Subscriber access provided by UCL Library Services
Article
Deconvolution of Utilization, Site Density, and Turn-Over-Frequency of FeNitrogen-Carbon ORR Catalysts Prepared with Secondary N-Precursors Nathaniel Leonard, Stephan Wagner, Fang Luo, Julian Steinberg, Wen Ju, Natascha Weidler, HUan Wang, Ulrike I. Kramm, and Peter Strasser ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.7b02897 • Publication Date (Web): 28 Dec 2017 Downloaded from http://pubs.acs.org on December 28, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Catalysis is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Deconvolution of Utilization, Site Density, and Turn-OverFrequency of Fe-Nitrogen-Carbon ORR Catalysts Prepared with Secondary N-Precursors Nathaniel D. Leonard,1 Stephan Wagner,2 Fang Luo,1 Julian Steinberg,1 Wen Ju,1 Natascha Weid‐ ler,2 Huan Wang,1 Ulrike I. Kramm,2 Peter Strasser*1 1. Department of Chemistry, Technical University Berlin, 10623 Berlin, Germany 2. Department of Chemistry and Department of Materials‐ and Earth Sciences, Graduate School of Excellence Energy Science and Engineering, Technical University Darmstadt, Otto‐Berndt‐Str. 3, 64287 Darmstadt, Germany ABSTRACT: Metal‐nitrogen‐carbon (M‐N‐C) catalysts represent a potential means of reducing cathode catalyst costs in low temperature fuel cell cathodes. Knowledge‐based improvements have been hampered by the difficulty to deconvolute active site density and intrinsic turn over frequency. In the present work, M‐N‐C catalysts with a variety of secondary nitrogen precursors are addressed. CO Chemisorption in combination with Mossbauer spectroscopy are utilized in order to unravel previously inaccessible relations between active site density, turn‐over‐frequency, and active site utilization, which provide a more fundamental description and understanding of the origin of the catalytic reactivity; they also provide guidelines for further improvements. Secondary nitrogen precursors impact quantity, quality, dispersion, and utilization of active sites in distinct ways. Secondary nitrogen precursors with high nitrogen content and micropore etching capabilities are most effec‐ tive in improving catalysts performance.
KEYWORDS: Oxygen Reduction Reaction, Non‐precious metal Catalysts, Fuel Cells, Mössbauer Spectroscopy, Electroca‐ talysis Building on this PANI process, other works have shown
Introduction A significant impediment of commercialization of proton
that adding a secondary nitrogen precursor to a metal‐
exchange membrane fuel cells (PEMFCs) is the cost of
nitrogen‐carbon (MNC) catalyst synthesis process im‐
catalysts. Most of the cost is at the cathode, where the
proves performance.7‐8 Similarly, Lefevre et al and others
sluggish oxygen reduction reaction (ORR) is usually cata‐
also used two nitrogen precursors for a different MNC
lyzed with the use of precious
metals.1‐4
One of the re‐
synthesis processes.9‐11 For example, in Lefevre et al. an
search directions to solve this problem is the search for
initial heat treatment with iron acetate, phenanthroline,
suitable non‐precious metal alternatives. The most promis‐
and a carbon black was followed by a second heat treat‐
ing alternatives are in a broad class of catalysts called
ment under ammonia atmosphere.9 For both of these
metal‐nitrogen‐carbon (MNC) catalysts. Specifically, a
works, the hypothesis was that one of the roles of the sec‐
promising non‐precious metal alternative has been syn‐
ondary precursor, cyanamide/ammonia, was etching the
thesized by pyrolysis of polyaniline with a metal salt on a
primary nitrogen precursor, PANI/phenanthroline. This
carbon support (PANI). Effectually, aniline is polymerized
etching process is believed to create the pores that might
around the metal and carbon precursors during a wet‐
host additional active sites. In the current work we test
impregnation procedure. Following this mixing step, py‐
this hypothesis by varying the secondary nitrogen precur‐
rolysis in nitrogen atmosphere at 900 °C incorporates the
sor.
nitrogen and metal precursors into the carbon matrix.5‐6
One counter argument to this hypothesis is that highly active MNC catalysts have been synthesized without a
ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
second nitrogen precursor or even highly volatile precur‐ sors. For example, catalysts obtained by the sacrificial
Page 2 of 12
Experimental Methods Materials
support method (SSM) or by the oxalate‐supported pyroly‐
Catalysts were synthesized similarly to other processes
sis have been synthesized with nitrogen precursors such
in the literature.6‐7, 16 Ketjen 600 EJ was pretreated by with
as e.g. aminoantipyrine, nicarbazin or porphyrins.12‐14 For
0.5 M HCl for 24 hours. After vacuum filtration, the result‐
these catalysts, the important factor has been mesoporosi‐
ing powder was mixed into HNO3 at 90 °C under reflux for
ty, even when a volatile precursor was also used.14 Similar‐
seven hours. After a second vacuum filtration, the resulting
ly, a closed pyrolysis method also showed a correlation
powder is considered pretreated carbon.
between mesoporosity and activity despite having a more
For the catalysts synthesis, iron (III) chloride (5 g), ani‐
For these cata‐
line (3 ml), and ammonium peroxidisulfate (5 g) are mixed
lysts, one of the challenges to synthesizing high perfor‐
in 1 M HCl. Optionally 0.333 moles of nitrogen were added
mance MNC catalysts has been incorporation of nitrogen
in the form of a secondary nitrogen precursor, based on a
species into the carbon matrix. Although it has been shown
optimization of cyanamide as a secondary nitrogen precur‐
that high nitrogen content is important for high activity,15
sor.7‐8, 17 These were selected from common MNC catalyst
the ambiguity of the active site has prevented quantifica‐
nitrogen precursors: cyanamide (PANI‐CM),7‐8, 18 melamine
tion of the relationship between metal and nitrogen con‐
(PANI‐Mel),15,
tent and active site quantity. Ultimately, the question re‐
(PANI‐NCB).22 Labels of the resulting catalysts are in the
mains as to identify the role of secondary precursors in
parentheses; the catalyst synthesized without secondary
MNC catalysts. Specifically, to what extent does the sec‐
nitrogen precursors is labeled PANI. After 4 hours of mix‐
ondary nitrogen precursor etch pores for active sites, and
ing, the pretreated carbon is added, and the mixture is
how do they participate in the reactions associated with
stirred for 48 hours. After this stirring period, the mixture
active site formation or later the performance of active
is dried at 110 °C. The dried powder is ball milled and
sites in the oxygen reduction reaction. This is especially an
pyrolized for one hour at 900 °C. After the first heat treat‐
aspect of crucial interest for such synthesis protocols
ment, the catalyst is washed with 2 M H2SO4 for 16 hours
where iron is not directly coordinated by a sufficient num‐
at 90 C under reflux. Then the catalyst is vacuum filtered
ber of nitrogen ligands, as it is the case for PANI‐based
and dried. After drying the catalyst undergoes a second
FeNC.
heat treatment at 900 °C. A second acid wash and third
volatile nitrogen precursor,
melamine.15
19‐20
urea (PANI‐Urea),21 and nicarbazin
In the present study we use carbon monoxide adsorp‐
heat treatment are performed. At this point XRD data is
tion to quantify potential active sites.6 These numbers are
used to assess whether a third acid wash is necessary.
correlated with chemical information to allow for quantita‐
After each acid wash and heat treatment, XRD is used to
tive observation of changing surface chemistry as the ni‐
assess whether inorganic iron species have been adequate‐
trogen content is tuned by varying the precursors. Fur‐
ly removed. Often times the catalysts is clean after two acid
thermore, these catalysts are analyzed with nitrogen phy‐
washes, but if inorganic iron species can still be observed
sisoprtion to measure microporosity. Ultimately, the data
via XRD, a third acid wash and fourth heat treatment were
will allow a comparison of the roles of secondary nitrogen
performed. This additional cycle was necessary for the
precursors on porosity and active site creation while keep‐
control catalyst (PANI) as well as the nicarbazin based
ing the primary nitrogen precursor and carbon substrate
catalyst (PANI‐NCB). The final XRD signature for each
constant.
catalyst is included in Figure S1 in the supplemental in‐ formation. Multiple batches were made of all catalysts except the NCB, and the RDE activity was found to be re‐ producible. For the remaining characterization techniques, a single batch was chosen.
ACS Paragon Plus Environment
Page 3 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Physical Characterization
Electrochemical Characterization
Nitrogen physisorption data acquired on an Autosorb‐1
Catalysts were electrochemically characterized with ro‐
(Quantachrome Instruments) were used to determine the
tating‐disk electrodes (RDE). Initially an ink was prepared
surface area. Pore size distributions were calculated from a
by sonicating a mixture of 7.85 mg catalysts in 760 μl of
non‐local density functional theorem (NLDFT) pore model
deionized water, 190 μl isopropanol, and 50 μl of 5 wt%
based on carbon pores with both slit and cylindrical geom‐
Nafion solution. After sonicating for 15 minutes 10 μl were
etries. Bulk iron content was measured by ICP and ele‐
drop cast on a 5 mm diameter glassy carbon disk. After
mental analysis was used for bulk nitrogen and sulfur
drying for at least 10 minutes, electrodes were ready for
contents. The thermal decomposition of the catalyst precur-
use.
sors were investigated by a thermal gravimetric analyzer
Catalysts deposited RDEs were inserted in oxygen satu‐
(Perkin Elmer, STA 8000) under N2 atmosphere (10
rated 0.1 M HClO4 electrolyte. Initially electrodes were
ml/min). The samples were kept at 30°C for 1hour to clean
conditioned by cycling ten times between 1 and 0.2 V vs
the atmosphere, then heated at 30 °C /min ramp from 30°C
RHE with a scan rate of 100 mV/s. After this conditioning,
to 900 °C, holding for 1 hour.
a 5 mV/s cycle was collected between 1.05 and 0 V vs RHE.
Mössbauer Spectroscopy
Following the characterization cycle, the electrolyte was
The Mößbauer measurements were performed in
bubbled with nitrogen for 15 minutes, and a capacitive
transmission mode with a Mössbauer spectrometer
scan was collected. Oxygen polarization curves were cor‐
“MS96” from RCPTM with a 100 mCi 57Co/Rh source. Cali‐
rected for capacitive current using the nitrogen data, and
bration of the velocity scale was made using the sextet
the iR‐corrected based on high frequency resistance as
lines of alpha‐iron. Measurements were performed at
measured by Galvano‐controlled EIS at 0 A vs RHE.
room temperature with a scintillation detector in a velocity range of ± 8 mm s‐1. PTFE sample holders were filled with the required catalyst mass and closed with TESA tape. Spectra were analysed using the program Recoil.
Results and Discussion Role of secondary N-precursor on microporosity A number of catalysts were made using a single primary nitrogen precursor and four different secondary nitrogen
For determining the the amount of iron assigned to the
precursors as shown in Table 1: cyanamide, melamine,
different iron species the Lamb‐Mössbauer factors as de‐
urea, and nicarbazin. An overview of the physical and
termined by Sougrati et al. were used.23
chemical characteristics of these catalysts are provided in
X-ray induced photoelectron spectroscopy
the supplemental information (Table S1).
X‐ray photoelectron spectra (XPS) were measured with a Specs Phoibos 150 hemispherical analyzer and a Specs XR50M Al Ka X‐ray source. The system was calibrated with
Table 1. List of Secondary Nitrogen Precursors Used Nitrogen Precursor
Formula
MW / g N/C mol‐1
Precursor Mass Added
Cyanamide
CH2N2
42
2/1
7 g
Melamine
C3H6N6
126
6/3
7 g
Urea
CH4N2O
60
2/1
10g
Nicarbazin
C19H18N6O6 426
6/19
23.7 g
a silver reference sample. For the measurements catalyst powder is pressed on an indium foil and transferred into the high‐vacuum system. Spectra were analyzed using CasaXPS. Peaks were fitted using a Shirley background and mixed Gauss/Lorentz peaks. On the basis of the different elemental regions for Fe 2p, S 2p, N 1s, O 1s and C1s the elemental composition was determined. Measurements were performed for the four catalysts prepared under addition of a secondary N‐ precursor.
The catalysts that were synthesized using these precur‐ sors were characterized using nitrogen physisorption to better understand porosity and to check the literature theories about porosity created by secondary nitrogen
ACS Paragon Plus Environment
ACS Catalysis
precursors. The pore size distributions shown in Fig. 1 are
with PANI‐CM and PANI‐Mel showing the highest perfor‐
relatively similar for the different catalysts. Comparison of
mance and PANI and PANI‐NCB showing the lowest per‐
the catalysts to the support (black line), indicates loss in
formance. At lower potentials, most of the catalysts exhibit
mesoporosity and gain in microporosity. The exception is
a similar mass transport limiting plateau.
the PANI‐NCB sample, which appears to lose both micro‐ and mesoporosity. The exceptionally low porosity of the has the highest carbon content of the secondary nitrogen precursors. It is likely that this excess carbon material remains after the pyrolysis and blocks pores. This is con‐ sistent with the work of Nallathambi et al in which precur‐ sors with lower N/C ratios produced catalysts with lower porosity and performance.19 Thermogravimetric data
0 Current Density / mA cm-2 (geo.)
PANI‐NCB catalyst is not surprising considering that NCB
-2 -3 -4 -5
supports the idea of pore blockage by excess carbon mate‐ rial. It is evident that the NCB does not decompose to a similar mass as the other catalysts. This indicates a large amount of the secondary nitrogen precursor is left behind. The large mass left behind may also be due to the large amount of NCB that was required in the synthesis process
0.4 0.6 0.8 Potential / V vs RHE
1.0
To account for the mass transport effects, the kinetic current were calculated with the Koutecky‐Levich expres‐ sion: 1
other secondary precursors. It is also possible that the plays a role in their lower porosity.
0.2
Figure 2. Polarization curves from rotating disk electrode studies of various catalysts. Conditions: 0.1 M HClO4, O2 satu‐ rated, room temperature, 1500 RPM, loading: 0.4 mg/cm2
to attain a similar amount of added nitrogen as with the fourth heat treatment used for the PANI and PANI‐NCB
PANI PANI-NCB PANI-Urea PANI-Mel PANI-CM
-1
-6 0.0
shown in the supplementary information Figure S5, also
1
1
Where ik, the kinetic current can be calculated from the measured current, i, and the mass transport limited cur‐ rent, il. The resulting kinetic curves are shown in the Tafel
400
plot (Fig. 3). The initial slopes are similar suggesting simi‐ Support PANI PANI-NCB PANI-Urea PANI-Mel PANI-CM
350 300 250 200
lar reaction kinetics. Again the exception is the PANI‐NCB which has a higher Tafel slope. This higher Tafel slope could indicate a change in the mechanism.
150 1.0
100 50 0
1
10 Pore Width / nm
Figure 1. Pore size distribution of support and catalysts.
The catalysts were electrochemically characterized via rotating disk electrode, as shown in Fig. 2. The samples
Potential / V vs RHE
Differential Pore Distribution / m2 g-1nm-1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 12
0.9
0.8
PANI PANI-NCB PANI-Urea PANI-Mel PANI-CM
0.7
0.6 0.1
show varying catalytic performance as indicated by their behavior in the kinetic regime of the polarization curves,
ACS Paragon Plus Environment
1 10 Kinetic Current / mA cm-2 (geo.)
100
Page 5 of 12
Figure 3. Kinetic Current as calculated from rotating disk electrode data. Conditions: 0.1 M HClO4, O2 saturated, room temperature, 1500 RPM, loading: 0.4 mg/cm2
other Mössbauer spectra are provided in the supplemen‐
The impact of secondary nitrogen precursors on pore
It is found that the relative content of FeN4 sites in rela‐
etching and consequently on electrochemical performance
tion to the overall iron‐signature of the catalysts is rela‐
can be probed by comparison of Fig. 2 with the pore size
tively high, especially when comparing to other PANI‐
distribution in Fig. 1. One can see from Fig. 4 that all the
based catalysts.24‐25, 27 It is particularly striking that similar
catalysts exhibit higher microporosities than the carbon
to Kramm et al.25 and Zitolo et al.26 we managed to produce
support, which supports the theory that secondary nitro‐
a catalyst with exclusive presence of FeN4 sites as found
gen precursors are involved in etching of the catalyst sur‐
for PANI‐Urea.
tary information, Figure S2.
face. On the other hand, the poor correlation between
The D2 related FeN4 site is attributed to a structure simi‐
microporosity and current indicates that this etching pro‐
lar to FePc with ferrous iron in the mid‐spin state.14, 24
cess is only part of the story concerning how secondary
Beside this, some catalysts indicate the presence of iron
nitrogen precursor impact performance.
carbide (Sext1),27 troillite (Sext2),28 and most probably very small iron nitride particles29‐30 (Sext3 in PANI‐CM). The high level of troillite in the PANI‐NCB is corroborated
carbon support microporosity
3.0 Current Density / mA cm-2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
2.5 2.0 1.5 1.0 0.5 0.0
0
100
200
by the XRD data and the high sulphur content. This is in‐ dicative of the poor iron utilization in the PANI‐NCB cata‐
PANI-Mel
lyst.
PANI-CM PANI-Urea PANI PANI-NCB
300 400 500 600 Micropore / m2 g-1
700
800
Figure 4. Current density varying with microporous area of the various catalysts. Current density is taken at 0.8 V vs RHE from the rotating‐disk electrode data.
In order to explore the relationship between secondary nitrogen precursors and performance, this work will dis‐ cuss three different aspects of the catalysts: number of active sites, number of exposed active sites, and quality of active sites. The relation of surface exposed active sites towards the total number of active sites is defined as cata‐ lyst utilization.
Effect of secondary N-precusor on quantity of active sites. Mössbauer spectroscopy was used to measure total number of active sites. That means the sum of surface accessible sites and those hidden in the bulk .6 The Möss‐ bauer spectra of the two most active catalysts are shown in Fig. 5, and the associated results shown in Table 2. All
ACS Paragon Plus Environment
ACS Catalysis
A bulk mass‐based site density (MSD) was calculated a)
from Mössbauer results assigned to the D1 doublet (FeD1) that was previously identified as ORR active site.22,23,25,40
0.0
The following equation was used for determination:
Abs / %
0.2
wt % / 100
g mol
N
where MFe represents the molecular mass of iron and NA
PANI-Mel D1 D2 Sext1
0.8
Fe
sites per g
MSD
0.4 0.6
is Avagadro’s Number. This is similar to previous work by Sahraie et al, in which CO‐adsorption data were used to quantify adsorption sites on the catalyst surface (surface exposed active sites).6 For reasons of comparison also
b)
mass‐based site densities related to FeN4 site assigned to 0.0
doublet D2 as well as to the overall number of FeN4 sites were calculated and the related comparison to CO sorption
0.1
data are given in the supplementary information, Figure S3 (MSDD1 replotted for comparison reasons). We will come
0.2 0.3
back to this, later.
PANI-Urea D1 D2
0.4 -8
10
-6
-4
-2
0
2
4
6
8
doppler / mms-1 Figure 5: Mössbauer Spectra of PANI‐Mel (a) and PANI‐Urea (b) catalysts.
Table 2: Mössbauer Parameters for iron sites as found in the five catalysts, the catalysts are ordered with increasing ORR activity. / ∆ / / mms‐1 mms‐1 T / mms‐1
D1 D2 Sext1 Sext2
0.41 0.51 0.16 0.51
D1 D2 Sext1 Sext2
0.38 0.48 0.16 0.76
D1 D2
0.38 0.49
D1 D2 Sext 1
0.38 0.49 0.20
D1 D2 Sext 3
0.40 0.56 0.27
PANI 1.06 2.35 ‐0.09 20.7 ‐ 31.1 PANI‐NCB 0.98 2.10 0.03 20.8 ‐0.09 31.2 PANI‐Urea 0.98 2.09 PANI‐Mel 1.05 2.10 0.00 20.8 PANI‐CM 1.08 2.55 0.07 13.1
.
Kinetic Current / mA mg-1
Abs / %
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 12
48 41 2.9 7.9
0.56 1.44 0.26 0.36
25.5 39.6 7.1 27.8
0.58 1.48
36.3 64.0
0.62 1.38 0.26
33.4 50.5 16.2
0.60 1.06 0.26
57.4 36.8 5.8
8 PANI-CM
7 6
PANI-Mel
5 PANI-Urea
4
PANI
3 2
PANI-NCB
1
/ %
0.66 2 (fixed) 0.20 0.34
9
0
0
50 100 COTPD / nmol mg-1
150
Figure 6. Kinetic current density at 0.8 V vs RHE varying with adsorbed CO (adsorbed at 193 K via CO pulses).
In the present work it is assumed that CO‐adsorption sites are indicative of O2 adsorption sites. End‐on O2 bind‐ ing is assumed based on modeling results.31‐34 With end‐on binding, as opposed to side‐on binding, a single CO‐ adsorption site will correspond to one oxygen adsorption site due to the similar orientation of the CO and O2 mole‐ cule. Additional concerns that should be addressed are that it is possible that some CO‐biding sites are not suitable for ORR, and some ORR sites don’t bind CO. The first concern is difficult to control for, and requires the reader to under‐
ACS Paragon Plus Environment
Page 7 of 12
stand that our TOF should be thought of as an average TOF for all possible ORR site. This is with the assumption that CO‐binding sites can be considered to be at least a possible
CO‐adsorption and Mössbauer spectroscopy. Catalysts ranked by bulk site density as measured by Mössbauer spectroscopy.
ORR site. The authors feel this a good assumption consid‐
Due to this relationship the ratio of MSD/MSDbulk repre‐
ering the success with which CO‐adsorption has been used
sents the catalysts utilization factor. As would be expected,
as sensor in precious‐metal catalysis. The second concern
this utilization is a strong function of porosity as shown in
mentioned, i.e. that it is possible that some ORR sites do
Fig. 8. Effectively, increasing porosity controls access to the
not bind CO, is addressed by referring readers to the body
catalysts surface. In this way the porosity induced by the
of literature that has indicated time and time again the
secondary nitrogen precursor plays a large role in catalyst
importance of metal content for ORR activity in acidic
utilization. Similar plots for the utilization of FeN4 centers
media
This observation is important because CO
assigned to D2 or the sum D1+D2 versus the BET surface
binds strongly to metals. Although it is true that non‐
area are shown in supporting information, Figure S3 d and
metallic catalysts can contain some ORR activity, their
f. It becomes evident from the comparison of all three pos‐
performance is generally limited to potentials below 0.7 V
sibilities that the FeN4 centers assigned to D1 gives by far
vs. RHE. Considering that our potential of interest is 0.8 V,
the best correlation. This is another indication, that it is
these active sites should not impact our calculations of TOF
(mainly) the D1‐related FeN4 site that is probed by CO
33, 35.
36. As further evidence of CO as a successful indicator of
sorption. This is understandable, taking into account that
ORR sites, we see that ORR activity increases with increas‐
the Mössbauer parameters of the D2 related FeN4 site are
ing CO‐adsorption as shown in Fig. 6. Representative ad‐
similar to iron phthalocyanine where a pseudo‐six‐fold
sorption traces for the various catalysts are shown in the
coordination leads to the quadrupole splitting. hence, it
supplementary information, Figure S4. Assuming that each
can be assumed that also for our D2 site the axial positions
active site adsorbs one CO molecule a mass‐based site
might
density (MSD) can be calculated from CO adsorption data
N
adsorption are shown in Fig. 7, in which the Mössbauer represents a bulk quantification of active sites and the CO‐ adsorption represents a surface sensitive quantification.
1.0
Utilization (D1) / %
10
The MSD values as calculated for Mössbauer and CO‐
1.2
be
available
for
gas
adsorption.
PANI-Urea
nmol mg
MSD sites per g
not
100
(nCO) by the following equation:
MSD / 1020 sites g-1
80 60
MSDsurf. (CO)
PANI PANI-Mel
40 PANI-NCB
PANI-CM
20 0
MSDbulk (D1)
0
100 200 300 400 500 600 700 800 900 1000 BET Surface Area / m2 g-1
Figure 8. Variation of catalyst utilization with porosity as
0.8
measured by BET surface area.
0.6
Although porosity plays a significant role in utilization,
0.4
there are other important factors to performance. For
0.2
example, from Fig. 8, we can see that a good catalyst (PANI‐CM) has low utilization, whereas a mediocre cata‐ PA N I-M el PA N I-N C B
PA N I-C M
PA N I
0.0 PA N I-U re a
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
Figure 7. Mass based site density of catalysts as measured by
lyst (PANI‐Urea) has high utilization. Part of the answer to this question is the role of the secondary nitrogen precur‐ sor in creating active sites.
ACS Paragon Plus Environment
ACS Catalysis
Role of secondary N-precursor on the quality of active sites
1.5
Some secondary nitrogen precursors result in high quantities of active sites such as PANI‐Mel, PANI‐NCB, and PANI‐CM as shown by the Mössbauer data in Fig. 7 by the MSDbulk parameters. PANI‐Mel and PANI‐CM likely create a lot of active sites
TOF / e- site-1 s-1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 12
PANI-CM
1.0
PANI-Urea
PANI-Mel
PANI
0.5
PANI-NCB
due to their high nitrogen contents. NCB may create a lot of active sites for a number of reasons. One hypothesis is that the low porosity of the material traps reactive gases during
0.0 3.0
3.5
the heat treatment. The smaller release of gaseous prod‐ ucts is also visible in the thermogravimetric analysis
4.0 4.5 N content / wt %
5.0
shown in the supporting information, Figure S5. The
Figure 9. Turn‐over‐frequency varying with bulk nitrogen content as measured from elemental analysis.
trapped reactive gases build up, creating high partial pres‐
Although it is difficult to ascertain why the TOF of the
sures which have been shown to improve ORR activity.37
PANI‐CM is significantly higher than the other catalysts,
But ultimately, these active sites are useless in RDE condi‐
the answer can be obtained partially from nitrogen con‐
tions due to the extremely low utilization. None‐the‐less, it
tent. A plot of TOF vs bulk nitrogen (Fig. 9) allows compar‐
is evident from these catalysts that the secondary nitrogen
ison of the intrinsic activity of the active sites to catalyst
precursor plays a role in both active site synthesis as well
nitrogen contents. TOF at 0.8 V vs RHE were calculated
as active site utilization as was discussed earlier.
from the adsorbed CO (nCO) and kinetic current (ik) by the
Now that active site synthesis and utilization have been
following equation:
discussed, turn‐over‐frequencies will be calculated to in‐ crease understanding of how the secondary nitrogen pre‐ cursors influence the intrinsic performance of active sites. A plot of kinetic current density from Fig. 3 vs CO adsorp‐ tion data indicates the relative intrinsic activity of these catalysts (Fig. 6). Specifically, the slope of a line from the origin to any of the data points represents a measure for the mean turn‐over‐frequency (TOF) of the CO adsorption sites for ORR. The lines in the Figure 6 represent the range of TOF that is evident in this work. For example, it is ob‐ servable that PANI‐CM has the highest TOF due to the slope of the line drawn in Fig. 6 through the PANI‐CM data point. Additionally, it is clear that the other catalysts have a similar, but lower TOF. The high TOF of PANI‐CM is calcu‐ lated due to a high ORR activity but low MSD (from CO sorption).
∙
∙
where n is electrons per mole equivalent (assumed to be four for ORR), and F is Faraday’s constant. It is evident that the PANI‐CM is significantly different than the other cata‐ lysts not only in TOF, but also in bulk nitrogen content. In fact, the best explanation for the high TOF of the cyana‐ mide catalyst is its high nitrogen content with respect to the other catalysts. Multiple works have indicated that nitrogen species in the vicinity of the active site may play an important role in iron center electronics.38‐41 Hence, nitrogen may play a subtle role in active site electronics as suggested by Kramm et al.41 and Ramaswamy et al.40. To further evaluate this, X‐ray induced photoelectron spec‐ troscopy was performed for the PANI‐catalysts prepared with secondary N‐precursor. The convoluted N1s spectra are shown in the Supporting Information, Figure S6. In Figure 10 the reference catalyst PANI is compared in terms of intensity with the four catalysts prepared with a sec‐ ondary N‐precursor.
ACS Paragon Plus Environment
Page 9 of 12
ing Information, Figure S7. Due to the use of a carbon sup‐ port (free or almost free of nitrogen) the nitrogen content
500 PANI (ref. 6)
Intensity / cps*
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
400 300
Npyrrol
PANI-CM PANI-MEL PANI-Urea PANI-NCB
is higher on the surface (XPS) compared to the bulk. How‐
Npyrid.
NMe-N
ever, while the slope is similar for the four catalysts, in case of PANI‐CM a different relation is observed. This indi‐ cates that the carbon burn‐off (related to the support) was
Ngraph
200
larger in comparison to the other samples. As TGA (Figure Noxid.
100
S5) indicates the same overall mass loss in comparison to PANI‐Urea, one has to conclude that the thermal induced
0 410
405
400
395
reactions appearing for PANI‐Urea (rather then the carbon in this sample) must have been more pronounced com‐
Binding energy / eV Figure 10. N1s finescans of the four catalysts prepared with a secondary N‐precursor in comparison to the PANI reference catalyst. Please note, XPS data on PANI were already reported in ref. 6, but measured with another spectrometer. For com‐ parison, the count rate per unit nitrogen was therefore changed accordingly to enable a direct comparison to the other samples (therefore the count rate is given as cps*).
pared to catalyst PANI‐CM. Maybe the changed ratio of N contents had a positive effect on the electronic interaction of active sites with the carbon. If so, however, one would not expect the correlation of TOF with the bulk nitrogen content. Further work will be required to identify the origin of improved TOF for PANI‐CM.
In this graph also the energetic positions for the obser‐ vation of the different nitrogen functionalities are indicat‐ ed. While the nitrogen content related to graphitic and oxi‐ dic nitrogen remains the same for all catalysts, mayor changes are visible in the lower binding energy region, related to Npyrrolic, NMe‐N and Npyrid.. It is interesting to note, that the PANI catalyst and PANI‐NCB signatures are close in their qualitative and quantitative behavior. Such a simi‐ larity can also be concluded from Figure 9. The other three samples lead to the incorporation of significant larger fractions of Npyrrolic, NMe‐N and Npyrid.. While all three of them can ‐ to some extend ‐ be attributed to different types of FeN4 sites, particularly pyridinic nitrogen is discussed to promote the ORR by an assisted proton transfer 42‐43. In‐ deed, the intensity in the Npyrid. region is one of the highest for PANI‐CM, but on the basis of the overall data set it can hardly be at the origin of the significant improved TOF. Further work will be required to get a better understand‐ ing of the origin of improved TOF for using cyanamide as secondary N‐precursor. In anyway, its very different be‐ havior in comparison to the other four catalysts also be‐ comes apparent from the comparison of bulk and “surface‐ related” (XPS) nitrogen contents, as shown in the Support‐
Conclusions Four different secondary nitrogen precursors were used to explore and deconvolute the complex relations between active site density, active site utilization, and active site turn‐over‐frequency of iron‐based Fe‐N‐C ORR fuel cell electrocatalysts. Using Mössbauer spectroscopy, X‐ray induced photoelectron spectroscopy, CO and N2 sorption studies we demonstrated how close to perfect utilization (active site dispersion) is offset by intrinsically less active catalytic surface sites (TOF values) and vice versa. We also demonstrated that the total nitrogen content not only increases the number of FeN4 sites, but in addition modu‐ lates their intrinsic activity (TOF) in agreement with pre‐ vious studies. Furthermore a few basic chemical mechanisms were un‐ covered in which such secondary precursors enhance the catalytic activity. One such mechanism consisted in signifi‐ cant etching of micropores (Urea and Melamine). Others work by creating a high density of active sites (Melamine and Cyanamide). Finally, it was discovered that one sec‐ ondary nitrogen precursor (Cyanamide) also incorporates significantly more nitrogen into the bulk catalyst frame‐ work, leading to higher TOF values, which explains the high values obtained for PANI‐CM catalyst. Our study clari‐
ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
fies how secondary nitrogen precursors serve as a flexible means of improving MNC catalyst performance by a num‐ ber of different phenomena.
ASSOCIATED CONTENT Supporting Information. Additional physical characteriza‐ tion, CO‐adsorption data, thermogravimetric data, and the remaining Mössbauer data are available in the supporting information.
AUTHOR INFORMATION Corresponding Author * Peter Strasser, email:pstrasser@tu‐berlin.de
ACKNOWLEDGMENT The authors would like to acknowledge Dr.‐Ing. Ralph Kräh‐ nert, Dr. Denis Bernsmeier, and Huan Wang, for their help with the nitrogen physisorption experiments. UIK and SW acknowledge financial support by the BMBF via the contract 05K16RD1 and by the graduate school of excellence Energy Science and Engineering (GRC1070). PS acknowledges sup‐ port by the BMWI and BMBF via contract “ChemEflex”.
REFERENCES
1. Strasser, P.; Kuhl, S., Nano Energy 2016, 29, 166-177. 2. Kuhl, S.; Strasser, P., Top Catal 2016, 59 (17-18), 1628-1637. 3. Beermann, V.; Gocyla, M.; Willinger, E.; Rudi, S.; Heggen, M.; Dunin-Borkowski, R. E.; Willinger, M. G.; Strasser, P., Nano Lett 2016, 16 (3), 1719-1725. 4. Tuaev, X.; Rudi, S.; Strasser, P., Catal Sci Technol 2016, 6 (23), 8276-8288. 5. Gang Wu, K. L. M., Christina M. Johnston, Piotr Zelenay, Science 2011, 332, 443447. 6. Sahraie, N. R.; Kramm, U. I.; Steinberg, J.; Zhang, Y. J.; Thomas, A.; Reier, T.; Paraknowitsch, J. P.; Strasser, P., Nat Commun 2015, 6. 7. Chung, H. T.; Holby, E. F.; Purdy, G. M.; Komini Babu, S.; Litster, S.; Cullen, D. A.; More, K. L.; Zelenay, P., Meeting Abstracts 2015, MA2015-02 (37), 1278. 8. Hoon Taek Chung, P. Z. Non-precious metal catalysts. 2014. 9. Lefevre, M.; Proietti, E.; Jaouen, F.; Dodelet, J. P., Science 2009, 324 (5923), 71-4. 10. Proietti, E.; Jaouen, F.; Lefevre, M.; Larouche, N.; Tian, J.; Herranz, J.; Dodelet, J. P., Nat Commun 2011, 2.
Page 10 of 12
11. Shui, J. L.; Chen, C.; Grabstanowicz, L.; Zhao, D.; Liu, D. J., P Natl Acad Sci USA 2015, 112 (34), 10629-10634. 12. Serov, A.; Robson, M. H.; Halevi, B.; Artyushkova, K.; Atanassov, P., Electrochem Commun 2012, 22, 53-56. 13. Herrmann, I.; Kramm, U. I.; Fiechter, S.; Bogdanoff, P., Electrochimica Acta 2009, 54 (18), 4275-4287. 14. Kramm, U. I.; Herrmann-Geppert, I.; Bogdanoff, P.; Fiechter, S., J Phys Chem C 2011, 115 (47), 23417-23427. 15. Leonard, N.; Nallathambi, V.; Barton, S. C., J. Electrochem. Soc. 2013, 160 (8), F788F792. 16. Wu, G.; More, K. L.; Johnston, C. M.; Zelenay, P., Science 2011, 332 (6028), 443-447. 17. Chung, H. T.; Zelenay, P. Non-Precious Metal Catalysts. 2014. 18. Chung, H. T.; Johnston, C. M.; Artyushkova, K.; Ferrandon, M.; Myers, D. J.; Zelenay, P., Electrochem Commun 2010, 12 (12), 1792-1795. 19. Nallathambi, V.; Leonard, N.; Kothandaraman, R.; Barton, S. C., Electrochem Solid St 2011, 14 (6), B55-B58. 20. Leonard, N. D.; Barton, S. C., J. Electrochem. Soc. 2014, 161 (13), H3100-H3105. 21. Nallathambi, V.; Leonard, N.; Barton, S. C., Abstr Pap Am Chem S 2011, 241. 22. Workman, M. J.; Dzara, M.; Ngo, C.; Pylypenko, S.; Serov, A.; McKinney, S.; Gordon, J.; Atanassov, P.; Artyushkova, K., J Power Sources 2017, 348, 30-39. 23. Sougrati, M. T.; Goellner, V.; Schuppert, A. K.; Stievano, L.; Jaouen, F., Catal Today 2016, 262, 110-120. 24. Koslowski, U. I.; Abs-Wurmbach, I.; Fiechter, S.; Bogdanoff, P., J Phys Chem C 2008, 112 (39), 15356-15366. 25. Kramm, U. I.; Herrmann-Geppert, I.; Behrends, J.; Lips, K.; Fiechter, S.; Bogdanoff, P., J Am Chem Soc 2016, 138 (2), 635-640. 26. Zitolo, A.; Goellner, V.; Armel, V.; Sougrati, M. T.; Mineva, T.; Stievano, L.; Fonda, E.; Jaouen, F., Nat Mater 2015, 14 (9), 937-+. 27. Kramm, U. I.; Herrmann-Geppert, I.; Fiechter, S.; Zehl, G.; Zizak, I.; Dorbandt, I.; Schmeisser, D.; Bogdanoff, P., J Mater Chem A 2014, 2 (8), 2663-2670.
ACS Paragon Plus Environment
Page 11 of 12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Catalysis
28. Ferrandon, M.; Kropf, A. J.; Myers, D. J.; Artyushkova, K.; Kramm, U.; Bogdanoff, P.; Wu, G.; Johnston, C. M.; Zelenay, P., The Journal of Physical Chemistry C 2012, 116 (30), 1600116013. 29. Borsa, D. M.; Boerma, D. O., Hyperfine Interact 2003, 151 (1), 31-48. 30. Gajbhiye, N. S.; Panda, R. N.; Ningthoujam, R. S.; Bhattacharyya, S., physica status solidi (c) 2004, 1 (12), 3252-3259. 31. Ramaswamy, N.; Tylus, U.; Jia, Q.; Mukerjee, S., J Am Chem Soc 2013, 135 (41), 15443-9. 32. Ju, W.; Bagger, A.; Hao, G. P.; Varela, A. S.; Sinev, I.; Bon, V.; Roldan Cuenya, B.; Kaskel, S.; Rossmeisl, J.; Strasser, P., Nat Commun 2017, 8 (1), 944. 33. Calle-Vallejo, F.; Martínez, J. I.; GarcíaLastra, J. M.; Abad, E.; Koper, M. T. M., Surface Science 2013, 607, 47-53. 34. Hammer, B.; Morikawa, Y.; Norskov, J. K., Phys Rev Lett 1996, 76 (12), 2141-2144. 35. Calle-Vallejo, F.; Martinez, J. I.; Rossmeisl, J., Physical chemistry chemical physics : PCCP 2011, 13 (34), 15639-43. 36. Dodelet, J.-P., The controversial role of the metal in Fe- or Co- based electrocatalysts for the oxygen reduction reaction in acid medium. In Lecture Notes in Energy - Electrocatalysis in
Fuel Cells, Shao, M., Ed. Springer: New York, 2013; Vol. 9, pp 271–338. 37. Gumeci, C.; Leonard, N.; Liu, Y. C.; McKinney, S.; Halevi, B.; Barton, S. C., J Mater Chem A 2015, 3 (43), 21494-21500. 38. Kramm, U. I.; Herranz, J.; Larouche, N.; Arruda, T. M.; Lefevre, M.; Jaouen, F.; Bogdanoff, P.; Fiechter, S.; Abs-Wurmbach, I.; Mukerjee, S.; Dodelet, J. P., Physical Chemistry Chemical Physics 2012, 14 (33), 11673-11688. 39. Kramm, U. I.; Lefevre, M.; Larouche, N.; Schmeisser, D.; Dodelet, J. P., J Am Chem Soc 2014, 136 (3), 978-985. 40. Ramaswamy, N.; Tylus, U.; Jia, Q. Y.; Mukerjee, S., J Am Chem Soc 2013, 135 (41), 15443-15449. 41. Kramm, U.; Abs-Wurmbach, I.; Herrmann-Geppert, I.; Radnik, J.; Fiechter, S.; Bogdanoff, P., J. Electrochem. Soc. 2011, 158 (1), B69-B78. 42. Busch, M.; Halck, N. B.; Kramm, U. I.; Siahrostami, S.; Krtil, P.; Rossmeisl, J., Nano Energy 2016, 29, 126–135. 43. Herranz, J.; Jaouen, F.; Lefevre, M.; Kramm, U. I.; Proietti, E.; Dodelet, J.-P.; Bogdanoff, P.; Fiechter, S.; Abs-Wurmbach, I.; Bertrand, P.; Arruda, T.; Mukerjee, S., J. Phys. Chem. C 2011, 115, 16087–16097.
ACS Paragon Plus Environment
ACS Catalysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 12 of 12
TOC Graphic
12
ACS Paragon Plus Environment
