Project Topics
Get Complete Project Material File(s) Now! »
Quantum-structured materials
For the development of novel PV concepts, we want to use quantum-nanostructured heterostructures that are especially important for bandgap engineering of the absorbing material. For III-V materials, the implementation of these nanostructures usually takes place during the growth step.
Two methods are considered in our study for growing III-V semiconductor thin films: molecular beam epitaxy (MBE) and metalorganic vapor phase epitaxy (MOVPE). The latter is also known as metalorganic chemical vapor deposition (MOCVD).
Quantum nanostructured heterostructures for bandgap engineering
The main goal behind the growth of quantum heterostructures is to confine the electrical wave and trigger new properties inside the absorbing material. The difficulty is to obtain a monolayer precision with the same interface quality as in typical heterostructures. Confinement can be implemented in:
– One-direction, as it is the case for quantum wells (QWs).
– Two-direction, as it is the case for quantum wires (QWRs).
– Three-direction, as it is the case for quantum dots (QDs).
Usually the confining material with the highest bandgap energy is called the “barrier” for QWs and “host” material in the case of QDs.
Light management constraint on thickness and epitaxial transfer
In order to achieve light management in quantum-structured solar cells, we decided to focus on the multi-resonant approach discussed earlier using nanophotonics. In this approach, we wish to use a combination of resonant optical modes to enhance the light path for a certain spectral domain. We can make use of at least four different optical effects:
– Fabry-Pérot (FP) resonances corresponding to round-trips of light between the front and back surface of the cell.
– Guided mode resonances induced by periodical nanopatterns allowing incident light to couple to waveguide modes.
– Localized Mie resonance induced by the geometry of nanostructures.
– Near-field resonant effect (plasmonic resonances) in metallic nanostructures.
Thin FP cavity as a first step
For subwavelength thicknesses, the vertical confinement can be properly tuned to induce broad resonances in the structure thanks to FP vertical cavity. Afterwards, other resonant effects can be added by implementing dielectric or metallic nanostructures. The addition of different resonant effects contribute to the realization of a multi-resonant light management strategy.
Accordingly, a special care has to be taken to design, control and check the thickness of the layer stacking in the PV device in order to tune FP resonances in the spectral range of interest. In our study, we worked with samples having a FP cavity thickness below or equal to 500 nm. The constraint of having to work with very thin films is another difficulty for the fabrication of our device. We need precise control on the growth and fabrication processes.
Back mirror implementation
In Figure 2.1-4, we summarize two different ways of implementing a back mirror for solar cells by epitaxial growth or by epitaxial transfer. In both cases, we need to use a “target layer” that is grown between the substrate and PV active layers.
Metal deposition and lift-off
For metal deposition, we used three different types of machines: sputtering, thermal evaporation and electron-beam evaporation. We notice that the thermal evaporation machine we used could heat our samples up to 200°C, which damaged a serie of our transferred sample. The adhesion of metallic gold mirror on semiconductor surface was always checked (e.g. using commercial Kapton® tape) and we tried different techniques to enhance such adhesion. The best one we could find was to use a 5-10 minutes deoxidation bath with 37%-HCl commercial solution diluted at 20% in deionized water just before putting our samples in the vacuum chamber.
After metal deposition, we used a lift-off process as described in Figure 2.1-8. We used acetone to dissolve AZ5214 resist.
Rapid thermal annealing
Rapid thermal annealing (RTA) is used as a heat treatment for example in order to activate dopants, change interfaces, densify deposited films or change states of grown films. The annealing duration can range from a few seconds to several minutes. In our case, the RTA is made by an oven with lamp-based heating under reduced pressure with an argon-hydrogen flow. Heating results from the temperature conduction to graphite sample holder and protecting cap, which absorb the lamps radiation.
During the course of our work we tested RTA to try to get better ohmic contacts. However, we did not obtain any enhancement and even in some cases, we noticed a decrease in the performance of our cells. This decrease was attributed to defaults induced by metallic atom diffusion with heat. The Fabrication and nanofabrication for quantum structured solar cells metal diffusion length is thought to be non-negligible for our thin junctions. The RTA process is a difficulty that needs to be considered especially for solar cells transferred on a metallic mirror. The increase in temperature can have an effect on the interdiffusion of metal atoms in the semiconductor and induces defects or even be an issue for the adhesion of the epitaxial layers on the supporting substrate. For simplicity, we decided to stop using RTA for all of our samples (non-transferred and transferred). For 1-sun or low concentration PV irradiation conditions, annealing is thought to be unnecessary if the doping level of contact layers is high enough.
Anti-reflection coating
The goal of anti-reflection coating (ARC) is to increase the coupling between the incident light and the absorbing layers and therefore reduce the reflection. Many coatings consist of transparent thin film structures with alternating layers of contrasting refractive index. In our study, we used three different types of machines to deposit SiO2, SiNx and TiO2 layers: sputtering, e-beam evaporation and ion beam assisted deposition (IBAD). These deposition methods imply different temperatures but also different qualities for the deposited material or surface. As we showed in Figure 2.1-10 top right, the deposition of layers can induce strains and even damage the solar cell.
Wire-bonding
In the case of a collaboration, laboratories do not all have the same setup for characterizing solar cells sample. For electrical characterization, wire bonding can help to facilitate the connection and use of a sample. We used three different machines with ball bonding technique where the wire is attached using a combination of downward pressure, ultrasonic energy and heat. For nontransferred samples, this technique was well-controled. However, in the case of transferred solar cells even after optimization of different parameters we found a success ratio less than one out of five. We consider the failure is because of the very thin transferred layers. Another way of connecting wires to the cell was attempted with epitaxy glue. However, after the solidification, the glue seems to induce strain that also damage the very thin layers.
Finally, we consider that wire-bonding needs to be better optimized and should be reduced to its minimum use. Therefore, we advice that during the fabrication metal contacts should be made large enough so that it is not a limitation for the characterization. We reckon that a 200 μm metal pad is large enough for spring probing by eyes and can be easily found in many laboratories.
Challenges of the WI-ELO process
In Figure 2.2-5, we show a photo of our WI-ELO setup. A weight is bending the flexible supporting substrate while we see the HF droplet etching the edge of the sample to detach the epitaxial layers from the GaAs substrate. A close P.E.T. beaker resistant to HF is used to avoid evaporation of HF. Many parameters need to be optimized to have a fast and succesfull WI-ELO process for example the thickness of the sacrificial layer, the weight, the HF volume and concentration of the droplet. After careful investigation, we found that the sacrificial layer thickness should be around a dozen of nanometers. This result is coherent with Schermer’s group findings. For hundreds of nanometers the sacrificial layer is not etched homogeneously and some cracks are appearing that are damageable for the fabrication of solar cells as shown in Figure 2.2-6.
Table of contents :
Contents
Acknowledgements
Abstract
Résumé
General introduction
Introduction to photovoltaics and light management
1.1 Photovoltaics: towards high-efficiency solar cells
1.1.1 Principle of photovoltaic conversion of solar energy
1.1.2 Novel concepts for high-efficiency PV conversion of solar energy
1.2 Light management for quantum-structured IBSCs
1.2.1 Lambertian scattering PV light trapping
1.2.2 Application to quantum-structured solar cells
Conclusion
Fabrication and nanofabrication for quantum structured solar cells
2.1 Sample fabrication
2.1.1 Stacking design and growth
2.1.2 Device fabrication
2.1.3 Additional processes
2.2 Epitaxial transfer processes
2.2.1 Etch stop layer technique
2.2.2 Sacrificial layer technique
2.3 Nanofabrication
2.3.1 Soft nanoimprint lithography
2.3.2 Application to solar cells
QD-IBSC system: from ideal to experimental PV material
3.1 Description of QD-IBSCs system for light management implementation
3.1.1 Choice of material: In(Ga)As QDs in Al0.2GaAs host material
3.1.2 Position of QD layers
3.2 Fabrication of QDs absorber suitable for IBSCs
3.2.1 Growth of QDs by MBE
3.2.2 Control and improvement of QDs growth
3.3 Investigation and modeling of QDs properties
3.3.1 QDs interactions with host material: thermal and electrical properties
3.3.2 Simulation of different QDs: shape, aspect ratio, QWIs, WL, In content .
3.4 QDs absorption: simulation and experimental determination
3.4.1 Absorption from k.p calculation
3.4.2 Absorption measurements
Intermediate band dynamic study to evidence novel PV concept
4.1 Calibrated photoluminescence
4.1.1 Characterization setup
4.1.2 Balance of photons
4.1.3 Quasi-fermi level splitting
4.2 Two-color excitation photoluminescence
4.2.1 Characterization setup
4.2.2 IR pump effect on photoluminescence
4.2.3 Steady-state rate equation model
4.3 Potential of a hot carrier thermally activated QD-IBSC or IB-HCSC
4.3.1 IB-CB QFL splittings
4.3.2 Fermi-Dirac distribution inside QDs
4.3.3 Potential efficiency of a thermally activated QD-IBSC
Light management strategies applied to quantum-structured solar cells
5.1 Application to multi-quantum well solar cells
5.1.1 Fabry-Pérot cavity effect on MQW solar cells
5.1.2 Nanopatterns effect on MQW solar cells
5.2 Application to quantum dot solar cells
5.2.1 Fabry-Pérot effect on QDSCs
5.2.2 Nanopatterns strategy for QDSCs
Conclusion
Conclusion and perspectives
Appendices
A. Fabrication protocols
B. InGaAs QD growth issues
C. Method to fit the PL from a QD ensemble
References
