Phosphorus Diffusion Profiles Research Articles

Phosphorus emitter profile control for silicon solar cell using the doss diffusion technique

The Doped Oxide Solid Source (DOSS) diffusion technique is well suited for fine-tuning of the surface concentration. The dopant surface concentration is important during phosphorus emitter diffusion due to the opposite requirements of a lowly doped emitter for good blue response and a sufficiently high surface concentration for a good ohmic contact. The sources are made in the laboratory using the standard POCl3 diffusion technique. DOSS Diffusions were carried out in the temperature range 850-1050°C using sources with different doping levels obtained by varying the POCl3 partial pressure from 0.004 % to 4.28 %. The electrical profiles were measured using the Stripping Hall profiling technique. Phosphorus diffusion profiles with the complete elimination of the dead layer have been obtained over a large range of source concentrations for all investigated diffusion temperatures. The residual diffusion oxide thickness increased with both temperature and source doping level within the range 7.5-30 nm. XPS profiling indicated that the composition of the residual glass was a mixture of P2O5 and SiO2.

Coordinate dependent diffusion analysis of phosphorus diffusion profiles in gallium doped germanium

We have analyzed phosphorus diffusion profiles in an In0.01Ga0.99As/In0.56Ga0.44P/Ge germanium structure during phosphorus co-diffusion with gallium for synthesis of the germanium subcell in multi-junction solar cells.. Phosphorus diffused from the In0.56Ga0.44P layer simultaneously with gallium diffusion into the heavily gallium doped germanium substrate thus determining the specific diffusion conditions. Most importantly, gallium and phosphorus co-diffusion produces two p–n junctions instead of one. The phosphorus diffusion profiles do not obey Fick’s laws. The phosphorus diffusion coefficient DP depth distribution in the specimen has been studied using two methods, i.e., the Sauer–Freise modification of the Boltzmann–Matano method and the coordinate dependent diffusion method. We show that allowance for the drift component in the coordinate dependent diffusion method provides a better DP agreement with literary data. Both methods suggest the DP tendency to grow at the heterostructure boundary and to decline closer to the main p–n junction. The DP growth near the surface p–n junction the field of which is directed toward the heterostructure boundary and its decline near the main p–n junction with an oppositely directed field, as well as the observed DP growth with the electron concentration, suggest that the negatively charged VGeP complexes diffuse in the heterostructure by analogy with one-component diffusion.

Diffusion of phosphorus in olivine and molten basalt

The diffusivity of phosphorus in San Carlos olivine (SCO) was measured at near-atmospheric pressure and 650–850 °C by in-diffusion of P from a surface powder source consisting of pre-reacted SCO and AlPO 4 . The experiments were conducted in evacuated silica-glass ampoules at oxygen fugacities fixed by solid-state buffers, generally Ni-NiO but also including two experiments buffered at wustite-magnetite. Phosphorus uptake profiles were characterized by Rutherford backscattering spectroscopy (RBS) and nuclear reaction analysis (NRA). The temperature dependence of P diffusion in SCO conforms to the expected Arrhenius relation D = D 0 exp(− E a /R T ), where the constants are as follows: log( D 0 , m 2 /s) = −10.06 ± 0.80 and E a = 229 ± 16 kJ/mol. These values characterize P as a relatively slow diffuser in olivine—slower by about an order of magnitude than Cr and Ca at basalt near-liquidus temperatures—but substantially faster than Si. With a view toward modeling P uptake during rapid growth of natural olivines, P diffusion was also characterized in dry MORB basalt melt over the temperature range 1250–1500 °C at 1 GPa, using traditional diffusion couples contained in graphite. Phosphorus diffusion profiles in the quenched and depressurized samples were quantified by laser-ablation ICP/MS. Phosphorus diffusion in basaltic melt is similar to that of Si, with log( D 0 , m 2 /s) = −6.30 ± 0.7 and E a = 147 ± 22 kJ/mol. The new data for P diffusion in olivine and basalt melt can be used to explore the acquisition of fine-scale zoning in natural olivine phenocrysts through kinetic models, as well as the survival of P zoning in olivine with time spent at elevated temperature. Models of growth entrapment of a P-enriched near-surface layer in the olivine lattice indicate that crystal growth at plausible sustained rates is indeed likely to result in regions of anomalously high P content in the resulting crystal. Phosphorus concentrations above the equilibrium partitioning value can also result from development of a diffusive boundary layer in the melt against a rapidly growing crystal, but this mechanism is ineffective at typical sustained olivine growth rates, requiring dendrite-forming growth speeds. Preservation of P zoning on the scale of a few micrometers apparently requires cooling within a few months of formation of the zoning.

Influence of Implantation Damages and Intrinsic Dislocations on Phosphorus Diffusion in Ge

Influence of the implant-induced damages and the threading dislocations during germanium (Ge) epitaxy was investigated on the phosphorus diffusion in Ge. An adequate n-type junction was formed by 650°C rapid thermal annealing of the implanted bulk Ge. It is observed that for epitaxial Ge on Si substrate, there was an enhanced local phosphorus diffusion approaching the Ge/Si interface. A diffusion model considering the enhancement by defects is proposed to sufficiently explain the phosphorus diffusion profiles in Ge. The intrinsic dislocation-enhanced phosphorus diffusivity found at least three orders of magnitude larger than that in the absence of dislocations.

Effect of implantation damage on transient loss of phosphorus in silicon

In this study, samples were cycled between silicon implantation and rapid thermal annealing to understand the impact of implantation damage on the phosphorus dose loss due to interface segregation. The total trap density for interface segregation was not affected by implantation damage because silicon implantation did not change the maximum dose loss in samples initially implanted by phosphorus at a high fluence of 2×1015cm−2. However, the shape of phosphorus diffusion profiles was changed with silicon implantation, implying that the implantation damage promoted the diffusion of phosphorus toward surface. The promotion of the diffusion toward surface increased the dose loss in samples with a low initial phosphorus fluence of 7×1013cm−2 because most of the segregation sites at the interface were empty.

Simplified Front Surface Field Formation for Back Contacted Silicon Solar Cells

We investigate the formation of deep Phosphorus diffusion profiles with low surface concentrations in one single diffusion step. Such processes are suited for front surface field formation for n-type back-contact back-junction (BC- BJ) silicon solar cells or p-type silicon solar cells with alternative metallization techniques. The deposition temperature allows accurate control of the sheet resistance whereas the temperature of the in-situ oxidation strongly influences the profile-depth and surface concentration. The newly developed process leads to low dark saturation current densities well below 30 fA/cm2 on alkaline-textured surfaces and low short circuit current loss at the front side. Due to their depth and adjustable surface concentration, the obtained profiles are promising for front-surface- fields (FSF) and novel n-type emitters. A first implementation in BC-BJ solar cells with aluminum-alloyed emitter shows a good blue response of the solar cells and an improved efficiency compared to the reference process by 0.6%abs.

Non-Gaussian Diffusion of Phosphorus and Arsenic in Silicon with Local Density Diffusivity Model

In the light of published phosphorus and arsenic diffusion profiles [1,2] a non-Gaussian mathematical diffusion model is developed in this work based on separate forward and reflected impurity diffusion flows and called local density diffusion (LDD) model. The LDD model includes the rational function diffusion (RFD) model published in [3] and represents an improvement for near surface and tail concentration profile slope approximation by introducing just one single empirical fit parameter “r”. This single fit parameter is related to the given combination of impurity species (phosphorus: r=0.18; arsenic: r=0.43) in the applied host lattice system (silicon), but does not vary while approximating different experiments with different impurity surface concentrations and penetration depths [1,2]. Based on the LDD approximation in this work a surface enhanced diffusivity for phosphorus and a tail decelerated diffusion for arsenic is suggested in comparison to RFD model approximation only. The local density diffusivity is found to be non-linear along the penetration path and reaches its maximum at a distance LLDD from the surface.

A Novel Approach Toward the Simultaneous Diffusion of Boron and Phosphorus in Silicon

A novel technique is described for simultaneously diffusing boron and phosphorus on opposite sides of a silicon wafer for the fabrication of simple, high‐efficiency solar cells. The solid doping sources used in this approach are fabricated using standard and liquid sources and can be tailored to produce a wide range of boron and phosphorus diffusion profiles for a fixed thermal diffusion cycle. Uniformity of sheet resistance across a 49 cm2 area is greater than 95% in many cases. Unique to this approach is that the resulting diffusion glass is extremely thin, which allows for the in situ growth of a thin thermal oxide for surface passivation, without appreciably increasing the reflectance of the solar cell after applying the antireflection coating. Measurements of the saturation current density (Jo) by the photoconductance decay technique gave a low Jo of 33.3 fA/cm2 for a phosphorus‐diffused sample (70 Ω/□) with in situ oxide surface passivation, and a Jo of 292 fA/cm2 for a boron‐diffused sample (18.9 Ω/□). Initial solar cell fabrication experiments with simple, untextured n+pp+ structures have produced efficiencies as high as 17%, displaying fill factors of 0.79 and shunt resistances greater than 65,000 Ω‐cm2, demonstrating that no detectable cross‐doping takes place during this new simultaneous boron and phosphorus diffusion technique.

A Quantitative Model for the Coupled Diffusion of Phosphorus and Point Defects in Silicon

In this paper, we develop and analyze models for the coupled diffusion of dopants and point defects, since such models have been observed to display the qualitative aspects of high concentration phosphorus diffusion profiles such as the characteristic “kink and tail.” We begin by describing a general model for phosphorus diffusion via dopant/defect pairs assuming local equilibrium for electronic processes, but not for chemical processes. Using this system, along with parameters based on experimental data previously reported in the literature, we test common model assumptions. Our analysis concludes that dopant/defect pairing reactions are near local equilibrium, but that defect recombination reactions are not. We then use a simplified model, based on the assumption that the pairing reactions are near equilibrium to simulate phosphorus profiles. By including diffusion of phosphorus via negatively‐charged phosphorus/vacancy pairs, in addition to diffusion via phosphorus/interstitial pairs which dominates in intrinsic materials, we are able to match experimental phosphorus diffusion profiles with surface concentrations ranging from intrinsic to solid‐solubility at 900 and 1000°C, while remaining consistent with the broad range of previously observed behavior of phosphorus and point defects in silicon.

Comparison of Measured and Simulated Two‐Dimensional Phosphorus Diffusion Profiles in Silicon

The two‐dimensional diffusion of phosphorus has been studied by comparing profiles measured experimentally after furnace and rapid thermal annealing with simulated profiles obtained with the two‐dimensional simulators SUPREM IV and PREDICT 2. It was first shown experimentally that the effect of stress due to the window edge on chemical staining used in the measurement process is not significant to within the experimental error. It was then observed that there is considerable disagreement between measured and simulated profiles in the lateral direction, especially in the region directly under the mask edge. Possible explanations for the differences are discussed in terms of the inaccuracy of the initial implant simulation, the absence of damage annealing modeling in the simulation, and the effect of stress in the window edge vicinity on diffusion. When vacancy generation resulting from the reaction of the oxide with the underlying silicon substrate was included in the simulations, better agreement was observed with the experimental profiles.

Analytical Model for Phosphorus Diffusion in Silicon

The problem of diffusion of phosphorus in silicon is investigated using an analytical approach based on similarity methods. The analysis results in simple explicit solutions, which accurately describe all relevant features of the phosphorus diffusion profiles including the properties of the surface region, the kink, and the tail region.

Glow discharge optical spectroscopy as an analytical depth profiling technique

Glow discharge optical spectroscopy has been demonstrated to be capable of obtaining chemical composition vs depth information in microvolume solid samples. Sample geometry and sputtering pressure were found to have a large effect on measured multilayer interface sharpness. The use of a sample backing plate which has a similar electrical conductivity and a larger cross−sectional area than the sample to be analyzed was found to greatly increase the interface depth resolution. Results are presented for GDOS−determined boron implantation and phosphorus diffusion profiles in Si. The measured profiles appear somewhat broader than expected from theory and possible explanations for this discrepancy are discussed.