Raman scattering and optical absorption studies of Ar+ implanted CdS thin films

Beam Interactions with Materials 8 Atoms

ELSEVIER

Nuclear

Instruments

and Methods

in Physics

Research

B 132 (1997) 61-67

Raman scattering and optical absorption studies of Ar+ implanted CdS thin films K.L. Narayanan

‘, K.P. Vijayakumar ‘, K.G.M. Nair bT*,B. Sundarakkannan G.V. Narasimha Rao b, R. Kesavamoorthy b

b,

” Depurtment of Physics. Cochin Unirersity of Scierzce and Technology. Co&in - 682 022. Indiu h Materials Science Division. Indira Gandhi Centre .fbr Atonlic Rrsrurch. Kalpakkam - 603 103. In&r Received

3 April

1997; revised form received

21 May 1997

Abstract The effect of argon ion implantation on chemical bath deposited Cadmium sulphide (CdS) thin films is investigated by X-ray diffraction, Raman scattering and optical absorption techniques. The X-ray diffraction pattern of the As-deposited CdS thin films shows the presence of both sphalerite (cubic) and wurtzite (hexagonal) phases. Phase transition from the As-deposited mixed phase to the more stable hexagonal phase along with grain growth is observed on post implantation annealing. Optical absorption studies of the implanted films reveal a reduction in the band gap on implantation and its recovery to As-deposited values on post implantation annealing. A decrease in the intensity of the Raman peak of CdS Ai mode is seen on implantation and on post implantation annealing, the intensity is found to increase. A drastic reduction in the full width at half maximum (FWHM) value of the films subjected to post implantation annealing compared to that of As-deposited or implanted films suggests the removal of defects and strain during annealing. The peak position of the Raman mode of CdS remains more or less the same. 0 1997 Elsevier Science B.V. PACS:

85.40.R~: 8 I .05.Dz; 63.20.-e CdS; Chemical bath deposition: Implantation:

Kqword~:

Raman scattering

1. Introduction Ion implantation is technologically important in the fabrication of semiconductor devices requiring controlled dopant profiles. However, ion implantation introduces lattice damage which can significantly influence the optical and electrical *Corresponding author. Tel.: 0091 4114 40381; 4114 40360: e-mail: [email protected].

fax: 0091

0168-583X/97/317.00 0 1997 Elsevier Science B.V. All rights reserved. PIIS0lh8-58?X(97)00391-1

properties. Thermal annealing subsequent to the implantation is normally made use of to remove the lattice damage caused by ion implantation. Cadmium sulphide (CdS) is one of the many well studied II-VI semiconductors on account of its applications in solar cells and other devices [1,2]. CdS films have been grown by a variety of techniques such as thermal evaporation [3]_ sputtering [4], pulsed laser evaporation [5], chemical vapour deposition [6], and chemical bath deposition

62

K.L. Nuruyanan et al. I Nucl. Instr. and Meth. in Phyx Rex B 131 (1997) 6147

(CBD) [7]. Among these. the CBD is quite inexpensive and suitable for the production of large area thin films. The films grown by CBD technique are generally microcrystalline, made up of the cubic sphalerite phase [7], the hexagonal wurtzite phase [8], or a mixture of both the phases [9]. The metastable cubic phase gets transformed to the more stable hexagonal phase on thermal annealing [lO,l 11. There are a number of reports in the literature on the doping of CdS thin films by ion implantation [12-161. However, there is very limited information available on the role of implantation induced lattice damage on the properties of CdS thin films. Since defects generated during implantation significantly alter the optical properties, optical absorption spectrometry forms a sensitive technique to investigate the lattice disorder [17]. Raman scattering is yet another technique ideally suited for probing the implantation induced lattice damage in semiconductors. Several studies in the past [18-231 have shown that Raman scattering is a very sensitive technique for studying the lattice damage produced during implantation since the thickness of the damaged layer and the optical skin depth are essentially of the same order. The annealing effects of As+-implanted CdTe epilayers have been studied using Raman spectroscopy [24]. Yedave et al. [25] have studied ion beam induced structural transformation in hydrogenated microcrystalline silicon using X-ray diffraction and Raman spectroscopy. One of the most striking features of the Raman spectrum of CdS is the remarkable overtones series of the longitudinal optical (LO) phonons [26,27]. Both cubic and hexagonal phases of CdS give rise to the Ai (LO) phonon mode at about 305 cm-‘. Raman scattering studies of CdS in the form of thin films [28], doped glass [29], and colloid aqueous suspensions [30] have been reported. Effects observed during implantation arise from (i) the electrical activation produced by the implanted species and (ii) the lattice disorder introduced by implantation. In the present study, we chose argon ion implantation as argon is not electrically active in CdS and as a result the effects observed can be attributed exclusively to lattice disorder. Understanding of these effects helps in delineating the effects of lattice disorder from that

of electrical activation when electrically active species are implanted. In the present paper, we report the investigations on the As-deposited, argon ion implanted and annealed CdS thin films prepared by CBD using X-ray diffraction, Raman spectroscopy and optical absorption spectrometry.

2. Experimental details CdS thin films were prepared by CBD technique. The deposition was carried out by the decomposition of cadmium chloride (1 M) and thiourea (1 M) in alkaline medium to yield cadmium and sulphur ions. Triethanolamine was added to control the release of the cadmium ions. The bath temperature during deposition was maintained at 333 K and the pH of the solution was slightly above 10. The glass substrates onto which the CdS thin films were to be deposited were washed with detergent solution, then with acid and base solutions, and finally was ultrasonically cleaned in deionized water and dried. Thin films of CdS were deposited on these glass substrates kept vertically in the bath. The thickness of CdS thin films was measured using a Sloan Dek-Tak 3030 surface profilometer. Mass analysed beam of 80 keV Ar+ ions from a low energy accelerator was used for carrying out the implantations. The beam current was maintained around 0.5-0.8 uA to avoid the local heating effect during ion implantation. The implantation was carried out at room temperature and at a vacuum of the order of lo-’ mbar to a dose of 5 x lOI ions/cm’. Post implantation annealing was carried out at 673 K for 2 h in flowing argon atmosphere. The unpolarised Raman spectra of CdS thin films were recorded in a Raman spectrometer built around a double grating monochromator, SPEX model 14018. An argon laser, lasing at 4880 A with 50 mW power was used as the source. This beam was focussed on the CdS thin films and the scattered light in the backscattering geometry was collected using a camera lens and a focussing lens. A thermoelectrically cooled photomultiplier tube model ITT-FW 130 was used to detect the scattered light. A slit width of the monochromator

K.L. Nurqanan

63

et al. I Nucl. Instr. and Meth. in Phys. Rex B 132 (1997) 6147

corresponding to 4.2 cm-’ in terms of full width at half maximum (FWHM) of the resolution function was employed throughout. The spectra were recorded digitally using a microprocessor based automated data collection system with a step of 0.5 cm-’ and a collection time of 5 s. X-ray diffraction(XRD) pattern of these films was recorded using a Siemens D-500 X-ray diffractometer on the samples using Cu-K, radiation. Optical absorption studies were carried out on these films at room temperature in the range of 4001100 nm using the Chimito UV-VIS spectrometer to measure the optical band gap of the material.

6;‘500

20

30 20

40

50

60

(DEGREES)

M 0

'

5

I

3. Results and discussions 3.1. As-deposited film

The thickness of the films was found to be 1.3 urn. Fig. l(a) shows the X-ray diffraction pattern of the As-deposited CdS thin film prepared by CBD technique. It is found from the figure that the films are crystalline and have both cubic (sphalerite) and hexagonal (wurtzite) phases. The peaks in the diffraction pattern are quite broad, possibly due to small crystallite sizes. In Fig. l(a), the peak at 28.4” belongs to thiourea (TU). The thiourea contamination in CdS thin films prepared by the CBD technique is inevitable as it is one of the main ingredients in the bath. However, thiourea is volatile and gets removed during annealing at high temperature. A typical Raman spectrum of the As-deposited CdS thin film is shown in Fig. 2(a). The Raman peak arises from the AI mode of CdS. A lorentzian line shape is fitted to the Raman spectrum from which the peak position. FWHM and the area under the Raman peak have been obtained. No asymmetry in the lower wavenumber side of the peak is seen in the spectrum. The observed Raman spectrum is the convolution between the true Raman spectrum and the spectrometer resolution function having a FWHM of 4.2 cm-‘. The true FWHMs of the Raman modes and the true height were deduced from the observed values using the empirical relations given by Arora and Umadevi [311.

0 20

I I L 30 40 50 20 (DEGREES)

t0

(cl i;; ki

5 2* 600 s t; 5:

400

2 E -

0 20

30 20

Fig. 1. (a) X-ray film prepared by for the diffraction planted CdS thin films subjected to

40

50

60

(DEGREES)

diffraction pattern of As-deposited CdS thin CBD technique. (Cu-Ku radiation was used studies) (b) X-ray diffraction pattern of imfilm. (c) X-ray diffraction pattern of CdS thin post implantation annealing.

Raman peak of CdS thin film prepared by CBD technique appears at 300.3 cm-’ and the true FWHM is 19.14 cm-’ whereas in the vacuum eva-

K.L. Narayanan et al. I Nucl. Ins@. and Meth. in Phys. Rex B I32 (1997) 61-67

E -

I

10

350 300 250 RAMAN SHIFT (CM-') (b)

porated CdS films, the AI mode appeared at about 303 cm-’ with FWHM of 8.3 cm-’ [32]. The low value of peak position and the high value of FWHM in the CBD CdS thin films as compared to those of vacuum evaporated CdS thin films are probably due to poor crystalline quality, It is reported that the defect-free crystalline CdS films result in 305 cm-’ peak position and about 8 cm-’ FWHM [33]. Concerning the influence of cubic and hexagonal phases of CdS on Raman spectrum, it has been reported that it is very difficult to distinguish between these two modifications, the Al(L0) mode being common in both cases [34]. The Raman frequency of CBD CdS thin films fall in the range as appeared in the earlier reports [X,27,35]. In Fig. 3, curve a shows a typical optical absorption spectrum of As-deposited CdS thin film. CdS is a direct band gap semiconductor and for a direct allowed transition, the optical absorption coefficient, cx,is given as hc y-Es

c(=ao

[

250

350 300 RAMAN SHIFT (CM-')

25 3 g 20

250

/”

1 I/?

.

(1)

where CQis a constant, /1 the wavelength of the incident light, c the velocity of light, Eg the band gap of the material, and h the Planck constant. The optical band gap (E,) of the CdS film can be deter-

350 300 RAbfAN SHIFT (CM-')

Fig. 2. (a) Raman spectrum of the As-deposited CdS thin film. The spectrum is fitted to a lorentzian function (dashed line indicates the fitted curve) (b) Raman spectrum of the implanted CdS thin film. The spectrum is fitted to a Iorentzian function (dashed Iine indicates the fitted curve) (c) Raman spectrum of the CdS films subjected to post implanted annealing. The spectrum is fitted to a lorentzian curve (dashed line indicates the fitted curve).

400

600 800 1000 WAVELENGTH (nm)

Fig. 3. Optical absorption spectrum of (a) As-deposited, (b) 5 x 10’” ions/cm’ Ar- implanted and (c) subsequently annealed CdS thin films.

PL.

1.0

1.5

Naravanan et al. I Nucl. Instr. and Meth. ill Phys. Rex B 13.? I 1997) 61-67

2.0 ENERGY

2.5

3.0

3.5

(eV)

Fig. 4. A plot between x2 and energy for the (a) As-deposited, (b) 5 x 10” ions/cm’ Ar’ implanted and (c) subsequently annealed CdS thin films.

mined by plotting a graph between a2 and hv and by extrapolating the linear portion of the curve to the X-axis (Fig. 4). The As-deposited sample is found to have an optical band gap of 2.35 eV. 3.2. l$fkt

of imphntation

The samples were bombarded with argon ions of SO keV at room temperature to a dose of 5 x lO’(’ ions/cm’. Fig. 2(b) shows the Raman spectrum of the argon implanted CdS thin film and its fit to a lorentzian line shape. On argon implantation, the intensity of the Raman peak has come down drastically as compared to that of the As-deposited film. This can be attributed to the ion implantation induced lattice damage. The FWHM of the implanted film (18.31 cm-‘) remained more or less the same as that ofAs-deposited film. The peak position increased from 300.3 cm-’ in the As-deposited film to 301.2 cm-’ in the implanted film. This marginal increase in the peak position is not significant. Fig. l(b) shows the X-ray diffraction pattern of the argon implanted CdS thin film. It shows that the implanted CdS films have mixed phases of cubic sphalerite and hexagonal wurtzite structures. Thiourea peak also remains on implantation. In addition, there is a peak at 38.4”. From careful analysis of the XRD pattern, it is seen that a peak appearing at 20 = 38.4” matches well with Cd as

65

well as Cd0 and hence it is difficult to assign the peak. However, the formation of Cd0 phase is unlikely because properly mass analysed argon ion beam was used for the implantation and the vacuum during the implantation was better than lo-’ mbar. Mady et al. [36] have investigated the effect of nitrogen implantation in ZnS films and observed the formation of zinc clusters on nitrogen irradiation. On argon implantation, there is a possibility of loss of sulphur atoms, and cadmium clusters could form in the CdS matrix. Hence it is probable that the peak at 30 = 38.4” is due to metallic cadmium clusters. Implantation with Ar+ ions produces vacancies and interstitials causing lattice damage. Defect clusters such as argon bubbles, dislocation loops, etc., also form during implantation. Ion irradiation induced lattice disorder changes the optical properties of the system. On implantation, defect levels are produced in the band gap. In Fig. 3. the curve b shows optical absorption spectrum of ion implanted CdS thin film. Curve b in Fig. 4 shows r2 versus the photon energy of incident light. The extrapolation of the linear portion of this curve gives the band gap of 1.75 eV in ion implanted film. The observed reduction in the band gap is attributed to the effect of band tailing due to the defects produced during implantation. 3.3. @f&t of’ post implantation

annruling

Fig. 2(c) shows the Raman spectrum of CdS thin films subjected to post implantation annealing fitted to a lorentzian line shape. The Raman mode peaks at 301.7 cm- ’ and has a FWHM of 12.45 cm-‘. The Raman peak position does not change much on annealing. It just increased from 301.3 cm-’ in the implanted film to 301.7 cm-’ in the post implanted annealed film. The peak position does not change significantly on implantation or on annealing. It is observed that the relative intensity of the Ai mode of CdS film has increased drastically when compared to that of the implanted film. On annealing, some of the lattice defects are annealed out and hence the defect density reduces, which causes an increase in the Raman intensity. The FWHM of the Raman spectrum of CdS thin film subjected to post implanta-

66

K.L. Narayanan et al. I Nucl. Instr. and Meth. in Phys. Res. B 132 (1997) 6147

tion annealing is found to be around 12 cm-’ which is even smaller than that of the as-deposited film which was around 20 cm-‘. Such an observation suggests that thermal annealing not only leads to the removal of the implantation produced damage but also anneals out the defects and strain present in the As-deposited film. X-ray diffraction pattern of CdS thin film after post implantation annealing is shown in Fig. l(c). The peak that was exclusively due to the cubic phase at 28 = 30.4” disappeared completely. The peak appearing at 28= 28.4” corresponding to thiourea also disappeared. Thiourea being volatile would have been evaporated at the annealing temperature of 673 K. Many new peaks exclusively attributable to the hexagonal phase have appeared in the diffraction pattern of the films subjected to post implantation annealing. These observations suggest that the cubic phase which is metastable has transformed to the stable hexagonal phase. Previous investigators have reported similar transition from cubic to hexagonal phase on thermal annealing [lO,l 11. In addition we observed that on annealing the implanted films, there is a pronounced sharpening of the peaks probably as a consequence of grain growth. The optical absorption spectrum of the CdS thin film subjected to post implantation annealing is shown as curve c in Fig. 3. It is observed that absorption coefficient falls in between those of as-deposited and implanted films. The absorption coefficient increases on implantation, but decreases on annealing. This tendency of decrease of absorption coefficient on post implantation annealing indicates the removal of strain and defects present in the CdS matrix. The band gap of the films after post implantation annealing is found to be around 2.3 eV. There is a recovery of the band gap value on post implantation annealing treatments from 1.75 to 2.3 eV.

4. Conclusion The CdS thin films, prepared by CBD technique, were characterised using the optical absorption spectrometry, Raman spectroscopy and Xray diffraction. The optical absorption measure-

ments reveal that the band gap of the As-deposited CdS film is 2.35 eV. On implantation, it reduces to 1.75 eV, due to the production of defect levels in the band gap. On annealing, the band gap of the implanted films increased back to 2.3 eV. The recovery of the band gap observed is due to annealing of the defects present in the CdS matrix. Raman peak of the as-deposited CdS thin film appears at 300 cm-’ due to the CdS Ar(L0) mode. The relative intensity of the Raman peak decreases on argon implantation due to implantation induced lattice disorder. On annealing the implanted films, there is a recovery of the relative intensity of the Raman peak with no appreciable change in the peak position. X-ray diffraction studies of the Asdeposited films reveal the presence of both cubic and hexagonal phases. On annealing the implanted films, it is found that the mixed phase gets transformed into hexagonal phase accompanied by substantial grain growth. Acknowledgements

This work has been done under the IUC-DAEF Scheme and one of the authors (KLN) acknowledges Dr. R. Rajaraman for useful discussions.

References [‘I B.M. Basol, Sol. Cells. 23 (1988) 69.

121S.K.

Das. G.C. Morris, J. Appl. Phys. 73 (1993) 782. T. Nishikura, T. Suzuki. Y Ema, J. Appl. Phys. 64 (1988) 3542. F. Sanchez-Quesada. J. Vat. [41 I. Martil. G. Gonzalez-Diaz. Sci. Technol. A 2 (1984) 1491. P. Mattocks. L. PI H.S. Kwok. J.P. Zheng, S. Witanachchi, Shi. Q.Y. Ying, X.W. Wang, D.T. Shaw. Appl. Phys. Lett. 52 (1988) 1095. PI H. Uda, S. Ikegami, H. Sonomura. Jpn. J. Appl. Phys. 29 (1990) 30. [71 M. Nagao, S. Watanabe, Jpn. J. Appl. Phys. 7 (1968) 684. PI I. Kaur. D.K. Pandya. K.L. Chopra, J. Electrochem. Sot. 127 (1980) 943. P. Pramanik. Sol. Energy. [91 A. Mondal, T.K. Chaudhuri. Mater. 7 (1983) 431. 0. de Melo. J.J. PO1 0. Zelaya Angel, L.H. Emandez. Alvarado-Gil, R. Lozado-Morales, C. Falcony. H. Vargas, R. Ramirez-Bon, Vacuum 46 (1995) 1083. R. Lozado-Morales, 1111 0. Zelaya Angel, J.J. Alvarado-Gil, H. Vargas. Appl. Phys. Lett. 64 (1994) 291.

[31 T. Hayashi.

K.L. Nuruyanan et al. I Nucl. Instr. and Meth. in Phvs. Res. B 133 (1997) 6147 [12] W.W. Anderson, J.T. Mitchell, Appl. Phys. Lett. 12 (1968) 334. [13] S.L. Hou. J.A. Marley, Jr., Appl. Phys. Lett. 16 (1970) 467. [14] Y. Shiraki. T. Shimada, K.F. Komatsubara. J. Appl. Phys. 43 (1972) 710. [IS] F. Chernow. G. Rose. L. Wahhn, Appl. Phys. Len. 12 (1968) 339. [I61 N.R. Parikh. D.A. Thompson G.J.C. Carpenter. Rad. Effects 98 (1986) 289. [17] K.L. Narayanan. K.P. Vijayakumar, K.G.M. Nair, N.S. Thampi. Physica B, to be published. [18] P.S. Peercy, Appl. Phys. Lett. 18 (1971) 574. [19] S. Ushioda. Solid. State. Commun. 15 (1974) 149. [20] T. Nakamura. T. Katoda. J. Appl. Phys. 53 (1982) 5870. [21] A. Deneuville. C.H. Park, P. Ayyub, T. Anderson, P. Lowen, K. Jones. P.H. Holloway. Appl. Surf. Sci. 50 (1991) 308. [22] C.S. Ran-id Rao. S. Sundaram. R.L. Schmidt, J. Comas. J. Appl. Phys. 54 (1983) 1808. [23] S.G. Kim, H. Asahi, M. Seta, T. Takizawa, S. Emura, R.K. Soni. S. Gonda. H. Tanoue, J. Appl. Phys. 74 (1993) 579. [24] J.M. Zhang, SC. Shen, S.P. Guo, S.X. Yuan, J. Phys. D: Appl. Phys. 29 (1996) 2162.

61

[25] S.N. Yedave, S.B. Ogale, R.O. Dusane, S.V. Rajarshi. V.G. Bhide, S.T. Kshrisagar. J. Appi. Phys. 71 (1992) 3773. [26] R.C.C. Liete, J.F. Scott. T.C. Damem, Phys. Rev. Lett. 22 (1969) 780. [27] M.V. Klein. S.P.S. Porto, Phys. Rev. Lett. 22 ( 1969) 782. [28] J.F. Scott, T.C. Damem. Opt. Commun. 5 (1972) 410. [29] A.V. Baranov, Y.S. Bobovich, N.I. Grebenshchikova, V.I. Petrov. M.Y. Tsenter. Opt. Spectrosc. 60 (1986) 685. (301 R. Rosetti, S. Nakahara, L.E. Brus, J. Chem. Phys. 79 (1983) 1086. [31] A.K. Arora. V. Umadevi, Appl. Spectros. 36 (1982) 424. [32] K.L. Narayanan. K.P. Vijayakumar. R. Kesavamoorthy. K.G.M. Nair. N.S. Thampi. Solid State Phys. (India) 39C (1996) 159. [33] M. Froment. M.C. Bernard, R. Cortes. B. Mokili. D Lincot, J. Electrochem. Sot. 142 (1995) 2642. [34] D.R.T. Zahn. C. Maierhofer, A. Winter, M. Reckzugel. R. Srama, A. Thomas. K. Horn. W. Richter, J. Vat. Sci. Technol. B 29 (1991) 2206. [35] D.S. Chuu. C.M. Dai. W.F. Hsieh. C.T. Tsai. J. Appl. Phys. 69 (1991) 8402. [36] KH.A. Mady, Z.S. Moustafa. M.S. Selim. A.A. Gabr. J. Mater. Sci. 23 (1988) 2403.