Recombination properties of a diffused pn junction determined by spectral response measurements

The theoretical spectral responsivity of a diffused pn junction is computed in the case of a silicon n⁺p junction which employes a rather deep (4, 7 μ) and lightly doped N⁺ front region.Comparing experimental results with theoretical predictions the diffusion length L and surface recombination velocity S₀ can be determined. Several cases are examined: the influence of an oxide layer on the front and of gettering processes on L and S₀ are presented and the overall sensitivity of the method is discussed.     

Enhanced lateral current transport via the front N+ diffused layer of n‐type high‐efficiency back‐junction back‐contact silicon solar cells

N‐type back‐contact back‐junction solar cells were processed with the use of industrially relevant structuring technologies such as screen‐printing and laser processing. Application of the low‐cost structuring technologies in the processing of the high‐efficiency back‐contact back‐junction silicon solar cells results in a drastic increase of the pitch on the rear cell side. The pitch in the range of millimetres leads to a significant increase of the lateral base resistance. The application of a phosphorus doped front surface field (FSF) significantly reduces the lateral base resistance losses. This additional function of the phosphorus doped FSF in reducing the lateral resistance losses was investigated experimentally and by two‐dimensional device simulations. Enhanced lateral majority carrier's current transport in the front n⁺ diffused layer is a function of the pitch and the base resistivity. Experimental data show that the application of a FSF reduces the total series resistance of the measured cells with 3.5 mm pitch by 0.1 Ω cm² for the 1 Ω cm base resistivity and 1.3 Ω cm² for the 8 Ω cm base resistivity. Two‐dimensional simulations of the electron current transport show that the electron current density in the front n⁺ diffused layer is around two orders of magnitude higher than in the base of the solar cell. The best efficiency of 21.3% was obtained for the solar cell with a 1 Ω cm specific base resistivity and a front surface field with sheet resistance of 148 Ω/sq. Copyright © 2008 John Wiley & Sons, Ltd.     

Silicon solar cells using natural inversion layers found in thermally-oxidized p-silicon

Inversion layer silicon solar cells are described which employ the natural inversion layer occurring at the surface of thermally-oxidized p-type silicon as one side of an induced n-p junction. Very shallow junctions are predicted theoretically with high electric fields in a direction to aid the collection of carriers generated by light of ultra-violet wavelengths. Collection efficiency calculations show the inversion layer cell to be less sensitive to lifetime and surface recombination velocity variations than diffused junction cells. Experimental 2 cm × 2 cm cells have been fabricated with the inversion layer contacted via a fine diffused n⁺ grid overlaid with a Ni-Cu-Au contact. The contact grid, specially designed to minimize the effect of the high inversion layer sheet resistance, produced a total shading of 16%. Illuminated I-V measurements confirm the induced junction to be near ideal, with an ideality factor $\text{A ? 1.05}$ and a reverse saturation current approaching that predicted theoretically. Conversion efficiencies of ? 8% have been obtained, with no special precautions being taken to reduce the series resistance of the back contact, or reflections at the front surface.     

A low-high-junction solar-cell model developed for use in tandem cell analysis

A comprehensive low-high (L-H) junction solar cell model has been developed. It accounts for actual solar spectrum related photogeneration of carriers in all regions of the n-p-p⁺ cell and allows for any value of rear surface-recombination-velocity (SRV). In typical GaAs L-H junction solar cells, photogeneration in the p⁺ region, but not the p region, is found to be negligible. The L-H junction's space-charge-layer recombination current density is also negligible. Assigning a non-infinite value of rear surface SRV makes this model applicable to tandem multi-junction structures made from materials with different band gaps.     

Effect of junction depth on the performance of a diffused np silicon solar cell

A detailed numerical analysis of the influence of the junction depth on the performance of a diffused n⁺p silicon solar cell is presented. The analysis includes the effects of Fermi-Dirac statistics, band gap narrowing, a finite surface recombination velocity and the built-in field due to the impurity profile. The recombination mechanism plays a dominant role in the performance of the solar cell. The ideality factor, “a”, varies from 1.006 for 0.1 μm junction depth, to 1.0135 for 2 μm junction depth. The saturation current density, J_o increases with the junction depth showing that the recombination increases in the heavily doped diffused layer of the device. The variation of the light generated current, J_L, the open-circuit voltage, V_oc, efficiency, η and the ideality factor, “a” are reported and analysed.     

Optimizing annealing steps for crystalline silicon solar cells with screen printed front side metallization and an oxide‐passivated rear surface with local contacts

Silicon solar cells that feature screen printed front contacts and a passivated rear surface with local contacts allow higher efficiencies compared to present industrial solar cells that exhibit a full area rear side metallization. If thermal oxidation is used for the rear surface passivation, the final annealing step in the processing sequence is crucial. On the one hand, this post‐metallization annealing (PMA) step is required for decreasing the surface recombination velocity (SRV) at the aluminum‐coated oxide‐passivated rear surface. On the other hand, PMA can negatively affect the screen printed front side metallization leading to a lower fill factor. This work separately analyzes the impact of PMA on both, the screen printed front metallization and the oxide‐passivated rear surface. Measuring dark and illuminated IV‐curves of standard industrial aluminum back surface field (Al‐BSF) silicon solar cells reveals the impact of PMA on the front metallization, while measuring the effective minority carrier lifetime of symmetric lifetime samples provides information about the rear side SRV. One‐dimensional simulations are used for predicting the cell performance according to the contributions from both, the front metallization and the rear oxide‐passivation for different PMA temperatures and durations. The simulation also includes recombination at the local rear contacts. An optimized PMA process is presented according to the simulations and is experimentally verified. The optimized process is applied to silicon solar cells with a screen printed front side metallization and an oxide‐passivated rear surface. Efficiencies up to 18.1% are achieved on 148.8 cm² Czochralski (Cz) silicon wafers. Copyright © 2009 John Wiley & Sons, Ltd.     

Analysis of silicon solar cells with stripe geometry junctions

Silicon solar cells with stripe geometry junctions are analyzed. The base region collection efficiency is found to be insensitive to the transmissivity of the electrode and/or the diffused layer but quite sensitive to the width and separation of the stripe junctions. The additional losses of carriers are mainly due to the increased bulk recombination rather than the surface recombination. In one case analyzed the collection efficiency decreases by 22% when the junctions are separated by about one sixth of the diffusion length with the surface recombination velocity at 300 cm/sec. Possible uses of the stripe-junction design in p-n junction cells and Schottky barrier cells are re-examined in the light of the new calculation and found to be less attractive than previously suggested.     

Two-dimensional analysis of the interdigited back-contact solar cell

Using two-dimensional computer analysis, the interdigited back contact silicon solar cell (IBC) was analyzed at high illumination levels and the results were compared with the conventional Front Junction cell. The change of effectiveness of Chockley-Read-Hall bulk and surface recombination centers at high currents as well as the induced internal electric field are argued to explain the improved efficiency predicted for IBC cells at high illumination levels. For a 100 μm cell thickness and lifetime τ_p0 = 10 μsec the efficiency is indicated to increase from 7.5% at 1 sun to 14.0% at 100 suns AMO, when a surface recombination velocity (s₀) equal to 1000 cm/sec is assumed. The substrate thickness to provide maximum efficiency was found to be approximately 50 μm. It is confirmed that the surface lifetime is a significant factor in determining the device conversion efficiency. Since surface recombination dominates the efficiency, a new IBC cell design with a front doping gradient has been introduced to suppress the surface recombination. The IBC cell with 10¹⁸/cm³ front surface n⁺-doping concentration is optimum for an impurity diffusion depth of 10 μm, s₀ = 1000 cm/sec, τ_p0 = 10 μsec, for which an efficiency of 12% is computed at 1 sum AMO. A useful efficiency of about 8% at 1 sun AMO, even with s₀ = 10⁵ cm/sec, is predicted with front doping.     

Progress on large area n‐type silicon solar cells with front laser doping and a rear emitter

We report on the progress of imec's n‐type passivated emitter, rear totally diffused rear junction silicon solar cells. Selective laser doping has been introduced in the flow, allowing the implementation of a shallow diffused front surface field and a reduction of the recombination current in the contact area. Simplifications have been implemented towards a more industrial annealing sequence, by replacing expensive forming gas annealing steps with a belt furnace annealing. By applying these improvements, together with an advanced texturing process and emitter passivation by atomic layer deposition of Al₂O₃, 22.5% efficient cells (three busbars) have been realized on commercial 156 · 156 mm² Czochralski‐Si. This result has been independently confirmed by ISE CalLab. Copyright © 2016 John Wiley & Sons, Ltd.     

Effects of grain boundaries on the current-voltage characteristics of polycrystalline silicon solar cells

The current components associated with the grain boundaries of diffused p/n junction polysilicon solar cells made on n- and p-type Wacker substrates are analyzed and experimentally identified. New electrical methods for determining the presence or absence of preferential diffusion along the grain boundaries and for determining the average doping density of preferentially diffused regions along the grain boundaries are described. For p-type substrates, these methods revealed preferential phosphorus diffusion along grain boundaries; no preferential boron diffusion along grain boundaries was observed. The recombination current components were analyzed for the cells in which preferential diffusion occurred. The analysis shows that the dominant current component at small bias levels (0-300 mV) is the recombination current at the grain boundaries within the p/n junction space-charge region. At higher bias levels (V simeq V_{OC} simeq 500-600mV), both this current component and the current component due to recombination at that part of the grain boundary below the preferentially diffused region are important. The grain-boundary shunt resistance does not contribute a significant current component. It is shown that the preferential diffusion makes negligible the recombination current injected into the sidewall of the preferentially diffused region. This is consistent with a model in which the phosphorus diffusion significantly lowers the surface recombination velocity at the grain boundaries and in which the retarding built-in electric field further decreases the recombination current.     

Modified drift field model for high-low transition in solar cells

A theoretical discussion is presented for the understanding of the back surface pp⁺ transition used in solar cells. It is shown that quasi-neutrality of space charge is a good approximation if the back surface diffusion is fairly deep. The error involved in the drift field model developed by assuming a quasi-neutrality of the space charge is compared with that inherent in the abrupt high-low junction model. The analysis shows that the back surface boundary, when measured from the heavily doped p⁺ side, effectively exists at a distance much larger than the impurity diffusion depth and the recombination current in the base is always less than its value estimated from the abrupt junction model. The voltage in the pp⁺ transition is due to the change in electric field by the excess carriers injected by light and drops across those regions of the cell where the injected carrier density is appreciable.     

All‐screen‐printed back‐contact back‐junction silicon solar cells with aluminum‐alloyed emitter and demonstration of interconnection of point‐shaped metalized contacts

In the following, high‐efficiency back‐contact back‐junction silicon solar cells with aluminum‐alloyed emitter are described. First, the theoretical background for the cell concept is explained. To that purpose, the bulk lifetime and the front surface field characteristics are considered. Three different process sequences for the phosphorus‐diffused profiles on the front and back surfaces are depicted: One exhibits a shallow field, and two sequences have deeper, driven‐in profiles. For realizing high efficiencies, such cell structures must meet several prerequisites, such as firing‐stable front and rear passivations, and functional small screen‐printed Al structures. Furthermore, it must be possible to create contacts on the Si surfaces using the driven‐in P‐profiles. With such a structure, cell efficiencies of 20.0% are reached. An analysis of the series resistance and area‐weighted recombination is performed. The results are compared with the measured cell parameters. Two‐dimensional simulations show the efficiency potential when decreasing the width of the backside field and when a cell structure, which would inhibit a passivated aluminum‐alloyed p⁺‐emitter, is created. Also, an advanced concept is demonstrated where a point array of both polarities on the cell backside is interconnected externally on module level. To that purpose, the cell is soldered to a printed wiring circuit board by using a reflow soldering process. Copyright © 2013 John Wiley & Sons, Ltd.     

Reverse bias light emission from GaAs1-xPx diodes

The reverse bias light emission originating at microplasmas was investigated in GaAs_1-xP_x diodes and compared with the forward bias emission. Both spectra were found to be almost identical and could be explained by the same radiative recombination processes. The presence of the strong electric field in the junction gave rise to the Franz-Keldysh effect manifested by the uniform shift (～3 meV) of the reverse bias emission towards longer wavelengths with the Stark effect broadening the free-exciton emission peak P₁. Measurement of the shift indicated that the electric field responsible for this, though high (～10³ V/cm), was considerably lower than that prevailing in the centre of the junction (～5×10⁵ V/cm). This pointed to recombination and the concomitant radiation occurring at the edge of the depletion region, the high field in the centre of the junction inhibiting the recombination of electron hole pairs.     

Influence of the collector resistance on the performance of accumulation channel driven bipolar transistor

In this paper, the physical mechanism limiting the maximum controllable current density (J_mcc) and safe operating area (SOA) of the accumulation channel driven bipolar transistor (ACBT) is identified and analyzed for the first time. According to our analysis, the hole current flowing into the P⁺ collector at the shallow trench creates a bias opposing the built-in potential of the P⁺ collector/N-drift junction due to a finite resistance associated with the contact to the diffused P⁺ collector region. This lowers the potential barrier established in the narrow mesa region (in between the self-aligned trenches) by the built-in potential of the P⁺ collector/N-drift junction and the control gate potential and promotes electron injection from the N⁺ emitter into the N-drift region over the potential barrier. At the onset of electron injection over the potential barrier, gate control over the base drive of the vertical wide base PNP transistor in the ACBT is lost. This hypothesis has been verified through numerical simulations and confirmed by experimental measurements which indicated an increase of over 300% in J_mcc due to a reduction in the contact resistance to the P⁺ collector region after a post metallization anneal of the fabricated ACBT designs. Based upon these observations, a new ACBT structure with a Schottky collector junction is proposed. It is demonstrated that the proposed ACBT structure has a higher J_mcc and wider SOA in comparison to the ACBT structure with P⁺ collector region.     

On the forward current-voltage characteristics of p+-n-n+ (n+-p-p+) epitaxial diodes

The forward-biased current-voltage characteristics of p⁺-n-n⁺ and n⁺-p-p⁺ epitaxial diodes are derived theoretically. Effects of the energy-gap shrinkage, the high-low junction built-in voltage, the high-level injection, and the minority-carrier life time on the forward-biased current-voltage characteristics are included. Good agreements between the theoretically derived results and the experimental data of Dutton et al. are obtained. The developed theory predicts that the leakage of the high-low junction is dominated by the recombination of minority carriers in the highly doped substrate, not by the recombination of minority carriers in the high-low space charge region, which is opposite to the previous prediction of Dutton et al.     

Computer analysis of induced-inversion layer MOS solar cells

A computer analysis of induced inversion layer MOS solar cells is described. The analysis simultaneously solves Poisson's equation and the continuity equation in one dimension and provides a very effective method for solar cell evaluation. Numerical solutions of the carrier continuity equation in the inversion layer illustrate how cell designs may be improved in order to obtain higher short wavelength spectral response. Very shallow junctions (on the order of 0.07-0.1 μm) are shown to be optimum with higher electric fields in a direction to aid the collection of carriers generated by very high energy photons. The results also indicate that induceed inversion layer cells are less sensitive to surface recombination velocity variations than diffused p-n junction cells and have higher minority carrier lifetime. Furthermore, the effect of a p-p⁺ low-high junction on the back surface is examined and the results indicate that it is insignificant when the substrate doping concentration is optimized. High inversion layer sheet resistance values are evaluated and minimized with the contact diffusion used in the analysis designed to reduce the high inversion layer sheet resistance. Design improvements in cell performance are evaluated and identified with further improvement possible here. Conversion efficiency for silicon of 17.3% at AMO in the inversion layer solar cell is predicted assuming 95% transmission through the transparent conductor.     

Opto‐electrical modelling and optimization study of a novel IBC c‐Si solar cell

Interdigitated back contact (IBC) crystalline silicon (c‐Si) solar cells are attracting a lot of attention because of their capability to reach world record conversion efficiency. Because of the relatively complex contact pattern, their design and optimization typically require advanced numerical simulation tools. In this work, a TCAD‐based simulation platform has been developed to account accurately and in detail the optical and passivation mechanisms of front texturization. Its validation has been carried out with respect to a novel homo‐junction IBC c‐Si solar cell based on ion implantation and epitaxial growth, comparing measured and simulated reflectance, transmittance, internal quantum efficiency, external quantum efficiency spectra, and current density–voltage characteristics. As a result of the calibration process, the opto‐electrical losses of the investigated device have been identified quantitatively and qualitatively. Then, an optimization study about the optimal front surface field (FSF) doping, front‐side texturing morphology, and rear side geometry has been performed. The proposed simulation platform can be potentially deployed to model other solar cell architectures than homo‐junction IBC devices (e.g., passivated emitter rear cell, passivated emitter rear locally diffused cell, hetero‐IBC cell). Simulation results show that a not‐smoothed pyramid‐textured front interface and an optimal FSF doping are mandatory to minimize both the optical and the recombination losses in the considered IBC cell and, consequently, to maximize the conversion efficiency. Similarly, it has been showed that recombination losses are affected more by the doping profile rather than the surface smoothing. Moreover, the performed investigation reveals that the optimal FSF doping is almost independent from the front texturing morphology and FSF passivation quality. According to this result, it has been demonstrated that an IBC cell featuring an optimal FSF doping does not exhibit a significant efficiency improvement when the FSF passivation quality strongly improves, proving that IBC cell designs based on low‐doped FSF require a very outstanding passivation quality to be competitive. Deploying an optimization algorithm, the adoption of an optimized rear side geometry can potentially lead to an efficiency improvement of about 1%_abs as compared with the reference IBC solar cell. Further, by improving both emitter and c‐Si bulk quality, a 22.84% efficient solar cell for 280‐μm thick c‐Si bulk was simulated. Copyright © 2017 John Wiley & Sons, Ltd.     

20.7% efficient ion‐implanted large area n‐type front junction silicon solar cells with rear point contacts formed by laser opening and physical vapor deposition

In this work, we report on ion‐implanted, high‐efficiency n‐type silicon solar cells fabricated on large area pseudosquare Czochralski wafers. The sputtering of aluminum (Al) via physical vapor deposition (PVD) in combination with a laser‐patterned dielectric stack was used on the rear side to produce front junction cells with an implanted boron emitter and a phosphorus back surface field. Front and back surface passivation was achieved by thin thermally grown oxide during the implant anneal. Both front and back oxides were capped with SiN_x, followed by screen‐printed metal grid formation on the front side. An ultraviolet laser was used to selectively ablate the SiO₂/SiN_x passivation stack on the back to form the pattern for metal–Si contact. The laser pulse energy had to be optimized to fully open the SiO₂/SiN_x passivation layers, without inducing appreciable damage or defects on the surface of the n⁺ back surface field layer. It was also found that a low temperature annealing for less than 3 min after PVD Al provided an excellent charge collecting contact on the back. In order to obtain high fill factor of ~80%, an in situ plasma etching in an inert ambient prior to PVD was found to be essential for etching the native oxide formed in the rear vias during the front contact firing. Finally, through optimization of the size and pitch of the rear point contacts, an efficiency of 20.7% was achieved for the large area n‐type passivated emitter, rear totally diffused cell. Copyright © 2014 John Wiley & Sons, Ltd.     

Measurement of space charge generation-recombination current in Hg1−xCdxTe photodiodes by deep level transient spectroscopy

Deep Level Transient Spectroscopy (DLTS) measurements have been used to characterize n⁺ ? pHg_1?xCd_xTe junction photodiode performance. Deep level results obtained on a x = 0.320 liquid phase epitaxial grown photodiode and a x = 0.219 bulk quench anneal-grown photodiode have identified deep Shockley-Read recombination centers. Detailed characterization of trap energy, trap density, and capture cross sections for these traps located within the diode depletion region have been used to predict a space charge generation-recombination current and dynamic resistance-area product at zero bias voltage. This paper presents for the first time a direct correlation of DLTS parameters with photodiode device performance.