In situ stress measurements during Al UPD onto (111)-textured Au from AlCl3-EMImCl ionic liquid

Article abstract of DOI:10.1149/1.2894202

In situ stress measurements were made during Al underpotential deposition (upd) onto (111)-textured Au from Lewis acidic aluminum chloride, 1-ethyl-3-methylimidazolium chloride (Al Cl3 -EMImCl), using the wafer curvature method. The surface stress response consists of three distinct features. In the potential range of 1.2-0.6 V the surface stress moves in the tensile (positive) direction from a value arbitrarily chosen as zero. This likely involves the desorption of Al Cl4- from the Au surface and is consistent with adsorbate-induced stress models that appear in the literature. At the start of Al upd, the surface stress moves in the compressive direction, in contrast to the tensile stress expected based on the positive lattice misfit. We attribute this compressive stress to the formation of Al-Au bonds which partially satisfy the bonding requirements of the Au surface atoms, thereby reducing the tensile surface stress inherent to the clean Au surface. In the latter stages of Al upd, the surface stress once again moves in the tensile direction, which we attribute to Al-Au alloying. The magnitude of the tensile stress change is close to that estimated from the elastic strain associated with the change in molar volume, using reaction kinetics reported for surface alloy formation in the upd region.

Full text of DOI:10.1149/1.2894202

D408
Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
0
013-4651/2008/155͑5͒/D408/6/$23.00 © The Electrochemical Society
In Situ Stress Measurements During Al UPD onto
111)-Textured Au from AlCl –EMImCl Ionic Liquid
(
3
,
z
G. R. Stafford* and C. R. Beauchamp
Materials Science and Engineering Laboratory, National Institute of Standards and Technology,
Gaithersburg, Maryland 20899, USA
In situ stress measurements were made during Al underpotential deposition ͑upd͒ onto ͑111͒-textured Au from Lewis acidic
aluminum chloride, 1-ethyl-3-methylimidazolium chloride ͑AlCl –EMImCl͒, using the wafer curvature method. The surface stress
3
response consists of three distinct features. In the potential range of 1.2–0.6 V the surface stress moves in the tensile ͑positive͒
−
4
direction from a value arbitrarily chosen as zero. This likely involves the desorption of AlCl from the Au surface and is consistent
with adsorbate-induced stress models that appear in the literature. At the start of Al upd, the surface stress moves in the
compressive direction, in contrast to the tensile stress expected based on the positive lattice misfit. We attribute this compressive
stress to the formation of Al–Au bonds which partially satisfy the bonding requirements of the Au surface atoms, thereby reducing
the tensile surface stress inherent to the clean Au surface. In the latter stages of Al upd, the surface stress once again moves in the
tensile direction, which we attribute to Al–Au alloying. The magnitude of the tensile stress change is close to that estimated from
the elastic strain associated with the change in molar volume, using reaction kinetics reported for surface alloy formation in the
upd region.
©
2008 The Electrochemical Society. ͓DOI: 10.1149/1.2894202͔ All rights reserved.
Manuscript submitted November 29, 2007; revised manuscript received February 8, 2008.
Available electronically March 20, 2008.
2
8
The underpotential deposition ͑upd͒ of metal monolayers onto
W͑110͒. As expected, when the misfit is negative ͑large adsor-
bate͒, elasticity theory predicts compressive stress which is gener-
ally supported by experiment, although there is some discrepancy
foreign metal substrates is important to the electrodeposition com-
munity. In metal deposition involving Stranski–Krastanov nucle-
ation and growth, the upd layer forms prior to the formation of
three-dimensional ͑3-D͒ crystals so that an understanding of depo-
sition processes in the upd region may help us better understand the
growth and subsequent properties of bulk thin films. For this reason,
upd processes have been extensively examined by in situ techniques
2
9,30
regarding the magnitude of the compressive stress.
The various thermodynamic contributions for an epitaxial bilayer
3
1,32
have been described by Cammarata et al.,
where the surface
stress change includes contributions from the substrate–adlayer in-
terface, the adlayer surface, and the lattice misfit between the metal
adlayer and the substrate. This has been used to obtain the critical
thickness associated with the loss of coherency for thin-film
1-6
7-9
such as electrochemical voltammetry/coulometry, spectroscopy,
1
0-15
11,16,17
scanning probe microscopy,
crystal nanogravimetry
X-ray scattering,
and quartz
1
8-20
32
in order to better understand the forma-
epitaxy. A similar treatment has been extended to epitaxial systems
tion and structure of the upd adlayer.
formed electrochemically where an electrocapillary correction is
An additional in situ probe gaining popularity in electrochemis-
try involves the measurement of surface and growth stress. Surface
stress is the reversible work required to elastically deform a surface.
The loss of bonds at a clean metal surface causes a redistribution of
charge density between the remaining surface atoms, thereby in-
creasing their attractive interaction and causing a decrease in their
equilibrium interatomic spacing. Because the surface atoms are held
in place by the bulk lattice, they are stretched from their equilibrium
lattice positions and a surface stress arises. The sign of the surface
stress is positive ͑tensile͒ if the surface would like to contract under
its own stress. The adsorption of species on the surface can be ex-
pected to alter the surface stress, because the local interaction of
each adsorbate alters the bond strength between neighboring atoms
on the surface. The sensitivity of surface stress to both ionic and
fully discharged adsorbates such as metal monolayers makes this
measurement particularly relevant for upd studies, where both pro-
cesses tend to occur simultaneously. For instance, in the case of Pb
upd onto ͑111͒-textured Au in HClO4 supporting electrolyte, the
stress transient clearly shows regions of ClO −4 desorption, Pb–Au
bond formation, stress relaxation due to island coalescence, and
compression of the monolayer at potentials just positive of bulk Pb
added to account for the change in the point of zero charge ͑pzc͒
2
6,33
with the addition of the adlayer.
Wafer-curvature experiments
have been conducted on a variety of upd systems, namely,
2
+
2+
+
Pb /Au͑111͒, Pb /Ag͑111͒, and Ag /Au͑111͒, and the measured
stress changes were used to estimate the interface stress of the
substrate/adlayer interface. A similar treatment has been applied to
the Cu/Au͑111͒ system where the contribution of anion adsorption
2
6,33
was considered as well.
It is not clear whether use of an epitax-
ial bilayer model, where both film and substrate are considered to
have bulklike properties, properly describes upd where the adlayer is
restricted to a monolayer.
Leiva et al. have used the embedded-atom method to calculate
the surface stress due to epitaxial monolayer adsorption for a variety
3
4
of face-centered cubic adsorbate–substrate combinations. When
adsorption energy is considered together with lattice misfit, they find
qualitative agreement with experimental data. The adsorption con-
tribution is particularly apparent for systems with positive misfit
͑small adsorbate͒, such as Cu on Au. The embedded-atom calcula-
tion predicts compressive stress, consistent with experimental
2
3,25-27
data,
whereas considering the +12.8% misfit alone leads to
tensile stress. Leiva et al. have further shown that the adsorption-
energy contribution of the monolayer to the change in surface stress
is negative for all of the adsorbate–substrate combinations
2
1,22
deposition.
It is readily apparent from experimental data reported in the lit-
erature that the surface stress change following growth of a pseudo-
morphic ͑1 ϫ 1͒ metal monolayer cannot be estimated from con-
tinuum elasticity theory using the difference between the bulk lattice
parameters to determine a misfit strain. When the misfit is positive
3
4
examined. The surface stress change is then a balance between an
adsorption contribution which is negative and a misfit term which
can be positive or negative.
Because upd is at least partially driven by the free energy of
dissimilar bonding, one would always expect the adsorption contri-
bution to favor compressive stress. In addition, because upd pro-
cesses often result in a negative shift of the pzc, anion readsorption
onto the newly formed adlayer generally results in an additional
compressive contribution to the surface stress, similar to that ob-
͑
small adsorbate͒, elasticity theory predicts a tensile stress; however,
there are several examples where the opposite is true, namely the
23-27
upd of Cu on Au͑111͒
and the vapor deposition of Fe on
2
3-27
served for Cu/Au͑111͒.
One is then left to question how signifi-
*
Electrochemical Society Active Member.
E-mail: gery.stafford@nist.gov
z
cant the lattice misfit contribution is toward the surface stress
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
change during upd processes. One could argue that in the case of
D409
Cu/Au͑111͒, the lattice misfit is too large to make a meaningful
contribution to the stress. Obtaining surface stress data on other upd
systems that have a positive but smaller lattice mismatch, such as
35
Pd/Au͑111͒ ͑+4.9% mismatch͒ or Al/Au͑111͒ ͑+0.7% mis-
match͒, might facilitate the discussion. In a previous paper, we ex-
amined the growth stress associated with Al deposition on ͑111͒-
textured Au, a process described by Stranski–Krastanov 3-D
3
6
growth. In that paper we briefly described the upd of Al on both
111͒-textured Cu and ͑111͒-textured Au cantilever electrodes and
͑
noted some anomalous behavior in the case of Al–Au. The purpose
of this paper is to examine the surface stress changes associated with
Al upd on ͑111͒-textured Au in more detail, paying particular atten-
tion to differentiating the stress behavior attributed to Al monolayer
formation and Al–Au alloy formation in the upd region.
The electrodeposition of aluminum must be carried out from
nonaqueous or aprotic solvents, because hydrogen is evolved in
aqueous solutions before aluminum can be deposited. The direct
electrodeposition of aluminum from ambient-temperature ionic liq-
uids has received considerable attention. These electrolytes are pre-
Figure 1. Linear sweep voltammetry associated with the upd of Al onto
pared by combining AlCl with certain unsymmetrical quaternary
3
͑111͒-textured Au in 55:45 mole ratio of AlCl -EMImCl. Sweep rate was
3
−
1
ammonium chloride salts such as 1-ethyl-3-methylimidazolium
20 mV s .
37
chloride ͑EMImCl͒. They are liquid at room temperature and dis-
play adjustable Lewis acidity. Melts prepared with a molar excess of
AlCl ͑Ͼ50% mole fraction AlCl ͒ are Lewis acidic due to the pres-
3
3
−
7
optical bench. The electrolyte was a 55:45 mole ratio of
ence of the coordinately unsaturated Al Cl ion, whereas those for-
2
AlCl –EMImCl. It was prepared and purified using the procedures
3
mulated with a molar excess of the quaternary ammonium chloride
outlined in Ref. 37. The counter electrode was an aluminum wire
placed parallel to and in the same solution as the working electrode.
The reference electrode was also an aluminum wire placed in the
same solution as the working electrode and positioned between the
working and counter electrodes. All potentials reported here are with
respect to the Al reference. The electrolyte was added to the cell
while in a dry box containing less than 2 ppm oxygen and was then
sealed. The cell was then removed from the dry box, placed in the
optical mount, and positioned on the optical bench. Potential control
was maintained using an EG&G Princeton Applied Research Corp.
model 273 potentiostat/galvanostat that was controlled by a Dell
Pentium 4 computer and LabView software. A more-detailed de-
scription of the optical bench and stress measurement can be found
in Ref. 25 and 36.
salt ͑Ͻ50% mole fraction AlCl ͒ are termed Lewis basic because
they contain a chloride ion that is not covalently bound to alumi-
3
num. The cathodic limit of the acidic AlCl –EMImCl melts is gov-
3
erned by the electrodeposition of aluminum according to the follow-
ing reaction
−
−
−
4
4
Al Cl + 3e → Al + 7AlCl
͓1͔
2
7
Aluminum cannot be electrodeposited from basic melts due to the
preferential reduction of the 1-ethyl-3-methylimidazolium cation
−
4
over AlCl . Numerous reports describing the electrodeposition of
aluminum and its alloys from the Lewis acidic AlCl –EMImCl ionic
3
38-41
liquids appear in the literature.
Recently it has been shown that
aluminum can be electrodeposited from the new generation of air-
and water-stable ionic liquids following controlled addition of
4
2,43
AlCl3.
Results and Discussion
Figure 1 shows the linear sweep voltammetry for the upd of Al
Experimental
In situ stress measurements were made on an optical bench using
onto ͑111͒-textured Au from a 55:45 mole ratio of AlCl -EMImCl
3
electrolyte. The voltammetry shows two cathodic peaks at 0.5 and
0.1 V as well as two anodic peaks at 0.26 and 0.54 V. This volta-
mmetry is similar to that reported by Lee et al. on polycrystalline Au
21,25
the wafer-curvature method.
The cantilever was a borosilicate
glass strip ͑D 263, Schott͒ measuring 60 ϫ 3 ϫ 0.108 mm. The
4
5
9
−2
glass had a Young’s modulus of 72.9 ϫ 10 N m and a Poisson
ratio of 0.208, as specified by the vendor. Onto one side of the
cantilever a 4 nm thick adhesion layer of titanium and a subsequent
in 60:40 mole ratio of AlCl3-EMImCl. Reports in the literature
45-48
indicate that Al–Au alloys can be formed in the upd region.
The
general consensus is that stripping of the upd layer occurs in the
potential range of 0.4–0.6 V and that the first anodic peak is asso-
ciated with dealloying rather than partial removal of the upd layer.
This conclusion is based on the fact that the height of the first
oxidation wave increases with the time at which the electrode is held
2
50 nm film of gold were vapor-deposited by electron-beam evapo-
ration. The glass–metal interface provided the reflective surface for
the laser beam. Prior to use, the electrodes were cleaned in piranha
solution ͑3:1 volume mixture of concentrated H SO :30% H O ͒.
2
4
2
2
4
5
The Au electrodes had a strong ͑111͒ crystallographic orientation.
The 200 reflection was not apparent in ␪–2␪ X-ray scans, and rock-
ing curves of the 111 reflection generally yielded a full width at
half-maximum on the order of 2°. These films have a fiber texture,
i.e., there is no in-plane orientation. Field-emission scanning elec-
tron microscopy shows the Au grain size to be on the order of
at +0.10 V. Zell et al. have shown by in situ scanning tunneling
microscopy ͑STM͒ that in the potential region of 0.2–0.4 V two-
dimensional ͑2-D͒ aluminum clusters measuring 1–2 nm in diameter
4
8
are deposited on the bare Au surface. When the potential is further
decreased to 0.1 V, the number of 2-D clusters increases and 3-D
growth is observed. When the Al is anodically removed at 1.1 V,
holes are left in the Au surface, indicating that surface alloying and
2
5 nm. The curvature of the substrate was monitored while under
4
8
potential control by reflecting a HeNe laser off of the glass/metal
interface onto a position-sensitive detector. The relationship between
the surface stress and the radius of curvature of the cantilever is
perhaps compound formation occur in the upd region. Similar sur-
face alloying has been reported for evaporated Al on Au͑111͒ at and
4
9
below room temperature.
44
given by Stoney’s equation.
We first examine the upd process by limiting the cathodic poten-
tial to +0.3 V in order to avoid significant alloy formation. Figure 2
shows the voltammetry ͑a͒ and surface stress response ͑b͒ for sweep
rates ranging from 10 to 500 mV s− from a starting potential of
The electrochemical cell was an air-tight single-compartment
Pyrex cell. A glass disk was joined to the back of the cell to allow
the cell to be held and positioned by a standard mirror mount on the
1
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D410
Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
to the onset of Al upd in Fig. 2a, the surface stress moves in the
compressive direction. The stress maximum occurs well into the upd
−
region, suggesting that perhaps two competing processes, i.e., AlCl4
desorption and Al upd, occur simultaneously in the early stages of
adlayer formation. Similar behavior has been observed for Bi and Pb
2
1,22
upd on Au.
The magnitude of the compressive stress at the
+
0.3 V cathodic limit is a function of sweep rate, indicating some
kinetic limitations in the stress response. When the potential scan is
reversed, the stress response for the return sweep follows a path
similar to that of the forward scan, although significant hysteresis is
present in the deposition region, reflecting the quasi-reversibility of
the voltammetry.
As mentioned previously, Al has a positive misfit ͑0.74%͒ with
respect to the Au substrate. As a consequence, one would expect a
tensile stress change for a pseudomorphic ͑1 ϫ 1͒ adlayer consid-
ering misfit alone. This can be quantified by the following
5
6
expression
Y͑111͒
␶
=
␧ d
mf ͑111͒
␪
͓2͔
1
− ␯͑111͒
where Y͑111͒ is Young’s modulus for the ͑111͒ surface of Al, ␯͑111͒ is
the Poisson ratio, ␧mf is the misfit strain, d͑111͒ is the height of the
monolayer, and ␪ is the number of monolayers on the surface. The
5
7
elastic constants, calculated from the elastic compliances for Al,
are Y͑111͒ = 72 GPa and ␯͑111͒ = 0.365. Inserting these values into
Eq. 2 for a monolayer of aluminum results in a misfit stress of
−
1
+
0.24 N m . It is clear from Fig. 2 that the upd of Al results in a
compressive stress rather than the tensile stress predicted from lat-
tice misfit.
We now examine the stress response for more cathodic potentials
where surface alloys are known to form in the upd region. These
transients are shown in Fig. 3 for a variety of cathodic limits extend-
ing down to 0 V. If the cathodic limit does not go negative of
Figure 2. ͑a͒ Linear sweep voltammetry and ͑b͒ surface stress response for
the upd of Al ͑limited to the first voltammetric wave͒ onto ͑111͒-textured Au
+
0.2 V, then the stress response is identical to that shown in Fig. 2b
in 55:45 mole ratio of AlCl -EMImCl. Sweep rate was varied from
3
and the cantilever position returns to its initial zero value at the
completion of the transient. As the potential is reduced to less than
−
1
1
0 to 500 mV s . The inset in ͑a͒ shows the peak current density for both
deposition and stripping as a function of sweep rate.
+
0.15 V, the stress begins to move in the tensile direction. The
stress continues to increase on the return sweep until the first anodic
wave occurs, after which the stress returns to a value somewhat
higher than its initial value. However, the cantilever returns to its
original position if left for several minutes at a potential of 1.2 V.
We attribute this tensile transient and remnant tensile stress at 1.2 V
to Al–Au surface alloying.
In order to examine the kinetics of stress evolution, potential-step
experiments were conducted to complement the potentiodynamic
scans of Fig. 3. Selected transients are shown in Fig. 4a for pulses
initiated at 1.2 V. When the potential is stepped to values positive of
Al upd ͑+0.6 V͒, the stress change in the tensile direction is rapid
and the stress remains constant for the 20 s duration of the pulse.
1
.2 V. The voltammetric waves appear to be quasi-reversible, based
50
on the sweep-rate dependence of the peak potential. The inset
shows that the peak current density, for both cathodic and anodic
processes, varies linearly with sweep rate, indicating a surface-
−
2
limited process. The charge is estimated to be about 1.8 mC cm ,
which is approximately factor of 2.5 higher than the
.67 mC cm expected for a pseudomorphic monolayer of Al on
a
−
2
0
Au͑111͒. A typical roughness factor for the as-evaporated Au canti-
lever is 1.3, based on the charge necessary to completely reduce the
51
Au oxide in acidic aqueous electrolyte. The higher roughness fac-
tor, as determined from the Al upd charge, suggests that cycling the
electrode in the upd region causes additional roughness due to the
formation and removal of the surface alloy.
The surface stress response for Al upd is shown in Fig. 2b. In the
potential range of +1.2 to + 0.55 V the surface stress moves in the
tensile ͑positive͒ direction from a value arbitrarily chosen as zero.
The stress response is likely due to both electrocapillarity as well as
the desorption of AlCl from the gold surface. An ordered adlayer of
AlCl has been observed by STM at potentials positive of Al upd on
both Cu͑111͒ in 55:45 mole ratio AlCl -EMImCl electrolyte and
The change in surface stress in this potential region is due to elec-
−
4
trocapillarity and AlCl desorption; both of these processes are ex-
pected to be rapid, which explains the shape of the transient in Fig.
4a. When the potential is pulsed into the upd region but positive of
alloy formation ͑+0.2 V͒, the stress first moves in the tensile direc-
−
4
tion, reflecting AlCl desorption, and then moves compressively as
the Al monolayer forms on the Au. The compressive stress change
due to Al upd is rather slow, which is consistent with the sweep-rate
dependence of the stress response in Fig. 2b. The total stress change
with respect to the clean Au surface at +0.6 V is approximately
−
4
−
4
5
2
3
−
1
Au͑111͒ in 58:42 mole ratio AlCl -1-butyl-3-methylimidazolium
−0.9 N m . When the potential is stepped more cathodically into
3
5
3
chloride. The tensile stress change in response to the desorption is
consistent with Ibach’s surface-induced charge-redistribution model,
where electron acceptors such as adsorbed anions cause compressive
stress because they reduce the electron density in the surface.
Haiss has observed a linear correlation between surface stress and
surface charge for the adsorption/desorption of a variety of anions
the region of alloy formation, the stress follows the same progres-
−
4
sion, i.e., rapid tensile for AlCl desorption, compressive for Al upd,
followed by a second tensile transient. When the potential is pulsed
to more cathodic potentials, it becomes clear that the time constant
for the second tensile transient is potential dependent. The potential-
step data is summarized in Fig. 4b, where the stress change, re-
corded after 20 s, for both the cathodic and anodic pulses is plotted
5
4
5
5
on Au. In the potential range of +0.55 to + 0.30 V, corresponding
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
D411
Figure 3. ͑a͒ Linear sweep voltammetry and ͑b͒ surface stress response for
the upd of Al onto ͑111͒-textured Au in 55:45 mole ratio of AlCl -EMImCl.
3
The starting potential was +1.2 V while the cathodic limit was varied from
+
0.6 to 0 V. Sweep rate was 50 mV s−1
.
Figure 4. ͑a͒ Surface stress-time transients for the upd of Al onto ͑111͒-
textured Au in 55:45 mole ratio of AlCl -EMImCl in response to the ca-
3
thodic potential steps indicated in the figure. ͑b͒ Stress change as a function
of step potential, recorded after 20 s, for the cathodic step ͑—b—͒ and the
return anodic step to 1.2 V ͑—᭿—͒. Potentiodynamic scan recorded at a
as a function of step potential. For comparison, a potentiodynamic
−
1
scan using a sweep rate of 50 mV s is also plotted. The agreement
−
1
−
4
sweep rate of 50 mV s ͑—͒.
is excellent for both the AlCl desorption and Al monolayer regions,
whereas the stress response for the potentiodynamic scan slightly
lags behind the potential step response in the surface alloy region.
The stress transients shown in Fig. 4b consist of three distinct
1
90.3 GPa. If we arbitrarily allow 10% of the surface Au atoms to
be replaced by Al, then the tensile stress generated in the Au–Al
alloy surface layer is +0.04 N m , only a fraction of that measured
features. The tensile stress observed at positive potentials
−1
−
4
͑
1.2–0.6 V͒ that we ascribe to AlCl desorption as well as the com-
experimentally. This is a clear indication that the decrease in molar
volume of the surface alloy layer must be more substantial than that
described by a simple place exchange of Al and Au atoms on the
surface.
STM studies of surface alloying of evaporated Al on Au͑111͒ at
room temperature have shown lattice contractions in the surface
alloy that are unusually large in light of the similar atomic radii of
Al and Au. It is also well known that large negative deviations
from a rule of mixtures addition of the constituent molar volumes
exist in several binary Al-transition-metal alloys. The high lattice
contraction of the Al–Au surface alloy observed by STM is then
rationalized as the 2-D analogue of the Al atom contraction in the
pressive stress associated with Al monolayer formation are common
features of stress transients reported in the literature for a variety of
upd processes. However, the potential-dependent tensile transient
observed in the latter stages of Al upd ͑0.15–0.0 V͒ is quite unusual
and warrants additional discussion. Stress arises when a film under-
goes any dynamic microstructural evolution process that changes its
density while rigidly attached to the substrate. Tensile stress gener-
ated during film growth is an indication that the density of the film
is increasing, i.e., there is a net decrease in molar volume. The initial
stages of alloy formation in the upd region involve place-exchange
processes between the Al adlayer and surface Au atoms. The stress
response to replacing some of the surface Au atoms with smaller Al
atoms can be estimated using an expression similar to Eq. 2
4
9
4
9
binary bulk alloys. For example, the molar volume contraction,
⌬V/V, associated with the formation of orthorhombic AlAu is
2
−
0.056. The stress change associated with bulk alloy formation can
␶
= EЈ ␧ d
␪e
͓3͔
1
11 mf ͑111͒
then be estimated from the molar volume contraction
where E1Ј11 is the biaxial modulus for the ͑111͒ surface of Au ͓i.e.,
⌬
V
Y
͑111͒/͑1 − ␯͑111͔͒͒, ␧mf is the +0.74% misfit strain between Al and
␶imc = − EiЈmc dimc
͓4͔
3
V
Au, d͑111͒ is the height of the Au surface layer, and ␪ is the fraction
e
of the Au surface that has been exchanged by Al atoms. The biaxial
modulus, calculated from the elastic compliances for Au, is
where EiЈmc is the biaxial modulus of the intermetallic and dimc is the
intermetallic thickness. If one assumes an EЈimc of 200 GPa and in-
5
7
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

D412
Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
cludes the ⌬V/V of −0.056, then the volume contraction due to
and Ole Kongstein. We are particularly grateful to Tetsuya Tsuda
−
1
surface alloying produces a stress change of +3.7 N m for each
nm of intermetallic formed. Lattice contractions of this magnitude
can certainly explain the tensile-stress changes observed experimen-
tally.
and Charles Hussey for providing the AlCl -EMImCl.
3
National Institute of Standards and Technology assisted in meeting the
publication costs of this article.
We now examine the kinetics of surface alloy formation. The
stress transients in Fig. 4a indicate that both the time constant and
the magnitude of the alloy-induced tensile stress are potential-
dependent. Similar potential dependence for alloy formation in the
References
1
. A. Hamelin, J. Electroanal. Chem. Interfacial Electrochem., 165, 167 ͑1984͒.
2. A. Hamelin and J. Lipkowski, J. Electroanal. Chem. Interfacial Electrochem., 171,
317 ͑1984͒.
58
3. E. Kirowa-Eisner, Y. Bonfil, D. Tzur, and E. Gileadi, J. Electroanal. Chem., 552,
upd region has been observed for Cd on Ag. Specifically, the loga-
rithm of the rate constant was found to increase linearly as the
potential was made more negative, approaching the bulk deposition
of Cd. This was attributed to a Nernst-type dependence of the Cd
activity at the surface on electrode potential. Garcia et al. have fur-
1
71 ͑2003͒.
4
. B. K. Niece and A. A. Gewirth, Langmuir, 12, 4909 ͑1996͒.
5. S. Sayed and K. Juttner, Electrochim. Acta, 28, 1635 ͑1983͒.
6. T. Solomun and W. Kautek, Electrochim. Acta, 47, 679 ͑2001͒.
. R. Adzic, V. Jovancicevic, and M. Podlavicky, Electrochim. Acta, 25, 1143 ͑1980͒.
8. R. Adzic, E. Yeager, and B. D. Cahan, J. Electrochem. Soc., 121, 474 ͑1974͒.
. M. Futamata, Chem. Phys. Lett., 333, 337 ͑2001͒.
10. C.-H. Chen and A. Gewirth, J. Am. Chem. Soc., 114, 5439 ͑1992͒.
11. C.-H. Chen, K. D. Kepler, A. A. Gewirth, B. M. Ocko, and J. Wang, J. Phys.
Chem., 97, 7290 ͑1993͒.
7
59
ther characterized the Cd–Ag system by STM. A thin surface alloy
film forms as the result of structural changes in the full monolayer
that are initiated by place-exchange processes between the Cd ad-
layer and surface Ag atoms. An effective diffusion coefficient of
9
−
16
2
−1
12. C.-H. Chen, N. Washburn, and A. A. Gewirth, J. Phys. Chem., 97, 9754 ͑1993͒.
3. M. Hara, Y. Nagahara, J. Inukai, S. Yoshimoto, and K. Itaya, Electrochim. Acta,
1
.5 ϫ 10 cm s was obtained from charge-time transients using
1
an alloying potential of +10 mV with respect to bulk Cd. Based on
these kinetics, one can expect a Cd–Ag surface alloy measuring
approximately 0.5 nm in thickness to form after 20 s at room tem-
perature.
51, 2327 ͑2006͒.
14. C. A. Jeffrey, D. A. Harrington, and S. Morin, Surf. Sci., 512, L367 ͑2002͒.
5. N. J. Tao, J. Pan, Y. Li, P. I. Oden, J. A. DeRose, and S. M. Lindsay, Surf. Sci. Lett.,
1
2
71, L338 ͑1992͒.
1
1
6. K. Tamura, J. Wang, R. Adzic, and B. Ocko, J. Phys. Chem. B, 108, 1992 ͑2004͒.
7. M. F. Toney, J. N. Howard, J. Richer, G. L. Borges, J. G. Gordon, O. R. Melroy, D.
Yee, and L. B. Sorenson, Phys. Rev. Lett., 75, 4472 ͑1995͒.
The Cd–Ag data cited above suggests that the kinetics of alloy
formation in the upd region is sufficient to cause the stress transients
shown in Fig. 4. Although similar room-temperature data is not
available for the formation of Al–Au alloys in the upd region, this
system has been examined in some detail at higher temperatures in
1
8. M. R. Deakin and O. R. Melroy, J. Electroanal. Chem. Interfacial Electrochem.,
2
39, 321 ͑1988͒.
1
2
9. M. Hepel, K. Kanige, and S. Bruckenstein, Langmuir, 6, 1063 ͑1990͒.
0. O. Melroy, K. Kanazawa, J. G. G. Gordon II, and D. Buttry, Langmuir, 2, 697
͑1986͒.
1. G. R. Stafford and U. Bertocci, J. Phys. Chem. B, 110, 15493 ͑2006͒.
2. G. R. Stafford and U. Bertocci, J. Phys. Chem. C, 111, 17580 ͑2007͒.
3. W. Haiss and J. K. Sass, J. Electroanal. Chem., 386, 267 ͑1995͒.
4. W. Haiss, R. J. Nichols, and J. K. Sass, Surf. Sci., 388, 141 ͑1997͒.
47
AlCl –NaCl electrolyte. The presence of both AlAu and Al Au
3
2
2
5
2
2
2
2
has been confirmed after 2 h at 250°C. An Arrhenius plot of the
high-temperature data suggests that at 25°C, an Al–Au diffusion
zone of 0.27 nm forms after 60 s. This is in reasonable agreement
with the ambient-temperature surface alloy kinetics described above
for Cd–Ag. Phase-formation studies on Al–Au thin films have
25. O. E. Kongstein, U. Bertocci, and G. R. Stafford, J. Electrochem. Soc., 152, C116
2005͒.
͑
6
0
26. T. Trimble, L. Tang, N. Vasiljevic, N. Dimitrov, M. van Schilfgaarde, C. Friesen, C.
shown that both Al Au and AlAu also form at room temperature.
2
5
2
V. Thompson, S. C. Seel, J. A. Floro, and K. Sieradzki, Phys. Rev. Lett., 95,
Based on the molar volume-induced stress change described by Eq.
1
66106 ͑2005͒.
4
, the 0.27 nm of AlAu formed in 60 s, estimated from the high-
27. M. Seo and M. Yamazaki, J. Solid State Electrochem., 11, 1365 ͑2007͒.
2
2
2
8. H. L. Meyerheim, D. Sander, R. Popescu, and J. Kirschner, Phys. Rev. B, 64,
045414 ͑2001͒.
9. T. A. Brunt, E. D. Chabala, T. Rayment, S. J. O’shea, and M. E. Welland, J. Chem.
Soc., Faraday Trans., 92, 3807 ͑1996͒.
temperature kinetic data, would result in a surface stress of
−
1
+
1.0 N m , which is in the same range as that observed experimen-
tally. Although these kinetic studies suggest that surface alloying
may be responsible for the tensile stress observed in the latter stages
of Al upd on Au, the kinetics of Al–Au alloy formation at room
temperature, as well as the stress changes associated with this pro-
cess, will have to be evaluated more thoroughly.
30. A. Grossmann, W. Erley, J. B. Hannon, and H. Ibach, Phys. Rev. Lett., 77, 127
1996͒.
͑
3
3
3
1. R. C. Cammarata and K. Sieradzki, Annu. Rev. Mater. Sci., 24, 215 ͑1994͒.
2. R. C. Cammarata, K. Sieradzki, and F. Spaepen, J. Appl. Phys., 87, 1227 ͑2000͒.
3. C. Friesen, N. Dimitrov, R. C. Cammarata, and K. Sieradzki, Langmuir, 17, 807
͑
2001͒.
4. E. P. M. Leiva, M. G. Del Popolo, and W. Schmickler, Chem. Phys. Lett., 320, 393
2000͒.
5. L. A. Kibler, M. Kleinert, R. Randler, and D. M. Kolb, Surf. Sci., 443, 19 ͑1999͒.
36. G. R. Stafford, O. E. Kongstein, and G. M. Haarberg, J. Electrochem. Soc., 153,
C207 ͑2006͒.
Conclusion
3
͑
In situ stress measurements were made during Al upd onto ͑111͒-
3
textured Au from Lewis acidic AlCl -EMImCl using the wafer-
3
curvature method. The surface stress response consists of three dis-
tinct features. In the potential range of 1.2–0.6 V the surface stress
3
7. J. S. Wilkes, J. A. Levisky, R. A. Wilson, and C. L. Hussey, Inorg. Chem., 21,
263 ͑1982͒.
8. R. T. Carlin, W. Crawford, and M. Bersch, J. Electrochem. Soc., 139, 2720 ͑1992͒.
1
moves in the tensile ͑positive͒ direction from a value arbitrarily
3
−
4
chosen as zero. This likely involves the desorption of AlCl from the
39. Q. Liao, W. R. Pitner, G. Stewart, C. L. Hussey, and G. R. Stafford, J. Electrochem.
Au surface and is consistent with adsorbate-induced stress models
that appear in the literature. At the start of Al upd, the surface stress
moves in the compressive direction, in contrast to the tensile stress
expected based on the positive lattice misfit. We attribute this com-
pressive stress to the formation of Al–Au bonds which partially
satisfy the bonding requirements of the Au surface atoms, thereby
reducing the tensile surface stress inherent to the clean Au surface.
In the latter stages of Al upd, the surface stress once again moves in
the tensile direction, which we attribute to Al–Au alloying. The
magnitude of the tensile-stress change is close to that estimated from
the elastic strain associated with the change in molar volume, using
reaction kinetics reported for surface alloy formation in the upd
region.
Soc., 144, 936 ͑1997͒.
4
4
0. J. Robinson and R. A. Osteryoung, J. Electrochem. Soc., 127, 122 ͑1980͒.
1. G. R. Stafford and C. L. Hussey, Advances in Electrochemical Science and Engi-
neering, Vol. 7, R. C. Alkire and D. M. Kolb, Editors, Wiley-VCH, Weinheim,
Germany ͑2002͒.
42. S. Z. El Abedin, E. M. Moustafa, R. Hempelmann, H. Natter, and F. Endres,
ChemPhysChem, 7, 1535 ͑2006͒.
3. E. M. Moustafa, S. Z. El Abedin, A. Shkurankov, E. Zschippang, A. Y. Saad, A.
Bund, and F. Endres, J. Phys. Chem. B, 111, 4693 ͑2007͒.
4. G. G. Stoney, Proc. R. Soc. London, Ser. A, 82, 172 ͑1909͒.
45. J.-J. Lee, I. T. Bae, D. A. Scherson, B. Miller, and K. A. Wheeler, J. Electrochem.
Soc., 147, 562 ͑2000͒.
6. J.-J. Lee, Y. Mo, D. A. Scherson, B. Miller, and K. A. Wheeler, J. Electrochem.
Soc., 148, C799 ͑2001͒.
7. B. S. Radovic, R. A. H. Edwards, and J. N. Jovicevic, J. Electroanal. Chem., 428,
113 ͑1997͒.
8. C. A. Zell, F. Endres, and W. Freyland, Phys. Chem. Chem. Phys., 1, 697 ͑1999͒.
4
4
4
4
4
Acknowledgments
49. B. Fischer, J. V. Barth, A. Fricke, L. Nedelmann, and K. Kern, Surf. Sci., 389, 366
1997͒.
͑
The authors gratefully acknowledge the technical contributions
of Charles Hussey, Ugo Bertocci, Thomas Moffat, Jonathon Guyer,
5
0. A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and Ap-
plications, John Wiley and Sons, New York ͑1980͒.
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 155 ͑5͒ D408-D413 ͑2008͒
D413
5
5
1. S. Trasatti and O. A. Petrii, Pure Appl. Chem., 63, 711 ͑1991͒.
56. H. Ibach, Surf. Sci. Rep., 29, 193 ͑1997͒.
2. G. R. Stafford, V. D. Jovic, T. P. Moffat, Q. Zhu, S. Jones, and C. L. Hussey, in
Molten Salts, PV 99-41, p. 535, The Electrochemical Society Proceeding Series,
Pennington, NJ ͑1999͒.
57. G. Simmons and H. Wang, Single Crystal Elastic Constants and Calculated Ag-
gregate Properties: A Handbook, MIT Press, Cambridge, MA ͑1971͒.
58. H. Bort, K. Juttner, W. J. Lorenz, and G. Staikov, Electrochim. Acta, 28, 993
͑1983͒.
59. S. G. Garcia, D. R. Salinas, and G. Staikov, Surf. Sci., 576, 9 ͑2005͒.
60. G. Majni, C. Nobili, G. Ottaviana, M. Costato, and E. Galli, J. Appl. Phys., 52,
4047 ͑1981͒.
5
5
5
3. O. Mann and W. Freyland, J. Phys. Chem. C, 111, 9832 ͑2007͒.
4. H. Ibach, J. Vac. Sci. Technol. A, A12, 2240 ͑1994͒.
5. W. Haiss, R. J. Nichols, J. K. Sass, and K. P. Charle, J. Electroanal. Chem., 452,
1
99 ͑1998͒.
Downloaded on 2015-01-28 to IP 132.174.255.116 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).