive hours. Interestingly, at pH 1, after the first hour,
the rate remained constant over 24 hr of electrolysis.
DISCUSSION
The mechanism for the ORR with metal-free catalysts is still under discussion. It is
clear that N-doped catalysts can evolve the reduction through two different path-
ways based on either two-electron or four-electron reduction, generating H
2
O
2
or
two-electron
reduction is less frequent, and the selective production of hydrogen peroxide is
very challenging, also considering that, even when formed, H can undergo
3
0
31–33
2
H O, respectively. Reduction to water is reported more often,
2
O
2
further in situ two-electron reduction to water. Two key factors are the porosity of
the material and the type and amount of N atoms. The porosity of the material is
strongly correlated with the conductivity of the catalyst, influencing also mass and
10
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
3
4
electron-transfer processes. In particular, the microporosity of the material also
plays a crucial role, because the presence of micropores decreases the residence
time of the H
2
O
2
on the material, thus contributing to prevent further reduction to
In more detail, a likely explanation is that the competition between
and the O molecule for the active sites is exacerbated by the smaller size of
the pores. Smaller pores would cause an enhanced relative concentration of the
smaller O around the active sites, thus decreasing the availability of H for further
3
3,35
water.
H
2
O
2
2
2
2 2
O
reduction. On the basis of this explanation, the CNHs contribute to define the overall
textural properties of the final material, in particular of the microporous volume,
which is instrumental to gaining a critical amount of micropores. By varying the
core carbon nanostructures, the microporous volume is sensibly reduced, in parallel
with a reduction in selectivity under acidic conditions. This suggests that there is a
dependence between the microporosity and the actual mechanism of reduction,
such that enhanced microporosity favors two-electron reduction. We measured
both lower surface area and lower microporous volume for all g-N-CNT, g-N-
CNC, and g-N-GO samples than for the g-N-CNHs (see Table S1). However, there
are several other factors likely influencing the mechanism, making the overall picture
rather complex. For example, N doping changes the basicity and hydrophilicity of
3
6
the nanocarbon, as well as the surface electronic properties. On the basis of pre-
vious studies, we also propose that, in our system, there is an increase in the charge
mobility of the nanocarbon, effected by the presence of N heteroatoms onto the sur-
3
7
face. Overall, compared with undoped nanocarbons, the modified band structure
2
2
resulting from N doping lowers the work function at the C/O
ported studies on ORR by N-doped carbon material point toward a four-electron
reduction of O . It has been previously reported that pyridinic N sites are the
most active N sites for ORR to H O because they can weaken O–O bonds by impart-
ing a relatively high positive charge density on the adjacent carbon atoms forming
2
interface. The re-
2
2
3
8
N–Ca bond polarization. In addition, it has been claimed that delocalized lone
pairs from the doped nitrogen (pyridinic) atoms facilitate charge transfer from the
p orbital to the antibonding orbitals in O
2
, also resulting in weakening of the O–O
bond and driving the reactivity through the four-electron reaction pathway to
3
9
water formation. XPS analysis showed that all three reference materials g-N-CNTs,
g-N-CNCs, and g-N-GO have a lower Npyridinic/Npyrrole ratio than g-N-CNHs. The
N
pyridinic/Npyrrole ratio decreases in the order g-N-CNCs (0.81) < g-N-GO (0.72) <
g-N-CNTs (0.56), which follows the same trend of selectivity decrease under acidic
regimes (g-N-CNCs, FE = 39%; g-N-GO, FE = 25%; g-N-CNTs, FE = 20%), indicating
that in parallel with the microporosity, the relative types of N atoms influence elec-
trocatalysis. In agreement with former literature hypotheses and with other previous
4
0
studies, it seems reasonable in our case to discern a mechanism dependent on the
different pH regimes applied. Under acidic conditions, the resulting effective pro-
tonation of the pyridinic N compromises the ability to weaken the O–O bond, lead-
ing to preferential formation of H
decrease of selectivity toward H
2
O
2
. In contrast, under neutral or alkaline pH, a
2
O
2
was observed as H O formation through rupture
2
of O–O became progressively more available, in strong agreement with the
observed FE% in each electrolyte, with the RRDE experiments, and with the Tafel
analysis (described below). This strongly suggests a prominent role of the pyridinic
ꢀ
N in our carbon nanomaterials. XPS data indicate that thermal treatment at 700 C
does not increase the relative percentage of quaternary or oxidized atoms, as
2
1
reported for other similar materials. On the contrary, no quaternary or oxidized
N-atom contribution was found for g-N-CNHs, and the total of N atoms consisted
4
1
of the catalytically active pyridinic (and to a less extent pyrrolic) N. The crucial
involvement of N atoms into the ORR mechanism is further corroborated by the chro-
noamperometric experiments conducted with g-N-CNHs800 and g-N-CNHs900
Chem 4, 1–18, January 11, 2018
11

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Figure 7. Tafel Analysis of the Catalyst under Different pH Regimes
Tafel plots of logI
K
versus E (V) for the ORR on the g-N-CNH electrode in oxygen-saturated solution
ꢂ1
at pH 1.0 (red line), at pH 7.4 (green line), and at pH 13.0 (blue line) under O
RDE rotation rate, 1,200 rpm.
2
. Scan rate, 0.05 V s
;
where the N content was lower, resulting in progressively lower FEs (Figure S19). In
further support of this hypothesis, we noted that a change in the enwrapping poly-
meric precursor, namely either polypyrrole or polyaniline, resulted in a different dis-
tribution of the N atoms, such that the percentage of pyridinic N was higher than that
of pyrrolic N. This also affected the selectivity, which was diminished for both
CNHs@pyrr-700 (FE = 6% at pH 1) and CNHs@PANI-700 (FE = 60% at pH 1). Hence,
it is plausible to partly relate the high selectivity of our catalyst under acidic condi-
tions to the enhanced amount of pyridinic N atoms. Obviously, the overall anatomy
of the mechanistic features determining a switch of selectivity can be dependent
upon several other parameters, calling for additional experiments, particularly den-
sity functional theory calculations. However, it is clear from the above experiments
that both microporosity and the amount of pyridinic N play a crucial role. In addition
to the textural and atomic characteristics, the presence of CNHs guarantees an
easier electron transfer onto the material surface where the active N sites are mainly
distributed, promoting the ORR at lower overpotentials. We believe that one other
2 2
key aspect of the current selective process (almost totally selective toward H O at
acidic pH) is also partly related to the operative conditions applied, in particular the
fact that the catalyst is active at very low overpotentials, which is a distinctive param-
eter contributing to high selectivity. Most of the other studies operated at more
2
negative applied potentials, and thus obtained mixed products or only H O. This
is proved by the drop in FEs when the bulk electrolysis experiments were conducted
at less positive potential (Figure 6), where the competing four-electron reduction
process starts to be effective. The pH strongly influences the mechanism, which
could be different when passing from acid to alkaline electrolytes. To prove this,
we calculated Tafel plots from RDE experiments at the three pH values. Figure 7
demonstrates that the catalyst shows linear Tafel plots in acidic, neutral, and alkaline
ꢂ1
media. The corresponding Tafel slopes were found to be 95 mV dec for pH 1.0,
ꢂ1
ꢂ1
8
4 mV dec for pH 7.4, and 71 mV dec for pH 13.0. At lower potentials, the curves
are bent downward, and there is no well-defined linear region found in the plot.
The Tafel plots clearly show the differences in electrode activity in the three different
pH environments and provide further insight on the dependence of the ORR
12
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
mechanism on the nature of the electrolyte; pH can also be investigated on the basis
of Tafel analysis. The different Nernstian slopes obtained agree well with previous
data reported in the literature on heteroatom-doped carbon-based materials as
ꢂ1 42–44
advanced electrocatalysts for the ORR process (60–120 mV dec ).
ꢂ1
According to the RRDE results, a decrease in the Tafel slope from 95 mV dec (acid
ꢂ1
environment) to ca. 70 mV dec (alkaline environment) is consistent with the change
ꢂ
ꢂ
in the reaction pathway from a 2e to (2 + 2)e transfer process and the correspond-
ing rate-determining step, thus further corroborating our interpretation based on
the strong pH dependence of the present catalyst.
In order to gain a better insight into the ORR catalysis, we measured EIS in an
2 2 4
O -saturated 0.1 M X solution (X = H SO , KOH, or PBS) at the potential value cor-
responding to the maximum FE obtained in the bulk electrolysis experiments by
superimposing an AC signal of 10 mV amplitude over the frequency range from
5
ꢂ2
45,46
1
0 to 10 Hz.
Nyquist plots show two different behaviors depending on the nature of the electro-
lyte solution. It is observed in Figure S21 that the real and imaginary components of
the Nyquist diagrams diminish in the sequence 0.1 M H
NaOH. On the other hand, the Nyquist and Bode plots show different shapes, which
can be related to different electrochemical processes. Unlike H SO - and PBS-based
2 4
SO > 0.1 M PBS > 0.1 M
2
4
electrolytes, the case of NaOH is described by two loops and two time constants in
Nyquist and phase diagrams, respectively. According to the K-L results, it is reason-
able to suggest that at pH 13.0, a first time constant is associated with the reduction
of O
second electron-transfer step, presumably the reduction of H
of H O, in agreement with the lower FE in NaOH solution. In contrast, the further
2
and the formation of H
2
O
2
; the second time constant can be attributed to a
2
O
2
to the formation
2
reduction of H
electrolytes.
2 2 2 2 4
O to H O is strongly disfavored in H SO - and PBS-based
In summary, we confined N-doped porous carbon obtained through thermal anneal-
ing of a PDA polymer to the nanoscale through coating of CNHs. The nanocatalyst is
2 2
able to trigger the ORR to H O with record performance (very positive onset poten-
tial in the 1–13 pH range, almost total FE in the best case, and long-term stability) as
a result of a well-suited set of properties: (1) a high surface area and suitable porosity
dictated by the CNH scaffold; (2) a convenient distribution of different types of N
atoms, the content of which is not much affected by thermal annealing; and (3)
improved conductivity and facilitation of electron-transfer processes. We did
not discern in detail the highly complex mechanism, but thanks to additional
experiments with reference materials, we could pinpoint microporosity and specific
N
pyridinic/Npyrrolic ratios as two key parameters affecting the reduction route. The re-
sults are very reproducible, in contrast with the general perception of nanocarbons
as highly batch-dependent materials. Given the remarkable catalytic performance
and the lower cost in comparison with traditional precious metal-based catalysts,
this study opens the door to the implementation of viable new strategies for the sus-
2 2
tainable production of H O .
EXPERIMENTAL PROCEDURES
Determination
The amount of H after the electrolysis experiments was detected by permanga-
nometric titration. The value reported is the average of nine independent samplings;
2 2
H O
2
O
2
Chem 4, 1–18, January 11, 2018
13

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
the reproducibility of the material performance was tested with three different
batches of material. The amount of H produced was further validated by an elec-
2
O
2
4
7
tro-sensing method developed recently within our group, and the two methods
are in excellent agreement.
Synthesis of the Materials
Ox-CNHs
3
p-CNHs (100 mg) were sonicated in 100 mL of HNO (69%) for 15 min. The resulting
ꢀ
dispersion was stirred for 3 hr at 35 C. Cold water (200 mL) was then added to the
reaction and the crude mixture was filtered through a Millipore membrane (JHWP,
0
.45 mm) and washed with deionized water until neutralization of the washings.
The black powder was then washed with MeOH and dried with Et O, affording
9 mg of ox-CNHs.
2
9
Ox-CNHs@PDA
Dopamine hydrochloride (400 mg, 2.11 mmol) and ox-CNHs (10 mg) were ultraso-
nicated in deionized water (40 mL) for 30 min. Tris buffer (pH 8.5, 10 mL, 50 mM)
was then added, and the reaction was stirred for 16 hr at room temperature. The
crude mixture was filtered through a Millipore membrane (JHWP, 0.45 mm) and
washed with cycles of ultrasonication and filtration in water (2 3 50 mL) and ethanol
(
2
2 3 50 mL). The black product was rinsed with Et O and dried under a vacuum,
affording 50 mg of ox-CNHs@PDA.
g-N-CNHs
Ox-CNHs@PDA (140 mg) were annealed under Ar for 3 hr to afford g-N-CNHs
92 mg). No further purification was required.
(
Ox-MWCNTs
2 4
HNO (69%, 4 mL) and H SO (95%, 12 mL) were mixed in a round-bottom flask.
3
Then, 100 mg of pristine MWCNTs was added on top. The resulting mixture was son-
icated for 5 hr at room temperature. The reaction was then quenched with cold water
(
200 mL). The crude mixture was filtered through a Millipore membrane (JHWP,
.45 mm) and washed with deionized water until neutralization of the washings.
The black powder was then washed with MeOH and dried with Et O, affording
8 mg of Ox-MWCNTs.
0
2
9
Ox-MWCNTs@PDA
Dopamine hydrochloride (400 mg, 2.11 mmol) and ox-MWCNTs (10 mg) were ultra-
sonicated in deionized water (40 mL) for 30 min. Tris buffer (pH 8.5, 10 mL, 50 mM)
was then added, and the reaction was stirred for 16 hr at room temperature. The
crude mixture was filtered through a Millipore membrane (JHWP, 0.45 mm) and
washed with cycles of ultrasonication and filtration in water (2 3 50 mL) and ethanol
(
2
2 3 50 mL). The black product was rinsed with Et O and dried under a vacuum,
affording 43 mg of Ox-MWCNTs@PDA.
g-N-CNTs
Ox-MWCNTs@PDA (154 mg) were annealed under Ar for 3 hr to afford g-N-CNTs
91 mg). No further purification was required.
(
GO
2 2 8 2 5
In a 100 mL round-bottom flask containing K S O (200 mg, 0.37 mmol), P O
(
2 4
100 mg, 0.35 mmol), and H SO (10 mL), pristine G (200 mg) was added. The reac-
ꢀ
tion was stirred at 80 C for 4 hr. The crude mixture was cooled down, diluted with
14
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
deionized water (50 mL), filtered through a Millipore membrane (JHWP, 0.45 mm),
and washed with deionized water until neutralization of the washings. The black
powder collected from the Millipore membrane was transferred into a round-bottom
ꢀ
flask and KMnO
4
(100 mg, 0.63 mmol) and H
2
SO
4
(20 mL) were added at 0 C. The
ꢀ
mixture was stirred at 35 C for 2 hr. After that, deionized water (20 mL) and 30%
2 2
H O (2.4 mL) were added, and the reaction was stirred for 15 min. The crude
mixture was filtered through a Millipore membrane (JHWP, 0.45 mm) and washed
with HCl 1 M (100 mL) and deionized water until neutralization of the washings. Even-
tually the black powder was washed with MeOH and dried with Et
17 mg of GO.
2
O, affording
2
GO@PDA
Dopamine hydrochloride (400 mg, 2.11 mmol) and GO (10 mg) were ultrasonicated
in deionized water (40 mL) for 30 min. Tris buffer (pH 8.5, 10 mL, 50 mM) was
then added, and the reaction was stirred for 16 hr at room temperature. The
crude mixture was filtered through a Millipore membrane (JHWP, 0.45 mm) and
washed with cycles of ultrasonication and filtration in water (2 3 50 mL) and ethanol
(
2
2 3 50 mL). The black product was rinsed with Et O and dried under a vacuum,
affording 47 mg of GO@PDA.
g-N-GO
GO@PDA (156 mg) was annealed under Ar for 3 hr to afford g-N-GO (91 mg). No
further purification was required.
Ox-CNCs@PDA
Oxidized carbon nanocones (ox-CNCs) were supplied by n-TEC (http://www.n-tec.
no), with an approximate composition of 20% cones, 70% disks, and 10% amorphous
carbon.
Dopamine hydrochloride (1.6 g, 8.44 mmol) and ox-CNCs (40 mg) were ultrasoni-
cated in deionized water (160 mL) for 30 min. Tris buffer (pH 8.5, 40 mL, 50 mM)
was then added, and the reaction was stirred for 16 hr at room temperature. The
crude mixture was filtered through a Millipore membrane (JHWP, 0.45 mm) and
washed with cycles of ultrasonication and filtration in water (2 3 200 mL) and ethanol
(
2
2 3 200 mL). The black product was rinsed with Et O and dried under a vacuum,
affording 175 mg of Ox-CNCs@PDA.
g-N-CNCs
Ox-CNCs@PDA (130 mg) were annealed under Ar for 3 hr to afford g-N-CNCs
69 mg). No further purification was required.
(
Ox-CNHs@PANI
Ox-CNHs (35 mg) and aniline (200 mg, 2.14 mmol) were ultrasonicated in HCl
(
10 mL, 2 M) for 15 min. Then, HCl (40 mL, 2 M) was added and the reaction was
ꢀ
ꢀ
4 2 2 8
cooled down to 0 C–4 C in an ice bath. (NH ) S O (735 mg, 3.2 mmol) was slowly
added to the reaction, which was mechanically stirred at room temperature for 16 hr.
Then, the crude mixture was filtered through a Millipore membrane (JHWP,
0
.45 mm), washed with HCl 1 M (2 3 100 mL), deionized water (3 3 100 mL), and
2
ethanol (1 3 100 mL), and dried with Et O, affording 108 mg of ox-CNHs@PANI.
CNHs@PANI-700
Ox-CNHs@PANI (156 mg) were annealed under Ar for 3 hr to afford CNHs@PANI-
00 (91 mg). No further purification was required.
7
Chem 4, 1–18, January 11, 2018
15

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
Ox-CNHs@polypyrrole
Ox-CNHs (30 mg) were ultrasonicated in water (30 mL) for 5 min. Pyrrole (200 mg,
3
2
.0 mmol), and FeCl (16 mg, 0.13 mmol) were added and the mixture was sonicated
for 5 min. Hydrogen peroxide (30%, 1 mL) was added to trigger the polymerization.
The reaction was mechanically stirred for 16 hr at room temperature. Then, the crude
mixture was filtered through a Millipore membrane (JHWP, 0.45 mm), washed
with deionized water (3 3 100 mL) and ethanol (1 3 100 mL), and dried with Et
affording 86 mg of Ox-CNHs@polypyrrole.
2
O,
CNHs@Pyrr-700
Ox-CNHs@polypyrrole (70 mg) were annealed under Ar for 3 hr to afford
CNH@Pyrr-700 (34 mg). No further purification was required.
Electroactive Surface Area
Calculation of the electroactive surface area (EASA) was done as follows: CV from
0
.15 to 1.65 V versus RHE was recorded in 0.5 mM of hydroquinone (HQ) in 0.1 M
PBS, and the peak potential of the oxidation peak was set as the working potential
for chrono-amperometry experiments 20 s long. This is the time necessary to allow
the formation of a monomolecular layer of HQ onto the modified GCE electrode.
The EASA was calculated according to the following equation:
nFACD1=2
Q =
;
(Equation 2)
1=2
ðptÞ
where n is the number of electrons involved, F is the Faraday constant (96,485 C
ꢂ1
2
mol ), A is the geometric area of the GCE (0.07 cm ), C is the concentration of
ꢂ3
5
2
ꢂ1
HQ in mol cm , D is the diffusion coefficient (0.85 3 10 cm
s
), and t is
time.
SUPPLEMENTAL INFORMATION
Supplemental Information includes Supplemental Experimental Procedures, 21
figures, and 5 tables and can be found with this article online at https://doi.org/
1
0.1016/j.chempr.2017.10.013.
AUTHOR CONTRIBUTIONS
D.I. performed the synthesis and physical characterization of the material. A.G. per-
formed the electrochemical characterization and catalytic tests. A.C. performed the
XPS characterization. L.N. performed the EDX and TEM characterization. M.B. per-
formed the impedance measurements. M.M. organized and prepared the manu-
script with input from S.M., M.B., C.T., F.V., P.F., and M.P.
ACKNOWLEDGMENTS
The research leading to these results received funding from the Seventh Frame-
work Programme (FP7/2007–2013) under grant agreement no. 310651 (SACS proj-
ect). The work also used funding from the Italian Ministry of Education, University,
and Research (MIUR) PRIN Giovani 2015 program under grant agreement no.
2
015TWP83Z. We thank Prof. G. Adami and Dr. M. Crosera for the ICP-MS
analysis.
Received: May 18, 2017
Revised: September 11, 2017
Accepted: October 25, 2017
Published: December 14, 2017
16
Chem 4, 1–18, January 11, 2018

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
REFERENCES AND NOTES
1
. Pfleiderer, G., and Rledl, H.J. (1939).
Production of hydrogen peroxide. US patent
US2158525A, filed October 3, 1936, and
published May 16, 1939.
arrays with high electrocatalytic activity for
oxygen reduction. Science 323, 760–764.
J.-C., Schaaf, P., and Ball, V. (2009).
Characterization of dopamineꢂmelanin
growth on silicon oxide. J. Phys. Chem. C 113,
1
4. Liu, X., and Dai, L. (2016). Carbon-based metal-
free catalysts. Nat. Rev. Mater. 1, 16064.
8234–8242.
2
. Yi, Y., Zhou, J., Guo, H., Zhao, J., Su, J., Wang,
L., Wang, X., and Gong, W. (2013). Safe direct
27. Bernsmann, F., Voegel, J.-C., and Ball, V.
(2011). Different synthesis methods allow to
tune the permeability and permselectivity of
dopamine–melanin films to electrochemical
probes. Electrochim. Acta 56, 3914–3919.
1
5. Dai, L., Xue, Y., Qu, L., Choi, H.-J., and Baek,
J.-B. (2015). Metal-free catalysts for oxygen
reduction reaction. Chem. Rev. 115, 4823–
synthesis of high purity H
plasma reaction. Angew. Chem. Int. Ed. 52,
446–8449.
2 2 2 2
O through a H /O
4892.
8
1
6. Qu, L., Liu, Y., Baek, J.-B., and Dai, L. (2010).
Nitrogen-doped graphene as efficient metal-
free electrocatalyst for oxygen reduction in fuel
cells. ACS Nano 4, 1321–1326.
3
. Zhou, J., Guo, H., Wang, X., Guo, M., Zhao, J.,
Chen, L., and Gong, W. (2005). Direct and
continuous synthesis of concentrated
28. Yang, J., Niu, L., Zhang, Z., Zhao, J., and Chou,
L. (2015). Electrochemical behavior of a
polydopamine nanofilm. Anal. Lett. 48, 2031–
2039.
hydrogen peroxide by the gaseous reaction of
2 2
H /O non-equilibrium plasma. Chem.
Commun. (Camb.), 1631–1633.
1
7. Iijima, S., Yudasaka, M., Yamada, R., Bandow,
S., Suenaga, K., Kokai, F., and Takahashi, K.
2
9. Bard, A.J., and Faulkner, L.R. (2001).
Electrochemical Methods: Fundamentals and
Applications, Second Edition (John Wiley),
p. 874.
(
1999). Nano-aggregates of single-walled
graphitic carbon nano-horns. Chem. Phys. Lett.
09, 165–170.
4
. Edwards, J.K., Pritchard, J., Lu, L., Piccinini, M.,
Shaw, G., Carley, A.F., Morgan, D.J., Kiely, C.J.,
and Hutchings, G.J. (2014). The direct synthesis
of hydrogen peroxide using platinum-
promoted gold–palladium catalysts. Angew.
Chem. Int. Ed. 53, 2381–2384.
3
3
0. Guo, D., Shibuya, R., Akiba, C., Saji, S., Kondo,
T., and Nakamura, J. (2016). Active sites of
nitrogen-doped carbon materials for oxygen
reduction reaction clarified using model
catalysts. Science 351, 361–365.
18. Kong, J., Yee, W.A., Yang, L., Wei, Y., Phua,
S.L., Ong, H.G., Ang, J.M., Li, X., and Lu, X.
(2012). Highly electrically conductive layered
carbon derived from polydopamine and its
5
6
. Edwards, J.K., Solsona, B., N, E.N., Carley, A.F.,
Herzing, A.A., Kiely, C.J., and Hutchings, G.J.
functions in SnO -based lithium ion battery
2
anodes. Chem. Commun. (Camb.) 48, 10316–
10318.
3
1. Zhang, P., Sun, F., Xiang, Z., Shen, Z., Yun, J.,
and Cao, D. (2014). ZIF-derived in situ nitrogen-
doped porous carbons as efficient metal-free
electrocatalysts for oxygen reduction reaction.
Energy Environ. Sci. 7, 442–450.
(2009). Switching off hydrogen peroxide
hydrogenation in the direct synthesis process.
Science 323, 1037.
19. Ding, W., Wei, Z., Chen, S., Qi, X., Yang, T., Hu,
J., Wang, D., Wan, L.-J., Alvi, S.F., and Li, L.
(2013). Space-confinement-induced synthesis
of pyridinic- and pyrrolic-nitrogen-doped
graphene for the catalysis of oxygen reduction.
Angew. Chem. Int. Ed. 52, 11755–11759.
. Edwards, J.K., Ntainjua, N.E., Carley, A.F.,
Herzing, A.A., Kiely, C.J., and Hutchings, G.J.
3
2. Wei, W., Liang, H., Parvez, K., Zhuang, X., Feng,
X., and M u¨ llen, K. (2014). Nitrogen-doped
carbon nanosheets with size-defined
mesopores as highly efficient metal-free
catalyst for the oxygen reduction reaction.
Angew. Chem. Int. Ed. 53, 1570–1574.
2 2 2 2
(2009). Direct synthesis of H O from H and O
over gold, palladium, and gold–palladium
catalysts supported on acid-pretreated TiO
Angew. Chem. Int. Ed. 48, 8512–8515.
2
.
2
0. Yang, S., Feng, X., Wang, X., and M u¨ llen, K.
2011). Graphene-based carbon nitride
(
7
8
9
. Yamanaka, I., Onizawa, T., Takenaka, S., and
Otsuka, K. (2003). Direct and continuous
production of hydrogen peroxide with 93 %
selectivity using a fuel-cell system. Angew.
Chem. Int. Ed. 42, 3653–3655.
nanosheets as efficient metal-free
electrocatalysts for oxygen reduction
reactions. Angew. Chem. Int. Ed. 50, 5339–
5343.
3
3. Liang, J., Du, X., Gibson, C., Du, X.W., and
Qiao, S.Z. (2013). N-doped graphene natively
grown on hierarchical ordered porous carbon
for enhanced oxygen reduction. Adv. Mater.
2
1. Koos, A.A., Murdock, A.T., Nemes-Incze, P.,
Nicholls, R.J., Pollard, A.J., Spencer, S.J.,
Shard, A.G., Roy, D., Biro, L.P., and Grobert, N.
25, 6226–6231.
. Jirkovsk y´ , J.S., Panas, I., Ahlberg, E., Halasa,
M., Romani, S., and Schiffrin, D.J. (2011). Single
atom hot-spots at Au–Pd nanoalloys for
electrocatalytic H O production. J. Am.
2 2
Chem. Soc. 133, 19432–19441.
3
4. Yu, H., Shang, L., Bian, T., Shi, R., Waterhouse,
G.I.N., Zhao, Y., Zhou, C., Wu, L.-Z., Tung,
C.-H., and Zhang, T. (2016). Nitrogen-doped
porous carbon nanosheets templated from
g-C3N4 as metal-free electrocatalysts for
efficient oxygen reduction reaction. Adv.
Mater. 28, 5080–5086.
(2014). Effects of temperature and ammonia
flow rate on the chemical vapour deposition
growth of nitrogen-doped graphene. Phys.
Chem. Chem. Phys. 16, 19446–19452.
. Siahrostami, S., Verdaguer-Casadevall, A.,
Karamad, M., Deiana, D., Malacrida, P.,
Wickman, B., Escudero-Escribano, M., Paoli,
E.A., Frydendal, R., Hansen, T.W., et al. (2013).
2
2
2
2
2. Dorjgotov, A., Ok, J., Jeon, Y., Yoon, S.-H., and
Shul, Y.G. (2013). Activity and active sites of
nitrogen-doped carbon nanotubes for oxygen
reduction reaction. J. Appl. Electrochem. 43,
35. He, W., Jiang, C., Wang, J., and Lu, L. (2014).
High-rate oxygen electroreduction over
graphitic-n species exposed on 3D
Enabling direct H
rational electrocatalyst design. Nat. Mater. 12,
137–1143.
2 2
O production through
387–397.
1
hierarchically porous nitrogen-doped carbons.
Angew. Chem. Int. Ed. 53, 9503–9507.
3. Dumitru, A., Mamlouk, M., and Scott, K. (2014).
Effect of different chemical modification of
carbon nanotubes for the oxygen reduction
reaction in alkaline media. Electrochim. Acta
1
1
0. Marchesan, S., Melchionna, M., and Prato, M.
2015). Wire up on carbon nanostructures! how
to play a winning game. ACS Nano 9, 9441–
450.
(
36. Arrigo, R., H a¨ vecker, M., Wrabetz, S., Blume, R.,
Lerch, M., McGregor, J., Parrott, E.P.J., Zeitler,
J.A., Gladden, L.F., Knop-Gericke, A., et al.
(2010). Tuning the acid/base properties of
nanocarbons by functionalization via
135, 428–438.
9
4. Sing, K.S.W. (1985). Reporting physisorption
data for gas/solid systems with special
reference to the determination of surface area
and porosity (Recommendations 1984). Pure
Appl. Chem. 57, 603.
1. Liu, Y., Quan, X., Fan, X., Wang, H., and Chen,
S. (2015). High-yield electrosynthesis of
hydrogen peroxide from oxygen reduction by
hierarchically porous carbon. Angew. Chem.
Int. Ed. 5, 6837–6841.
amination. J. Am. Chem. Soc. 132, 9616–9630.
37. Strelko, V.V., Kuts, V.S., and Thrower, P.A.
(2000). On the mechanism of possible influence
of heteroatoms of nitrogen, boron and
phosphorus in a carbon matrix on the catalytic
activity of carbons in electron transfer
5. Hashimoto, A., Yorimitsu, H., Ajima, K.,
Suenaga, K., Isobe, H., Miyawaki, J., Yudasaka,
M., Iijima, S., and Nakamura, E. (2004).
Selective deposition of a gadolinium(III) cluster
in a hole opening of single-wall carbon
nanohorn. Proc. Natl. Acad. Sci. USA 101,
8527–8530.
1
2. Qiao, M., Tang, C., He, G., Qiu, K., Binions, R.,
Parkin, I.P., Zhang, Q., Guo, Z., and Titirici,
M.M. (2016). Graphene/nitrogen-doped
porous carbon sandwiches for the metal-free
oxygen reduction reaction: conductivity versus
active sites. J. Mater. Chem. A 4, 12658–12666.
reactions. Carbon 38, 1499–1503.
38. Tuci, G., Zafferoni, C., D’Ambrosio, P.,
Caporali, S., Ceppatelli, M., Rossin, A., Tsoufis,
T., Innocenti, M., and Giambastiani, G. (2013).
Tailoring carbon nanotube N-dopants while
designing metal-free electrocatalysts for the
1
3. Gong, K., Du, F., Xia, Z., Durstock, M., and Dai,
L. (2009). Nitrogen-doped carbon nanotube
26. Bernsmann, F., Ponche, A., Ringwald, C.,
Hemmerl e´ , J., Raya, J., Bechinger, B., Voegel,
Chem 4, 1–18, January 11, 2018
17

Please cite this article in press as: Iglesias et al., N-Doped Graphitized Carbon Nanohorns as a Forefront Electrocatalyst in Highly Selective O
Reduction to H , Chem (2017), https://doi.org/10.1016/j.chempr.2017.10.013
2
2 2
O
oxygen reduction reaction in alkaline medium.
carbon nanotubes in the oxygen
J. (2016). Hierarchical metal-free nitrogen-
ACS Catal. 3, 2108–2111.
reduction reaction. J. Phys. Chem. C 113,
doped porous graphene/carbon composites
as an efficient oxygen reduction reaction
catalyst. ACS Appl. Mater. Interfaces 8, 1415–
1423.
14302–14310.
3
9. Wan, K., Long, G.-F., Liu, M.-Y., Du, L., Liang,
Z.-X., and Tsiakaras, P. (2015). Nitrogen-doped
ordered mesoporous carbon: synthesis and
active sites for electrocatalysis of oxygen
reduction reaction. Appl. Catal. B 165,
42. Feng, X. (2015). Nanocarbons for Advanced
Energy Conversion (Wiley-VCH Verlag).
4
6. Hou, Y., Wen, Z., Cui, S., Ci, S., Mao, S., and
Chen, J. (2015). An advanced nitrogen-doped
graphene/cobalt-embedded porous carbon
polyhedron hybrid for efficient catalysis of
oxygen reduction and water splitting. Adv.
Funct. Mater. 25, 872–882.
43. Shinagawa, T., Garcia-Esparza, A.T., and
Takanabe, K. (2015). Insight on Tafel slopes
from a microkinetic analysis of aqueous
electrocatalysis for energy conversion. Sci. Rep.
5, 13801.
566–571.
4
0. Wan, K., Yu, Z.-P., Li, X.-H., Liu, M.-Y., Yang, G.,
Piao, J.-H., and Liang, Z.-X. (2015). pH effect on
electrochemistry of nitrogen-doped carbon
catalyst for oxygen reduction reaction. ACS
Catal. 5, 4325–4332.
44. Lopes, T., Kucernak, A., Malko, D., and
Ticianelli, E.A. (2016). Mechanistic insights into
the oxygen reduction reaction on metal–N–C
electrocatalysts under fuel cell conditions.
ChemElectroChem 3, 1580–1590.
4
7. Bracamonte, M.V., Melchionna, M., Giuliani, A.,
Nasi, L., Tavagnacco, C., Prato, M., and
Fornasiero, P. (2017). H2O2 sensing
enhancement by mutual integration of single
walled carbon nanohorns with metal oxide
4
1. Kundu, S., Nagaiah, T.C., Xia, W., Wang, Y.,
Dommele, S.V., Bitter, J.H., Santa, M.,
Grundmeier, G., Bron, M., Schuhmann, W., and
Muhler, M. (2009). Electrocatalytic activity
and stability of nitrogen-containing
catalysts: the CeO
2
case. Sens. Actuators B
45. Men, B., Sun, Y., Li, M., Hu, C., Zhang, M.,
Wang, L., Tang, Y., Chen, Y., Wan, P., and Pan,
Chem. 239, 923–932.
18
Chem 4, 1–18, January 11, 2018