Sunteți pe pagina 1din 5

ARTICLE IN PRESS

Materials Science in Semiconductor Processing 10 (2007) 128132

Short communication

Optoelectronic properties of ZnSe thin lms


S. Venkatachalama,b,, S. Agilanb, D. Mangalarajb,, Sa.K. Narayandassb
a

Department of Mechanical System Engineering, University of Yamanashi, 4-3-11 Takeda, Kofu 400-8511, Japan b Thin Film Laboratory, Department of Physics, Bharathiar University, Coimbatore 641 046, TN, India Available online 10 July 2007

Abstract ZnSe/Si Schottky diodes were fabricated by the vacuum evaporation technique at different substrate temperatures (303 and 483 K) and the results are discussed. Compositions of the deposited lms were measured by Rutherford backscattering analysis. The atomic percentages of Zn and Se were evaluated as 0.50 and 0.50, respectively. Structural analysis shows that deposited lms posses good crystalline nature with a cubic structure oriented along the (1 1 1) direction. In the Raman analysis, the presence of the LO mode conrms that the deposited lms have the crystalline nature. Photoconduction analyses were performed on the prepared device. The prepared device exhibits a strong photoresponse in the visible region. r 2007 Published by Elsevier Ltd.
PACS: 73.61.Ga; 68.55.Nq; 78.66.Hf; 68.55.Jk; 73.50.Pz Keywords: Thin lms; Structural; Raman; Optical; Schottky diodes

1. Introduction Recently, wide band gap IIVI compound semiconductors have attracted much attention because of their properties and their applicability in several types of optoelectronic devices. Particularly, Schottky photodiodes with fast responses in the UV visible region are more in focus, because Schottky barriers result in both very fast switching times and low forward voltage drop. Silicon photodiodes will give more response in the IR region; the main reason for this is the band gap value. The reported
Corresponding authors. Thin Film Laboratory, Department of Physics, Bharathiar University, Coimbatore 641 046, TN, India. Tel.: +91 422 2425458; fax: +91 422 2422387. E-mail addresses: svchalam23@yahoo.co.in (S. Venkatachalam), dmraj800@yahoo.com (D. Mangalaraj).

band gap value of Si is 1.1 eV. By providing an overlayer on the silicon surface, the silicon Schottky diodes will give more response in the UVvisible region. Due to their direct energy gap in the visible region, ZnSe would be perfectly suitable for this. Also, the lattice match between the substrate (Si) and overlayer (ZnSe) is very important. ZnSe has either a sphalerite structure with lattice parameter a 5.668 A or a wurtzite structure with lattice parameters a 3.820 A and c 6.626 A. The lattice constant value of cubic Si is reported as 5.6576 [1]. Lattice mismatch between these two materials is 4.4%. Wen-Shiung Lour and Chung-Cheng Chang [2] have rst successfully fabricated a new PIN-like (Si (p)/ZnSe (n)/ZnSe (n+)) visible photodiode in 1996 using vapour-phase epitaxy. They have used a two-step growth method to overcome the problem of lattice mismatch existing between ZnSe and Si.

1369-8001/$ - see front matter r 2007 Published by Elsevier Ltd. doi:10.1016/j.mssp.2007.06.001

ARTICLE IN PRESS
S. Venkatachalam et al. / Materials Science in Semiconductor Processing 10 (2007) 128132 129

Ullrich [3] has fabricated ZnSe/Si diodes by using the molecular beam epitaxy technique. In order to bypass the lattice mismatch, he used InSe (25 A) as a buffer layer between ZnSe and Si. These two techniques are very expensive. But vacuum evaporation is very simple and inexpensive. To our knowledge, very few reports are available on the preparation and characterization of the optoelectronic properties of ZnSe/Si photodiodes. In particular, there is no report on the preparation of ZnSe/Si photodiodes using the vacuum evaporation technique. It is mainly used for the fabrication of large-area lm deposition. Hence, an attempt has been made on the preparation and characterization of ZnSe/Si photodiodes. In this paper, we report the preparation and characterization of ZnSe/Si photodiodes using the vacuum evaporation technique. 2. Experimental analysis

normal transmittance (T) was measured by UV VISNIR spectrophotometer over the wavelength range 2002500 nm. Raman experiments were carried out using the exciting source of wavelength 488 nm with a power of 30 mW. In the spectral response analysis, an Oriel 150 W xenon lamp was used as a light source, along with a monochromator from Acton Research Corporation. At each wavelength, the power P of the light incident on the Au mesa was carefully measured using a Newport 835 optical power meter that uses a UV-enhanced Si photodiode (818 UV) as the detector. The shortcircuit photocurrent Isc was measured as a function of the wavelength of the incident photons using a digital current meter (Keithley Model 237) with high sensitivity and high precision. 3. Results and discussion 3.1. Composition analysis

Normalized Yield

Pure ZnSe compound (99.995%, Semiconductor Material Corporation, Korea) was thermally evaporated from a molybdenum boat onto well-cleaned high resistivity (10 O cm) p-type silicon (1 0 0) silicon and glass substrate at room and elevated substrate temperatures (483 K) under a vacuum of 3 105 Torr. After depositing the ZnSe layer, gold (Au) was deposited on the front surface of the device. Here, aluminium was used as back contact. Au and aluminium electrodes were deposited by the thermal evaporation method. The active area of the prepared device is 3 106 m2. The deposition parameters are given in Table 1. Rutherford backscattering (RBS) analysis was done using a 2 MV tandem Pelletron accelerator from High Voltage Engineering in Europe (University of Surrey Ion Beam Centre). Typically, a beam of 1.5 MeV He+ ions is directed perpendicularly on the samples surface. Thicknesses of the deposited lms were measured by RBS and the thickness of the deposited lm was calculated as 230 nm. X-ray diffraction (XRD) studies were carried out using an X-ray diffractometer in the 2y range 20801. The spectral
Table 1 Deposition parameters of ZnSe thin lms Substrate Substrate temperature Source Evaporant Source-to-substrate distance Silicon (1 0 0) and glass 303 and 483 K Molybdenum and tungsten boat ZnSe 17 cm

Compositions of the deposited lms were measured by RBS. Fig. 1 shows the RBS spectrum of ZnSe thin lms deposited on silicon substrates at 303 K. The details of the experimental set-up can be found elsewhere [4]. The collected RBS spectra were then tted by the code Rump [5] to nd the relative concentrations of various elements in the lm. The atomic percentage of the deposited lms is calculated as 0.50% (Zn):0.50% (Se). It shows that the composition of the deposited lms is found to be nearly stoichiometric.
Energy (MeV) 1.0

60 50 40 30 20 10

0.5

1.5

0 100

200

300 Channel

400

500

Fig. 1. Rutherford backscattering spectrum of ZnSe lms deposited onto silicon at 303 K.

ARTICLE IN PRESS
130 S. Venkatachalam et al. / Materials Science in Semiconductor Processing 10 (2007) 128132

3.2. Structural properties X-ray diffractograms of ZnSe lms deposited on Si (1 0 0) substrates are shown in Fig. 2. The appearance of many diffraction peaks conrms that the ZnSe lms deposited on Si substrates have a polycrystalline nature. The XRD patterns indicate the predominant zinc blende structure. The peak intensity increases with an increase of substrate temperature. The XRD patterns exhibit reections at 27.241 (303 K), and correspond to (1 1 1) phase. Similar results have been observed by Bong Hyun Boo et al. [6]. The peak at $691 corresponds to silicon. The particle size (D) is calculated using Scherrers formula from the full-width at halfmaximum (b) (FWHM) D 0:94l , b cos y (1)

l . 2 sin y

(2)

The observed lattice spacing value (d 3.257 A) coincides with the reported value. The lattice constant (a) is determined by the following expression: 1 h2 k 2 l 2 , a2 d2 (3)

where h, k and l represent the lattice planes. The dislocation density (d) has been calculated from the following relation: d 15 . aD (4)

where l is the wavelength of the X-ray used, b is the FWHM and y is the angle between the incident and the scattered X-rays. The lattice spacing (d) is calculated from Braggs formula

600 500 Intensity (arb. units) 400 300 200 100


303 K
(111)

Si

483 K

0 20

30

40

50 2 (degrees)

60

70

80

Fig. 2. X-ray diffractograms of ZnSe thin lms deposited onto silicon at different substrate temperatures (Ts 303 and 483 K).

The calculated values of particle size, strain, dislocation density and lattice constant values are given in Table 2. The calculated lattice constant value coincides with the reported value [7]. Fig. 3 shows the Raman spectra of ZnSe thin lms deposited on glass substrates at 303 and 483 K. Generally, the peak positions of longitudinal optical (LO) phonon and transverse optical (TO) phonon modes for bulk ZnSe were positioned at 252 and 205 cm1, respectively [8]. In Fig. 3, the peak positions of LO and TO modes for the as-deposited (303 K) lms are at 248.38 and 203.02 cm1. If we compare the peak position of the LO mode with the reported bulk ZnSe, the LO mode shifts towards the lower wave number side. It is attributed that the deposited lms are affected by tensile stress [9]. The peak position of LO mode shifts from 248.38 to 249.208 cm1, with an increase of substrate temperature from 303 to 483 K. It shows that the effect of stress decreases with an increase of substrate temperature from 303 to 483 K. The FWHM value of LO modes decreases from 16.37 to 8.59 cm1, with an increase of substrate temperatures. The presence of the longitudinal mode conrms that the deposited lms have a crystalline nature [10,11].

Table 2 Structural parameters of ZnSe thin lms deposited onto glass substrates Substrate temperature (K) 303 483 Particle size (nm) 20.12 23.93 Strain 103 (lin2 m4) 1.95 1.644 Lattice spacing (A) Lattice constant (A) Dislocation density 1014 (lin/m2) 2.577 1.815

(220)

(311)

3.257 3.278

5.641 5.677

ARTICLE IN PRESS
S. Venkatachalam et al. / Materials Science in Semiconductor Processing 10 (2007) 128132 131

LO

0.4

TO Intensity (a. u)

Photocurrent x 10-6 (A)

0.3

0.2
4.5 V

483 K

0.1
4V 3.5 V

303 K

0.0

400

500 600 Wavelength (nm)

700

200

250 300 Raman shift (cm-1)

350

Fig. 3. Raman spectra of ZnSe thin lms deposited onto glass substrates at different substrate temperatures (Ts 303 and 483 K). (Inset shows a plot of (ahn)2 versus hn of ZnSe thin lms deposited on glass substrates at 303 and 483 K.)

Fig. 4. Variation of photocurrent with wavelength at different bias voltages (3.5, 4 and 4.5 V) for ZnSe thin lms deposited at 303 K.

3.3. Optical properties The relation between absorption coefcient (a) and incident photon energy (hn) is given by [12] ahn Chn E g n , (5)

where C is a constant, n 1/2 for direct transitions and n 2 for indirect transitions. The inset in Fig. 3 shows the plot between hn versus (ahn)2. The optical band gap of the deposited lms was determined by extrapolating the linear portion of the curve to the energy axis at (ahn)2 0. The optical band gap values for the lms deposited at 303 and 483 K were calculated as 2.726 and 2.67 eV, respectively. The band gap value decreases with an increase of substrate temperature. This may be attributed to the increase of particle size. The calculated optical band gap value well coincides with the reported value of ZnSe lms (2.62.7 eV) prepared by the vacuum evaporation method [13]. 3.4. Spectral response measurements In the present work, the ZnSe/Si Schottky diodes were prepared at 303 K. Variation of photocurrent with wavelength at different bias voltages (3.5, 4 and 4.5 V) is shown in Fig. 4. The photocurrent increases with an increase of bias voltage. The higher value of peak current is observed at 470 nm, because the signal around $470 nm corresponds to

the absorption edge of ZnSe lms. At 470 nm, the peak current is calculated as 0.214 mA. The peak current is measured at 3.5 V. The device prepared using pure ZnSe shows that the devices provide no signal if they were not biased. The prepared device shows mainly a visible response when it was biased at a voltage of 3.5 V. In order to bypass the lattice mismatch, Wen-Shiung Lour and Chung-Cheng Chang [2] and Ullrich [3] have used two-step growth techniques. But they have obtained results for multilayer structured Si (p)/ZnSe (n)/ZnSe (n+) photodiodes. We have obtained results for singlelayer structured Si/ZnSe diodes. But in the present work, the performance of the prepared device is poor. The efciency of the peak response was below 2% (just estimated by comparing with our UV diodes). 4. Conclusion ZnSe thin lms were prepared on silicon and glass substrates. The composition of the deposited lms is found to be nearly stoichiometric, which was conrmed by Rutherford backscattering analysis. The optical band gap value decreases with the increase in substrate temperatures. From X-ray diffraction and Raman analysis, it was conrmed that the deposited lms have a crystalline nature. The particle size of the deposited lms increases with the increase in substrate temperatures. ZnSe/Si Schottky photodiodes were successfully prepared without any buffer layer between the substrate (Si) and overlayer (ZnSe). In the photoresponse

ARTICLE IN PRESS
132 S. Venkatachalam et al. / Materials Science in Semiconductor Processing 10 (2007) 128132 [7] Khlyap G, Andrukhiv M. Cryst Res Technol 1999; 34(56):751. [8] de Moraes AR, Mosca DH, Schreiner WH, Mattoso N, Silveira E. Braz J Phys 2002;32(2A):383. [9] Rho H, Jackson HE, Lee S, Dobrowolska M, Furdyna JK. Phys Rev B 2000;61(23):641. [10] Sarigiannis D, Peck JD, Mountziaris TJ, Kioseoglou G, Petrou A. MRS Proc 2000;41:616. [11] Ja-Chin J, Shou-Yi K, Sun-Bin Y, Wen-Feng H. Chin J Phys 2001;39:90. [12] Rusu GI, Ciupina V, Popa ME, Prodan G, Rusu GG, Baban C. J Non-Cryst Solids 2006;352:15258. [13] Rusu GI, Popa ME, Rusu GG, Salaoru I. Appl Surf Sci 2003;218:222.

analysis, the prepared device shows a good response in the visible region. References
[1] ASTM JCPDS, X-ray powder diffraction le, Joint Committee for Powder Diffraction Standards, Swarthmore, PA, 1990, Card Number 37. [2] Lour W-S, Chang C-C. Solid State Electron 1996;39:1295. [3] Ullrich B. Mater Sci Eng B 1998;56:6971. [4] Jamieson DN. Nucl Instrum Methods B 1998;136138:1. [5] Doolittle LR. Nucl Instrum Methods B 1985;9(3):344. [6] Boo BH, Xu N, Lee JK. Vacuum 2002;64:14551.

S-ar putea să vă placă și