Materials issues in very thin film CdTe for photovoltaics

doi:10.1016/j.tsf.2004.11.130

Thin Solid Films

Volumes 480–481, 1 June 2005, Pages 76-81

https://doi.org/10.1016/j.tsf.2004.11.130 Get rights and content

Abstract

A study is made of extremely thin absorber layers of CdTe deposited onto planar substrates in order to characterise their properties. Metal organic chemical vapour deposition (MOCVD) was used to deposit the CdS and CdTe films under controlled conditions of substrate temperature and VI/II organometallic ratio. In situ laser reflectance was used to monitor the thickness of the films and also the roughening from loss of reflected intensity. The roughness of the final structure was measured with atomic force microscopy (AFM) and compared with that of both indium tin oxide (ITO)/glass substrates and CdS deposited on them. Absorber layer thicknesses of between 50 and 500 nm were deposited, the thicker layers being grown in two stages, with a highly As doped cap layer to act as a contact. A degradation of the current–voltage characteristics was observed from good rectification at 500 nm thick absorbers to an ohmic characteristic at 100 nm thickness. This could not be simply explained by the non-coalescence of grains, based on the AFM results, but attributed to defects on the ITO/glass substrate. Improvements in the cleaning procedure resulted in good rectification for 100 nm absorber layer. It was demonstrated that CdTe covers non-planar CdS conformally within 50 nm.

Introduction

The absorber layer thickness for CdTe is normally between 2 and 10 μm, but most of the band gap radiation is absorbed in less than 2 μm in a planar structure. Thicker films are often used to avoid pinholes through the absorber layer that can lead to shorting from the back contact. The pinholes can arise from incomplete coalescence of the CdTe grains during deposition, or due to defects in the underlying surface, and are a continuing concern in thin film polycrystalline devices. A further problem is encountered in very thin CdTe where the minority carrier diffusion length is considerably less than the absorber thickness as this will reduce the photocurrent and photoconversion efficiency of the device. Successful PV devices rely on a fortuitous passivation of the grain boundaries that repel the minority carriers and increase the probability of diffusion to the junction [1], [2], [3]. Other factors that can affect the collection efficiency are built-in fields, junction location and carrier diffusion barriers.

The extremely thin absorber (ETA) cell has been proposed to improve the electrical collection efficiency of polycrystalline thin film photovoltaics [4]. The loss in optical absorption efficiency can be compensated by using a porous or nano-porous structure. For example, Ernst et al. [4] used a micro-porous TiO₂ substrate for the CdTe ETA layer and coated with a total thickness of 500–2000 nm of CdTe. The estimated coating thickness on the TiO₂ micro-pores was 100–200 nm. More recent work by Ernst et al. [5] has reduced this to 150 nm, using electrodeposition of CdTe onto nano-porous TiO₂. The coverage was very good with local variations in thickness from 100 to 200 nm. However, the quantum efficiency was low in this structure, falling to below 40% at 800 nm. Mercury was alloyed with the CdTe in an attempt to improve the absorption of the film. The device characteristics of these ETA cells have been to have a low short circuit current and very low fill factor. This was attributed to the large band offset between TiO₂ and CdTe. However, these authors were able to show an enhancement over the planar equivalent structure of a factor of 50 in the quantum efficiency.

Konenkamp et al. [6] have investigated different nano-structured substrates, including TiO₂ and ZnO. The ZnO produced a very columnar structure and it was concluded that this might be a good morphology for ETA cells. The same approach was taken by Levy-Clement et al. [7] who produced a columnar ZnO film by electrodeposition. These columns were subsequently coated with CdTe by vapour phase epitaxy under a dynamic vacuum. The coating uniformity appeared to be excellent with an average thickness of just 50 nm.

This paper investigates some of the materials issues in depositing ETA layers of CdTe by depositing in a conventional planar geometry onto a CdS window layer. The chosen method for depositing the CdS and CdTe layers was metal organic chemical vapour deposition (MOCVD) as this has shown to be a useful method for controlling grain size, VI/II stoichiometric ratio and intentional introduction of dopants [8], [9], [10], [11]. The latter two advantages offer the opportunity for experimenting with the characteristics of nucleation, development of surface morphology and electrical characteristics of the films. In this study, a series of layers were grown where the absorber layer thickness was changed in each experimental structure. It was also possible to investigate the CdS window layer characteristics and produce films with different degrees of roughening. The film deposition was monitored in situ using laser interferometry (635 nm diode laser) and ex situ measurements of the films included atomic force microscopy (AFM) and measurement of I–V characteristics. A particular aspect to this work was to investigate the effects of very thin absorber layers without the complication of deposition onto a nano-structured surface. Issues such as shunting due to pinholes and non-coalescence of grains were studied.

Section snippets

Experimental

Details of the experimental procedure for the MOCVD growth of CdS/CdTe have been given elsewhere [8]. The substrates were indium tin oxide (ITO) coated glass substrates supplied by Merck and Delta Technology. The organometallics used in these experiments were dimethylcadmium (DMCd), ditertiarybutylsulphide (DTBS) and diisopropyltelluride (DIPTe), all supplied by Epichem. The dopant precursors were n-hexylchloride (n-hexCl) for the CdS layer and tris-dimethylaminoarsenic (DMAAs) for the CdTe

Characteristics of thin absorber layers

The thicknesses of the MOCVD grown layers were measured in situ using laser reflectometry and a typical interferometer trace is shown in Fig. 1 for layer 2. The first broad peak is approximately one interference oscillation of CdS which was grown for 20 min. The relatively slow growth rate relates to the low temperature of 290 °C that is needed to obtain high quality CdS layers. The growth process was then stopped while the substrate temperature was reset to 320 °C under a flow of pure

Conclusions

Planar CdTe solar cell structures with extremely thin absorbers have been studied to gain more insights into the relationship of device to materials characteristics.

Using MOCVD, it was possible to grow a two-level CdTe structure with an undoped layer followed by a highly As doped cap layer.

Comparison of CdTe/CdS/TCO devices showed an increasing tendency for shorting in the structures with thinner CdTe. Structures were grown under conditions to give a smoother CdTe layer but the same tendency

Acknowledgements

The Authors gratefully acknowledge the support of EPSRC through GR/R20816/01 for the support of Dr Anne Stafford. This research was supported by Epichem who is gratefully acknowledged for their generosity and encouragement. The authors also wish to thank John Cambridge for the AFM measurements and Steve Jones for technical support.

References (12)

K. Durose et al.
J. Cryst. Growth
(1999)
R. Konenkamp et al.
Physica, E, Low-dimens. Syst. Nanostruct.
(2002)
C. Levy-Clement et al.
Physica, E, Low-dimens. Syst. Nanostruct.
(2002)
R.A. Berrigan et al.
J. Cryst. Growth
(1998)
A. Hartley et al.
Thin Solid Films
(2001)
S.J.C. Irvine et al.
J. Cryst. Growth
(2000)

There are more references available in the full text version of this article.

Cited by (24)

Combined experimental and DFT-TDDFT computational, structural and study effect of inter-diffusion Cu into CdTe thick film by annealing for optoelectronics
2021, Journal of Molecular Structure
Citation Excerpt :
Britt and Ferekides [6] have reported that the efficiency of CdTe/CdS solar cell 16.8 percent. Due to its low evaporation temperature, polycrystalline CdTe films can be equipped using a number of techniques [7–13]. One of the different techniques that have shown promising results is the technique of close spaced sublimation (CSS).
In the present study, thermal evaporation techniques were used to prepare Cadmium Telluride (CdTe) films with thickness 1 μm on a glass substrate (at constant temperature150 °C). By immersing these films in a solution Cu(NO₃)₂ (1g/1000ml) for 30 min, the prepared films were doped by ion exchange with Cu. Furthermore, the doped films were annealed in vacuum at different temperatures (25, 125, 250, 375, and 500 °C) for the diffusion of dopant Cu content, for 30 min. Using the method of Swanepoel, the film thickness as well as the refractive index of films was identified. The density functional theory (DFT-TDDFT) by DMol³ was used for the optimization of CdTe and Cu:CdTe as an isolated molecule in the gaseous phase. From using the DMol³/(GGA-PW91) method with DFT-TDDFT simulation, the HOMO and LUMO values for CdTe and Cu:CdTe isolated molecule are 2.41 eV and 1.66 eV, respectively. The MEP computations and Quantum-chemical calculations properties are computed and combined with the experimental study by using DMol³/(GGA-PW91) software. The possible transition in these as deposited and treated films are found to allow direct transition with reduction of the band gap from 1.50 to 1.37 eV with an increase of the temperature of the annealing. Dielectric constant, Energy loss functions (VELF and SELF) for CdTe as deposited and treated were addressed in terms of temperature rise in the annealing. The energy gap is also determined using photoluminescence (PL) and its values the corresponding wavelength peaks correspond roughly to the energy obtained through optical studies.
Effects of Cd<inf>1-x</inf>Zn<inf>x</inf>S alloy composition and post-deposition air anneal on ultra-thin CdTe solar cells produced by MOCVD
2017, Materials Chemistry and Physics
Citation Excerpt :
The use of less material will also have a positive impact for reducing manufacturing costs and carbon footprint providing that the PV conversion efficiency does not deteriorate significantly. However, ultra-thin (≤1 μm) CdTe solar cells are more susceptible to lateral inhomogeneity across the device [9,10] which is a limiting factor for industrial production. There have been several studies on ultra-thin CdTe PV cells [6,11–13], reporting a best cell efficiency of 11.2% for close spaced sublimated (CSS) CdTe cells with an absorber thickness of 0.6 μm [11] and an 11% best and 10% mean efficiency over 25 dot cells (0.06 cm2 cell area) for 0.5 μm thick CdTe produced by sputtering [13].
Ultra-thin CdTe:As/Cd_1-xZn_xS photovoltaic solar cells with an absorber thickness of 0.5 μm were deposited by metal-organic chemical vapour deposition on indium tin oxide coated boro-aluminosilicate substrates. The Zn precursor concentration was varied to compensate for Zn leaching effects after CdCl₂ activation treatment. Analysis of the solar cell composition and structure by X-ray photoelectron spectroscopy depth profiling and X-ray diffraction showed that higher concentrations of Zn in the Cd_1-xZn_xS window layer resulted in suppression of S diffusion across the CdTe/Cd_1-xZn_xS interface after CdCl₂ activation treatment. Excessive Zn content in the Cd_1-xZn_xS alloy preserved the spectral response in the blue region of the solar spectrum, but increased series resistance for the solar cells. A modest increase in the Zn content of the Cd_1-xZn_xS alloy together with a post-deposition air anneal resulted in an improved blue response and an enhanced open circuit voltage and fill factor. This device yielded a mean efficiency of 8.3% over 8 cells (0.25 cm² cell area) and best cell efficiency of 8.8%.
Physical device modeling of CdTe ultrathin film solar cells
2016, Solar Energy
The device physics of CdS/CdTe Ultrathin Film solar cells (UTF-SCs) is modeled using Sah–Noyce–Shockley (SNS) theory for the generation/recombination in the Space Charge Region (SCR) of an asymmetric ${pn}^{+}$ junction. This theory is based on Shockley–Read–Hall (SRH) recombination mechanism related to midgap defects in the band gap. The maximum effect of SRH is within SCR where the carriers are mainly collected by the electric field. Since the depletion width in a UTF extends to the back region, a UTF resembles an asymmetric ${pn}^{+}$ junction, which is most commonly modelled by SNS theory. Several materials and electrical parameters dominate the device physics of a fully depleted CdTe layer with $d_{CdTe} ⩽ 1$ μm. The uncompensated defect density, recombination lifetime, and CdTe thickness are parameters which effectively control the current–voltage characteristics and quantum efficiency of UTF devices. Simulation results indicate that any parameter that narrows the SCR of a UTF will reduce the open-circuit voltage whereas only thinning the CdTe layer lowers the short-circuit current density. The electrical resistivity and quantum efficiency are a measure of defect density/level and $d_{CdTe}$ or band bending at the junction, respectively. A practical takeaway of our simulation results is to measure the electrical resistivity of the CdTe layer before metallization of UTF solar cells. Band bending is a measure of junction quality. The latter parameters are usually ignored in device fabrication and modeling. However, any parameter related to defect density in UTF devices must be considered in the modeling since the front and back contacts are closer here and the degradation rate is rather probable. The degradation of UTF devices based on carrier collection mechanisms dominated by drift component is briefly discussed.
Influence of CdCl<inf>2</inf> activation treatment on ultra-thin Cd<inf>1-x</inf>Zn<inf>x</inf>S/CdTe solar cells
2015, Thin Solid Films
Citation Excerpt :
In order to produce CdTe solar cells without the effects of pinholes, the absorber thickness is typically 2–7 μm [5,6], but the majority of photons are absorbed in the first 1 μm [7,8]. However, the performance of ultra-thin CdTe solar cells (absorber thickness ≤ 1 μm) can be optically limited with loss of longer wavelength photons towards the CdTe band gap [8–10]. For future large scale production of CdTe solar modules to remain sustainable, consideration has to be given to the global availability of tellurium [1,8,11].
Ultra-thin CdTe photovoltaic solar cells with an absorber thickness of 0.5 μm were produced by metal organic chemical vapour deposition onto indium tin oxide coated boroaluminosilicate glass. A wide band gap Cd_1−xZn_xS alloy window layer was employed to improve spectral response in the blue region of the solar spectrum. X-ray photoelectron spectroscopy, X-ray diffraction and scanning electron microscopy were used to monitor changes in the chemical composition and microstructure of the Cd_1−xZn_xS/CdTe solar cell after varying the post-deposition CdCl₂ activation treatment time and annealing temperature. The CdCl₂ treatment leached Zn from the Cd_1−xZn_xS layer causing a redshift in the spectral response onset of window absorption. S diffusion occurred across the Cd_1−xZn_xS/CdTe interface, which was more pronounced as the CdCl₂ treatment was increased. A CdTe_1−yS_y alloy was formed at the interface, which thickened with CdCl₂ treatment time. Small concentrations of S (up to 2 at.%) were observed throughout the CdTe layer as the degree of CdCl₂ treatment was increased. Greater S diffusion across the Cd_1−xZn_xS/CdTe interface caused the device open-circuit voltage (V_oc) to increase. The higher V_oc is attributed to enhanced strain relaxation and associated reduction of defects in the interface region as well as the increase in CdTe grain size.
Pinhole treatment of a CdTe photovoltaic device by electrochemical polymerization technique
2012, Solar Energy Materials and Solar Cells
Citation Excerpt :
Pinholes can arise from random incomplete coalescence of the CdTe grains during deposition, or due to defects in the underlying surface [1]. Examination of the surface of a typical conductive oxide glass substrate using optical microscopy showed particulates consistent with surface contamination that can lead to pinholes during the growth of a thin film [1]. If these pinholes are left untreated, they lead to short-circuit or to shunting of the photovoltaic after deposition of the metal back contact.
An electrochemical treatment was developed to passivate pinhole defects in photovoltaic devices consisting of the device stack: glass substrate/SnO₂:F/CdS/CdTe/Cu/Au (metal/back contact). Samples were carefully examined and only those possessing a high density of pinholes through the CdS and CdTe layers were chosen for this study. The presence of a pinhole defect through these layers would lead to the formation of a wire-like connection through the photovoltaic device (shunt), which would seriously degrade the overall performance of the photovoltaic device, on deposition of the metal back contact. We report a solution to this problem wherein the pinhole defects are selectively filled with a resistive material such as polyaniline, using an electrochemical polymerization of the aniline monomer. After the polymer film deposition, this new layer acts as an insulator plug, and prevents the metallic back contact (Cu–Au) from making electrical contact with the transparent conductive oxide (SnO₂:F) layer, thus avoiding shunting of the photovoltaic device. Characterization techniques employed in this study are current–voltage measurements to measure the photovoltaic performance of treated and untreated devices and SEM–EDS microscopy and Grazing Angle X-ray diffraction to study the CdTe surface chemistry.
MOCVD of Cd <inf>(1-x)</inf>Zn <inf>(x)</inf>S/CdTe PV cells using an ultra-thin absorber layer
2012, Solar Energy Materials and Solar Cells
Citation Excerpt :
Hence, CdTe thicknesses from 2 μm to 10 μm are typically adopted. To further reduce the CdTe thickness without compromising the (ultra) thin film quality, a controllable process capable of producing uniform layers, such as metal organic chemical vapour deposition (MOCVD) [2–8], must be used. Irvine et al. [2] investigated the potential towards reducing absorber layer thicknesses in CdS/CdTe solar cells concluding that comprehensive substrate cleaning was required to minimise pin-hole formation.
Ultra-thin Cd_(1−x)Zn_(x)S/CdTe devices were produced by atmospheric pressure metal organic chemical vapour deposition (AP-MOCVD) with varying CdTe absorber thicknesses ranging from 1.0 to 0.2 μm and compared to baseline cells with total CdTe thickness of 2.25 μm. The ultra-thin CdTe layers (≤1 μm) were intentionally doped with As to induce p-type conductivity in the absorber. Cell performance reduced with CdTe thickness, with the magnitude of photo-current generation loss becoming more significant for the very thin CdTe layers. The decline in cell performance was lower than the optically limited performance relating to a decrease in shunt resistance, R_sh, especially for the thinnest cells due to areas of incomplete CdTe coverage and large presence of pin-holes leading to micro-shorts. Incorporation of Zn into the CdS window layer improved cell performance for all devices except when 0.2 μm thick CdTe was used. This improvement was markedly in the blue region owing to enhanced optical transparency of the window layer. External quantum efficiency (EQE) measurements showed a red-shift of the window layer absorption edge due to leaching out of Zn during the CdCl₂ treatment. Reduction of the CdCl₂ deposition time was demonstrated to recover the blue response of the ultra-thin cells.

View all citing articles on Scopus

View full text

Materials issues in very thin film CdTe for photovoltaics

Abstract

Introduction

Section snippets

Experimental

Characteristics of thin absorber layers

Conclusions

Acknowledgements

J. Cryst. Growth

Physica, E, Low-dimens. Syst. Nanostruct.

Physica, E, Low-dimens. Syst. Nanostruct.

J. Cryst. Growth

Thin Solid Films

J. Cryst. Growth