Back Surface Reflection Research Articles

Optical Path Length Factor at Near-Bandgap Wavelengths in Si Solar Cells

By using the Rand and Basore (R&B) physical model for light trapping, a particular analytical expression for near-bandgap optical path length factor Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> is derived for the case where back surface reflectivity R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> is equal to unity (R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> = 1). This expression shows that according to the R&B physical model, at R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> = 1, Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> has a finite value. At R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> = 1, the R&B expression for Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> is shown to diverge. By using the original R&B derivation procedure, a new expression for Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> is obtained, which at R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> = 1, reduces to the aforementioned particular expression for ideal back surface reflectance. The new expression is immune from the physical incongruity and tendency to overestimate Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> that affect the R&B expression. By using the new expression, a relationship between Z <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">o</sub> and the Yablonovitch and Cody (Y&C) absorption-enhancement factor is derived for devices where parasitic absorption predominantly occurs at backside reflector. This relationship is shown to be in agreement with the literature. In the derivation, a connection between the approaches to light trapping of R&B and Y&C is exploited. This connection allows obtaining a new method to evaluate R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">BACK</sub> at near-bandgap wavelengths in most Si solar cells. The new method is checked on two real Si solar cells and found to provide plausible results. Compared to the method used by R&B, the method presented in this paper is shown to be more reliable

Attenuation coefficient estimation using experimental diffraction corrections with multiple interface reflections

The ultrasonic attenuation coefficient of a fluid or solid is an acoustic parameter routinely estimated for the purpose of materials characterization and defect/disease detection. This paper describes a broadband attenuation coefficient estimation technique that combines two established estimation approaches. The key elements of these two approaches are: (1) the use of magnitude spectrum ratios of front surface, first back surface, and second back surface reflections from interfaces of materials with plate-like geometries, and (2) the use of an experimental diffraction correction approach to avoid diffraction losses. The combined estimation approach simplifies the attenuation coefficient estimation process by eliminating the need to explicitly make diffraction corrections or calculate reflection/transmission coefficients. The approach yields estimates of the attenuation coefficient, reflection coefficient, and material density. Models, experimental procedures, and signal analysis procedures, which support implementation of the approach, are presented. Attenuation coefficient and reflection coefficient estimates are presented for water and solid samples with estimates based on measurements made with multiple transducers.

On the frustration of back-surface reflection from transparent substrates in ellipsometry

The effects of back-surface reflection from transparent substrates on ellipsometricmeasurements are discussed and it is shown that their effects can be minimized by placingthe substrate on a commercially available product, Blu-tack. The ellipsometric properties ofBlu-tack in the range 250–850 nm are presented and the frustration of the back-surfacereflection is discussed in terms of index matching between substrate and Blu-tack.

03/01842 Structural and electrical properties of CuGaS 2 thin films by electron beam evaporation : Jeong, W.-J. and Park, G-C. Solar Energy Materials and Solar Cells, 2003, 75, (1–2), 9–100

In order to manufacture high-efficiency Si solar cells with a passivated rear surface and local contacts, it is necessary to develop both an excellent rear-passivation scheme compatible with screen-printing technology and a robust patterning technique for local contact formation. In this work, we have fabricated Si solar cells on ∼130 μm thick substrates using manufacturable processing, where rear side was passivated with a plasma-enhanced chemical vapor deposited SiOx/SiNx/SiOxNy stack and local back contacts using laser. As a result of both the rear surface passivation stack and the laser-fired local contacts, cell efficiencies of up to 17.6% on a 148.6 cm2 Float-zone Si wafer and 17.2% for a 156.8 cm2 multicrystalline Si wafer were achieved. PC-1D calculations revealed that the cells had a back surface recombination velocity (BSRV) of ∼400 cm/s and a back surface reflectance (BSR) of over 90%, as opposed to standard full Al-BSF cells having a BSRV of ∼800 cm/s and a 70% BSR. This result clearly indicates that the new technique of the passivation scheme and the patterning using laser developed in this study are promising for manufacturing high-efficiency PERC-type thin Si solar cells.

Textured monocrystalline thin‐film Si cells from the porous silicon (PSI) process

AbstractWe have fabricated a textured monocrystalline Si solar cell with a thickness of 15·5 μm and a confirmed efficiency of 12·2% using porous silicon (PSI) for layer transfer. The PSI process avoids photolithography and high‐temperature oxidation. The cell has a surface that is textured with randomly positioned inverted pyramids for light trapping. The device does not yet fully exploit the light‐trapping capability of this film shape, owing to a small back‐surface reflectance. Copyright © 2001 John Wiley &amp; Sons, Ltd.

Photon recycling and recombination processes in 0.53 eV p-type InGaAsSb

Transient radio frequency photoreflectance measurements were performed on 0.53 eV p-type InGaAsSb double heterostructures, grown by organometallic vapor phase epitaxy on lattice matched GaSb substrates, for determining excess carrier lifetime. Direct evidence of photon recycling was observed by changing the GaSb backsurface reflectivity and observing the change in excess carrier lifetime. Consistent with theory developed for this type of structure, effective lifetimes increased by 30%–40% when the backsurface was changed from an absorbing to a reflecting surface. The theory develops a closed-form expression for the total radiative recombination rate, starting with continuity equations for both excess minority carriers and the photon density. Lifetime measurements in these p-InGaAsSb structures with different doping concentrations and epitaxial layer thicknesses allow the extraction of minority carrier recombination parameters. Measurements indicate a value for the radiative recombination coefficient of B=5–6×10−11 cm3/s, for the Auger recombination coefficient of C=2–5×10−29 cm6/s, for the Shockley–Read–Hall lifetime of 100–150 ns and for the surface recombination velocity of 1–2×103 cm/s.

Measurement of transparent plates with wavelength-tuned phase-shifting interferometry

A wavelength-tuned Fizeau interferometer is applied to the problem of flatness testing of transparent plates. When the plate is positioned at a specific distance from the reference surface and an integer-math 13-frame phase-shifting algorithm is applied, the system directly filters out unwanted interference arising from backsurface reflections. The resulting front-surface profile exhibits less than 2 nm of residual error attributable to spurious reflections from within the plate.

Effect of substrate thickness, back surface texture, reflectivity, and thin film interference on optical band-gap thermometry

A model for the temperature-dependent shape of the absorption edge in GaAs is used to determine the effect of substrate thickness, back surface texture, and deposited dielectric films, on the temperature measurement accuracy, for both transmission and diffuse reflectance geometries of optical band-gap thermometry. A 5% change in substrate thickness, a 20% change in light scattering at the back of the substrate, or a 40% change in the residual absorption below the band edge all cause temperature errors of about 1°C. Band-gap thermometry is also sensitive to thin film interference during the growth of wide band-gap layers, which can cause the apparent temperature to oscillate. An algorithm is presented that uses the width of the knee in the spectrum to correct temperature errors caused by interference oscillations. This algorithm is suitable for real-time applications as the information needed to correct the knee position is obtained from the spectrum itself. The correction procedure is tested on data taken during growth of AlGaAs on GaAs. The interference oscillations in the deposited thin film reduced apparent temperature oscillations from about 1.5°C at 570°C to about 0.7°C (RMS).

Piezoreflectance study of InP near the absorption edge

Piezoreflectance measurements were carried out on InP samples with different doping concentrations near the absorption edge in the temperature range between 20 and 300 K. We show that the large features often found below the bandgap in piezomodulated spectra of InP samples are related not only to the impurity states but also to the back-surface reflection effects. Their proximity to the band edge depends on temperature, the impurity species and the geometry of the species in which the multiple reflections occur. At the band edge, the Burstein - Moss shift for the heavily S-doped InP and a pronounced excitonic effect for the Fe-doped semi-insulating InP substrates are observed. In addition, the temperature dependence of the energy position and broadening parameter of the excitonic and direct-bandgap transition features are also evaluated. The expression eV is proposed for the temperature dependence of the semi-insulating InP bandgap over the temperature range 20 - 300 K.

Fabrication of Polycrystalline Si Thin Film for Solar Cells

AbstractThe a-Si/poly-Si thin film tandem solar cell is a promising candidate for low-cost solar cells. We have conducted R&amp;D on poly-Si thin film using the Solid Phase Crystallization (SPC) method from amorphous silicon (a-Si). To improve the film quality of SPC poly-Si, we have developed a new SPC method called the partial doping method. This method features two stacked starting a-Si layers, a P-doped layer and a non-doped layer. Nucleation occurs in the P-doped layer, and the non-doped layer is the crystal growth layer. For the nucleation layer, we developed a Si film with a unique structure, which features relatively large crystallites (-1000A) embedded in a matrix of amorphous tissue. By combining these technologies, a conversion efficiency of 9.2% was obtained for poly-Si thin-film solar cells. For further improvement in the conversion efficiency, based on the concept of “independent control of nucleation and crystal growth”, it is necessary to combine the best fabrication methods for each layer. A high conversion efficiency of more than 12% was found possible by using the CVD method and a new back surface reflection structure.

Spectroscopic ellipsometry of thin films on transparent substrates: A formalism for data interpretation

By reviewing the operational principles of modern rotating analyzer ellipsometry, we propose general analytical formalism which permits the analysis of spectroscopic ellipsometry data for thin films on transparent substrates. The inclusion of incoherent reflection from the backsurface of an optically thick substrate changes the usual definitions of Δ and ψ. Experimental data on both glass and ZnO-coated glass are modeled to confirm the validity of this new formalism. We also simulated the ellipsometry spectra of amorphous silicon on glass to show the potential for in situ studies. The results show that the accuracy of ellipsometry data can be improved by including this incoherent backsurface reflection in the measurement, as opposed to roughening the back of the substrate.

Below gap photoreflectance of semi-insulating GaAs

Below-gap structures in photoreflectance spectra of semi-insulating GaAs were observed originating from electroabsorption of light reflected from the back surface. The peak amplitude of these structures was strongly temperature dependent within the range (78–290) K. The different lineshape at low temperature in comparison to that at room temperature suggested different physical mechanisms. This conclusion was proven experimentally using a two phase lock-in technique by which a fast (X) and slow (Y) component of the spectra could be separated. A fit procedure provides the pure back surface reflection effect for both components. The X mode dominating at low temperatures is attributed to electromodulation of the field-broadened excitonic absorption tail whereas the Y mode originates from the Franz–Keldysh electroabsorption.

LBIC and diffusion length mapping applied to the investigation of gettering by aluminium diffusion in multicrystalline silicon

Abstract In this work the effect of aluminium treatment on the device properties of highly disordered multicrystalline silicon wafersis analysed using the light beam induced current method at different wavelengths. The results show an improvement of the diffusion length L measured on samples submitted to aluminium diffusion at 900 °C to 4 h. Owing to the low initial values of diffusion lengths measured on wafers containing a high density of crystallographic defects (grain boundaies, dislocations, etc.), we assume that the main cause of enhancement is the gettering effect. However, back surface field and back surface reflection can occur if the initial value of L approaches the thickness of the wafer.

Dependence of laser characteristics on distributed Bragg reflector pairs in vertical-to-surface transmission electrophotonic devices

The threshold current dispersion, slope efficiency and lasing wavelength of vertical-to-surface transmission electrophotonic devices with vertical cavities are investigated. The critical number of output distributed Bragg reflectors (DBRs) required to overcome the disturbance due to back-surface reflections for surface-emitting-laser (SEL) structure devices is clarified. >

Modulation effects near the GaAs absorption edge

We show that the large features often found below the band gap in modulated spectra of GaAs samples are linked not only to impurities but also to back-surface reflection effects. Their proximity to the band edge depends on the geometry of the cavity in which the multiple reflections occur, the impurity species, and the temperature. Impurity contributions to the energy-dependent absorption coefficient become manifest in a multiple-reflection-enhanced reflectance edge. Externally applied perturbations modulate this reflectance edge, thereby producing first-derivative-like spectral features. At 5 K the extrema of the impurity-related features roughly correspond to the energies associated with donor and acceptor states. However, at room temperature they disagree significantly and therefore do not specify an impurity species with certainty.

Analysis of buried layers from high dose oxygen ion implantation

Silicon-on-insulator buried layers have been fabricated using implantation of 160 keV O + ions into n-type (100) silicon with a dose of 1.7 × 10 18 atoms/cm 2 at a wafer temperature of 550°C. The 5 mA oxygen beam was electrostatically scanned on two axes into a specially designed end station. A fixed average deflection of 5° was introduced to minimize impurity contamination and neutral beam. The wafers were not processed with oxide coatings. The as-implanted structures have been analyzed using SIMS and optical reflectometry. The optical data exhibit the expected maxima and minima due to interference plus incoherent back surface reflection. An optical model of the buried layer geometry has been developed using the technique of multiple discrete layers to simulate the observed physical regions. In order to investigate the presence and distribution of contaminating species, the SIMS technique has been used to provide depth profiles of oxygen, silicon, carbon, aluminum, iron, and copper. It is observed that, when present, iron is preferentially enhanced at the buried layer surfaces while aluminum is not. The iron concentration was calibrated using a 54Fe + implant in silicon rather than 56Fe + in order to avoid mass interference from 28Si 2+. The data show that despite the lack of a protective capping layer, very low impurity concentrations can be achieved.

Computer-Controlled System for Nondestructive Thickness Measurement of Corroded Steel Structures

Over a period of time, corrosion attacks the surface of offshore and waterfront steel structures and can cause catastrophic failures. To date, thickness measurement to determine loss from corrosion has been performed visually; however, this method provides only limited assessment [1], A computer-controlled ultrasonic system to more accurately measure corroded steel structure thickness is described. The system is based on an innovative concept using an ultrasonic focused transducer aligned directly over the center of corrosion pits. At each of these positions, a corrosion pit acts as an acoustic lens, collimating the beam inside the steel specimen. Using this approach, the reflection from the back surface of the steel specimen is strong. The time difference between the back surface and front-surface reflections is used to calculate the thickness of the plate. The prototype system developed for implementing this concept is capable of automatically locating the pit centers, acquiring waveforms, and determining the thickness. During this process, the system rejects waveforms with spurious signals using simple pattern recognition approaches. Results have shown that the thickness of corroded plates can be measured with an average error of 4 percent. The system not only is accurate for thickness measurement, but also is easy to operate. Commercially available equipment used for the system makes it simple to build and less expensive than a fully integrated custom system.

Variable birefringent beam splitter

A sapphire disk is used as a variable birefringent beam splitter for the linearly polarized pump beam in an oscillator-amplifier dye laser system. Rotation of the disk varies the extent of retardation. The induced elliptical polarization results in a variable back surface reflectance between approximately 0.02-0.15 for the green spectral region. Reflectances are measured on two disks with different thicknesses and compared with an approximate model describing the reflectance as a function of angle of rotation of the disk.

Submillimeter spectroscopy of TO-phonon mode softening in PbTe

Reflectivity spectra of lead telluride epitaxial layers on BaF2 and NaCl substrates have been investigated in the wave-number range of 5–100 cm−1. Due to multiple reflections and interference effects in the sandwich samples, weak structures in the reflectivity spectra are accentuated. This leads to a more precise determination of the lattice parameters, especially the TO-phonon mode frequency as compared to measurements using bulk samples. By measuring the back-surface (NaCl or BaF2) reflectivity, an additional advantage caused by a kind of index matching is achieved. The temperature dependence (T = 4.2–300 K) and the carrier concentration dependence of the PbTe TO-phonon mode frequency was determined.

New dithiene complexes for Q-switching and mode-locking infrared lasers

Computer results show high efficiencies can be obtained for solar energy conversion with thin back-surface field (BSF) cells. A large effect of the back-surface reflectivity is also predicted for such thin cells. The quality of the back-surface diffused layer is found to be an important parameter.