Third batch of samples: removing the N and thickening the spacers

The layer structure of the last fabricated batch of samples (referred to as “third batch”) is shown in Fig. 5.23. The new layer structure is designed to avoid two major problems encountered in the InAs/GaAsN QD cells: 1) shrinking of the bandgap produced by the dilute nitride and 2) quantum tunneling produced between QD layers. The first problem is tackled by replacing the strain-compensation GaAsN layers by thick GaAs layers, which dilute the strain throughout their 60 nm GaAs layer, thus preventing the appearance of defects derived from the accumulation of strain. The thick spacers also prevent (block) the quantum tunneling produced between contiguous QD layers as well as with the barrier material CB. This tunneling contributes to the so-called carrier escape [Antol´ın et al.,

(a) (b)

Figure 5.23: Layer structures of the third batch of samples. (a) 25 InAs/GaAs QD layer cell with thick spacers. (b) p-i-n GaAs reference cell.

2010a], making its associated rate (together with the thermal escape component) to be much higher than the photogenerated rate and causing the IB and CB electronic popula- tions to be indistinguishable in practice, which ultimately reduces the solar cell effective bandgap. The QD region is in this case 1,500 nm thick (25 layers times 60 nm per layer), as it is the intrinsic region of the GaAs reference sample. Regarding the rest of the layers of the structure, the solar cell is identical to the previous batch, including the fabrication of one doped and one undoped QD samples.

Figure 5.24: External QE of the devices from the third batch of samples at room temperature. Courtesy of Mrs. Esther L´opez.

At the time of writing this Thesis, very few characterization experiments have been carried out on this batch of samples. The most representative may be again the external QE, which is performed at room temperature and which results are shown in Fig. 5.24. In this case, the three devices seem to have suffered from problems at processing level, which

explains the low absolute QE values and the very different behavior in the supra-bandgap response.

The most significant improvement of this new designs is that both QD cells are now only limited by the GaAs bandgap and not by the smaller GaAsN bandgap, which implies that the below-bandgap response is exclusively originated by absorption in the QDs (and in the WL). This sub-bandgap absorption is again higher in the undoped sample, which is consistent with the previous corresponding explanation. The pick at λ=920 nm (E=1.35 eV), corresponds to the WL absorption.

5.4.4 Future design: InGaAs/AlGaAs QD solar cell grown on GaAs(311)B substrate with thick spacers

The designs shown in Fig. 5.25 upgrade the previous QD cell structures and propose a couple of important improvements regarding the important tasks corresponding to the enhancement of the below-bandgap absorption and the separation of the IB electronic population from that of the CB, which should ultimately allow the preservation of the voltage, even at room temperature. In this case, the solar cell will be grown on top of a GaAs(311)B substrate, so that a higher QD density can be achieved [Akahane et al., 2002] thanks to the higher crystal index, as explained in section 5.2. Another important feature will also have to be implemented: in this case, AlxGa1-xAs layers with a relatively high Al

content (close to the direct-to-indirect threshold) will have to be used for the spacer layers as well as for the p- and n-emitters.

(a) (b)

Figure 5.25: Sketch of the last layer structures proposed in the framework of this collaboration research program with the RCAST. (a) The proposed QD cell is grown on top of GaAs(311)B substrates. It consists of 25 stacked InGaAs/AlGaAs QD layers with 60 nm thick spacers. (b) p-ν-n AlGaAs reference cell.

The use of AlGaAs barriers theoretically allows a higher CBO which can produce an IB well separated (in energy) from the CB [Linares et al., 2011], leading to isolated electronic populations in both bands. Furthermore, IBSCs with larger fundamental bandgaps (i.e. closer to the optimum of 1.95 eV) account for larger limiting efficiencies. A BSF layer is also proposed as another improvement with respect to previous designs. It has to be implemented by means of a highly-doped AlGaAs layers with an increased Al content in order to create the appropriate minority carrier potential barrier (the same reasoning has to be followed for the window layer). A thicker emitter is also prescribed so that the cell has a better collection of the supra-bandgap part of the solar spectrum. The last modification affects the reference cell, which is meant to reproduce the conditions of the QD cell by hosting a 1,500 nm thick ν-region, instead of an intrinsic one. This ν-region is a lightly-doped region with the same doping level as the equivalent volumetric density of the direct-doped IB material region.

5.5

Summary

The fundamentals of the IBSC implemented with QDs are reviewed in this chapter, where a research line including the growth, processing and characterization of strain-compensated QD solar cells is carried out. Several batches of InAs/GaAs(N) stacked QD layer solar cells have been fabricated, processed and tested, leading to the general conclusion that the nitrogen introduced in the spacer layers limits the performance of the QD-IBSC because it reduces the effective bandgap of the device from approximately 1.4 eV to 1.2 eV. This strain-compensation layers can be substituted by thicker GaAs layers which dilute the strain and block the electronic tunnel escape. The performance of Si-direct doping in this structures is also studied. Finally, a new QD-IBSC structure based on an InGaAs/AlGaAs QD solar cell grown on a GaAs(311)B substrate is proposed to enhance the IR absorption and to effectively reduce the thermal escape component.

Low temperature concentrated

light characterization system

applied to IBSCs

6.1

Introduction

CPV is nowadays an increasingly important field for both research and industry [Sala and Ant´on, 2011]. The use of concentrated light techniques allows for significant reduction of the solar cell size, leading to a dramatic decrease of the amount of semiconductor used in a PV system [Swanson, 2003]. Within this paradigm, cheaper PV module components (e.g. lenses) are used in exchange for a significant part of solar cell area, together with the yet-large room for improvement of the concentrator solar cell itself (and of the whole system).

The specific conditions of operation of concentrator solar cells require the use of ad-hoc I-V characterization tools, that will be described in this chapter. Electrical characteriza- tion applied to CPV is useful and necessary for the improvement of the concentrator solar cell performance in general, both at material growth and processing levels. Besides, for reasons that have been discussed elsewhere throughout this document, the characterization of IBSCs under concentrated light can also provide valuable information on whether these devices operate in the way the IB theory predicts. Thus, the electrical characterization under concentrated light is an indispensable tool for the research on IBSCs.

The use of ad-hoc concentrated light characterization techniques is not new in itself [Emery, 1986]. Furthermore, many research groups worldwide have focused on this topic throughout recent years and even some companies that belong to the PV industry are willing to have reliable concentrator cell and CPV module characterization set-ups [Keogh

et al., 2004, Pravettoni et al., 2010, Dom´ınguez et al., 2008]. Therefore, after a brief description of the concentrator cell characterization technique and the different possible strategies for implementation, this chapter will deal with the specific features that our concentration set-up incorporate in order to adapt it to the IBSC research needs.