1. Introduction

Cadmium telluride (CdTe) is an attractive material for fabrication of thin film solar cells due to its suitable band gap energy (~1.5 eV) and high absorption coefficient (> 5 × 10⁵ cm⁻¹) [1–5]. Advantages of CdS/CdTe solar cells include their lowest unit cost for generating electricity, simple device structure and good stability [6,7]. These factors have driven substantial technological progress over the past decades, achieving the cell efficiency as high as 21.5% [8]. There is still an ongoing research on CdTe solar cells to attain the high specific power (or power per mass) and flexibility to diversify their uses in mobile and aerospace applications. To this day, the best efficiencies for CdTe solar cells have been obtained from superstrate structure (back contact/p-CdTe/n-CdS/transparent conductive layer/glass) because of higher conversion efficiency in superstrate configuration as opposed to substrate configuration [7,9]. However, using conventional glass as substrate presents issues of heavy weight and rigid structure, not allowing it to be applicable in portable and form-factor-free applications. Previously, polymers were used as substrates to fabricate flexible and light-weight CdS/CdTe cells in superstrate configuration. Mathew et al. did a pioneering work on fabricating flexible, lightweight CdTe solar cells with notable conversion efficiencies [10]. However, polymer substrates that can endure high process temperature are generally opaque, which is not desirable for cells with superstrate configuration [11].

Besides acquiring optimal conditions for film growth with high quality, development of post-growth process is critical for enhancing the conversion efficiency of CdTe solar cells due to the nature of its polycrystalline structure. The conventional post-deposition processes include CdCl₂ treatment (also called an activation process) and nitric-phosphoric (NP) etch step [12–15]. Although it is still controversial, there has been a lot of research on the effects of CdCl₂ treatment, including grain growth, recrystallization, passivation of the grain boundary, intermixing of CdS and CdTe, etc [5,16–20]. NP etch process is as essential as the activation process for enhancement of the cell performance. Bromine methanol (BrMeOH) is an alternative etching solution to the NP etch solution. However, it was found that BrMeOH solution has some issues. Tiwari et al. reported that etching with BrMeOH solution broadens grain boundaries and Tellurium layer that forms along these boundaries provides a conductive link between grains [15]. Also, by using BrMeOH solution, bromine accumulates at the interface of CdS/CdTe [21]. Therefore, NP etch is preferred in recent years. Durose et al. reported that NP etch process combined with the activation process removes native oxides (TeO₂) on CdTe grains and forms a Te-rich layer [12–14]. Hence, optimization of NP etch process is critical to obtain highly efficient CdS/CdTe solar cells. In this study, we investigated the effects of pre- and post-NP etch processes on the photovoltaic properties of ultra-lightweight and flexible CdS/CdTe solar cells in a superstrate configuration.

2. Experimental details

A schematic diagram of the fabrication process of flexible and ultra-lightweight CdS/CdTe thin film solar cells with superstrate configuration is shown in Fig. 1(a). Thin glass (borosilicate) with thickness of 200 μm was used as substrate, followed by deposition of 200 nm-thick indium tin oxide (ITO) as transparent conducting oxide layer. 150 nm of CdS thin films were deposited by sputtering method as a window layer. Sputtering process was carried out using a single target source at a pressure of 1 mTorr, under Ar ambient and RF power of 120 W. Afterwards, close spaced sublimation (CSS) method was used to deposit 5 μm of polycrystalline CdTe thin films with an area of 1 cm x 1 cm, using the CdTe powder source (99.999%, Johnson Matthey Co.) under Ar ambient. Chamber pressure was controlled to 0.5 Torr. The SiC-coated carbon block was used to keep 3 mm distance between the substrate and the CdTe powder source, and the temperatures were maintained at 540 °C and 600 °C, respectively throughout the deposition period of 5 min at a growth rate of 1 µm/min.

 

Fig. 1 (a) schematic diagram of fabrication of a flexible and ultra-lightweight CdTe thin film solar cell with superstrate configuration, (b) flowchart illustrating three different types of post-deposition treatment processes.

Download Full Size | PDF

Post-deposition processes include CdCl₂ activation and NP etch steps. For CdCl₂ activation, samples (CdTe/CdS/ITO/glass) were dip-coated with a saturated CdCl₂ solution and annealed in a furnace at 385 °C, under air ambient condition for 30 min. NP etch process was executed by immersing the samples into the NP etch solution, which was made from mixing nitric acid and phosphoric acid (HNO₃: H₃PO₄: H₂O = 1/3: 71: 29 in volumetric ratio, Sigma-Aldrich). After NP etching, the samples were rinsed and dried by N₂ gas. For a comparative study on the extent of NP etch process on the photovoltaic properties of our CdTe solar cells, three types of samples were prepared, each with different post-deposition treatment procedure (Fig. 1(b)): (i) no NP etch; (ii) no pre-NP etch; (iii) both pre-NP and post-NP etch process. The first sample was subjected to CdCl₂ activation treatment only. The second sample followed the conventional post-deposition treatment procedure: CdCl₂ activation step, followed by the NP etch process. The process of post-deposition treatment of the last sample was different from the second sample. The as-deposited sample was treated with NP etch solution prior to CdCl₂ treatment. Then the CdCl₂ activation step and post-NP etch step followed.

Subsequent to post-deposition treatments, 30/120 nm of Cu/Au back-contact electrodes were deposited by electron-beam evaporation method. Then the samples were annealed in a furnace under N₂ ambient for approximately 10 min at 300 °C. To investigate the effects of NP etch process on the cell performance, optical, structural, and morphological analysis were performed. Micro-photoluminescence (micro-PL) spectroscopy was used to study the optical properties of the CdTe thin films after they were treated with three different post-deposition processes mentioned above. Micro-PL spectra at room temperature were obtained using the 532 nm line of a diode-pumped solid-state laser (Omicron). Structural characteristics of the CdTe thin films were attained by x-ray diffraction (XRD) (Cu target, 2θ mode, ATX-G, Rigaku). Scanning electron microscope (SEM) (Hitachi S-4300) was used to observe the effects of NP etch process on the morphology of CdTe thin films. Then, the photovoltaic characteristics were measured by using a solar simulator (WACOM WXS-155S-10, AM1.5G, 100 mW/cm²). Lastly, arch-shaped pipe with radius of curvature of 7.9 cm was used to test the flexibility and stability of the cells after multiple cycles of bending.

3. Results and discussion

Light current-voltage (I-V) characteristics of each type of cells with best conversion efficiencies are displayed with the device parameters in Fig. 2. Process (i) shows the lowest parameter values out of the three types of cells, with V_OC = 456.9 mV, J_SC = 24.72 mA/cm², fill factor = 43.39% and efficiency = 4.9%. Process (iii) exhibited dominant photovoltaic values over the other two samples: V_OC = 780.7 mV, J_SC = 29.95 mA/cm², fill factor = 58% and efficiency = 13.56%. Process (ii) presented comparable results to process (iii): V_OC = 767.6 mV, J_SC = 26.70 mA/cm², fill factor = 55.92% and efficiency = 11.46%. The improvement of the photovoltaic parameters of process (ii), compared to process (i), is attributed to post-NP etch process, which is known to remove native oxides (TeO₂) and to form Te-rich layer on CdTe thin film surface, and as a consequence, total resistance decreases by the formation of p⁺ region. Moreover, as shown in Fig. 2(a), the rollover that is observed for process (ii) can’t be seen with process (iii), which can be explained by the effect of pre-NP etch step. As the cell is etched with NP solution prior to CdCl₂ activation, the surface of the CdTe thin films are populated with tellurium, which is beneficial for CdCl₂ activation by enhancing the in-diffusion of chlorine, compared to the surface of CdTe without pre-NP etch process [12]. As a result, the electrical properties can be greatly improved through pre-NP etch, CdCl₂ activation, and post-NP etch steps. We believe that pre-NP etch treatment is a critical post-deposition process as much as the post-NP etch treatment step in achieving high conversion efficiencies in a flexible and light-weight CdTe solar cell.

 

Fig. 2 (a) Light I-V curves of the CdTe solar cells with the highest efficiencies of each process, (inset) photograph of the fabricated ultra-lightweight, flexible CdTe thin film solar cell, (b) summary of the best photovoltaic parameters of each type of cells.

Download Full Size | PDF

To analyze the effects of NP etch process in-depth, SEM was used to observe the surface morphologies and cross-sectional images of the CdTe thin films for each process (Fig. 3). From the top and cross-sectional views of sample (i), bumps on the surface can be observed. These bumps are formed during the CdCl₂ activation step due to recrystallization. The only difference between the fabrication procedures of process (i) (Figs. 3(a) and 3(d)) and process (ii) (Figs. 3(b) and 3(e)) is the presence of post-NP etch step. Correspondingly the roughness of the surface decreased as it is shown in cross-sectional SEM image of sample (ii). Edges of the grains became round and the gaps between the grains disappeared, resulting in elimination of shunt path and thus enhancing the electrical contact between electrode and CdTe. Consequently, the cell performance was advanced, as it was previously shown in electrical properties data (Fig. 2(b)). The surface morphology and cross-sectional image of CdTe thin films made by process (iii) are distinctly different from CdTe thin films fabricated by processes (i) and (ii). Sample (iii) (Figs. 3(c) and 3(f)), which was etched prior to CdCl₂ activation and once more afterwards, shows noticeable increase in CdTe grain sizes and smoother surface. Pre-NP etch effect, which modified the surface of CdTe thin films susceptible to the subsequent activation process has an important role in promoting CdCl₂ activation step. XRD (Fig. 4(a)) was used to analyze crystal orientation of each type of CdTe thin films. Micro-PL spectra in Fig. 4(b) shows that the optical bandgap of CdTe thin films did not change when subjected to processes (i), (ii) and (iii). There was no evident difference in CdTe-related peak intensities and distribution of the peaks, which indicates that pre-NP etch does not affect crystal orientation of CdTe. Based on the changes in optoelectrical properties, grain size and morphology, high efficiency of CdTe solar cell can be attributed to the pre-NP etch step.

 

Fig. 3 (a, b, c) top and (d, e, f) cross-sectional SEM images of the three different types of CdTe thin films. Insets of (a, b, c) are high resolution SEM image.

Download Full Size | PDF

 

Fig. 4 (a) XRD images and (b) micro-PL spectra of the fabricated cells (processes (i), (ii), and (iii)).

Download Full Size | PDF

By using the setup shown in Fig. 5(b), a tensile strain of 0.13% was repeatedly applied to the fabricated cells to assess their flexibility and stability. Following equation was used to calculate the strain: S = t_glass/(2 × R), where R stands for the radius of curvature and t_glass is the thickness of glass substrate [22]. Light I-V curves were obtained each time after applying the strain. As shown in Fig. 5(a), there is barely any difference among the curves for 10 cycles of bending. Also, all the photovoltaic properties of the device show relatively consistent values throughout 10 cycles of bending (Figs. 5(c) and 5(d)). In addition, our flexible CdTe solar cell has a high specific power (about 254 W/kg) based on the weight of substrate and device structure, enhancing the stowability and an ease of installation. Note that a 15% CdTe cell with a conventional glass substrate (~1.5 mm thick) was estimated to have a specific power of 40 W/kg [19]. We believe that our approach can be used to fabricate ultra-lightweight and flexible CdTe thin films solar cells for mobile and space applications [23].

 

Fig. 5 (a) Light I-V curves after repeated bending test under tensile strain of 0.13%, (b) photograph image of the bended CdS/CdTe cell at a radius of curvature of 7.9 cm (0.13% strain), (c) photovoltaic parameters of the CdTe solar cells after applying tensile strain of 0.13% up to 10 cycles.

Download Full Size | PDF

4. Conclusion

A comparative study of post-deposition process was investigated to demonstrate flexible and ultra-lightweight CdTe thin film solar cells in superstrate configuration. NP etch step in conjunction with CdCl₂ treatment were carried out to the CdTe thin films grown on ultra-thin and flexible glass substrates in different sequences. Pre-NP etch step is essential to improve the conversion efficiency for flexible CdS/CdTe solar cells as high as 13.56%. Bending the cells for multiple times with a tensile strain of 0.13% did not cause the degradation of photovoltaic performances, confirming flexibility and durability of the cells. Both flexibility (up to 0.13% strain) and high specific power (~254 W/kg) were achieved by our approach.