A general design guideline for strain-balanced quantum-wells toward high-efficiency photovoltaics

Highlights

• We propose a general design guideline of SB-MQWs for photovoltaics.

• Our design flow is demonstrated by applying to a top-cell and middle-cell example.

• Mobility degradation is analyzed with a tunneling and thermionic emission model.

• The impacts of carrier lifetime and background doping are investigated.

• This work is applicable to energy-conversion devices based on SB-MQWs.

Abstract

Due to the tunable optical absorption threshold and several intriguing physical properties, multiple-quantum-well (MQW) structures have been playing a crucial role in achieving ultra-high efficiency in tandem photovoltaic cells and various optoelectronic technologies. However, devices incorporating such nano-structures suffer from poor perpendicular carrier transport, which hinders device performance. On the other hand, designing MQWs is challenging. Simultaneously optimizing the structural parameters and constituent materials suggests a high design complexity with a large parameter space. In this report, on the basis of the quantitative analysis on the degradation of carrier transport, we propose a general design guideline for MQWs by which the structures and constituent materials can both be optimized targeting at a particular absorption threshold. We firstly demonstrate our design flow by optimizing an InGaP/InGaP strain-balanced (SB) MQW at 1.91 eV to replace In_0.49Ga_0.51P bulk serving as the top cell of dual-MQW triple-junction solar cells. Secondly, we apply the design concept to a conventional InGaAs/GaAsP SBMQW targeting at 1.23 eV, which is significant for not only high-efficiency single-junction photovoltaics but also current-matched tandem cells. The results illustrate the possibility of further improving current devices in terms of their constituent materials for QWs. Our design optimization method is considered to be applicable to any device based on QW structures.

Introduction

Thanks to highly-radiative and tunable optical property, multiple-quantum-well (MQW) structures have been extensively applied to various optoelectronic devices, such as lasers (Halliday et al., 1992, Nagarajan et al., 1992, Tessler and Eistenstein, 1993), LEDs (Bedair et al., 1984), optical amplifiers (Tiemeijer et al., 1993), modulators (Wood, 1988), photodetectors (Andrews and Miller, 1991), and solar cells (Adams et al., 2011, Ekins-Daukes et al., 1999, Fujii et al., 2014a, Fujii et al., 2012a, Razykov et al., 2011, Yamaguchi et al., 2008). In the tandem solar cells technology in particular, by dividing incoming solar spectrum equally between subcells with optimum absorption thresholds, the conversion efficiency is anticipated to be able to go beyond 50% (Geisz et al., 2018a, Geisz et al., 2018b). For instance, it has been suggested that a quad-junction solar cell with ideal bandgap combinations of 1.99/1.51/1.15/0.66(Ge) eV can achieve an efficiency of 50.1% by concentrator systems under 1000-suns (Thomas et al., 2014). To attain bandgap as low as ~1.0 eV for the middle cell, dilute-nitride alloy GaInNAs has been considered as a good approach for monolithic growth on germanium substrate (Courel et al., 2012a, Courel et al., 2012b, Friedman et al., 1998, Tukiainen et al., 2016, Yamaguchi et al., 2005). Additionally, a strain-balanced (SB) InGaAs/GaAsP MQW, a promising alternative approach, has been developed with an engineered effective bandgap down to the targeted value 1.15 eV successfully (Toprasertpong et al., 2016). However, when incorporating MQW structures into solar cells or the other aforementioned devices, carrier transport in the perpendicular direction suffers due to the introduction of the potential barriers at the heterointerfaces. Without being carefully designed, the collection probability of the photogenerated carriers tends to drop drastically, leading to limited energy yields.

Optimizing an MQW structure is challenging. There are typically four parameters that can be controlled when designing MQW structures based on compound III-V semiconductors: well thickness t_w, barrier thickness t_b, well material, and barrier material, as illustrated in Fig. 1a. Conventional design techniques involve strain-balance conditions (Ekins-Daukes et al., 2002), reducing the number of free parameters from four to three. This represents a large parameter space, which is challenging to explore systematically. In previous work, the optimization was done only partially, e.g. in reference (Kim et al., 2018) only the well material and the barrier thickness are modified while the well thickness is fixed; in reference (Fujii et al., 2014a) only the barrier material and barrier thickness are optimized, whereas the well thickness is fixed. Those past optimization schemes are essentially not satisfying for three reasons:

• Not all parameters are explored comprehensively. The optimum design can be found unlikely.

• It was trial-and-error. Because there were no comprehensive models to aim at a favorable design beforehand, the QW design needed to be further adjusted at the fabrication stage. Such a method is time-consuming and unprecise.

• MQWs with different absorption thresholds were compared. However, this is the most significant material requirement when designing a subcell with an optimum bandgap for tandem photovoltaic cells.

On the other hand, understanding the carrier dynamics behind is crucial to further improve optoelectronic devices based on QW. So far, conventional diagnosis of such degraded carrier transport in MQWs on the microscopic level has been relying mainly on the escape probability function derived from tunneling and thermionic-emission lifetime models (Fox et al., 1991, Nelson et al., 1993, Schneider and Klitzing, 1988). It was revealed as a good technique interpreting the transport dynamics involved in carrier collection as a solar cell from the macroscopic experiment results qualitatively (Fujii et al., 2012a, Sayed and Bedair, 2019). Nevertheless, the relative contribution of thermionic emission and carrier tunneling remains unclear. Furthermore, without estimating carrier transport in terms of their equivalent mobilities, it is difficult to quantitatively analyze how poor the perpendicular carrier transport could be. This prohibits one to fully exploit the tunability of transport properties of MQWs from the microscopic point of view.

In this research, to tackle these design and analysis conundrums, we propose a novel design flow including three essential steps:

• Step 1: Reduce the parameter space from 4 to 2, by setting two important design requirements for well thickness t_w and barrier thickness t_b: a strain-balance condition and an effective bandgap equal to a target value E_g,target. Then, with their critical layer thicknesses, assess the difficulty of performing heteroepitaxy given the design structures.

• Step 2: Use a recently-developed model allowing to quantitatively resolve effective mobilities for MQW structures as a function of the last 2 free parameters: well material and barrier material. This enables mapping of the effective mobilities with different material combinations.

• Step 3: By considering probable minority-carrier-lifetime ranges, deduce the cell efficiency. Based on several additional design constraints according to applications, we are able to pin down the optimum range for the composed materials and thicknesses and therefore derive the optimum MQW design.

This flow will define an optimum region to be explored experimentally, in terms of the 4 free parameters. It solves the three drawbacks highlighted above. By comparing only MQWs with the target absorption threshold, we are able to determine the optimum subcell design for high-efficiency tandem photovoltaic cells.

In the following, we first illustrate our design strategy in the specific case of InGaP/InGaP MQW (depicted in Fig. 1b) as a candidate absorber for the top cell in dual-MQW triple-junction solar cells. Such an architecture unveils enhanced tolerance against the fluctuation of the incoming spectrum owing to photon-coupling between MQW cells (Lee et al., 2012). However, replacing the conventional top In_0.49Ga_0.51P absorber by MQWs implies worse carrier transport. A detailed quantitative investigation of degradation mechanisms has never been reported and is in urgent demand. With this example, we present our full design procedures and resolve the carrier transport mechanism in MQW compared to the In_0.49Ga_0.51P bulk with identical absorption thresholds for the first time. Secondly, we apply the design guideline to optimize a 1.23 eV-InGaAs/GaAsP MQW absorber (depicted in Fig. 1c), which is significant for not only high-efficiency single-junction photovoltaics but also current-matched tandem cells (Barnham and Duggan, 1990, Katsuyama et al., 1987). Yet, the MQW absorber was studied within a small parameter space in previous research (Fujii et al., 2014a). With this example, we discuss the possibility of further improving the efficiency of the conventional design in terms of its constituent materials. Additional design constraints, such as background doping that hinders carrier collection and requirement for current matching to an ideal In_0.49Ga_0.51P top cell, are investigated as well. Our design technique is considered to be applicable not only to MQW photovoltaics but also to any optoelectronic device based on MQWs to explore the optimum combination of constituent materials toward better performance.