computacionales que nos permitieron simular teóricamente tanto los espectros de transmitancia espectral como los de XRD. Estos resultados demostraron que ...
REVISTA MEXICANA DE F´ISICA S 53 (7) 265–269
DICIEMBRE 2007
Effect of synthesis parameters on the optical and structural properties of CuGaSe2 thin films E. Romeroc,∗ , J. Clavijoc , J.S. Oyolab , and G. Gordilloa Departamento de F´ısica, Universidad Nacional de Colombia, Bogot´a,Colombia. b Universidad de la Amazon´ıa, Florencia, Caquet´a, Colombia. c Departamento de Qu´ımica, Universidad Nacional de Colombia, Bogot´a, Colombia. a
Recibido el 30 de noviembre de 2006; aceptado el 8 de octubre de 2007 CuGaSe2 (CGS) thin films grown on soda-lime and boro-silicate glass substrates by a chemical reaction of the precursor species, which were sequentially evaporated in a two stage process, were studied through spectral transmittance and x-ray diffraction (XRD) measurements, in order to investigate the effect of the substrate temperature and the selenium flow on the optical constants (refractive index n, absorption coefficient α and optical gap Eg) and on the structural properties (phase, crystalline structure and lattice parameters). Using the experimental measurements and computational tools that allow us simulating theoretically the transmittance and XRD spectra, the optical constants and the crystallographic parameters of the CGS films were calculated. These results showed that the mentioned synthesis parameters significantly affect the optical constants as well as the crystallographic properties of the CGS films. Keywords: Optical and structural properties; thin films; X-Ray diffraction; solar cells. Pel´ıculas delgadas de CuGaSe2 (CGS) crecidas en sustratos de vidrio soda-lime y boro-silicato mediante una reacci´on qu´ımica de sus especies precursoras en un proceso de dos etapas, fueron estudiadas a trav´es medidas de transmitancia espectral y de difracci´on de rayos X (XRD), con el prop´osito de investigar el efecto de la temperatura de substrato y el flujo de selenio sobre las constantes o´ pticas (´ındice de refracci´on n, del coeficiente de absorci´on y gap o´ ptico Eg), y sobre las propiedades estructurales (fase, estructura cristalina y par´ametros de red). Tanto las constantes o´ pticas como las propiedades estructurales fueron determinadas usando las medidas experimentales y herramientas computacionales que nos permitieron simular te´oricamente tanto los espectros de transmitancia espectral como los de XRD. Estos resultados demostraron que los mencionados par´ametros de s´ıntesis afectan significativamente las constantes o´ pticas as´ı como tambi´en las propiedades cristalogr´aficas de las pel´ıculas de CGS. Descriptores: Propiedades o´ pticas y estructurales; pel´ıculas delgadas; difracci´on de rayos X; celdas solares. PACS: 78.66.Li; 79.60.Dp; 68.37.Ps; 89.30.Cc
1.
Introduction
Wide-band gap chalcopyrite solar cells based on CuIn1−x Gax Se2 (CIGS) are sought for their high voltage and low current output [1, 2]. This type of electrical characteristics should, in principle, significantly reduce resistive losses in the device. Another area of interest for such widegap materials is the fabrication of tandem solar cells where these CIGS absorbers can be used for the top cell in either a mechanically stacked or monolithically grown structure. The NREL group has reported that the optimum bandgaps for a series-connected two junction tandem cell are 1.72 eV for the top cell and 1.44 eV for the bottom cell [3]. These band-gaps are ideally matched to CuGaSe2 (CGS) (Eg=1.68 eV) and Cu(In0.7 Ga0.3 )Se2 (Eg=1.15 eV) solar cells. However, it is a well-known fact that cells made from a CIGS alloy with x>0.3 do not benefit from the increased band-gap because they are dominated by bulk recombination whenever their composition is Cu-poor [4]. This paper reports the improvements we have achieved in the CGS properties through careful choice of growth parameters. We selected as substrates for the experiment traditional soda-lime glass and alternative 7059 corning glass. These alternative substrates allowed us to explore higher processing temperatures (>600◦ C). Another variable of interest in this
study, in addition to higher processing temperatures, is the Se overpressure during film growth.
2.
Experimental
The CIGS films were grown by selenization of the precursors, which are evaporated sequentially using a system constituted by an evaporation chamber connected to a vacuum system working at pressures of about 10−6 Torr, three boats (used to evaporate Se, Ga and Cu, respectively) and a thickness monitor (Maxtec TM-400) with a quartz crystal as sensor, which was used for measuring the flux of the evaporated elements. The substrate temperature is controlled with a programmable PID controller (Eurotherm 900C). Since the reproducibility of the selenium flux is easy to achieve by controlling the evaporation temperature of Se, the CGS films were grown keeping constant the Se- evaporation temperature using a PID controller. However, the Se-flux determined through the deposition rate reported by the thickness monitor was correlated with the Se-evaporation temperature. Fig. 1 shows the deposition rate of Se as a function of the evaporation temperature of Se. The deposition of the CGS films was accomplished in two stages:
266
E. ROMERO, J. CLAVIJO, J.S. OYOLA, AND G. GORDILLO
TABLE I. List of deposition parameters and variation ranges used to perform the parameter study Deposition parameters ˚ Flux of Ga (A/s)
1st stage
2nd stage
4
-
˚ Flux of Cu (A/s)
-
3
˚ Flux of Se (A/s)
2 - 40
2 - 40
Substrate temperature (◦ C)
300
450 - 680
Evaporated mass of Ga (gr)
0.3
-
Evaporated mass of Cu (gr)
-
0.2
• In the second stage the substrate temperature is elevated to values ranging from 500 to 630◦ C and Cu is evaporated in a Se environment. During this stage the ˚ and the flux Se flux was varied between 3 and 40 A/s ˚ of Cu was kept at a value of about 3 A/s. Table I lists the deposition parameters and the variation range used to perform the parameter study. The elemental evaporated metal flux and the substrate temperature profile that led to CGS films with good properties for solar cells fabrication is shown in Fig. 2. The transmittance measurements were carried out with a VIS - IR Oriel spectrophotometer and the morphological characterization was performed using a PSI AFM microscope. The XRD measurements were performed using the Cu Kα radiation of a Shimadzu-6000 diffraction-meter.
F IGURE 1. Variation of the Se-deposition rate as a function of the evaporation temperature of Se.
F IGURE 3. Transmittance measurements of CGS thin films grown on a) soda-lime and b) boron-silicate glass substrates varying de Se-flux.
F IGURE 2. Substrate temperature profile and metal flux used to prepare the CGS films with good properties for solar cells fabrication.
• In the first stage a Gax Sey layer is grown by simultaneous evaporation of Ga and Se, keeping the substrate temperature at 30◦ C. During this stage the Se flux is ˚ and the flux of Ga at a value of kept around 16 A/s ˚ about 4 A/s.
F IGURE 4. Transmittance measurements of CGS thin films grown on a) soda-lime and b) boron-silicate substrate varying the substrate temperature.
Rev. Mex. F´ıs. S 53 (7) (2007) 265–269
EFFECT OF SYNTHESIS PARAMETERS ON THE OPTICAL AND STRUCTURAL PROPERTIES OF CuGaSe2 THIN FILMS
267
TABLE II. Lattice constant of CGS thin films grown on boronsilicate substrate varying the Se-flux and substrate temperature ˚ Lattice Constant (A)
F IGURE 5. Curves of n vs.λ, α vs. λ and (αhν)2 vs. hν, corresponding to CGS thin films deposited on soda-lime and boronsilicate glass substrates respectively.
F IGURE 6. Typical XRD spectra of CGS thin films grown on a) soda-lime and b) boron-silicate substrate, varying the Se-flux.
3.
˚ Rate (As-1)
a ± 0.005
c ± 0.005
8
5.613
10.999
21
5.579
11.007
40
5.580
10.982
Tsus
a ± 0.005
c ± 0.005
550
5.615
11.005
600
5.614
11.002
630
5.610
10.995
Results and discussion
3.1.
Optical characterization
Figures 3 and 4 show the effect of the Se flux and the substrate temperature on the transmission spectra of CGS films deposited in two stages on soda-lime and boron-silicate glass substrates, respectively. It is observed in Fig. 3 that the decrease of both, the Se flux and the substrate temperature, leads to a decrease of the transmittance, especially at λ values close to the cut-off wavelength. This behavior seems to be caused by absorption through states within the energy band gap associated to native defects (vacancies and antisite defects) [5]. Absorption via extended states associated to tail bands could be another possible cause of the transmittance decrease observed when the Se-flux and/or the substrate temperature decrease. Therefore, to get CGS films with low density of states associated to native defects or to tail bands, it is recommendable to prepare them, under Se-overpressure and high substrate temperature conditions. Similar results were reported elsewhere [6]. The optical constants (refractive index n, absorption coefficient α and optical gap Eg ) were determined using the transmission spectrum and calculations based on a procedure described in detail in reference [7]. Figure 5 shows typical curves of n vs. λ, α vs. λ and (αhν)2 vs. hν, obtained for CGS films deposited on sodalime and boron-silicate glass substrates. From the results shown in Fig. 5, the following facts can be remarked: • The refractive index of the CGS films grown on boronsilicate glass substrate is greater than that of the films deposited on soda-lime glass substrates. This behavior could be attributed to the fact that the CGS films grown on boron-silicate glass present, generally, an electrical conductivity greater than the other CGS films, taking into account that the refractive index can be expressed by the relation, n = 4πσ/cα [8].
F IGURE 7. Typical XRD spectra of CGS thin films grown on a) soda-lime and b) boron-silicate substrate varying the substrate temperature.
• The absorption coefficient of the CGS films grown on both types of substrates is greater than 104 cm−1 , indi-
Rev. Mex. F´ıs. S 53 (7) (2007) 265–269
268
E. ROMERO, J. CLAVIJO, J.S. OYOLA, AND G. GORDILLO
cating that these films are adequate as absorber layers in thin film solar cells. • The linear behavior of the (αhν)2 vs. hν curves indicates that the CGS thin films have a direct energy band gap. Using the relation (αhν)2 =A(Eg -hν), which is valid for this type of materials, Eg values of 1.68 and 1.76 eV were found for the CGS films deposited on soda-lime glass and boron-silicate glass substrates respectively. 3.2. Structural characterization The effect of the Se- flux and the substrate temperature on the structural properties of the CGS films was also studied through XRD measurements. Figures 6 and 7 show typical XRD spectra of CGS films deposited under different Seflux and substrate temperature conditions, on soda-lime and boron-silicate glass substrates, respectively. The XRD measurements indicate that the CGS films grow with a chalcopyrite type structure and, in general, independently of the substrate type used, the CGS films deposited at low Se-flux (atomic flux ratio of anion to cation < 2.5) and
low substrate temperature (< 550◦ C) grow with a preferential orientation along the (112) plane. The increase of the Se-flux and/or the substrate temperature leads to the growth of more randomly oriented CGS films. The data regarding structure and lattice constants associated to the CuGaSe2 phases, which are listed in the Table II, were obtained using the data reported in the JCPDS data base for these types of compounds. 3.3.
Morphological characterization
The CGS films were characterized through AFM measurements in order to determine the effect of the Se flux, substrate temperature and type of glass substrate on the morphology. Figure 8 shows typical AFM images of CGS films deposited on soda-lime and boron-silicate glass substrates, under different substrate temperatures and Se-flux. It is clear from these results that the deposition conditions and substrate type significantly affect the films’ morphology. Table III lists the average values of the grain size and roughness of the CGS films deposited on soda-lime and boron-silicate glass substrates, as functions of the Se-flux and substrate temperature, respectively.
F IGURE 8. AFM images indicating the influence of the substrate temperature and Se-flux on the morphology of CGS thin films deposited a) on soda-lime and b) on boron-silicate glass substrate. Rev. Mex. F´ıs. S 53 (7) (2007) 265–269
EFFECT OF SYNTHESIS PARAMETERS ON THE OPTICAL AND STRUCTURAL PROPERTIES OF CuGaSe2 THIN FILMS
269
TABLE III. Variation of the roughness and average grain size of the CGS films deposited on soda-lime and boron-silicate glass substrates, as functions of the Se-flux and substrate temperature. Se-flux ˚ (A/s)
Soda-lime
Boro-silicate
Soda-lime
Boro-silicate
8
71.0
79.2
0.53
0.23
21
52.2
50.4
0.63
0.26
40
36.7
26.5
0.81
0.28
ccc TSus ( C)
Rms rough (nm)
Grain size (µm)
500
24.5
0.30
550
13.6
0.15
580
11.5
0.11
◦
Rms rough. (nm)
Grain size (µm)
It is observed that the grain size of the CGS films increases by increasing the Se-flux, whereas the roughness decreases when the Se-flux is increased. On the other hand, it was found that the increase of the substrate temperature leads to a decrease of both, the average roughness and the average grain size. It was also found that the grain size and roughness of the CGS films deposited on soda-lime glass substrates are greater than those of the films deposited on boron-silicate substrates. Until now, we have not investigated the causes of this behavior.
4.
grown through two stage processes. Characterizations carried out using spectrophotometry measurements as well as XRD and AFM measurements have revealed that the substrate type as well as the substrate temperature and Se-overpressure significantly affect the optical, structural and morphological properties. In general, the films grown at higher substrate temperatures and under Se-overpressures present better properties to be used as absorber layer in solar cells.
Acknowledgements
Conclusions
CuGaSe2 thin films with adequate properties to be used as absorber layer in the top solar cell of tandem structures were
This work was supported by the Universidad Nacional de Colombia and Colciencias.
1. R. Nuofi et al., Proc. 3rd World Conference on Photovoltaic Energy Conversion, 2003, Osaka, Japan.
5. R.H. Bube, Series on properties of Semiconductor materialsVol. 1: Photovoltaic materials, (Imperial College Press, 1998).
2. M. Symko-Davies, 31st IEEE Photovoltaic Specialists Conference, 2005, Lake Buena Vista, Florida USA.
6. M. Contreras, M. Romero, and D. Young, 3rd World Conference on Photovoltaic Energy Conversion, Papan (2003).
3. T.J. Coutts, K.A. Emery, and J.S. Wand, Prog. Photovolt.: Res. Appl. 10 (2002) 195.
7. M. Gracia, F. Rojas, and G. Gordillo, 20th European photovoltaic Solar Energy Conference (Barcelona-Espa˜na, 2005).
4. M. Turcu, O. Pakma, and U. Rau, Appl. Phys. Lett. 80 (2002) 2598.
8. J.I. Pankove, Optical processes in semiconductors (Denver publications, New York, 1971).
Rev. Mex. F´ıs. S 53 (7) (2007) 265–269
