Cz Wafers Research Articles

Atomic-layer-deposited BOx/Al2O3 stack for crystalline silicon surface passivation

Surface passivation is a crucial factor in improving the efficiency of c-Si solar cells. In this work, we develop a boron oxide/aluminum oxide stack (BOx/Al2O3) using the atomic layer deposition technique, and investigate the passivation quality and mechanism on c-Si surfaces. The BOx/Al2O3 stacks display excellent surface passivation on c-Si surfaces after annealing, superior to that of Al2O3 single layers. With the optimal BOx/Al2O3 passivation stack, we obtained a very low dark current density (J0) of 7.8 fA/cm2 (iVoc 725 mV) and 4.7 fA/cm2 (iVoc 714.6 mV) on industrial n-type (5.3 Ω cm) and p-type (3.0 Ω cm) Cz wafers, respectively. The BOx/Al2O3 stack also shows excellent passivation performance on boron-doped p+ emitters (110 Ω/sq), achieving a low J0 of 15.1 fA/cm2. The enhanced passivation quality of the BOx/Al2O3 stack can be attributed to the increased fixed negative charge density after thermal annealing. BOx/Al2O3 passivation stacks also can potentially act as an effective p-type dopant source for laser processing, which would be very desirable for high-efficiency c-Si solar cells.

Application of n‐Polysilicon Rear Emitter for High‐Efficiency p‐TOPCon Solar Cells

AbstractThis study entails the examination of tunnel oxide passivated contact on p‐type silicon wafers (p‐TOPCon) passivated with n‐polysilicon for a solar cell by using Quokka‐3, a numerical simulation program. The effects of the thickness, bulk lifetime, resistivity, and selectivity of charge carriers due to the polysilicon passivated contact are investigated. Through such n‐polysilicon passivated contact, the back‐emitter solar cells engender higher internal power owing to enhanced surface passivation. This further reduces the shading loss due to front metallization; however, the reduced minority carrier lifetime of the p‐type Czochralski (Cz) wafer restricts the possibilities for high efficiency. Subsequently, the minority charge carrier lifetime of the p‐type wafer conceivably becomes an obstacle to realizing TOPCon solar cells with a high conversion efficiency. This study demonstrates that a configuration suitable for the industrial manufacturing of high‐efficiency solar cells is a crystalline silicon solar cell on a p‐type wafer through a rear‐emitter n‐polysilicon passivated contact. A roadmap toward 24.87% of the p‐TOPCon solar cells through the n‐type polysilicon passivated contact is also devised.

Two-dimensional BMD, BSF, ring-OISF and COP distributions in a dislocation-free ingot grown by the NOC method

The noncontact crucible (NOC) method has a large and deep low-temperature region in Si melt to grow a uniform large Si single ingot. The theoretical simulation pointed out that the lowest vacancy or interstitial Si atom area appears in the region with the smaller ratio between temperature gradients at melt surface and growing interface. The growth condition to obtain this region is suitable for the NOC method. The quality of a NOC ingot was first evaluated by determining the distribution of bulk micro defect (BMD), bulk stacking fault (BSF), oxidation induced stacking fault (OISF) and crystal originated particle (COP) in the cross section of the dislocation-free ingot. A ring-OISF or BSF can be clearly observed on the cross section cut from the middle and tail parts of the ingot. BMD can be observed inside and outside the ring-OISF. By estimated comparison to the none-doping case, CV (=CI) at cross point clearly has a trend to increase for the high B-doing case (5 × 1019 cm−3), where CV and CI are concentration of vacancy and interstitial Si atom, respectively. Nevertheless, in comparison with the distribution of COP in Cz and NOC ingots, COP cannot be detected inside the ring-OISF in the NOC wafer under the same etching process as many COP can be observed in the Cz wafer. This may be a unique feature for NOC growth.

Design of Large Poly‐Si on Oxide Interdigitated Back Contact (POLO IBC) Silicon Solar Cells with Local Al–p+ Contacts in the Constraints of Measurement and Module Integration

Interdigitated back contact (IBC) silicon solar cells with a passivating n‐type poly‐Si on oxide emitter and an aluminum‐doped p+ base contact on M2‐sized Ga‐doped p‐type Cz wafers are reported. The Al‐doped base contact forms during the firing of the printed contacts and allows for a lean process flow. The device optimization balances recombination at the base contacts against resistive losses and respects constraints set by the need of interconnecting cells in a module and contacting the cells temporally by a measurement chuck. A special sample holder is designed for measuring the Isc–Voc curve of the IBC cell with a busbar‐less metal grid. The pseudo‐efficiency is 24.7%. All fingers of each polarity are connected with wires and an efficiency of 22.3% is measured. The comparison of simulations and measurements reveals that the cell has 23.4% efficiency without the series resistance losses due to the wires. A huge part of the resistive losses in the cell are the transport losses of the majorities in the base dissipating a power that corresponds to 0.76%abs efficiency and the resistive losses at the Al‐doped base contact (0.29%abs).

Annealing-temperature-dependent evolution of hydrogen-related donor and its strong correlation with X-photoluminescence center in proton-irradiated silicon

We have investigated the formation and decay of hydrogen-related donors (HDs) and irradiation-induced intrinsic defects. N-type m:Cz and FZ silicon wafers, which were irradiated with 2 MeV protons and subsequently annealed at 100–600 °C, were analyzed using spreading resistance profiling and photoluminescence (PL). HDs formed at 260 °C and then disappeared in two stages at 400–440 and 500–540 °C. This decay behavior indicates the existence of two types of HDs with different thermal stabilities. PL measurements showed interstitial silicon clusters (W and X center), a carbon–oxygen complex (C center), and exciton lines bound to unknown shallow centers. The origin of the HDs was investigated based on the correlation of the formation and decay temperatures between HDs and irradiation-induced defects. The predominant defects at the early stage of annealing, such as the C and W centers, are ruled out as candidates for the core defects of HDs because annealing above 260 °C is indispensable for the HD formation. In contrast, the X center was found to be thermally generated above 200 °C and disappeared at 580 °C. The similarity of the formation and decay temperatures between the X and HD centers suggests that HDs are associated with the formation of the interstitial silicon-related defects attached to hydrogen. Our results suggest that controlling the formation of interstitial silicon-related defects is important for realizing desirable doping profiles with high accuracy and reproducibility for power devices. Annealing above 400 °C exclusively provides thermally more stable HDs, leading to the realization of more rugged power devices.

A route towards high‐efficiency silicon heterojunction solar cells

AbstractIn this work, we propose a route to achieve a certified efficiency of up to 24.51% for silicon heterojunction (SHJ) solar cell on a full‐size n‐type M2 monocrystalline‐silicon Cz wafer (total area, 244.53 cm2) by mainly improving the design of the hydrogenated intrinsic amorphous silicon (a‐Si:H) on the rear side of the solar cell and the back reflector. A dense second intrinsic a‐Si:H layer with an optimized thickness can improve the vertical carrier transport, resulting in an improved fill factor (FF). In order to reduce the plasmonic absorption at the back reflector, a low‐refractive‐index magnesium fluoride (MgF2) is deposited before the Ag layer; this leads to an improved gain of short circuit current density (Jsc). In total, together with MgF2 double antireflection coating and other fine optimizations during cell fabrication process, ~1% absolute efficiency enhancement is finally obtained. A detailed loss analysis based on Quokka3 simulation is presented to confirm the design principles, which also gives an outlook of how to improve the efficiency further.

Effect of carbon, oxygen, and intrinsic defects on hydrogen-related donor concentration in proton irradiated n-type silicon

We investigated the effect of the concentration of carbon, oxygen, and irradiation-induced intrinsic defects on hydrogen-related donor (HD) concentration. Several n-type silicon wafers having different carbon and oxygen concentrations were irradiated with 2 MeV protons, subsequently annealed at 300–400 °C, and analyzed by spreading resistance profiling. The HD concentration had no correlation with carbon and oxygen concentration. Additionally, the HD concentration showed a strong increasing linear dependence with proton-irradiation dose at 350 and 400 °C and a square root dependence at 300 °C. In the decay process of HD concentration at 400 °C, fast- and slow-decay components were observed regardless of wafer type. Our results show that the HD formation is based on the interactive process of irradiation-induced intrinsic defects and hydrogen, rather than hydrogen-catalyzed thermal double donor formation. Magnetic-field-applied Czochralski (m:Cz) wafers with 300 mm diameter, which are critical for the production scaling of power devices, have a relatively higher oxygen concentration than conventional floating-zone wafers. Our results further suggest that controlling the intrinsic defect formation, rather than oxygen impurity concentration, is more important in realizing designed doping profiles with high accuracy and reproducibility for next-generation power devices using large-diameter m:Cz wafers as a standard starting material.

24.58% efficient commercial n‐type silicon solar cells with hydrogenation

AbstractTunnelling oxide passivated contact (TOPCon) solar cells are gaining significant commercial interest, due to the potential for high efficiency. Industrially, this passivated contact scheme is typically coupled with an n‐type Czochralski (Cz) wafer. JinkoSolar Holding Co., Ltd. is one of the leading manufacturers that are producing n‐type TOPCon solar cells (referred to as ‘HOT’ cells) on a commercial scale. In this work, the influence of a post‐cell hydrogenation step, using illumination from an LED light source, on the performance and stability of n‐type TOPCon solar cells is investigated. The incorporation of this additional hydrogenation treatment led to an average efficiency enhancement of 0.64%abs on a batch of 50 cells made in an industrial environment. This significant improvement was caused by a 6.9 mV and 1.04%abs increase in open‐circuit voltage (VOC) and fill factor (FF), respectively. We also assessed the stability and found almost no light‐ and elevated temperature‐induced degradation (LeTID) in hydrogenated n‐type TOPCon cells. Testing at 70 ± 5°C under 1‐sun illumination revealed that the maximum degradation is limited to 0.06%rel. Following further stability testing, the efficiency increased beyond the initial value, up to 0.4%rel increase after 120 h. By incorporating this hydrogenation process into the production, an average line efficiency of 24.08% and VOC of 707.5 mV was achieved. The champion cell from the batch displayed an efficiency of 24.58%, as certified by measurement at the Chinese Academy of Sciences.

Progress with Defect Engineering in Silicon Heterojunction Solar Cells

Due to the low temperature processing constraint in silicon heterojunction (SHJ) solar cells, no defect engineering to improve silicon wafer quality is typically incorporated during cell fabrication. This means that high‐quality n‐type Cz wafers are required to produce high‐efficiency cells. Herein, recent demonstrations of incorporating defect engineering approaches, such as gettering and hydrogenation, into the SHJ process flow for both n‐type and p‐type wafers are surveyed. Defect engineering on wafers before cell fabrication can significantly improve the quality of low‐lifetime p‐type wafers, making them much more suitable for SHJ cells. Interestingly, these same approaches are also very effective in improving the cell performance in the n‐type wafers that are conventionally used in industry. Post‐cell illuminated annealing processes have been shown to eliminate boron–oxygen light‐induced degradation (LID) in p‐type wafers, leading to stable V OC exceeding 735 mV. New results indicate that the hydrogen naturally incorporated during SHJ processing is sufficient to passivate these defects. Similar illuminated annealing processes also lead to substantial improvements in n‐type SHJ cells. With these findings considered, it is demonstrated that a modified SHJ process flow, which includes defect engineering on a wafer level and post‐cell hydrogen passivation, is beneficial for SHJ production, regardless of the wafer type.

Influence of oxygen on trap-limited diffusion of hydrogen in proton-irradiated n-type silicon for power devices

The growing demand for power devices has led to the use of magnetic field-applied Czochralski (m:Cz) wafers owing to the limited production capacity and available diameters of the traditionally used floating zone (FZ) wafers. Consequently, the influence of oxygen impurities in the wafers on the electrical properties of devices, regardless of the growth method, needs to be investigated to achieve a stable fabrication process for power devices. Using the proton irradiation doping process and spreading resistance profiling technique, we evaluated the effective diffusion coefficient (Deff) related to trap-limited diffusion of hydrogen and the effects of impurities on diffusivity. We irradiated n-type silicon wafers, which have different carbon, oxygen, and phosphorus concentrations, with 2 MeV protons and annealed them at 300–400 °C. By analyzing the width of the n-type region, where hydrogen-related shallow donors (HDs) are induced, we estimated Deff to be five to six orders of magnitude lower than the intrinsic diffusion coefficient, indicating that hydrogen motion is highly trap-limited. Deff was significantly dependent on the oxygen concentration, and the activation energy of hydrogen diffusion varied from 0.57 ± 0.15 eV (pure epitaxial wafer) to 2.19 ± 0.15 eV (m:Cz wafer). This trend suggests that oxygen-related defects preferentially trap the mobile hydrogen released from thermally dissociated HDs. This study also reveals that the diffusion coefficients of different materials when annealed at 400 °C are comparable. This information is essential to realize the cost-effective production of power devices because we can treat m:Cz and FZ wafers equivalently during the doping process.

Improving surface passivation on very thin substrates for high efficiency silicon heterojunction solar cells

Silicon solar cells are now less than 3% absolute from the theoretical efficiency limit. Advanced passivated contact architectures have demonstrated surface saturation current densities close to 1 fA/cm2. We have optimized the thin intrinsic hydrogenated amorphous silicon layer by controlling the deposition temperature and the silane-to-hydrogen dilution ratio. Thin wafers were used as a testbed to increase the sensitivity to surface passivation. By optimizing the intrinsic layer, we reduced the surface saturation current densities from 1.7 fA/cm2 to 0.6 fA/cm2 on textured wafers with thicknesses ranging between 40 and 180 μm. Implied open-circuit voltages over 760 mV were accomplished on p-i/c-Si/i-n structures deposited on n-type CZ wafers with wafer thicknesses below 50 μm. Further, we demonstrated experimentally the potential of using very thin wafers by manufacturing screen-printed silicon heterojunction solar cells on 40 μm thick standalone wafers while achieving an efficiency of 20.48%.

Spatial Atomic Layer Deposition of Aluminum Oxide as a Passivating Hole Contact for Silicon Solar Cells

Herein, tunneling aluminum oxide (Al2O3) passivation layers are demonstrated to be a candidate for hole collecting, passivating contacts when coupled with a boron‐doped surface. These very thin Al2O3 films (1.5–3 nm) are deposited using spatial atomic layer deposition (ALD) on boron diffused (110–115 Ω □−1) hydrophilic surfaces operating as metal–insulator–semiconductor (MIS) contacts. The emitter saturation current density values of ≈57 fA cm−2 are achieved for the Al2O3 film thicknesses of 2 nm before metallization. At this same thickness, the contact resistivity values of 33 mΩ cm2 are obtained after metallization. Most importantly, the boron diffusion, wet chemical surface preparation, and spatial ALD of Al2O3 used to create these MIS structures are all performed with industrially relevant equipment on Cz wafers.

Industrial TOPCon solar cells on n-type quasi-mono Si wafers with efficiencies above 23%

Since 2017, the efficiency record for cast multi-crystalline silicon (mc-Si) solar cells has been 22.3% using a TOPCon design on a 2 × 2 cm2 area of an n-type wafer processed in the laboratory. As the efficiency of industrial PERC and TOPCon cells on monocrystalline wafers has been higher than that, there is widespread expectation that multi-crystalline materials will be used less and less. Recently, we applied our industrial TOPCon (i-TOPCon) technology to 157.4 mm × 157.4 mm quasi-mono n-type wafers and achieved an efficiency of 23.22% with Voc of 711.9 mV with illumination for the front side, independently confirmed by ISFH CalTeC in Germany. To the best of our knowledge, this is the first solar cell based on cast-mono material that exceed 23% efficiency. This gives scope that cast materials may still be used in future cell fabrication because such high efficiencies and lower cost may offset the lower production yield compared to Czochralski (Cz) wafers. We report on details about our cell structure and processing, we analyze the bulk lifetime before and after cell fabrication, and give details about the optimization of the rear passivating contact.

Large‐Area Boron‐Doped 1.6 Ω cm p‐Type Czochralski Silicon Heterojunction Solar Cells with a Stable Open‐Circuit Voltage of 736 mV and Efficiency of 22.0%

Herein, large‐area defect‐engineered p‐type silicon heterojunction (SHJ) solar cells using standard 1.6 Ω cm commercial‐grade boron‐doped Czochralski (Cz) silicon wafers are fabricated. It is demonstrated that despite achieving an open‐circuit voltage of 735 mV with an efficiency of 21.6% for gettered samples, without appropriate treatment, the cells are heavily susceptible to boron–oxygen‐related light‐induced degradation (LID), with the effective lifetime at maximum power point decreasing to 13 μs. This degradation results in a loss of efficiency of more than 3.1%abs (14.3%rel) after 48 h of light soaking. However, the addition of an advanced hydrogenation postcell fabrication process increases the efficiency by 0.2%abs to 21.8%, and dramatically reduces susceptibility of LID, decreasing the extent of degradation to 0.2%abs (0.9%rel). A peak stable independently measured efficiency of 22.0% with an open‐circuit voltage (VOC) of 736 mV is achieved with the addition of a dedicated high‐temperature prefabrication hydrogenation. These results indicate that p‐type Cz wafers can be used to fabricate stable, next‐generation high‐efficiency solar cells using silicon heterojunctions or other passivated contact architectures requiring VOCS well above 700 mV.

Carrier‐Selective Contact Based Silicon Solar Cells Processed at Room Temperature using Industrially Feasible Cz Wafers

For the broad use of solar photovoltaic devices, the device fabricated at commercially viable silicon wafers at room temperature is more preferable to harvest abundant solar energy. Silicon heterojunction solar cells at room temperature, based on carrier‐selective layers without using any specified surface passivation layer on the silicon wafer is fabricated. Industrially feasible Cz n‐type non‐textured silicon wafers having the bulk lifetime of 300 µs are used for cell fabrication. The molybdenum oxide (MoOx) and lithium fluoride (LiFx) are used as hole‐ and electron‐selective layers, respectively. The highest conversion efficiency of >15% from the simple architecture of Ag/TCO/MoOx/n‐Si/LiFx/Al is achieved. The internal quantum efficiency of ≈96% is observed in the shorter wavelength region, whereas to understand relatively less response between 800 and 1100 nm wavelength region; effective minority carrier diffusion lengths are estimated. The authors also confirm the inversion layer formation in the silicon after MoOx contact from temperature dependent capacitance‐voltage measurements, and quantify the crucial built‐in‐potential of ≈0.69 V from the cell structure due to the high work function of MoOx layer.

26.1%‐efficient POLO‐IBC cells: Quantification of electrical and optical loss mechanisms

AbstractWe present experimental results for interdigitated back contacted (IBC) solar cells with passivating POLO contacts for both polarities with a nominal intrinsic poly‐Si region between them. We reach efficiencies of 26.1% and 24.9% on a 1.3 Ω cm and 80 Ω cm p‐type FZ wafer and 24.6% on a 2 Ω cm n‐type Cz wafer, respectively. The initially measured implied efficiency potentials of the cells after passivating the surfaces are very similar, namely, 26.8%, 26.8%, and 26.4%, respectively. We attribute the difference between the efficiency potential and the final current‐voltage measurement to degradation, perimeter, and series and shunt resistance losses, which we quantify by lifetime measurements. With these measurements in combination with a finite element simulation, we determine the surface recombination velocity in the nominal intrinsic poly‐Si region to be in the range from 13 to 21 cm s−1. Using the same approach, we analyze the increase of the front surface recombination velocity during cell processing from 2 to 10 cm s−1 for the 1.3 Ω cm and from 0.5 to 2.3 cm s−1 for the 80 Ω cm. This leads to the fact that cells fabricated on lowly doped bulk material are more vulnerable to a process‐induced degradation of the surface passivation quality. We further determine the theoretical limits of the cells by firstly idealizing the recombination (28% for 1.3 Ω cm and 28.2% for 80 Ω cm) and secondly also idealizing the optics of the solar cells (29.4% and 29.5%).

Numerical exploration for structure design and free-energy loss analysis of the high-efficiency polysilicon passivated-contact p-type silicon solar cell

In this work, the application of the p-type and n-type polysilicon passivated contact on industrial-level p-type silicon solar cell is studied using numerical simulation. The effects of (i) the structure design, (ii) the bulk lifetime and resistivity of the p-type wafer, and (iii) the carrier selectivity of polysilicon passivated contact on cell performances are investigated. Furthermore, the corresponding energy-loss pathways are classified by using free energy loss analysis (FELA). In essence, the rear-junction solar cell with the n-type polysilicon passivated-contact generates more internal power because of the better surface passivation and less front metallization shading, but the efficiency potential is limited by the low lifetime of the state-of-the-art p-type Czochralski (Cz) wafer. Thus, the p-type polysilicon passivated contact serving as the back-surface field would be more favorable if the lifetime of the p-type Cz silicon were less than 350 μs. Over the long term, the lifetime of the p-type wafer possibly becomes the bottleneck of the high-efficiency polysilicon passivated-contact solar cells. Finally, we present the roadmap toward the 23% industrial p-type silicon solar cell with the p-type or n-type polysilicon passivated contact.

N-type silicon solar cell with rear tunnel oxide combined with rear screen-printed electrodes

In this article, we describe an n-type silicon solar cell with a conversion efficiency of 21.9% on a 156 mm pseudosquare Czochralski (Cz) wafer with a rear tunnel oxide-passivated structure and rear screen-printed electrodes. The thermal degradation of the rear structure during electrode firing was solved by adding a SiNx film on the rear surface. Not only the SiNx film but also the tunnel oxide structure was chemically etched using fire-through screen-printed electrodes, so that the electrodes contacted directly to the bulk silicon. A phosphorus-diffused layer appeared at the bulk silicon immediately below the tunnel oxide, and this shallowly and densely diffused layer reduced the contact resistivity between the electrodes and the bulk silicon layer. Degradation of rear passivation caused by etching on the tunnel oxide structure was improved by adding a diffused layer using POCl3 before chemical oxidation. The diffused layer using POCl3 was deep and had a low phosphorus concentration, thus the field effect worked successfully while Auger recombination was not enhanced.

Evaluation of oxygen precipitation behavior in n-type Czochralski-Si for photovoltaic by infrared tomography: Effects of carbon concentration and annealing process conditions

In this study, we evaluated the effects of carbon and annealing process conditions on the oxygen precipitation in an n-type Czochralski (Cz) wafer for solar cells by infrared light scattering tomography (IR-LST). It was confirmed that precipitates grow larger and denser as carbon concentration increase. Thus, carbon promotes oxygen precipitation. We also evaluated the effect of oxygen precipitation on the minority carrier lifetime by photoluminescence (PL) imaging. It was confirmed that the interface between the precipitate and the Si matrix is a dominant recombination center since the surface area of the precipitate obtained by IR-LST measurement and the PL intensity show good correlation. In addition, it was also confirmed that carbon is involved in the supply of interstitial oxygen to precipitates through the formation and extinction of the thermal donor. We believe that it is important to understand and control the effect of carbon of controlling the oxygen precipitation behavior.

Kerfless epitaxial silicon wafers with 7 ms carrier lifetimes and a wide lift-off process window

Silicon wafers contribute significantly to the photovoltaic module cost. Kerfless silicon wafers that grow epitaxially on porous silicon (PSI) and are subsequently detached from the growth substrate are a promising lower cost drop-in replacement for standard Czochralski (Cz) wafers. However, a wide technological processing window appears to be a challenge for this process. This holds in particularly for the etching current density of the separation layer that leads to lift-off failures if it is too large or too low. Here we present kerfless PSI wafers of high electronic quality that we fabricate on weakly reorganized porous Si with etch current densities varying in a wide process window from 110 to 150 mA/cm2. We are able to detach all 17 out of 17 epitaxial wafers. All wafers exhibit charge carrier lifetimes in the range of 1.9 to 4.3 ms at an injection level of 1015 cm−3 without additional high-temperature treatment. We find even higher lifetimes in the range of 4.6 to 7.0 ms after applying phosphorous gettering. These results indicate that a weak reorganization of the porous layer can be beneficial for a large lift-off process window while still allowing for high carrier lifetimes.