Dark Annealing Research Articles

Investigation of light-induced degradation in N-Type silicon heterojunction solar cells during illuminated annealing at elevated temperatures

In this paper, we study a light-induced degradation (LID) mechanism observed in commercial n-type silicon heterojunction (SHJ) solar cells at elevated temperatures using dark- and illuminated annealing for a broad range of illumination intensities (1–40 kWm−2) at temperatures from 25 to 180 °C. Three key results are identified. Firstly, an increase in solar conversion efficiency (η) of up to 0.3% absolute is observed after 3 min of dark annealing at 160 °C, attributed to improved surface passivation and a reduction in series resistance. Secondly, a temperature-dependent light-induced degradation behaviour is observed at temperatures as low as 85 °C under 1-sun equivalent illumination, with increasing degradation extent and rate for increasing temperatures. At 160 °C, an average η loss of 0.5 ± 0.3% absolute is observed after only 5 min and exceeding 0.8% in some cells. Thirdly, a subsequent light intensity-dependent recovery occurs with continued illumination exposure. Under 1-sun illumination at 160 °C, a reduction in net η loss up to 0.05 ± 0.1% absolute is observed after 2 h. Increasing the illumination intensity to 40 kWm−2 accelerates the recovery and can result in a net η improvement of 0.2% absolute at 150 °C within 100 s. The results suggest that attempts to improve the efficiency of SHJ solar cells using illuminated annealing could be detrimental to cell performance if not carefully optimised. Further investigation is required to identify the exact nature of the underlying defect mechanism(s) and develop appropriate mitigation strategies on commercially suitable timescales.

Long-term stability study of the passivation quality of polysilicon-based passivation layers for silicon solar cells

We investigate the stability of the effective lifetime τeff and the recombination current density parameter J0 in n-type silicon samples with symmetric phosphorus doped poly-Si/SiOx structures, and identify factors that contribute to the passivation degradation behavior. It is found that the surface passivation quality of phosphorus doped polysilicon passivating contacts degrades upon dark annealing and light soaking at temperatures between 75 °C and 200 °C, which can lead to a pronounced increase of the recombination current density parameter J0 (one-side) from below 10 fA/cm2 to 50 fA/cm2 or above. The degradation is only detected on fired wafers, whereas the surface passivation quality is found to be stable in the non-fired sister samples. Surprisingly, a recovery of τeff and J0 is observed after the degradation. The degradation and regeneration behaviors depend strongly on temperature and light intensity, and the presence of silicon nitride (SiNx) capping layers during the light soaking. Increasing the annealing temperature dramatically increases the rate of the degradation and the regeneration process, and at the same time reduces the magnitude of the degradation. The regeneration process appears to be affected by the presence of SiNx films during the light soaking treatment. Samples with SiNx films removed after firing suffer a significantly larger degradation upon light soaking without any lifetime regeneration. Grazing incidence X-ray diffraction measurements reveal negligible change in the structural property and crystalline quality of the polysilicon layer during the degradation.

The Role of Dark Annealing in Light and Elevated Temperature Induced Degradation in p-Type Mono-Like Silicon

We have studied lifetime instabilities in p-type boron-doped mono-like silicon during light soaking (LS) and dark annealing (DA) at different temperatures, and their behavior upon LS/DA cycling at various degradation and regeneration stages. Despite having similar capture cross section ratios, it is found that the defects responsible for the degradation under illumination and in the dark could stem from two separate reactions, with hydrogen being the common precursor. A model for light and elevated temperature induced degradation (LeTID) is presented based on our experimental findings. It is proposed that hydrogen atoms originally bound in the silicon nitride layer are released into the silicon bulk above a certain firing temperature, which then interact with some other species in the silicon bulk under illumination, causing the LeTID degradation. During the cooling ramp of the firing process or extended DA, hydrogen in the silicon bulk starts to effuse into the ambient, reducing the amount of hydrogen remaining in the silicon bulk, and correspondingly affecting their LeTID behavior. The proposed model provides new insights to help understand complex LeTID behaviors reported in the literature, including its dependence on the firing profile, sample thickness, dopant type, and DA pretreatment.

Fast and Slow Stages of Lifetime Degradation by Boron–Oxygen Centers in Crystalline Silicon

A conflict between previous and recently published data on the two‐stage light‐induced degradation (LID) of carrier lifetime in boron‐doped oxygen‐containing crystalline silicon is addressed. The previous experiments showed the activation of two boron–oxygen recombination centers with strongly differing recombination properties for the fast and slow stages of LID, whereas more recent studies found only a single center for both stages. To resolve this controversy, the historic silicon samples of these previous examinations are re‐examined in this study after more than one decade. It is found that, in the historic samples, the fast stage can be either described by two different centers or a mixture of the two, depending on the duration of previous dark annealing. A possible solution is suggested based on the involvement of different activating impurities in the boron–oxygen defect. In dark‐annealed samples, the defect consisting of boron, oxygen, and the activation impurity is present in two latent configurations, which reconfigure during LID at a fast and a slow stage. In the examined historic silicon samples, which did not undergo a gettering pretreatment, a significant concentration of an additional boron–oxygen defect with a different kind of activating impurity attached exists. The historic and modern results are thus reconciled.

Excessive light-induced degradation in boron-doped Cz silicon PERC triggered by dark annealing

This work investigates the impact of annealing at elevated temperatures on the light-induced degradation (LID) of passivated emitter and rear cells (PERC) processed on boron-doped Czochralski-grown silicon substrates. The boron-oxygen (BO) defect has been stabilised prior to annealing and subsequent LID treatment. Excessive LID of up to 19.1 %rel. is observed upon illumination after extended dark annealing at 150 °C for 552 h, which is well above the BO defect-related LID of 5.6 %rel. measured upon illumination after cell processing if BO is not stabilised. Light and elevated Temperature Induced Degradation (LeTID), iron-boron pairs and surface recombination are excluded as root causes for the observed increased LID, which shows a similar behaviour as the BO defect but which cannot be explained by the well-established three-state model of the BO defect with the assumption of an empty regenerated state prior to BO stabilisation. Two speculative hypothesis for an explanation are (i) that further BO defect precursors are formed, which could be described via a reservoir or (ii) that a high percentage of the in-principle available BO defects are already in the stabilised state even without dedicated BO defect stabilisation. This increased LID does not occur when at least a small level of excess carrier concentration is induced during extended annealing and, hence, is expected not to occur during field operation. However, the observed behaviour is highly relevant for accelerated aging testing such as, e.g., damp heat testing during IEC and UL certification.

Degradation and Recovery of n-Type Multi-Crystalline Silicon Under Illuminated and Dark Annealing Conditions at Moderate Temperatures

Recently, an n -type multi-crystalline silicon (mc-Si) was observed to be susceptible to degradation under illumination at elevated temperatures with similarities to carrier-induced degradation in p -type mc-Si. In this study, we demonstrate degradation and regeneration of the effective lifetime of non-diffused n -type mc-Si wafers using illuminated and dark annealing conditions at moderate temperatures. Under illuminated annealing conditions, the degradation and regeneration rates of the n -type mc-Si are observed to be slower than those of the p -type mc-Si; however, the opposite trend was observed under dark annealing conditions. The carrier-induced degradation kinetics of the n -type wafers can be described by degradation and regeneration that occur simultaneously, and the activation energies have been identified to be 1.23 ± 0.16 eV for the degradation process and 1.34 ± 0.08 eV for the regeneration. Surprisingly, no degradation was observed in n -type mc-Si under dark annealing above 160 °C. Rather, at these conditions, a two-stage improvement in the lifetime was observed. Although degradation occurs after a subsequent laser treatment, the stable lifetime at the end of the degradation is still slightly higher than its initial value.

Impact of Dark Annealing on the Kinetics of Light- and Elevated-Temperature-Induced Degradation

Light- and elevated-temperature-induced degradation (LeTID) has been shown to have a significant detrimental impact on p-type multicrystalline silicon solar cells and, in particular, on passivated emitter and rear cells. Previous studies have shown that defect kinetics can be modulated for samples that are dark annealed prior to light soaking at elevated temperature. In this work, we show that while short annealing durations help accelerate both degradation and recovery rates to different extents, extended annealing instead instigates a retarding effect. Our results confirm that thermally induced degradation and regeneration mechanisms can be observed during dark annealing. The results also suggest that the response to this yet undetermined defect mechanism not only depends on the initial dark annealing temperature, but it is also highly dependent on the stage of the dark annealing degradation and regeneration cycle reached before beginning light soaking. Finally, we propose a refined model of three generalized modes to describe the changes in LeTID kinetics after dark annealing.

Evidence for Molybdenum‐Hydrogen Bonding in p‐Type Silicon upon Annealing under Illumination

Transition metals (TMs) are common contaminants in solar silicon produced by block casting or using lower grade polysilicon feedstock. Gettering can effectively remove certain metals, but slow diffusing TMs can remain and act as highly efficient recombination centers. One method to reduce their recombination activity is hydrogen passivation, however, this is generally ineffective in p‐type silicon, in which hydrogen and TMs are typically both in positive charge states. Recent studies have suggested that under illumination, the charge state of hydrogen in silicon can be altered, and its bonding with certain defects can be enhanced. In this work, deep level transient spectroscopy (DLTS), Laplace DLTS, and capacitance‐voltage measurements are used to study TM‐H bonding in p‐type silicon intentionally contaminated with Mo and hydrogenated via remote H plasma. It is found that heat‐treatments at moderate temperatures (<250 °C) under illumination with an LED at 850 nm result in a reduction in the concentration of electrically active Mo, with no similar reduction seen in samples annealed in the dark. The passivated fraction of Mo can be recovered after dark annealing at temperatures >200 °C. The results obtained provide evidence that upon heat‐treatments with illumination Mo‐H bonding in p‐type Si occurs. Possible mechanisms of the observed effects are discussed.

On the impact of dark annealing and room temperature illumination on p-type multicrystalline silicon wafers

In the past few years, carrier-induced degradation (CID) in p-type multicrystalline silicon (mc-Si) has been receiving significant attention. Recently, it has been reported that this material is also susceptible to degradation under dark anneal at moderate temperatures. In the first part of this study, we investigate the impact of the dark anneal temperature on mc-Si wafers. We identify both degradation and regeneration of the effective lifetime, where higher temperatures lead to faster rates and lower degradation extent. A fitting model is developed to describe the kinetics of these processes, where the degradation and regeneration process are assumed to happen simultaneously. An Arrhenius analysis of the degradation and regeneration rates, extracted from the proposed model, determines activation energies of 1.08 ± 0.05 eV for the degradation process and 1.11 ± 0.04 eV for the regeneration one. An improvement of the minority carrier effective lifetime of up to 40% is observed after a long dark anneal process. This improvement is associated with enhancement of both the bulk and surface passivation. Temperature- and injection-dependent lifetime spectroscopy measurements indicate that the recombination parameters of the associated defect causing the degradation in the dark are similar to those determined for the CID-related defect; therefore, it seems both defects have a similar nature.In the second part of the study, the effect of the illumination intensity at room temperature on the degradation/regeneration is studied. Surprisingly, an improvement in the effective lifetime is found, followed by a very slow degradation. The proposed model is found to be suitable to fit these measurements. The extracted rates suggest that the observed behavior is due to a regeneration that is much faster than the degradation.The reported findings provide new insights into CID in p-type mc-Si that will help improve understanding and assist in developing mitigation solutions.

A four-state kinetic model for the carrier-induced degradation in multicrystalline silicon: Introducing the reservoir state

In this work, we present new insight into the multi-crystalline silicon carrier-induced defect (CID) by performing multiple degradation and regeneration cycles and further investigation on the partial recovery of mc-CID through extended dark annealing (DA). The maximum normalised defect density was found to decay exponentially with the number of cycles, suggesting that the defect precursors were slowly depleted by DA. A four-state kinetic model is proposed by introducing a reservoir state. Simulation results generated by mathematical modelling based on the proposed state diagrams exhibited good agreement with the experimental results. Extended DA on a partially recovered sample combined with simulation results suggests that the capability of defect formation through DA and the existence of the reservoir state proposed herein were the root causes for the partial recovery reported in the literature. Finally, the change in bound hydrogen state is speculated to cause the modulation of mc-CID formation. A qualitative reservoir model based on the interaction between hydrogen molecules (H2), boron-hydrogen pairs (B-H) and free hydrogen (H+, H°) is proposed and further discussed.

Impact of interstitial iron on the study of meta-stable B-O defects in Czochralski silicon: Further evidence of a single defect

We explore the influence of interstitial iron (Fei) on lifetime spectroscopy of boron-oxygen (B-O) related degradation in p-type Czochralski silicon. Theoretical and experimental evidence presented in this study indicate that iron-boron pair (Fe-B) related reactions could have influenced several key experimental results used to derive theories on the fundamental properties of the B-O defect. Firstly, the presence of Fei can account for higher apparent capture cross-section ratios (k) of approximately 100 observed in previous studies during early stages of B-O related degradation. Secondly, the association of Fe-B pairs can explain the initial stage of a two-stage recovery of carrier lifetime with dark annealing after partial degradation. Thirdly, Fei can result in high apparent k values after the permanent deactivation of B-O defects. Subsequently, we show that a single k value can describe the recombination properties associated with B-O defects throughout degradation, that the recovery during dark annealing occurs with a single-stage, and both the fast- and slow-stage B-O related degradation can be permanently deactivated during illuminated annealing. Accounting for the recombination activity of Fei provides further evidence that the B-O defect is a single defect, rather than two separate defects normally attributed to fast-forming recombination centers and slow-forming recombination centers. Implications of this finding for the nature of the B-O defect are also discussed.

Impact of annealing on the formation and mitigation of carrier-induced defects in multi-crystalline silicon

Carrier-induced degradation (CID) of multi-crystalline silicon (mc-Si) wafers is a major problem currently affecting the photovoltaic industry. A large number of studies investigating this phenomenon have provided important clues regarding the identification of the defect, however, as of yet, none have isolated a specific cause. In this work, we provide further insight into the kinetics of CID in mc-Si by presenting a detailed study of the impact of dark annealing on the formation and subsequent mitigation of the carrier-induced defect. Previous work has shown that such anneals can modulate the kinetics of the defect. Here, we extend that work and demonstrate that dark annealing can result in accelerated defect formation and extended degradation throughout a subsequent light soaking cycle, irrespective to when the dark annealing was applied. It is suggested that dark annealing could release extra defect precursors into mc-Si, which then become recombination active upon illumination. Therefore, the subsequent degradation after dark annealing might not necessary involve a reverse reaction. Through multiple dark anneal and light soak cycles, the extent of degradation in each cycle continues to reduce. A direct reverse reaction (de-stabilisation) alone does not explain this observation. We suggest that this effect could be explained by the presence of a reservoir of defect precursors, which are gradually depleted throughout the dark annealing processes. Finally, it is demonstrated that dark annealing alone could potentially cause a similar degradation to illumination at elevated temperature.

Effect of low-temperature annealing on defect causing copper-related light-induced degradation in p-type silicon

Copper is a common impurity in photovoltaic silicon. While reported to precipitate instantly in n-type Si, copper causes light-induced degradation (Cu-LID) in p-type Si. Recently, partial recovery of Cu-LID was observed after only few minutes of dark annealing at 200 °C. In this contribution, we investigate the effects of the dark anneal on Cu-LID-limited minority carrier lifetime both experimentally and by simulations. Surprisingly, after initial recovery, the dark anneal results in further degradation corresponding to a many-fold increase in recombination activity compared to the degraded state after illumination. This anneal-induced degradation can potentially cause additional losses in accidentally Cu-contaminated devices when exposed to elevated temperatures, for example during recovery and regeneration treatments of solar cells. Transient ion drift measurements confirmed that the anneal-induced degradation cannot be attributed to residual interstitial Cu after illumination. After hundreds of hours of annealing, the samples showed another recovery. To analyze these experimental results, a comparison to simulations is performed at the end of the paper.

Evidence of an identical firing-activated carrier-induced defect in monocrystalline and multicrystalline silicon

While progress has been made in understanding the behaviour of the recently identified carrier-induced degradation mechanism in p-type multicrystalline silicon solar cells, little is currently known about the root cause of the defect or its possible existence in other materials. In this work, we present evidence suggesting that the defect also exists in Czochralski grown monocrystalline silicon wafers. For both mono- and multicrystalline silicon we demonstrate: 1) the presence of a degradation and recovery of bulk minority carrier lifetime induced by either illuminated or dark annealing; 2) a modulation in the magnitude of degradation by varying the firing conditions; and 3) capture cross-section ratios of 39.4 ± 4.9 and 33.4 ± 1.5 in monocrystalline and multicrystalline silicon, respectively. The results indicate that the recently identified degradation mechanism does not only occur in multicrystalline silicon from illuminated annealing at elevated temperatures, but it is also induced by dark annealing at elevated temperatures, and that the degradation can occur in Czochralski grown silicon.

Modulating the extent of fast and slow boron-oxygen related degradation in Czochralski silicon by thermal annealing: Evidence of a single defect

The fast and slow boron-oxygen related degradation in p-type Czochralski silicon is often attributed to two separate defects due to the different time constants and the determination of different capture cross section ratios (k). However, a recent study using high lifetime samples demonstrated identical recombination properties for the fast and slow degradation and proposed an alternative hypothesis that these were in fact due to a single defect. The study presented in this article provides further experimental evidence to support the single defect hypothesis. Thermal annealing after light soaking is used to investigate the behaviour of subsequent boron-oxygen related degradation. Modifying the temperature and duration of dark annealing on pre-degraded samples is observed to alter the fraction of fast and slow degradation during subsequent illumination. Dark annealing at 173 °C for 60 s is shown to result in almost all degradation occurring during the fast time-scale, whereas annealing at 155 °C for 7 h causes all degradation to occur during the slow time-scale. This modulation occurs without changing the total extent of degradation or changing the capture cross-section ratio. The results are consistent with the fast decay being caused by defect formation from immediately available defect precursors after dark annealing, whereas the slow degradation is caused by the slow transformation of another species into the defect precursor species before the more rapid reaction of defect formation can proceed.

Modulation of Carrier‐Induced Defect Kinetics in Multi‐Crystalline Silicon PERC Cells Through Dark Annealing

In this letter, we report on significant changes caused after dark annealing to the kinetics of the carrier‐induced defect, present in p‐type multi‐crystalline silicon PERC cells. The characteristic shapes of the degradation and regeneration curves under light soaking at 75 °C are dramatically altered, depending on the temperature of an initial dark anneal on the non‐degraded cell. Dark annealing for a fixed time (2.5 h) at temperatures of 200 °C or below, is found to accelerate both the subsequent degradation and regeneration rate and the degradation severity, while at higher temperatures it appears that a possible second defect with a significantly longer degradation and regeneration rate is activated. Through a further increase of the dark annealing temperature, the magnitude of this slow degradation is suppressed. This data provides essential information into the role that thermal history plays in the behavior of the still unidentified defect or defects, which is crucial for future studies of the degradation and methods to mitigate it.

The physics of photon induced degradation of perovskite solar cells

Lead-trihalide perovskite solar cells are an important photovoltaic technology. We investigate the effect of light induced degradation on perovskite solar cells. During exposure, the open-circuit voltage (Voc) of the device increases, whereas the short-circuit current (Isc) shows a decrease. The degradation can be completely recovered using thermal annealing in dark. We develop a model based on light induced generation of ions and migration of these ions inside the material to explain the changes in Isc, Voc, capacitance and dark current upon light exposure and post-exposure recovery. There was no change in defect density in the material upon exposure.

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 σn/σp = 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 (Oi) exists in a certain state (the precursor state) that can rapidly form defects (fast degradation) and another population of Oi 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)

Reproducible growth of p-type ZnO:N using a modified atomic layer deposition process combined with dark annealing

Nitrogen doped ZnO (ZnO:N) films were deposited by atmospheric atomic layer deposition (ALD) between 100 and 300°C. Postannealing was required to remove compensating defects. After a low temperature dark annealing, originally n-type films became p-type. Films deposited at low temperatures (⩽150°C) have low hole mobilities (μ) of 0.2–0.4cm2V−1s−1 and moderate hole concentrations (np) of around 1×1015cm−3. Higher temperature deposited films (⩾200°C) have higher μ values (6cm2V−1s−1) but np values <1×1013cm−3. This crossover in transport properties can be explained by the opposing effects of deposition temperature on nitrogen doping level and distribution, and film crystallinity.

Two Intermediate States I and J Trapped at Low Temperature in the Photocycles of two BLUF Domain Proteins of Cyanobacteria Synechocystis sp. PCC6803 and Thermosynechococcus elongatus BP-1†

We identified the two intermediate states, I and J, that are common in the photocycles of the cyanobacterial BLUF (sensor of Blue Light Using Flavin) domain proteins of Slr1694 of Synechocystis sp. PCC6803 and Tll0078 of Thermosynechococcus elongatus BP-1 by analyzing the absorption spectra at 5 K. Illumination at 5 K accumulated intermediate forms (designated as I5 and I9), which showed 5 and 9 nm redshifts of the absorption bands of flavin in the Tll0078 and Slr1694 proteins, respectively. I5 (I9) was converted into the next intermediate, which have 11 nm (14 nm) red-shifted absorption bands J11 (J14) after dark annealing at 230 K (240 K). Further dark annealing at 280 K (270 K) of J11 (J14) produced the signal-transmitting final form F490 (F495), with a small increase in the absorption at around 490 nm (495 nm). The results indicate that the BLUF proteins of Tll0078 and Slr1694 exhibit the common photocycle of D471 (D467) --> I5 (I9) --> J11 (J14) --> F490 (F495) at low temperature. The transition temperatures for these intermediate forms differ for two proteins. The amount of I5 (I9) accumulated at 5 K was small and increased at a higher temperature, suggesting heterogeneity of the protein structure that determines the reaction pathway.