In situ Ru K-edge EXAFS of CO adsorption on a Ru modified Pt/C fuel cell catalyst

Article abstract of DOI:10.1016/j.electacta.2009.03.060

The Ru-CO bond of CO adsorbed on a Ru modified Pt/C fuel cell catalyst has been directly probed by in situ EXAFS at the Ru K-edge, providing evidence of a CO:metal surface atom ratio greater than 1:1 and that CO is adsorbed at bridging sites associated wi

Full text of DOI:10.1016/j.electacta.2009.03.060

Electrochimica Acta 54 (2009) 5262–5266
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
In situ Ru K-edge EXAFS of CO adsorption on a Ru modified Pt/C fuel cell catalyst
Abigail Rosea,1, Robert Bilsborrowc, Colin R. Kinga, M.K. Ravikumarb, Yangdong Qianb,
Richard J.K. Wiltshirea, Eleanor M. Crabbb, Andrea E. Russella,∗
a School of Chemistry, University of Southampton, Highfield, Southampton, SO17 1BJ, UK
b Department of Chemistry, The Open University, Walton Hall, Milton Keynes, MK7 6AA, UK
c Science and Technology Facilities Council, Daresbury Laboratory, Warrington, WA4 4AD, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The Ru–CO bond of CO adsorbed on a Ru modified Pt/C fuel cell catalyst has been directly probed by in situ
EXAFS at the Ru K-edge, providing evidence of a CO:metal surface atom ratio greater than 1:1 and that CO
is adsorbed at bridging sites associated with Ru atoms at the surface of the catalyst nanoparticles. This
result illustrates the limitations of single crystal models as representations of the bonding of adsorbed
species at nanoparticle surfaces.
Received 23 November 2008
Received in revised form 28 February 2009
Accepted 21 March 2009
Available online 28 March 2009
© 2009 Elsevier Ltd. All rights reserved.
Keywords:
PtRu
Fuel cell catalysts
CO adsorption
Surface area
EXAFS
1. Introduction
fore, a 1:1 ratio between the number of adsorbed CO molecules and
metal surface atoms exists with a charge of 420 ␮C cm−2 being used
Assessment of the intrinsic activities of carbon supported fuel
cell catalysts relies on the in situ measurement of the true/accessible
surface area of the metal nanoparticles when formulated into an
ionomer bound electrocatalyst layer. In situ measurements are
essential, as variations in the preparation method and ionomer con-
tent as well as hydration of the electrocatalyst layer will determine
the fraction of the particles that are electrochemically accessible.
For Pt catalysts measurement of the charge associated with the
adsorption of a monolayer of hydrogen atoms is routinely used, with
one H atom corresponding to one surface Pt atom and an associated
charge of 210 ␮C cm−2 being used to calculate the Pt surface area [1].
As discussed by Green and Kucernak [2], H adsorption and stripping
is not suitable for the assessment of the area of many alloy catalysts
such as PtRu, as the charge associated with H and other oxidation
processes overlap or the H adsorption reaction is more complex due
to the formation of bronzes or the absorption of H into the metal
lattice. Thus, the electrooxidation of an adsorbed monolayer of CO
is used as the preferred method for the in situ measurement of the
surface area of such electrocatalysts [3]. It is generally assumed that
the CO is bonded linearly to a single metal surface atom and, there-
which is double that of the adsorbed H monolayer. This assumption
has recently been called into question by Green and Kucernak [2]
and ourselves [4], with a higher site occupancy of CO at the Ru
surface atoms being suggested.
In this paper in situ EXAFS is used to directly probe the bonding
of CO to Ru surface atoms on a Ru modified Pt/C electrocatalyst to
probe the validity of the 1:1 CO to surface metal atom assumption.
2. Experimental
Preparation of the Ru modified Pt/C catalyst by the reac-
tion of reduced 19.5 wt% Pt/C catalyst powder with a solution of
ruthenocene in n-heptane was conducted, as has been described
previously [4,5], to provide a coverage equivalent to 1.5 monolay-
ers of Ru on the Pt surface. Briefly, the 19.5 wt% Pt/C catalyst powder
was reduced by heating in H2 at 200 ◦C for 3 h and then exposed to
the required amount of ruthenocene dissolved in n-heptane to yield
a coverage of 0.75 monolayers of Ru based on the known dispersion
(fraction of surface atoms) of the base Pt/C catalyst. H2 was then
purged through the resulting dispersion for 24 h at room tempera-
ture, followed by 8 h at 98 ◦C. The mixture was filtered, washed with
clean n-heptane, and then dried in air. The process was repeated
to give a final (theoretical) coverage of 1.5 monolayers of Ru on
the Pt/C. The final ruthenium modified catalysts composition was
determined by ICP-AES and found to be 17.86 wt% Pt, 5.29 wt% Ru,
∗
Corresponding author. Tel.: +44 0 23 8059 3306; fax: +44 0 23 8059 3781.
E-mail address: a.e.russell@soton.ac.uk (A.E. Russell).
Present address: Physical Sciences Department, Dstl Porton Down, Salisbury SP4
1
0JQ, UK.
0013-4686/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2009.03.060

A. Rose et al. / Electrochimica Acta 54 (2009) 5262–5266
5263
which corresponds to a coverage of 1.4 monolayers and a Pt:Ru
atomic ratio of 1:0.59.
The resulting catalyst and the base Pt/C were characterised by
XRD, TEM, and CO chemisorption. The XRD patterns corresponded
to an fcc structure, characteristic of a PtRu alloy. No evidence of
a separate amorphous or crystalline Ru phase was found. Sherrer
analysis of the XRD pattern, corrected for the instrumental broad-
ening, of the Ru/Pt/C catalyst yielded an average crystallite size of
4.9 nm based on the arithmetic mean from four reflections, com-
pared to a diameter of 3.8 nm for the Pt/C. The average particle sizes,
determined by analysis of the TEM images (not shown), were 4.2
and 2.8 nm for the Ru/Pt/C and Pt/C catalysts, respectively. The CO
chemisorption measurements were conducted as follows. A quan-
tity of the dry catalyst powder was loaded into a quartz tube and
precisely weighed; 0.0607 g of the Pt/C and 0.0584 g of the Ru/Pt/C
catalyst was used. The tubes were loaded into the CO chemisorp-
tion apparatus, purged with He for a minimum of 30 min and then
pretreated by heating to 50 ◦C whilst purging with H2 for 30 min
followed by cooling to room temperature whilst maintaining the
H2 flow to remove any surface oxidation of the metal nanoparti-
cles, which occurs upon exposure and subsequent storage of the
catalysts in air. CO was then introduced and the uptake measured
volumetrically.
Electrodes were prepared from an ink containing the catalyst
powder (either the 19.5 wt% Pt/C base catalyst or the Ru modified
Pt/C) and an alcoholic solution of Nafion as a binder (to provide
15 wt% Nafion in the final electrode and a loading of 5 mg Pt cm−2
of prepared electrode area) which was spread onto wet proofed
carbon paper (E-TEK TGPH-90), and the resulting sheet pressed at
10 kg cm−2 for 3 min at 100 ◦C. Circular electrodes were cut from
the compressed sheet and boiled in triply distilled water to ensure
a fully flooded state prior to both the electrochemical and EXAFS
measurements.
The cells and methods used for collection of cyclic voltammo-
grams (CVs) and in situ EXAFS measurements (station 16.5 of the
SRS, Daresbury), as well as the data analysis techniques have been
described previously [4,6]. Both CVs and the EXAFS measurements
were conducted in 1 mol dm−3 H2SO4 solution. The EXAFS data was
collected in fluorescence mode using a 30-element Ge solid-state
detector. CO monolayer oxidation CVs were obtained by bubbling
CO through the electrolyte for 30 min at 0.05 V, followed by purg-
ing with N2 for a further 30 min before the CV was collected. EXAFS
data were obtained prior to CO exposure in un-purged electrolyte,
with CO adsorbed in the presence of CO saturated solution, and
after oxidative stripping of the adsorbed CO in un-purged solution
by exchanging the electrolyte in the EXAFS cell.
Fig. 1. Cyclic voltammograms of (a) Pt/C and (b) Ru/Pt/C electrodes recorded at
10 mV s−1 in 1.0 mol dm−3 H2SO4 at 25 ◦C. Black solid line before CO adsorption,
red dashed line CO stripping voltammogram and blue dotted line second cycle after
CO stripping. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of the article)
and Pt/C catalysts and a series of bimetallic PtRu/C catalysts, we
have shown that the peak potential for CO monolayer oxidation
decreases from 0.57 V for Ru/C to 0.48 V for 1:1 PtRu/C and then
increases again to 0.79 V for Pt/C [7]. The large capacitative current
and the 0.62 V peak potential for the oxidation of the CO monolayer
observed in the CVs of the Ru/Pt/C catalyst are, therefore, indicative
of a Ru rich PtRu surface. This suggests that some of the Ru is driven
below the surface of the metal particle following reduction of the
modified surface, either electrochemically or during the reduction
steps in the two deposition steps. This result is not surprising, as
it has been shown by Ruban et al. [8] that the surface segregation
energy for a Ru overlayer on a Pt substrate is such that the Ru will
preferentially move away from the surface, and is in agreement with
our earlier study of the effects of reduction on the composition of
the Ru modified Pt/C catalysts [4,5].
The CO monolayer oxidation CVs were used to determine the
surface areas of the catalyst electrodes by subtraction of the sec-
ond cycle CVs (black lines in Fig.1 a and b) from the first cycle
(red lines in Fig. 1a and b), in which the monolayer of adsorbed
CO was oxidised, thereby reducing the effects of anion adsorption
and capacitance, which have been commented on by Jusys et al. [9].
A charge of 420 ◦C cmP−t1 was assumed. The surface area normalised
to the mass of Pt on the electrode of the Pt/C base catalyst was
102 m2 gP−t1 and 154 m2 gP−t1 for the Ru modified Pt/C catalyst, which
3. Results and discussion
Cyclic voltammograms of both the Pt/C and the modified,
Ru/Pt/C, catalyst electrodes in 1 mol dm−3 H2SO4 are shown in Fig. 1.
As reported in an earlier study of such a Ru/Pt/C catalyst, the CV of
the modified catalyst is dominated by a relatively large capacita-
tive current compared to that of the Pt/C catalyst [5]. The hydrogen
adsorption/desorption peaks between 0.0 and 0.25 V are also much
less distinct in the Ru/Pt/C CV and the peak associated with oxide
stripping is shifted to 0.45 from 0.75 V observed in the Pt/C CV
(compare the black lines in Fig. 1a and B). As discussed in the intro-
duction, this loss of resolution in hydrogen region is characteristic
of PtRu alloy catalysts. The shift in the oxide stripping peaks is also
indicative of the presence of Ru at the surface of the catalyst par-
ticles. The position of the CO stripping peak (maximum) provides
further evidence of the surface composition of the Ru/Pt/C catalyst;
0.62 V for Ru/Pt/C and 0.79 V for Pt/C. In a recent study of the effects
of the composition of conventionally prepared monometallic Ru/C
corresponds to 110 m2 gP−t1+Ru. The surface areas were also deter-
mined by CO chemisorption. A value of 92.3 m2 gP−t1 was found for
the Pt/C, corresponding to a value of CO:M of 0.37. For the Ru/Pt/C
catalyst a value of 180.4 m2 gP−t1, corresponding to a CO:M value of
−1
0.66 based on the Pt content, which equates to 128.8 g
or a
Pt+Ru
value of 0.47 based on the Pt + Ru content. Both results indicate
a greater dispersion/decrease in particle size following modifica-
tion of the parent Pt/C catalyst by Ru, which is opposite to increase
in particle or crystallite size observed in both TEM and XRD data,
respectively. Previously we have interpreted this discrepancy as

5264
A. Rose et al. / Electrochimica Acta 54 (2009) 5262–5266
evidence of a breakdown of the assumption of a 1:1 CO:metal sur-
face atom ratio which is commonly employed in the analysis of both
the CO monolayer oxidation CVs and chemisorption data [5], an
assumption that has also been questioned by Green and Kucernak,
in a study of unsupported PtRu catalyts [2] where they attributed
the increase in surface area/dispersion to the adsorption of more
than one CO per Ru surface atom.
The coordination of CO to Ru in the Ru/Pt/C catalysts was further
investigated using in situ EXAFS measurements. Fig. 2 shows the Ru
K-edge XAS data collected at 0.0 V vs. RHE before purging the solu-
tion with CO, in the presence of a saturated CO solution, and after
removing the adsorbed CO from the electrode surface by oxida-
tive stripping. The inset highlights the differences in the relative
amplitudes of the edge peak, at 22,150 eV, and the peak of the first
oscillation, at 22,175 eV, as the conditions were varied. The increase
in the relative amplitude of the peak at 22,175 eV in the presence of
CO agrees well with an increase in the number of neighbours with
low atomic mass in the first coordination shell of the Ru, such as
the C of adsorbed CO and is in excellent agreement with the Ru K-
edge XANES study of CO adsorbed on a Pt50Ru50 catalyst electrode
surface by Aberdam et al. [10].
Fig. 2. Ru K edge XAS for Ru modified Pt/C at 0.0 V vs. RHE in 1 mol dm−3 H2SO4. (a)
before CO exposure, (b) with adsorbed CO, (c) after CO stripping.
Fig. 3. k weighted Ru K edge EXAFS (a, c, and e) and the corresponding Fourier transforms (b, d, and f) for Ru modified Pt/C at 0.0 V vs. RHE in 1 mol dm−3 H2SO4. (a and b)
before CO exposure, (c, d) with adsorbed CO, (e, f) after CO stripping. Data shown as thin solid lines and fits as dotted lines.

A. Rose et al. / Electrochimica Acta 54 (2009) 5262–5266
5265
Table 1
EXAFS fitting parameters for the Ru modified Pt/C catalyst electrode at 0.0 V vs. RHE as a function of the CO exposure. Data collected at the Ru K edge in situ in 1 mol dm−3
H2SO4. Standard deviations of the parameters given in brackets.
CO exposure
Before
Neighbour
N
R (Å)
2ꢀ2 (Å)2
Ef (eV)
Rexaf (%)
Ru–Ru
Ru–Pt
Ru–Pt
3.1 (0.2)
2.1 (0.2)
0.5 (0.9)
2.63 (0.01)
2.69 (0.01)
3.83 (0.16)
0.015 (0.002)
0.010 (0.002)
0.018 (0.033)
−12.1 (0.5)
37
Solution saturated with CO
After CO stripping
Ru–C
1.4 (0.3)
3.7 (0.3)
2.0 (0.3)
3.2 (1.6)
2.04 (0.02)
2.71 (0.01)
2.74 (0.01)
3.91 (0.04)
0.024 (0.008)
0.015 (0.001)
0.010 (0.002)
0.024 (0.008)
Ru–Ru
Ru–Pt
Ru–Pt
−1.5 (0.9)
−2.1 (0.6)
39
43
Ru–Ru
Ru–Pt
Ru–Pt
3.1 (0.2)
2.2 (0.2)
1.0 (1.1)
2.64 (0.01)
2.70 (0.01)
3.80 (0.07)
0.017 (0.001)
0.010 (0.002)
0.019 (0.015)
The isolated Ru K-edge EXAFS data and the corresponding
Fourier transforms are shown in Fig. 3. The data collected before
CO exposure and following the removal of the adsorbed CO yield
EXAFS and Fourier transforms that are identical within experimen-
tal error, indicating that no significant restructuring of the catalyst
particles accompanied the adsorption and removal of CO. Analysis
of the data, shown in Table 1, indicates that the average coordi-
nation surrounding the Ru atoms is composed of 3.1 Ru atoms at
2.63 Å and 2.1 Pt atoms at 2.70 Å in the first coordination shell. The
presence of Ru and Pt neighbours in the first coordination shell and
the absence of O or other neighbours of low atomic mass in the
EXAFS of the electrochemically reduced catalyst demonstrates that
the Ru is selectively deposited onto the supported Pt particles dur-
ing the preparation of the catalyst and is not present as a separate
amorphous Ru oxide phase within the experimental error [4].
In the presence of CO, the coordination numbers of the Ru and
Pt neighbours remain unchanged within the experimental error
from those in the absence of CO. In addition, 1.4 low atomic num-
ber neighbours, whom we have attributed to C of adsorbed CO,
are found at 2.04 Å. The structural parameters obtained from the
analysis of EXAFS data provide the average, per atom, coordination
environment surrounding the absorber atoms. Thus, a full mono-
layer of linearly bound CO adsorbed on particles with a surface of
equal numbers of Pt and Ru atoms should have a C coordination
number equal to the fraction of metal atoms on the surface based
on the assumption that there is one linearly bound CO per surface
metal atom; for example a M–C coordination number equal to 0.5
is expected for a catalyst with a dispersion of 50%, as we found
previously for CO adsorbed on a Pt/C catalyst [6]. As indicated by
the CO monolayer oxidation peak, the surface of the Ru/Pt/C cat-
alyst is enriched in Ru compared to 1:1 PtRu/C alloy catalyst and,
thus, the predicted coordination number is expected to be slightly
greater than the CO:M ratios determined from the CO chemisorp-
tion measurements. Even if all the Ru atoms were segregated to the
surface (an assumption not supported by the metal coordination
numbers obtained by analysis of the EXAFS data) the Ru–C coor-
dination number cannot be greater than 1 for a full monolayer of
linearly adsorbed CO molecules.
of linear and bridging CO ligands encapsulate the metal cluster. Such
bridging CO ligands, in particular, stabilise the formation of M–M
bonds. The metal nanoparticles comprising the Ru/Pt/C catalyst in
the current study are considerably larger than the cores of such
cluster compounds and it is hard to envisage a surface covered by
bridge bound CO species. However, if the adsorbed CO molecules
at the Ru atoms at the edges of the faces of the particles were to
be bridge bound, this would make up a considerable fraction of the
total number of Ru–C interactions. Additionally, more than one CO
molecule may be bound to the individual Ru atoms at the verticies
of the nanoparticles, as is the case for the cluster compounds. Such
a high coverage of CO is likely to have an effect on the metal–metal
distance of the outermost layer of metal atoms in the nanoparticle
and, indeed, we observe a slight increase in the Ru–Ru and Ru–Pt
distances compared to those obtained in the absence of adsorbed
CO.
4. Conclusions
The surface areas obtained using either CO chemisorption or
the electrochemical stripping of an adsorbed CO monolayer are
commonly used to characterise carbon supported Pt and Pt based
alloy electrocatalysts and the results are often used to normalise
electrocatalytic activity to obtain specific activities. In calculating
the surface areas from such measurements assumptions are made
regarding the adsorption site and/or site occupancy of adsorbed
CO and the coverage that constitutes a full monolayer. The num-
bers employed are based on careful studies of the adsorption of
CO on single crystal surfaces, both in UHV and in the electro-
chemical environment. The combined electrochemical and EXAFS
results presented above provide direct evidence that the picture
of adsorbed CO derived from such single crystal studies may not
be directly translatable to the realm of supported nanoparticle
electrocatalysts and may provide inaccurate determinations of the
surface area and/or the size of the nanoparticles comprising the
catalyst. In the case of the Ru/Pt/C catalyst described above, the
areas determined from both CO chemisorption and the CO stripping
voltammogram are likely to be over-estimations of the electro-
chemically active area of the catalyst.
The distances obtained from the EXAFS analysis provide fur-
ther insights into the coordination of CO to the catalyst surface.
In our previous study of CO adsorption on a Pt/C catalyst, a Pt–C
distance of 1.85 Å was found and attributed to linearly bound CO.
The 2.04 Å distance for the Ru–C shell found in the current study
is more indicative of bridge bound than linearly bound CO. In an
EXAFS study of a ruthenium carbonyl hydroformylation catalyst,
Evans et al. obtained Ru–C distances of 1.94 Å for linearly bound CO
and 2.17 Å for the bridge bound species [11]. Such bridging coordina-
tion is typical of metal carbonyl clusters [12], where a combination
Acknowledgements
The EPSRC are thanked for funding studentships for AR, CRK, and
RJKW and postdoctoral fellowship for MKR, and CLRC for access to
the SRS. The Open University is thanked for funding the studentship
for YQ. Johnson Matthey is also thanked for their financial sup-
port. The technical assistance of Chris Corrigan at the SRS is also
appreciated.

5266
A. Rose et al. / Electrochimica Acta 54 (2009) 5262–5266
[6] S. Maniguet, R.J. Mathew, A.E. Russell, J. Phys. Chem. B 104 (2000) 1998.
References
[7] R.J.K. Wiltshire, C.R. King, A. Rose, P.P. Wells, H. Davies, M.P. Hogarth, D.
Thompsett, B. Theobald, F.W. Mosselmans, M. Roberts, A.E. Russell, Phys. Chem.
Chem. Phys. 11 (2009) 2305.
[8] A.V. Ruban, H.L. Skriver, J.K. Norskov, Phys. Rev. B 59 (1999) 15990.
[9] Z. Jusys, J. Kaiser, R.J. Behm, Electrochim. Acta 47 (2002) 3693.
[10] D. Aberdam, R. Durand, R. Faure, F. Gloaguen, J.L. Hazemann, E. Herrero, A.
Kabbabi, O. Ulrich, J. Electroanal. Chem. 398 (1995) 43.
[11] J. Evans, G. Jingxing, H. Leach, A.C. Street, J. Organomet. Chem. 372 (1989) 61.
[12] C. Femoni, M.C. Iapalucci, F. Kaswalder, G. Longoni, S. Zacchini, Coord. Chem.
Rev. 250 (2006) 1580.
[1] A. Essalik, K. Amouzegar, O. Savagogo, J. Appl. Electrochem. 25 (1995) 404.
[2] C.L. Green, A. Kucernak, J. Phys. Chem. B 106 (2002) 1036;
C.L. Green, A. Kucernak, J. Phys. Chem. B 106 (2002) 11446.
[3] F.C. Nart, W. Vielstich, Normalisation of porous active surface areas, in: W. Viel-
stich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells Electrocatalysis,
vol. 2, John Wiley and Sons, N.Y. USA, 2003, pp. 302–315.
[4] E.M. Crabb, M.K. Ravikumar, D. Thompsett, M. Hurford, A. Rose, A.E. Russell,
Phys. Chem. Chem. Phys. 6 (2004) 1792.
[5] A. Rose, E.M. Crabb, Y. Qian, M.K. Ravikumar, P.P. Wells, R.J.K. Wiltshire, J. Yao,
R. Bilsborrow, F. Mosselmans, A.E. Russell, Electrochim. Acta 52 (2007) 5556.