Light‐induced degradation in indium‐doped silicon

Light‐induced degradation of charge carrier lifetime was observed in indium‐doped silicon. After defect formation, an annealing step at 200 °C for 10 min deactivates the defect and the initial charge carrier lifetime is fully recovered. The observed time range of the defect kinetics is similar to the well known defect kinetics of the light‐induced degradation in boron‐doped samples. Differences between defect formation in boron‐ and indium‐doped silicon are detected and discussed. A new model based on an acceptor self‐interstitial A_Si–Si_i defect is proposed and established with experimental findings and existing ab‐initio simulations.

     

A 3‐state defect model for light‐induced degradation in boron‐doped float‐zone silicon

We report on a light‐induced bulk defect activation and subsequent deactivation in boron doped float‐zone silicon that can be described by a 3‐state model. During treatment at elevated temperature and illumination, a sample first converts from an initial high lifetime state into a degraded low lifetime state and then shows a recovery reaction leading to a third high lifetime state that is then stable under degradation conditions. Furthermore, it is shown that reverse reactions into the initial state appear to be possible both from the degraded as well as the regenerated state. An injection dependent analysis of lifetime data yields a defect capture cross section ratio of ～20 suggesting a positively charged defect. (© 2016 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Light‐induced degradation of PECVD aluminium oxide passivated silicon solar cells

Light‐induced degradation (LID) has been identified to be a critical issue for solar cells processed on boron‐doped silicon substrates. Typically, Czochralski‐grown silicon (Cz‐Si) has been reported to suffer from stronger LID than block‐cast multicrystalline silicon (mc‐Si) due to higher oxygen concentrations. This work investigates LID under conditions practically relevant under module operation on different cell types. It is shown that aluminium oxide (AlOx) passivated mc‐Si solar cells degrade more than a reference aluminium back surface field mc‐Si cell and, remarkably, an AlOx passivated Cz‐Si solar cell. The defect which is activated by illumination is shown to be doubtful a sole bulk effect while the AlOx passivation might play a certain role. This work may contribute to a re‐evaluation of the suitability of boron‐doped Cz‐ and mc‐Si for solar cells with very high efficiencies. (© 2015 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Light‐induced degradation and regeneration of multicrystalline silicon Al‐BSF and PERC solar cells

Light‐induced degradation (LID) is a well‐known problem faced by p‐type Czochralski (Cz) monocrystalline silicon (mono‐Si) wafer solar cells. In mono‐Si material, the physical mechanism has been traced to the formation of recombination active boron‐oxygen (B–O) complexes, which can be permanently deactivated through a regeneration process. In recent years, LID has also been identified to be a significant problem for multicrystalline silicon (multi‐Si) wafer solar cells, but the exact physical mechanism is still unknown. In this work, we study the effect of LID in two different solar cell structures, aluminium back‐surface‐field (Al‐BSF) and aluminium local back‐surface‐field (Al‐LBSF or PERC (passivated emitter and rear cell)) multi‐Si solar cells. The large‐area (156 mm × 156 mm) multi‐Si solar cells are light soaked under constant 1‐sun illumination at elevated temperatures of 90 °C. Our study shows that, in general, PERC multi‐Si solar cells degrade faster and to a greater extent than Al‐BSF multi‐Si solar cells. The total degradation and regeneration can occur within ～320 hours for PERC cells and within ～200 hours for Al‐BSF cells, which is much faster than the timescales previously reported for PERC cells. An important finding of this work is that Al‐BSF solar cells can also achieve almost complete regeneration, which has not been reported before. The maximum degradation in Al‐BSF cells is shown to reduce from 2% (relative) to an average of 1.5% (relative) with heavier phosphorus diffusion.     

On the mechanism of potential‐induced degradation in crystalline silicon solar cells

Multicrystalline standard p‐type silicon solar cells, which undergo a potential induced degradation, are investigated by different methods to reveal the cause of the degradation. Microscopic local ohmic shunts are detected by electron‐beam‐induced current measurements, which correlate with the sodium distribution in the nitride layer close to the Si surface imaged by time‐of‐flight secondary ion mass spectroscopy. The results are compatible with a model of the formation of a charge double layer on or in the nitride, which inverts the emitter. (© 2012 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Acceleration and mitigation of carrier‐induced degradation in p‐type multi‐crystalline silicon

Recently, a new carrier‐induced defect has been reported in multi‐crystalline silicon (mc‐Si), and has been shown to be particularly detrimental to the performance of passivated emitter and rear contact (PERC) cells. Under normal conditions, this defect can take years to fully form. This Letter reports on the accelerated formation and subsequent passivation of this carrier‐induced defect through the use of high illumination intensity and elevated temperatures resulting in passivation within minutes. The process was tested on industrial mc‐Si PERC solar cells, where degradation after a 100 hour stability test was suppressed to only 0.1% absolute compared to 2.1% for non‐treated cells. (© 2016 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Fast and slow lifetime degradation in boron‐doped Czochralski silicon described by a single defect

A demonstration that boron–oxygen related degradation in boron‐doped Czochralski silicon could be caused by a single defect with two trap energy levels is presented. In this work, the same two‐level defect can describe the fast and slow lifetime decay with a capture cross‐section ratio of electrons and holes for the donor level of σ_n/σ_p = 19 ± 4. A model is proposed for the multi‐stage degradation involving a single defect, in which the product of the slow reaction is a reactant in the fast reaction. After thermal processing, a population of interstitial oxygen (O_i) exists in a certain state (the precursor state) that can rapidly form defects (fast degradation) and another population of O_i exists in a state that is required to undergo a slow transformation into the precursor state before defect formation can proceed (slow degradation). Kinetic modelling is able to adequately reproduce the multi‐stage degradation for experimental data. Dark annealing is also shown to impact the extent of ‘fast’ degradation. By decreasing the dark annealing time on pre‐degraded wafers, a more severe ‘fast’ degradation of the samples can be enabled during subsequent illumination, consistent with this theory. The paper then discusses possible candidates for the chemical species involved. (© 2016 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

硅片及其太阳电池的光衰规律研究

采用氙灯模拟太阳光源,将光强调至1000 W/m2,研究常规太阳能级单晶硅片、多晶硅片和物理提纯硅片的原片、去损减薄片、热氧化钝化片、双面镀氮化硅(SiN x:H)膜钝化片、碘酒钝化片以及太阳电池的光衰规律.利用WT-2000少子寿命测试仪以及太阳电池I-V特性测试仪分别对硅片的少子寿命和太阳电池的I-V特性参数随光照时间的变化进行了测试.结果表明:所有硅片以及太阳电池在光照的最初60 min内衰减很快随后衰减变慢,180 min之后光衰速率变得很小,几乎趋于零.     

Raman spectroscopy of gallium‐ and zinc‐based hydrotalcites

Insight into the unique structure of hydrotalcites (HTs) has been obtained using Raman spectroscopy. Gallium‐containing HTs of formula Zn₄ Ga₂(CO₃)(OH)₁₂ · xH₂O (2:1 ZnGa‐HT), Zn₆ Ga₂(CO₃)(OH)₁₆ · xH₂O (3:1 ZnGa‐HT) and Zn₈ Ga₂(CO₃)(OH)₁₈ · xH₂O (4:1 ZnGa‐HT) have been successfully synthesised and characterised by X‐ray diffraction (XRD) and Raman spectroscopy. The d(003) spacing varies from 7.62 Å for the 2:1 ZnGa‐HT to 7.64 Å for the 3:1 ZnGa‐HT. The 4:1 ZnGa‐HT showed a decrease in the d(003) spacing, compared to the 2:1 and 3:1 compounds. Raman spectroscopy complemented with selected infrared data has been used to characterise the synthesised gallium‐containing HTs. Raman bands observed at around 1050, 1060 and 1067 cm⁻¹ are attributed to the symmetric stretching modes of the (CO₃²⁻) units. Multiple ν₃ (CO₃²⁻) antisymmetric stretching modes are found between 1350 and 1520 cm⁻¹, confirming multiple carbonate species in the HT structure. The splitting of this mode indicates that the carbonate anion is in a perturbed state. Raman bands observed at 710 and 717 cm⁻¹ and assigned to the ν₄ (CO₃²⁻) modes support the concept of multiple carbonate species in the interlayer. Copyright © 2010 John Wiley & Sons, Ltd.     

Thermal deactivation of lifetime‐ limiting grown‐in point defects in n‐type Czochralski silicon wafers

In this study, we uncover a recombination‐active grown‐in defect reducing the minority carrier lifetime of Czochralski grown n‐type silicon from 5 ms to below 2 ms. We also demonstrate that the defect can be de‐activated by annealing between 300 °C and 360 °C. Our experimental findings suggest that vacancy‐related pairs incorporated during ingot growth may be responsible for the decreased minority carrier lifetime. (© 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Microstructural identification of Cu in solar cells sensitive to light‐induced degradation

Light‐induced degradation (mc‐LID or LeTID) can lead to a severe efficiency loss in multi‐crystalline solar cells. The underlying mechanism clearly distinguishes from known mechanisms as B‐O‐LID and Fe‐B‐LID. Various defect models have been suggested for mc‐LID mainly based on metal impurities, including Cu which is known to cause light‐induced degradation. We investigate mc‐LID sensitive PERC cells that show an efficiency degradation of 15%_rel. The weaker degradation of the grain boundaries (GBs) typical for mc‐LID is identified and further investigated from front and rear side with respect to recombination activities. The combination of local electrical measurements (LBIC), target preparation (REM, FIB) and element analysis (EDX, TEM) unveil Cu‐containing precipitates at the rear side of the solar cells. They accumulate at grain boundaries and at the rear surface of the Si‐bulk material where the passivation stack is damaged. We conclude that Cu originates from the cell material and discuss its relation to mc‐LID.

LBIC mapping (EQE at fixed wavelength) of a degraded mc‐Si PERC cell from front and rear side results in qualitatively different appearance of GBs.     

Impact of the firing temperature profile on light induced degradation of multicrystalline silicon

Light‐ and elevated temperature‐induced degradation in multicrystalline silicon can reduce the efficiency of solar cells significantly. In this work, the influence of the firing process and its temperature profile on the degradation behaviour of neighbouring mc‐Si wafers is analysed. Five profiles with measured high peak temperatures ≥800 °C and varying heating and cooling ramps are examined. With spatially resolved and lifetime calibrated photoluminescence images, normalized defect concentrations N^*_t are calculated to determine the degradation intensity. Wafers that underwent a fast firing process typical for industrial solar cell production show a significantly stronger degradation than samples that were subjected to the same peak temperature but with slower heating and cooling rates. A spatially resolved analysis of the carrier lifetime in the whole wafer shows that the degradation begins in low lifetime areas around dislocation clusters, spreading into good grains after several hours. By the use of optimized ramp‐up and/or ramp‐down rates during the firing even at very high peak temperatures, light and elevated temperature induced degradation can be suppressed. (© 2016 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Reduced pore diameter fluctuations of macroporous silicon fabricated from neutron‐transmutation‐doped material

The precision of photo‐electrochemical etching of perfectly‐ordered macropores in single‐crystalline silicon is limited by pore diameter fluctuations due to doping variations of the starting wafer (striations). The doping variation originates from the rotation during crystal growth in the float‐zone or Czochralski process, respectively. Experimentally, variations of the pore diameter up to 7% can occur. These so‐called striations limit performance of possible applications of macroporous silicon. As doping inhomogeneities are the reason for the striations, uniformly doped silicon wafers by neutron transmutation doping were used for the first time. Photoelectrochemical etching of neutron transmutated silicon has been carried out and the pore diameter fluctuation has been reduced by about 40% compared to standard doped float‐zone wafers. (© 2010 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)     

Micro‐photoluminescence spectroscopy on heavily‐doped layers of silicon solar cells

We report and explain the photoluminescence spectra emitted from silicon solar cells with heavily‐doped layers at the surface. A micro‐photoluminescence spectroscopy system is employed to investigate the total spectrum emitted from both the heavily‐doped layer and the silicon substrate with micron‐scale spatial resolution. The two regions of the device give rise to separate photoluminescence peaks, due to band‐gap narrowing effects in the highly‐doped layer. Two key parameters, the absorption depth of the excitation wavelength, and the sample temperature, are shown to be critical to reveal the separate signatures from the two regions. Finally, this technique is applied to locally diffused and laser‐doped regions on silicon solar cell pre‐cursors, demonstrating the potential value of this micron‐scale technique in studying and optimizing locally doped regions. (© 2015 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Atom probe tomography of size‐controlled phosphorus doped silicon nanocrystals

Doping of silicon nanocrystals is essential to control their electronic and optical properties. The incorporation of an impurity into a silicon nanovolume is a nontrivial task due to the self‐purification effect. Here, a systematic atom probe tomography study of the phosphorus distribution and incorporation in size‐controlled silicon nanocrystals embedded in silicon dioxide is presented. Qualitatively, it turns out that the phosphorus distribution in the system follows a universal, nanocrystal‐size independent trend: phosphorus‐enrichment at the interface with a substantial phosphorus‐incorporation in the silicon nanocrystal as small as 2 nm in diameter. This clearly contradicts strict self‐purification. These observations are explained by the bulk‐solubility and ‐segregation behaviour, kinetic effects related to the diffusion lengths, and nanoscale interface strain. The quantitative determination of the amount of phosphorus atoms per quantum dot enables a systematic understanding of phosphorus‐induced effects on optical and electronic properties of silicon nanovolumes.     

Raman study of Fano interference in p‐type doped silicon

As the silicon industry continues to push the limits of device dimensions, tools such as Raman spectroscopy are ideal to analyze and characterize the doped silicon channels. The effect of inter‐valence band transitions on the zone center optical phonon in heavily p‐type doped silicon is studied by Raman spectroscopy for a wide range of excitation wavelengths extending from the red (632.8 nm) into the ultra‐violet (325 nm). The asymmetry in the one‐phonon Raman lineshape is attributed to a Fano interference involving the overlap of a continuum of electronic excitations with a discrete phonon state. We identify a transition above and below the one‐dimensional critical point (E = 3.4 eV) in the electronic excitation spectrum of silicon. The relationship between the anisotropic silicon band structure and the penetration depth is discussed in the context of possible device applications. Copyright © 2010 John Wiley & Sons, Ltd.     

Accelerated formation of the boron–oxygen complex in p‐type Czochralski silicon

This Letter reports on the acceleration of the rate of formation of the boron–oxygen defect in p‐type Czochralski silicon with illumination intensities in excess of 2.1 × 10¹⁷ photons/cm²/s. It is observed that increased light intensities greatly enhance the rate of defect formation, without increasing the saturation concentration of the defect. These results suggest a dependence of the defect formation rate upon the total majority carrier concentration. Finally, a method using temperatures up to 475 K and an illumination intensity of 1.68 × 10¹⁹ photons/cm²/s is shown to result in near‐complete defect formation within seconds. (© 2015 WILEY‐VCH Verlag GmbH &Co. KGaA, Weinheim)     

Nanoporous silicon doped by Cu for gas-sensing applications

The effect of hydrogen sulphide on the current–voltage characteristics of metal–insulator–semiconductor (MIS) structures based on nanoporous silicon (Si_nanopor) under copper doping has been investigated. Scanning electron (SEM), atomic force (AFM) and optic microscopes and/or secondary ion mass spectroscopy (SIMS) were used to obtain detailed characterisation of copper cluster distribution present at the surface and pores, respectively. SIMS spectra reveal that finite gradient in copper distribution along the pores and oxidation of nanoporous silicon simultaneously can be obtained successfully under electroless deposition process. It was also shown that the doping of nanoporous silicon by Cu leads to enhanced hydrogen sulphide sensitivity of MIS structures even without catalytic active top electrodes (for example, Pd) at room temperature. Furthermore, for different types of familiar MIS structures based on nanoporous silicon, e.g., MIS structures doped or undoped by copper and by using Pd metal electrodes, the hydrogen sulphide detection at room temperature mainly depends on the modification in the height of barrier of hetero- (Al–Cu–Si_nanopor–c-Si) or Schottky-like (Pd–Cu–Si_nanopor–c-Si) structures resulting the chemical interaction of molecular H₂S gas with copper clusters at the surface and in the pores. It is demonstrated that MIS structures based on the nanoporous silicon with copper doping are more sensitive to H₂S action at room temperature. In addition, the physical mechanism explaining the observed phenomena is also discussed.     

Fano resonance in heavily doped porous silicon

Unexpectedly, the Fano resonance caused by the interference of continuum electron excitations with the longitudinal optical (LO) phonons was observed in random porous Si by Raman scattering. The analysis of the experimental data shows that the electron states trapped at the Si SiO₂ interface dominate in the observed Raman scattering. The gap energy associated with the interface states was determined. Copyright © 2011 John Wiley & Sons, Ltd.