Optimization studies of HgSe thin film deposition by electrochemical atomic layer epitaxy (EC-ALE)

Article abstract of DOI:10.1016/j.electacta.2005.12.012

Studies of the optimization of HgSe thin film deposition using electrochemical atomic layer epitaxy (EC-ALE) are reported. Cyclic voltammetry was used to obtain approximate deposition potentials for each element. These potentials were then coupled with their respective solutions to deposit atomic layers of the elements, in a cycle. The cycle, used with an automated flow deposition system, was then repeated to form thin films, the number of cycles performed determining the thickness of the deposit. In the formation of HgSe, the effect of Hg and Se deposition potentials, and a Se stripping potential, were adjusted to optimize the deposition program. Electron probe microanalysis (EPMA) of 100 cycle deposits, grown using the optimized program, showed a Se/Hg ratio of 1.08. Ellipsometric measurements of the deposit indicated a thickness of 19 nm, where 35 nm was expected. X-ray diffraction displayed a pattern consistent with the formation of a zinc blende structure, with a strong (1 1 1) preferred orientation. Glancing angle fourier transform infrared spectroscopy (FTIR) absorption measurements of the deposit suggested a negative gap of 0.60 eV.

Full text of DOI:10.1016/j.electacta.2005.12.012

Electrochimica Acta 51 (2006) 4347–4351
Optimization studies of HgSe thin film deposition by
electrochemical atomic layer epitaxy (EC-ALE)
Venkatram Venkatasamya, Mkhulu K. Mathea, Stephen M. Coxb,
Uwe Happekb, John L. Stickneya,∗
a Department of Chemistry, University of Georgia, Athens, GA, United States
b Department of Astronomy and Physics, Athens, GA, United States
Received 10 October 2005; received in revised form 9 December 2005; accepted 10 December 2005
Available online 19 January 2006
Abstract
Studies of the optimization of HgSe thin film deposition using electrochemical atomic layer epitaxy (EC-ALE) are reported. Cyclic voltammetry
was used to obtain approximate deposition potentials for each element. These potentials were then coupled with their respective solutions to deposit
atomic layers of the elements, in a cycle. The cycle, used with an automated flow deposition system, was then repeated to form thin films, the
number of cycles performed determining the thickness of the deposit. In the formation of HgSe, the effect of Hg and Se deposition potentials, and a
Se stripping potential, were adjusted to optimize the deposition program. Electron probe microanalysis (EPMA) of 100 cycle deposits, grown using
the optimized program, showed a Se/Hg ratio of 1.08. Ellipsometric measurements of the deposit indicated a thickness of 19 nm, where 35 nm was
expected. X-ray diffraction displayed a pattern consistent with the formation of a zinc blende structure, with a strong (1 1 1) preferred orientation.
Glancing angle fourier transform infrared spectroscopy (FTIR) absorption measurements of the deposit suggested a negative gap of 0.60 eV.
© 2005 Elsevier Ltd. All rights reserved.
PACS: 81.05. D; 81.05. B, C, D, E, G; 81.15. P; 82.45. Q; 61.10. N; 78.66. H
Keywords: B1. HgSe; A3. EC-ALE; A3. UPD; A3. Electrodeposition; A1. XRD; A1. EPMA; A1. FTIR; A3 ALE; A3 ALD
1. Introduction
layer by layer, using surface limited reactions, under potential
deposition (UPD) in the case of EC-ALE. UPD refers to the elec-
trodeposition of an atomic layer of a first element on a second,
at a potential prior to, under, that needed to deposit the element
on itself [22–26]. A number of II–VI compounds such as CdTe,
CdS and ZnSe have been successfully grown using EC-ALE
[27–30], as well as some III–V compounds such as GaAs, InAs
and InSb [28,29,31,32]. Recently PbSe [33], PbTe and Bi2Te3
[34] have also been grown using EC-ALE.
This paper is an extension of previous work on HgSe depo-
sition by this group [11]. Attempts have been made to optimize
the deposition cycle of HgSe, by adjusting the potentials for Se
and Hg used in the EC-ALE cycle.
Mercury selenide, HgSe, is a II–VI compound semicon-
ductor. Based on its electrical properties, it is classified as a
semimetal or degenerate semiconductor [1–3]. It is of special
interest for fundamental studies because of its inverted band
structure [4]. Some of the possible applications for mercury
selenide lie in optoelectronics. Molecular beam epitaxy (MBE)
[5,6], chemical bath deposition [7–9] and the cold traveling
heater method (CTHM) [10] are some of the methods previ-
ously used to deposit HgSe.
Recently, this group reported the first deposits of mercury
selenide formed via. electrochemical atomic layer epitaxy (EC-
ALE) [11]. EC-ALE is the electrochemical analog of atomic
layerepitaxy(ALE)[12–17], andatomiclayerdeposition(ALD)
[18–21]. All are methods for which the deposits are grown
2. Experimental
Depositionswereperformedusingathinlayerflowelectrode-
position system [16,35,36], which consisted of pumps, valves,
a flow cell and a potentiostat. All components were computer
controlled using a LABVIEW program.
∗
Corresponding author. Tel.: +1 706 542 2726; fax: +1 706 542 9454.
E-mail address: stickney@chem.uga.edu (J.L. Stickney).
0013-4686/$ – see front matter © 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2005.12.012

4348
V. Venkatasamy et al. / Electrochimica Acta 51 (2006) 4347–4351
The flow cell has been previously described [37], with minor
design changes to the reference compartment. A Teflon tube
fitting was changed to a simple O-ring, which provided a bet-
ter seal. The auxiliary electrode was an ITO glass slide, and
the reference electrode was Ag/AgCl (3 M NaCl) (Bioanalyti-
cal Systems Inc., West Lafayette, IN). Substrates consisted of
300 nm thick gold films on glass. The substrates were annealed
at 400 ◦C for 12 h under a vacuum of 10−6 Torr, after Au vapor
deposition, resulting in a (1 1 1) habit.
The solutions used were 0.2 mM HgO, pH 2 and 0.5 mM
SeO2, pH 3. Both solutions contained 0.5 M Na2SO4 as a sup-
porting electrolyte. The blank solution contained only the 0.5 M
Na2SO4, at pH 4. Solution pH was adjusted using H2SO4. The
water used to make solutions was supplied from a Nanopure
water filtration system (Barnstead, Dubuque, IA) attached to the
house DI water system. Chemicals were reagent grade or better.
The EC-ALE cycle used to deposit HgSe was performed
as follows: the Se solution was flushed into the cell for 2 s
(40 mL/min), and then held quiescent for 15 s, all at the cho-
sen Se deposition potential. Blank solution was then flushed
through the cell for 3 s. This was followed by filling the cell
with the Hg solution for 2 s, and holding quiescent for 15 s for
deposition. The cycle was then completed by flushing with blank
solution for 3 s. This cycle was repeated 100 times for each
experiment.
Fig. 1. Cyclic voltammogram of Au electrode in 0.5 mM HSeO3−, pH 3 (elec-
trode area: 4 cm2; scan rate: 5 mV/s).
A series of experiments were then performed where the depo-
sition potentials for Se were varied from 0.20 to −0.20 V, while
keeping the same deposition potential for Hg at 0.45 V (Fig. 3).
No deposits were evident until the potential for Se deposi-
tion was −0.10 V or below. Optical microscopy and EPMA
were performed on each deposit to determine homogeneity
and stoichiometry, respectively. Based on deposit homogene-
ity, −0.15 V was selected as a good Se deposition potential. The
resulting deposits appeared homogeneous and stoichiometric,
however, ellipsometric data suggested the deposits were close
Deposit thickness was monitored using a single wavelength
ellipsometer (Sentech SE 400). A Scintag, PAD-V diffractome-
˚
ter with Cu K␣ radiation (λ = 1.5418 A), was used to obtain the
glancing angle X-ray diffraction patterns. Electron probe Micro-
analysis (EPMA) was run on a Joel 8600 wavelength dispersive
scanning electron microprobe.
Glancing angle absorption measurements were performed,
using an FTIR spectrophotometer (Bruker FTS-66v, Bruker
optics Inc.).
3. Results and discussions
Cyclic voltammetry was used to determine approximate start-
ing deposition potentials for the EC-ALE cycle. The voltammet-
ric behaviors of HSeO3− and Hg2+ with gold on glass substrates
are shown in Figs. 1 and 2. Potentials of 0.20 V for Se and 0.45 V
for Hg were identified as reasonable initial potentials for the
EC-ALE cycle. Thus, the initial program went as follows: the
cell was filled with the Se solution, pumped for 2 s at 0.20 V,
no flow (static) for 15 s for deposition, then blank solution was
flushed for 3 s at 0.2 V. The Hg solution flush in, pumped for
2 s at 0.45 V, static for 15 s, and blank solution flush for 3 s at
0.45 V. The intent was that this cycle would ideally result in the
deposition of one compound monolayer.
Using these conditions, no deposition was observed by visual
inspection after 100 cycles. Evidently the potentials chosen were
insufficiently negative to drive the deposition of HgSe, even
though UPD of these elements on Au looked promising. This
points out that the potentials derived from voltammetry on the
substrate are simply a good first approximation for the potentials
needed for a cycle, but do not represent the potentials needed to
deposit the elements on each other.
Fig. 2. Cyclic voltammogram of Au electrode in 0.2 mM Hg2+, pH 2 (electrode
area: 4 cm2; scan rate: 5 mV/s).

V. Venkatasamy et al. / Electrochimica Acta 51 (2006) 4347–4351
4349
Fig. 3. Effect of Se deposition potential on HgSe deposition (Hg deposition
potential: 0.48 V; Se stripping potential: −0.63 V).
Fig. 5. Effect of Se stripping potential on deposit thickness.
to 100 nm thick, significantly thicker than expected for a sim-
ple model of one compound monolayer per cycle, where a 100
cycle deposit was expected to result in a 37.4 nm thick film. This
model assumes one Hg–Se bi-layer grows each cycle, with the
(1 1 1) orientation [37].
Previous studies of Se deposition [38,39] have shown that Se
does not result in a classic underpotential deposition process. On
the contrary, Se deposition requires an over potential, but bulk
deposition of Se is so slow that a surface limited feature is still
visible. The result is that along with the surface limited reaction,
formation of an atomic layer of Se, a small amount of bulk Se
is deposited as well. This can be seen from the current time pro-
file, during Se deposition (Fig. 4). The current for Se deposition
does not drop down to zero after the surface is covered, that sug-
gests that it is not a surface limited process and the amount of
bulk Se deposited depends on the deposition time. The surface
limited process is fast, reaching completion quickly, while the
bulk deposition is slow, resulting in a small steady state current
[40]. Thus, one explanation for the excess deposit thicknesses
is the formation of some bulk Se on the surface. Deposition of
bulk Se can be minimized by using a short deposition time, just
long enough for the majority of the surface limited deposition to
complete. Alternatively, bulk Se can be removed by introduction
of an extra step designed to reduce excess bulk Se to a soluble
selenide species [27]. This second methodology was selected
for the present study. After 15 s of Se deposition, the cell was
rinsed with blank solution for 3 s, at which point the potential
was shifted negatively, such that bulk Se was reduced to HSe−.
HSe− is a soluble species, which diffuses away, leaving only
upd Se on the surface. A set of experiments were performed by
varying this reductive stripping potential for Se from −0.55 to
−0.65 V, whilekeepingalltheotherparametersconstant(Fig. 5).
Based on the resulting deposit stoichiometry data from EPMA,
and optical microscopy observations of the deposit thickness
and homogeneity, −0.63 V was picked as the reductive strip-
ping potential for the cycle.
To optimize the deposition potential for Hg, another series of
experiments were conducted, where only the deposition poten-
tial for Hg (Fig. 6) was varied. The best Hg deposition poten-
tial appeared to be 0.48 V, based on stoichiometry and optical
microscopy. However, it is of note that the coulometry for Hg, at
this potential, suggested a coverage of only 0.10 ML1, instead of
the 0.44 ML ideally expected. Previous results have suggested
that the formation of one bi-layer of Cd and Se from the (1 1 1)
plane of zinc blende CdSe would require 0.44 ML of Cd and
0.44 ML of Se. It is not clear why the coulometry for Hg was so
1
Coverages was used in this paper are relative to the number of Au surface
Fig. 4. Current–time profile during one cycle of HgSe deposition using ideal
deposition program.
atoms. Hence, 1 ML here would refer to one deposited metal atom for each Au
surface atom.

4350
V. Venkatasamy et al. / Electrochimica Acta 51 (2006) 4347–4351
Fig. 6. Effect of Hg deposition potential on HgSe deposition (Se deposition
potential: −0.15 V; Se stripping potential: −0.63 V).
Fig. 8. XRD diffraction pattern of 100 cycle HgSe thin film.
low, 0.10 ML versus 0.44 ML. One possibility is that some of the
deposited Se was oxidized at the positive potentials required for
surface limited Hg deposition. The coulometry observed was
thus the net charge for Hg2+ reduction and Se oxidation. In
addition, the resulting films were about half as thick as ide-
ally expected, where deposition of a compound monolayer is
expected with each cycle, see below.
for 2 s at −0.15 V. The solution was held static for 15 s for depo-
sition. The cell was then flushed with blank solution for 3 s at
−0.15 V, at which point, the potential was changed to −0.63 V
for 5 s. After this, the Hg solution was filled for 2 s, and deposited
for 15 s at 0.48 V. This was followed by another blank rinse at
0.48 V for 3 s (Fig. 7).
Overall, the optimal EC-ALE cycle for HgSe deposition
appeared to be as follows: Se solution was rinsed into the cell
Fig. 7. Optimal deposition program.
Fig. 9. Absorption spectrum of 100 cycles HgSe thin film.

V. Venkatasamy et al. / Electrochimica Acta 51 (2006) 4347–4351
4351
Ellipsometric measurements of the resulting deposit indi-
cated that the film was 19 nm thick. EPMA of the deposit
indicated a Se/Hg atomic ratio of 1.08. Fig. 8 shows the X-
ray diffraction pattern for the deposit. Peaks corresponding to
(1 1 1), (2 2 0) and (3 1 1) planes of HgSe (JCPDS 8-469) were
evident, and no elemental peaks for Hg and Se were observed.
The deposit showed a predominant (1 1 1) orientation as the ratio
of the intensity between the (1 1 1) and (2 2 0) peak came out to
be 10.2, which is considerably higher than the literature value
of 2.
[9] Y. Li, Y. Ding, H. Liao, Y. Qian, J. Phys. Chem. Solids 60 (1999)
965.
[10] C. Reig, Y.S. Paranchych, V. Munoz-Sanjose, Cryst. Growth Design 2
(2002) 91.
[11] S.M. Cox, M.K. Mathe, V. Venkatasamy, U. Happek, J.L. Stickney, J.
Electrochem. Soc. 152 (2005) C751.
[12] S. Bedair, Atomic Layer Epitaxy, Elsevier, Amsterdam, 1993.
[13] T.F. Kuech, P.D. Dapkus, Y. Aoyagi, Atomic Layer Growth and Pro-
cessing, Materials Research Society, Pittsburgh, 1991.
[14] H.L. Colin, Goodman, M.V. Pessa, J. Appl. Phys. 60 (1986) R65.
[15] W. Faschinger, Phys. Scr. T49B (1993) 492.
[16] J.L. Stickney, Electroanalytical Chemistry, vol. 21, Marcel Dekker, New
York, 1999, p. 75.
[17] B.W. Gregory, J.L. Stickney, J. Electroanal. Chem. 300 (1991) 543.
[18] M. Leskela, M. Ritala, Thin Solid Films 409 (2002) 138.
[19] E.B. Yousfi, B. Weinberger, F. Donsanti, P. Cowache, D. Lincot, Thin
Solid Films 387 (2001) 29.
Room temperature IR absorption studies of HgSe were per-
formed using a glancing angle of 85◦ from the surface normal.
Fig. 9 shows a plot of the square of the absorption data versus
energy for a 100-cycle deposit of HgSe. An absorption edge was
VB
CB
found at −0.60 eV. This value corresponds to the Γ 6 → Γ 8
[20] V. Sammelselg, A. Rosental, A. Tarre, L. Niinisto, K. Heiskanen, K.
Ilmonen, L.-S. Johansson, T. Uustare, Appl. Surf. Sci. 134 (1998) 78.
[21] M. Ylilammi, Thin Solid Films 279 (1996) 124.
[22] D.M. Kolb, M. Przasnys, H. Gerische, J. Electroanal. Chem. 54 (1974)
25.
transition of the inverted band structure of HgSe, reported to
be −0.51 eV from theoretical calculations [41] and −0.46 eV
experimentally [4].
[23] D.M. Kolb, Advances in Electrochemistry and Electrochemical Engi-
neering, vol. 11, John Wiley, New York, 1978, p. 125.
[24] K. Juttner, W.J. Lorenz, Z. Phys. Chem. N. F. 122 (1980) 163.
[25] A.T. Hubbard, V.K.F. Chia, D.G. Frank, J.Y. Katekaru, S.D. Rosasco,
G.N. Salaita, B.C. Schardt, D. Song, M.P. Soriaga, New Dimensions in
Chemical Analysis, Texas A&M University Press, College Station, TX,
1985, p. 135.
[26] A.A. Gewirth, B.K. Niece, Chem. Rev. 97 (1997) 1129.
[27] L.P. Colletti, B.H. Flowers Jr., J.L. Stickney, J. Electrochem. Soc. 145
(1998) 1442.
[28] T.L. Wade, B.H. Flowers Jr., U. Happek, J.L. Stickney, Presented at
the National Meeting of the Electrochemical Society, Spring, Seattle,
Washington, 1999.
[29] T.L. Wade, B.H. Flowers Jr., R. Vaidyanathan, M.K. Mathe, C.B. Mad-
dox, U. Happek, J.L. Stickney, Presented at the Materials Research
Society, 2000.
4. Conclusion
The influence of the deposition potentials for Hg and Se, as
well as that for a reductive Se stripping step, has been reported.
The optimal deposition cycle devised includes deposition of Se
at −0.15 V, stripping of excess Se at −0.63 V, and deposition
of Hg at 0.48 V. The resulting deposit was a little over half of
that expected from the ideal model of one compound monolayer
for each cycle, but the deposit was stoichometric, and showed
strong preferential (1 1 1) deposition. The absorption spectrum
for this deposit appears consistent with the literature: an inverted
band structure and a negative gap of 0.6 eV.
Acknowledgements
[30] J.L. Stickney, K. Varazo, L.C. Ward, M.D. Lay, T. Sorenson, Encyclo-
pedia of Surface and Colloid Science, Marcel Dekker Inc., New York,
2002.
[31] T.L. Wade, L.C. Ward, C.B. Maddox, U. Happek, J.L. Stickney, Elec-
trochem. Solid State Lett. 2 (1999) 616.
The authors would like to acknowledge the support of NSF
divisions of Chemistry and Material science.
[32] T.L. Wade, R. Vaidyanathan, U. Happek, J.L. Stickney, J. Electroanal.
Chem. 500 (2001) 322.
References
[33] R. Vaidyanathan, J.L. Stickney, U. Happek, Electrochim. Acta 49 (2004)
1321.
[34] J.Y. Yang, W. Zhu, T.J. Zhang, Electrochim. Acta 50 (2005) 4041.
[35] T.L. Wade, T. Sorenson, J.L. Stickney, Interfacial Electrochemistry, Mar-
cel Dekker, New York, 1999, p. 757.
[1] T.C. Harman, A.J. Strauss, J. Appl. Phys. 32 (1961) 2265.
[2] P.W. Kurse, M.D. Blue, J. Phys. Chem. Solids 23 (1963) 577.
[3] K.U. Gawlik, L. Kipp, M. Skibowski, N. Orlowski, R. Manzke, Phys.
Rev. Lett. 78 (1997) 3165.
[4] S. Einfeldt, F. Goschenhofer, C.R. Becker, G. Landwehr, Phys. Rev. B
51 (1995) 4915.
[36] T.L. Wade, B.H. Flowers Jr., K. Varazo, M. Lay, U. Happek, J.L.
Stickney, Presented at the Electrochemical Society National Meeting,
Washington, D.C., 2001.
[37] B.H. Flowers Jr., T.L. Wade, J.W. Garvey, M. Lay, U. Happek, J.L.
Stickney, J. Electroanal. Chem. 524 (2002) 273.
[38] B.M. Huang, T.E. Lister, J.L. Stickney, Surf. Sci. 392 (1997) 27.
[39] N.N. Greenwood, A. Earnshaw, Chemistry of the Elements, Pergamon
Press, Oxford, 1984.
[40] T.E. Lister, J.L. Stickney, Appl. Surf. Sci. 107 (1996) 153.
[41] M. Rohlfing, S.G. Louie, Phys. Rev. B 57 (1998) R9392.
[5] S. Einfeldt, H. Heinke, M. Behringer, C.R. Becker, E. Kurtz, D. Hom-
mel, G. Landwehr, J. Cryst. Growth 138 (1994) 471.
[6] C.R. Becker, L. He, S. Einfeldt, Y.S. Wu, G. Lerondel, H. Heinke,
S. Oehling, R.N. Bicknell-Tassius, G. Landwehr, J. Cryst. Growth 127
(1993) 331.
[7] P.P. Hankare, V.M. Bhuse, K.M. Garadkar, A.D. Jadhav, Mater. Chem.
Phys. 71 (2001) 53.
[8] P. Pramanik, S. Bhattacharya, Mater. Res. Bull. 24 (1989) 945.