THE HALL EFFECT IN CdO:Sn THIN FILMS

Transcription

1 ANALELE ŞTIINŢIFICE ALE UNIVERSITĂŢII AL.I.CUZA DIN IAŞI Tomul XLIX-L, s.1.b.fasc.2 Fizica Stării Condensate, THE HALL EFFECT IN CdO:Sn THIN FILMS BY R.S. RUSU, G.I. RUSU KEYWORDS Thin films, CdO:Sn, Hall effect, carrier mobility ABSTRACT CdO:Sn thin films were deposited onto glass substrates by thermal evaporation under vacuum. The studied films are polycrystalline and have a sodium chloride structure. The Hall effect is investigated for films with different thickness, substrates maintained at different temperatures. The Hall voltage linearly depends on the magnetic induction. In the studied temperature range, ΔT= K, an exponential increase of Hall coefficient with increasing temperature is observed. The temperature dependence of the Hall mobility is also investigated. Some correlations between film structure and electronic transport properties are established. 1. INTRODUCTION Cadmium oxide (CdO) is an important semiconductor material for the development of various technologies of solid state devices, (panel display, optoelectronic components, thermally insulating glass, etc.) [1-3]. Some electrical properties of this oxide have been investigated by several authors [4-7]. The experimental data concerning the Hall effect of CdO are rather poor. A variety of methods has been used to prepare thin films of cadmium oxide such as thermal evaporation, oxidation of cadmium films, spray pyrolysis, metalorganic chemical deposition, etc. [5-9]. It was experimentally found that the electronic transport and optical properties of CdO thin films strongly depend on the preparation method and deposition conditions [6-9]. On the other hand, we remark, that in the above-mentioned papers on the electronic transport in CdO and CdO:Sn thin films, the experimental data were discussed with respect to a relative, small number of samples prepared under different conditions. Consequently, it is difficult to compare the results presented by different authors. D. Mangeron High School of, Iasi, Romania Al.I. Cuza University, Iasi, Romania

2 HALL EFFECT IN CDO:SN THIN FILMS 31 In a series of previous papers [10-11], we have studied temperature dependence of the electrical conductivity of CdO and CdO:Sn thin films. It was experimentally established that the stable structure of films can be obtained if, after preparation, they are submitted to a heat treatment. In present paper, the Hall coefficient in CdO:Sn thin films are investigated as a function of deposition conditions, magnetic induction and film temperature. 2. EXPERIMENTAL PROCEDURE CdO:Sn thin films were deposited onto glass substrates by physical vapour deposition under vacuum of high purity CdO and Sn. The quasi-closed volume technique was used [12]. The deposition equipment is described in detail in [10,11]. The experimental conditions were settled for obtaining films of a uniform thickness on large areas of the substrate surface. The samples under study were prepared using the following deposition parameters: the source evaporator-substrate distance was D=50mm; source temperature was T v =1100K; deposition rate was r d =10 /s; substrate temperature T S = K. Source temperature was monitored by a Pt/Pt-Rh thermocouple. A special holder for substrates was made of aluminium and permits to prepare four samples simultaneously. A chromel-alumel thermocouple, which monitored substrate temperature, was attached to the front surface of the substrate. The film thickness was measured by using multiple-beam Fizeau s fringe method [13] at reflection of monochromatic light, λ=550nm. The X-ray diffraction (XRD) patterns of the films were recorder with a DRON-2 diffractometer operating with CuK α radiation (λ= ). For measurement of the Hall coefficient and its temperature dependence a special arrangement has been used [10,11]. The indium thin film electrodes were used. The Hall voltage was determined by a standard DC potentiometric method. Two Keithley instruments (Model 6517 electrometer and a Model 487 picoampermeter) were used for measurements. The Hall coefficient was calculated from the following expression [14,15] U H d R = (1) BI where U H is Hall voltage, d represents film thickness, B is magnetic induction and I denotes is intensity of total current which passes through the film.

3 32 R.S. RUSU et. al. The magnetic induction (produced by an electromagnet) was measured by a teslameter with Hall sonde. The temperature dependence of the electrical conductivity, σ, was studied using surface-type cells. The optical bandgap of CdO:Sn thin films was determined from optical absorption spectra. The absorption coefficient, α, was calculated from the expression [16,17] 2 1 (1 R) α = ln (2) d T where d is film thickness, and R is reflection coefficient and T represents transmission coefficient. Reflection and transmission spectra (in the spectral range from 400nm to 1800nm) were recorded using a PMQ II (C. Zeiss, Jena) spectrophotometer and an ETA-STO (Etaoptic Steag) spectrometer. 3. RESULTS AND DISCUSSION The XRD patterns indicate that the studied samples are polycrrystalline and have a cubic (sodium chloride) structure. For as-deposited samples of CdO:Sn films the crystallites are preferentially oriented with the (111) planes parallel to the substrate. A typical XRD pattern is presented in Fig. 1. intensity (u.a.) sample A21 d=0.22μm; T S =473K λ CuKα = nm CdO (111) β - Sn Θ(deg) Fig. 1 XRD-pattern for CdO:Sn thin films.

4 HALL EFFECT IN CDO:SN THIN FILMS 33 A peak characteristic for crystalline tin (β-sn) is observed from these patterns. It is known, that β-sn has a tetragonal crystalline structure and is characterized by metallic properties [18,19]. It can be supposed that excess tin atoms are distributed within the film crystallites and also form the crystalline tin precipitates at intercrystalline boundaries. The crystallite size have been determined from diffraction peaks using the Debye-Scherrer expression [12,20,21]. It was established that the crystallite size ranged between 20nm and 37nm for films with thickness in the range nm. A linear dependence of the voltage, U H, as a function of magnetic induction, B, has been observed for films with different thickness, deposited onto the substrates maintained at different temperatures during the film growth (Figs. 2-4.) sample A21 d=0.22μm; T S =473K U H (mv) I=5mA I=10mA I=15mA I=20mA B(T) Fig. 2 The dependence of Hall voltage on the magnetic induction It is known that in polycrystalline thin films electronic transport mechanism is influenced by the intercrystalline boundaries. This mechanism is based upon the consideration that the crystallite boundaries have an inherent charge region due to the interface. Consequently, energy band bending occurs, and potential barriers to the transport of charge carriers result. For investigated samples, the activation energy of electrical conduction, E a, calculated from the temperature dependence of the electrical conductivity can vary from 0.30 to 0.70eV. These values strongly differ from energy gap, E g,

5 34 R.S. RUSU et. al. of CdO crystals (the values of E g reported by different authors for cubic CdO crystals are, usually, ranged between 2.70 and 3.20eV [1,2,4,8]) sample A23 d=0.68μm; T S =473K U H (mv) I=5mA I=10mA I=15mA I=20mA B(T) Fig. 3 The dependence of Hall voltage on the magnetic induction sample A31 (d=0.19μm) sample A32 (d=0.46μm) sample A33 (d=0.62μm) (T S =523K; I=12.5mA) 0.7 U H (mv) B(T) Fig. 4 The dependence of Hall voltage on the magnetic induction

6 HALL EFFECT IN CDO:SN THIN FILMS 35 Clearly, that the conduction mechanism in investigated samples can be explained by applying the models developed for the semiconducting discrete (polycrystalline) structure [13,22,23]. In [24,25] we found that Seto s model [26,27] could explain the electronic transport mechanism in ZnTe, Sb 2 O 3 and Sb 2 S 3 polycrystalline films. On the other hand, the linear dependence of the Hall voltage, U H, as function of magnetic induction, B, (Fig. 2-4) shows that in the CdO:Sn thin films, boundary domains little influence on the carrier transport in magnetic field. We consider that presence of tin crystalline (metallic) precipitate in the intercrystallite domains determines a important decrease of electrical resistivity of these domains. Crystallite size and film thickness can also influenced the Hall coefficient. One of the wellknown models accounting for the Hall coefficient in the semiconducting thin films is that developed by Amith [28]. According to this model, the effective Hall coefficient, R H, is related to Hall coefficient intracrystalline domains R Hb through the expression λ RH = RHbη (3) d where η(λ/d) is a function depending on the ratio λ/d (λ denotes the mean scattering length and d is film thickness). For smaller values of λ/d (λ/d<2), the function η(λ/d) 1 and R H is little influenced by film thickness and mean scattering length. For bulk CdO crystals the mean scattering length have lower values (λ 10-15nm). Fig. 5 shows that Hall coefficient not depend on the magnetic induction B. This behavior is justified, because the condition μ 2 B 2 <<1 (where μ is carrier mobility) is true for investigated samples (μ= m 2 /(Vs)). The temperature dependence of the Hall coefficient is presented in Figs.6-7. An exponential increase of Hall coefficient with increasing temperature is observed in all investigated temperature range. Clearly, the extrinsic conduction prevails in studied samples. Impurity atoms (tin atoms) introduce energy levels within the energy gap and act as donors. The activation energy of the impurities, E d, calculated from the dependences ln(r H T 3/2 ) =f(10 3 /T) [14,15,18], generally ranged between 0.020eV and 0.070eV.

7 36 R.S. RUSU et. al. 6.0x x x x10-8 sample A21 (d=0.22μm) sample A22 (d=0.51μm) sample A23 (d=0.68μm) (T S =473K; I=12.5mA) R H (m 3 /C) 4.0x x x x x B(T) Fig. 5 The dependence of Hall coefficient on the magnetic induction 4.5x10-8 proba A21 (d=0.22μm) proba A22 (d=0.51μm) I=15mA 4.0x10-8 proba A23 (d=0.68μm) (T S =473K; B=0,22T) 3.5x x10-8 R H (m 3 /C) 2.5x x x x T(K) Fig. 6 Temperature dependence of the Hall coefficient for films with various thickness

8 HALL EFFECT IN CDO:SN THIN FILMS ln[r H T 3/2 (m 3 K 3/2 /C) I=10mA sample A11 (d=0.17μm) sample A12 (d=0.39μm) sample A13 (d=0.57μm) (T S =293K; B=0.22T) /T(K -1 ) Fig. 7 Temperature dependence of the Hall coefficient for films with various thickness 8.0x x10-5 sample A23: T S =473K; d=0.68μm sample A22: T S =473K; d=0.51μm sample A21: T S =473K; d=0.22μm I=20mA μ(m 2 /Vs) 4.0x x T(K) Fig. 8 Temperature dependence of the Hall mobility for films with various thickness

9 38 R.S. RUSU et. al. In Fig. 8 the dependence of the Hall mobility on the temperature is illustrated for three studied samples. The Hall mobility μ H has estimated by means of the expression μ H = R H / ρ (4) where R H is Hall coefficient and ρ denotes the electrical resistivity. It is known that carrier mobilities depend on the predominant scattering mechanism in studied sample. In bulk semiconductors two scattering mechanisms are dominant: the scattering of the carriers by crystal lattice and scattering by ionized impurities. The carrier mobility due to lattice vibrations can be expressed by [18,19] 5/ 2 3/ 2 μ L = A L m ef T (5) and charged ionized centers affect mobility as follows [18,19]. 1 / 2 3/ 2 μ I = A I m ef T (6) where A L and A I are the characteristic parameters, m ef denotes scalar effective mass of charge carriers and T is absolute temperature. In the polycrystalline thin films, free carriers are scattered by the crystallite boundary surface in addition to the scattering mechanisms observed in respective bulk materials. Fig. 8 shows temperature dependence of Hall mobility for three studied samples. It can be observed that these dependences present two parts. We suppose that in the first parts the carrier scattering by ionized impurities (tin donors) prevails. According to Eq. (6), in this case an increase of carrier mobility with temperature is observed. The decrease of mobility in higher temperature is caused by lattice scattering of charge carriers. Also, the large value of the carrier concentration determines a decrease of the mobility [19]. By studying transmission and absorption spectra of CdO thin films, very useful information can be obtained about the energy gap, position of impurity levels in forbidden band, characteristics of optical transitions, etc. For investigated samples, the values of optical energy band gap have been calculated using absorption spectra. The determination of E g from the absorption spectra seems to be more appropriate. An important characteristic of absorption spectra for polycrystalline semiconducting films is an additional maximum (compared to single crystals) for photon energies less than band gap. At higher photon energy the absorption spectra are little influenced by polycrystalline structure of the films.

10 HALL EFFECT IN CDO:SN THIN FILMS 39 For allowed direct band-to-band transitions (neglecting exciton effects), the energy dependence of absorption coefficient, α, near the band edge, is given by [17,22] Aa 1/ 2 α = ( hν Eg ) (7) hν where hν denotes the photon energy and Aa is the characteristic parameter (independent to photon energy) for respective transitions. According to Eq. (7), the (αhν) 2 =f(hν) dependences are linear. Consequently, the optical band gap, E g0, can be determined by extrapolating the linear portions of mentioned dependences to (αhν) 2 =0. 1.5x10 13 sample A22 (CdO:Sn) d=0.51μm; T S =473K 1.0x10 13 (αhν) 2 (m -1 ev) 2 5.0x E g d =3,29eV hν(ev) Fig. 9 Absorption spectra for CdO:Sn thin films. A typical absorption spectra is presented in Fig. 9 For studied samples the values of E g0 ranged between 2.90eV and 3.30eV, which is in a good agreement with the E g0 values obtained for bulk CdO:Sn samples. 4. REFERENCES [1] Tsuda N, Nasu K, Fujimori A, Siratori K (2000) Electronic Conduction in Oxides, Springer- Verlag Berlin-Heidelberg-New York [2] Chopra CL, Das SR (1983) Thin Film Solar Cells, Plenum Press, New York-London [3] Maissel LI, Glang R (1970) Handbook of Thin Films Technology, McGraw Hill New York [4] Champness CH, Chan CH (1995) Sol. Energy Mater. Sol. Cells 37:75 [5] Kondo R, Okhimura H, Sakai Y (1971) Jpn. J. Appl. Phys. 10:1547 [6] Lokhande BJ, Uplane MD (2001) Mater. Res. Bull.. 36:439

11 40 R.S. RUSU et. al. [7] Dakhel AA, Henari FZ (2003) Cryst. Res. and Technol. 38:979 [8] Zhao Z, Morel DL, Ferekides CS (2002) Thin Solid Films 413:203 [9] Metin H, Esen R (2003) Semicond. Sci. Tehnol. 18:647 [10] Rusu RS, Rusu GI (2004) On the Electrical and Optical Properties of Polycrystalline CdO Thin Films, 7 th Intern. Conf. Phys. Adv. Mater., June, 10-12, Iassy-Romania [11] Rusu RS, Rusu GI (2004), On the Hall Effect in CdO Thin Films, Workshop Fundam. Appl. Res. Phys., November,3, Iassy-Romania [12] Rusu M, Nicolaescu II, Rusu GG (2000) Appl.Phys. A70:565 [13] Chopra KL (1969) Thin Film Phenomena, McGraw Hill New York [14] Putley EH (1969) The Hall Effect and Semiconductor Physics, Dover Publications, New York [15] Kuchis EV (1974) Investigation methods of Hall effect (in Russian), Sov. Radio, Moskva [16] Hodgson JN (1970) Optical Apsorption and Dispersion in Solids, Chapman&Hall London [17] Heavens O (1965) Optical Properties of Thin Solid Films, Dover NewYork [18] Seeger K (1999) Semiconductor Physics, Springer, Berlin [19] Wolf HF (1971) Semiconductors, Wiley New York [20] Flewitt PEJ, Wild KR (1994) Physical Methods for Materials Characterization, IOP Publishing Bristol-Philadelphia [21] Tigau N, Ciupina V, Prodan G, Rusu GI, Vasile E (2004) J. Cryst. Growth 269:392 [22] Kazmersky LL (1980) Polycristalline and Amorphous Thin Films and Devices, Academic Press New York [23] Harbeke G (1985) Polycrystalline Semiconductors: Properties and Application, Springer Berlin [24] Rusu GI, Popa ME, Rusu GG, Salaoru I (2003) Appl.Surf. Sci. 218:222 [25] Tigau N, Gheorghies C, Rusu GI, Condurache-Bota S (2004) J-Non. Cryst. Solids (in press) [26] Seto JYW (1975) J. Elecrochem. Soc. 122:701 [27] Seto JYW (1975) J. Appl. Phys. 46:5247 [28] Amith A (1960) J. Phys. Chem. Solids 14:271