The low resistive and transparent Al-doped SnO
2
films: p-type
conductivity, nanostructures and photoluminescence
C.E. Benouis
a, M. Benhaliliba
a,⇑, Z. Mouffak
b, A. Avila-Garcia
c, A. Tiburcio-Silver
d, M. Ortega Lopez
e,
R. Romano Trujillo
e, Y.S. Ocak
fa
Department of Material Technology, Physics Faculty, USTOMB University, BP1505 Oran, Algeria
bDepartment of Electrical and Computer Engineering California State University, Fresno, CA, USA c
Cinvestav-IPN, Dept. Ingeniería Eléctrica-SEES, Apdo. Postal 14-740, 07000 México, D.F., Mexico
d
ITT-DIE, Apdo, Postal 20, Metepec 3, 52176 Estado de Mexico, Mexico
e
Centro de Investigación en Dispositivos Semiconductores, Instituto de Ciencias-BUAP, 14 Sur y Av. San Claudio, C.U. Puebla, Pue., Mexico
f
Dicle University, Education Faculty, Science Department, 21280 Diyarbakir, Turkey
a r t i c l e
i n f o
Article history:
Received 7 December 2013
Received in revised form 2 March 2014 Accepted 6 March 2014
Available online 18 March 2014 Keywords:
Sprayed Al:SnO2films
ITO coated glass substrate AFM
Photoluminescence Nanostructures n/p-Type conductivity
a b s t r a c t
In this work, we study the crystalline structure, surface morphology, transmittance, optical bandgap and n/ p type inversion of tin oxide (SnO2). The Nanostructured films of Al-doped SnO2were successfully produced
onto ITO-coated glass substrates via the spray pyrolysis method at a deposition temperature of 300 °C. A (1 0 1) and (2 1 1)-oriented tetragonal crystal structure was confirmed by X-ray patterns; and grain sizes varied within the range 842 nm. The films were polycrystalline, showing a high transparency in the visible (VIS) and infrared (IR) spectra. The optical bandgap was estimated to be around 3.4 eV. The atomic force microscopy (AFM) analysis showed the nanostructures consisting of nanotips, nanopatches, nanopits and nanobubbles. The samples exhibited high conductivity that ranged from 0.55 to 104(S/cm) at ambient
and showed an inversion from n to p-type as well as a degenerate semiconductor characters with a bulk concentration reaching 1.7 x 1019cm3. The photoluminescence measurements reveal the detection of
violet, green and yellow emissions.
Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction
For the last three decades, tin oxide (SnO2) has become one of
the most studied semiconductor materials. Tin oxide is a II–VI semi-conductor with a wide band gap (Eg= 3.5 eV) and it belongs to a
variety of transparent conductive oxides known as TCOs[1,2]. Re-cently, researchers have claimed that SnO2could exhibit a p-type
behavior when doped with metal elements such as aluminium
[3], indium[4,5], antimony[6], zinc[7]and iron[8]. Spray pyrolysis deposition (SPD) [9], polymer-pyrolysis deposition method [6], sputtering[10], hydrothermal method[11], colloidal solution route
[12], are all deposition techniques that have been used to synthe-size SnO2films onto substrates. In the literature various
applica-tions of SnO2have been reported like light emitting diodes[13],
and gas sensors[12–14]. It should be noted that the efficient doping effects are obtained when the ionic radius of the dopant is the same as or smaller than of the host ion (Sn4+ (r = 0.71 Å) and Al3+
(r = 0.51 Å)) . To our knowledge, there has been no work reported on the structure, surface, Hall measurement, p-type conductivity
and optical properties of Al-doped SnO2 thin films grown onto
ITO coated glass substrate by spray pyrolysis process at fixed sub-strate temperature. As indicated in prior reports, important param-eters for the preparation of a p-type TCO are: deposition technique, type of crystalline lattice, atomic size, doping level, valence and electronic configuration of dopants. A lower valency cation as acceptor impurity such as Al3+in tin oxide decreases n-type
con-ductivity and increases the hole concentration and hence the p-con-ductivity. However, it is to be noted that in a successful acceptor doping process, besides doping level, the atomic or cationic size of the acceptor dopant is very important. The current investigation was done on the fully transparent conductive sprayed Al:SnO2/ITO
nanostructured thin films. The n-type SnO2inverted to p-type
con-ductivity prepared by SPD was evidenced. In this report, we used Al3+ions as a preferred acceptor dopant to investigate effect of Al
doping level on the electrical properties of the SnO2thin films. In
or-der to investigate the electrical properties and the defects of Al:SnO2layers deposited onto ITO substrate, we have achieved such
films. The as-synthesized Al:SnO2/ITO films will be used in the
microelectronic device applications. Intrinsic oxides don’t exhibit high electrical conductivity and optical transparency. To overcome this inability, it is to create electron degeneracy in a wide band gap
http://dx.doi.org/10.1016/j.jallcom.2014.03.046
0925-8388/Ó 2014 Elsevier B.V. All rights reserved.
⇑ Corresponding author. Tel.: +213 772211491; fax: +213 41429212. E-mail address:mbenhaliliba@gmail.com(M. Benhaliliba).
Contents lists available atScienceDirect
Journal of Alloys and Compounds
oxide by controlable non-stoichiometry and suitable doping. Our aim is how to produce p-type conductivity from n-type doped tin oxide by a facile deposition route and to fabricate later a p-n junc-tion which will be used in microptoelectronic devices. To strength-en the knowledge in conduction type conversion mechanism in such oxides layers.
2. Experimental procedures
The film fabrication was carried out by a homemade SPD technique. Firstly, pure and Al-doped SnO2layers were sprayed on ITO substrates. The precursor was
(SnCl45H2O) and the doping material was aluminum (III) chloride (AlCl3) which
were dissolved in methanol. The starting material concentration was 0.2 M and the concentrations of the dopants Al/Sn in the solution were 0.5, 2, 3 and 8.5%. Spray rate and substrate to nozzle distance were maintained respectively at 20 ml/min and 25 cm. The glass substrate was heated at 300 °C which was con-trolled by a digital thermometer connected to the heater. Finally, the film thick-nesses ranged from 230 to 300 nm. Al-doped SnO2films were sprayed onto a
rectangular (ITO)-coated glass slide. This substrate has a sheet resistance of 1525X/Sq supplied by Sigma–Aldrich. The concentrations of the dopants Al/Sn in the solution were 0.5%, 2%, 3% and 8.5%. The structural properties of our samples were carried out by a Rigaku X-ray diffractometer, model DMAX 2200 with a copper anticathode (Cu Ka, k = 1.54 ÅA0), with an angle range (2h) of 30–70°. The optical parameters of the as-synthesized Al:SnO2/ITO films were measured using a
Shima-dzu UV-3101PC double beam spectrophotometer. The electrical resistivity, mobility and bulk concentration of charge carriers were measured using a Hall Effect mea-surement system (HMS) ECOPIA-3000 at room temperature. DC current was applied through the sample; a magnetic field, of 0.58 T, was applied perpendicular to
sample. VMQand VNPare the applied voltages. The samples’ surface morphology
was analyzed by atomic force microscope (AFM) analysis using a Quesant Model 250 system having an (80lm 80lm) head, in the wave mode in air. For the (3lm 3lm) square images the resolution was (300 300) pixels at a fixed scan rate of 2 Hz. All analyses were performed with the the WSXM system software. The room temperature photoluminescence (PL) analysis was carried out in an experi-mental setup consisting of a 325 nm, 15 mW He–Cd laser (Kimmon type), a 0.85 m double monochromator (SPEX, model 1404), and a GaAs photon counting photomultiplier (Hamamatsu). The band explored was from 350 to 600 nm, in steps of 0.5 nm and at a speed of 0.2 s per measured point.
3. Results and discussions
The studied samples are: pure (0%), 0.5% (sample A), 2% (sample B), 3% (sample C) and 8.5% (sample D) Al-doped tin oxide films respectively.
3.1. X-rays studies
Fig. 1A depicts X-ray patterns of as-sprayed Al:SnO2/ITO at
300 °C as a function of Al content. The spectra confirm the presence of tetragonal crystal structure in most samples and a preferred direction along the (1 0 1) and (2 1 1) planes. It is shown that the Al:SnO2films, analyzed by X-rays in the 2h-range of 20–80°, are
polycrystalline and exhibit several significant reflections. The main
20 30 40 50 60 70 80 0 100 200 300 400 500
*
*
(3 0 1 ) (3 1 0 ) (2 2 0 ) (2 1 1 ) (2 0 0 ) (1 0 1 ) In ten s it y (a. u .) 2-Theta (Deg.) (1 1 0 )O
A B C D*
ITO O OrthoO
Fig. 1A. X-ray patterns of sprayed Al:SnO2films produced at 300 °C (tetragonal phase) (A: 0.5%, B: 2%, C: 3%, D: 8.5% Al doped SnO2), range of angle 2h° = 20–80°,
orthorhombic phase was displayed by letter O and ITO peaks were signed by star (
). The inset showed the angle shift and the discrepancy in intensity of main peaks, (black, red, magenta, blue curve correspond respectively to sample A, B, C and D). The lines of JCPDS card 41-1445 of SnO2are displayed. The X-ray patterns of pure SnO2is displayed
at right. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 1
(h k l) Planes indexation, Bragg angle (2h°) and angle shift (D2h), of sprayed Al:SnO2grown onto ITO glass substrate at 300 °C (sample A:0.5%, B:2%, C:3% and D:8.5%), JCPDS card of
tetragonal SnO2phase.
(h k l) Samples JCPDS 2-Theta (Deg.)
A B C D 2h° (D2hÞ 2h° (D 2hÞ 2h° (D 2hÞ 2h° (D 2hÞ (1 1 0) 26.1 0.4 26.7 0.2 – – – – 26.5 (1 0 1) 35.5 0.1 33.9 0.6 – – – – 34.5 (2 0 0) – – 37.8 0.4 – – – – 38.2 (2 1 1) 51.1 0.9 51.5 0.5 50.5 1.5 – – 52 (2 2 0) 55.9 0.9 54.7 0.3 54.9 0.1 54.9 0.1 55 (3 1 0) 60.5 1.5 60.7 1.3 – – – 62 (3 0 1) 64.6 0.4 65.2 0.2 63.4 1.6 63.4 1.6 65
planes for the 0.5%, 2% and 3% Al-doped tin oxide films were (1 1 0), (1 0 1), (2 0 0), (2 1 1), (2 2 0), (3 1 0) and (3 0 1) that were nearly peaked at 26.6°, 34.2°, 38°, 50.8°, 55°, 61° and 63° respectively. Moharrami has found the same configuration of reflections[15]. Positions of reflections and angle shift are gathered in Table 1. Orthorhombic phase of SnO2remains particularly in films doped
with 0.5% and 2% Al, and their corresponding reflections’ planes are (1 1 3) and (0 2 6) respectively peaked at 30.5° and 45.5° as signed by ‘‘O’’ inFig. 1A. Recent studies have shown the formation of the orthorhombic phase of SnO2 under a variety of synthesis
processes as reported by Komen et al.[15]. Overall, the phase of ITO was identified by two reflections positioned respectively at 21.6° and 40.5° (signed by star inFig. 1A), with significant peaks in the A, B and C cases[17–19]. The JCPDS card 41-1445 of SnO2
phase was displayed inset ofFig. 1A. Moreover, the X-rays spec-trum did not show the existence of any aluminum (Al2O3phase)
or/and indium (In2O3 phase) peaks, indicating that aluminum
was incorporated in the substitutional sites within the tin oxide lattice. In literature, the Al2O3phase was absent in the Al-doped
sample with an Al content less than 10%[15]. Based on our X-ray patterns, we see that sample D has a much smaller peak compared to samples A, B and C which suggests a poor crystalline structure, the intensity of reflections of sample D is nearly 10 times less than that of samples with a doping level less than 8.5%. A pure tin oxide layer shows a (2 0 0) peak, as a preferential orientation, which is lo-cated at 37.82° (seeFig. 1Aright). While this orientation is not seen in the samples except the sample B. Both the pure and the doped film (sample B) revealed a slight angle shift of the main orientation (D2h°~0.4°). The doping level has an influence on both peak posi-tions and intensities as sketched in the inset ofFig. 1A. When the Al doping increases in the SnO2 lattice, its crystalline structure
tends to be deteriorated and the intensity of main peaks decreases, the peak of ITO is apparent as observed in the case of sample C. Along the (1 1 0) direction: a high peak intensity is detected for the sample B, a discrepancy between the strong peak and the peak of sample C is clearly observed as indicated by an arrow in the inset ofFig. 1A. Furthermore, the difference is even greater than with peaks of samples A and D. Along the direction (1 0 1): an angle shift is detected between the samples A and B as seen in the inset of
Fig. 1A, while the peaks of samples C and D are completely absent. The lattice mismatch could cause this angle shift. Along the (2 1 1) plane: there is no big difference in position and intensity of peaks for the samples A et B but the both samples C and D showed slightly weaker peaks. The intensity corresponding to the main peaks gets enhanced as Al content in the precursor was lowered (63%). Moreover, the broadening of peaks decreased gradually with an increase in the Bragg angle (from 26.5°, 35° to 50°) which corresponds respectively to the (1 1 0), (1 0 1) and (2 1 1) planes as seen in the inset ofFig. 1A. The size of crystallites, calculated using
Scherrer’s formula as cited in our previous paper[20], ranged from 8 to 42 nm as indicated inFig. 1B. Similar result was reported by Chan who has found the grain size of SnO2to be within the range
1030 nm[21]. Our results confirm the nanostructure character of
-2 0 2 4 6 8 10 0 10 20 30 40 Grain size (nm) Al content (%) According to (211) plane
Fig. 1B. Grain size along (2 1 1) plane was plotted against Al content for the sample 0.5%, 2%, 3% and 8.5% Al doped SnO2films grown by spray-pyrolysis at 300 °C.
Fig. 2. AFM pictures (10lm 10lm) of sprayed Al:SnO2/ITO film produced at
300 °C (A: 0.5%, B: 2%, C: 3%, D: 8.5% Al doped SnO2). (Top) z range is displayed in
2D-view, histograms. Histogram of diameter and height distribution. Power PSD against frequency of sprayed Al:SnO2thin films (samples A, B, C and D). (Bottom)
3D-Images A, B, C and D displayed respectively nanotips, nanopatches, nanopits and nanobubbles.
the as-grown samples. The pure SnO2film prepared by spray
pyro-lysis technique presents a grain size of 20.98 nm according to (2 0 0). The plane (2 0 0) is the main and intense peak of the pure SnO2film while Al doping improve the (2 1 1) orientation as shown
in X rays spectra.
3.2. Al-doped SnO2layers morphology
To characterize the topography and confirm the nanostructure
property, we performed the AFM analysis on the Al-doped SnO2
films.Fig. 2depicts the 2D and 3D AFM views of sprayed Al:SnO2
films grown onto ITO corning glass substrate at a fixed substrate temperature of 300 °C. The films of 0.5% Al:SnO2revealed roughly
a homogenous surface consisting of crystallites. The crystallites
Fig. 2 (continued)
have grown from the inner towards the surface. The AFM observa-tion showed a large amount of grain agglomeraobserva-tion that looked like assembled nanotips (see 3D-AFM picture A) with few voids. A topography parameters such as a diameter (U), a perimeter (P) and a mean height (Z) of nanograins are summarized inTable 2.
The average size was evaluated at 65 nm as described in 2D view, and the root mean square roughness (RMS) of films was deter-mined as summarized inTable 3. In addition, the obtained nano-grains were concentrated in an unique cluster with no well boundaries. Similar morphology of SnO2nanostructures was found
in the literature[22]. It is confirmed that the Al doping level influ-enced the grain structures as sketched inFig. 2(A–D). Nanograin-like assembled tips were shaped in the 0.5% Al-doped tin oxide film. Overall, the surface of sample A was homogenous, and grains have different sizes, shaped like tips with an almost circular base and a very sharp peak. Moreover, nanotips are smooth (RMS 12.5 nm) compared to those obtained from the other samples as listed inTable 3, and grown along the z-axis. The nanotips grew side by side with a high surface density (number of nanotips/
l
m2)leaving almost no voids. These configurations are in well agree-ment with those obtained by Chacko et al.[22]. Bar graphs show the distribution of the diameter and height of the grain inFig. 2. The distribution of nanotips as a function of diameter was depicted in a histogram view (insetFig. 2A), and the power spectral distribu-tion (PSD) that described the percent of crystallites versus fre-quency (1/
l
m) was depicted. With an increase of Al content the nanograins tend to reach quasi-spherical shapes as seen in 3D-AFM images B, C and D. In contrast, the 3D-AFM view of the sample B exhibited different surface topographies, and the crystallites seem to be patches, with an average height of 677.3 nm, and a smaller amount of voids. The patches grew from the inner to the surface, and were rough as listed inTable 3(RMS > 200 nm), with a slight tilt as shown inFig. 2B, and no well-defined boundaries, this is in accord with previous results cited by Kaur et al. [23]. The surface of sample C was not homogeneous, and consisted of grains looking like nanopits, and having an average height of 180 nm. In previous studies, nanograins with size up to 100 nm were synthesized, it is mentioned that films break-up into large isolated islands[24]. The Rmsparameter is given by the followingexpression[25], Rms¼ Pn 1ðZi ZÞ 2 N ð1Þ
where Ziis the value of each point, Z is the average of the Z values
and N is the number of points. The grains were smooth (25.5 nm seeTable 3) with different sizes and shapes as shown inFig. 2C. The statistical distribution showed a difference in intensity of bars (%), the power spectrum density is PSD. The 3D-view (Fig. 2D) shows
nanobubbles with various diameters and morphologies. These nanobubbles grew along different orientations, contrary to the pre-ceding cases, as indicated inFig. 2D. The sample D has grown ran-domly which confirms the tendency to the disordered or/and amorphous character, confirmed by the X-ray analysis. Moreover the detailed AFM data confirmed the nanostructure presence in our samples which is in agreement with those given above by X-ray patterns. In order to further explore the topography of films’ surface, we determined the following parameters: Rms (nm), the
max peak height Sp(nm) and the mean summit curvature Ssc(1/
nm) (seeTable 3). We conclude that the crystallite size measured from the XRD data of Al:SnO2was smaller than the grain size
deter-mined by AFM analysis. This might be attributed to the fact that the AFM measurement is more susceptible to the surface topography while the X-ray patterns analysis is susceptible to the inner of the bulk. PSDof the loaded image is obtained from Fourier transform
(FT) and reflects the Rmsroughness of the sample surface. PSD, FT,
and Rmsroughness are related as follows,
PSD¼ R2ms¼ F 2
T ð2Þ
The power spectral density (PSD) of a surface is equal to the square
of its Fourier transform (FT) and to the Rmsroughness value squared.
PSDis one of many parameters that are used to represent the sample
surface roughness. PSD has advantages over the other surface
roughness factors such as the RMS parameter because it contains information of how each frequency components contributes to the total roughness of the surface. The pure SnO2sprayed films show
a roughness Rms of 10.41 nm , Al doping increases the roughness of films.
3.3. UV–VIS–IR studies
The pure tin oxide presents only one oscillation in visible range of 88.25 %, at 550 nm, followed by a slight decay and an increase in infrared spectrum from the wavelength of 1200 nm. In the visible band, the transmittance of sample A reached a maximum of 81% around a wavelength of 600 nm, followed by a decay of about 5% from 600 nm to the visible band edge. A second maximum of trans-mittance was obtained in the IR spectrum, and found to be 89% and the transmittance also decreased rapidly by 15% in medium and far-infrared band. Sample B exhibited another variation of trans-mittance. This latter increased rapidly in UV–VIS shaped two oscil-lations and highest transmittance was around 88%, and then decreased rapidly by 38% in the VIS–IR band. High transmittance
(90%) of SnO2 was obtained by Moure-Flores et al. [10]. We
obtained a similar increase of transmittance in UV–VIS for sample
Table 2
U: diameter, P: perimeter and Z mean height (The parameters are expressed in nm) of sprayed Al:SnO2grown onto ITO glass substrate at 300 °C (sample A:0.5%, B:2%, C:3% and
D:8.5%).
Result Sample A Simple B Simple C Sample D
U P Z U P Z U P Z U P Z
Min 65.36 267.56 19.68 65.36 267.56 264.86 65.36 267.56 30.50 75.47 344.45 432.66 Max 688.66 3946.49 38.39 1327.30 6488.29 480.02 1129.63 6488.92 85.20 1485.28 6889.63 664.54 Mean 181.15 826.23 22.46 402.43 1960.47 286.57 192.59 1039.43 36.84 348.96 1754.69 463.49
Table 3
Grain size by X-ray patterns, transmittance at wavelength of 550 nm, transmittance at wavelength of 550 nm, optical band gap, thickness of films and figure of merit, root mean square roughness RMS (nm) and peak height max Sp(nm), mean summit curvature Ssc(1/nm) of sprayed Al:SnO2grown onto ITO glass substrate at 300 °C (sample A:0.5%, B:2%,
C:3% and D:8.5%). Samples (Al %)
Grain size by XRD (nm) along (2 1 1) plane T (550 nm) (%) Eg (eV) Thick (nm) Figure of merit (103X1 ) RMS (nm) Sp (nm) Ssc(103 1/ nm) A:0.5 11.4 80.74 3.46 231.5 12 12.52 64.19 1.55 B:2 11.3 83.51 3.55 297.6 0.1 209.63 677.29 5.71 C:3 8.1 85.15 3.45 238.1 49 25.55 180.07 6.21 D:8.5 42.6 74.85 3.25 232.7 7 104 348.70 911.6 14.69 500 1000 1500 2000 2500 0 10 20 30 40 50 60 70 80 90 100 %0.5 % 2 % 3 % 8.5 Pure SnO2 Transmittance (%) Wavelength (nm)
Fig. 3A. The UV–VIS–IR transmittance dependence on the photon wavelength of sprayed Al:SnO2films produced at 300 °C. Black arrow showed the oscillations.
(pure, A: 0.5%, B: 2%, C: 3%, D: 8.5% Al doped SnO2.) (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 3B. Transmittance plotted versus wavelength in UV, VIS and IR ranges. Arrow showed the oscillations in UV range, arrows signed the maxima of transmittance in visible range, the highest transmittance and rapid decay in IR spectrum.
C, around 590 nm, with a maximum of 86%, while the decay occu-red in the IR range from 86% to 62%. A rapid decay in IR band was also reported by Fukano and Babar[26,27]. The transmittance of sample D obeys to different paths, the VIS and IR highest transmit-tance was lower (<80%) than those obtained for the previous cases. There was a 30% rapid decrease of transmittance in the IR range. We conclude that Al doping decreases the transmittance in visible band. Consequently, the decrease in electrical resistivity, studied in Section3.3, is a result of this decay. As the concentration of dopant increases sufficiently as the structure is deteriorated, so the film surface is not homogenous then the transmittance decreased. Fur-thermore, significant oscillations were detected in the VIS range as shown inFig. 3A. Similar trends are observed by Moharrami and Martínez[15,28]. Discernible interferences in UV-band are empha-sized for the sprayed 0.5%, 2% and 3% Al doped SnO2as shown,
indi-cated by arrows, in Fig. 3B. These interferences might be the characteristic of the nanostructures’ production. The transmittance of as-deposited Al:SnO2 films as a function of incident photon
wavelength was plotted inFig. 3A. Our films exhibited a high trans-parency up to 85% as indicated inFigs. 3A and 3B. The transmit-tance grew rapidly in UV–VIS range, then it decreased slowly in IR spectrum. The average transmittance was around 83% in visible and IR ranges. Ions of Al dopant locate in defects centers such as impurities which cause the scattering of incident photon which then lead to a transmittance decay. Our results were in well agree-ment with those found by other researchers[29]. The absorption coefficient
a
is calculated using Lambert’s law[30],a
¼ ln Tt ð3Þ
where T is the transmittance and t is the film’s thickness. The fol-lowing formula expressed the optical band gap[31],
ð
a
hm
Þ2¼ ðhm
EgÞ ð4Þwhere
a
is the absorption coefficient, h is Planck’s constant,m
is the photon frequency, and Egis the band gap energy. By extrapolation ofthe linear part of the curve, the intersection with energy axis deter-mined the band gap Eg(not shown here) as cited in Ref.[4]. The
optical band gap Egof our films was calculated with an accuracy
of 1% as listed inTable 3.Fig. 3Cdepicts the films’ thickness and optical bandgap (in double y-axis) as a function of Al content, and the highest value of Eg(3.55 eV) was found for sample B. Ikhmayies
et al. have found a similar range of Eg[32]. The discrepancy in band
gap is due to Burnstein–Moss (BM) effect, expressed in earlier re-port as, Eg= Eg0+ (h2n2/3/8
p
2/3meff), where Eg0= 3.2 eV is theintrin-sic band gap and (Eg–Eg0) represents the difference of band gap or
called BM effect[33]. It causes the variation of band gap due to dop-ing. The difference between measured, determined from absorption plot, and calculated band gap, determined from BM relation is dis-played inFig. 3D. The films’ thickness was obtained using the fol-lowing formula[34], t ¼ 1 2 1 k1 1 k2 n0 ð5Þ
where t is the film’s thickness, k1and k2are the wavelengths of two
consecutive maxima in the transmittance plot. n0 is the refractive
index of SnO2found to be equal to 2[34]. The obtained results
re-vealed that the films’ thickness ranged from 230 nm to 300 nm as sketched inFig. 3C.
3.4. Electrical measurements
The sample was held between four gold contacts M, N, P and Q as shown inFig. 4A. Hall measurements were performed in order to determine the carriers type, and any n to p conductivity inversion.
Fig. 4Bshows the charge carrier concentration, the resistivity and the mobility as a function of Al content. As Hall mobility (
l
), according to the relation 1/q
= nql
, is inversely proportional tore-sistivity (q
=1/r
, wherer
is the electrical conductivity), the mobil-ity value varies roughly in thereverse tendency as the resistivmobil-ity, as sketched inFig. 4B(I and II). The pure SnO2layer is n-typesemi-condcutor which exhibits a resistivity around 90X.cm. The electri-cal parameters were influenced by aluminum incorporation in tin oxide as shown in Figs. 4B–C. The charge carrier concentration was found to be high, around 1020cm3, the mobility ranged from
2 to 23 cm2/Vs and electrical resistivity of our sprayed films was found to be lower around 9.105
Xcm, the samples have p and
n-type conductivity as listed in Table 4. As previously reported, with inserting certain amounts of acceptor dopants to SnO2layers,
the electrical conductivity changes from n to p type. In a specific acceptor doping rate, the number of electrons becomes less than holes and p type conduction mechanism dominates. Babar et al. found a mobility of around 20 cm2/Vs, and a resistivity in the
vicin-ity of 104
Xcm, as well as a carrier concentration of 1020cm3
[27]. K. Ravichandran et al. reported that the nature of conductivity changes from n-type to p-type when the Al doping level is 10 at.%.
A bulk density of holes (p type) variying from 2.9 x 1018 to
Fig. 3C. Films’ thickness (left) and optical bandgap Eg(right) plots against Al
content of sprayed Al:SnO2film produced at 300 °C, red lines showed the error bars.
(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.).
0.0 5.0x1013 1.0x1014 1.5x1014 3.20 3.25 3.30 3.35 3.40 3.45 3.50 3.55 3.60 Eg (e V ) (n or p) 2/3 (cm-2) measured Eg calculated Eg
Fig. 3D. The measured and calculated band gap as function of charge carrier (n or p) concentration (Egmeasured is determined using absorption plots) and (Eg
calcu-lated is determined according to Burnstein–Moss relation), p-type value is indicated by the open red circle. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
5.02 x 1019 cm3, and the resistivity ranged within the gap
0.4 x 102–8.5 x 102X.cm as a result of Al doping were reported
[35]. While Moharrami obtained also a lower concentration around 1018cm3for both n and p-type conductivity[15]. Elangovan
re-ported that a bulk concentration of electrons (n-type semiconduc-tor) was close to 1021cm3[36]. A bias voltage was measured as
follows[37],
q
¼VSlI ð6Þ
V is the measured potential drop across the sample, I is the current through the sample, S is the cross section area through which the current runs and l is the distance separating the voltage leads. The Hall coefficient RHis given by[37],
RH¼
VHS
I
l
0Hlð7Þ Where VHis the Hall voltage and
l
0H is the magnetic field. Usingthe HMS measurement, we applied a current of 1 nA to sample B, which exhibited a bulk carrier density and a mobility slighter by 102and 0.3 respectively than those found for samples A and C,
whereas it is more resistive (r>4 x 102X.cm). The sheet
concentra-tion Ns (cm2) is found to be: +1.7 x 1014
(B, p-type) and: 1.2 x 1016(A) , 3.1 x 1016(C), 1.6 x 1013(D) (n-type), when the
Al content exceeds 3 %, the sheet concentration increased as seen inFig. 4C. This fact was explained by the change in charge carrier concentration at the film surface. The average Hall coefficient differs from p to n-type by 102as listed inTable 4. Sample D was n-type
and showed the lowest bulk density and the mobility of charge car-rier giving the largest resistivity of 1.8X.cm and the disorder even the amorphous behavior occurs. As well as the doping amount in-creases in the host structure as well as the resistivity inin-creases. From these measurements, it may be confirmed that the conductiv-ity of tin oxide for acceptor doping concentration up to 2% is actu-ally an n-type semiconductor; and it becomes p type when the carrier concentration comes close to 2%. Al3+has a low valence than Sn4+, so it is considered as acceptor in the case of SnO
2, inducing
Fig. 4A. Hall measurement apparatus of Al doped SnO2grown onto ITO glass substrate; the films were kept by four Au probes (red arrow) as signed by yellow arrow (left). As
shown at right, the cheme of the samples, the four contacts M, N, P, Q, and the magnetic field was applied perpendicularly to the sample. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
0 2 4 6 8 10 -3.2x1021 -2.4x1021 -1.6x1021 -8.0x1020 0.0 Ch ar ge car ri er d e n s it y ( cm -3) Al content (%)
I
0 5 10 15 20 25 M o b ility (c m ²/V s ) 0 2 4 6 8 10 -3.2x1021 -2.4x1021 -1.6x1021 -8.0x1020 0.0 Al content (%)(b)
Char ge car ri er densi ty ( cm -3)II
-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Resi st iv it y ( Ω . cm )Fig. 4B. The plots of Hall Effect parameters: charge carrier and mobility (I), resistivity (II) versus Al content of sprayed Al:SnO2film grown at 300 °C onto ITO
glass substrate. Ω 0 2 4 6 8 10 -3x1016 -2x1016 -1x1016 0 Al content (%) Ns ( c m -2) 0.7 0.8 0.9 1.0 Sheet resistance ( Ω /Sq.)
Fig. 4C. The sheet concentration (left) and the sheet resistance (right) as function of Al doping content of the samples A, B, C and D.
holes’ concentration increase in the lattice. As mentioned in litera-ture, if the Sn tetravalent ions were substituted by Al trivalent ions, one broken bond (hole) is produced, which acts as an acceptor en-ergy level near the valence band and accepts an electron from it, thus increasing the hole concentration or the p-conductivity[38]. M.M. Bagheriet al. reported that sprayed Al-doped tin oxide, depos-ited onto glass substrate, exhibdepos-ited a change in resistivity as a result of varying the acceptor dopant from 10-2to 0.9
X.cm[38]. The sheet resistance (Rs was expressed as Rs=
q
/t[39], where t is the layer’s thickness. The parameter Rs of our films reached a lower point at acceptor concentration of 2% Al and then it increased drastically till 0.99X/sq (8.5 Al % doped sample), because of the substitution of Al3+cations in Sn4+lattice, consequently creating a number of holes.Based on the X-ray patterns, 8.5% Al doped SnO2tends to exhibit an
amorphous structure, which we believe has created, consequently increased Rs. Furthermore, Bagheri et al. reported that Rs goes from 0.6 to 15.6 kXfor a fixed SnO2:Al film thickness of 450 nm[38], a
higher value of Rs ~33 kX was found by Chantarat et al. [40]. However, the carreful analysis of our samples revealed a n or p-type conductivity. With a charge carrier density of 1.7 1019cm3, a
mobility of 7.6 cm2/Vs and a resistivity around 0.05
X.cm the sam-ple B exhibits a p-type conductivity while the n-type, 0.5 and 3% Al doped SnO2layers, have a lower resistivity and a higher bulk
con-centration respectively as indicated in Table 4. The ability of sample to turn into p-type may be due to doping with Al trivalent element in the case of our study. The inversion of SnO2semiconductor doped
with Zn (II) from n to p-type was achieved by Miao Ni et al. and a p-type resistivity of 3.58–13Xcm was reported [41]. Generally, our results are in well agreement with those obtained previously, a resistivity range of 0.189–78 (104Xcm), a mobility range of
2.230.7 (cm2/Vs) and a bulk concentration range of 937
(1019cm3) [26,4247]. The problem in the doping method is
the concentration level of acceptor dopant in the host lattice. It is clear that in the high acceptor doping condition, amorphous phase and atomic disorder increase as well as the electrical resistivity. Consequently, in the chemical design of a p-type TCO (from n-type TCO), the optimum doping level is necessary to be determined. Be-low this optimum level, TCO remains an n-type material[48]. In our samples, with increasing Al concentration, the film’s sheet resis-tance increases roughly because of the substitution of Al3+in Sn4+
sites and consequently the production of holes. The Hall voltage is negative and the majority carriers are electrons. However, the sheet resistance of the films is increasing with the Al content and the maximum of Rsoccurs in 8.5 Al% doped samples at 80 kX/Sq.
Fur-thermore, based on X-ray patterns, an increase in resistance may be due to increasing in the disordered or amorphous character of the film. Al3+ is the suitable dopant for acceptor doping process,
with the susbstitution of the lower valency of cation in Sn4+
posi-tions of tin oxide lattice structure, one or a few holes may be pro-duced and then decreases the n-type conduction or increases hole density and thus p-conductivity. Compared to the obtained values for the samples A and C, a high value of resistivity of sample B is reached (
q
> 0.04X.cm). The acceptor cation can be mixed with oxygen 2p states, thus a metal-oxygen bonding in lattice of SnO2is made. Another parameter, can play a role in the transition phase (n to p) of conduction mechanism in such oxides, is the cationic size of Al3+. The lattice disorder at higher doping levels is a characteristic nature of SnO2films as mentioned in previous papers[49]. The SnO2
transparent conducting thin films are n-type semiconductors with a direct optical band gap about 3.44.2 eV. In the tin oxide layers, the valence band is a closed shell of oxygen 2s22p6states, mixed with
some Sn states. The conduction band consists of a wide minimum of empty Sn 5s states with a low effective mass of 0.3 m0. The
struc-ture of tin oxide in its bulk form is tetragonal rutile with lattice parameters a = 4.74 Å and c = 3.19 Å. But in thin film form, depend-ing on the production technique, their structure can be polycristal-line or amorphous. The grain size is typically 20-40 nm, which is highly dependent on the deposition temperature, process, post-treatment and doping rate. Therefore, it is possible to increase the electrical conductivity by substrate heating during deposition or post heattreatment, in order to improve the crystal structure and increase grain size. A good criterion to define the quality of trans-parent films is the figure of merit / (X1) expressed as follows[37],
/¼T
10
Rs ð8Þ
where T is the transmittance in the visible range and Rs is the sheet resistance (Rs
q
/t), the values of figure of merit were gathered inTable 3. The figure of merit ranged within 0.1-49 (103X1) as
listed inTable 3. Our figure of merit is in the same order of magnitude as reported previously (106–103
X1)[26,27]. Ravichandran has
Fig. 5. Room temperature photoluminescence spectra of sprayed Al:SnO2films
produced at 300 °C as function of the photon wavelength. Gaussian deconvolution is shown (blue curves), emission peaks a (2.80 eV), b (2.30 eV) and c (2.11 eV) are displayed. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Table 4
Bulk concentration of charge carrier Nb, mobilityl, resistivityq, average Hall coefficient, conductivity type and sign of Hall voltage of sprayed Al:SnO2grown onto ITO glass
substrate at 300 °C (sample A: 0.5%, B: 2%, C: 3% and D: 8.5%).
Samples (Al%) Nb(cm3) l(cm2/Vs) q(Xcm) RH(cm3/C) Type VHsign
A: 0.5 1.25 1021 22.79 2.18 10-4 4.98 10-3 n Negative B: 2 +1.69 1019 7.57 4.87 10-2 +3.69 10-1 p Positive C: 3 3.12 1021 20.32 9.8 10-5 2.00 10-3 n Negative D: 8.5 1.58 1018 2.16 1.82 3.95 n Negative A: 0.5 1.25 1021 22.79 2.18 10-4 4.98 10-3 n Negative B: 2 +1.69 1019 7.57 4.87 10-2 +3.69 10-1 p Positive C: 3 3.12 1021 20.32 9.8 10-5 2.00 10-3 n Negative D: 8.5 1.58 1018 2.16 1.82 3.95 n Negative
found a good quality factor around 2x104X1for a high Al (20%)
doped tin oxide[35].
3.5. Photoluminescence spectra investigation
The room temperature photoluminescence spectra of as-grown films is depicted in Fig. 5. The peaks labeled (a): 2.80 eV, (b): 2.30 eV and (c): 2.11 eV are the main observed PL emissions. As shown inFig. 5, no significant change in peak position is observed. Strong PL emissions in the visible band were revealed while no UV emissions, originating from the near-band edge, were detected. The Al3+ions can go to the interstitial sites as donnors and might oc-cupy Sn4+sites as acceptors. In the 8.5% Al doped SnO
2, the peaks a, b
and c correspond respectively to violet, green and yellow emissions (442, 540 and 590 nm), the same behavior, with a slight differ-ence in peak intensity, is observed for the 3% Al-doped SnO2sample.
The peaks a, b and c are positioned at the same wavelengths but the peak intensity decreases for the 2% and 0.5% Al-doped samples as indicated by an arrow inFig. 5. Both, the preferential orientation and abundant grain boundaries in the as-grown SnO2films might
cause the presence of the violet emission, the fitting peaks are given inFig. 5. Intensive violet emission peak is observed in the region
393395 nm for Sb-doped SnO2 nanostructured films produced
by spray pyrolysis as reported by Babar et al.[50]. As shown in
Fig. 5, the intensity of PL emission increases with Al doping. Extrin-sic impurity or defects might cause the yellow emission 590 nm. Similar strong emission, around 600 nm indicating oxygen vacan-cies, has been detected by Luo et al.[51]. Furthermore, lesser PL bands are respectively blue emission (500–510 nm) and green emission (535–540 nm). Whereas, intense blue luminescence band around 545 nm was observed in tin oxide nanoblades prepared at low temperature[52]. In our films, the green emission 540 nm presence is due to the concentration of defects responsible of deep energy level in SnO2films.
Conclusions
Nanostructures of aluminum doped tin oxide sprayed onto ITO glass substrate were successfully synthesized. A (1 0 1)-oriented tetragonal crystalline structure was confirmed by X-ray patterns analysis. According to (2 1 1) orientation the obtained grain size of the as-grown films is lower <45 nm. Thin films, with thicknesses in the range 230300 nm, showed nanostructures with size 70 nm. Nanotips and nanopatches were also observed, grown from inner to the surface, having an average diameter of 180 and 400 nm respectively. The films’ typical transmittance in the visible range is 85% and the optical bandgap is found to be 3.7 eV. Through this study, we reported that the lowest resistivity of SnO2is less
than 104
Xcm, which is very good and competitive with other
leading material candidates. Moreover, a very low resistivity (9.85 105Xcm) and a high transmittance in visible spectrum
give Al:SnO2the characteristics of a great TCO. Overall, the samples
were n-type degenerate semiconductors (n 1020cm3) and
man-ifested inversion from n to p-type conductivity as a result of Al content (2% Al). More interestingly, the conductivity of tin oxide (SnO2) semiconductor is improved by the aluminium doping. The
high electrical conductivity and the high carrier concentration that we obtained are quite promising for photovoltaic devices. PL emis-sion have been measured, and violet, green and yellow emisemis-sions have been detected.
Acknowledgements
The work is included in the PNR projects under contract number 8/U311/R77 and U311/R81, supported by ‘‘agence thématique de recherche en science et technologie’’ (ATRST)
http://www.atrst.dz/, and national administration of scientific research (NASR)http://www.nasr.dz. This work is a part of
CNE-PRU project NerD01920120039 supported by Oran University of
Sciences and Technology. The first authors are grateful for the assistance of The Head of DUBTAM-Renewable Energy Research
Laboratory http://www.dicle.edu.tr/dubtam/ and the virtual
library of SNDLhttps://www.sndl.cerist.dz.
References
[1] Cleber A. Amorim, Cleocir J. Dalmaschio, Edson R. Leite and Adenilson J. Chiquito, J. Phys. D: Appl. Phys. 47 (2014) 045301 (6pp).
[2]F.I. Chowdhury, T. Blaine, A.B. Gougam, Energy Procedia 42 (2013) 660–669.
[3]J.S. Bhat, K.I. Maddani, A.M. Karguppikar, S. Ganesh, Nucl. Instrum. Methods
Phys. Res. B 258 (2007) 369–374.
[4]C.E. Benouis, M. Benhaliliba, F. Yakuphanoglu, A. Tiburcio Silver, M.S. Aida, A.
Sanchez Juarez, Synth. Met. 161 (2011) 1509–1516.
[5]Zhenguo Ji, Zhenjie He, Yongliang Song, Kun Liu, ZhiZhen Ye, J. Cryst. Growth
259 (2003) 282–285.
[6]Yuan-Qing Li, Jian-Lei Wang, Fu Shao-Yun, Shi-Gang Mei, Jian-Min Zhang, Kang
Yong, Mater. Res. Bull. 45 (2010) 677–681.
[7]R.K. Mishra, P.P. Sahay, Ceram. Int. 38 (2012) 2295–2304.
[8]M.A. Batal, F. Haj Jneed, Energy Procedia 6 (2011) 1–10.
[9]C. Luangchaisri, S. Dumrongrattana, P. Rakkwamsuk, Procedia Eng. 32 (2012)
663–669.
[10]F. de Moure-Flores, J.G. Quiñones-Galván, A. Hernández-Hernández, A.
Guillén-Cervantes, M.A. Santana-Aranda, M.de la L. Olvera, M. Meléndez-Lira,
Appl. Surf. Sci. 258 (2012) 2459–2463.
[11] D. Liu et al., Gas sensing mechanism and properties of Ce-doped SnO2sensors
for volatile organic compounds, Mater. Sci. Semicond. Process. (2012),http://
dx.doi.org/10.1016/j.mssp.2012.02.015.
[12]Wen Zeng, Tianmo Liu, Zhongchang Wang, Physica E 43 (2010) 633–638.
[13]Sangyub Ie, Ji-Hwan Kima, Byeong Taek Bae, Dong-Hee Park, Ji-Won Choi,
Won-Kook Choi, Thin Solid Films 517 (2009) 4015–4018.
[14]T. Rui-qin, G. Yan-qun, Z. Jun-hua, L. Yue, X. Tie-feng, S. Wei-jie, Trans.
Nonferrous Metal Soc. China 21 (2011) 1568–1573.
[15]F. Moharrami, M.-M. Bagheri-Mohagheghi, H. Azimi-Juybari, M.
Shokooh-Saremi, Phys. Scr. 85 (2012) 015703. 6.
[16]C. Van Komen, A. Punnoose, M.S. Seehra, Solid State Commun. 149 (2009)
2257–2259.
[17]H. Cho, Y.H. Yun, Ceram. Int. 37 (2011) 615–619.
[18]A. Pohl, B. Dunn, Thin Solid Films 515 (2006) 790–792.
[19]J. Gao, R. Chen, D.H. Li, L. Jiang, J.C. Ye, X.C. Ma, X.D. Chen, Q.H. Xiong, H.D. Sun,
T. Wu, Nanotechnology 22 (2011) 195706. 10.
[20]M. Benhaliliba, C.E. Benouis, A. Tiburcio-Silver, F. Yakuphanoglu, A.
Avila-Garcia, A. Tavira, R.R. Trujillo, Z. Mouffak, J. Lumin. 132 (2012) 2653–2658.
[21]E. Chan y Díaz, A. Duarte-Moller, Juan M. Camacho, R. Castro-Rodríguez, Appl.
Phys. A (2012) 619–624. doi: 10.1007/s00339-011-6630-7.
[22]S. Chacko, N.S. Philip, K.G. Gopchandran, P. Koshy, V.K. Vaidyan, Appl. Surf. Sci.
254 (2008) 2179–2186.
[23]J. Kaur, R. Kumar, M.C. Bhatnagar, Sens. Actuators B 126 (2007) 478–484.
[24]W. Hamd, Y.C. Wu, A. Boulle, E. Thune, R. Guinebretière, Thin Solid Films 518
(2009) 1–5.
[25]J. Faucheu, K.A. Wood, L.P. Sung, J.W. Martin, JCT Res. 3 (1) (2006) 29–39.
[26]T. Fukano, T. Motohiro, Sol. Energy Mater. Sol. Cells 82 (2004) 567–575.
[27]A.R. Babar, S.S. Shinde, A.V. Moholkar, C.H. Bhosale, J.H. Kim, K.Y. Rajpure, J.
Semicond. 32 (5) (2011). 3001(8).
[28] D.Y.T. Martínez, R.C. Pérez, G.T. Delgado, O.Z. Angel, J. Photochem. Photobiol. A: Chem. (2010),http://dx.doi.org/10.1016/j.jphotochem.2012.03.009.
[29]E. Elangovan, M.P. Singh, M.S. Dharmaprakash, K. Ramamurthi, J. Optoelectron.
Adv. Mater. (6 1) (2004) 197–203.
[30]M. Benhaliliba, C.E. Benouis, M.S. Aida, F. Yakuphanoglu, A. Sanchez, Juarez, J.
Sol Gel Sci. Technol. 55 (2010) 335–342.
[31] S.S. Pan, C. Ye, X.M. Teng, H.T. Fan, G.H. Li, Appl. Phys. A 85 (2006) 21–24,
http://dx.doi.org/10.1007/s00339-006-3659-0.
[32]S.J. Ikhmayies, R.N. Ahmad-Bitar, Phys. Scr. 84 (2011) 055801. 7.
[33]C.E. Benouis, M. Benhaliliba, A. Sanchez Juarez, M.S. Aida, F. Chami, F.
Yakuphanoglu, J. Alloys Comp. 490 (2010) 62–67.
[34]A. Gholizadeh, N. Tajabor, M.R. Alinejad, Iranian J. Cryst. Miner. 16 (2009) 665–
675.
[35]K. Ravichandran, K. Thirumurugan, J. Mater. Sci. Technol. 30 (2) (2014) 97–102.
[36]E. Elangovan, S.A. Shivashankar, K. Ramamurthi, J. Cryst. Growth 276 (2005)
215–221.
[37]Q.H. Luo, Y.H. Lu, Y.Z. Lou, Y.B. Wang, D.L. Yu, Thin Solid Films 518 (2010)
7038–7043.
[38]M.M. Bagheri Mohagheghi, M. Shokooh Saremi, J. Phys. D: Appl. Phys. 37
(2004) 1248–1253.
[39] G. Haacke, J. Appl. Phys. 47 (1976) 4086,http://dx.doi.org/10.1063/1.323240.
[40]N. Chantarat, Yu-Wei Chen, Chin-Ching Lin, Mei-Ching Chiang, Yu-Chun Chen,
San-Yuan Chen, J. Phys. Chem. C 115 (2011) 23113–23119.
[41]J. Miao Ni, X.J. Zhao, J. Zhao, J. Inorg. Organomet. Polym. (2012) 21–26. doi:
10.1007/s10904-011-9619-y.
[43]A.V. Moholkar, S.M. Pawar, K.Y. Rajpure, C.H. Bhosale, Mater. Lett. 61 (2007)
3030.
[44]A.I. Martinez, D.R. Acosta, Thin Solid Films 483 (2005) 107.
[45]M. Mwamburi, E. Wackelgrd, A. Roos, Thin Solid Films 374 (2000) 1.
[46]V. Vasu, A. Subrahmanyam, Thin Solid Films 189 (1990) 217.
[47]D.R. Acosta, E.P. Zironi, F. Montoya, W. Estrada, Thin Solid Films 288 (1996) 1.
[48] M.M. Bagheri Mohagheghi, M. Shokooh Saremi, in: Thomas B. Elliot (Ed.), Trends in Semiconductors Research, 191316 ISBN 1-59454-414-X.
[49]E. Elangovan, K. Ramamurthy, J. Appl. Sci. 247 (2005) 183–196.
[50] A.R. Babar, S.S. Shinde, A.V. Moholkar, C.H. Bhosale, J.H. Kim, K.Y. Rajpure, J.
Alloys Comp. 509 (2011) 3108–3115.
[51]S.H. Luo, Q. Wan, W.L. Liu, M. Zhang, Z.T. Song, C.L. Lin, Paul K. Chu, Prog. Solid
State Chem. 33 (2005) 287–292.
[52]Yung-Chiun Her, Wu Jer-Yau, Yan-Ru Lin, Song-Yeu Tsai, Appl. Phys. Lett. 89
