Templating of electrodeposited platinum group metals as a tool to control catalytic activity

Article abstract of DOI:10.1016/j.electacta.2007.06.043

The present study is focused on the synthesis and investigation of Pt and Pt-Ru nanostructures templated by anodic aluminum oxide films. Samples of this sort may be considered in future as model electrode materials of high roughness and specific controllable nanostructural features. Nanostructure and morphology of electrodeposited materials are characterized using scanning electron microscopy and scanning tunneling microscopy techniques. A possibility to control catalytic activity of electrodeposits by means of templating is discussed as well. Higher catalytic activity of nanostructured Pt-Ru as compared to usual electrodeposit is very promising for further clarification of structural effects in electrocatalysis. In particular, newly proposed materials combine high roughness and the increased number of grain boundaries known to provide high catalytic activity of platinum group metals in certain processes. Future prospects of obtaining materials with even higher roughness factors are discussed.

Full text of DOI:10.1016/j.electacta.2007.06.043

Electrochimica Acta 52 (2007) 7910–7919
Templating of electrodeposited platinum group metals as a
tool to control catalytic activity
a,∗
b
a
a
K.S. Napolskii , P.J. Barczuk , S.Yu. Vassiliev , A.G. Veresov ,
a,1
b,1
G.A. Tsirlina , P.J. Kulesza
a
Moscow State University, Moscow 119992, Russia
b
University of Warsaw, Pasteura 1, PL-02-093 Warsaw, Poland
Received 17 February 2007; received in revised form 19 May 2007; accepted 18 June 2007
Available online 23 June 2007
Abstract
The present study is focused on the synthesis and investigation of Pt and Pt-Ru nanostructures templated by anodic aluminum oxide films.
Samples of this sort may be considered in future as model electrode materials of high roughness and specific controllable nanostructural fea-
tures. Nanostructure and morphology of electrodeposited materials are characterized using scanning electron microscopy and scanning tunneling
microscopy techniques. A possibility to control catalytic activity of electrodeposits by means of templating is discussed as well. Higher catalytic
activity of nanostructured Pt-Ru as compared to usual electrodeposit is very promising for further clarification of structural effects in electrocataly-
sis. In particular, newly proposed materials combine high roughness and the increased number of grain boundaries known to provide high catalytic
activity of platinum group metals in certain processes. Future prospects of obtaining materials with even higher roughness factors are discussed.
©
2007 Elsevier Ltd. All rights reserved.
Keywords: Anodic aluminum oxide; Nanoporous alumina; Templated electrodeposition; Nanowire; Platinum; Pt-Ru; Electrocatalysis
1
. Introduction
of electrodeposited platinum metals is largely nanostructural.
Recent XRD and microscopic studies discovered at least two
rarely discussed features of these materials, namely the pro-
nounced lattice compression and very specific grain boundaries
keeping large arrays of nanocrystals altogether [11,12]. One of
the simplest tools to affect these nanostructural details is deposi-
tion potential, and for platinum it appears to be a good instrument
changing the compression and coalescence degree in a wide
range [11]. Unfortunately, the possibilities to control Pt-Ru
nanostructure by the same tool are rather limited because start-
ing from some potential Ru tends to be incompletely reduced
[12]. This is the reason why we apply below another tool to
affect nanostructure, namely templated deposition.
Electrodeposited noble metals are known to demonstrate
rather high electrocatalytic activity in fuel cell reactions. Among
these metals platinum shows the highest catalytic activity
towards methanol electrooxidation, but its activity suffers from
CO poisoning, therefore Pt-Ru bifunctional system is consid-
ered for a long time as more promising. Pt-Ru electrocatalysis
was first observed just on electrolytically co-deposited films [1].
Many attempts were reported to further improve Pt-Ru cata-
lyst for methanol fuel cells by changing metal morphology and
structure, i.e., comparing various alloys [2,3], Pt-Ru co-deposits
[
4,5], Ru adsorbed on Pt [6,7], Ru spontaneous adsorbed on Pt
single crystals [8,9] or platinum nanoparticles [10]. The major-
ity of these attempts have not resulted in higher activity towards
methanol electrooxidation as compared to activity of electrode-
posited Pt-Ru. The reason of the unique catalytic properties
Anodic aluminum oxide membranes formed by two-step
anodization technique [13] or nanoimprint technology [14] are
well-known to possess uniform pore structure with hexagonal
arrangement of cylindrical channels. The use of different elec-
trolytes, voltages and anodic oxidation times allows one to adjust
pore diameter Dp, interpore distance Dint and film thicknesses
Lf in a wide range (Dp = 15–200 nm; Dint = 50–500 nm; Lf up
to several hundreds of micrometers) [15–17]. Metal templat-
ing with the use of anodic alumina is a widely used approach
∗
Corresponding author. Tel.: +7 495 9395248; fax: +7 495 9390998.
E-mail address: napolsky@inorg.chem.msu.ru (K.S. Napolskii).
ISE member.
1
0
013-4686/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2007.06.043

K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
7911
[
18–22], but it was mostly applied to less noble metals forming
electrodeposits of low roughness. Pt deposition into alumina
matrix was recently reported [23], however the voltammogram
of the resulting material demonstrated low purity of metal sur-
face and negligible specific surface area even after removing
alumina matrix. It should be mentioned that Pt fabricated in
Ref. [23] contained rather low fraction of nanowires attached to
usual electrodeposit. In our approach electrodeposited platinum
metals form nanowires on gold substrate without any additional
fragments of different structure.
We believe that coexistence with inert oxide matrix does
not obligatory result in contamination of platinum group metal.
Actually when Pt-Ru nanoparticles are immobilized in tita-
nia nanotubes, a normal electrochemical behavior is observed
[
24,25]. We also hope that electrodeposited Pt metal can demon-
Fig. 1. SEM image of AAO membrane (top view). Inset: cross-section of porous
structure.
strate much higher roughness than that reported in Ref. [23].
We compare electrodeposited Pt and Pt-Ru on gold fabricated
withandwithoutaluminananotubularmatrix, byconsideringthe
surface areas and the specific activities towards methanol elec-
trooxidation, in order to estimate the prospects and limitations
of template procedure for design of well-characterized catalysts
of high roughness.
Au/Al2O3 thin porous film were isolated with Apiezon Wax W
(BOC-Edwards).
It should be mentioned that homogeneity of the pore fill-
ing by the metal strongly depends on availability of the pores
for electrodeposition. Immersing the porous electrode into elec-
trolyte can lead to formation of air bubbles inside the narrow
pores. Prior to metal electrodeposition, vacuum pre-treatment
of membrane in electrolyte solution was performed, in order to
prevent blocking of pores with air bubbles. The samples were put
into the test-tube with electrolyte solution connected to water-jet
pump (residual pressure ca. 10 mmHg). After 30 min of pump-
ing the electrodes were moved to electrochemical cell for metal
deposition.
2
. Experimental
2
.1. Preparation of porous alumina matrices
For preparation of anodic aluminum oxide (AAO) mem-
branes, two-step anodization technique was applied as described
elsewhere [15]. High purity aluminum foil (99.999%, 0.5 mm
thick, Goodfellow) was used as the starting material. Prior to
Electrodeposition of platinum was carried out in three-
electrode cell from 0.01 M Na2PtCl6 + 0.02 M HCl solution
◦
anodization aluminum was annealed at 500 C in air in order
2
to remove the mechanical stress and to enhance the grain
size in the metal. Subsequently, the foils were mechanically
polished to a mirror finish and cleaned repeatedly with ace-
tone and deionized water. The anodization was carried out
in two-electrode cell in 0.3 M oxalic acid at constant volt-
age 40 V using Pt foil as a counter electrode. The electrolyte
was rigorously stirred, and its temperature was kept in the
range of 0–5 C during anodization. After first anodization for
4
3
tion under the same conditions for 50 h the oxide layer with
thicknesses of about 100 ␮m was obtained. To separate the
oxide film from the substrate the remaining aluminum was dis-
solved using a solution of 10 vol.% Br2 in methanol at room
temperature. Subsequently, the pore bottoms were opened by
at Ed = 0.3 V versus RHE reference electrode using Autolab
PGSTAT 100 potentiostat. The counter electrode was a Pt wire,
and the reference electrode was saturated Ag/AgCl electrode
connected with a cell via a Luggin capillary. The obtained
nanocomposites of Pt/Al2O3 are denoted as AAO Pt.
The co-deposition of Pt-Ru was performed at con-
stant Ed = 0.05 and 0.2 V from 0.015 M Na2PtCl6 + 0.015 M
RuCl3 + 1 M HCl solution in three-electrode cell with SCE con-
nected with a cell via a Luggin capillary as a reference electrode
and platinum foil as a counter electrode using EG&G PARC 273
potentiostat. These samples are denoted below as AAO Pt-Ru.
◦
8 h alumina film was selectively etched away in a mixture of
◦
5 ml/l H3PO4 and 20 g/l CrO3 at 70 C. After second anodiza-
−2
The total charge spent for Pt deposition was 1–26 C cm
,
in order to prepare arrays of metal nanowires of various aspect
ratios. Under assumption of 100% current efficiency and deposit
−
3
density 21.5 g cm these charges correspond to the nanowire
◦
chemical etching in a 5 vol.% H3PO4 water solution at 60 C for
5
length from 0.8 to 20.6 ␮m. For Pt-Ru deposition, the charges
−
2
min.
were 3.9 or 7.8 C cm
.
2
.2. Preparation of Pt and Pt-Ru nanostructures inside
2.3. Samples characterization
anodic alumina membranes
ElectrochemicalcharacterizationofAAO PtandAAO Pt-Ru
samples was performed in 0.5 M H2SO4 solution after deaera-
In order to prepare Pt or Pt-Ru nanowire arrays by electrode-
position technique a layer of Au was sputtered onto one side of
the AAO template to serve as working electrode. Glassy carbon
plates used as current collectors and mechanical supports for
2
Here and below all potentials are reported in RHE scale.

7
912
K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
Fig. 2. SEM images of Pt nanowires: cross-sectional images (a and b) and images of nanowires after matrix dissolution (c–e). Deposition charge 26.1 C cm 2.
−
tion with Ar using Autolab PGSTAT 100 and EG&G PARC
73 potentiostats. True surface areas S of Pt for AA Pt samples
tial step was determined by reaching the steady-state (<1% per
min) current value.
2
were determined coulometrically from hydrogen desorption or
CO stripping regions of cyclic voltammograms measured in a
Highresolutionscanningelectronmicroscopywasperformed
using SEM LEO Supra 50 VP with microanalytical system
INCA Energy + Oxford.
−
1
0
.5 M H2SO4 solution at scan rates from 10 to 200 mV s [26].
To arrange the latter technique, CO was bubbled through the
electrolyte for 45 min while the electrode was kept at a con-
stant potential of 0.14 V versus RHE, then dissolved CO was
removed from electrolyte by purging the solution with Ar for
STM imaging was performed ex situ using commercial
scanning tunneling microscope with extended spectroscopic
facilities “UMKA” (Concern “Nanoindustry”, Moscow). Pt-Ir
tip (10% Ir) of 0.5 mm diameter was mechanically sharpened.
For topographic measurements, −0.1 V tunneling voltage was
applied (negative voltage corresponds to positive tip polar-
ization), in combination with 0.3–1.0 nA tunneling current
(specially chosen for different types of samples).
9
0 min.
Methanol electrooxidation tests were performed in 0.5 M
H2SO4 + 0.1 M CH3OH solution under chronoamperometric
mode in 0.3–0.8 V potential range. Switching to the next poten-

K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
7913
Fig. 3. Cyclic voltammograms recorded for platinum electrodeposited in alumina matrix: before CO adsorption (a) and CO stripping (b). Scan rate 20 mV s 1,
−
−
2
electrolyte 0.5 M H2SO4. Deposition charge 12.1 C cm
.
Prior to the all STM and several SEM measurements the
nanocompositesAAO PtandAAO Pt-Ruwereimmersedin1 M
NaOH or H2SO4 solution at least for 24 h in order to completely
dissolve alumina matrix, and thereafter carefully washed with
deionized water and dried in air.
It should be noted that in the course of matrix dissolution
and subsequent washing of samples with deionized water single
nanowires or their bunches appeared to break off, forming the
empty regions on Au support (Fig. 2d).
For electrochemical characterization of AAO Pt cyclic
voltammograms in 0.5 M H2SO4 solution were recorded. Before
adsorption of CO monolayer Pt surface was surely con-
taminated: CV curves with unseparated peaks of hydrogen
3
. Results
According to direct SEM observations all synthesized anodic
alumina membranes possess the ordered porous structure with
uniform pores aligned perpendicular to the film surface (Fig. 1).
The pores form two-dimensional hcp structure with the inter-
pore distance (D ) of 105 nm, that is typical for the applied
int
anodization conditions (0.3 M (COOH)2, 40 V) [17]. Besides
the film thicknesses depend on the second anodization time: the
longer the time of anodization, the porous film of higher thick-
ness is formed. It is well documented that the pore diameter in
oxide layer is approximately constant (Dp = 35 nm) throughout
all their length, with a slight increase in the upper part of mem-
brane. The procedure of etching of the barrier layer increases
the pore diameter slightly (up to 50 nm near external surface,
and up to 40 nm near the bottom for 100 ␮m thick membrane).
Deposition transients demonstrate a limited reproducibility
of current values, especially in initial period. However these
transients are always of similar smoothly growing shape, with
a sharp increase of current only after filling some pores com-
pletely (when metal growth starts at the external matrix/solution
interface). At this stage we consider only growth inside alumina
template.
Fig. 2a demonstrates height distribution of electrodeposited
platinum nanowires. In general all pores are filled simultane-
ously with close rates. However about 1–2% of wires probably
undergo much faster growth (Fig. 2b). According to SEM obser-
vations after deleting the matrix all synthesized samples consist
of Pt nanowires which diameters are very close to pore diameters
in Al2O3 matrix (see Fig. 2c–e). The average diameter of the Pt
nanowires is ca. 50–60 nm, and their length increases with the
charge spent for electrodeposition. We are inspired by mechan-
ical stability of nanowires, which remain stable without matrix
and demonstrate the pronounced elasticity (Fig. 2c).
Fig. 4. True surface area of platinum determined from hydrogen desorption
(
circles) and CO desorption (triangles) charge (a) and methanol electrooxidation
current normalized per true surface area (E = 0.65 V) (b) versus charge spent for
electrodeposition. Dashed line in Fig. 4(b) corresponds to the value for platinized
gold [29].

7
914
K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
adsorption/desorption were always registered (see Fig. 3a). We
assume that contaminants are various residual chemical sub-
stances used in the course of matrix preparation. After CO
stripping process typical CVs for pure Pt deposits are clearly
seen (Fig. 3b). Thereby the adsorption and subsequent stripping
of CO leads to purification of Pt surface. AAO Pt samples of
various loading demonstrated a slight difference in the shape
and position of CO stripping peak, but practically no difference
in hydrogen and oxygen adsorption regions. Using CV curves,
the true surface areas of electrodeposited Pt are determined
coulometrically from both adsorbed H and CO.
tration into Al2O3 matrix, but the thickness dependence can also
reflect heterogeneity of the wires structure at nanoscale, i.e., dif-
ferent degree of crystal coalescence or difference in crystal size
2
−1
along the wire. However S ≈ 5–7 m g for short nanowires can
be explained only under assumption that Pt inside the pores is
dispersed. For a set of smooth nanowires of 60 nm diameter, the
2
−1
expected S values do not exceed 3 m g . This value is rather
close to experimentally observed S for high loadings (Fig. 4a).
The reported S values can be underestimated. If being so,
one should assume that both short and long nanowires are dis-
persed. Unfortunately at this stage we failed to compare the
specific surface areas measured before and after alumina disso-
lution because it is difficult to determine carefully the amount
of deposit lost in the course of deleting matrix.
The properties of usual Pt deposited on gold are studied in
detail. The specific surface areas observed for these deposits are
2
−1
typically 10–20 m g [11]. The dependence of the true surface
area determined for platinum in AAO Pt versus charge spent for
its deposition is plotted in Fig. 4a. It is worth noting that the val-
ues of the specific surface area of Pt electrocrystallized inside the
alumina matrix are significantly lower than for usual Pt/Au elec-
trodeposits. Both low S values and their decrease with thickness
can be surely explained by some limitations of electrolyte pene-
In order to obtain more information about the size and mutual
location of Pt crystallites and the morphology of the electrode-
posited metal inside the porous matrix the STM investigation
was performed.
According to STM studies, two completely different motives
of crystal coalescence inside the Pt nanowires can be observed.
Fig. 5. STM images of Pt nanowires after matrix dissolution (a and b) and calculated particle size distribution (c and d).

K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
7915
Fig. 6. Electrooxidation of methanol on Pt electrodeposited in alumina matrix.
−
2
−2
−2
Deposition charge: 1.4 C cm
(circles), 1.6 C cm
(squares), 5.8 C cm
(diamonds). Deposition
−2
−2
(
stars), 12.1 C cm
(triangles) and 18.7 C cm
charges are normalized per geometric surface area. Dashed curve corresponds
to the most active Pt deposit on Au [29] (in the series of samples fabricated
at different potentials). Electrolyte 0.5 M H2SO4 + 0.1 M CH3OH. Currents are
normalized per true surface area.
Fig. 5 shows STM images (a and b) and particle size distribu-
tion obtained by statistical treatment of these images (c and d).
Usually Pt wires consist of small quasi spherical particles with
average diameter of 8 nm, but in some regions the layered struc-
ture was observed. Data on more than 400 spherical and 200
layered composite particles were treated to construct particle
size distribution. Average thickness of the layers is a little bit
lower than maximum of the size distribution of spherical par-
ticles. In general these distributions are similar to distributions
reported earlier for Pt/Pt [27], the material of higher specific
surface area. One can conclude that the main nanostructural dif-
ference of templated and usual electrodeposited Pt consist not
in crystal size, but in degree of coalescence.
Theformationoflayeredstructureisaninterestingnewobser-
vation which should be in future applied to better understanding
of secondary nucleation phenomenon. It is already quite clear
that the effect of diffusion limitations on the growth of primary
crystals and on the rate of secondary nucleation is quantitatively
different. Well defined layered fragments are formed if the for-
mation of 2D Pt particle with a size close to pore diameter from
primary crystals takes place faster than secondary nucleation
happens.
Fig. 7. SEM images of Pt-Ru nanowires electrodeposited at 0.05 V: top view
after matrix dissolution (a) and cross-sectional images (b and c). Deposition
−
2
charge 7.9 C cm
.
considered as additional tool to increase R ; even if nanocrystals
f
Unfortunately, we could not establish in what part of the
nanowires (near the bottom, in the mid, or in their top part)
the layered structure is formed. The true nature of the observed
indirect influence of porous matrix on the electrodeposits mor-
phology requires future investigation.
are completely coalesced, the expected S for the smallest pos-
sible 15 nm pore diameter is 12 m g (already very close to
2
−1
usual Pt electrodeposits). The highest possible R estimated for
f
realisticmatrixthicknessof200 ␮misca. 12,000. Thisvaluecor-
responds to Dint = 50 nm and Dp = 40 nm (parameters which can
be obtained with the use of sulfuric acid for aluminum anodiza-
tion). If to rely on pulse anodization techniques providing wider
Roughness factors R achieved for the longest nanowires in
f
AAO Pt do not exceed 60. The reasons are the relatively low
porosity of matrix and the relatively low nanowires length. How-
range of pore parameters [28], one can even expect R up to
f
ever we expect no problems with increasing R by applying
20,000–30,000.
f
higher deposition charges (larger nanowires length) as well as by
We should stress that all the estimates given above correspond
using matrices with smaller D . The specific surface area can be
to the lower limit of S and R values. Additional increase of both
f
int

7
916
K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
Fig. 8. Cyclic voltammetry response of Pt-Ru electrodeposited on the gold
at 0.05 V (solid) and 0.2 V (dashed). Inset: dashed line—fresh sample; solid
line—sample after 2 days in air (first cycle). Electrolyte 0.5 M H2SO4, scan rate
Fig. 9. Cyclic voltammetry response of the Pt-Ru electrodeposited into alu-
−
2
−2
mina matrix at charge 3.9 C cm (solid) and 7.8 C cm (dashed). Dotted curve
corresponds to cyclic voltammogram of the Pt-Ru electrodeposited on gold at
−
1
2
0 mV s
.
−
2
3
1
.9 C cm . Deposition potential 0.05 V. Electrolyte 0.5 M H2SO4, scan rate
0 mV s 1.
−
valuescanbeprovidedbyincompletecrystalscoalescenceinside
nanowires.
The prospects of AAO Pt composites as model materials
of high roughness for electrochemical surface science and
electrocatalysis depend on representativity of their electrocat-
alytic behavior. This is why we compare below the steady-state
methanol oxidation on AAO Pt and usual Pt electrodeposits.
by chemical etching of alumina in acidic solution and increase
of pores diameter in parallel with platinum and ruthenium co-
deposition.
Our attempts to image AAO Pt-Ru cross-sections were com-
plicated by low mechanical stability of matrix. Surprisingly,
the upper fragments with still empty pores were more stable,
when the bottom regions containing Pt-Ru deposit were mostly
crashed (Fig. 7b and c). We suppose that not simply acid, but (in
addition) the excess of chloride ions generated in the course of
deposition is responsible for alumina damage. These results are
very important for future optimization of deposition solutions.
As nanowires remain stable over after matrix dissolution
(Fig. 7a), one can be sure that they form well-ordered arrays
inside AAO Pt-Ru samples. That is why we assign electrochem-
ical results presented below to nanostructured Pt-Ru with taking
into account possible local break of walls between the neighbor-
Fig.
6 presents steady-state polarization curves of
methanol electrooxidation versus applied potential in 0.1 M
CH3OH + 0.5 M H2SO4 solution. Dashed curve corresponds to
the most active Pt deposit on Au [29] (from the series of samples
fabricated at different potentials). Rather high catalytic activity
confirms that alumina matrix does not poison nanowires at all.
Au contribution is negligible in potential range under consider-
ation, as it was checked experimentally in our previous studies.
The specific activity of Pt is practically independent on the
wires length (or Pt loading). For the most short nanowires, a
noticeable scatter was observed (Fig. 4b), but the mean val-
ues demonstrate no deviations from horizontal line, which is a
steady-state value of methanol electrooxidation current for pla-
tinized gold [29]. The latter, in its turn, is very close to activity
of platinized platinum [30], so there are no reasons to assume
any effect of Pt–Au alloy (known only for lower potentials (see
[
31] and References therein)). Figs. 4 and 6 confirm that sur-
face of the nanowires throughout all their length is involved in
electrooxidation process, and the specific activities of various
nanostructural fragments are close.
Further examination involved preparation of similar
nanostructured film by means of platinum and ruthenium co-
deposition. This attempt was risky because of necessity to use
more acidic deposition bath. Our observations confirm indirectly
that alumina matrix was less stable in this bath as compared to
weakly acidic platinum deposition solution. Fig. 7a illustrates
SEM data for AAO Pt-Ru nanocomposite after matrix dissolu-
tion. All synthesized samples consist of Pt-Ru nanowires with
uniform diameter and length. The average diameter of Pt-Ru
nanowires is ca. 70–80 nm, i.e., it exceeds the initial diameter
of pores in alumina matrix. This fact can be probably explained
Fig. 10. Electrooxidation of methanol on Pt-Ru electrodeposited on gold at
0.05 V (squares), 0.2 V (triangles) and inside the alumina matrix at 0.05 V (cir-
cles). Electrolyte 0.5 M H2SO4 + 0.1 M CH3OH. Currents are normalized per
surface area estimated from hydrogen desorption charge.

K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
7917
is independent on scan rate in the range of 5–100 mV s 1,
demonstrating fast reversible recharging of Pt-Ru. For a sam-
ple electrodeposited at 0.05 V we obtained sharper peaks in the
hydrogen adsorption region and higher current at the potentials
ofthe“doublelayer”regionthanforthesampleobtainedat0.2 V.
−
ing pores and corresponding coalescence of some certain wires;
deposition of small Pt-Ru portions between Au and damaged
bottom fragments of matrix also can not be excluded.
Cyclic voltammetry data of Pt-Ru electrodeposited on gold
are shown in Fig. 8. Charge in the hydrogen adsorption region
Fig. 11. STM images of Pt-Ru electrodeposited on gold at 0.05 V, fresh sample (a and b) and sample aged for 5 days (c and d); sample deposited at 0.2 V and aged
for few days (e and f).

7
918
K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
These can be explained by higher current efficiency of Pt-Ru
deposition at lower potential and probably by higher roughness
UPD technique within the accuracy of 10–20%. This is quite
enough to conclude the difference mentioned above, but proba-
bly insufficient for comparison with a sample electrodeposited
at 0.2 V.
Significance of the rough nanostructured catalytic surface
is even more pronounced if we compare methanol electroox-
idation currents for AAO Pt-Ru and colloid nanoparticles on
the TiO2 substrate [24]. Much higher current is observed for
nanostructured material.
WefailedtoobtainSTMimagesofPt-Ruaftermatrixdissolu-
tion. In order to obtain some information about the morphology
of the electrodeposited surface and the size of the particles the
STM investigation of deposits on Au foil was performed. In
the case of the surface electrodeposited at 0.05 V the distribu-
tion of the particles is homogenous and its average size is ca.
10 nm (Fig. 11a and b). Sample electrodeposited at 0.2 V con-
sists of slightly larger particles forming agglomerates (Fig. 11c
and d). From the STM studies we can see that electrodeposition
at more positive potential results in more ordered packing of
platinum-ruthenium particles, and probably this feature causes
higher catalytic activity.
of material obtained at E = 0.05 V.
d
For further comparative studies we used the samples
electrodeposited at 0.05 V exclusively. In contrast to recent
investigations [12] we limited our studies by the relatively
“fresh” deposits, however we observed slight voltammetric dif-
ference for the samples aged during few days in air and fresh
deposits. For aged samples, in the course of the first cycle the
reduction of Ru oxo-species takes place (see inset in Fig. 8, solid
line), but finally for the next cycles the total charge for both sam-
ples appeared to be approximately the same (dashed line). Hence
short-term ageing of catalyst should not have any pronounced
influence on its electrochemical behavior. These differences are
more distinct if to compare the samples aged for longer period,
especially for Pt-Ru electrodeposited at 0.2 V [12]. Strongly
aged samples are distinguished by lower reversibility in the
hydrogen adsorption region.
The chronoamperometric behavior observed in the course of
electrodeposition of Pt-Ru inside the alumina template is dif-
ferent as compared to the electrodeposition on gold foil. After
initial step we observe constant increase of the current whereas
in the case of gold foil current remains constant, with rather
small tendency to the decrease. The value of the current den-
sity (normalized per geometric area) during electrodeposition
inside anodic alumina is above 20 times smaller than on plain
gold. The cyclic voltammetry examination of two AAO Pt-Ru
samples (Fig. 9) demonstrates that the response is under the first
approximation proportional to deposition charge. This fact con-
firms indirectly that almost all pores are involved in deposition
process, and the increase in true surface area of Pt-Ru is con-
trolled by the amount of deposited metal. For one and the same
deposition charge, AAO Pt-Ru demonstrates lower currents on
the cyclic voltammogram than test sample electrodeposited on
the gold foil (Fig. 9). Like in the case of Pt, it may be caused
by hindrance in the transport of the electrolyte into the porous
matrix.
Note that templating of Pt-Ru into inert inorganic matrix by
electrodeposition leads to higher true surface area as compared
with previous study of assembling Pt-Ru nanoparticles colloid
into TiO2 matrix [24]. It is possible that during electrodeposition
more homogenous distribution of metal particles takes place,
and thus the larger amount of dispersed metal is involved in
electrochemical processes.
In addition AAO Pt-Ru shows even higher catalytic activity
towards electrooxidation of methanol than catalyst electrode-
posited on gold at the same potential (Fig. 10). This fact may
be explained by specific nanostructural features of Pt-Ru in
AAO Pt-Ru. This higher activity manifests itself in the whole
range of the applied potential, and it is even more distinct at
less positive potentials under real DMFC conditions. Another
aspect is the lower degree of poisoning, which is indicated by
less pronounced current hysteresis. When comparing the activ-
ities we normalized all currents per real surface area estimated
from the charge in the hydrogen adsorption region. According
to Ref. [12], for the samples of constant elemental composition
this estimate agrees with the surface area determined by copper
We also observed some changes in morphology resulting
from aging of Pt-Ru film (Fig. 11e and f). As ageing strongly
affects Pt-Ru electrochemistry [12], it requires special future
studies for AAO Pt-Ru.
4. Conclusions
Electrodeposition of Pt and Pt-Ru in the nanoporous alu-
mina template provides the unique possibility to obtain thick
but still dispersed metals, and this method may be used in future
to produce the electrodes of extremely high roughness. Usual
electrodeposits on plain supports remain mechanically stable
only when being of no more than several micrometer thick.
For alumina-templated deposits, no thickness limitations are
expected. Asthetubesappeartobehighlypermeableforsolution
even when already filled with dispersed metal, almost all sur-
face can be charged–discharged even in the presence of matrix.
Actually this fact is not more surprising than the permeability of
usual dispersed electrodeposited platinum, in which the pores
on nm-scale diameter always work, at least if the adsorbates are
small enough [32]. It is hardly possible to use deposit-matrix
nanocomposite as the electrode materials for acidic media, as
the matrix tends to dissolve slowly. However the presence of
alumina is not damaging for electrocatalysis, as confirmed by
rather high specific activities reported in this paper. The pos-
sibility to improve the stability of alumina membrane using
temperature pre-treatment or chemical modification of matrix
should be clarified in future experiments.
Besides apparent effect of a porous matrix on the geometry of
electrodeposited nanostructures (by means of restricted growth
in some directions) it is found that alumina matrix affects the
morphology of nanoparticle arrays. In particular the ordered lay-
ers of 5–12 nm particles are found, which were never observed
in usual electrodeposits with similar size distribution of crystals.
The true nature of the observed effect demands future investiga-
tion.

K.S. Napolskii et al. / Electrochimica Acta 52 (2007) 7910–7919
7919
Possibility to keep the deposits stable after deleting the matrix
References
requires further study. At the current stage we can conclude
that the cohesion of nm-size crystals inside separate nanowires
is rather good, and it allows to keep the wires as a whole
in the absence of alumina. The procedure of matrix dissolu-
tion should be improved to avoid the occasional break of some
wires.
[
[
1] O.A. Petrii, Doklady AN SSSR 160 (1965) 871.
2] N.M. Markovic, H.A. Gasteiger, P.N. Ross, X. Jiang, I. Villegas, M.J.
Weaver, Electrochim. Acta 40 (1995) 91.
[3] H.A. Gasteiger, N. Markovic, P.N. Ross, E.J. Cairns, J. Phys. Chem. 97
(1994) 12020.
[
[
4] M. Krause, W. Vielstich, J. Electroanal. Chem. 379 (1994) 307.
5] M.P. Hogarth, J. Munk, A.K. Shukla, A.J. Hamnett, Appl. Electrochem. 24
Thus, at present we showed that presence of porous matrix
can affect both microstructure and catalytic activity of elec-
trodeposited catalysts. In relation to important applied problems,
like DMFC improvement, new model material is of interest
in the context of so-called blocked catalysts containing inter-
grain disordered regions. Our templated Pt combines high
surface area and the increased concentration of grain bound-
aries formed due to coalescence of crystals. The higher degree
of coalescence follows from the specific surface area (it is
lower that for non-templated deposits prepared under sim-
ilar conditions). For usual deposits, a natural limitation to
combine high surface area and high concentration of grain
boundaries is a limited deposit thickness providing satisfactory
adhesion.
We hope that this study is also of general interest for
actively developing area, templated electrodeposition. The anal-
ysis of mean nanowires length dependence on deposition charge
demonstrate that the pores are completely filled. Density of dis-
persed metal estimated from these data appears to be about 5%
lower as compared to bulk metal density. This estimate agrees
satisfactorily with nanostructured metal porosity and demon-
strates the absence of prolonged empty regions. So one can
hardly assume the initial growth of nanorings along the pore
walls. This type of growth reported in Ref. [33] most proba-
bly requires some artificial nucleation centers near the walls
(
1994) 85.
[
6] E.R. Fachini, R. Diaz-Ayala, E. Casado-Rivera, S. File, C.R. Cabrera,
Langmuir 19 (2003) 8986.
[7] P.J. Kulesza, M. Matczak, A. Wolkiewicz, B. Grzybowska, M. Galkowski,
M.A. Malik, A. Wieckowski, Electrochim. Acta 44 (1999) 2131.
8] W. Chrzanowski, H. Kim, A. Wieckowski, Catal. Lett. 50 (1998) 69.
9] A. Crown, C. Johnston, A. Wieckowski, Surf. Sci. 506 (2002) L268.
10] M. Chojak, M. Mascetti, R. Wlodarczyk, R. Marassi, K. Karnicka, K.
Miecznikowski, P.J. Kulesza, J. Solid State Electrochem. 8 (2004) 854.
[
[
[
[11] L.M. Plyasova, I.Yu. Molina, A.N. Gavrilov, S.V. Cherepanova, O.V. Cher-
stiouk, N.A. Rudina, E.R. Savinova, G.A. Tsirlina, Electrochim. Acta 51
(
2006) 4477.
[
12] A.N. Gavrilov, O.A. Petrii, A.A. Mukovnin, N.V. Smirnova, T.V.
Levchenko, G.A. Tsirlina, Electrochim. Acta 52 (2007) 2775.
[13] H. Masuda, K. Fukuda, Science 268 (1995) 1466.
[14] J. Choi, R.B. Wehrspohn, U. G o¨ sele, Adv. Mater. 15 (2003) 1531.
15] S. Shingubara, J. Nanoparticle Res. 5 (2003) 17.
[
[
[
16] O. Jessensky, F. M u¨ ller, U. G o¨ sele, Appl. Phys. Lett. 72 (1998) 1173.
17] K. Nielsch, J. Choi, K. Schwirn, R.B. Wehrspohn, U. G o¨ sele, Nano Lett. 2
(
2002) 677.
[18] G. Sauer, G. Brehm, S. Schneider, H. Graener, G. Seifert, K. Nielsch, J.
Choi, P. G o¨ ring, U. G o¨ sele, P. Miclea, R.B. Wehrspohn, Appl. Phys. Lett
8
8 (2006) 023106.
[
[
[
19] R.E. Benfield, et al., J. Alloy Compd. 362 (2004) 18.
20] A. Jagminas, et al., Appl. Surf. Sci. 211 (2003) 194.
21] K.S. Napolskii, A.A. Eliseev, N.V. Yesin, A.V. Lukashin, Yu.D. Tretyakov,
N.A. Grigorieva, S.V. Grigoriev, H. Eckerlebe, Physica E 37 (2007) 178.
[22] H. Wang, Y. Cheng Wu, L. Zhang, X. Hu, Appl. Phys. Lett. 89 (2006)
32508.
23] Y. Piao, H. Lim, J. Young Chang, W.-Y. Lee, H. Kim, Electrochim. Acta
0 (2005) 2997.
2
[
[
(
like silver nanorings in Ref. [32]) and should be supported by
5
the presence of gas bubbles. However just the bottom region
of nanowires requires special attention, as it remains the most
mechanically weak fragment.
The problem of enhanced growth of some nanowires also
requiresattention, astheserarefragmentscaninitiatefastdeposit
growth on the top of alumina matrix when the majority of
nanowires still continue to grow slowly inside the pores. The
reason of sporadic fast growth of very few nanowires should be
clarified in future experiments.
24] J.M. Macak, P.J. Barczuk, H. Tsuchiya, M.Z. Nowakowska, A. Ghicov,
M. Chojak, S. Bauer, S. Virtanen, P.J. Kulesza, P. Schmuki, Electrochem.
Comm. 7 (2005) 1417.
25] M. Hepel, I. Kumarihamy, C.J. Zhong, Electrochem. Comm. 8 (2006) 1439.
26] S. Trasatti, O.A. Petrii, Pure Appl. Chem. 5 (1991) 771.
27] O.A. Petrii, G.A. Tsirlina, S.N. Pron’kin, F.M. Spiridonov, M.L. Khr-
uscheva, Russ. J. Electrochem. 35 (1999) 9.
[28] T. Inada, N. Uno, T. Kato, Y. Iwamoto, J. Mater. Res. 20 (2005) 114.
[29] O.V. Cherstiouk, A.N. Gavrlilov, L.M. Plyasova, I.Yu. Molina, G.A. Tsir-
lina, E.R. Savinova, J. Solid State Electrochem., submitted for publication.
[
[
[
[
30] O.A. Petry, B.I. Podlovchenko, A.N. Frumkin, H. Lal, J. Electroanal. Chem.
0 (1965) 253.
[31] L. Yang, J. Chen, X. Zhong, K. Cui, Y. Xu, Y. Kuang, Colloids Surf. A 295
2007) 21.
1
Acknowledgments
(
[
32] B.I. Podlovchenko, T.D. Gladysheva, O.V. Vyaznikovtseva, Yu.M.
Vol’fkovich, Elektrokhimiya 19 (1983) 424.
This work is supported by RFBR (95-03-32592a, 06-08-
0
1443). P.J. Kulesza and P.J. Barczuk were supported in part
[33] Y. Zhao, Y.-G. Guo, Y.-L. Zhang, K. Jiao, Phys. Chem. Chem. Phys. 6
(2004) 1766.
by Ministry of Science and Education, Poland.