Platinum electrodeposition in porous silicon: The influence of surface solvation effects on a chemical reaction in a nanospace

Article abstract of DOI:10.1016/j.cplett.2012.05.078

A chemical reaction in a nanospace is decelerated once a diffusion-limited condition is reached due to the difficulty in supply of reactants from the bulk. We illustrate how to overcome this problem for platinum electrodeposition within nanoporous silicon electrodes. The surface-induced hydration structure of reactants is essential. We make nanopore surfaces hydrophobic by covering them with organic molecules and adopt platinum complex ions with sufficiently large sizes as reactants. Due to the resulting fact that the ions are strongly excluded from the bulk solution to the surface, the ion concentration is greatly enriched within nanopores. This enrichment accelerates the reaction.

Full text of DOI:10.1016/j.cplett.2012.05.078

Chemical Physics Letters 542 (2012) 99–105
Contents lists available at SciVerse ScienceDirect
Chemical Physics Letters
journal homepage: www.elsevier.com/locate/cplett
Platinum electrodeposition in porous silicon: The influence of surface solvation
effects on a chemical reaction in a nanospace
a
a
a
a
b
Kazuhiro Fukami a, , Ryo Koda , Tetsuo Sakka , Tomoko Urata , Ken-ichi Amano , Hikaru Takaya ,
⇑
Masaharu Nakamura b, Yukio Ogata a, Masahiro Kinoshita a
a Institute of Advanced Energy, Kyoto University, Uji, Kyoto 611-0011, Japan
b Institute for Chemical Research, Kyoto University, Uji, Kyoto 611-0011, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
A chemical reaction in a nanospace is decelerated once a diffusion-limited condition is reached due to the
difficulty in supply of reactants from the bulk. We illustrate how to overcome this problem for platinum
electrodeposition within nanoporous silicon electrodes. The surface-induced hydration structure of reac-
tants is essential. We make nanopore surfaces hydrophobic by covering them with organic molecules and
adopt platinum complex ions with sufficiently large sizes as reactants. Due to the resulting fact that the
ions are strongly excluded from the bulk solution to the surface, the ion concentration is greatly enriched
within nanopores. This enrichment accelerates the reaction.
Received 7 May 2012
In final form 29 May 2012
Available online 7 June 2012
Ó 2012 Elsevier B.V. All rights reserved.
1. Introduction
visualize the past record of reaction process by investigating the
density and distribution of platinum deposit within the porous
A chemical reaction occurring in a nanospace is an important
process enabling development of novel science and technology or
highly functional materials and devices. However, the control of
the reaction is difficult to achieve in a conventional way using
the temperature and concentration of reactants in bulk solution
[1–7]. A typical problem is that the reaction in a nanospace can
remarkably be decelerated due to a diffusion-limited condition
reached (i.e., the difficulty in supply of reactants from the bulk
through mass transfer). In modeling and understanding the reac-
tion mechanism in a nanospace, the solvent confined by a surface
or surfaces should play crucially important roles. It has been
shown that the structure of ionic liquid or organic solvent near a
nanosurface is substantially different from that in bulk [8–10].
Here we show that the surface-induced solvation structure of reac-
tants has surprisingly large effects on the reaction process. By
adjusting the affinity of surfaces for the solvent and the solvation
properties of reactants, the reaction process can be controlled in
a variety of manners, for instance, to overcome the deceleration
problem mentioned above. A statistical-mechanical theory for con-
fined molecular liquids provides a useful guide to the control that
can be achieved by manipulating the affinity of electrode surfaces
for water and varying the ion size.
structure (deposited platinum is stable against the spontaneous
dissolution). In this case, the solvent is water and the reactants
are platinum complex ions. The important point is that the hydra-
tion properties of reactants in the aqueous solution confined on the
scale of a nanometer are substantially different from those in the
bulk aqueous solution. More specifically, the surface-induced
hydration structure of reactants, which has never been considered
explicitly in electrochemical research, is experimentally explored
by covering the nanopores with organic molecules to make their
surface hydrophilic or hydrophobic and by testing different ion
sizes of reactants. Theoretical analyses based on a molecular model
(not a dielectric continuum model) for water are also employed. As
a striking result, the physical factor that the ions are excluded from
the bulk water to the surface comes into play when the surface is
hydrophobic, leading to enrichment of the ion concentration with-
in nanopores. Moreover, the degree of this enrichment is remark-
ably influenced by the ion size.
2. Experimental method
Nanoporous silicon electrodes were prepared by anodization of
a p-Si (100) wafer (Shin-Etsu Astech Co. Ltd.). The resistivity of the
In the present Letter, porous silicon and platinum electrodepos-
ition are used as platforms of a nanospace and a chemical reaction,
respectively. The pore diameter is ꢀ5 nm at its maximum and
ꢀ3 nm on the average. With platinum electrodeposition, we can
silicon was 10–20
X cm. All chemicals purchased from Nacalai Tes-
que, Inc. with analytical grade were used without further purifica-
tion. The anodization was performed in 22 wt.% HF ethanolic
solution at current density of 2.0 mAcmꢁ2 for 20 min with a con-
ventional potentio/galvanostat (Hokuto Denko, HZ-3000). The
anodized area was 0.78 cm2 (1 cm in diameter).
⇑
Corresponding author.
E-mail address: fukami.kazuhiro.2u@kyoto-u.ac.jp (K. Fukami).
0009-2614/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cplett.2012.05.078

100
K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
Figure 1. Schematic illustration explaining how the TEM sample was prepared and which part of the sample was considered for the observation by TEM.
After washing the as-anodized nanoporous silicon with water
‘macroparticle’. The Ornstein–Zernike (OZ) equation for the mix-
ture comprising water molecules, cations, and anions can be writ-
ten as [12–17]
and ethanol, the substrate was immersed for 15 h in n-hexane con-
taining 0.2 M propiolic acid or 0.2 M methyl propiolate to cover the
pore-wall surfaces with the organic molecules [11]. We note that it
is better to graft almost similar molecules which show the opposite
hydration properties, i.e., hydrophobic and hydrophilic, to under-
stand the effect of surface-induced hydration structures. This is be-
cause the molecules on the pore wall strongly affect the
electrochemical behavior, meaning that analyses of the electro-
chemical property become quite complex if one uses molecules
with substantially different structures. Propiolic acid and methyl
propiolate form one of the best combinations for the present Letter
from this point of view. The chemically-modified nanoporous sili-
con substrate was analyzed using a Fourier-transform infrared
spectrometer (JASCO, FT/IR-460 Plus; FT-IR) with a diffuse-reflec-
tion mode. The contact angles of the porous layers were evaluated
using a contact-angle measure (KSV Instruments, CAM 200).
The deposition bath was aqueous solution containing 0.1 M
K2PtCl4 and 0.5 M NaCl. Aqueous solution containing 0.1 M K2PtBr4
and 0.5 M KBr was also used to examine the ion-size effect. The
platinum electrodeposition was carried out cathodically at current
Z
X
gabð12Þ ¼ f1=ð8p2Þg qc cacð13Þfgcbð32Þ þ ccbð32Þgdð3Þ ð1aÞ
c
ga ð12Þ ¼ habð12Þ ꢁ cabð12Þ;
a; b ¼ S;þ; ꢁ; M
ð1bÞ
b
where h and c are the total and direct correlation functions, respec-
tively, (ij) represents (rij, Xi Xj), rij is the vector connecting the cen-
,
ters of particles i and j, Xi denotes the three Euler angles describing
R
the orientation of particle i, d(3) represents integration over all po-
sition and angular coordinates of particle 3, and
q is the number
density. The closure equation is expressed by [12–17]
Z
1
cabð12Þ ¼
½habð12Þ@fwabð12Þ ꢁ babð12Þg=@r012ꢂdr10 2
r12
ꢁ uabð12Þ=ðkBTÞ þ babð12Þ
ð2aÞ
ð2bÞ
wabð12Þ ¼ ꢁga ð12Þ þ uabð12Þ=ðkBTÞ
b
density of ꢁ6.4
l
Acmꢁ2 (the minus sign means ‘cathodic’ current
where u is the pair potential, b is the bridge function, and r12 = |r12|.
In the present analysis, the hypernetted-chain (HNC) approxima-
tion is employed (b = 0). We assume that the macroparticle is pres-
density). The cross-sectional views of the platinum deposited
nanoporous silicon were observed with a field-emission scanning
electron microscope (FE-SEM) (JEOL, JSM-6500F; SEM) and ana-
lyzed by an energy-dispersive X-ray spectrometer (EDS) equipped
in the SEM. The nanoscopic images of the cross-section were ob-
served with a scanning transmission microscope (JEOL, JEM-
2200FS; STEM). To prepare the samples for the STEM observation,
the porous layers were sliced into thin films using a focused ion
beam (JEOL, JIB-4500; FIB). The schematic illustration shown in
Figure 1 explains how the STEM samples were prepared and which
part of the sample was considered for the observation by STEM.
ent at infinite dilution (qM ? 0). The calculation process can then be
split into two steps:
Step (i). Solve Eqs. (1 and 2) for bulk aqueous electrolyte solu-
tion. Calculate the correlation functions XSS, XS+, XSꢁ, X++, X+ꢁ, and
X
ꢁꢁ
(X = h, c).
Step (ii). Solve Eqs. (1 and 2) for the macroparticle-aqueous
electrolyte solution system using the correlation functions ob-
tained in step (i) as input data. Calculate the correlation functions
XMS, XM+, and XMꢁ (X = h, c).
For the numerical solution of Eqs. (1 and 2), the pair potentials
and correlation functions are expanded in a basis set of rotational
invariants (i.e., Wigner’s generalized spherical harmonics), and the
basic equations are reformulated in terms of the projections XmnlðrÞ
3. Theoretical method
It is required that a molecular model be employed for water to
investigate the effects of the ion size. A water molecule is modeled
as a hard sphere with diameter dS = 0.28 nm in which a point di-
pole and a point quadrupole of tetrahedral symmetry are embed-
ded [12]. The influence of molecular polarizability of water is
included by employing the self-consistent mean field (SCMF) the-
ory [12]. Hard spherical cations and anions with diameters d+
lm
occurring in the rotational-invariant expansion of X [12–17]. The
expansion considered for m, n 6 nmax = 4 gives sufficiently accurate
results. The basic equations are then numerically solved using the
robust, highly efficient algorithm developed by Kinoshita and
coworkers [14,18]. In the numerical treatment, a sufficiently long
range rL is divided into N grid points (ri = idr, i = 0, 1,. . ., Nꢁ1;
dr = rL/N) and all of the projections are represented by their values
on these points. The grid width and the number of grid points are
set at dr = 0.01dS and N = 4096, respectively.
and d , respectively, are immersed in water. The water-water
ꢁ
and water-ion correlations are then dependent not only on the dis-
tance between centers of the particles but also on the particle ori-
entations. We analyze the structure of electrolyte solution at an
extended, uncharged (hydrophobic) or charged (hydrophilic) sur-
face. The water-surface correlations are also dependent on the ori-
entations of water molecules.
In the real porous structure, the solution is confined within
pores having various sizes whose surfaces are concave. However,
the analyses on the solution confined by an extended surface pro-
vide fundamental information which can readily be applied to the
solution confined between two extended surfaces, between two
concave surfaces, or within a nanopore in a qualitative sense. The
microstructure (heterogeneity) of the surface is not taken into ac-
count in the theoretical calculation. However, it has been shown
that it has no essential effects on the conclusion as long as the aver-
aged properties of the surface-induced structure are discussed [19].
We employ the angle-dependent integral equation theory [12–
15],
a statistical-mechanical theory for molecular liquids. A
macroparticle with diameter dM = 30dS (dS is the diameter of water
molecules, 0.28 nm) mimicking an extended surface is immersed
in aqueous electrolyte solution. The subscripts, ‘S’, ‘+’, ‘ꢁ’, and ‘M’
respectively, represent ‘solvent (water)’, ‘cations’, ‘anions’, and

K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
101
terized by FT-IR measurements (see Supplementary content). As
shown in Figure 2c, e, and g, platinum deposits are mainly ob-
served on the top of the porous silicon when the hydrophilic por-
ous electrode is employed. On the contrary, in Figure 2d, f, and h,
platinum is uniformly deposited within the porous layer by the
use of the hydrophobic porous electrode.
Current density-potential (i-E) curve measurements were per-
formed to understand the electrodeposition behavior on various
surfaces (Figure 3). A platinum plate, a bare p-Si(100), and chem-
ically-modified p-type silicon (100) substrate were used as the
working electrodes. According to the i-E curves, the chemically-
modified electrodes require much higher overpotential than a
4. Results and discussion
4.1. Electrodeposition in aqueous solution of [PtCl4]2ꢁ
First, platinum electrodeposition was carried out in aqueous
solution of [PtCl4]2ꢁ using the hydrophobic and hydrophilic nano-
porous silicon electrodes. Figure 2 shows images of contact angle
measurements, cross-sectional SEM images, and mappings of Si
and Pt by EDS using the hydrophilic and hydrophobic porous sili-
con electrodes. The contact angles of the chemically-modified por-
ous silicon electrodes in Figure 2a and b are 78° and 122°,
respectively. The modifications of porous silicon were also charac-
Figure 2. Strong effect of surface properties of porous silicon electrodes on platinum electrodeposition in which [PtCl4]2ꢁ is used as platinum source. The electrode surfaces
are hydrophilic in (a), (c), (e), and (g) whereas they are hydrophobic in (b), (d), (f), and (h). Images of contact angle measurements are shown in (a) and (b), cross-sectional
SEM images in (c) and (d), EDS mappings of Si in (e) and (f), and EDS mappings of Pt in (g) and (h) after the platinum electrodeposition. The gradation bar in the bottom of the
Figure indicates the relative intensity of each element detected in the mapping. A marked difference in the electrodeposition behavior is observed between (g) and (h).

102
K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
Figure 4. Cross-sectional SEM image of as-anodized nanoporous silicon after
platinum electrodeposition. The condition for electrodeposition was the same as
that in Figure 2, but the pore surface was not modified with organic molecules.
Platinum is densely deposited within the porous layer. The pore surface is
hydrogen-terminated, meaning that it is more hydrophobic than methyl propiolate.
Figure 3. Current density-potential curves measured on various electrodes. The
black, black-dotted, red, and blue curves were measured using a platinum plate,
bare p-Si (1 0 0), propiolic acid-modified p-Si (100), and methyl propiolate-
modified p-Si (100), respectively. The measurements were carried out in aqueous
solution containing 0.1 M K2PtCl4 and 0.5 M NaCl. The green curve was also
measured with a methyl propiolate-modified p-Si (100), but in an aqueous solution
containing 0.1 M K2PtBr4 and 0.5 M KBr. The scan rate of potential was 10 mVsꢁ1
.
anions and g(h) ? 1 as h ? 1. Our experience in theoretical anal-
yses for alkali-halide ions has shown that the profile is rather
insensitive to the ion concentration in the bulk over its wide range
(<3 M) [15]. It is far more sensitive to the ion size. It is not signif-
icantly influenced by the species of the coexisting cations. On the
basis of this information, the aqueous solution is treated as water
containing K2PtCl4 at 1 mM. The anion, [PtCl4]2ꢁ, is modeled as a
spherical ion within which the point charge of ꢁ2|e| (e is the elec-
tronic charge) is placed at the center. The point charge of K+ placed
at its center is |e|. The cation size d+ is set at 0.3 nm. As the anion
platinum plate electrode. However, the difference in the overpo-
tential between the propiolic acid-modified p-Si and the methyl
propiolate-modified one is not large, meaning that the electro-
chemical properties of these electrodes are similar. The electrode-
position was carried out under a constant-current condition.
Therefore, the hydrophilic and hydrophobic electrodes share the
same total electric charge and duration of deposition. In addition,
the current efficiency for the platinum deposition is expected to
be nearly 100% because the current density is extremely small. It
follows that the amounts of deposited platinum are the same for
the hydrophilic and hydrophobic porous silicon electrodes but
the preferential deposition sites are completely different. This, in
turn, means that the difference between them in deposition behav-
ior (Figure 2) originates solely from the surface-induced hydration
structure of [PtCl4]2ꢁ which is dependent on the surface properties.
We have performed an additional experiment. The amount of plat-
inum deposited on as-anodized nanoporous silicon, whose surface
(it is not modified with the organic molecules) shows the highest
hydrophobicity, is drastically larger as depicted in Figure 4. Be-
cause the electrochemical propeties are different from those of
the modified electrodes (Figure 3), one cannot compare the results
in a simple way. However, it has been verified that the hydropho-
bic pore wall enhances the deposition reaction within the nanop-
ores. Further, as discussed in the latter part of this Letter (see
Figure 6), the electrodeposition with a different platinum complex
ion exhibits the same behavior as that with [PtCl4]2ꢁ, i.e., platinum
deposition is substantially enhanced within the hydrophobic por-
ous layer. These results indicate that the behavior is quite general
and the specific chemical properties of functional groups on the
modified surface and metal ions are not relevant.
size d , 0.60 nm is adopted for [PtCl4]2ꢁ. The size of [PtCl4]2ꢁ was
ꢁ
approximately calculated from the molecular weight and density
of K2PtCl4 in solid state with the tetragonal crystal structure.
First, we show g(h) near a hydrophobic surface in Figure 5. Two
more values, 0.65 nm and 0.70 nm, are tested for d to look at the
ꢁ
size effect. The anion concentration is enriched in the immediate
vicinity of the surface. The cations are highly hydrophilic due to
their small size and they are preferentially hydrated in the bulk,
with the result that they are depleted near the surface. Further,
the concentration of K2PtCl4 in the bulk is very low: 1 mM. The cat-
ions have essentially no effects on the anion behavior. Of course,
the local charge neutrality does not hold near the surface. Strong
enrichment of g(h) occurs only within the first layer of the particles
((h–d /2)/dS ꢀ 1). Even in Figure 5c, g(h) reduces quite rapidly as h
ꢁ
increases, and the values of g(h) are only 2.7 and 1.1 within the sec-
ond ((h–d /2)/dS ꢀ 2) and third ((h–d /2)/dS ꢀ 3) layers, respec-
ꢁ
ꢁ
tively. The physical interpretations of the enrichment and the
great dependence on the anion size are given as follows.
(1) d = 0.60 nm
ꢁ
We consider system 1 where the charges of ions and the point
dipole and quadrupole of water molecules are all switched to zero:
A mixture of hard spheres with three different diameters (dS, d+,
and d ) forms the solution. The surface is made by a hard wall.
ꢁ
In this system, all of the allowed system configurations share the
same energy and the system behavior is purely entropic in origin.
4.2. Theoretical analyses
The measure of the enrichment, g(d /2), is ꢀ30. This entropic
The above result was elucidated by our statistical-mechanical
theory for confined molecular liquids. Our principal concern is
ꢁ
enrichment arises from ‘the packing force’ (see Supplementary
content) [20]. We then consider system 2 where only the charges
of ions are set at zero. A binary mixture of hard spheres with differ-
the normalized number-density profile of anions g(h) =
q(h)/qꢁ
where h is the distance from the surface, (h) is the number-den-
q
ent diameters (d+ and d ), which form hydrophobic solutes, is
sity profile of anions, and qꢁ is the number density of anions in the
bulk: g(h) represents the surface-induced hydration structure of
ꢁ
immersed in water. g(d /2) reaches ꢀ67000. Energetically, water
ꢁ

K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
103
Figure 5. Remarkable effect of anion size on surface-induced hydration structure of anions: Normalized number-density profile of anions g(h) near an extended hydrophobic
surface for the three different values of d , 0.60 nm (a), 0.65 nm (b), and 0.70 nm (c). The anions in (a) and (c) correspond to [PtCl4]2ꢁ and [PtBr4]2ꢁ, respectively. The profile
ꢁ
exceeding 1 near the surface indicates the enrichment of anion concentration. Our experience in analyses on confined electrolyte solutions has shown the following [15]. Let
us consider the solution confined between two extended surfaces. When the surface separation H is large, the solution around the center is very much like that in the bulk and
g(h) near each of the surfaces is close to g(h) near a single surface. As H becomes smaller, the solution for all h is more influenced not only by the nearest surface but also by
the other surface. When H becomes smaller than a few times of the diameter of a water molecule, g(h) for all h exhibits an upward shift and the anion-size effect becomes
larger. The upward shift and the enhancement of anion-size effect are more appreciable when H is smaller or the surface is concave and its curvature is larger (see ‘Entropic
enrichment originating from packing force’ in Supplementary content). By the interplay of these physical factors, the enrichment of anion concentration and the anion-size
effect become remarkable when the solution is confined within a pore having the size of a nanometer.
Figure 6. Substantially large increase in the density of platinum nanoparticles in the case where [PtBr4]2ꢁ is used as platinum source: Cross-sectional STEM images of the
hydrophobic porous silicon after the electrodeposition of platinum. The samples of (a, b) and (c, d) are deposited in deposition baths containing [PtCl4]2ꢁ and [PtBr4]2ꢁ
respectively. The images (b, d) are magnified images of (a, c). The scale bars in (a, c) and (b, d) indicate 30 and 10 nm, respectively.
,
molecules do not want to accommodate a solute which cannot par-
ticipate in the hydrogen bonding. The number of water molecules
surrounding such a hydrophobic solute NC is an important quan-
tity. NC for the solute in contact with the surface is half of NC for
the solute in the bulk. Therefore, the solutes are strongly excluded
from water to the surface, which forms those with smaller NC.

104
K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
When the charges are given to the ions (system 3: the model aque-
electrochemical viewpoint. Nevertheless, [PtBr4]2ꢁ produces much
higher density of platinum particles within the hydrophobic por-
ous electrode than [PtCl4]2ꢁ. This result strongly suggests that in
cases of nanopores the surface-induced hydration structure of plat-
inum complex ions plays dominant roles in the reaction
mechanism.
ous solution specified in the last paragraph), g(d /2) ꢀ 5 as ob-
ꢁ
served in Figure 5a. This value is smaller than 30 in system 1,
and the anions with dꢁ = 0.60 nm are rather favored by water: they
are weakly hydrophilic.
(2) dꢁ = 0.70 nm
It has been found that g(d /2) ꢀ44 in system 1 and g(d /2)
ꢁ
ꢁ
ꢀ170 in system 3 as observed in Figure 5c. The latter value is larger
5. Concluding remarks
than the former one, and the anions with d = 0.70 nm are unfavor-
ꢁ
able for water: they are rather hydrophobic despite that the charge
carried by them is as large as ꢁ2|e|. This property is ascribed to the
large anion size.
As explained in the caption for Figure 5, the enrichment be-
comes larger as the pore diameter decreases. For a pore having
the size of a nanometer, the anion concentration within the whole
nanospace is made far higher than that in the bulk.
We have illustrated how to control a chemical reaction occur-
ring in a nanospace for platinum electrodeposition within nano-
porous silicon electrodes as an example. The surface-induced
hydration structure of reactants is shown to be crucially important.
A statistical-mechanical theory for confined molecular liquids pro-
vides a useful guide to the control that can be achieved by manip-
ulating the affinity of electrode surfaces for water and varying the
ion size.
We repeat that g(h) and g(d /2) remain roughly constant
ꢁ
against a change in the anion concentration in the bulk C . How-
ꢁ
The important point is that when d (size of platinum complex
ꢁ
ever, the number-density profile and the number density of anions
near the surface themselves (not normalized) increase in proportion
ions) is large enough, there is a physical factor excluding the anions
from the bulk to the surface. This factor becomes stronger as d in-
ꢁ
with C . Actually, in our experiments, the electrodeposition behav-
ꢁ
creases. Moreover, the strength is extremely sensitive to d . When
ꢁ
ior of platinum within the hydrophobic porous layer displayed a
rather abrupt change at around the [PtCl4]2ꢁ concentration of
ꢀ10 mM. Below this concentration, the electrodeposition occurred
preferentially on the top of the hydrophobic porous layer, which is
indicative that the concentration of [PtCl4]2ꢁ near the surface is not
high enough to deposit platinum despite almost the same degree of
the enrichment. It is interesting that such a threshold value exists
the surface is hydrophobic, this factor dominates with the result
that the anion concentration remains enriched in the immediate
vicinity of the surface. For a pore having the size of a nanometer,
the enrichment is kept remarkable within the whole nanospace.
Thus, by modifying the silicon surface to make it hydrophobic
and using platinum complex ions with considerably large d , we
ꢁ
have succeeded in promoting electrodeposition within pores
which are deep and extremely small in diameter. Without the
physical factor, the reaction within such pores would be deceler-
ated to a remarkable extent after a short time due to a diffusion-
limited condition reached. Even when a pore is considerably large
in diameter and the enrichment occurs only in the immediate
vicinity of the pore surface, the continuous supply of reactants into
the pore could be driven by surface diffusion arising from the con-
centration gradient formed along the surface. Once sufficiently
large nuclei of platinum are formed on the electrode surface, the
electrodeposition proceeds quite easily. The surface-induced struc-
ture could affect the electrostatic-potential profile across the elec-
trode-solution interface during the electrodeposition which is
important in understanding and controlling the electrochemical
reaction. This issue is to be pursued in a future study.
for C .
ꢁ
We then discuss the theoretical result for a hydrophilic surface.
When the surface is negatively charged (i.e., hydrophilic) as in our
experiments described above, within the surface-induced layer the
anions are largely depleted (g(d /2) ꢃ 1). The depletion becomes
ꢁ
more remarkable as the pore diameter decreases. There are two
reasons for this behavior. The first reason is that the anions are re-
pelled from the surface through electrostatic interaction. The sec-
ond one is that the number density of water molecules is higher
within a nanopore and the insertion of rather hydrophobic ions
into the nanopore is less allowed.
4.3. Electrodeposition in aqueous solution of [PtBr4]2ꢁ
Figure 5 indicates that the enrichment of anion concentration is
drastically enhanced by a minor increase in the anion size. To dem-
onstrate this size effect, the deposition experiment was carried out
in the bath containing [PtBr4]2ꢁ. The size difference between
[PtBr4]2ꢁ and [PtCl4]2ꢁ can be set at the double of that between
Brꢁ and Clꢁ with the result that the size of [PtBr4]2ꢁ is ꢀ0.7 nm
which corresponds to Figure 5c [21]. Figure 6 depicts the micro-
scopic structures of the deposited platinum in the [PtCl4]2ꢁ and
[PtBr4]2ꢁ baths. In Figure 6a and b, platinum is deposited as nano-
particles with ꢀ4 nm in diameter, which is almost the same as the
diameter of the original pores of porous silicon. These are images of
the interfaces between bulk silicon and the hydrophobic porous
layer (Figure 2), showing that platinum complex ions were contin-
uously supplied to the bottom and platinum electrodeposition oc-
curred smoothly. As predicted by our theory, the deposition
behavior exhibited a drastic change when platinum deposition
was carried out in the [PtBr4]2ꢁ bath. The density of platinum
nanoparticles is obviously increased when [PtBr4]2ꢁ is used as plat-
inum source (Figure 6c and d). Strikingly, the electrochemical
properties of [PtCl4]2ꢁ bath are different from those of [PtBr4]2ꢁ
bath in the respect that the latter requires much higher overpoten-
tial than the former (see Figure 3). It follows that the deposition of
[PtCl4]2ꢁ would be much easier than that of [PtBr4]2ꢁ from the
To date, a variety of porous structures and solvents have been
utilized in many applications. The present Letter, which has shown
that the solvation properties of reactants in solvent confined on the
scale of a nanometer play crucially important roles, sheds new
light on design and control of highly efficient, chemical systems
in a nanospace combined with proper solvents and reactants.
Acknowledgments
We thank Profs. K. Krischer (TU München) and S. Nakanishi
(Univ. of Tokyo) for useful comments and suggestions. This Letter
was supported by Grant-in-Aid for Young Scientists (B) (No.
23750239) and that for Scientific Research on Innovative Areas
(No. 20118004) from MEXT.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.cplett.2012.05.
078.

K. Fukami et al. / Chemical Physics Letters 542 (2012) 99–105
[11] A. Imanishi, S. Yamane, Y. Nakato, Langmuir 24 (2008) 10755.
105
References
[12] P.G. Kusalik, G.N. Patey, J. Chem. Phys. 88 (1988) 7715.
[13] M. Kinoshita, S. Iba, M. Harada, J. Chem. Phys. 105 (1996) 2487.
[14] M. Kinoshita, D.R. Bérard, J. Comput. Phys. 124 (1996) 230.
[15] M. Kinoshita, Condens. Matter. Phys. 10 (2007) 387.
[16] N.M. Cann, G.N. Patey, J. Chem. Phys. 106 (1997) 8165.
[17] D.R. Bérard, M. Kinoshita, N.M. Cann, G.N. Patey, J. Chem. Phys. 107 (1997)
4719.
[18] M. Kinoshita, M. Harada, Mol. Phys. 81 (1994) 1473.
[19] J.C. Shelley, G.N. Patey, D.R. Bérard, G.M. Torrie, J. Chem. Phys. 107 (1997)
2122.
[1] T.S. Koblenz, J. Wassenaar, J.N. Reek, Chem. Soc. Rev. 37 (2008) 247.
[2] K. Kurotobi, Y. Murata, Science 333 (2011) 613.
[3] Y.-W. Huang, Y.-C. Lai, C.-J. Tsai, Y.-W. Chiang, Proc. Natl. Acad. Sci. USA 108
(2011) 14145.
[4] S. Han, M.Y. Choi, P. Kumar, H.E. Stanley, Nat. Phys. 6 (2010) 685.
[5] Y. Zhang et al., Proc. Natl. Acad. Sci. USA 108 (2011) 12206.
[6] V. Garcia-Morales, K. Krischer, Proc. Natl. Acad. Sci. USA 107 (2010) 4528.
[7] V. Garcia-Morales, K. Krischer, Proc. Natl. Acad. Sci. USA 108 (2011) 19535.
[8] R.M. Lynden-Bell, A.I. Frolov, M.V. Fedorov, Phys. Chem. Chem. Phys. 14 (2012)
2693.
[9] A.I. Frolov, K. Kirchner, T. Kirchner, M.V. Fedorov, FaradayDiscuss. 154(2012) 154.
[10] A.I. Frolov, R.N. Arif, M. Kolar, A.O. Romanova, M.V. Fedorov, A.G. Rozhin, Chem.
Sci. 3 (2012) 541.
[20] C.-Y. Lee, J.A. McCammon, P.J. Rossky, J. Chem. Phys. 80 (1984) 4448.
[21] M. Kinoshita, F. Hirata, J. Chem. Phys. 106 (1997) 5202.