K. E.
Hnida
*a,
S.
Bäßler
b,
J.
Mech
c,
K.
Szaciłowski
a,
R. P.
Socha
d,
M.
Gajewska
a,
K.
Nielsch
be,
M.
Przybylski
af and
G. D.
Sulka
g
aAGH University of Science and Technology, Academic Centre for Materials and Nanotechnology, al. A. Mickiewicza 30, 30-059 Krakow, Poland. E-mail: khnida@agh.edu.pl; katarzyna.hnida@gmail.com; Tel: +48 12 617 52 82
bUniversity of Hamburg, Multifunctional Nanostructures, Institute of Nanostructure and Solid State Physics, Jungiusstrasse 11, 20-355 Hamburg, Germany
cAGH University of Science and Technology, Faculty of Non-Ferrous Metals, Al. A. Mickiewicza 30, 30-059 Krakow, Poland
dJerzy Haber Institute of Catalysis and Surface Chemistry Polish Academy of Sciences, Niezapominajek 8, 30-239 Krakow, Poland
eLeibniz Institute for Solid State and Materials Research Dresden, PO Box 270116, 01171 Dresden, Germany
fAGH University of Science and Technology, Faculty of Physics and Applied Computer Science, al. A. Mickiewicza 30, 30-059 Krakow, Poland
gJagiellonian University in Krakow, Department of Physical Chemistry and Electrochemistry, Ingardena 3, 30-060 Krakow, Poland
First published on 13th January 2016
We present an electrochemical route to prepare nanocrystalline InSb thin films that can be transferred to an industrial scale. The morphology, composition, and crystallinity of the prepared uniform and compact thin films with a surface area of around 1 cm2 were investigated. The essential electrical characteristics such as conductivity, Seebeck coefficient, the type, concentration and mobility of charge carriers have been examined and compared with InSb nanowires obtained in the same system for electrochemical deposition (fixed pulse sequence, temperature, electrolyte composition, and system geometry). Moreover, obtained thin films show much higher band gap energy (0.53 eV) compared to the bulk material (0.17 eV) and InSb nanowires (0.195 eV).
In this paper, we present the straightforward and applicable on an industrial scale electrochemical deposition of nanocrystalline InSb thin films showing widened band gap in comparison with the previously reported layers and bulk material. The characterization of the film morphology, composition, crystallinity and electric properties at room temperature was conducted. The proposed method of synthesis and characterization of the electrical properties of electrodeposited InSb thin films has not been reported so far.
The electrodeposition of indium and antimony was performed using 1.5 × 1.5 cm electropolished (H3PO4 + ethanol, 5:2 in volume, 80 mA cm−2) and chemically etched (0.5 M H2SO4) Cu substrates. Before electrodeposition, an Au layer (20 nm) was sputtered on Cu. All electrodepositions were performed at room temperature in a conventional three-electrode cell (100 cm3 in volume) with a platinum counter electrode and an Ag/AgCl electrode as a reference electrode. The potentiostatic pulse electrodeposition of binary thin films was carried out from a citrate bath (0.2 M citric acid, 0.15 M sodium citrate) containing 0.06 M In3+ and 0.045 M Sb3+. Using a Reference 3000 potentiostat (Gamry Instruments), the InSb synthesis was performed at 20 °C at Eon = −2.3 V and Eoff = −0.5 V vs. Ag/AgCl. The pulse duration and delay time were 1 ms and 5 ms, respectively.
To achieve a compact structure of thin films, during codeposition of In3+ and Sb3+ ions, a pulse electrodeposition technique with a pulse sequence shown in the inset of Fig. 1A was used. During the Eon pulse (ton = 1 ms) the reduction processes described elsewhere occurred.16 At the ‘off’ time (toff = 5 ms), there was almost no electrodeposition observed. What is more, a partial dissolution of the top film layer, which could have a non-stoichiometric composition, occurred. A potential value of the off pulse is slightly different from the open circuit potential (OCP) of the system, however using a fixed pulse sequence allows to achieve reproducible nanocrystalline layers with a stoichiometric composition (tested over 20 samples). The dependence between deposition time and film thickness is shown in Fig. 1A. A growth rate of films obtained by pulse deposition was ca. 29 nm min−1. The average thickness of InSb films which were used to determine electrical properties was 1.5 μm.
The structural and morphological characterization studies of binary thin films were performed using a field-emission scanning electron microscope (FE-SEM/EDS, Hitachi S-4700 with a Noran System 7). The crystallinity of the obtained InSb films was examined using a Rigaku Mini FlexII diffractometer. For the analysis of diffraction patterns, the ICDD (International Centre for Diffraction Data) database, PDF Card No. 01-089-3667, was used. A thin InSb foil for transmission electron microscopy imaging was cut from a thin film using focus ion beam (FIB) FEI Quanta 3D 200i. Transmission electron microscopy (TEM) investigations were carried out using a FEI Tecnai TF20 X-TWIN (FEG) microscope equipped with an EDAX system, at an accelerating voltage of 200 kV. The film thickness and the crystal diameter were measured by using a profilometer (DektakXT Bruker). The atomic composition and electronic states of the elements were analyzed by X-ray photoelectron spectroscopy (XPS). The XPS spectrometer was equipped with a hemispherical analyzer SES R4000 (Gammadata Scienta) and an Al Kα X-ray source (1486.6 eV). The maximum energy resolution of the spectrometer for the line Ag 3d5/2 was 1.0 eV (for an analyzer pass energy of 100 eV). The electron binding energy (BE) scale was calibrated at the Fermi edge. The curves were fitted with Voigt profiles (GL = 30) and a Shirley background using Casa XPS software 2.3.12. The reflectance spectra of prepared thin films were recorded using a FTIR THERMO/Nicolet 5700 spectrophotometer, excitonic peak was fitted with GRAMS 32 software.
Sheet resistivity measurements of prepared thin films (with defined geometry) were performed at least ten times on each sample using a non-contact sheet resistance tester (EDDyCus® TF lab 2020, Suragus Sensors & Instruments). The Seebeck coefficient of the prepared samples was measured using a Potential-Seebeck Microprobe (PSM). Cross plane temperature gradient (ΔT) was applied and thermovoltage U (U = −SΔT, where S – Seebeck coefficient) was measured. The Seebeck coefficient was measured with an uncertainty of <3%.
The XRD analyses allow a non-destructive study of the compositional variation of InSb thin films (Fig. 2A). To prevent surface oxidation of stoichiometric InSb, a 30 nm carbon layer was sputtered. The three reflections at 23.5, 39.4 and 46.7° (inset in Fig. 2A) are typical for indium antimonide ((111), (220) and (311)) and correspond to the regular structure of zinc blend. However, the intensity of peaks corresponding to the InSb phase is low. For this reason, additional selected area electron diffraction (SEAD) analysis was carried out. Fig. 2B shows the bright field TEM image with a marked area where the SAED analysis (inset) was performed. Fig. 2C shows a representative high resolution TEM image confirming the polycrystalline nature of the obtained thin film. The estimated crystallite size (d) was within the range of 20–30 nm.
Fig. 2 XRD diffraction pattern (A) and bright field TEM-SAED analysis (B) together with (C) high resolution TEM image for as-prepared InSb thin films with stoichiometric composition. |
For practical applications of nanocrystalline films, it is crucial for the material to contain a stoichiometric amount of appropriate atoms suitable for forming the desired compound but not an alloy. The surface states of the prepared InSb thin film were analyzed by XPS (Fig. 3). A typical In 3d core level spectrum (Fig. 3A) exhibits the In–Sb bonds identified at BE of 444.3 eV and 451.9 eV. These data are in agreement with the previously reported binding energies of In 3d doublet components at 444.2 and 451.8 eV for the bulk InSb.17 In InSb thin films we observe a slight shift (0.1 eV) towards higher energies. No other electronic states of indium are found by XPS. The Sb 3d and O 1s peaks are overlapped, however can be deconvoluted into two O 1s components and two doublet Sb 3d ones. The maximum of Sb 3d5/2 core excitation (component A) is found at 527.2 eV (46.8% of the Sb 3d spectrum intensity) that refers to an InSb compound.18 The other doublet component (B and B’) in the Sb 3d core level spectrum corresponds to the oxidized Sb3+ form of antimony (53.2% of intensity) – Fig. 3B. The XPS analyses show that the InSb film is partially oxidized. The calculated analytic depth of the XPS method is 7.3 nm for InSb and 5.8 nm for Sb2O3.19 Therefore, taking into account the intensity of both Sb 3d spectrum components, we can estimate the thickness of oxidized antimony species at the InSb layer to be lower than 4 nm. The In/Sb ratio in the InSb phase obtained from In 3d and Sb 3d excitation intensity ratios by subtracting the antimony oxide contribution is 1.03 that confirms the stoichiometry of the InSb layer. An antimony oxide layer with a few nanometer thickness on the surface of the sample is too thin to affect the optical (Sb2O3 absorption range of 325–500 nm)20 and electric properties of thin films. However, in order to avoid negative effects of the dielectric layer to the electric conductivity and to the Seebeck coefficient, the InSb films were soaked in 0.1 M NaOH (to remove the surface layer of Sb2O3) before each electrical measurement.
Fig. 3 In 3d (A) and Sb 3d (B) XPS high resolution spectra of the InSb thin film obtained via electrodeposition. |
Seebeck coefficient (S) (Table 1) is found to be negative indicating that majority charge carriers are electrons (in freshly prepared InSb films). Nagata and Yamaguchi reported negative values of S for stoichiometric Si-doped InSb thin films formed by MOCVD.21 Depending on the degree of Si-doping, S was reported to vary from −217 μV K−1 to −178 μV K−1 for the wholly doped InSb thin film and Si-doped near the surface of InSb, respectively. The undoped indium antimonide thin film showed S at around −160 μV K−1. Our films exhibit a S value of −380 μV K−1, more than twice as high in comparison with the InSb obtained by MOCVD. Such a highly negative S value in the electrochemically prepared thin film is caused probably by considerable polycrystallinity, the presence of defects and vacancies formed during the synthesis.
Nanocrystalline InSb thin film (1.5 μm in thickness) | n-InSb nanowire35 (67 nm in diameter, 8.5 μm in length) | |
---|---|---|
E g [eV] | 0.53 | 0.195 |
n [cm−3] | 1.68 × 1017 | 4.74 × 1019 |
σ [Ω−1 m−1] | 206 | 17.4 |
μ [cm2 V−1 s−1] | 76.5 | 4.1 |
Majority charge type | Electrons | Electrons |
S [μV K−1] | −380 | — |
P f [W m−1 K−2] | 2.97 × 10−5 | — |
To determine the band gap of the obtained InSb films, FTIR analysis was performed. The recorded diffuse reflectance spectrum i.e. intensity vs. energy is shown in Fig. 4 as an Intensity = f(energy) function. The spectrum shows three characteristic features: an absorption onset at ca. 0.51 eV, a sharp peak at ca. 0.56 eV and a plateau at the high energy region. The 0.56 eV peak is attributed to the excitonic transition.22,23 A pure absorption spectrum was obtained by numerical subtraction of the excitonic Gaussian peak from the experimental spectrum. The resulting differential spectrum was used to create the Tauc plot (Fig. 4B). Due to the fact that InSb is a direct band gap semiconductor, the (KM·E)2 = f(E) relationship is plotted. The optical band gap of the value 0.53 eV was obtained from the linear fit of the low energy portion of the spectrum. The resulting value of Eg is much larger than that for bulk indium antimonide (0.17 eV at 300 K).13 There are two different factors that can be responsible for such behaviour: quantum confinement24 and Burstein–Moss shift.25 In order to check the possible role of the quantum confinement, the size of crystallites was estimated from the high resolution TEM images. The average value of d = 25 nm was found. On the other hand the Bohr radius of an exciton in InSb can be calculated from (1):
(1) |
(2) |
(3) |
(4) |
Assuming ΔEBM to be equal to 0.19 eV, on the basis of eqn (4), one can estimate the concentration of electrons n in the InSb thin film. The obtained n value of 1.68 × 1017 cm−3 is one order of magnitude higher than for bulk InSb (2 × 1016 cm−3). Therefore, it may be assumed that the band gap widening is associated with the size of crystallites and an increase in concentration of charge carriers via doping. Moreover, quantum confinement and the Burstein–Moss effect have rather a similar magnitude of 0.17 and 0.19 eV, respectively.
In order to learn more about the electrical properties of electrochemically deposited thin films, sheet resistance measurements were performed. The resulting conductivity values, as well as other electrical parameters for a freshly prepared InSb thin film measured at 294 K are presented in Table 1.
Electrical characteristics of n-InSb thin films synthesized via magnetron sputtering, electron beam evaporation, metalorganic vapour phase epitaxy and many other methods have been reported in the literature. Carrier concentrations of one order of magnitude smaller than our value (Table 1) were reported for InSb epitaxial layers grown by MBE29 (1.6 × 1016 cm−3), MOCVD30 (2.48 × 1016 cm−3) and for Pb-doped InSb thin films31 (1.6 × 1016 cm−3). It was also shown that higher n values can be obtained by doping InSb with Sn32 and Si33 atoms or by using radio frequency magnetron sputtering.34 Knowing the n determined from eqn (4) (under the assumption that only quantum confinement and the Burstein–Moss effect are involved in widening of the energy band gap) and the measured conductivity σ we estimated charges mobility μ (μ = σ/ne). The obtained value (76.5 cm2 V−1 s−1) is lower than the value determined for polycrystalline n-InSb films,34 monocrystalline layers8,30 and bulk materials. This fact can be explained by significant contribution of increased electron scattering at the grain boundaries.
For the electrochemically synthesized InSb thin film, the RT value of thermoelectric power factor (Pf = S2σ) was calculated (Table 1). The power factor is an important parameter often used to express the energy conversion efficiency. The calculated power factor for the as-deposited InSb film, with a stoichiometric composition, equals to 2.97 × 10−5 W m−1 K−2. Ishii et al.30 showed that a thin monocrystalline InSb film can reach a power factor of about 4.56 × 10−5 W m−1 K−2. This value is in good agreement with our results.
The influence of reducing the dimensionality from a film to a nanowire on the electric properties is shown in Table 1. It can be seen that the nanowires show a two orders of magnitude higher concentration of charge carriers than our thin films. This can be attributed to a significantly higher amount of defects, which is the characteristic feature for the nanowires obtained by electrochemical deposition. Due to the fact that nanowires are composed of crystals with a small diameter (a few nanometers), increased scattering at the grain boundaries explains the strong reduction of charge carrier mobility (μ) in comparison with a thin film obtained via the same synthesis method. Due to the non-directional growth of thin films (in comparison with nanowires grown inside porous anodic alumina templates), larger crystal grains and less number of vacancies/defects are observed, which result in higher electrical conductivity (less number of interfaces and grain boundaries and less scattering of charge carriers are observed). In summary, reducing the dimensionality of the material while maintaining the same synthesis conditions lead to lower conductivity and electron mobility and increase the concentration of charge carriers as a result of increased defects concentration and reduced grain sizes.
This journal is © The Royal Society of Chemistry 2016 |