Journal of the Korean Physical Society, Vol. 53, No. 2, August 2008, pp. 680684
Investigation of CdS Films Prepared by Using Chemical Bath Deposition Arsen Babajanyan, Tigran Sargsyan and Kiejin Lee Department of Physics and Interdisciplinary Program of Integrated Biotechnology, Sogang University, Seoul 121-742
Deokjoon Cha Department of Physics, Kunsan National University, Kunsan 573-701
(Received 11 April 2008, in nal form 28 April 2008) Cadmium-sulphide (CdS) lms were prepared on glass substrates by using a chemical bath deposition method. The dependences of the microstructures and the morphologies of CdS lms on the annealing temperatures were investigated by using X-ray di raction, scanning electron microscopy and atomic force microscopy. The change in the sheet resistance due to di erent annealing temperatures was studied by using a near- eld microwave microprobe by measuring the re ection coecient S11 . As the annealing temperature increased from room temperature to 300 C, the sheet resistance of the CdS lms decreased, the surface roughness and grain size decreased and the surface showed a smoother morphology. PACS numbers: 68.37.-d, 68.37.Uv, 68.55.-a, 68.60.-p Keywords: Chemical bath deposition, CdS, Annealing, Microwave microprobe
Contactless, nondestructive characterization techniques are very useful for these applications. The NFMM technique, which directly measures the physical properties such as the surface resistance of thin lms, shows practical promise. In this paper, we report on the crystal and the band gap structure of CdS lms that were prepared by using the chemical bath deposition method and on measurement of the sheet resistance on the annealing temperature by using a NFMM. The change in the absorption spectrum of CdS lms was measured as a function of annealing temperatures. The surface roughness and the sheet resistance of CdS lms with di erent microstructures and morphologies were investigated by using X-ray di raction (XRD), scanning electron microscopy (SEM), atomic force microscopy (AFM) and a NFMM. We used a NFMM coupled to a high-quality dielectric resonator with a distance regulation system at an operating frequency f = 4.1 GHz. The changes in the sheet resistance of the CdS lms due to di erent annealing temperatures were investigated by using a NFMM to measure the microwave re ection coecient S11.
I. INTRODUCTION
Cadmium-sulphide (CdS), a member of the II-VI semiconducting compounds, is used extensively in photovoltaic and photoconducting cells, photosensors, transducers and optical detectors [1{4]. Normally asdeposited CdS lms have a carrier concentration in the range of 1014 { 1016 cm 3, a mobility of 300 cm2/Vs and a direct bandgap of 2.42 eV at room temperature [5{8]. CdS is a suitable window material for many heterojunction solar cells. Since the heterojunctions based on CdS layers are very promising structures for solar cells, a comprehensive characterization is required of CdS lms obtained under various experimental conditions. The basic requirements of these lms for solar cell applications are high optical transparency, low electrical sheet resistance and better crystallinity. Even though the sheet resistance and the optical transmittance strongly depend on the preparation conditions, pure CdS lms generally show high electrical resistivity. Thus, it is dicult to produce undoped CdS lms with good electro-optical properties just by controlling the preparation conditions. Near- eld microwave microprobe (NFMM) techniques with high sensitivity have been developed for the microwave and the millimeter-wave ranges [9{13]. An important ability of the NFMM is the contactless, nondestructive characterization of thin lms, in particular, the characterization of the electrical properties of the lms. E-mail:
II. EXPERIMENTS
[email protected]
-680-
CdS lms were prepared on glass substrates by using the chemical bath deposition (CBD) method [14{16]. CdCl2 and Thiorea (H2NCSNH2 ) were mixed with distilled water. The ratio between Cd and S was 1 : 1. The solution was prepared in quantities of about 2000 ml in a
Investigation of CdS Films Prepared by Using { Arsen Babajanyan et
-681-
al.
Fig. 1. Experimental setup of the NFMM system.
beaker. Substrates of slide glass were immersedinto the solution in the beaker and dried at around 350 C in an electric furnace for a few seconds. The glass substrate was pulled out and immersed again into the beaker solution. The processes were repeated to deposit a CdS lm on the glass substrate by pyrolysis. The thickness of the CdS lm was 2 m. After nishing the deposition process (as-grown sample), samples were annealed for 2 hours at 200 C, 300 C and 400 C to analyze crystallinity. To characterize how the electric properties of CdS lms depended on the annealing conditions, we measured the changes in the microwave re ection coecient S11 . The microstructure changes of the CdS lms were observed by using XRD. SEM and AFM were used to evaluate the surface morphology and the surface roughness of the CdS lms. The optical absorption of the CdS lms was measured in the range 400 { 600 nm by using a Sinco UV-VIs spectrometer. The experimental setup of our NFMM is described in Ref. 12 and is presented in Figure 1. The probe tips were made of stainless-steel wires with diameters of 0.05 mm. The dielectric resonator used in this experiment was a the Ba(ZrTa)O3 dielectric cylindrical resonator with a dielectric constant = 29, inner and outer diameters of 2 mm and 14 mm, respectively and a height of 5.8 mm [12,17]. The resonance frequency of a given TE011 mode was calculated by using Ansoft/HFSS software and was measured by using a network analyzer (Agilent 8753ES). In addition, the network analyzer was used in tuning the dielectric resonance cavity. The operation frequency and impedance of the NFMM can be precisely tuned over a range of 500 MHz with an unloaded Q factor of 24,000. The probe tip was glued onto one of the prongs of a tuning fork and was oriented perpendicular to the sample surface. The other end of the tip was connected to the resonator loop directly, with the probe tip to sample distance being kept at about 10 nm. The sample was mounted on an x-y-z translation stage for coarse adjustment, which was driven by a computer-controlled microstepping motor with a resolution of 10 nm whereas
Fig. 2. Measured microwave re ection coecients S11 of CdS lms: (a) as-grown and annealed at (b) 200 C, (c) 300 C and (d) 400 C for 2 h. The inset shows (a) the measured re ection coecient S11 near the resonant frequency and (b) the sheet resistance Rs plotted against the annealing temperature.
ne movement of the sample was controlled by using a PZT tube. During all measurements, the temperature was kept at 25 C. III. RESULTS
To characterize the dependence of the sheet resistance of the CdS lms on the annealing temperature, we measured the re ection coecient S11 near the resonance frequency. The resonant frequency is sensitive to the near- eld interaction of the probe tip with the CdS lms. By measuring the resonance frequency and the re ection coecient of the microwave resonator probe, one can determine the sheet resistance of the CdS lms. Figure 2 is a plot of the measured microwave re ection coecient S11 pro le of CdS lms: (a) as-grown and at annealed at (b) 200 C, (c) 300 C, (d) 400 C for 2 h. The matched resonant curve of the (a) as-grown sample has a minimum level of 43.06 dB, which is the reference level of S11 of our measurements. As the annealing temperatures was increased up to 300 C, the minimum re ection coecient S11 decreased from 43.55 dB to 44.24 dB and increased againup to 43.25 dB for an annealing temperature of 400 C, as shown in the inset to Figure 2. The amount of change in the re ection coecient depended on the sheet resistance. An expression of how the re ection coecient S11 depends on the electric sheet resistance of the CdS lms can be derived by using stan-
-682-
Journal of the Korean Physical Society, Vol. 53, No. 2, August 2008
Fig. 3. Optical absorption spectra of CdS lms: (a) asgrown and annealed at (b) 200 C, (c) 300 C and (d) 400 C for 2 h. The inset shows the bandgap energy as a function of the annealing temperature.
dard transmission line theory and is given by assuming impedance matching between the microwave probe and the microwave source [18]: S11
= 20log
R in Z0 ; R + Z0 Zin
Z
(1)
where Z0 is the impedance of the probe (Z0 = 50 ), R is the real part of the complex impedance of the Zin CdS/glass system and can be calculated as 1 + Zg =Rs R (2) Zin = Zg 1 + 2Zg =Rs ; with Zg being the impedance of glass (Zg = 169 ) and Rs the sheet resistance of CdS lm (Rs = 1/(c tc ), where s is the conductivity and ts is the thickness). Note, that the sheet resistance of the as-grown CdS lms is about 5 108 ohm (the electrical conductivity of 0.001 S/m for a lm thickness of 2 m) [4]. We found that the re ection coecient S11 decreased as the annealing temperature was increased up to up to 300 C. The changes in the re ection coecient S11 depends on variations in the sheet resistance of the CdS lm as we can see from Eqs. (1) and (2). Note that the CdS lms sheet resistance changes on the order of 0.016 %/C up to of 300 C annealing temperature. The inset of Figure 2 (rectangle; right axis) shows the dependence of the estimated sheet resistance on the annealing temperature. As the annealing temperatures was increased up to 300 C, the sheet resistance decreased from about 19 Mohm to 8 Mohm and increased again to about 45 Mohm as the annealing
Fig. 4. X-ray di raction spectra of CdS lms: (a) as-grown and annealed at (b) 200 C, (c) 300 C and (d) 400 C for 2 h.
temperature was increased to 400 C. Over the investi gated range (up to 300 C annealing temperature), the empirical dependence of the changes of re ection coecient, S11, on annealing temperature variation T can be approximated as a line with a slope of S11/T = 0.0068 (dB/C) or 5 10 6 C 1 on a linear scale. For our system at maximum sensitivity, the root-meansquare (rms) statistical noise on a linear scale was about 5.4 10 7 [19]. Thus, the signal-to-noise 6ratio (SNR)7 of the CdS lms was SNR = 20 log (5 10 /5.4 10 ) 20 dB. Figure 3 shows typical optical absorption spectra of CdS lms on glass substrates for various annealing tem C and (d) peratures: (a) as-grown, (b) 200 C, (c) 300 400 C. As can be seen, when the annealing temperature was increased up to 300 C, the optical absorption spectrum decreased gradually compared to the as-grown lm and increased for a 400 C annealing temperature. The band-gap energy increased from 2.432 eV to 2.437 eV as the annealing temperature was increased up to 300 C and decreased from 2.437 eV to 2.43 eV for a 400 C annealing temperature, as shown in the inset of Figure 3. Note that CdS lms with a hexagonal phase transformed into a wurtzite phase when the annealing temperature was increased to over 400 C. This fact indicates that the band gap su ers a narrowing e ect due to the transition from a hexagonal to a wurtzite phase. Figure 4 shows the XRD patterns for CdS lms: (a) C, (c) 300 C, (d) 400 as-grown and annealed at (b) 200 C for 2 h. The three main sharp peaks correspond to the (002), (101) and (100) orientations. The structure of the as-grown CdS lm is known to a hexagonal having
Investigation of CdS Films Prepared by Using { Arsen Babajanyan et
Fig. 5. SEM micrographs of CdS lms: (a) as-grown and annealed at (b) 200 C, (c) 300 C and (d) 400 C for 2 h.
-683-
al.
nealing temperature was increased up to 300 C, the surface roughness of the CdS lms become atter and the morphology become denser. At a 400 C annealing temperature the surface roughness increased again. AFM images also con rm these considerations. Figure 6 shows the AFM images of CdS lms: (a) C, (d) as-grown and annealed at (b) 200 C, (c) 300 400 C for 2 h. For the as-grown sample, a rough-grain morphology of the CdS lm was observed. As the annealing temperature was increased up to 300 C, the grain size become larger and the surface roughness become smoother. The decrease in the sheet resistance may be due to the change in the crystal surface morphology as the annealing temperature was increased. Further increases in the annealing temperature brings deterioration of the conducting and the optical properties of the CdS lms. Similarly, increasing the annealing time to more than 2 hours does not have a positive e ect. The optimal annealing condition for CdS lms seems to be 300 C for 2 h. IV. CONCLUSIONS
Fig. 6. AFM images of CdS lms: (a) as-grown and annealed at (b) 200 C, (c) 300 C and (d) 400 C for 2 h.
unit cell dimensions of a = 0.4135 nm and c = 0.6731 nm. As the annealing temperature was increased up to 300 C, the most intense peak at 2 = 26.6 increased and decreased again at 400 C. This peak is attributed to the (002) hexagonal or (111) cubic structure. In order to further study the crystallinity of the CdS lms, we observed the full width at half maximum (FWHM) of the (002) peak orientations. The FWHM value of the (002) orientations decreased from 0.24 to 0.20 as the sheet resistance decreased due to the annealing temperature being increased to 300 C. This result indicates that the lower FWHM of (002) orientations can be correlated with the lower sheet resistance of the CdS lms. Figure 5 shows the SEM images of CdS lms: (a) as C, (c) 300 C, (d) 400 grown and annealed at (b) 200 C for 2 h. For the as-grown CdS lm, a rough-grain morphology of the CdS lm was observed. As the an-
We observed the dependence of the sheet resistance of the CdS lms on annealing temperature by using a NFMM to measure the microwave re ection coecient S11 . The sheet resistance of the CdS lms depends on the morphology and the crystal alignment. As the annealing temperature was increased up to 300 C, the sheet resistance of the CdS lms decreased. Further temperature increases caused an increase in the sheet resistance and a deterioration of the optical properties. The best conducting properties of bath-deposited CdS lms was reached for annealing at 300 C for 2 h. Quantitative characterizations of the dependences of the electric properties of CdS on the substrate heating temperature and time are under ongoing. ACKNOWLEDGMENTS
This work was supported by the Small and Medium Business Administration's Consortium Fund (2007), by Sogang University (2007), by the Korea Research Foundation (KRF-2005-042-C00058; KRF-2002-005-CS0003), by the Seoul Research and Business Development Program (10816) and by the Korea Science & Engineering Foundation (F01-2004-000-1082-0; R01-2006-000-112270). REFERENCES
[1] D. Yang, S. Xu, Q. Chen and W. Wang, Colloids Surf. A 299, 153 (2007).
-684-
Journal of the Korean Physical Society, Vol. 53, No. 2, August 2008
[2] P. Chowdhury, P. Ghosh and A. Patra, J. Lumin. 124, 327 (2007). [3] C. Thanachayanont, K. Inpor, S. Sahasithiwat and V. Meeyoo, J. Korean Phys. Soc. 52, 1540 (2008). [4] B. Pradhan, A. Sharma and A. Ray, J. Crystal Growth 304, 388 (2007). [5] J. Lee, J. Yi, K. Yang and D. Mangalaraj, J. Korean Phys. Soc. 40, 877 (2002). [6] S. Tiwari1 and S. Tiwari, Cryst. Res. Technol. 41, 78 (2006). [7] R. Grecu, E. Popovici, M. Ladar, L. Pascu, E. Indrea and J. Optoelectron. Adv. Mater. 6, 127 (2004). [8] I. Shim, D. Choi, Ch. Kim, J. Pyun and S. Bowles, J. Korean Phys. Soc. 52, 322 (2008). [9] S. Dutta, C. Vlahacos, D. Steinhauer, A. Thanawalla, B. Feenstra, F. Wellstood, S. Anlage and H. Newman, Appl. Phys. Lett. 74, 156 (1999). [10] B. Knoll, F. Keilmann, A. Kramer and R. Guckenberger, Appl. Phys. Lett. 70, 2667 (1997).
[11] J. Lee, Y. Hong, K. Kim, J. Joo, Y. Yoon, S. Kim, Y. Kim and K. Kim, J. Korean Phys. Soc. 48, 1534 (2006). [12] B. Friedman, M. Gaspar, S. Kalachikov, K. Lee, R. Levisky, G. Shen and H. Yoo, J. Am. Chem. Soc. 127, 9666 (2005). [13] B. Friedman, B. Oetiker and K. Lee, J. Korean Phys. Soc. 52, 588 (2008). [14] M. Rami, E. Benamar, M. Fahoume and A. Ennaoui, Phys. Stat. Sol. A 172, 137 (1999). [15] D. Cha, S. Kim and N. Huang, Mater. Sci. Eng. B 106, 63 (2004). [16] B. Pradhana, A. Sharmab and A. Ray, J. Cryst. Growth 304, 388 (2007). [17] A. Babajanyan, J. Kim, S. Kim, K. Lee and B. Friedman, Appl. Phys. Lett. 89, 183504 (2006). [18] D. Pozar, Microwave Engineering (Anderson-Wesley, New York, 1990). [19] J. Kim, M. Kim, H. Kim, D. Song, K. Lee and B. Friedman, Appl. Phys. Lett. 83, 1026 (2003).
