- Email: [email protected]
Effect of fluorine doping on structural, electrical and optical properties of ZnO thin films
Materials Science and Engineering B 117 (2005) 307–312
Effect of fluorine doping on structural, electrical and optical properties of ZnO thin films P.M. Ratheesh Kumara , C. Sudha Karthaa , K.P. Vijayakumara,∗ , F. Singhb , D.K. Avasthib a
Thin Film Photovoltaic Division, Department of Physics, Cochin University of Science and Technology, Kochi 682022, India b Nuclear Science Centre, Aruna Asaf Ali Marg, New Delhi 110067, India Received 20 August 2004; received in revised form 22 December 2004; accepted 22 December 2004
Abstract We studied effect of fluorine doping on ZnO thin films prepared using chemical spray pyrolysis technique. Structural, electrical and optical characterization techniques were used for the study. Both doped and undoped films showed preferential orientation along (0 0 2) plane. Optical absorption studies indicated a decrease of energy band gap and optical transmission increased with increase in fluorine doping. PL spectrum of undoped sample has a single broad peak at 517 nm while three additional peaks appeared for fluorine-doped film. Electrical resistivity and photosensitivity decreased considerably on doping. Temperature dependent conductivity studies revealed that prominent shallow donor levels are Zn in interstitial and regular lattice positions. © 2004 Elsevier B.V. All rights reserved. Keywords: Thin films; Zinc oxide; Doping effects; Optical properties
1. Introduction Zinc oxide (ZnO) has wide range of technological applications as transparent conducting electrodes in solar cells, flat panel displays, surface acoustic wave devices and sensors. Moreover there is growing interest in this material recently, due to its potential applications in optoelectronic devices such as light emitting diodes and laser diode having emission in short wavelength region due to its large band gap. Very recently several groups in different research labs are trying to get p-type ZnO through different techniques. Research work in this direction is still in its infancy [1,2]. For wide band gap semiconductors, addition of impurities often induces dramatic change in their electrical and optical properties. Physical properties of undoped and doped ZnO films have been widely reported. However there are not much publications describing variations in luminescence properties, photoresponse and defect states induced by impurity
∗
Corresponding author. Tel.: +91 484 2577404; fax: +91 484 2577595. E-mail address: [email protected] (K.P. Vijayakumar).
0921-5107/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2004.12.040
atoms. In this paper, we report variations in photoluminescence, electrical conductivity, photosensitivity, structure and optical properties of ZnO films due to fluorine doping. In spite of extensive studies on preparation, properties and effects of dopants on the properties of ZnO, certain effects of either some dopants or preparation procedures are still remaining unclear. Unlike the effects of indium and other group III elements in ZnO, effect of fluorine is less discussed. Sanchez et al. reported the role of F/Zn ratio on structural and optical properties of ZnO thin films prepared using chemical spray pyrolysis technique [3]. They also studied variation of refractive index with fluorine concentration [4]. Olvera et al. studied effects of fluorine concentration, substrate temperature and acidity of spray solution on electrical and optical properties of ZnO films [5,6]. Here starting solution was zinc acetyl acetonate and ammonium fluoride. Santiago et al. reported effect of ageing time of starting solution and substrate temperature on the properties of F-doped ZnO films [7]. They used hydrofluoric acid for fluorine doping. Hichou et al. reported the cathodoluminescence characteristics of spray pyrolysed F-doped ZnO thin film using ZnCl2
308
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
as starting solution [8]. Recently Kityk et al. reported a giant linear electro-optic (Pockels) effect in ZnO doped with fluorine and they interpreted the effect as due to substantial non-centrosymmetric charge density distribution between the ZnO wurtzite like films and the bare glass substrate as well as by additional charge density polarization caused by fluorine atoms [9]. Our aim was to characterize the fluorine doped ZnO (F:ZnO) thin films using different techniques. We characterized F:ZnO films with XRD, optical absorption and transmission, photoluminescence, photosensitivity and temperature dependent electrical conductivity measurements. To the best our knowledge, not much publications are there on photoluminescence properties, temperature dependence of conductivity and photoresponse of F:ZnO thin films. Moreover, we could study the variations of band gap of ZnO thin films due to fluorine doping.
2. Experimental details ZnO thin films were prepared using chemical spray pyrolysis (CSP) technique. Zinc acetate, dissolved in equal volume of water and ethanol, was used as precursor solution. Required quantity of ammonium fluoride was dissolved in this solution for doping with fluorine. Volume of sprayed solution was 200 ml. Temperature of the substrate (400 ◦ C) and spray rate (10 ml/min) were fixed. In aqueous media, Zn can precipitate in the form of zinc hydroxide (Zn(OH)2 ). Acetic acid was added (5 ml) to the solution in order to prevent the formation of milky Zn(OH)2 . Quantity of acetic acid added to the solution was also a key parameter for the film deposition. Depending upon the pH of the solution, Zn2+ was progressively converted into complexes, which were Zn(Ac)+ and Zn(Ac)2 . Concentration of these species played crucial role in the film formation. It is preferable to have neutral zinc acetate (Zn(Ac)2 ) in solution in order to get film deposition [10]. But addition of acetic acid locks the pH in the range of 4–5, resulting Zn(Ac)2 as dominating species in the solution, enhancing the film deposition. Structural properties of intrinsic and fluorine doped samples were determined through XRD studies (Rigaku ˚ (D.Max.C)) having Cu K␣ radiation with λ = 1.5405 A. Optical properties were studied using optical absorbance and transmittance spectrum (UV–vis NIR spectrophotometerHitachi U-3410 model). Photosensitivity and resistivity measurements were performed using Keithley 236 Source Measure Unit. Electrical contacts were made using silver paint in the form of two end contacts with a distance of 5 mm between them. For photosensitivity measurements, sample was illuminated using tungsten halogen lamp of power 100 mW/cm2 . Photoluminescence (PL) studies were performed using excitation with 325-nm line from Kimmon He–Cd laser and Michelle 900 spectrograph.
3. Results and discussion 3.1. Structural characterization Fig. 1 depicts XRD pattern of both undoped and F-doped samples. All samples showed preferential orientation along (0 0 2) plane. The (0 0 2) direction corresponds to the c-axis of the crystal lattice, normal to the substrate plane. A weak XRD peak corresponding to the plane (1 0 1) was also observed for the undoped and 0.5 at% doping samples. This peak completely vanished for the sample with 1 at% F. One can assume the substitution process of O by F species and this process could only be partial leading to the formation of a specific configuration like ZnFx O1−x [8]. Probably this could be the reason for reduced intensity of XRD peak of F:ZnO. Mean crystallite size was calculated for the (0 0 2) diffraction peak, using Scherrer formula D = (0.9λ)/(β cos θ), where D is the diameter of crystallites, λ the wavelength of Cu K␣ line, β is full width at half maximum (FWHM) in radians and θ is Bragg’s angle. The values were found to decrease with increase in fluorine percentage and are tabulated (Table 1). Higher doping concentration did not have considerable effect on the grain size. 3.2. Optical absorption and transmission studies Optical absorption spectra of doped and undoped samples were recorded in the wavelength range 350–900 nm and are
Fig. 1. XRD pattern of doped and undoped films.
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
309
Table 1 Variation in grain size, electrical resistivity and photosensitivity with doping concentration Sample name
Grain ˚ size (A)
Resistivity ( cm)
Photosensitivity
Undoped F:ZnO (0.5%) F:ZnO (1%) F:ZnO (2%) F:ZnO (5%)
158 150 135.7 133.5 128
1298 59.9 43 41.3 151.8
200 4.19 2.83 4.5 11.2
depicted in Fig. 2. Optical band gap was determined from the plot of (αhν)2 versus hν graph (Fig. 3). Undoped sample had band gap of 3.3 eV. We observed band-gap narrowing (BGN) for the F:ZnO samples. Band gap was reduced to 3.19 eV for the sample having 0.5 doping and 3.1 eV for the one having 1% doping. Again it increased to 3.18 eV for 2% doping and then decreased to 3.09 eV for 5% (not shown in figure) doping. There were two broad peaks corresponding to two defect levels in absorption spectrum of fluorinedoped samples (Fig. 2). In 0.5% doping these levels were at 476 nm and 646 nm, while for 1% doping these peaks are at 455 nm and 586 nm. The blue shift observed for the 1% doped sample was not clearly understood. The appearance of two broad absorption peaks centred at 455 and 476 correspond the presence of large number of defect levels just below conduction band. Probably these defects levels near to the conduction band might be the reason for the small reduction in band gap. Absorption peak at 650 nm was observed for the sample doped with 2% F. Photoluminescence measurements also showed two emissions at 455 nm and 670 nm
Fig. 2. Absorption spectrum of doped and undoped samples.
Fig. 3. (αhν)2 vs. hν plot of doped and undoped ZnO thin films.
for the doped sample, which will be discussed in detail in Section 3.4. Optical transmission spectra were recorded in the wavelength range 350–1500 nm (Fig. 4). Both undoped and doped films showed interference fringe pattern in transmission
Fig. 4. Transmission spectra of the films.
310
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
spectrum. This revealed the smooth reflecting surfaces of the film and there was not much scattering loss at the surface. Interestingly, doped sample exhibited increased optical transmission in the visible and NIR region and this was good for device fabrication. In transparent metal oxides, metal to oxygen ratio decides the percentage of transmittance. A metal rich film usually exhibits less transparency [11]. Fluorine doped samples were showing higher transmittance than the undoped samples. This may be due to the decrease in the Zn/[O + F] ratio in the film. Amplitude of interference fringes decreased for higher doping concentration and this indicated the loss surface smoothness leading to a slight scattering loss.
current. There was significant decrease in photosensitivity on fluorine doping which may be due to the increase in majority carrier concentration. For higher doping concentration, as resistivity was high, the films showed an increase in photosensitivity. 3.4. Photoluminescence measurements
Electrical resistivity was measured using two-probe method. It was found that resistivity decreased considerably with fluorine doping (103 –101 cm) (Table 1). Variations in resistivity and photosensitivity are plotted (Fig. 5). This might be due to the replacement of oxygen by fluorine. This was also clear from the XRD results. Fluorine required only one electron for making covalent bond with zinc and the other electron was set free which enhanced the conductivity. Higher doping concentration resulted an increase in the resistivity. When doping concentration increased, fluorine turned into interstitials and since fluorine has more electro negativity than oxygen, it might cause an increase in the electrical resistivity. Photosensitivity was calculated using the formula (IL − ID )/ID , where IL is illuminated current and ID is dark
Photoluminescence measurements were done on both doped and undoped samples at room temperature (Fig. 6). Undoped ZnO film gave single broad emission peak centered at 517 nm (2.4 eV). It was the characteristic blue green emission of ZnO thin films and origin of which is still controversial. Exact mechanism responsible for blue–green emission is still not clearly understood. It had been suggested that this emission might be due to the substitution of Zn2+ by Cu2+ [12,13]. Another suggestion was that it might be related to radiative transition between deep acceptors and shallow donors [14]. Kang et al. suggested that the green emission might be due to the transition from deep donor levels of oxygen vacancies to valence band [15]. In the present case, no copper was added to our samples intentionally. Thus the green emission could not be related to Cu impurity. Using full-potential linear muffin–tin orbital method, the energy levels of intrinsic defects were calculated [16]. It was showed that the energy interval from the bottom of conduction band to the antisite oxygen (OZn ) level was 2.38 eV [17]. On fluorine doping, intensity of this peak was getting decreased. When fluorine was incorporated, this may be replacing oxygen. Hence one could assume that the level corresponding to the antisite oxygen
Fig. 5. Variation in resistivity and photosensitivity with doping concentrations.
Fig. 6. Photoluminescence spectra of doped and undoped films.
3.3. Electrical resistivity and photosensitivity measurements
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
(OZn ) was also getting depleted on fluorine doping and we suggest that the green emission may be due to the radiative transition from conduction band to the OZn level. PL emission spectrum of F:ZnO exhibited four emissions at λ = 405, 455, 517 and 670 nm. The violet emission at 405 nm (3.07 eV) might be due to the radiative transitions from defect levels related to interface traps existing at the grain boundaries to valence band [18]. We found that relative intensity of emission at 405 nm, increased with fluorine doping. This may be due to the fact that grain size of the films was getting reduced with increase in fluorine doping making grain boundary area larger and hence the intensity was higher for higher doping concentration. Another emission was at 455 nm (2.73 eV). The relative intensity of this emission also increased on fluorine doping. This emission might be due to the lattice modification of Zn2+ environment in the film because of the incorporation of fluorine atoms [8]. Peak at ∼670 nm (1.85 eV), might be corresponding to the transition from the levels of oxygen interstitial [19]. This was prominent for the sample having 1% doping and was not seen in undoped and 0.5% doped samples. This gives a strong evidence for the replacement of oxygen by fluorine. This oxygen might be occupying the interstitial position, giving red emission at 670 nm.
311
and 65 meV) for the undoped samples. The existence of an intrinsic donor level, 150–170 meV below the conduction band was reported earlier [20]. There were also reports on theoretical calculations of another native donor with energy 61 meV that originated from octahedral zinc interstitials [21]. These values were in good agreement with our experimental data. When ZnO was doped with fluorine, intrinsic donor level at 148 meV got suppressed. Instead activation energies 89 and 30 meV were obtained for 0.5% doped samples and a slightly different set of values (65 and 22 meV) were obtained for 1% doped samples. Also activation energies, 65 and 33 meV were obtained for 5% doped samples. Here we can assume reasonably well that the higher values (89 and 65 meV) correspond to octahedral zinc interstitials while the lower values (30 and 22 meV) correspond to zinc atoms in tetrahedral positions [21,22]. The former one was present in both undoped and doped samples while latter one became evident only in doped samples. This may be resulting from fluorine doping. As fluorine enters into ZnO lattice, zinc in regular lattice site is (tetrahedral) also becoming a donor, as only one electron is needed for bonding.
4. Conclusion 3.5. Temperature dependent conductivity measurements Temperature dependence on electrical conductivity of the samples was measured in the range 100–300 K (Fig. 7). Sample was placed on cold finger of liquid helium cryostat using thermal grease to avoid any thermal gradient along or across the sample. Cryostat was evacuated to a pressure of 10−5 mbar. We obtained two activation energies (148 meV
Fluorine doping leads to considerable reduction in electrical resistivity, enhancement of optical transmission, and reduction in photosensitivity. Reduction in resistivity may be due to contribution of electron from Zn at tetrahedral sites also. PL studies revealed levels of interstitial oxygen. Moreover, levels due to antisite oxygen are getting depleted on fluorine doping. Hence electrical and optical properties of ZnO film can be very well modified through fluorine doping.
Acknowledgment One of the authors (PMR) is grateful to Ministry of Nonconventional Energy Sources for providing research fellowship and another author (KPV) is thankful to Cochin University of Science and Technology for granting sabbatical leave for conducting this study. The authors would like to thank UGC for financial assistance through the DSA COSIST programme.
References
Fig. 7. Variation of conductivity with temperature for different fluorine concentration.
[1] J.M. Bian, X.M. Li, X.D. Gao, W.D. Yu, L.D. Chen, Appl. Phys. Lett. 84 (2004) 541. [2] D.C. Look, B. Claflin, Phys. Stat. Sol. (b) 241 (2004) 624. [3] A. Sanchez-Juarez, A. Tiburcio-Silver, A. Ortiz, Sol. Energy Mat. Sol. Cells 52 (1998) 301. [4] A. Sanchez-Juarez, A. Tiburcio-Silver, A. Ortiz, E.P. Zironi, J. Rickards, Thin Solid Films 333 (1998) 196. [5] M. de la, L. Olvera, A. Maldonado, R. Asomoza, M. Melendez-Lira, Sol. Energy Mat. Sol. Cells 71 (1998) 301.
312
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
[6] M. de la, L. Olvera, A. Maldonado, R. Asomoza, Sol. Energy Mat. Sol. Cells 73 (2002) 425. [7] A. Guillen-Santiago, M. de la L. Olvera, A. Maldonado, R. Asomoza, D.R. Acosta, Phys. Stat. Sol. 201 (2004) 952. [8] A. El Hichou, A. Bougrine, J.L. Bubendorff, J. Ebothe, M. Addou, M. Troyon, Semicond. Sci. Technol. 17 (2002) 607. [9] I.V. Kityk, J. Ebothe, A. El Hichou, M. Addou, A. Bougrine, B. Sahraoui, J. Phys.: Condens. Matter. 14 (2002) 5407. [10] A. Smith, Thin Solid Films 376 (2000) 47. [11] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, IOP Publishing, Bristol and Philadelphia, 1995, p. 244. [12] R. Dingle, Phys. Rev. Lett. 23 (1969) 579.
[13] N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, G. Cantwell, Appl. Phys. Lett. 81 (2001) 2285. [14] E.G. Bylander, J. Appl. Phys. 49 (1978) 1188. [15] H.S. Kang, J.S. Kang, J.W. Kim, S.Y. Lee, J. Appl. Phys. 95 (2004) 1246. [16] Y.M. Sun, Ph.D. thesis, University of Science and Technology of China, 2000. [17] B. Lin, Z. Fu, Y. Jia, Appl. Phys. Lett. 79 (2001) 943. [18] B.J. Jin, S. Im, S.Y. Lee, Thin Solid Films 366 (2000) 107. [19] Y.G. Wang, S.P. Lau, X.H. Zhang, H.W. Lee, H.H. Hng, B.K. Tay, J. Cryst. Growth 252 (2003) 265. [20] V. Srikant, D.R. Clarke, J. Appl. Phys. 83 (1998) 5447. [21] Y. Sun, H. Wang, Physica B 325 (2003) 157. [22] D.C. Look, J.W. Hemsky, J.R. Sizelove, Phys. Rev. Lett. 82 (1999) 2552.
Effect of fluorine doping on structural, electrical and optical properties of ZnO thin films P.M. Ratheesh Kumara , C. Sudha Karthaa , K.P. Vijayakumara,∗ , F. Singhb , D.K. Avasthib a
Thin Film Photovoltaic Division, Department of Physics, Cochin University of Science and Technology, Kochi 682022, India b Nuclear Science Centre, Aruna Asaf Ali Marg, New Delhi 110067, India Received 20 August 2004; received in revised form 22 December 2004; accepted 22 December 2004
Abstract We studied effect of fluorine doping on ZnO thin films prepared using chemical spray pyrolysis technique. Structural, electrical and optical characterization techniques were used for the study. Both doped and undoped films showed preferential orientation along (0 0 2) plane. Optical absorption studies indicated a decrease of energy band gap and optical transmission increased with increase in fluorine doping. PL spectrum of undoped sample has a single broad peak at 517 nm while three additional peaks appeared for fluorine-doped film. Electrical resistivity and photosensitivity decreased considerably on doping. Temperature dependent conductivity studies revealed that prominent shallow donor levels are Zn in interstitial and regular lattice positions. © 2004 Elsevier B.V. All rights reserved. Keywords: Thin films; Zinc oxide; Doping effects; Optical properties
1. Introduction Zinc oxide (ZnO) has wide range of technological applications as transparent conducting electrodes in solar cells, flat panel displays, surface acoustic wave devices and sensors. Moreover there is growing interest in this material recently, due to its potential applications in optoelectronic devices such as light emitting diodes and laser diode having emission in short wavelength region due to its large band gap. Very recently several groups in different research labs are trying to get p-type ZnO through different techniques. Research work in this direction is still in its infancy [1,2]. For wide band gap semiconductors, addition of impurities often induces dramatic change in their electrical and optical properties. Physical properties of undoped and doped ZnO films have been widely reported. However there are not much publications describing variations in luminescence properties, photoresponse and defect states induced by impurity
∗
Corresponding author. Tel.: +91 484 2577404; fax: +91 484 2577595. E-mail address: [email protected] (K.P. Vijayakumar).
0921-5107/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.mseb.2004.12.040
atoms. In this paper, we report variations in photoluminescence, electrical conductivity, photosensitivity, structure and optical properties of ZnO films due to fluorine doping. In spite of extensive studies on preparation, properties and effects of dopants on the properties of ZnO, certain effects of either some dopants or preparation procedures are still remaining unclear. Unlike the effects of indium and other group III elements in ZnO, effect of fluorine is less discussed. Sanchez et al. reported the role of F/Zn ratio on structural and optical properties of ZnO thin films prepared using chemical spray pyrolysis technique [3]. They also studied variation of refractive index with fluorine concentration [4]. Olvera et al. studied effects of fluorine concentration, substrate temperature and acidity of spray solution on electrical and optical properties of ZnO films [5,6]. Here starting solution was zinc acetyl acetonate and ammonium fluoride. Santiago et al. reported effect of ageing time of starting solution and substrate temperature on the properties of F-doped ZnO films [7]. They used hydrofluoric acid for fluorine doping. Hichou et al. reported the cathodoluminescence characteristics of spray pyrolysed F-doped ZnO thin film using ZnCl2
308
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
as starting solution [8]. Recently Kityk et al. reported a giant linear electro-optic (Pockels) effect in ZnO doped with fluorine and they interpreted the effect as due to substantial non-centrosymmetric charge density distribution between the ZnO wurtzite like films and the bare glass substrate as well as by additional charge density polarization caused by fluorine atoms [9]. Our aim was to characterize the fluorine doped ZnO (F:ZnO) thin films using different techniques. We characterized F:ZnO films with XRD, optical absorption and transmission, photoluminescence, photosensitivity and temperature dependent electrical conductivity measurements. To the best our knowledge, not much publications are there on photoluminescence properties, temperature dependence of conductivity and photoresponse of F:ZnO thin films. Moreover, we could study the variations of band gap of ZnO thin films due to fluorine doping.
2. Experimental details ZnO thin films were prepared using chemical spray pyrolysis (CSP) technique. Zinc acetate, dissolved in equal volume of water and ethanol, was used as precursor solution. Required quantity of ammonium fluoride was dissolved in this solution for doping with fluorine. Volume of sprayed solution was 200 ml. Temperature of the substrate (400 ◦ C) and spray rate (10 ml/min) were fixed. In aqueous media, Zn can precipitate in the form of zinc hydroxide (Zn(OH)2 ). Acetic acid was added (5 ml) to the solution in order to prevent the formation of milky Zn(OH)2 . Quantity of acetic acid added to the solution was also a key parameter for the film deposition. Depending upon the pH of the solution, Zn2+ was progressively converted into complexes, which were Zn(Ac)+ and Zn(Ac)2 . Concentration of these species played crucial role in the film formation. It is preferable to have neutral zinc acetate (Zn(Ac)2 ) in solution in order to get film deposition [10]. But addition of acetic acid locks the pH in the range of 4–5, resulting Zn(Ac)2 as dominating species in the solution, enhancing the film deposition. Structural properties of intrinsic and fluorine doped samples were determined through XRD studies (Rigaku ˚ (D.Max.C)) having Cu K␣ radiation with λ = 1.5405 A. Optical properties were studied using optical absorbance and transmittance spectrum (UV–vis NIR spectrophotometerHitachi U-3410 model). Photosensitivity and resistivity measurements were performed using Keithley 236 Source Measure Unit. Electrical contacts were made using silver paint in the form of two end contacts with a distance of 5 mm between them. For photosensitivity measurements, sample was illuminated using tungsten halogen lamp of power 100 mW/cm2 . Photoluminescence (PL) studies were performed using excitation with 325-nm line from Kimmon He–Cd laser and Michelle 900 spectrograph.
3. Results and discussion 3.1. Structural characterization Fig. 1 depicts XRD pattern of both undoped and F-doped samples. All samples showed preferential orientation along (0 0 2) plane. The (0 0 2) direction corresponds to the c-axis of the crystal lattice, normal to the substrate plane. A weak XRD peak corresponding to the plane (1 0 1) was also observed for the undoped and 0.5 at% doping samples. This peak completely vanished for the sample with 1 at% F. One can assume the substitution process of O by F species and this process could only be partial leading to the formation of a specific configuration like ZnFx O1−x [8]. Probably this could be the reason for reduced intensity of XRD peak of F:ZnO. Mean crystallite size was calculated for the (0 0 2) diffraction peak, using Scherrer formula D = (0.9λ)/(β cos θ), where D is the diameter of crystallites, λ the wavelength of Cu K␣ line, β is full width at half maximum (FWHM) in radians and θ is Bragg’s angle. The values were found to decrease with increase in fluorine percentage and are tabulated (Table 1). Higher doping concentration did not have considerable effect on the grain size. 3.2. Optical absorption and transmission studies Optical absorption spectra of doped and undoped samples were recorded in the wavelength range 350–900 nm and are
Fig. 1. XRD pattern of doped and undoped films.
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
309
Table 1 Variation in grain size, electrical resistivity and photosensitivity with doping concentration Sample name
Grain ˚ size (A)
Resistivity ( cm)
Photosensitivity
Undoped F:ZnO (0.5%) F:ZnO (1%) F:ZnO (2%) F:ZnO (5%)
158 150 135.7 133.5 128
1298 59.9 43 41.3 151.8
200 4.19 2.83 4.5 11.2
depicted in Fig. 2. Optical band gap was determined from the plot of (αhν)2 versus hν graph (Fig. 3). Undoped sample had band gap of 3.3 eV. We observed band-gap narrowing (BGN) for the F:ZnO samples. Band gap was reduced to 3.19 eV for the sample having 0.5 doping and 3.1 eV for the one having 1% doping. Again it increased to 3.18 eV for 2% doping and then decreased to 3.09 eV for 5% (not shown in figure) doping. There were two broad peaks corresponding to two defect levels in absorption spectrum of fluorinedoped samples (Fig. 2). In 0.5% doping these levels were at 476 nm and 646 nm, while for 1% doping these peaks are at 455 nm and 586 nm. The blue shift observed for the 1% doped sample was not clearly understood. The appearance of two broad absorption peaks centred at 455 and 476 correspond the presence of large number of defect levels just below conduction band. Probably these defects levels near to the conduction band might be the reason for the small reduction in band gap. Absorption peak at 650 nm was observed for the sample doped with 2% F. Photoluminescence measurements also showed two emissions at 455 nm and 670 nm
Fig. 2. Absorption spectrum of doped and undoped samples.
Fig. 3. (αhν)2 vs. hν plot of doped and undoped ZnO thin films.
for the doped sample, which will be discussed in detail in Section 3.4. Optical transmission spectra were recorded in the wavelength range 350–1500 nm (Fig. 4). Both undoped and doped films showed interference fringe pattern in transmission
Fig. 4. Transmission spectra of the films.
310
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
spectrum. This revealed the smooth reflecting surfaces of the film and there was not much scattering loss at the surface. Interestingly, doped sample exhibited increased optical transmission in the visible and NIR region and this was good for device fabrication. In transparent metal oxides, metal to oxygen ratio decides the percentage of transmittance. A metal rich film usually exhibits less transparency [11]. Fluorine doped samples were showing higher transmittance than the undoped samples. This may be due to the decrease in the Zn/[O + F] ratio in the film. Amplitude of interference fringes decreased for higher doping concentration and this indicated the loss surface smoothness leading to a slight scattering loss.
current. There was significant decrease in photosensitivity on fluorine doping which may be due to the increase in majority carrier concentration. For higher doping concentration, as resistivity was high, the films showed an increase in photosensitivity. 3.4. Photoluminescence measurements
Electrical resistivity was measured using two-probe method. It was found that resistivity decreased considerably with fluorine doping (103 –101 cm) (Table 1). Variations in resistivity and photosensitivity are plotted (Fig. 5). This might be due to the replacement of oxygen by fluorine. This was also clear from the XRD results. Fluorine required only one electron for making covalent bond with zinc and the other electron was set free which enhanced the conductivity. Higher doping concentration resulted an increase in the resistivity. When doping concentration increased, fluorine turned into interstitials and since fluorine has more electro negativity than oxygen, it might cause an increase in the electrical resistivity. Photosensitivity was calculated using the formula (IL − ID )/ID , where IL is illuminated current and ID is dark
Photoluminescence measurements were done on both doped and undoped samples at room temperature (Fig. 6). Undoped ZnO film gave single broad emission peak centered at 517 nm (2.4 eV). It was the characteristic blue green emission of ZnO thin films and origin of which is still controversial. Exact mechanism responsible for blue–green emission is still not clearly understood. It had been suggested that this emission might be due to the substitution of Zn2+ by Cu2+ [12,13]. Another suggestion was that it might be related to radiative transition between deep acceptors and shallow donors [14]. Kang et al. suggested that the green emission might be due to the transition from deep donor levels of oxygen vacancies to valence band [15]. In the present case, no copper was added to our samples intentionally. Thus the green emission could not be related to Cu impurity. Using full-potential linear muffin–tin orbital method, the energy levels of intrinsic defects were calculated [16]. It was showed that the energy interval from the bottom of conduction band to the antisite oxygen (OZn ) level was 2.38 eV [17]. On fluorine doping, intensity of this peak was getting decreased. When fluorine was incorporated, this may be replacing oxygen. Hence one could assume that the level corresponding to the antisite oxygen
Fig. 5. Variation in resistivity and photosensitivity with doping concentrations.
Fig. 6. Photoluminescence spectra of doped and undoped films.
3.3. Electrical resistivity and photosensitivity measurements
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
(OZn ) was also getting depleted on fluorine doping and we suggest that the green emission may be due to the radiative transition from conduction band to the OZn level. PL emission spectrum of F:ZnO exhibited four emissions at λ = 405, 455, 517 and 670 nm. The violet emission at 405 nm (3.07 eV) might be due to the radiative transitions from defect levels related to interface traps existing at the grain boundaries to valence band [18]. We found that relative intensity of emission at 405 nm, increased with fluorine doping. This may be due to the fact that grain size of the films was getting reduced with increase in fluorine doping making grain boundary area larger and hence the intensity was higher for higher doping concentration. Another emission was at 455 nm (2.73 eV). The relative intensity of this emission also increased on fluorine doping. This emission might be due to the lattice modification of Zn2+ environment in the film because of the incorporation of fluorine atoms [8]. Peak at ∼670 nm (1.85 eV), might be corresponding to the transition from the levels of oxygen interstitial [19]. This was prominent for the sample having 1% doping and was not seen in undoped and 0.5% doped samples. This gives a strong evidence for the replacement of oxygen by fluorine. This oxygen might be occupying the interstitial position, giving red emission at 670 nm.
311
and 65 meV) for the undoped samples. The existence of an intrinsic donor level, 150–170 meV below the conduction band was reported earlier [20]. There were also reports on theoretical calculations of another native donor with energy 61 meV that originated from octahedral zinc interstitials [21]. These values were in good agreement with our experimental data. When ZnO was doped with fluorine, intrinsic donor level at 148 meV got suppressed. Instead activation energies 89 and 30 meV were obtained for 0.5% doped samples and a slightly different set of values (65 and 22 meV) were obtained for 1% doped samples. Also activation energies, 65 and 33 meV were obtained for 5% doped samples. Here we can assume reasonably well that the higher values (89 and 65 meV) correspond to octahedral zinc interstitials while the lower values (30 and 22 meV) correspond to zinc atoms in tetrahedral positions [21,22]. The former one was present in both undoped and doped samples while latter one became evident only in doped samples. This may be resulting from fluorine doping. As fluorine enters into ZnO lattice, zinc in regular lattice site is (tetrahedral) also becoming a donor, as only one electron is needed for bonding.
4. Conclusion 3.5. Temperature dependent conductivity measurements Temperature dependence on electrical conductivity of the samples was measured in the range 100–300 K (Fig. 7). Sample was placed on cold finger of liquid helium cryostat using thermal grease to avoid any thermal gradient along or across the sample. Cryostat was evacuated to a pressure of 10−5 mbar. We obtained two activation energies (148 meV
Fluorine doping leads to considerable reduction in electrical resistivity, enhancement of optical transmission, and reduction in photosensitivity. Reduction in resistivity may be due to contribution of electron from Zn at tetrahedral sites also. PL studies revealed levels of interstitial oxygen. Moreover, levels due to antisite oxygen are getting depleted on fluorine doping. Hence electrical and optical properties of ZnO film can be very well modified through fluorine doping.
Acknowledgment One of the authors (PMR) is grateful to Ministry of Nonconventional Energy Sources for providing research fellowship and another author (KPV) is thankful to Cochin University of Science and Technology for granting sabbatical leave for conducting this study. The authors would like to thank UGC for financial assistance through the DSA COSIST programme.
References
Fig. 7. Variation of conductivity with temperature for different fluorine concentration.
[1] J.M. Bian, X.M. Li, X.D. Gao, W.D. Yu, L.D. Chen, Appl. Phys. Lett. 84 (2004) 541. [2] D.C. Look, B. Claflin, Phys. Stat. Sol. (b) 241 (2004) 624. [3] A. Sanchez-Juarez, A. Tiburcio-Silver, A. Ortiz, Sol. Energy Mat. Sol. Cells 52 (1998) 301. [4] A. Sanchez-Juarez, A. Tiburcio-Silver, A. Ortiz, E.P. Zironi, J. Rickards, Thin Solid Films 333 (1998) 196. [5] M. de la, L. Olvera, A. Maldonado, R. Asomoza, M. Melendez-Lira, Sol. Energy Mat. Sol. Cells 71 (1998) 301.
312
P.M. Ratheesh Kumar et al. / Materials Science and Engineering B 117 (2005) 307–312
[6] M. de la, L. Olvera, A. Maldonado, R. Asomoza, Sol. Energy Mat. Sol. Cells 73 (2002) 425. [7] A. Guillen-Santiago, M. de la L. Olvera, A. Maldonado, R. Asomoza, D.R. Acosta, Phys. Stat. Sol. 201 (2004) 952. [8] A. El Hichou, A. Bougrine, J.L. Bubendorff, J. Ebothe, M. Addou, M. Troyon, Semicond. Sci. Technol. 17 (2002) 607. [9] I.V. Kityk, J. Ebothe, A. El Hichou, M. Addou, A. Bougrine, B. Sahraoui, J. Phys.: Condens. Matter. 14 (2002) 5407. [10] A. Smith, Thin Solid Films 376 (2000) 47. [11] H.L. Hartnagel, A.L. Dawar, A.K. Jain, C. Jagadish, Semiconducting Transparent Thin Films, IOP Publishing, Bristol and Philadelphia, 1995, p. 244. [12] R. Dingle, Phys. Rev. Lett. 23 (1969) 579.
[13] N.Y. Garces, L. Wang, L. Bai, N.C. Giles, L.E. Halliburton, G. Cantwell, Appl. Phys. Lett. 81 (2001) 2285. [14] E.G. Bylander, J. Appl. Phys. 49 (1978) 1188. [15] H.S. Kang, J.S. Kang, J.W. Kim, S.Y. Lee, J. Appl. Phys. 95 (2004) 1246. [16] Y.M. Sun, Ph.D. thesis, University of Science and Technology of China, 2000. [17] B. Lin, Z. Fu, Y. Jia, Appl. Phys. Lett. 79 (2001) 943. [18] B.J. Jin, S. Im, S.Y. Lee, Thin Solid Films 366 (2000) 107. [19] Y.G. Wang, S.P. Lau, X.H. Zhang, H.W. Lee, H.H. Hng, B.K. Tay, J. Cryst. Growth 252 (2003) 265. [20] V. Srikant, D.R. Clarke, J. Appl. Phys. 83 (1998) 5447. [21] Y. Sun, H. Wang, Physica B 325 (2003) 157. [22] D.C. Look, J.W. Hemsky, J.R. Sizelove, Phys. Rev. Lett. 82 (1999) 2552.