Interface Study of ITO/ZnO and ITO/SnO₂ Complex Transparent Conductive Layers and Their Effect on CdTe Solar Cells

Abstract

Transparent ITO/ZnO and ITO/SnO₂ complex conductive layers were prepared by DC- and RF-magnetron sputtering. Their structure and optical and electronic performances were studied by XRD, UV/Vis Spectroscopy, and four-probe technology. The interface characteristic and band offset of the ITO/ZnO, ITO/SnO₂, and ITO/CdS were investigated by Ultraviolet Photoelectron Spectroscopy (UPS) and X-ray Photoelectron Spectroscopy (XPS), and the energy band diagrams have also been determined. The results show that ITO/ZnO and ITO/SnO₂ films have good optical and electrical properties. The energy barrier those at the interface of ITO/ZnO and ITO/SnO₂ layers are almost 0.4 and 0.44 eV, which are lower than in ITO/CdS heterojunctions (0.9 eV), which is beneficial for the transfer and collection of electrons in CdTe solar cells and reduces the minority carrier recombination at the interface, compared to CdS/ITO. The effects of their use in CdTe solar cells were studied by AMPS-1D software simulation using experiment values obtained from ZnO, ITO, and SnO₂. From the simulation, we confirmed the increase of 𝐸_ff, FF, 𝑉_oc, and 𝐼_sc by the introduction of ITO/ZnO and ITO/SnO₂ layers in CdTe solar cells.

1. Introduction

Transparent conducting oxide (TCO) layers have been extensively studied because of their use as transparent electrodes in displays and in photovoltaic devices [1]. By incorporating a high resistance layer, the thickness of a conducting cadmium sulfide (CdS) layer can be reduced, which significantly improves the blue response of CdTe devices [2] and makes CdTe thin-film solar cells more competitive [3]. Wu has reported the efficiency of 16.5% with <100 nm CdS thickness [4]. However, as the thickness of CdS is decreased, the films would become discontinuous leading to the formation of localized CdTe/TCO junction, which leads to excessive shunting and therefore lowers the solar cell efficiency [5]. Using complex transparent conductive layers is known as a feasible method to improve the characteristics of CdTe thin films solar cells. Indium-tin oxide (ITO) systems, SnO₂, and ZnO have been used as the high resistance layer because of their excellent electrical and optical properties [6, 7] and the improvement of device performance [8].

The complex transparent conductive layers are always heterojunction structure, which are rather complicated systems for their different electron affinities, band gaps. The band offset and interface properties of a heterostructure are some of the most important properties. Sheng et al. have studied the n-layer/transparent conducting oxide (n/TCO) interfaces in amorphous silicon (a-Si:H) and microcrystalline silicon (mc-Si:H) materials by XPS [9]. Liu et al. studied the interface properties and band alignment of Cu₂S/CdS heterojunction, and the band offsets are obtained [10]. Horn studied electronic structure at the interface, relating to band bending and the evolution of transport barriers such as the Schottky barrier and the heterojunction band offset [11]. Bernède and Marsillac. have measured the band offsets of SnO₂/γ-In₂Se₃ heterojunction by XPS and estimated the conduction band discontinuity Δ𝐸_𝑐 to be −0.3±0.3 eV [12]. Mönch discussed the electronic properties and chemical interactions at GaAs (110) and InP (110) surfaces [13].

In this present work, ITO, ZnO, and SnO₂ films have been successfully prepared on ITO coated glass substrate by DC- and RF-magnetron sputtering and characterized by XRD, UV/Vis spectra, and four-probe apparatus. UPS and XPS were used to characterize the band offset of ITO/CdS, ITO/ZnO, and ITO/SnO₂ layers and the influence on the transfer of electrons in CdTe thin film solar cells. AMPS-1D simulations, based on the Poisson equation and the hole-electron continuity equations in one dimension [14], were used to study the device performance of the CdTe solar cells with these different transparent conductive layers.

2. Experiment

DC magnetron was used to sputter the ITO and ZnO films in this paper. The targets were ceramic ITO (In₂O₃ : Sn₂O₃ = 90 : 10) and metallic Zn (99.999%).The substrate was held at 300^∘C while the sputtering pressure was 1.0 Pa for ITO and 2.4 Pa for Zn, of high purity argon (99.999%) mixed with 10% oxygen. The SnO₂ films were prepared on an ITO substrate by RF-magnetron sputtering. Sputtering was conducted at 267^∘C and 1.0 Pa of argon (99.999%) mixed with 1% oxygen. The ITO layers were about 400 nm and the ZnO (SnO₂) were about 150 nm. The CdS layers were deposited by chemical bath deposition (CBD) and were about 150 nm thick.

The TCO/CdS structure was analyzed by X-ray diffraction (DX-2500, Dandong Fangyuan Instrument LLC) using Cu Κa radiation (𝜆=0.15405nm). The sheet resistance was measured with a Digital Four-Probe Tester (SZT-2, Suzhou Tongchuan Electronics). The thickness of each film was measured with a stylus profiler (XP-2, Ambios Technology Inc.). The optical transmission was measured by UV/Vis spectrometer (Perkin Elmer Inc., Lambda-950). The XPS and UPS were measured by the multifunctional X-ray Photoelectron Spectroscopy (AXIS Ultra^DLD, Kratos Analytical Inc.) with the base pressure ~5.0 × 10⁻⁹ Torr, the X-ray of Al K_𝛼, and the X-ray tube power of 130 W. The samples were etched by 21.2 eV He⁺ beam and were biased with 7.36 volts to obtain reproducible cut-off results. The work functions were determined from the low-kinetic energy cut-off in the UPS spectra; that is, the intersection of the linear extrapolation with the baseline. In this experiment, samples were cleaned and thinned by sputtering with He⁺ ions in HUV.

AMPS-1D has been employed to model and analyze the CdTe solar cells, and the different cells configuration is shown in Figure 1. The parameters used in the simulation are shown in Table 1. The electron affinity energies and mobility were obtained from [15–17]. The thickness of CdTe film was set as 6 μm.

Table 1 

The parameters used in the AMPS-1D simulation.

Figure 1 

Structural view of CdTe solar cells with different configuration.

3. Results and Discussion

3.1. XRD

Figure 2 shows XRD patterns of ZnO and SnO₂ films deposited on ITO films. The main diffraction peaks (400), (440), and (222) and so on come from the ITO films. Only one weak peak of ZnO was observed in the spectra labeled as (101), and one weak peak of SnO₂ was observed labeled as (200) at 36.10^∘ diffraction angle. This indicates that the ZnO and SnO₂ films have been deposited successfully onto the ITO films.

(a)

(b)

Figure 2 

XRD spectrum of ITO/ZnO and ITO/SnO2 films.

3.2. Transmittance

The optical and electrical properties of ITO, ITO/ZnO, and ITO/SnO₂ films were measured. The thickness was 400 nm for ITO films and 150 nm for ZnO and SnO₂ films. Figure 3 shows the optical transmission of the as-deposited ITO/ZnO and ITO/SnO₂ films. At the wavelength from 500 nm to 850 nm, the average transmittance is 82% for ITO/ZnO and 81.64% for ITO/SnO₂ films, respectively, which are not lower than ITO films (85%) too much. On the other hand, in the blue region, the red shifts of the effective absorption edge of ITO/ZnO and ITO/SnO₂ films are clearly observable. The sheet resistance of the ZnO was obtained as 108 Ω/□ while 104 Ω/□ for the SnO₂ films, which are higher than ITO (13.2 Ω/□). The deposition of ZnO or SnO₂ films as high resistance transparent (HRT) on ITO films can passivate the CdS surface than ITO, which could eliminate the leakage current caused by the pinhole effects of CdS [15] and thus improves the short circuit current remarkably.

Figure 3 

Transmittance spectra of the ITO, ITO/ZnO, and ITO/SnO2 films.

3.3. XPS/UPS

Before analysis with XPS/UPS, all of the samples were cleaned by sputtering with He⁺ ions for 1 minute in HUV to eliminate surface effects. The layers were profiled using XPS and UPS by taking spectra after every profiling time intervals until it revealed to the ITO film by He⁺ sputtering.

3.3.1. ITO/CdS System

Figure 4 shows the UPS of ITO/CdS structure at different times (9–12 min). The “valence band offset” corresponds to be (𝐸_𝐹−𝐸_VBM) and provides a direct measure of the Fermi level at the sample surface. The results show that the valence band maximum (VBM) in the interface of  ITO/CdS films increases from 2.14 eV to 3.10 eV. The 𝐸_cutoff (secondary electron onset) is the abscissa value on the left side when the intensity is 0 in the later UPS, so the work function can be obtained by Φ=ℎ𝑣−𝐸_cutoff−𝐸_Fermi. (1)

Figure 4 

UPS profiling spectra of ITO/CdS films.

The ℎ𝜈 is the excitation energy of Helium (21.2 eV), and the 𝐸_Fermi is set to be 0, so the Φ would be concluded to be 4.7 eV for CdS and 4.2 eV for ITO. And the interface dipole 𝛿 of ITO and CdS films can be obtained as 0.41 eV and 0.28 eV by 𝛿=𝐸_𝑔−VBM. (2)

The XPS at various times in Figure 5 shows the band energy variation at the interface of the ITO/CdS films. The profiling of samples starts CdS films and ends at 46 min, when the intensity peak of Cd3d is zero. The O1s emission gradually increases in intensity and the Cd3d decreases during the profiling process. At 1–9 min the intensity of O1s is weak and the B.E. is of no change because the content of oxygen atoms is very lower when the CdS films have been etched for 1 min. At 9–11 min, the intensity of O1s increases while the Cd3d decreases and the B.E. of Cd3d increases about 0.14 eV. That is to say, that the surface profiling occurs at 9–11 min. The thickness of CdS films is 150 nm, so the profiling speed is estimated as 15 nm/min. The intensity of Cd3d decreases gradually and passes off at 46 min. The results show a shift in all the core level lines to larger binding energies, which indicates the formation of a space charge layer (band bending) in the substrate.

(a)

(b)

Figure 5 

XPS spectrum of the O1s and Cd3d reigns for n-ITO/n-CdS isotype heterojunction at various profiling times.

In order to construct the band energy diagram, the position of 𝐸_𝐹 within the bulk must be known. The difference between the vacuum level Δ𝐸_vac can be obtained by subtracting the overall band bending from the difference of the work functions by (𝑥 is electron affinity) Δ𝐸_vac=𝜒₂+𝛿₂−𝜒₁+𝛿₁=0.56eV. (3)

Then the valence band offset Δ𝐸_𝑉 can be determined by using the band energy difference between the O1s and Cd3d at intermediate coverage and the binding energies of the core levels with respect to the valence band. One has Δ𝐸_𝑉=VBM₁−VBM₂−Δ𝐸_vac=0.61eV. (4)

Using the Δ𝐸_𝑉 and the band gaps given before, the conduction band offset Δ𝐸_𝑐 was calculated to be Δ𝐸_𝑐=Δ𝐸_vac+𝛿₁−𝛿₂=0.9eV. (5)

Having determined band bending, band offset, and interface dipole, the final band alignment at the interface ITO/CdS heterojunction is presented in Figure 6. The conduction band bends upward in the ITO layer at the surface while the CdS layer bends downward. The electrons need to overcome a huge energy barrier (about 0.9 eV) when transferring from the CdS to ITO films.

Figure 6 

Energy band diagram of the ITO/CdS isotype heterojunction.

3.3.2. ITO/ZnO System

The characterization of the layer at different profiling times is illustrated in Figure 7. The VBM increases from 3.11 eV to 3.21 eV at the interface of the ITO/ZnO films. And the Φ would be calculated to be 3.8 eV for ZnO and 4.1 eV for ITO. And the interface dipole δ of ITO and ZnO fims can be obtained as 0.51 eV and 0.16 eV.

Figure 7 

The UPS spectra of ITO/ZnO thin films.

The XPS at various profiling times in Figure 8 shows the variations in the band energy of the ITO/ZnO films at the interface (20–23 min). The profiling of samples starts ZnO films and ends at 61 min. The In3d (ITO) emission gradually increases in intensity and the Zn2p (ZnO) intensity decreases along the profiling progress. The thickness of ZnO films is 150 nm, so the profiling speed is estimated as 7 nm/min, which is obviously lower than CdS. And Δ𝐸_vac=0.05 eV was obtained and the valence band offset Δ𝐸_𝑉 and conduction band offset Δ𝐸_𝑐 were calculated to be 0.4 eV and 0.05 eV, respectively.

(a)

(b)

Figure 8 

XPS spectrum in the In3d and Zn2p reigns for n-ITO/n-ZnO isotype heterojunction under various profiling time.

The final band alignment at the interface ITO/ZnO heterojunction is presented in Figure 9. The conduction band bends downward in the ITO layer while ZnO layer bends upward at the interface. The barrier energy is about 0.4 eV, which is lower than ITO/CdS heterojunction potential barrier. That is to say the introducing ZnO film is beneficial for the transfer and collection of electrons.

Figure 9 

Energy band diagram of the ITO/ZnO isotype heterojunction.

3.3.3. ITO/Sn𝑂₂ System

The interface profiling from 9 to 20 min is presented in Figure 10. At the interface of ITO/SnO₂, the VBM value decreases from 3.38 eV to 3.28 eV. The Φ of SnO₂ films would be calculated to be 4.14 eV and 4.06 eV for ITO films. And the interface dipole 𝛿 of ITO and SnO₂ films can be obtained as 0.44 eV and 0.73 eV.

Figure 10 

The UPS spectra of ITO/SnO2 complex layers.

The XPS at various profiling times is showed in Figure 11. The profiling of samples starts SnO₂ films and ends at 42 min, when the intensity of Sn3d passes off. It confirms that the In3d (ITO) emission increases in intensity and the Sn3d (SnO₂) decreases during the profiling process. At 20–23 min, the profiling came to the interface, and the profiling speed is estimated as 7 nm/min and the thickness of SnO₂ film is 150 nm. The adding of SnO₂ films between ITO and CdS films also changes the energy structure. The Δ𝐸_vac=0.13 eV, Δ𝐸_𝐶=0.44 eV, and Δ𝐸_𝑉=−0.06 eV were also obtained.

(a)

(b)

Figure 11 

XPS spectrum in the In3d and Sn3d regions of an n-ITO/n-SnO2 isotype heterojunction at various profiling times.

The energy band diagram is presented in Figure 12. The conduction band bends downward in the ITO layer at the surface while SnO₂ layer bends upward. The results shows that electrons must overcome an energy barrier (0.44 eV) when transferring from the SnO₂ to ITO films, which is also less than that in the ITO/CdS heterojunction. Thus adding SnO₂ layer is also beneficial for the transfer and collection of electrons.

Figure 12 

Energy band diagram of the ITO/SnO2 isotype heterojunction.

3.4. Device Simulation

Based on the previous analysis, we have simulated the effect of ITO/ZnO and ITO/SnO₂ films in CdTe cells by AMPS-1D. Figure 13 shows the electric fields, dark I-V curves, and simulated output performance of different cells. The results show that inserting of ZnO or SnO₂ films changes the electric field distribution, with the electric field strength decreasing at the ITO/CdS interface and a new electric field appearing at the ZnO (or SnO₂)/CdS interface. These electric field distributions are beneficial for the transfer and collection of electrons in CdTe cells. The introduction of ITO/ZnO or ITO/SnO₂ films in CdTe solar cells improves the efficiency (𝐸_ff), open voltage (𝑉_oc), and short circuit current (𝐼_sc) significantly.

(a)

(b)

(c)

(d)

Figure 13 

The simulated output performance for different structure: the electric field distribution (a), the dark IV curve (b), 𝐽sc and 𝐸ff, (c) and FF & 𝑉oc (d).

Also, we fabricated CdTe solar cells with or without HRT films. The cells with ZnO films have the efficiency of 12.17% (𝑉_oc=742 mV, 𝐽_sc=26 mA/cm², FF=62.6%, and area=0.5 cm²) and the sample with SnO₂ films has the efficiency of 11.4% (𝑉_oc=724mV, 𝐽_sc=25.8 mA/cm², FF=61.2%, and area=0.5 cm²), while the sample without HRT layers has a much lower efficiency of 8.7% (𝑉_oc=689 mV, 𝐽_sc=23.55 mA/cm², FF=53.6%, and area=0.5cm²). The results also show that the introduction of HRT layers decreases the series resistance, which was partly attributed to good interface properties between HRT and CdS layers.

4. Conclusion

The ITO/ZnO and ITO/SnO₂ films were successfully deposited on a glass substrate by DC- and RF-magnetron sputtering. The optical transmittance of the ITO/ZnO and ITO/SnO₂ as complex TCO layers was 82% and 81.64% from 500 nm–850 nm, respectively. The measured sheet resistances of ITO/ZnO and ITO/SnO₂ layers were 10⁵ Ω/□ and 37.5 Ω/□, respectively. The interface compositions of the TCO layers were characterized by UPS and XPS, and the energy band diagrams were determined. The energy barriers at the interface of ITO/ZnO and ITO/SnO₂ layers are almost 0.4 and 0.44 eV, which are lower than those at ITO/CdS heterojunctions (0.9 eV). The ITO/ZnO and ITO/SnO₂ as complex transparent conductive benefit the transfer and collection of electrons in CdTe solar cells and reduce the minority carriers recombination at the interface, compared to CdS/ITO.

Furthermore, we have also simulated and analyzed the effects of the ITO/ZnO and ITO/SnO₂ films on CdTe cells by AMPS-1D. The results show that the electric field distribution changes a lot by the introduction of ZnO and SnO₂ films between ITO and CdS. The 𝐸_ff, FF, 𝑉_oc, and 𝐼_sc are improved significantly, that is to say, the ITO/ZnO and ITO/SnO₂ complex transparent conductive layers are beneficial for the performance of CdTe solar cells.

Acknowledgments

This research was supported by the National Basic Research Program (973 Program) of China under Grant no. 2011 CBA00708.