Few subjects trigger as much passion as alternative or, more specifically, ‘green’ energy. Economical use of sunlight has been a key goal of scientists, engineers, and governments for over 50 years, given that the sun provides us hourly with enough energy to meet the annual electricity requirements of the entire world population.1 Solar cells (or photovoltaics: PVs) convert sunlight directly into electricity. Until recently, most PVs were based on single-crystal silicon. Yet, lack of grid parity (which refers to equal cost compared with existing, grid-based electricity sources) has discouraged their use. Recently, the PV industry has shifted towards adopting more exotic, thin-film materials, driven by the high sunlight-to-electricity conversion efficiencies attainable for cadmium telluride2 (CdTe) and copper indium gallium selenide—Cu(In,Ga)Se2—cells.3 Grid parity for thin-film modules (a packaged set of series-connected cells) is now almost certain.4 However, questions persist about their long-term reliability and, ultimately, the true cost of electricity: the longer a PV module functions, the cheaper it is.
To achieve lower costs, materials like CdTe and Cu(In,Ga)Se2 are deposited quickly, often using low growth (or fabrication) temperatures. These conditions result in polycrystalline structures with sizes and thicknesses on the order of microns, which (although packaged using technologically advanced materials) are inherently more sensitive to long-term, outdoor exposure (10–20 years) than their monocrystalline counterparts.
We employ a design-of-experiment approach combined with accelerated lifetime testing (ALT) to correlate the type of cell-fabrication process with performance and reliability.5 ALTs enable us to estimate the long-term robustness of solar cells within a short period of time. Most current ALTs apply heat to accelerate degradation and measure changes in either the module's power or the cell's current-density/voltage (J–V) output characteristics to quantify reliability.
Although simple in concept, such information is useful to evaluate different thin-film manufacturing processes, layer materials, and device structures. When performed at different temperatures, ALT studies reveal mechanistic activation energies associated with cell-degradation processes. For example, degradation in CdTe cells used to be associated with diffusion of both Cu (a commonly used CdTe dopant) and sulfur (S) from the cadmium sulfide (CdS) n-type window layer.6 Cu diffusion (activation energy Ea∼0:6eV) dominates degradation at higher temperatures, while S diffusion (Ea∼2:9eV) dominates at lower temperatures (see Figure 1). Prior to our study, the diffusion of S and the resulting Kirkendall void formation in CdS were unknown as a potential degradation pathway.
Figure 1. Arrhenius plot for calculating degradation-activation energies in cadmium telluride (CdTe) solar cells. Time to fail was set arbitrarily as the time in which cells lose 10% of their initial performance, which is a function of temperature, T. Ea: Diffusion-activation energy.
Capacitance-voltage (C–V) measurements are commonly used to determine doping profiles in semiconductor p-n-junction devices. In polycrystalline thin-film devices, these calculations are confounded by hysteresis associated with pre-measurement bias. Hysteresis has traditionally been attributed to majority-carrier trapping.7 However, contribution by an ionic-diffusion component has recently been suggested.8 We now incorporate bi-directional C–V data scans in ALT studies of cells and modules to capture information associated with hysteresis. These techniques are quick, easily incorporated, nondestructive, and sensitive to chemical and electronic changes in PV products.
Figure 2. Capacitance-voltage (C–V) hysteresis observed in CdTe cells (in μm) stressed at 100°C using two different window-layer materials. The greater rate of hysteresis observed in cadmium/zinc stanate (CTO/ZTO) cells correlates with increased cell degradation. ΔWd: Depletion-width difference.
Using this bi-directional-scan approach, we discovered that decreasing J–V performance in stressed CdTe cells correlates well with increasing C–V hysteresis9 (defined as the depletion-width difference, ΔWd, between forward- and reverse-direction voltage scans at zero volts). The latter increases more quickly when using next-generation cadmium/zinc stanate (CTO/ZTO), transparent-oxide, n-type window layers relative to more conventional tin-oxide films (see Figure 2). The faster CTO/ZTO-layer degradation and the resulting increase in deep-level trapping states appear responsible for the increased PV degradation observed when using CTO/ZTO layers.
C–V hysteresis is also employed to study industrial thin-film module reliability and the nature of their degradation mechanism. We recommend that certain stabilization procedures be adopted to consistently measure module performance indoors, in essentially the same way that irradiance and temperature are normalized to reference conditions when comparing outdoor performance.10 Hysteresis effects in thin-film modules can be considerable because of the additive voltage effect of series-connected cells.
In summary, we used ALT studies to estimate the reliability of different thin-film module fabrication processes, materials, and structures. These studies identify how cells and modules degrade. Our future efforts will focus on using C–V hysteresis to evaluate and improve the robustness of new thin-film cells. We plan to use these techniques in module-reliability studies to develop better thin-film module stabilization protocols. The first such improvement aims to determine whether expensive light-exposure protocols can be replaced by less costly, module-qualification procedures using forward-biased, dark exposures at elevated temperatures.
This work was supported by the US Department of Energy under contract DOE-AC36-08G028308 with the National Renewable Energy Laboratory (NREL).
David Albin, Joe del Cueto
National Center for Photovoltaics
National Renewable Energy Laboratory (NREL)
David Albin is a senior scientist with a PhD in materials science and a minor in electrical and computer engineering. He has more than 20 years experience in polycrystalline thin-film solar-cell development, over 100 publications, and holds three patents.
Joe del Cueto received his PhD degree from the State University of New York at Stony Brook in 1984. He joined NREL in 1997 and works on long-term, outdoor PV-module performance and reliability. His industry experience includes research and development in thin-film PV devices, modules, transparent-conducting oxides, and process development.
3. M. A. Contreras, K. Ramanathan, J. AbuShama, F. Hasoon, D. L. Young, B. Egaas, R. Noufi, Diode characteristics in state-of-the-art ZnO/CdS/Cu(In1-xGax)Se2 solar cells, Prog. Photovolt. 13, pp. 209, 2005.
http://188.8.131.52/uploads/media/11_First_Solar.pdf. First Solar company overview given by Dave Eaglesham (2007). Accessed 23 July 2009.