Surface passivation schemes for high‐efficiency n‐type Si solar cells

An effective passivation on the front side boron emitter is essential to utilize the full potential of solar cells fabricated on n‐type silicon. However, recent investigations have shown that it is more difficult to achieve a low surface recombination velocity on highly doped p‐type silicon than on n‐type silicon. Thus, the approach presented in this paper is to overcompensate the surface of the deep boron emitter locally by a shallow phosphorus diffusion. This inversion from p‐type to n‐type surface allows the use of standard technologies which are used for passivation of highly doped n‐type surfaces. Emitter saturation current densities (J_0e) of 49 fA/cm² have been reached with this approach on SiO₂ passivated lifetime samples. On solar cells a certified conversion efficiency of 21.7% with an open‐circuit voltage (V_oc) of 676 mV was achieved. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

21.0%‐efficient co‐diffused screen printed n‐type silicon solar cell with rear‐side boron emitter

Plasma enhanced chemical vapor deposition (PECVD) is applied to deposit boron silicate glasses (BSG) acting as boron diffusion source during the fabrication of n‐type silicon solar cells. We characterize the resulting boron‐diffused emitter after boron drive‐in from PECVD BSG by measuring the sheet resistances R_sheet,B and saturation current densities J_0,B. For process optimization, we vary the PECVD deposition parameters such as the gas flows of the precursor gases silane and diborane and the PECVD BSG layer thickness. We find an optimum gas flow ratio of SiH₄/B₂H₆= 8% and layer thickness of 40 nm. After boron drive in from these PECVD BSG diffusion sources, a low J_0,B values of 21 fA/cm² is reached for R_sheet,B = 70 Ω/□. The optimized PECVD BSG layers together with a co‐diffusion process are implemented into the fabrication process of passivated emitter and rear totally diffused (PERT) back junction (BJ) cells on n‐type silicon. An independently confirmed energy conversion efficiency of 21.0% is achieved on 15.6 × 15.6 cm² cell area with a simplified process flow. This is the highest efficiency reported for a co‐diffused n‐type PERT BJ cell using PECVD BSG as diffusion source. A loss analysis shows a small contribution of 0.13 mW/cm² of the boron diffusion to the recombination loss proving the high quality of this diffusion source. (© 2016 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

High efficiency screen-printed n-type silicon solar cell using co-diffusion of APCVD boron emitter and POCl3 back surface field

We present the fabrication and analysis of Passivated Emitter and Rear Totally Diffused (PERT) solar cells on n-type silicon using a co-diffusion process. In a single high temperature step, a BSG/SiO_x stack deposited by APCVD and a POCl₃ back surface field diffuse into the wafer to form the boron doped emitter and phosphorus doped back surface field. The SiO_x layer on top of BSG acts as a masking layer to prevent cross-doping of phosphorus as well as a blocking layer for boron out-diffusion. This resulted in an initial sheet resistance of 76 Ω/□ with good uniformity and a final p⁺ emitter sheet resistance of 97 Ω/□ after boron rich layer removal. Additionally, bulk lifetime was investigated before and after the high temperature step that resulted in an increase from 1.2 ms to 1.5 ms due to a POCl₃ gettering effect. A peak cell efficiency of 20.3% was achieved and each recombination component in terms of saturation current density was calculated and analyzed to understand the cell for further efficiency enhancement.     

Electrical measurements of boron implanted silicon on sapphire and bulk silicon

Hall effect and sheet resistivity measurements have been performed on boron implantations in 1μm silicon layers on sapphire (SOS), and in bulk silicon. The doses used were 10¹⁴, 10¹⁵ and 10¹⁶ ions/cm², and implantation energies were 150 and 300 keV. The samples were annealed at temperatures between 300 and 800°C. As a rule the effective number of carriers in SOS was found to be about twice the number of carriers in bulk silicon. However, the mobility is lower in bulk silicon, resulting in a sheet resistivity almost the same in boron implanted SOS and bulk silicon.     

19.4%‐efficient large‐area fully screen‐printed silicon solar cells

We demonstrate industrially feasible large‐area solar cells with passivated homogeneous emitter and rear achieving energy conversion efficiencies of up to 19.4% on 125 × 125 mm² p‐type 2–3 Ω cm boron‐doped Czochralski silicon wafers. Front and rear metal contacts are fabricated by screen‐printing of silver and aluminum paste and firing in a conventional belt furnace. We implement two different dielectric rear surface passivation stacks: (i) a thermally grown silicon dioxide/silicon nitride stack and (ii) an atomic‐layer‐deposited aluminum oxide/silicon nitride stack. The dielectrics at the rear result in a decreased surface recombination velocity of S_rear = 70 cm/s and 80 cm/s, and an increased internal IR reflectance of up to 91% corresponding to an improved J_sc of up to 38.9 mA/cm² and V_oc of up to 664 mV. We observe an increase in cell efficiency of 0.8% absolute for the cells compared to 18.6% efficient reference solar cells featuring a full‐area aluminum back surface field. To our knowledge, the energy conversion efficiency of 19.4% is the best value reported so far for large area screen‐printed solar cells. (© 2011 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Passivation of black silicon boron emitters with atomic layer deposited aluminum oxide

The nanostructured surface – also called black silicon (b‐Si) – is a promising texture for solar cells because of its extremely low reflectance combined with low surface recombination obtained with atomic layer deposited (ALD) thin films. However, the challenges in keeping the excellent optical properties and passivation in further processing have not been addressed before. Here we study especially the applicability of the ALD passivation on highly boron doped emitters that is present in crystalline silicon solar cells. The results show that the nanostructured boron emitters can be passivated efficiently using ALD Al₂O₃ reaching emitter saturation current densities as low as 51 fA/cm². Furthermore, reflectance values less than 0.5% after processing show that the different process steps are not detrimental for the low reflectance of b‐Si. (© 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Influence of laser power on POCl3 diffused back surface field on n-type PERT silicon solar cells with local back contact

We report n-type passivated emitter rear totally diffused (PERT) silicon solar cells with local back contacts (LBC) formed by laser process. With passivated back surface field (BSF), the PERT cell design shows an improved open circuit voltage (V_oc) with reduced recombination at the rear due to improved optical confinement. The rear side was diffused by POCl₃ diffusion with low sheet resistance (R_s) BSF and passivated using SiN_x. Laser ablation was used to open the SiN_x on the rear for LBC. The Nd:YAG laser power (mW) parameters and POCl₃ doping temperature were varied to obtain the BSF with lower sheet resistance. Laser power of 44 mW with 10 kHz resulted in 30 Ω/sq BSF layer with effective lifetime (τ_eff) of 90 μs and a higher V_oc of 646 mV. With the optimized laser parameters the best electrical results yielded a short circuit current density (J_sc) of 36 mA/cm² and efficiency of 18.54%.     

Contact passivation in silicon solar cells using atomic‐layer‐deposited aluminum oxide layers

Atomic‐layer‐deposited aluminum oxide (AlO_x) layers are implemented between the phosphorous‐diffused n⁺‐emitter and the Al contact of passivated emitter and rear silicon solar cells. The increase in open‐circuit voltage V_oc of 12 mV for solar cells with the Al/AlO_x/n⁺‐Si tunnel contact compared to contacts without AlO_x layer indicates contact passivation by the implemented AlO_x. For the optimal AlO_x layer thickness of 0.24 nm we achieve an independently confirmed energy conversion efficiency of 21.7% and a V_oc of 673 mV. For AlO_x thicknesses larger than 0.24 nm the tunnel probability decreases, resulting in a larger series resistance. (© 2011 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Low energy boron implantation profiles in silicon from junction depth measurements

Some methods have been recently developed to investigate the distribution of implanted ions in semiconductors, especially into silicon. Generally, these techniques are not valid for boron due to the absence of convenient radioactive isotopes, or to a too small sensitivity when the lower part of the distribution is of interest. This corresponds to our problem, since boron implanted nuclear particle detectors prepared with high resistivity material (up to 50,000 ω.cm) are needed. The properties of these P-N junctions depend in a certain amount on the impurity distribution existing several orders of magnitude below the top of the distribution. Therefore, only the junction location method can be employed. In this method a series of N-type silicon samples, differing each from the other by an increase in resistivity are implanted with boron. The depth of the P-N junction corresponds to the point of the profile where the concentration N_A is equal to that of the substrate N_D (i.e. this latter being well known from the resistivity of the starting material). If the location of the junction can be measured, the profile can then be constructed point by point. The junction location is visualized generally by copper staining. Roosild,⁽¹⁾ Kleinfelder,⁽²⁾ Fairfield⁽³⁾ and D. E. Davies⁽⁴⁾ have used this procedure for boron implantations at energies higher than 50 keV. There is a problem due to the small penetration of the boron ion, and, for high resistivity materials, it is difficult to know the true limits of the zones stained with copper.

In our problem, when heavy particle detectors are desired, it is necessary to implant at lower energies than those indicated previously (< 20 keV). We have developed a new technique derived from the junction depth method, which is useful even at very low implant energy (≈ 10 keV). It consists in measuring the energy loss by 100 keV protons when crossing the entrance window of the P-N junctions used as detectors.

In the first part of this paper the method is described and the possible errors are analyzed and evaluated. In the second part, the distribution of 15 keV boron ions implanted under several experimental conditions is studied. Emphasis is given to the defects resulting from the silicon bombardment.     

Optical management in high‐efficiency thin‐film silicon micromorph solar cells with a silicon oxide based intermediate reflector

In the effort to increase the stable efficiency of thin film silicon micromorph solar cells, a silicon oxide based intermediate reflector (SOIR) layer is deposited in situ between the component cells of the tandem device. The effectiveness of the SOIR layer in increasing the photo‐carrier generation in the a‐Si:H top absorber is compared for p–i–n devices deposited on different rough, highly transparent, front ZnO layers. High haze and low doping level for the front ZnO strongly enhance the current density (J_sc) in the μc‐Si:H bottom cell whereas J_sc in the top cell is influenced by the angular distribution of the transmitted light and by the reflectivity of the SOIR related to different surface roughness. A total J_sc of 26.8 mA/cm² and an initial conversion efficiency of 12.6% are achieved for 1.2 cm² cells with top and bottom cell thicknesses of 300 nm and 3 μm, and without any anti‐reflective coating on the glass. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Characteristics of silicon solar cell emitter with a reduced diffused phosphorus inactive layer

The diffusion of phosphorus using a phosphorous oxychloride (POCl₃) source in silicon has been used widely in crystalline silicon solar cells. The thermal diffusion process in the furnace consists of two steps: pre-deposition and drive-in. The phosphorous doping profile via thermal diffusion often exhibits high concentrations in the surface-near emitter, which result in a recombination increase. This layer, called the dead layer, should be inhibited in order to fabricate high efficiency silicon solar cells. In this paper, the amount of the POCl₃ flow rate was varied during the pre-deposition process in order to minimize the dead layer, and the characteristics of the phosphosilicate glass (PSG) and emitter were analyzed. From the secondary ion mass spectroscopy (SIMS) and electrochemical capacitance–voltage profiler (ECV) measurements, the emitter formed using a POCl₃ flow rate of 1000 sccm contained the least amount of inactive dopant and resulted in reasonable performance in the silicon solar cell. As the POCl₃ flow rate increased, the doped silicon wafer included electrically inactive P near the surface, which functions as a defect degrading the electrical performance of the emitter. As a result of this, the removal of the dead layer containing the inactive P was attempted through dipping the doped wafer in a HF solution. After this process, the emitter saturation current density and implied V_oc were improved. The completed solar cells and their external quantum efficiencies at a short wavelength also demonstrated improved performance. A quantitative analysis of the emitter can provide a deeper understanding of methods to improve the electrical characteristics of the silicon solar cell.     

Plasma enhanced atomic layer deposition of gallium oxide on crystalline silicon: demonstration of surface passivation and negative interfacial charge

Herein we report on the passivation of crystalline silicon by gallium oxide (Ga₂O₃) using oxygen plasma as the oxidizing reactant in an atomic layer deposition (ALD) process. Excess carrier lifetimes of 2.1 ms have been measured on 1.75 Ω cm p‐type silicon, from which a surface recombination current density J₀ of 7 fA cm^–2 is extracted. From high frequency capacitance‐voltage (HF CV) measurements it is shown that, as in the case of Al₂O₃, the presence of a high negative charge density Q_tot/q of up to –6.2 × 10¹² cm^–2 is one factor contributing to the passivation of silicon by Ga₂O₃. Defect densities at midgap on the order of ～5 × 10¹¹ eV^–1 cm^–2 are extracted from the HF CV data on samples annealed at 300 °C for 30 minutes in a H₂/Ar ambient, representing an order of magnitude reduction in the defect density compared to pre‐anneal data. Passivation of a boron‐diffused p⁺ surface (96 Ω/□) is also demonstrated, resulting in a J₀ of 52 fA cm^–2. (© 2015 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Effective hydrogenation and surface damage induced by MW-ECR plasma of fine-grained polycrystalline silicon

This work reports the investigations on the effects of the hydrogenation process of thin film polycrystalline n⁺pp⁺ mesa silicon cells using MW-ECR plasma in a conventional PECVD system. Different operating parameters such as MW-ECR power, annealing temperature and the doping level of the emitter region were varied. The n⁺-type emitter regions were obtained by phosphorus diffusion in a conventional furnace using an oxide doping source containing phosphorus (P507 or P509 solutions, from Filmtronics Inc.). The MW hydrogenation was carried out at a sample temperature of 400°C for 60 min. Both types of emitters formed from P507 and P509 showed V _oc of 155 mV and 206 mV, which increased linearly to 305 mV and 331 mV, respectively, after hydrogenation when the MW power varied from 200 to 650 W. However, the sheet resistances of the n⁺ emitter region showed a slight increase depending upon hydrogenation power because of its etching. In a further study, hydrogenated samples were annealed in neutral or forming gas (FG) and we observed interesting results on V _oc in the presence of FG. The FG annealing temperature study revealed a strong dependence of V _oc on MW power, which affected the etching level of emitter and emitter dopant concentration, which controls the diffusion of hydrogen ions during post-hydrogenation step. The results were explained in detail by combining the effects of MW power and dopant level of the emitter.     

Chemical etching to dissolve dislocation cores in multicrystalline silicon

Multicrystalline silicon wafers are used for approximately half of all solar cells produced at present. These wafers typically have dislocation densities of up to ∼10⁶ cm⁻². Dislocations and associated impurities act as strong recombination centres for electron–hole pairs and are one of the major limiting factors in multicrystalline silicon substrate performance. In this work we have explored the possibility of using chemical methods to etch out the cores of dislocations from mc-Si wafers. We aim to maximise the aspect ratio of the depth of the etched structure to its diameter. We first investigate the Secco etch (1K₂Cr₂O₇ (0.15 M): 2HF (49%)) as a function of time and temperature. This etch removes material from dislocation cores much faster than grain boundaries or the bulk, and produces tubular holes at dislocations. Aspect ratios of up to ∼7:1 are achieved for ∼15 μm deep tubes. The aspect ratio decreases with tube depth and for ∼40 μm deep tubes is just ∼2:1, which is not suitable for use in bulk multicrystalline silicon photovoltaics. We have also investigated a range of etches based on weaker oxidising agents. An etch comprising 1I₂ (0.01 M): 2HF (49%) attacked dislocation cores, but its etching behaviour was extremely slow (<0.1 μm/h) and the pits produced had a low aspect ratio (<2:1).     

Extremely low surface recombination in 1 Ω cm n‐type monocrystalline silicon

A key requirement in the recent development of highly efficient silicon solar cells is the outstanding passivation of their surfaces. In this work, plasma enhanced chemical vapour deposition of a triple layer dielectric consisting of amorphous silicon, silicon oxide and silicon nitride, charged extrinsically using corona, has been used to demonstrate extremely low surface recombination. Assuming Richter's parametrisation for bulk lifetime, an effective surface recombination velocity S_eff = 0.1 cm/s at Δn = 10¹⁵ cm^–3 has been obtained for planar, float zone, n ‐type, 1 Ω cm silicon. This equates to a saturation current density J_0s = 0.3 fA/cm², and a 1‐sun implied open‐circuit voltage of 738 mV. These surface recombination parameters are among the lowest reported for 1 Ω cm c‐Si. A combination of impedance spectroscopy and corona‐lifetime measurements shows that the outstanding chemical passivation is due to the small hole capture cross section for states at the interface between the Si and a‐Si layer which are hydrogenated during nitride deposition. (© 2016 The Authors. Phys. Status Solidi RRL published by WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Theoretical studies on dielectronic recombination of neonlike gold and its effects on plasma ionization balance and radiation energy

The dielectronic recombination (DR) of neonlike gold ions is investigated employing the flexible atomic code based on the relativistic configuration interaction method, and its influence on the ionization balance and radiation energy in high-temperature plasma is also studied. The total resonance strength for LMM configuration complex is in a good agreement with the experimental measurement and other theoretical works. The DR rate coefficients are calculated and compared with the three-body recombination and radiative recombination rate coefficients. The DR process is the dominant recombination mechanism of Ne-like gold ions for plasma with temperature T_e≥6.5 keV and density n_e≤2×10²² cm^-3, which is close to the condition of X-ray conversion region in the inertial confinement fusion. Moreover, the DR satellite spectra of LMM, LMN and LMO resonances are simulated, and compared with the intensities of the corresponding resonance lines induced by the electron impact excitation. The intensity ratio of the satellite line 3D’ [(2p⁵_3/23d_3/23d_5/2)_J=5/2[(2p^{5}_{3/2}3d_{3/2}3d_{5/2})_{J=5/2}–(2p⁶3d_3/2) _J=3/2](2p^{6}3d_{3/2})_{ J=3/2}] and the resonance line 3D [(2p⁵_3/23d_5/2)_J=1[(2p^{5}_{3/2}3d_{5/2})_{J=1}–(2p⁶)_J=0](2p^{6})_{J=0}] is given, which can be applied for diagnostics of plasma temperature.     

Ion implantation into amorphous Si layers to form carrier‐selective contacts for Si solar cells

This paper reports our findings on the boron and phosphorus doping of very thin amorphous silicon layers by low energy ion implantation. These doped layers are implemented into a so‐called tunnel oxide passivated contact structure for Si solar cells. They act as carrier‐selective contacts and, thereby, lead to a significant reduction of the cell's recombination current. In this paper we address the influence of ion energy and ion dose in conjunction with the obligatory high‐temperature anneal needed for the realization of the passivation quality of the carrier‐selective contacts. The good results on the phosphorus‐doped (implied V_oc = 725 mV) and boron‐doped passivated contacts (iV_oc = 694 mV) open a promising route to a simplified interdigitated back contact (IBC) solar cell featuring passivated contacts. (© 2014 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

The p recombination layer in tunnel junctions for micromorph tandem solar cells

A new tunnel recombination junction is fabricated for n–i–p type micromorph tandem solar cells. We insert a thin heavily doped hydrogenated amorphous silicon (a-Si:H) p + recombination layer between the n a-Si:H and the p hydrogenated nanocrystalline silicon (nc-Si:H) layers to improve the performance of the n–i–p tandem solar cells. The effects of the boron doping gas ratio and the deposition time of the p-a-Si:H recombination layer on the tunnel recombination junctions have been investigated. The current-voltage characteristic of the tunnel recombination junction shows a nearly ohmic characteristic, and the resistance of the tunnel recombination junction can be as low as 1.5 ·cm 2 by using the optimized p-a-Si:H recombination layer. We obtain tandem solar cells with open circuit voltage V oc = 1.4 V, which is nearly the sum of the V oc s of the two corresponding single cells, indicating no V oc losses at the tunnel recombination junction.     

Heavily doped Si:P emitters of crystalline Si solar cells: recombination due to phosphorus precipitation

The measured saturation current density J_0e of heavily phosphorus‐doped emitters of crystalline Si solar cells is analysed by means of sophisticated numerical device modelling. It is concluded that Shockley–Read–Hall (SRH) recombination exceeds Auger recombination significantly; it is caused by inactive phosphorus. This explains the large discrepancies between measured and simulated J_0e values, observed persist‐ently over the last two decades in industrially fabricated Si solar cells. As a consequence, the heavily phosphorus‐diffused emitters still bear a significant potential to contribute to higher Si solar cell efficiency levels, if the amount of inactive phosphorus can be reduced. (© 2014 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Some studies of BF2 and B+F implanted Silicon

Ion-implanted shallow junctions have been investigated using BE₂ (molecular ions) by the anodic oxidation method coupled with a four-point probe technique. BF₂ ions were implanted through screen oxide at doses of 3–5 × 10¹⁵ ions/cm² and energies of 25 and 45 keV which is equivalent to 5.6 keV and 10 keV of boron ions. The effect of energy, dose and annealing temperature on shallow junctions is presented in this paper. The shallow junctions in the range of 0.19 μm to 0.47 μm were fabricated.

The effect of fluorine on sheet resistivity of boron implanted silicon at various doses, treated with two-step and three-step annealing, is also presented for comparison in the paper.