Abstract: In order to determine the suitable source temperatures to deposit CuInSe2 thin films for solar cells, absorption coefficient α and energy gap Eg of this compound were evaluated. The films were grown by vacuum free CSVT technique for 3 h. The source temperatures of deposition ranged from 400 to 600°C. Absorption coefficient and energy gap were determined from transmittance (T) and reflectance (R) of obtained films. Transmittance (T) and reflectance (R) were characterized using a spectrophotometer with wavelength γ range 1-2.5 μm. Curves (α) and (hvα)2 versus photon energy (hv) were analysed. At source temperatures of 450-500°C, the value of α was found to be about 104 cmG1. At source temperature higher than 500°C, α dropped to lower values of 104 cmG1. CuInSe2 thin films were found to have a direct allowed transition and the best Eg value was estimated between 0.94-0.96 eV at source temperatures of 450-500°C. A decrease in Eg value was observed when the source temperature became higher than 500°C. In view of these results, the source temperature range of 450-500°C was found to be more appropriate to grow CuInSe2 films for solar cells. Indeed, α and Eg values can be increased by introducing Gallium into ternary CuInSe2 for a better alignment with the solar spectrum, e.g., a CuIn1-xGaxSe2 heterojunction device.
INTRODUCTION
Copper indium diselenide (CuInSe2 or CIS) is one of the most promising materials studied for solar cell applications due to its direct band gap of about 1.02 eV at room temperature and its high absorption coefficient in the range of 104-105 cm–1.
CuInSe2-based thin films solar cells exhibit high conversion efficiencies (Ramanathan et al., 2003; Contreras et al., 1999; Rau and Schock, 1999; Tuttle et al., 1995). Its high absorption coefficient allows the reduction of required thickness of absorber layer down to about 0.5-1.5 μm.
CuInSe2 thin films have been deposited by various processing techniques, like solution growth technique (Chavhan and Sharma, 2006), RF sputtering (Müller et al., 2006), chemical spray pyrolysis (Terasako et al., 2006), multi-source co-evaporation (Chityuttakan et al., 2006), UV laser ablation (Tverjanovich et al., 2006), metal-organic decomposition (Nakamura and Ando, 2005), electrodeposition (ED) technique (Huang et al., 2004), Close-Spaced Vapour Transport (CSVT) under vacuum (Massé et al., 2004; Moukadir et al., 2004; Kannan et al., 2004; El Hadj Moussa et al., 2002), electron beam and flash evaporation (Castaòeda and Rueda, 2000).
However, the characteristics of the films obtained depend strongly on the growth process and conditions.
The present study reports absorption coefficient and energy gap of CuInSe2 thin films deposited by vacuum free CSVT technique at different source temperature. Absorption coefficient and energy gap were evaluated versus the source temperature in the range of 400 to 600°C. Then, optimal source temperature to deposit CuInSe2 films for solar cells was determined.
MATERIALS AND METHODS
Figure 1 shows the vacuum free CSVT device for the preparation of CuInSe2 thin films as already used (Konan et al., 2007, 2006). It consists of an open quartz tube reactor set horizontally without vacuum, which has a length of 40 cm and an interior diameter of 3.5 cm. The entrance of the quartz tube is composed of two openings for argon flow which can be closed by valves V1 and V2.
| Fig. 1: | Schematic diagram of the CSVT device |
The argon pressure through V1 and V2 were identical and adjusted between 1.5 and 2 bar to allow the transport of iodine vapour into the reactor. The end of the quartz tube is composed of a gas outlet which was kept opened during films deposition process. The source is a powder of CuInSe2 placed in a graphite crucible and hand pressed. This powder with the ratio (Cu/In) = 0.99 was obtained from bulk CuInSe2 synthesized from elements of high purity (6N) of Cu, In and Se. The substrates used are flat pyrex glass of 1 mm thick. The substrates were successively well cleaned in alcohol, de-ionised water and acetone baths and then well dried. The substrates were placed above the source and separated from it by an 2 mm thick glass spacer. A U-form Kanthal resistor placed just under the crucible is used to heat both the source and the substrate. The source temperature indicated by a regulator varied between 400 and 600°C. The substrate temperature directly measured using a chromel-alumel thermocouples, is colder than the source of about 100-180°C. Solid iodine used as a transporting agent was heated to the remaining temperature of 30°C. At the early stage of the heating up of the system, valve V1 was opened and V2 closed. The first argon-flow was bypassed the iodine source and removed the air from the reactor. Therefore, no iodine vapour was present in the reactor. After the source temperature was stabilized, valve V1 was closed and V2 opened so that the second argon-flow transported iodine vapour into the reactor. After the deposition time of about three hours, the two valves were closed during cooling period. Therefore, no more iodine vapour was transported into the reactor.
CuInSe2 thin films were deposited at source temperatures of 400, 450, 500, 550 and 600°C. The films thicknesses were measured by a surface profilometer (DEKTAK 3). A spectrophotometer (Brucker IFS 66v) was used to characterize transmittance (T) and reflectance (R) of films with wavelength (λ) range 1-2.5 μm.
RESULTS
The samples of films obtained at source temperatures of 400, 450, 500, 550 and 600°C were, respectively noted S1, S2, S3, S4 and S5, as shown in Table 1.
Evaluation of absorption coefficient α: The absorption coefficient α was calculated with the following formula (Huang et al., 2004; Djessas et al., 2000):
| (1) |
| d | = | The film thickness. |
| T | = | The films transmittance |
| R | = | Reflectance data, respectively. |
Transmittance (T) and reflectance (R) depend on wavelength λ. Transmittance (T) and reflectance (R) spectra of thin films are shown respectively in Fig. 2 and 3. Wavelengths were converted into incident photon energy using the relationship below:
| (2) |
where, hv is expressed in eV and λ in μm.
The Eq. 1 gives then curves α as a function of hv (Fig. 4).
Each curve exhibits a linear portion that contains an inflexion point. The fundamental value of α was noted αo. The values of αo were the intercept points between the α axis and a horizontal straight-line starting from the inflexion point. Thus, αo was evaluated for each sample.
| Table 1: | Samples obtained at different source temperatures |
| Fig. 2: | Transmittance (T) spectra of CuInSe2 films. (a) Tsource = 400°C; (b) Tsource = 450°C; (c) Tsource = 500°C; (d) Tsource = 550°C and (e) Tsource = 600°C |
| Fig. 3: | Reflectance (R) spectra of CuInSe2 films. (a) Tsource = 400°C; (b) Tsource = 450°C; (c) Tsource = 500°C; (d) Tsource = 550°C and (e) Tsource = 600°C |
Table 2 gives values of αo of the films deposited at different source temperatures. The results showed that at source temperatures of 400-500°C, the values of αo were found to be in the range of 104-2.3 104 cm–1. This range values is in a good agreement with the absorption coefficient values reported by other researchers such as Massé et al. (1997) and Zouaoui et al. (1999). At source temperatures above of 500°C, αo became lower than 104 cm–1 at 550°C. At source temperature of 600°C, the corresponding curve (curve e) in Fig. 4 has not an inflexion point. αo could not be so clearly evaluated.
| Table 2: | Absorption coefficient α and energy gaps Eg estimated from Fig. 4 and 5 |
| Fig. 4: | Absorption coefficient α as a function of photon energy hv for samples. (a) Sample S1; (b) Sample S2; (c) Sample S3; (d) Sample S4 and (e): Sample S5 |
Evaluation of energy band-gap Eg: The energy gaps Eg and the absorption coefficient α are correlated according to the following equation (Kannan et al., 2004; Tverjanovich et al., 2006):
| (3) |
Where:
| k | = | A constant. |
| h | = | The Plank’s constant. |
| hv | = | The incident photon energy. |
As the electronic transition between the valence and conduction bands can be direct or indirect, the exponent β is a number which characterizes the transition process: β = 1/2 for direct allowed transitions, β = 3/2 for direct forbidden transitions, β = 2 for indirect allowed transitions and β = 3 for indirect forbidden transitions. As CuInSe2 is a direct band-gap, the corresponding Eg is obtained with β = 1/2 (Chavhan et al., 2006; Muller et al., 2006; Huang et al., 2004; Djessas et al., 2000; Massé et al., 1997). Therefore the formula used is:
| (4) |
or
| Fig. 5: | Plot of (hvα)2 versus photon energy hv for samples. (a) Sample S1; (b) Sample S2; (c) Sample S3; (d) Sample S4 and (e) Sample S5 |
| (5) |
where, K is a constant equal to k2.
The formula (5) gives curves (hvα)2 as a function of hv (Fig. 5).
All the curves of Fig. 5 exhibit a nonlinear and linear portion, which is characteristic of direct allowed transitions. Such transitions have been reported by Chavhan et al. (2006), Muller et al. (2006), Huang et al. (2004), Djessas et al. (2000) and Massé et al. (1997). According to Muller (2006), the nonlinear portion corresponded to a residual absorption which could be attributed to factors: which are absorption involving impurity states, interference due to multiple reflections that take place at the surface and backside of the films. The straight line portion characterizes the fundamental absorption.
The fundamental value of Eg was noted Ego. The energy gaps Ego of the samples was evaluated from the intercept of the linear portion of each curve (Fig. 5) with the hv axis. Table 2 gives values of Ego of the films deposited at different source temperatures. At source temperatures of 400°C, Ego value of 0.90 eV was lower than the expected value. At source temperatures of 450-500°C, Ego values were in the order of 0.94-0.96 eV. This range values agree well with results early reported by Muller et al. (2006), Chavhan et al. (2006), Huang et al. (2004), Djessas et al. (2000) and Massé et al. (1997) from Eq. 4. These Ego values corroborated with the αo values in the same range of source temperatures 450-500°C.
When the source temperatures became higher than 500°C, Ego values dropped to 0.83 and 0.57 eV, respectively at 550 and 600°C.
The αo and Ego values obtained versus source temperatures of CuInSe2 thin films deposition were in good agreement with the results of structural and morphological analysis reported in the same range of source temperatures (Konan et al., 2007). In both case of study, the optimal source temperature to deposit suitable CuInSe2 thin films for solar cells was about 500°C. At this source temperature, thin films obtained can be used as CuInSe2-based solar cells. Indeed the corresponding αo and Ego values, respectively 104 cm–1 and 0.96 eV can be increased by introducing Gallium into ternary CuInSe2 for a better alignment with the solar spectrum, e.g., a CuIn1-xGaxSe2 heterojunction device (Ramanathan et al., 2003; Contreras et al., 1999; Rau et al., 1999; Tuttle et al., 1995).
CONCLUSIONS
CuInSe2 thin films were deposited by vacuum free CSVT technique at different source temperatures. Absorption coefficient and energy gap of the obtained thin films were determined versus source temperatures through measurements of transmittance (T) and reflectance (R) of films. For films deposited at source temperatures of 450 and 500°C, the value of absorption coefficient was about 104 cm–1. Above 500°C, it lowered down to 104 cm–1at 550°C and could not be clearly evaluated at 600°C.
CuInSe2 thin films were found to have a direct allowed transitions and the best energy gap value was estimated in the order of 0.94-0.96 eV at source temperatures of 450-500°C. At source temperature higher than 500°C, a decrease in energy gap value was observed. These results, compared to structural and morphological analysis previously reported in the same range of source temperatures, confirmed that the source temperature of 450-500°C was required to deposit CuInSe2 films for solar cells.