Ignacio Rey-Stolle

A GaAs solar cell with an efficiency of 26.2% at 1000 suns and 25.0% at 2000 suns

A GaAs solar cell without prismatic covers, with the highest efficiency known to the authors in the range of 1000-2000 suns for a single junction, is presented. Low temperature liquid phase epitaxy is used for its growth. In addition to improvements such as the achievement of a good quality material or a low contact resistance, this solar cell exhibits specific enhanced aspects. Among the most noticeable are: (1) an innovative design; (2) a double and gradual emitter layer; (3) a small size: 1 mm/sup 2/, (4) a finger width of the front metal grid of 3 /spl mu/m; and (5) a tailored ARC deposition based on a nondestructive and accurate AlGaAs window layer characterization. As a consequence, an efficiency of 26.2% at 1000 suns and 25.0% at 2000 suns AM1.5D (standard conditions) is achieved thanks mainly to a short-circuit current density at 1000 suns of 26.8 A/cm/sup 2/ (and 53.6 A/cm/sup 2/ at 2000 suns) with a simultaneous series resistance of 3 m/spl Omega//spl middot/cm/sup 2/.

A 3-D model for concentrator solar cells based on distributed circuit units

A three-dimensional (3-D) distributed model for high-concentrator solar cells based on elementary units made up of electrical circuits is presented. The recombination mechanisms are dealt with in detail, paying special attention to the perimeter properties. No ohmic effect is omitted making this a powerful simulation tool for concentrator solar cells. A shunt resistance is also included. The model allows the simulation of the external connections and nonuniform illumination profiles making this model very useful for optimizing future structures and technological processes. The proposed 3-D model is compared with a lumped, two-diode model in the simulation of a GaAs solar cell operating from 1 to 2000 suns. It is found that the 3-D distributed model agrees satisfactorily with the experimental data for all concentrations. The agreement cannot be made simultaneously for both low and high concentrations for the lumped model.

Study of non-uniform light profiles on high concentration III–V solar cells using quasi-3D distributed models

The quasi-3D models based on distributed circuit units are a powerful tool to analyse the performance of a solar cell from the point of view of its electrical behavior. Quite accurate models have been developed in the past that reproduce the experimental data of single-junction solar cells very closely. These models help in the determination of the origin of the peculiarities of the dark and one sun or high concentration experimental I–V curves. They also allow the design of the front grid, the analysis of the impact of the electrical parameters of the solar cell on the performance of the final device, etc. In this work, these models are used to study the effect of non-uniform profiles, generated by the concentrator optics, on the performance of a concentrator solar cell. The design of the front grid is then optimized to minimize the losses introduced by the light distribution, even taking into account the effect of the tracking system misalignment. As an introductory application example of multijunction solar cells analysis with this kind of modeling, the effect of the chromatic aberration on a double junction solar cell is presented.

A comparative study of BSF layers for GaAs-based single-junction or multijunction concentrator solar cells

An effective back surface field is a key structural element for a high-efficiency GaAs concentrator solar cell, either in a multijunction or in a single-junction device. In this paper, several BSF materials are analysed, namely: (1) p++GaAs(Zn), (2) p+Ga0.5In0.5P(Zn) and (3) p++Al0.2Ga0.8As(C). The results of the comparison demonstrate that the best option is C-doped Al0.2Ga0.8As, which exhibits a low series resistance and behaves as an excellent minority carrier mirror; p++GaAs(Zn) shows reduced minority carrier mirror properties resulting from Zn diffusion and p+Ga0.5In0.5P(Zn) is shown to produce important series resistance problems because of an unfavourable heterojunction with GaAs.

Modeling of the resistive losses due to the bus-bar and external connections in III-V high-concentrator solar cells

A model for the analysis of the resistive losses due to current flow through the bus-bar and external connections, mainly focused on III-V high-concentrator solar cells, is presented. Initially, a formulation that takes two-dimensional current flow into account in the bus-bar is proposed for calculating the bus-bar equivalent resistance. Next, a simplification assuming a one-dimensional (1-D) current flow is considered and a fully analytical model is obtained. Then, both models are compared and the applicability range of the 1-D approximation is established. The model is then applied to the analysis of the inverted square grid with several configurations for the external connections. Finally, the potential of the model is illustrated with a set of simulations carried out using a 1000/spl times/ concentrator GaAs solar cell and a number of conclusions of practical and technological interest are extracted.

The Interest and Potential of Ultra-High Concentration

The benefits of the ultra-high concentration (>1,000 suns) are shown in terms of cost reduction, raw material availability and efficiency increase. The main challenges for the operation at such high concentrations, namely, (a) to keep a low series resistance; (b) to manufacture tunnel junctions with high peak currents and low series resistance; (c) to design and build advanced characterisation methods and tools; and (d) to increase the reliability of the devices are addressed and solutions for them are proposed. As examples of success in manufacturing multijunction solar cells at ultra-high concentration, we have developed and manufactured aGaInP/GaAs dual-junction device, which exhibits an efficiency of 32.6% for a concentration range going from 499 to 1,026 suns. This efficiency is the worldrecord efficiency for a dual-junction solar cell. Besides, the efficiency is still as high as 31% at 3,000 suns. We have extended this strategy tolattice-matched GaInP/Ga(In)As/Ge triple junction solar cells. At the current state of development, such cells show an efficiency of 36.2% at 700 suns. The theoretical optimisation shows that an efficiency well over 40% at 1,000 suns is achievable.

III–V multijunction solar cells for ultra-high concentration photovoltaics

In this paper, the benefits of the ultra high concentration (¿ 1000 suns) are shown in terms of cost reduction and efficiency increase. Accordingly, the strategy followed at IES-UPM for more than 15 years is the development of III-V solar cells suitable for operation at 1000 suns or more. Recently, we have developed and manufactured a GaInP/GaAs dual-junction cell which results in an efficiency (measured at the Calibration Laboratory, CalLab, of Fraunhofer Institute in Freiburg) of 32.6% for a concentration range going from 499 to 1026 suns. This efficiency is the world record efficiency for a dual-junction solar cell. Besides, the efficiency is still as high as 31% at 3000 suns. The theoretical optimization of this solar cell (based on an accurate modelling) shows a potential efficiency over 36% at 1000 suns. We have extended this strategy to lattice-matched GaInP/Ga(In)As/Ge triple junction solar cells. First manufactured cells exhibit an efficiency of 31.5% at 1000 suns. The theoretical optimization of this cells show that an efficiency over 43% at 1000 suns is achievable. A roadmap has been established in order to reach this value.

Influence of PH3 exposure on silicon substrate morphology in the MOVPE growth of III–V on silicon multijunction solar cells

Dual-junction solar cells formed by a GaAsP or GaInP top cell and a silicon bottom cell seem to be attractive candidates to materialize the long sought-for integration of III–V materials on silicon for photovoltaic applications. One of the first issues to be considered in the development of this structure will be the strategy to create the silicon emitter of the bottom subcell. In this study, we explore the possibility of forming the silicon emitter by phosphorus diffusion (i.e. exposing the wafer to PH3 in a MOVPE reactor) and still obtain good surface morphologies to achieve a successful III–V heteroepitaxy as occurs in conventional III–V on germanium solar cell technology. Consequently, we explore the parameter space (PH3 partial pressure, time and temperature) that is needed to create optimized emitter designs and assess the impact of such treatments on surface morphology using atomic force microscopy. Although a strong degradation of surface morphology caused by prolonged exposure of silicon to PH3 is corroborated, it is also shown that subsequent anneals under H2 can recover silicon surface morphology and minimize its RMS roughness and the presence of pits and spikes.

Optimization of the silicon subcell for III-V on silicon multijunction solar cells: Key differences with conventional silicon technology

Dual-junction solar cells formed by a GaAsP or GaInP top cell and a silicon (Si) bottom cell seem to be attractive candidates to materialize the long sought-for integration of III-V materials on Si for photovoltaic (PV) applications. Such integration would offer a cost breakthrough for PV technology, unifying the low cost of Si and the efficiency potential of III-V multijunction solar cells. The optimization of the Si solar cells properties in flat-plate PV technology is well-known; nevertheless, it has been proven that the behavior of Si substrates is different when processed in an MOVPE reactor In this study, we analyze several factors influencing the bottom subcell performance, namely, 1) the emitter formation as a result of phosphorus diffusion; 2) the passivation quality provided by the GaP nucleation layer; and 3) the process impact on the bottom subcell PV properties.

In situ control of As dimer orientation on Ge(100) surfaces

We investigated the preparation of single domain Ge(100):As surfaces in a metal-organic vapor phase epitaxy reactor. In situ reflection anisotropy spectra (RAS) of vicinal substrates change when arsenic is supplied either by tertiarybutylarsine or by background As4 during annealing. Low energy electron diffraction shows mutually perpendicular orientations of dimers, scanning tunneling microscopy reveals distinct differences in the step structure, and x-ray photoelectron spectroscopy confirms differences in the As coverage of the Ge(100):As samples. Their RAS signals consist of contributions related to As dimer orientation and to step structure, enabling precise in situ control over preparation of single domain Ge(100):As surfaces.