Dendritic Web Silicon Research Articles

Resistivity dependence of minority carrier lifetime and cell performance in p-type dendritic web silicon ribbon

This study shows that the bulk lifetime in 95 μm thick p-type dendritic web silicon solar cells is a strong function of bulk resistivity. The higher the resistivity, the greater the bulk lifetime. This behavior is explained on the basis of dopant–defect interaction, which increases the lifetime limiting trap concentration with the addition of dopant atoms. Model calculations show that in the absence of doping dependence of bulk lifetime ( τ), ∼2 Ω cm web should give the best cell efficiency for bulk lifetimes below 30 μs. However, strong doping dependence of bulk lifetime in p-web cells shifts the optimum resistivity from 2 to 15 Ω cm. Bulk lifetime in the as-grown web material was found to be less than 1 μs for all the resistivities. After the cell processing which involves phosphorus gettering, aluminum gettering, and SiN induced hydrogen passivation of defects, the bulk lifetime increased to 6.68, 11, 31 and 68.9 μs in 0.62, 1.37, 6.45 and 15 Ω cm p-type web material, respectively. Therefore, cell process induced recovery of lifetime in web is doping dependent, which favors high resistivity. Solar cells fabricated on 95 μm thick web silicon by a manufacturable process involving screen-printing and belt-line processing gave 14.5% efficient 4 cm 2 cells on 15 Ω cm resistivity. This represents a record efficiency for such a thin manufacturable screen-printed cell on a low-cost PV grade Si ribbon that requires no wafering or etching.

Origins of defect structures in dendritic web silicon

Defects introduced into dendritic web silicon during growth have deleterious effects on electrical properties, and in extreme cases can lead to growth termination. This study used transmission X-ray topography and etch pitting techniques to elucidate the formation of several defect structures in dendritic web silicon. Results indicate that the formation of third dendrites is caused by liquid droplet entrapment within the web. Inverted U-shaped surface features have no correlation with internal dislocation structure, but are probably caused by a decrease in web thickness near the growth initiation. Stringers are formed by the reaction of two dislocations of different Burger's vectors, and the stringer density does not increase with ribbon length. Finally, web curvature is caused by plastic deformation due to a build up of glissile dislocations pinned by surface steps. An understanding of the causes of the formation of these defects can lead to improved processing for more efficient web growth.

Belt furnace gettering and passivation of n‐web silicon for high‐efficiency screen‐printed front‐surface‐field solar cells

AbstractSix different resistivities (0.32, 0.57, 1.2, 2.2, 9.1 and 20 Ω cm) were investigated to understand the dopant–defect interaction in n‐type, antimony‐doped, dendritic web silicon ribbon, and to study its response to gettering and passivation during belt furnace processing (BFP). The as‐grown lifetime was found to be a strong function of resistivity with higher resistivity displaying higher lifetime. Phosphorus gettering at 925° C/6 min raised the as‐grown lifetime of ∼1 μs in 20 Ω cm n‐web to 5.4 μs. A combination of phosphorus gettering followed by simultaneous Al gettering and SiN hydrogenation raised the 20 Ω cm n‐web lifetime to 78 μs. Unlike the as‐grown web, the processed lifetime was greater than 75 μs for all resistivities, with no clear doping dependence. This is attributed to the very effective gettering and passivation during the belt furnace processing. Front surface field (FSF) n+–n–p+ cells were fabricated by spin‐on phosphorus diffusion on the front and screen‐printed Al on the back. A lifetime value of over 100 μs was obtained in a 14.2% screen‐printed FSF n‐web solar cell fabricated on 100‐μm‐thick 20 Ω cm substrate. The screen‐printed FSF cell fabricated on (111) FZ gave an efficiency of 14.9% with a fill factor of 77.6%. These results are supported by model calculations, which revealed a maximum efficiency of ∼15% for 100‐μm‐thick planar screen‐printed FSF cells and their insensitivity to bulk lifetime above 60 μs. Copyright © 2001 John Wiley & Sons, Ltd.

Aluminum alloy back p–n junction dendritic web silicon solar cell

A new silicon solar cell structure is presented in which the p–n junction is formed by alloying aluminum with n-type silicon, and where this p–n junction is located at the back (unilluminated) side of the cell. With a phosphorus front diffusion, the resultant n +np + structure has been implemented using dendritic web silicon substrates which are 100 μm thick and doped with antimony to 20 Ω cm. Such a structure eliminates shunting of the p–n junction, provides an effective front surface field, enables a high minority carrier lifetime in the base, and is immune to light-induced degradation. Using only production-worthy, high-throughput processes, aluminum alloy back junction dendritic web cells have been fabricated with efficiencies up to 14.2% and with corresponding minority carrier (hole) lifetime in the base of 115 μs.

Dislocations in dendritic web silicon

Due to its peculiar laminated structure, dendritic web silicon presents one of the most interesting configurations of dislocations in Si crystals. On the way to growing dislocation-free Si webs, X-ray diffraction topography techniques were used to visualize the dislocations in the material. The results and the analysis of dislocation reactions reported here provide information on the mechanisms involved in the formation of dislocation networks in the material. Also, the applicability of synchrotron radiation white-beam topography for a rapid characterization of dislocations is investigated and discussed.

Back surface field effects in the 17.3% efficient n-type dendritic web silicon solar cells

A solar cell efficiency of 17.3% (4 cm 2 area) has been achieved on 11 Ω cm, n-type dendritic web silicon. This is the highest reported efficiency to date on any silicon ribbon material. The minority carrier lifetime in the bulk material was determined to be at least 150 μs in spite of the fact that dopant diffusions were done at high temperature in a rapid thermal processing unit (1000°C for 30 s with cooling to 825°C at 50°C/min). Detailed characterization and modeling show that due to the reduced substrate thickness (100 μm) and long minority carrier diffusion length (>400 μm), device performance is strongly dependent on the back surface recombination velocity ( S b). In this study, an n + phosphorus back surface field (BSF) was implemented to reduce the effective S b of a fully metallized rear surface to approximately 20 cm/s. This effective S b value was determined through measurement and numerical analysis of the BSF doping profile. Model calculations reveal that the n +–n BSF served to increase the device efficiency by nearly 4% (absolute) above the case of infinite S b. By extending these model calculations to “mirror” solar cells (analogously doped n +–p–p + and p +–n–n + devices with equivalent lifetimes), it is shown that substrate type plays only a minor role in determining the overall device efficiency.

Self-aligned locally diffused emitter (SALDE) silicon solar cell

This paper presents, for the first time, a low-cost, high-throughput manufacturing approach for fabricating n-base dendritic web silicon solar cells with selectively doped emitters and self-aligned aluminum contacts using rapid thermal processing (RTP) and screen printing. The self-aligned locally diffused emitter (SALDE) structure is p + nn ++ where aluminum is screen-printed on a boron-doped emitter and fired in a belt furnace to form a deep self-doped p +-layer and a self-aligned positive contact to the emitter according to the well-known aluminum-silicon (AlSi) alloying process. The SALDE structure preserves the shallow emitter (20.2 μm) everywhere except directly beneath the emitter contact. There the junction depth is greater than 5 μm, as desired, in order to shield carriers in the bulk silicon from that part of the silicon surface covered by metal where the recombination rate is high. This structure is realized by using n-base (rather than p-base) substrates and by utilizing screen-printed aluminum (rather than silver) emitter contacts. Prototype dendritic web silicon (web) cells (25 cm 2 area) with efficiencies up to 13.2% have been produced.

An X-ray topography study of the dendritic web silicon growth process

In this study, transmission X-ray topography was used for investigating the nature of dislocation networks in as-grown dendritic web silicon ribbons. We were able to correlate minority carrier diffusion length with the configuration of the networks that are strongly affected by twin planes lying midway across the web thickness. Analysis of the networks is also useful in providing information regarding regions of high stress levels associated with a given growth environment.

Rapid thermal processing of high-efficiency silicon solar cells with controlled in-situ annealing

Silicon solar cell efficiencies of 17.1%, 16.4%, 14.8%, and 14.9% have been achieved on FZ, Cz, multicrystalline (mc-Si), and dendritic web (DW) silicon, respectively, using simplified, cost-effective rapid thermal processing (RTP). These represent the highest reported efficiencies for solar cells processed with simultaneous front and back diffusion with no conventional high-temperature furnace steps. Appropriate diffusion temperature coupled with the added in-situ anneal resulted in suitable minority-carrier lifetime and diffusion profiles for high-efficiency cells. The cooling rate associated with the in-situ anneal can improve the lifetime and lower the reverse saturation current density ( J o), however, this effect is material and base resistivity specific. PECVD antireflection (AR) coatings provided low reflectance and efficient front surface and bulk defect passivation. Conventional cells fabricated on FZ silicon by furnace diffusions and oxidations gave an efficiency of 18.8% due to greater short wavelength response and lower J o.

Analysis of Dislocation Networks and Electronic Properties of Dendritic Web Silicon

AbstractDuring the growth of dendritic web silicon, an ideal material for fabrication of high efficiency solar cells, a thin ribbon of silicon single crystal is obtained. Due to thermal stresses characteristic in this growth process, dislocations and residual stresses are observed in most ribbons. In this study, transmission X-ray topography was used for analyzing dislocation networks in as-grown web silicon. We were able to correlate minority carrier diffusion length with the configuration of the networks that are strongly affected by twin planes lying midway across the web thickness. Analysis of the networks is also useful in providing information regarding regions of high stress levels associated with a given growth environment.

Dendritic web silicon solar cells

An overview of recent developments in the growth of web silicon and in the subsequent fabrication of solar cells is presented. Single-crystal ribbons having a maximum length of 17 m and a maximum width of 7 cm have been produced. In a prepilot facility, web cells with an area of 24.5 cm are made with an average efficiency of 14% (AM1.5,100 mW/cm, 25°C). Advanced processing techniques have been employed to produce cells (4 cm area) with efficiencies approaching 17%. Modules approximately 40 X 120 cm in size with efficiencies as high as 14% have been made using web cells. The primary defect responsible for limiting the efficiency in web cells has been identified as a dislocation decorated with SiOx precipitates. Atomic hydrogen, introduced by ion implantation, has been found to passivate this defect through the entire thickness of the cell. The response of dendritic web silicon cells to 1 MeV electron irradiation has been shown to be very similar to that of cells made from Czochralski silicon.

ChemInform Abstract: Fabrication and Analysis of Bifacial Solar Cells from Dendritic Web Silicon Substrates.

AbstractBifacial solar cells, i.e., both front and back surface have a current‐collecting grid, are fabricated using dendritic web silicon of 100‐160 μm thickness as substrate material.

The effect of doping density and injection level on minority-carrier lifetime as applied to bifacial dendritic web silicon solar cells

The measured short-circuit current density in bifacial dendritic web silicon solar cells has been found to decrease with decreasing base resistivity, particularly under back illumination. In addition, the ratio of short-circuit current under back illumination to short-circuit current under front illumination was observed to vary with light intensity. These observations reflect the fact that the minority-carrier lifetime in the base of these cells is a function of the base resistivity and the illumination level. The dopant was assumed to play only an indirect role in determining lifetime. This decrease in lifetime is shown to follow from a distribution of defect levels in the bandgap. These levels are a consequence of extended defects that have been observed in the web material, namely oxide precipitates and the dislocation cores that they decorate. The dopant, acts only in the indirect role of moving the Fermi level over an existing background distribution of defect levels that arise from the extended defects. Assuming a parabolic distribution of defect levels in the bandgap, the minority-carrier lifetime was calculated as a function of doping density and excess carrier concentration (illumination level) using the Shockley-Reed-Hall theory. The short-circuit current densities that were calculated using these lifetimes agreed reasonably well with measured values for bifacial dendritic web silicon solar cells. The measurements were made over a range of doping densities (6*10/sup 14/ to 3*10/sup 16/ cm/sup -3/) and illumination levels (0.001 to 1 sun) for both front and back illumination of the bifacial cells. >

The Fabrication and Analysis of Bifacial Solar Cells from Dendritic Web Silicon Substrates

Bifacial solar cells, which respond to light incident on either the front or the back surface, have been fabricated using dendritic web silicon as the substrate material. Both p‐type and n‐type substrates have been used, with resistivity ranging from 0.2 to 100 Ω‐cm for p‐type and from 0.2 to 10 Ω‐cm for n‐type. The thickness of the web substrates varied from 100 to 160 μm. The performance of the cells was found to depend strongly on substrate resistivity, with the highest resistivity giving the best results. Efficiencies in excess of 15% under front illumination, and approaching 15% under back illumination, have been measured. The efficiency under back illumination decreases dramatically with resistivity because of a decrease in minority carrier diffusion length, as deduced from measurements of internal quantum efficiency. The ratio of short‐circuit current under back illumination to short‐circuit current under front illumination also decreases with illumination level below 1 sun. The observed effects of resistivity and illumination level on cell performance are throught to be a consequence of the presence of extended defects (dislocations decorated with precipitates) in web material.

Effect of low-energy hydrogen ion implantation on dendritic web silicon solar cells

The effect of a low-energy (0.4 keV), short-time (2 min), heavy-dose (1018/cm2) hydrogen ion implant on dendritic web silicon solar cells and material was investigated. Such an implant was observed to improve the cell open-circuit voltage and short-circuit current appreciably for a number of cells. In spite of the low implant energy, measurements of internal quantum efficiency indicate that it is the base of the cell, rather than the emitter, which benefits from the hydrogen implant. This is supported by the observation that the measured minority-carrier diffusion length in the base did not change when the emitter was removed. In some cases, a threefold increase of the base diffusion length was observed after implantation. The effects of the hydrogen implantation were not changed by a thermal stress test at 250 °C for 111 h in nitrogen. It is speculated that hydrogen enters the bulk by traveling along dislocations, as proposed recently for edge-defined film-fed growth silicon ribbon.

Dislocation-Solute Interactions at Twin Boundaries in Dendritic Web Silicon

ABSTRACTDLTS, TEM and APFIM techniques have been used to study impurity-defect interactions responsible for reducing minority carrier diffusion lengths in solar cells of dendritic web silicon. Fine scale precipitates of SiOx, which decorate dislocations piled up at the twin planes in low eff clency cells, appear to be the dominant recombination centers.

Investigation of Liquid Dopants for the Production of High Efficiency Solar Cells from Dendritic Web Silicon

Conventional gaseous and CVD are presently used in the Westinghouse baseline process for the production of high efficiency (13%–15%) dendritic web solar cells. The objective of this study was to replace the CVD mask and the subsequent diffusion step with a liquid boron dopant diffusion process. The number of process steps are reduced significantly with the use of the liquid process. Data on time‐temperature relationship on sheet resistance, junction depth, etc., were obtained and are presented. Various application techniques for the liquid process were investigated. As a first step, a manually applied sponge‐squeegee applicator was used, which was later replaced with a meniscus coating machine. This equipment permitted the application of uniform layers of liquid dopants of controlled thickness on the dendritic web. Various runs were made to make direct comparison of the liquid and baseline processes. The results of cell efficiencies achieved using the liquid process indicate that high efficiencies (13%–15%) can be achieved with this process. Results of , , fill factor, etc., are presented in detail and compared with the baseline process. Dark IV measurements made on liquid‐processed cells indicate excellent quality junctions and high lifetime values. The series and shunt resistance measurements on typical cells show no difference as compared with the baseline process. Junction profiles are presented for both a gaseous diffused cell and a liquid processed cell. The profiles show no dissimilarities.