Project topics and materials
Get Complete Project Material File(s) Now! »
From atoms to semiconductors
Silicon is the most used semiconductor material. The characteristic properties of semicon-ductors come from the interaction of their constituting atoms when brought together into a crystal.
Energy bands formation
Having 14 protons, isolated silicon atoms are accompanied by the same amount of electrons, which are distributed into orbitals with the following configuration: 1 22 22 63 23 2. The Bohr model of the silicon atom is represented in Figure I•2, with its 4 valence electrons on the latest orbitals ( = 3). Each of these orbitals corresponds to one discrete energy and momentum states for their elec-trons. However, atoms are never actually isolated: crystal formation promotes their interac-tion. Figure I•3 is a representation of the energy levels in silicon as a function of interatomic space.
When atoms are brought close to each other, their valence orbitals are disrupted and start to allow an increasing number of energy levels: they become bands ①. If atoms are moved closer, the 3 and 3 orbitals hybridize as the bands merge together ②. Finally, what actually happens in crystalline materials occurs at interatomic space. Four cases are possible, as depicted in Figure I•4:
• The hybridized 3 − −3 bands split into one high and one low energy bands, respectively named conduction and valence bands, (CB and VB) which are separated by a forbidden energy band gap. This is the case for silicon (③ in Figure I•3), where lattice parameter is 5.43 Å and bandgap is 1.12 eV). If the bandgap is small, external energy excitations are enough for electrons to jump from VB up to CB: the material is semiconducting.
• However if the band gap is large and the valence band is filled, thermal energy and solar radiation are not high enough to overcome the barrier: the material is insulating.
• If band splitting does not occur, then the conduction and valence band overlap: the ma-terial is a semimetal.
• If band splitting occurs and the valence band is partially filled, charge carriers can move freely: the material is a metal.
Doping
At thermodynamic equilibrium and 0 K, all electrons in pure — namely intrinsic — semicon-ductors are in the valence band, making it insulating. At ambient temperature, even in dark, thermal energy is enough for a small amount of electrons to cross the bandgap, which allows some carrier conduction. To increase the conductivity of silicon, one can alter the density of free carriers by introducing impurities.
Such impurities are chosen from the adjacent columns of the periodic table, from column V and column III, because they respectively present one more and one less valence electron. The incorporation of these impurities, that can donate or accept electrons from their neighbouring silicon atoms, is called doping. Figure I•5 illustrates this concept.
Fermi level is result of the distribution probability of electrons. If a material were to allow a continuum of energy levels in its bandgap. The energy level at Fermi level would have 50% chance to be occupied. As shown in Figure I•6, p- and n-type have similar effects on the energy bands, their carrier population and the resulting conductivity and Fermi level.
• N-type doping is performed with donor atoms, such as phosphorus, as its energy level is close to the conduction band of silicon and thermal excitation is enough for its additional electron to jump into the conduction band. Conductivity is then favoured by the addi-tional electrons in the conduction band. As there are more electrons in the conduction band, Fermi level is risen.
• P-type doping is obtained with acceptor atoms, such as boron. Being close to the valence band, boron energy level allows its lack of electrons to be exchanged with electrons from the valence band. Being perpetually filled by neighbouring electrons, this lack of electrons moves and behaves like a quasiparticle named hole. Conductivity is then enhanced thanks to the holes in the previously filled valence band. As there are less electrons available, Fermi levels decreases closer to the valence band.
Direct and indirect bandgaps
Bandgaps can be indirect or direct, depending on whether electrons need a change of mo-mentum to cross the bandgap, or not. For example, silicon has an indirect bandgap, while the bandgap of indium arsenide is direct (see Figure I•7). Direct bandgap materials are useful for optoelectronics, as they can easily emit light. For indirect bandgap materials, electrons crossing the bandgap are assisted by the absorption of a photon and a phonon. Phonons are quasiparti-cles representing crystal vibrations, naturally occurring in crystalline materials thanks to the elasticity of the lattices and thermal excitation.
Light absorption and electron-hole pairs generation
Silicon bandgap is indirect and 1.12 eV at room temperature. This means that any photon whose wavelength is 1110 nm can be absorbed if silicon temperature is above 0 K — which is true for normal operating temperatures, ignoring bandgap contraction above room temperature. Higher energy photons are also absorbed, but excess energy is thermally lost (see the absorp-tion mechanism in Figure I•8).
Considering the whole flux of light that can be absorbed by silicon, the number of excess mi-nority carrier generated in p-type silicon, i.e. electrons, is the so-called injection level Δn. Once a photon is absorbed, the excited electron crosses the bandgap up to the conduction band, and leaves a hole in the valence band. Both are able to move freely in the crystal, forming an electron-hole pair.
Recombination mechanisms in silicon
Given enough time, the electron will fall back into the hole, releasing its excess energy in form of photon or phonon. This process is called recombination.
Table of contents :
Introduction
I Context and state of the art
I·A From atoms to semiconductors
I·A·1 Energy bands formation
I·A·2 Doping
I·A·3 Direct and indirect bandgaps
I·A·4 Light absorption and electron-hole pairs generation
I·A·4·a Recombination mechanisms in silicon
I·B From semiconductors to solar cell structures
I·C Overcoming the limitations of silicon with III-V materials
I·D GaP/Si heterojunctions
I·D·1 Advantages of GaP over a-Si:H, and history of GaP/Si devices
I·D·2 Challenges of epitaxy of GaP on silicon
I·E State of the art on GaP/Si solar cells
I·F Objectives and outlines
II Processes and methods
II·A Material processing
II·A·1 Overview of the process flow
II·A·2 Crystalline silicon substrates
II·A·3 Metalorganic chemical vapour epitaxy of gallium phosphide: MOCVD
II·A·4 Plasma-enhanced chemical vapour deposition of hydrogenated amorphous silicon
II·A·5 Physical vapour deposition of Transparent Conductive Oxide (TCO)
II·A·6 Metallization
II·A·6·a Front side: Screen printing
II·A·6·b Back side: Electron beam evaporation
II·A·7 Wafer surface preparation and cleaning
II·A·7·a RCA-HF-O3 clean for storage
II·A·7·b Deoxidation
II·A·7·b·ⅰ Siconi™ etch before MOCVD
II·A·7·b·ⅱ HF deoxidation before PECVD
II·A·7·c Material etching procedures
II·A·7·c·ⅰ GaP etching with HF
II·A·7·c·ⅱ Removing GaP and metals : Aqua regia etching
II·A·7·c·ⅲ Silicon etching with KOH
II·B Characterization techniques
II·B·1 Materials morphology
II·B·1·a Layer thickness: Ellipsometry
II·B·1·b Surface roughness: Atomic Force Microscopy (AFM)
II·B·1·c Layer thickness, crystalline defects: Electron Microscopy
II·B·1·c·ⅰ Scanning Electron Microscopy (SEM)
II·B·1·c·ⅱ Transmission Electron Microscopy (TEM)
II·B·2 Materials composition
II·B·2·a Chemical composition and bonds: X-ray photoelectron spectrometry
II·B·2·b Active dopant concentration: Electrochemical capacity-voltage (ECV)
II·B·2·c Chemical composition: Secondary Ion Mass Spectroscopy: SIMS
II·B·3 Electronic properties
II·B·3·a Bulk resistivity: 4-point probe
II·B·3·b Charge carrier density: Hall Effect
II·B·3·c Carrier effective lifetime
II·B·3·c·ⅰ Quasi steady-state photoconductance: QSSPC
II·B·3·c·ⅱ Microwave photoconductance decay: μWPCD
II·B·3·c·ⅲ Photoluminescence: PL
II·B·3·d Solar cells performances
II·B·3·e Current-density–voltage curve: J–V
II·B·3·e·ⅰ J–V under illumination
II·B·3·e·ⅱ J–V without illumination : Dark-J–V
II·B·3·f Pseudo-J–V curve: Suns-Voc
II·B·3·g Internal and external quantum efficiency: IQE and EQE
II·B·3·h Pseudo-efficiency
II·C Heterojunction solar cells simulation: AFORS-HET
III Minority carrier lifetime degradation during GaP/Si solar cells fabrication
III·A Solar cells measurements
III·B Carrier lifetime degradation origins
III·B·1 Lifetime vs epitaxy steps
III·B·2 Discussion on the origin of minority carrier lifetime degradation
III·C Silicon bulk degradation analysis
III·C·1 Annealing temperature dependency
III·C·2 Reproducibility of degradation
III·C·3 Minority carrier lifetime distribution
III·C·4 Preventing contaminants diffusion
III·C·5 Contaminants detection
III·C·5·a SIMS
III·C·5·b Hall Effect Spectroscopy
III·C·6 Minority carrier lifetime evolution under illumination
III·C·6·a Lifetime vs time monitored by μWPCD
III·C·6·b Iron in boron-doped silicon
III·C·6·c Lifetime vs time measured by QSSPC
III·C·6·d Effect of trap dissociation on IQE
III·C·7 Contaminants quantification
III·C·7·a Iron concentration calculation from carrier lifetimes limited by recombination through iron levels
III·C·7·b Iron concentration estimation through IQE simulation
III·C·7·c Discussion on iron concentration determination
III·D Interface passivation
III·D·1 Determination of Sa-Si:H
III·D·2 Determination of SGaP and Sox
III·D·3 Discussion on the surface recombination velocity
III·E Conclusion on the minority carrier lifetime degradation
IV GaP/Si interface passivation
IV·A Surface reconstruction annealing in non-contaminating chambers
IV·A·1 Annealing in RTP systems
IV·A·2 Annealing in SiGe-MOCVD system
IV·A·3 Surface reconstruction on 4°-offcut wafers
IV·A·4 Conclusion on surface reconstruction annealing in non-contaminating chambers
IV·B GaP/Si template precursors
IV·C Hydrogen implantation for GaP/Si interface passivation
IV·D GaP/Si interface and bulk GaP contaminants
IV·D·1 Effect of air exposure
IV·D·2 Surface vs bulk GaP
IV·E Wetting layers for GaP epitaxy
IV·F Conclusion on the GaP/Si interface passivation
V Solar cells fabrication and results
V·A Solar cells fabrication and performances
V·A·1 Integrating decontamination steps for GaP/Si solar cells
V·A·2 Tested gettering techniques
V·A·2·a Implantation of phosphorus
V·A·2·b Diffusion of phosphorus
V·A·2·b·ⅰ Adapting the process flow
V·A·2·b·ⅱ Lifetime of precursors
V·B Solar cells fabrication and measurement
V·B·1 SiOx diffusion barrier
V·B·1·a Passivation of the solar cells precursors
V·B·1·b J–V and efficiencies
V·B·1·c Internal Quantum Efficiency
V·B·1·d Photoluminescence
V·B·1·e Morphology of the materials after decontamination steps
V·B·1·f Conclusion on the SiOx barrier for gettering
V·B·2 SiOx/SiN diffusion barrier
V·B·2·a J–V and efficiencies
V·B·2·b Internal Quantum Efficiency
V·B·2·c Photoluminescence
V·B·2·d Morphology of the materials before and after gettering
V·B·2·e SIMS analysis of the solar cells
V·B·3 GaP on unannealed silicon and GaP as window layer
V·B·3·a J–V and efficiencies
V·B·3·b Photoluminescence
V·C Conclusion on the solar cells fabrication and their performances
Conclusion and perspectives
