Interface Optimization for Improved Photovoltaic Devices

Glaudell, "The Legacy of William Shockley: Racism and Ableism in STEM," in 2021 IEEE 48th Photovoltaic Specialists Conference (PVSC), R.G. The colored balls correspond to the relevant atoms as follows: purple = arsenic, black = carbon, green = gallium, pink = hydrogen, red = oxygen, yellow = sulphur, blue = nitrogen. a) The DTT binds to the GaAs surface at the sulfur atoms on either side of the molecule.

NOMENCLATURE

The process by which the atoms on the surface of a crystal take on a structure different from that of the bulk. A measurement technique in which the atoms of a crystal cause an interference pattern of waves present in an incident X-ray beam, providing information about the atomic spacing of the lattice.

INTRODUCTION

Battling the Climate Crisis with Photovoltaics
Crystalline Silicon Photovoltaics
Silicon Heterojunction Devices
GaAs Photovoltaics
Introduction to ZnS x Se 1 −x Deposition
Fundamentals of Molecular Beam Epitaxy
Non-epitaxial Deposition

One of the most important consequences of the energy revolution that enabled the industrial revolution is climate change. The equilibrium pressure 𝑝𝑒 𝑞within the effusion cell cavity is given by the vapor pressure of the heated material.

Figure 1.2: Average annual energy consumption per person in Great Britain by energy source for decades between 1561 and 1859 [2]

CHARACTERIZATION OF II-VI FILMS

Growth Rate and Film Morphology Thickness MeasurementsThickness Measurements

Ray Reflectivity

Uniform etchant coverage was impossible due to surface tension, and sometimes the surface tension broke and the etchant spread out of the substrate. In this work, RHEED imaging was used to evaluate surface preparation techniques and determine the type of thin film deposition.

Figure 3.1: Reflection high-energy electron diffraction images of < 211 > c-Si before (a) and after (b) deposition of ZnSe:Al

Ray Diffractometry

Electrical Properties Resistivity MeasurementsResistivity Measurements

Where 𝑤 is the width of the sample in the y direction, the electric field is related to the Hall voltage by 𝐸𝑦 = −𝑉𝐻. Specifically, the high temperatures (up to 550◦C) involved in the standard ZnSxSe1−x deposition preparation were expected to significantly degrade the lifetime of the passivated c-Si.

Figure 3.2: Diffracted x-ray intensity as a function of detector angle from 2 𝜃 − 𝜔 scan taken on ZnS x Se 1 − x films in the Bragg-Bretano geometry, aligned to the offcut Si -<001> substrate

BAND ENERGETICS OF PHOTOVOLTAIC CONTACTS

Experimental Determination of Energy Band Alignment Valence Band OffsetValence Band Offset

The binding energy position of the Si 2p 3/2 peak was used as a reference peak at the substrate core level. Depending on the conductive nature of the sample of interest, sample charging may occur during XPS. The aim remains to calibrate the binding energy scale of the obtained photoelectron spectra.

However, when multiple chemical states of the same element are present in the sample (e.g., Ga-As, Ga1+ oxide, and Ga3+ oxide on a GaAs surface), calibration of the binding energy can make chemical state identification easier.

Figure 4.1: The offset in valence band energy at the interface between two materials, given in the energy band scale, is calculated from the valence band maximum position of the materials in bulk, given in the binding energy scale, and shift in VBM positio

The discrepancy between the referenced Zn/Se and S data is due to the overlap of the S 2p core level peaks with the Se 3s peaks and a Si 2s plasmon loss peak. The relative intensity of the Se 3s doublet compared to the S 2p doublet increases with increasing Se fraction from top to bottom. As the Se 3s signal adds to the photoelectron signal mixture with increasing Se content, the double S 2p assembly becomes less certain.

The Se 3p doublet contributes signal to the S 2p region even when there is no sulfur content in the sample (ZnSe).

Figure 4.3: ZnS x Se 1 − x /Si conduction band offsets as a function of mole fraction, 𝑥

III-V Semiconductor Passivation Characterization

To relax the GaAs surface, we modeled two layers of the <100>-GaAs surface and attached the bottom layer to the undisturbed bulk structure. The colored balls correspond to the relevant atoms as follows: purple = arsenic, black = carbon, green = gallium, pink = hydrogen, red = oxygen, yellow = sulphur, blue = nitrogen. a) The DTT binds to the GaAs surface at the sulfur atoms on either side of the molecule. The center Ga atom of the displayed unit cell at the GaAs/NHC interface is significantly displaced by the bond to the carbene center.

The octanethiol layer is the best of the organic layers at room temperature, with an SRV of 250 cm/s.

Figure 4.6: Conduction band offset at the ZnS x Se 1 − x /Si interface vs mole fraction.

Conclusions

The radiative lifetime (no photon recycling) is shown for reference at 150 ns in gold triangles⊳. All organic chemical passivators achieve a minimum SRV of 200 cm/s, corresponding to a photon recycling factor of 2.7. To our knowledge, this is the first report of photon recycling in chemically passivated GaAs films.

Passivation fails with increasing temperature for each layer, starting with the NHC at 100◦C, the thiols at 150◦C and the sulfide above 200◦C.

Figure 4.18: TRPL-determined surface recombination velocities of GaAs passi- passi-vated with GaInP on the rear and GaInP (DHJ, grey square), sulfide (red circle), octanethiol (blue triangle △ ), dithiothreitol (green triangle ▽ ), and n-heterocyclic carbe

SIMULATION OF II-VI CARRIER-SELECTIVE CONTACTS

Transmission Probability at Contact Interfaces

2 is in the depletion region where the magnitude of the band bending brings the CSC conduction band below the electron energy. Prior to knowledge of the actual conduction band gap, or to explore the influence of a tunable conduction band gap (perhaps based on surface treatments), the exploration of transmission probability as a function of band gap and dopant concentration for a given material can guide the prioritization of experimental measurements. We can see a general tendency that when the conduction band offset is more than the electron energy (the electron must tunnel), the transmission probability is close to zero except at high doping.

After the conduction band shifts were measured, another specific layer was added to the model.

Introduction to Photovoltaic Device Simulation Sentaurus TCADSentaurus TCAD

The energies of the incident electrons were chosen as test cases based on the energy difference between the conduction bands of the bulk Si substrate and II-VI CSC. Of course, the energy of the incident electrons is not a single value, but a distribution based on the wavelength of the exciting photon where the light was absorbed in the device and any loss processes the electron may have undergone before reaching the contact. The transfer probability is close to 1 if the conduction band offsets are smaller than the electron energy, and close to 0 for larger offsets, except at high donor concentrations.

Most importantly, this involves parameterizing the properties of mole fraction materials such as ZnSxSe1−x as a function of 𝑥.

Figure 5.1: Electron transmission probabilities across ZnS (a, b) and ZnSe (c, d) conduction band barriers as a function of ZnS x Se 1 − x donor concentration 𝑁 𝐷 and conduction band offset for assumed incident electron energies of 1 eV (a, c) and 0.5 eV (

Second, an SHJ-style cell with a p-type c-Si absorber, p-type hole-selective a-Si back contact, and n-type electron-selective ZnSxSe1−xfront contact. Third, an SHJ-style cell with a p-type c-Si absorber, p-type hole-selective a-Si back contact, an intrinsic a-Si passivation layer on the front, and an n-type electron-selective ZnSxSe1−x front contact. In addition, modified versions of the contact-specific and device-level metrics for passivation, conductivity and selectivity, introduced in [41], were used to more specifically analyze the impact of design variation.

In both types of traps, 𝜎, or 𝐸 in Sentaurus, documentation is characteristic of the energy dissipation of the defects.

Standard Device Metrics

The main contributors to the series resistance are the contact resistance and the large absorber resistance. For a particular photovoltaic device, the short-circuit current depends on the surface area of the solar cell, the power (intensity) and spectrum (dependence on the wavelength) of the incident light, the optical properties of the solar cell, and the collection probability. of cargo carriers. The AM1.5G spectrum is calculated from the AM0 spectrum (no air between sunlight a cell) and is representative of sunlight striking a solar cell at sea level, where the sun shines at 11.2◦ from the cell normal vector, the cell faces a blue sky and light sandy soil with no light concentration (G for “global” condition), and the sun is 41◦ above the horizon.

The optical properties of the solar cell and the probability of collection of the charge carriers vary between the designs considered in this study.

Carrier-Selective Contact Metrics

In each contact there is positive current from the hole population and negative current from the electron population to the single Fermi level of the electrode. The 𝑥 position of the outer edge of the absorber is 𝐿𝑎 𝑏 𝑠 and 𝐿𝑎 𝑏 𝑠 + 𝐿𝐶 is the outer edge of the contact. Only in (5.22), 𝑥 represents the position of depth in the device rather than mole fraction composition of the II-VI CSC.

The validity of the flux-based method can be determined by comparing the overall device selectivity metric.

Simulated Device Performance Results

The first two use the partial-specific contact resistance ratio definition (right-hand side of (5.20)), where𝜌𝑙/𝑟. The second two use the ratio of the 𝑉𝑜𝑐 to 𝑖𝑉𝑜𝑐. The implied 𝑉𝑜𝑐 can be calculated from the qFL separation or the carrier density at the center of the absorber. 5.23) 𝑁𝐴 is the acceptor concentration, 𝛿𝑛 the excess carrier concentration, and 𝑛𝑖 the intrinsic carrier concentration, all for the c-Si absorber. When calculated from bulk values (voltage ratios 5.3c, 5.3d), the sign of the selectivity metric does not change with location (in front or behind) of the electron-selective contact. The ZnSxSe1−x-only CSCs increase𝐽𝑠𝑐 and𝑉𝑜𝑐 compared to the reference SHJ simulation, and are more efficient than the champion ZnSxSe1−x+a-Si CSC design.

Given the large conduction band shift between ZnS and Si, it appears that the main advantage of the wider bandgap of ZnS (compared to the lower mole fraction ZnSxSe1−x), outweighs the losses due to the transport barrier of electrons in the heterojunction, but not enough. to outperform a standard p-SHJ design in simulation.

Figure 5.3: Full device selectivity metric S tot calculated four ways for simulated SHJ devices with ZnSe (x = 0) carrier-selective contacts as a function of CSC donor concentration 𝑁 𝐷 and thickness: (a) S tot ,𝜎 from carrier conductivities, (b) 𝑆

Proposed Designs and Sensitivities

The opposite is true for ZnS CSCs, where almost all current comes from electrons tunneling through the conduction band barrier. The device efficiency is more sensitive to shell medium density with a ZnS CSC than a ZnSe CSC. 5.8, the efficiency of ZnS CSC devices, while higher than ZnSe CSC devices, starts to drop significantly at 𝑁𝐷 = 9×1018 cm−3.

From top to bottom ZnS CSC with tunnel (blue squares), ZnSe CSC with tunnel (black triangle △), ZnSe CSC without tunnel (red triangle▽), and ZnS CSC without tunnel (green circles).

Figure 5.4: Device performance for selected subsets of simulations: efficiency of devices with (a) ZnS x Se 1 − x doping 𝑁 𝐷 = 4

VI–ON–C-SI PHOTOVOLTAIC DEVICES

Fabrication
Characterization

V Curves

Transition to Industry

Since the ZnS CSC in this study is an n-doped electron-selective contact and the remaining a-Si CSC is hole-selective, the order of contact deposition is changed compared to a standard process (n then p instead of p then n). Fabrication would proceed as follows: test incoming wafer, saw damage removal and surface texturing in KOH bath, HF dip for oxide removal, atomic layer deposition of ZnS carrier selective contact, plasma enhanced chemical vapor deposition (PECVD) of 5 nm thick intrinsic a-Si- passivation layer and 5-nm thick p-doped a-Si hole-selective layer, sputter behind transparent conductive oxide for bifacial option, screen print mask on front and back, plate metallic contacts, remove contact masks and characterize via current-voltage measurements and sort cells then . It is estimated that the costs per unit area in high-volume fabrication of ALD ZnS will be equivalent to the PECVD a-Si it replaces, so the levelized energy cost benefits of a ZnS-SHJ over a standard SHJ will be primarily through the increased efficiency.

ALD deposition of ZnS instead of PECVD of a-Si:H (i-passivation and then p-doped layers at the front), 5.

Figure 6.2: Current-voltage curve (black, left ordinate) and power-voltage curve (blue, right ordinate) for ≈ 0

GROWTH OF 2D II-VI SEMICONDUCTOR LAYERS BY HYBRID LAMINATION

Introduction to 2D II-VI Semiconductor Layers
Fundamentals of Hybrid Lamination
Free-standing II-VI Layers
Band Alignment
Future Work
Conclusions

The reaction temperature and time were varied to optimize for the desired phase yield. With the weaker bonding between planes II-VI enabled by the organic ligands, separation of the layers can be accomplished simply through sonication in the solvent. The color of the powder product at different stages of processing changed from red to red, red-orange, brick, or even black.

Raman spectra were taken at different points in such a sample and the variation in features indicates the non-uniform composition of fully processed Zn2Se2.

Figure 7.1: X-ray diffraction spectra from 2 𝜃 - 𝜔 scan on [(Zn 2 Se 2 )(pa)]. Black line is raw data

FINAL REMARKS

For epitaxy-free GaAs photovoltaic devices, band alignment measurements have identified several candidates for electron- and hole-selective contacts. Furthermore, high resolution x-ray photoelectron spectroscopy studies of GaAs passivation layer bonding conditions showed that the GaAs surface can be passivated in the presence of Ga1+ oxide, leading to carrier lifetimes of 300 ns or more, stable under heating up to 200◦ C. Freestanding sheets of differently structured II-VI semiconductors have been studied as candidates for solution-processable carrier-selective contacts, where optoelectronic properties can be tuned with layer number and stacking partners.

High-resolution X-ray photoelectron spectroscopy measurements of Zn2Se2on c-Si showed a type III broken-gap band alignment with the Zn2Se2 valence band above the c-Si conduction band, suitable for a hole-selective switch in a photovoltaic device or for other electronics applications.

BIBLIOGRAPHY

Barraud, et al., “Kasaaraa humnaa fuuldura seelii aduu silikoonii heterojunction,” IEEE Joornaalii Footovoltaayikii, jildii. Nishikawa, et al., “Madda kompaawundii molakiyuulaa biimii epitaaksii caasaa laayizerii II-VI,” Joornaalii Guddina Kiristaal, jildii. Rodriguez, et al., “Qophii fuula silikoonii III-v molecular beam epitaxy,” Joornaalii Guddina Kiristaal, vol.

Sherwood, et al., "Practical Guide to Curve Fitting in X-ray Photoelectron Spectroscopy," Journal of Vacuum Science.