The underpotential deposition of bismuth and tellurium on cold rolled silver substrate by ECALE

Article abstract of DOI:10.1016/j.electacta.2005.03.028

Thin-layer electrochemical studies of the underpotential deposition (UPD) of Bi and Te on cold rolled silver substrate have been performed. Different approaches have been employed to investigate the influence of silver oxide film on Bi UPD. As a result, the precedent deposition of a little bismuth can effectively prevent silver from surface oxidation. The voltammetric analysis of underpotential shift demonstrates that the first Te UPD on Bi-covered Ag and Bi UPD on Te-covered Ag fit UPD dynamics mechanism. Thin film of bismuth telluride was formed using an automated flow deposition system, by alternately depositing Te and Bi. The electrochemical conditions necessary to form Bi ₂Te₃ deposits of 50 cycles on cold rolled silver by ECALE are described here. X-ray diffraction indicated the deposits were Bi ₂Te₃. EDX quantitative analysis gave the 2:3 stoichiometric ratio of Bi to Te, which is consistent with XRD result. Electron probe microanalysis of the deposits showed a worm-like network structure. The map of Te and Bi element indicated the distribution of both Te and Bi is homogeneous and locates the same sites, which is favorable to Te-Bi binary system. The composition analysis of structural expanded image also showed the approximately constant composition of Te:Bi ≈ 3:2 has taken place.

Full text of DOI:10.1016/j.electacta.2005.03.028

Electrochimica Acta 50 (2005) 5465–5472
The underpotential deposition of bismuth and tellurium on
cold rolled silver substrate by ECALE
∗
W. Zhu, J.Y. Yang , X.H. Gao, J. Hou, S.Q. Bao, X.A. Fan
State Key Laboratory of Die and Mould Technology, College of Materials Science and Engineering,
Huazhong University of Science and Technology, Wuhan 430074, PR China
Received 9 January 2005; received in revised form 3 March 2005; accepted 5 March 2005
Available online 31 May 2005
Abstract
Thin-layer electrochemical studies of the underpotential deposition (UPD) of Bi and Te on cold rolled silver substrate have been performed.
Different approaches have been employed to investigate the influence of silver oxide film on Bi UPD. As a result, the precedent deposition of a
little bismuth can effectively prevent silver from surface oxidation. The voltammetric analysis of underpotential shift demonstrates that the first
Te UPD on Bi-covered Ag and Bi UPD on Te-covered Ag fit UPD dynamics mechanism. Thin film of bismuth telluride was formed using an
automated flow deposition system, by alternately depositing Te and Bi. The electrochemical conditions necessary to form Bi
2 3
Te deposits of
5
0 cycles on cold rolled silver by ECALE are described here. X-ray diffraction indicated the deposits were Bi Te . EDX quantitative analysis
2
3
gave the 2:3 stoichiometric ratio of Bi to Te, which is consistent with XRD result. Electron probe microanalysis of the deposits showed a
worm-like network structure. The map of Te and Bi element indicated the distribution of both Te and Bi is homogeneous and locates the
same sites, which is favorable to Te–Bi binary system. The composition analysis of structural expanded image also showed the approximately
constant composition of Te:Bi ≈ 3:2 has taken place.
©
2005 Elsevier Ltd. All rights reserved.
Keywords: ECALE; UPD; Bismuth telluride; Thermoelectric material; Thin film
1
. Introduction
Atomic layer epitaxy (ALE) can form thin film by using
surface limited reactions. Electrochemical atomic layer epi-
taxy (ECALE) [9–17] is the electrochemical analog of ALE.
ECALE involves the alternated electrochemical deposition
of elements to form a compound. Epitaxial deposition is
achieved by using underpotential deposition (UPD) as the
means to achieve surface chemistry-limited growth [18–22].
The phenomenon of UPD involves the deposition of an
atomic layer of one element on a second, at a potential prior to
(under) that needed to form deposits of the element on itself.
The driving force is generally thermodynamic, involving the
Gibbs energy of formation of a surface compound.
IIB–VIA compounds such as CdTe [9,10], CdS [11,12],
and ZnSe [13] have been successfully formed by using EC-
ALE, as well as IIIA–VA compound InAs [14], IIIA–VIA
compound In2Se3 [15], IVA–VIA compound PbSe [16],
VA–VIA compound Bi2S3 [17]. In our previous work, a first
trialandpreliminaryresultwerereportedonformingaBi2Te3
Bismuth telluride belong to the class of VA–VIA bi-
nary chalcogenide compound semiconductor with a nar-
row optical energy band gap of 0.13 eV and a very high
figure of merit [1–3]. These materials are widely used
for thermoelectric and optoelectronic devices, for example
in solid-state refrigeration, heat pumps, subminiature elec-
tronic devices, infrared sensors and high efficiency photo-
voltaic solar cells. Thin films of bismuth telluride and related
compounds have already been elaborated by flash evapo-
ration [4], co-evaporation [5], molecular beam epitaxy [6]
and metal–organic chemical vapor deposition [7,8]. These
methods are thermal and generally performed in vacuum.
∗
Corresponding author. Tel.: +86 27 87540944; fax: +86 27 87543776.
E-mail addresses: wennar@sina.com (W. Zhu),
jyyang@public.wh.hb.cn (J.Y. Yang).
0
013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.03.028

5
466
W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
3
+
VA–VIA group compound film by ECALE [23–25]. As a
follow-up in subsequent work, we report the electrodepo-
sition of Bi2Te3 on cold rolled silver substrate at underpo-
tential. The crystal structure of bismuth telluride is rhom-
bohedral with the space group (R3m). Crystal structure of
Bi2Te3 along the XZ plane shows the quintuple layer leaves
of Te–Bi–Te–Bi–Te separated by a van der Waals gap. It is
easy to cleave the crystals along the c-plane as the mechani-
cal strength is weak. The weakness of the c-plane mechanical
strength is a problem for fabrication of the thermoelements
used in the micro-modules for optoelectronic telecommu-
nications. This issue may be solved by forming a perfect
in-plane alignment texture tape. It has been found that the
perfect in-plane alignment texture tape was suitable to im-
prove the mechanical properties and to reach high values of
figure of merit [26]. Experiments have shown that the tex-
tured tape could originate from the epitaxial growth on tex-
turing substrates [27]. It is also known that cold rolled silver
can obtain a high level of biaxial texture [28]; it is thus a
promising substrate for the second generation of optoelec-
tronic telecommunication thermoelectric devices. In contrast
to cold rolled silver, hot rolled silver is not a good choice.
Hot rolling can occur the dynamic recrystallization during
deformation whether the hot rolled temperature is high or
low and cannot obtain some amount of texture component.
In this letter, a biaxial alignment is achieved by deposition
of the Bi2Te3 thin film directly on cold rolled silver by using
ECALE and the effects of varying surface structure of silver
substrates and electrochemical factors on the underpotential
deposition processes is analyzed.
Bismuth is soluble as Bi , but only for pH < 2 according
to the thermodynamic Eq. (4) [29]:
Bi3 + H2O ⇔ BiOH2+ + 4H+
+
2+
Bi3
+
]
[
BiOH
]
pH = 2 + log
(4)
[
At higher pH (pH > 4.98) for a Bi concentration of 0.1 mM,
bismuth precipitates as Bi O (Eq. (5)):
2
3
+
+
2
BiOH2 + H2O ⇔ Bi2O3 + 4H
pH = 2.98 − 0.5 log[BiOH
2+
]
(5)
2
+
In the pH range of 2–4.98, bismuth is soluble as BiOH .
Therefore, in order to dissolve bismuth, it is necessary to
work at a pH below 2.
2. Experimental
Solutions were prepared with high purity reagents and
twice-distilled water. All bismuth solutions consisted of
0.1 mM Bi(NO3)3·5H2O, and using 0.1 M HClO4 as a sup-
porting electrolyte, pH 1.5. Tellurium solutions were all
0.1 mM in TeO2, and also used 0.1 M HClO4 as a support-
ing electrolyte. The pH 8.5 Te solution was pH adjusted with
ammonia. Various blank rinse solutions were also utilized,
with a pH analogous to its respective deposition solution. All
solutions were deaerated by blowing purified N2 gas through
and over the solution for 30 min. All experiments were per-
formed at room temperature. The working electrodes were
cold rolled silver (99.99%) substrates. The substrates were
mirror-like polished mechanically and then annealed in a
SincesilversubstrateoxidizesreadilyunderlowerpHcon-
dition, so we were interested in the range of pH over which it
is possible to dissolve bismuth and tellurium, meanwhile not
to disrupt the structure of the silver surface through the for-
mation of a silver oxide layer. Tellurium is soluble as HTeO2+
in a narrow range at low pH (−0.37 < pH < 1.93) for a Te con-
centration of 0.1 mM according to the thermodynamic Eqs.
◦
muffle furnace under vacuum for 30 min at 650 C. Before
depositions, the electrode was polished chemically for 5 s in
a silver etch solution, which consisted of a mixture of 0.1 M
CrO3 + 0.1 M HCl. After polishing, the electrode was soaked
first in concentrated ammonia for about 5 min, and then in
concentrated sulfuric acid for about 20 min. Finally, it was
rinsed thoroughly with water.
An automated deposition apparatus consisting of peri-
staltic pumps, valves, programmable logistic computer
PLC), electrochemical flow cell and potentiostat was used
under the control of a computer. The electrolytic cavity was
delimited by the working electrode and the auxiliary elec-
trode, a plate of Pt. These electrodes were held apart by a
5 mm-thick gasket, which defined a 0.7 cm × 3 cm rectangu-
lar opening. The reference electrode, saturated calomel elec-
trode (SCE), was positioned at the cavity outlet. Both the
subminiature valve bank and the cell were developed by our
group and were described in our previous work [23,24].
The deposit is annealed in Ar to convert precursor layers
(1) and (2) [29,30]:
Te4 + 2H2O ⇔ HTeO2+ + 3H+
+
+
[
HTeO2 ]
pH = −0.37 + 0.333 log
(1)
(2)
Te4+
[
]
(
HTeO2 ⇔ TeO2 + H+
+
+
pH = −2.07 − log[HTeO2 ]
and also at high pH as TeO32 (pH > 8.36, Eq. (3) [29,30]):
−
TeO2 + H2O ⇔ TeO3 + 2H+
pH = 10.355 + 0.5 log[TeO3
2
−
2
−
]
(3)
In the pH range of 1.93 to 8.36, tellurium precipitates as
TeO2. In order to dissolve tellurium, we used concentrated
perchloride acid to dissolve 0.1 mM TeO2, and then the solu-
tion was pH adjusted with ammonia to form a pH 8.5 solution
◦
of the elements into the compounds in 200 C. Deposit stoi-
chiometric ratio was measured using Oxford Inca energy dis-
persive X-ray spectrophotometer (EDX). The XRD pattern
was obtained with a Philips-PW 1710 X-ray diffractometer
−
of HTeO3 .

W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
5467
using Co K␣ radiation. Electron probe microanalysis studies
were performed using a Joel JCXA-733 super probe.
3
. Results and discussion
3
.1. The UPD of bismuth on cold rolled Ag surface
The voltammetric profile in Fig. 1 represents the under-
potential deposition of bismuth on a cold rolled silver sur-
face. The potential in this experiment was initially scanned
in the negative direction starting from a potential of 0.04 V.
This potential was chosen because it is necessary to start the
deposition process with as little bismuth on the surface as
possible. It was important not to disrupt the structure of the
silver surface through the formation of a silver oxide layer.
This deposition of a little bismuth was to prevent silver from
surface oxidation, and it appeared that at least initially, the
oxidation of silver was suppressed. The scan proceeded in the
negative direction and resulted in a reduction peak at−0.03 V,
labeled C1. The subsequent anodic stripping peaks were ob-
served at −0.024 and 0.08 V, labeled A2 and A1, respec-
tively (Fig. 1a). Fig. 1b showed the increase in the cathodic
electrode polarization limit. When the potential is scanned
towards more negative value (−0.4 V), peak A2 increased
Fig. 2. Cyclic voltammogram into the more negative region with a Bi-
covered silver electrode in blank rinse solution of pH 1.5. The silver electrode
was held at −0.2 V in bismuth solution, then the bismuth solution was re-
placed by the blank rinse solution and the potential sweep was performed.
The scanning rate is 10 mV/s.
and A1 remained unchanged. The charge corresponding to
2
the A1 peak was 227 ␮C/cm , assuming that the reduction
3+
of one Bi ion involves three electrons, this charge corre-
sponds to 0.4 monolayers. At this low concentration, there
are no discernible peaks related to the bulk deposition of Bi
3+
on silver surface. In fact, increasing Bi concentration be-
yond 0.4 mM just causes a reduction peak at −0.4 V, so we
labeled C2 at the position. Concentration dependent growth
examination also showed that peak C1 and A1 were indepen-
dent on Bi ion concentration whereas peak C2 and A2 was
considerablydependentonconcentration. Thissuggestedthat
peak C1 and A1 were corresponding to UPD Bi deposition
and stripping respectively while peak C2 and A2 were corre-
sponding to bulk Bi deposition and striping.
Potential cycling into the more negative region was also
performed with a Bi-covered silver electrode, and is depicted
in Fig. 2. This experiment was performed in a blank rinse
solution following an exposure of the silver electrode to bis-
muth solution (pH 1.5) at −0.2 V. These were two reductive
voltammetric features; the first peak occurs at about −0.57 V
(labeled C1). The fact that the C1 peak grew with prolonging
the polarized time at −0.2 V, would imply that the reduc-
tion peak was associated with Bi coverages on the electrode
surface. It is also noticeable that there is a re-reduction prop-
erty for bismuth when the cathodic electrode polarization was
rather negative at acid solutions [29]. For these reasons, we
assign the C1 peak as the reduction of Bi to BiH3 according
to the following general reaction:
+
−
Bi + 3H + 3e ⇒ BiH3
(6)
The second reduction feature C2 occurs near −0.7 V and
+
is the onset of reduction of H to H2. On the subsequent
oxidative scan, a single small peak A1 was observed. Because
BiH3 is generally not stable, so the peak A1 is responsible
for the Bi3 ions tend to escape from the electrode.
Potential cycling of Bi on silver surface with silver oxide
film was also performed in order to investigate the influence
−
Fig. 1. Cyclic voltammogram of Bi on cold rolled silver surface starting from
0
.04 V. (a) the scanning scope was from 0.25 to −0.25 V; (b) the scanning
scope was from 0.25 to −0.4 V. The scanning rate is 10 mV/s.

5
468
W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
Fig. 4. Cyclic voltammograms into the silver oxidation region with a Bi-
covered silver electrode in blank rinse solution of pH 1.5. The numeral indi-
cates consecutive scans, from 0.3 to −0.7 V. The scanning rate is 10 mV/s.
Fig. 3. Cyclic voltammogram of Bi on silver surface with silver oxide film.
The scanning rate is 10 mV/s.
of silver oxide film adsorbed on the surface of the electrode
on Bi UPD. Initially, this silver oxide adsorption/desorption
process was preformed in clean supporting electrolyte. When
more positive potentials were applied, a silver oxide film was
formed on the surface of the electrode. The silver oxide film
was subsequently reduced during the cathodic scan. Subse-
quently, a single aliquot of Bi solution was injected, and the
potential was cycled. The resulting voltammogram (Fig. 3)
is noticeable different from the voltammetry at electrode sur-
face without silver oxide film. Briefly, this is a negative shift
cyclic voltammogram there was a loss of Bi upon potential
cyclingintotheBire-reductionregion. Withrepeatedcycling,
the peak belonging to Bi deposition gradually decreased. Fol-
lowing the first adsorbate Bi remove, bare silver surface be-
came available for silver oxidation. This conclusion was in
line with previous studies, which have shown that this deposi-
tion of a little Bi could effectively prevent silver from surface
oxidation. Comparing with Fig. 3, the intensity of silver oxi-
dation/reduction in Fig. 4 is larger. One of the reasons for this
might be that the silver oxide adsorption/desorption process
has been preformed reproducibly prior to the curve equiva-
lent to that of Fig. 3 was obtained. The precedent silver oxide
adsorption/desorption process and the adsorbed little bismuth
on the surface of electrode during the injecting of Bi solution
resulted in the suppression of silver oxidation/reduction in
Fig. 3.
(by approximately 70 mV) in Bi deposition potential (labeled
C1), comparing with C1 in Fig. 1a. Anodic stripping peak A1
is broader and shifted to more positive potential, resulting in
the overlapping of the bulk and UPD stripping peak. The
splitting between the deposition and stripping peak has been
increased as compared with on a silver surface free from sur-
face oxides. It is believed that the presence of the oxide layer
inhibits the deposition of Bi, thus leading to the observed po-
tential shift. In such a case, the deposition of Bi is precluded
until the onset of silver oxide reduction at 0.07 V, yielding
bare silver surface sites on which Bi deposition may occur.
In addition to the change in potential, the Bi deposition peak
C1 was broader as well, suggesting that Bi deposition took
place over a number of energetically different sites, resulting
from the imposed surface disorder by silver surface oxide.
The current–potential curve into the silver oxidation re-
gion with a Bi-covered silver electrode was shown in Fig. 4.
This figure consists of three consecutive cyclic voltammo-
grams in a blank rinse solution following exposure of the
silver electrode to bismuth solution (pH 1.5) at −0.1 V. It
was apparent that surface oxidation of silver only took place
at third cyclic voltammogram. In fact, according to Eq. (6),
when potential cycled into the Bi re-reduction region, it
would occur the re-reduction of Bi to BiH3. However, the re-
reduction production BiH3 is extreme unstable. On the subse-
quent oxidative scan, BiH3 tend to escape from the electrode
surface and are decomposed as bismuth precipitates into the
blank rinse solution, and, therefore, it would not cause Bi re-
deposition on the electrode surface. So at the first and second
3.2. The UPD of tellurium on cold rolled Ag surface
The cyclic voltammogram of TeO in ammonia buffer so-
2
lutions of pH 8.5 was shown in Fig. 5. The scan proceeded
from −0.1 to −0.7 V and resulted in a reduction peak at
Fig. 5. Cyclic voltammograms of 0.1 mM TeO2 in an ammonia buffer solu-
tion of pH 8.5 on cold rolled Ag surfaces. The numeral indicates consecutive
scans. The scanning rate is 10 mV/s.

W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
5469
Fig. 6. The cyclic voltammogram curves, obtained in a blank rinse solution.
Fig. 7. Cathodic stripping of the previously deposited Te, after replacing the
tellurium solution by a 0.1 M KOH solution. The scanning rate is 10 mV/s.
The electrode is polarized at −0.6 V in tellurium solution for 20 s−1, 40 s−2,
6
0 s−3, then the tellurium solution is changed by the blank rinse solution and
the cathodal potential scan is performed from −0.7 to −1.5 V. The scanning
rate is 10 mV/s. Dot line was obtained after Te electrode had been held at
Fig. 7 shows distinct UPD and the bulk reduction peaks, de-
noted by A and C.
−1.4 V for 5 s.
In conclusion, the suggested procedure to obtain a UPD
layeroftelluriumconsistsofdepositinganexcessoftellurium
and then applying a potential of −1.4 V, which is sufficient to
reduce bulk tellurium, but not the underpotentially deposited
Te, and Te (UPD) is left.
−
0.37 V, labeled C1, which is only observed during the first
potential scan. The subsequent bulk reduction peaks was ob-
servedat−0.5 V, labeledC2. IntegrationofthepeakC1yields
2
a charge of about 353 ␮C/cm . Assuming that the reduction
of one Te(IV) atom involves four electrons, this charge corre-
sponds to 0.414 monolayers, indicating C1 was correspond-
ing to UPD Te deposition. The UPD nature of peak C1 is
also supported by the fact that it disappears in the successive
scannings, in which only bulk TeO2 reduction is observed.
Due to the high irreversibility of the system, the reoxidation
of the underpotentially deposited Te was prevented by sil-
ver oxidation, and the remained UPD layers blocked off the
formation of new UPD. This is in agreement with the result
in the Cd–Te system on Ag (1 1 1) reported by Foresti et al.
+
−
−
Te(UPD) + Te(bulk) + H + 2e ⇒ Te(UPD) + HTe
(7)
As mentioned above, in that program a little bulk Te and UPD
Te were first deposited at −0.6 V, followed by rinsing out of
−
the excess HTeO3 . No Te(IV) was present when the poten-
tial was subsequently shifted to −1.4 V in order to convert the
excess Te to Te(II), leaving the Te atomic layer. The reduction
−
must be performed after replacing the HTeO3 solution by a
solution of supporting electrolyte (ammonia buffer) to avoid
the conproportionation reaction between Te(II) ions formed
in the reduction and the tellurite ions present in the solution.
After holding the electrode at −1.4 V for 5 s, the potential
is cycled. Evidently the bulk reduction peak C disappeared
(Fig. 6, dot line), indicating that the reduction of bulk Te to
[10].
The cyclic voltammogram shown in Fig. 6, at more neg-
ative potentials, were obtained in a blank rinse solution af-
ter deposition of both UPD and bulk Te. Initially, Te was
deposited at −0.6 V for an interval, t. Subsequently, the tel-
lurium solution is changed by a blank rinse solution at −0.7 V,
and the potential was cycled between −0.7 and −1.5 V caus-
ing the appearance of a reduction peak at −1.3 V, labeled C.
When the duration of the deposition of both UPD and bulk
Te increases, the charge associated with the reduction peak
C becomes larger. According to Refs. [9,10], this peak cor-
−
HTe is a rapid process.
3.3. The UPD of bismuth on Te-covered Ag surface
The cyclic voltammogram of Bi at the Te-covered Ag sur-
face is quite different from the voltammetry at bare Ag sur-
face. Briefly, an initial peak, labeled C1 in Fig. 8, occurs at
about −0.1 V, is shifted by about 70 mV to a less positive
potential compared with bare Ag. Anodic stripping peaks A1
and A2 have a shift of 20 and 30 mV to more positive po-
tential respectively compared with the voltammogram of Bi
on bare Ag (Fig. 1). The underpotential shift, ꢀEp, the dif-
ference between the UPD potential and bulk deposition po-
tential, is smaller for C1 on Te-covered Ag than on bare Ag.
This observation indicates that the deposition of Bi at −0.1 V
on the Te-covered Ag surface requires less energy than the
deposition on bare Ag. The underpotential shift ꢀEp, is re-
lated to the difference, ꢀϕ, between the work function of the
−
responds to Te(0) re-reduction to HTe .
A second peak of constant height is observed at more neg-
ative potentials and is labeled A. This peak corresponds to
the reduction of the first atomic layer of tellurium, which is
more strongly bound to the silver substrate and can, there-
fore, be regarded as a UPD layer. The UPD reduction peak is
partially obscured by hydrogen evolution and it can be seen
more clearly if, after depositing Te(0), the tellurium solution
is replaced by a pH 13 supporting electrolyte solution alone,
and cathodic stripping is performed (Fig. 7). In this way, hy-
drogen evolution is shifted towards more negative potentials.

5
470
W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
Fig. 8. Cyclic voltammogram of Bi on Te-covered cold rolled silver surface.
The scanning rate is 10 mV/s.
depositing metal and that of the substrate. Therefore, inter-
actions of Bi with the substrate weaker than that with the
adsorbed Te atoms could shift the underpotential deposition
towards more negative potential. For the same reason, the
shift of anodic stripping peaks towards more positive poten-
tials must be caused by the interaction of Bi atoms with the
adsorbed Te atoms, which is obviously stronger than that with
the substrate atoms and makes the Bi deposit more stable.
Fig. 10. Diagram of an ECALE cycle for Bi2Te3 formation.
3
.5. Further UPD of bismuth and tellurium alternately
The growth of further UPD of bismuth and tellurium al-
ternately was obtained by keeping the electrode at −0.6 V for
3
.4. The UPD of tellurium on Bi-covered Ag surface
2
0 s in a Te solution, washing the cell with blank solution,
shifting the potential to −1.4 V for 5 s. The Bi blank solu-
tion was then introduced. The Te and Bi solutions differed in
pH by several units, so that rinsing with the Bi blank solu-
tion served to condition the cell and the tubing prior to the
introduction of the Bi solution, avoiding incidental stripping
of previous Te UPD layer in a higher pH. After the Bi blank
solution, the Bi solution was pumped in and held quiescent at
The cyclic voltammogram of Te at the Bi-covered Ag sur-
face is shown in Fig. 9. In comparison with Fig. 5, the Te
UPD peak has not been observed. However, the Te bulk de-
position peak C2 is obviously stronger than that on bare Ag
substrate. This means that the presence of bismuth somehow
hinders Te deposition. In other words, the interactions of Te
with the substrate, which are responsible for the underpoten-
tial deposition, become weaker in the presence of Bi, so that
Te UPD might shift towards more negative potentials. How-
ever, we can reasonably assume that bulk depositions are not
significantly affected by the substrate. Hence this UPD peak
is very close to the bulk reduction peak, which overlaps with
bulk reduction peak and results in the increase of C2.
−
0.18 V for 20 s, while Bi UPD layer was deposited, wash-
ing the cell and repeating this cycle as many times as desired.
The basic ECALE cycle for the deposition is schematized in
Fig. 10.
The current time traces for this procedure were shown in
Fig. 11. Using the current time traces, we can examine the
Fig. 9. Cyclic voltammogram of Te on Bi-covered cold rolled silver surface.
The numeral indicates consecutive scans. The scanning rate is 10 mV/s.
Fig. 11. Plots of the current involved in UPD of bismuth and tellurium
alternately as a function of deposition time.

W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
5471
substrate. In Fig. 13c and d, the bright areas represented Te
and Bi element, respectively. One explanation for the mor-
phology is that it results from the sequential preferential
deposition of monolayers of Bi2Te3 on nucleation sites, on
the surface. The preferential nucleation sites mainly originate
from the surface defects of the cold rolledsilver substrate. De-
fect sites such as steps and kinks have been found to possess a
significant energy of adsorption that may, in some instances,
favor the preferential nucleation. The random domains on the
polycrystalline substrate surface also may provide energeti-
cally distinct sites for the preferential adsorption. In addi-
tion, the lattice mismatch of Ag with Bi2Te3 is close to 7%,
which might result in the formation of an array of disloca-
tions, within the first few deposited Bi2Te3 monolayers.
It is interesting to note that the worm-like structure is
comprised simultaneously of Bi and Te, which seems to be
favorabletoTe–Bibinarysystem. Thishypothesiscanbecon-
firmed by the map of Te and Bi element on the surface layer
of sample shown in Fig. 13c and d. The bright areas observed
at the two panels were Te and Bi element, respectively. It is
found that in surface layer of sample, the distribution of both
Fig. 12. X-ray pattern of Bi2Te3 electrodeposited thin film on Ag.
deposition process for per step in real time. From integration
of the currents for deposition, the Bi and Te coverages per
cycle, from coulometry, averaged 0.6 mL of Te and 0.4 mL
of Bi. EDX quantitative analysis of the 50 cycles sample, ob-
tained by this procedure, gave the 2:3 stoichiometric ratio of
Bi to Te, as expected for the formation of the Bi2Te3 com-
pound. Fig. 12 is an X-ray diffraction pattern for a 50-cycles
deposit. Two peaks for Bi2Te3 are evident, with the [0 1 5]
reflection most prominent.
An electron probe microanalyzer (EPMA) attachment was
used for Bi and Te elemental analysis. Electron probe micro-
analysis images of the 50 cycles sample showed a clear two-
dimensional worm-like network structure (Fig. 13). Fig. 13a
showed the back-scattering electron image of the sample.
Fig. 13b–d were the map patterns of sample, showing the
distribution of silver, tellurium and bismuth, respectively. In
Fig. 13b the dark areas were corresponding to Bi2Te3 com-
pounds while the bright areas were corresponding the silver
Fig. 13. EPMA images of the 50 cycles sample: (a) back-scattering electron
image; (b–d) map patterns of sample, showing the distribution of silver,
tellurium and bismuth, respectively.
Fig. 14. A line scan of electron probe microanalysis on the worm-like struc-
ture island section.

5
472
W. Zhu et al. / Electrochimica Acta 50 (2005) 5465–5472
Te and Bi is homogeneous and their deposition positions are
almost the same.
the National Foundation of China for Returnee Student from
Abroad.
Fig. 14 shows an enlarged image of the worm-like struc-
ture island. The distributions of Te and Bi along a line in the
section are also shown at Fig. 14. It is quite apparent that
Te and Bi uptake is more pronounced in the island section.
While the average concentration of Te attains values above
References
[
1] J.Y. Yang, T. Aizawa, A. Yamamoto, T. Ohta, Mater. Chem. Phy. 70
2001) 90.
2] J.Y. Yang, T. Aizawa, A. Yamamoto, T. Ohta, Mater. Sci. Eng. B 85
2001) 34.
[3] J.Y. Yang, T. Aizawa, A. Yamamoto, T. Ohta, J. Alloys Comp. 312
2000) 326.
(
5
7.49% in the island region, the concentration of Bi in the is-
[
land part is about 42.51%, indicating that the approximately
constant composition of Te:Bi ≈ 3:2 has taken place. No sig-
nificant uptake of Te could be detected by EPMA outside the
island. The region outside the island mainly consists of silver
including a small quantity of Bi. However, there is no single
phase Bi exists in the XRD pattern. So the Bi element may
originates from the adsorption of Bi on Ag surface, which is
very slight and cannot be examined by XRD.
(
(
[4] V.D. Das, R.C. Mallik, J. Mater. Res. Bull. 37 (2002) 1961.
[5] H. Zou, D.M. Rowe, S.G. Williams, Thin Solid Films 408 (2002)
270.
[
[
[
[
6] H. Beyer, J. Nurnus, H. Bottner, A. Lambrecht, E. Wagner, G. Bauer,
J. Phys. E 13 (2002) 965.
7] B. Aboulfarash, A. Mzerd, A. Giani, A. Boulouz, F. Pascal-Delannoy,
A. Foucaran, A. Boyer, J. Mater. Chem. Phys. 62 (2000) 179.
8] A. Giani, A. Boulouz, F. Pascal-Delannoy, A. Foucaran, A. Boyer,
Thin Solid Films 315 (1998) 99.
9] B.H. Flowers Jr., T.L. Wade, J.W. Garvey, M. Lay, U. Happek, J.L.
Stickney, J. Electroanal. Chem. 524 (2002) 273.
4
. Conclusions
[
10] F. Forni, M. Innocenti, G. Pezzatini, M.L. Foresti, Electrochim. Acta
5 (2000) 3225.
A number of conclusions may be drawn from these ini-
4
tial studies of Bi and Te underpotential deposition. The first
involves that the underpotential deposition of Bi and Te on
cold rolled Ag substrate is a strong function of the initial state
of the electrode surface. The precedent deposition of a little
bismuth can effectively prevent silver from surface oxidation.
It also appears that the presence of the silver oxide layer can
inhibit the deposition of Bi. In such a case, the deposition
of Bi is precluded until the onset of silver oxide reduction.
Subsequently silver oxide reduction imposed the silver sur-
face disorder and resulted in Bi deposition took place over
a number of energetically different sites. In addition, on Te-
covered Ag surfaces, the interactions of Bi underpotential
deposit with the adsorbed Te atoms are obviously stronger
than that with the substrate atoms, which results in the depo-
sition of Bi on the Te-covered Ag surface requires less energy
and makes the Bi deposit more stable. The same result was
observed when Te UPD on Bi-covered Ag surface.
Results obtained from the alternated deposition studies
on Ag indicate that the deposits were Bi2Te3. The worm-
like network structure results from the surface defects of the
cold rolled Ag substrate and the lattice mismatch between
substrate and deposit.
Future work will be devoted to grow thin films with a
highernumberofcyclesandstudytheircompositionandmor-
phology and structure control on an ordered, lattice-matched
single-crystal substrate. These research works are now in
progress.
[
11] T. Cecconi, A. Atrei, U. Bardi, F. Forni, M. Innocenti, F. Loglio,
M.L. Foresti, G. Rovida, J. Electr. Spectrosc. Rel. Phenom. 114–116
(2001) 563.
[12] M. Innocenti, S. Cattarin, M. Cavallini, F. Loglio, M.L. Foresti, J.
Electroanal. Chem. 532 (2002) 219.
[13] G. Pezzatini, S. Caporali, M. Innocenti, M.L. Foresti, J. Electroanal.
Chem. 475 (1999) 164.
[14] M. Innocenti, F. Forni, G. Pezzatini, R. Raiteri, F. Loglio, M.L.
Foresti, J. Electroanal. Chem. 514 (2001) 75.
[15] R. Vaidyanathan, J.L. Stickney, S.M. Cox, S.P. Compton, U. Happek,
J. Electroanal. Chem. 559 (2003) 55.
[16] R. Vaidyanathan, J.L. Stickney, U. Happek, Electrochim. Acta 49
(2004) 1321.
[
[
17] T. Oznuluer, U. Demir, J. Electroanal. Chem. 529 (2002) 34.
18] J.L. Stickney, in: D.M. Kolb, R. Alkire (Eds.), Advances in Electro-
chemical Science and Engineering, Wiley, Weinheim, 2002, p. 1.
19] K. Varazo, M.D. Lay, T.A. Sorenson, J.L. Stickney, J. Electroanal.
Chem. 522 (2002) 104.
[
[
[
20] M. Innocenti, G. Pezzatini, F. Forni, M.L. Foresti, J. Electrochem.
Soc. 148 (2001) c357.
21] J.L. Stickney, T.L. Wade, B.H. Flowers Jr., R. Vaidyanathan, U.
Happek, in: E. Gileadi, M. Urbakh (Eds.), Encyclopedia of Electro-
chemistry, Marcel Dekker, New York, 2002.
[22] E.A. Streltsov, S.K. Poznyak, N.P. Osipovich, J. Electroanal. Chem.
518 (2002) 103.
[23] W. Zhu, J.Y. Yang, X.H. Gao, J. Hou, T.J. Zhang, K. Cui, Trans.
Nonferrous Met. Soc., China, in press.
[24] W. Zhu, J.Y. Yang, X.H. Gao, S.Q. Bao, X.A. Fan, T.J. Zhang, K.
Cui, Electrochim. Acta, in pres.
[
[
[
[
[
25] J.Y. Yang, W. Zhu, X.H. Gao, S.Q. Bao, X.A. Fan, J. Electroanal.
Chem. 577 (2005) 117.
26] M. Susumu, S. Yasunori, F. Katsushi, N. Keio, I. Keisuke, Mater.
Sci. Eng. A 277 (2000) 244.
27] E. Bellingeri, H.L. Suo, J.-Y. Genoud, M. Schindl, E. Walker, R.
Flukiger, IEEE Trans. Appl. Supercond. 11 (2001) 3122.
28] H.L. Suo, M.L. Zhou, L.H. Zhai, Z. Sun, X.F. X, W. Han, T.Y. Zuo,
Rare Met. 16 (1997) 221.
Acknowledgements
This work is co-financed by the National Basic Research
Program (2004CCA03200), the Chinese National Natural
Science Foundation (50401008), the National Foundation
for Quick Response to Fundamental Research of China, and
29] A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous
Solution, Marcel Dekker, New York, 1985.
[30] D.R. Lide, The CRC Handbook of Chemistry and Physics, vol. 80,
CRC Press Boca Raton, Florida 1999–2000.