Search

Crystalline silicon becomes photosensitive and absorbing in the sub-bandgap spectral region if hyperdoped, i.e. supersaturated to a level above the solubility limit in thermal equilibrium, by deep impurities, such as sulfur. Here we apply femtosecond laserpulses to crystalline silicon in a SF6 atmosphere as hyperdoping method. The ultrashort laser pulses cause crystal damage and amorphous phases that would decrease quantum efficiency in a potential optoelectronic device application. We investigate five different post-hyperdoping methods: three etching techniques (ion beam etching IBE, reactive ion etching RIE, and wet-chemical etching HNA) as well as ns-annealing and minute-long thermal annealing and study their impact on crystallinity by Raman spectroscopy and absorptance in the visible and near infrared wavelength regime. We use femtosecond laser hyperdoped silicon (fs-hSi) with two different levels of surface roughness to study a potential dependence on the impact of post-treatments. In our investigation, ns-annealing leads to the best results, characterized by a high Raman crystallinity and a high remaining absorptance in the sub-bandgap spectral region of silicon. Within the used etching methods IBE outperforms the other etching methods above a certain level of fs-hSi surface roughness. We relate this to the specific anisotropic material removal behavior of the IBE technique and back this up with simulations of the effect of the various etching processes.

Ultrashort pulse laser processed sulfur hyperdoped black silicon represents a promising silicon-based material for infrared optoelectronic applications due to its high sub-bandgap optical absorptance. Non-thermal melting and resolidification processes associated with such laser processing, however, result in amorphous and polycrystalline phases which may be detrimental for this purpose. Furthermore, the sulfur impurities are electrically inactive, impeding the formation of a rectifying junction. This work demonstrates an ultrafast laser heating process based on heat accumulation with laser pulses of 10 ps pulse duration at high repetition rates of 41 MHz and peak fluences between 33% and 66% of the ablation threshold as a method to (i) recrystallize the material and (ii) electrically activate the sulfur dopants while (iii) maintaining the sub-bandgap absorption. Furthermore, laser heating recovers the optical activity of sulfur states that have been previously deactivated by thermal annealing. The demonstrated process can have versatile applications in material functionalization due to its highly localized heat input accompanied by high cooling rates.

We set up an analytical optical model to emulate the absorptance spectra of light scattering, sulfur-hyperdoped silicon that we fabricate by using femtosecond laser pulses. The model allows us to distinguish between contributions to the absolute sub-bandgap absorptance from the path length enhancement of photons due to laser-induced surface roughness, on the one hand, and from the actual hyperdoped layer, on the other hand. Both effects are quantified via the two free parameters of the model. By varying the laser fluence and the areal pulse density, we create a range from almost planar to heavily structured hyperdoped Si samples that we show to behave almost like a Lambertian scatterer. The optical depth a1, i.e., the product of the absorption coefficient close to the Si bandgap energy and the effective thickness of the hyperdoped layer, scales with the surface area enhancement, which we identify as the main driving force for large sub-bandgap absorptances of this material type. It reaches maximum values of nearly a1 = 0.4, which refers to an absolute absorptance of 82% at a wavelength of 1450 nm. We furthermore discuss, quantify, and reduce possible error sources when determining the absorptance of such optically rough, hyperdoped samples with a spectrophotometer.

We present a simulation-based study for identifying promising cell structures, which integrate poly-Si on oxide junctions into industrial crystalline silicon solar cells. The simulations use best-case measured input parameters to determine efficiency potentials. We also discuss the main challenges of industrially processing these structures. We find that structures based on p-type wafers in which the phosphorus diffusion is replaced by an n-type poly-Si on oxide junction (POLO) in combination with the conventional screen-printed and fired Al contacts show a high efficiency potential. The efficiency gains in comparsion to the 23.7% efficiency simulated for the PERC reference case are 1.0% for the POLO BJ (back junction) structure and 1.8% for the POLO IBC (interdigitated back contact) structure. The POLO BJ and the POLO IBC cells can be processed with lean process flows, which are built on major steps of the PERC process such as the screen-printed Al contacts and the $$\text{Al}_\text{2 }\text{O}_\text{3 }/\text{SiN }$$ Al 2 O 3 / SiN passivation. Cell concepts with contacts using poly-Si for both polarities ($$\text{POLO}^2$$ POLO 2 -concepts) show an even higher efficiency gain potential of 1.3% for a $$\text{POLO}^2$$ POLO 2 BJ cell and 2.2% for a $$\text{POLO}^2$$ POLO 2 IBC cell in comparison to PERC. For these structures further research on poly-Si structuring and screen-printing on p-type poly-Si is necessary.

Wide-bandgap perovskite top solar cells (PSCs) with optimal bandgap (Eg) are key to boost the efficiency of perovskite/Si tandem devices beyond the Shockley-Queisser limit for single-junction solar cells. However, the large open circuit voltage (Voc) deficit in the optimal bandgap range and the poor transmission of the top semi-transparent perovskite solar cells (s-PSCs) restricts the development in this field. Here, we present a novel 2D/3D perovskite heterostructure architecture to reduce the voltage deficit in PSCs. The reduced voltage deficit is a result of the decreased non-radiative recombination losses at the perovskite/hole-transport layer interface. Employing the 2D/3D perovskite heterostructure, efficient four-terminal (4T) perovskite/Si tandem solar cells with a stabilized power conversion efficiency (PCE) of up to 25.7% is demonstrated. In order to improve the PCE further, we present alternative transparent conductive oxide electrodes that reduce the parasitic absorption and reflection losses and enhances the transmission in the near infrared wavelengths, leading to a potential PCE of 27.4% for 4T perovskite/c-Si tandem devices.

Unusually high carrier lifetimes are measured by photoconductance decay on n-type Czochralski-grown silicon wafers of different doping concentrations, passivated using plasma-assisted atomic-layer-deposited aluminum oxide (Al2O3) on both wafer surfaces. The measured effective lifetimes significantly exceed the intrinsic lifetime limit previously reported in the literature. Several prerequisites have to be fulfilled to allow the measurement of such high lifetimes on Al2O3-passivated n-type silicon wafers: (i) large-area wafers are required to minimize the impact of edge recombination via the Al2O3-charge-induced inversion layer, (ii) an exceptionally homogeneous Al2O3 surface passivation is required, and (iii) very thick silicon wafers are needed. Based on our lifetime measurements on n-type silicon wafers of different doping concentrations, we introduce a new parameterization of the intrinsic lifetime for n-type crystalline silicon. This new parameterization has implications concerning the maximum reachable efficiency of n-type silicon solar cells, which is larger than assumed before.

We demonstrate damage-free laser contact openings in silicon oxide layers on polycrystalline silicon on oxide (POLO) passivating contacts. A pulsed UV-laser evaporates the upper part of the polycrystalline silicon layer, lifting off the silicon oxide layer on top. On n-type POLO (and p-type POLO, respectively) samples a saturation current density of 2 fA cm−2 (6 fA cm−2) and an implied open-circuit voltage of 733 mV (727 mV) are achieved with a laser contact opening area fraction of 12.3% (8.7%). The application of this ablation process in an interdigitated back contact solar cell leads to an independently confirmed power conversion efficiency of 26.1%. The excellent contact quality of the laser contact openings is proven by the low series resistance of 0.1 Ω cm2 on the solar cell with a contact area of only 3%.

We report on an industrial large area, screen-printed, double-side contacted cell with polysilicon on oxide (POLO) junctions on both sides and an energy conversion efficiency of 22.3% ( A = 244.15 cm 2, V oc = 714 mV, FF = 81.1%, J sc = 38.5 mA/cm2, measured in-house). This cell shows an extraordinarily low series resistance below 0.05 Ω cm2. This confirms the low specific junction resistance observed recently for POLO junctions. The present cell suffers from 1) low short-circuit current due to parasitic absorption in the rather thick poly-Si (30 nm), as well as in the indium tin oxide, 2) deterioration of the recombination behavior upon sputter deposition of a transparent conductive oxide (TCO), and 3) shunts near the edge due to nonadapted TCO edge exclusion. We address all of these limitations experimentally. In particular, we developed a plasma-enhanced chemical vapor deposition process for ZnO:Al, which does not compromise the passivation of the POLO junctions underneath. An estimation of the efficiency potential (based on the two-diode model and the assumption that all these building blocks can be successfully combined on a cell level) shows that 25.3% can be achieved with this cell concept. We also look into potential cost advantages of the POLO junction scheme for this cell structure, such as the usage of p-type Cz-Si material and the omission of Ag fingers.

We introduce a method for the quantification of perimeter recombination in solar cells based on infrared lifetime measurements. We apply this method at a 25.0%-efficient interdigitated back contact (IBC) silicon solar cell with passivating contacts. The implied pseudo-efficiency determined by infrared lifetime mapping is 26.2% at an intermediate process step. The 1.2%abs loss is attributed to a process-related reduction in surface passivation quality, recombination in the perimeter area, and series resistance. The 2 × 2 cm2 -sized cell is processed on a 100 mm wafer. We determine the implied pseudo-efficiency with illuminated and with shaded perimeter area during infrared lifetime mapping. The difference between both implied pseudo-efficiencies yields the efficiency loss by perimeter recombination, which is determined to be 0.4%abs for a wafer resistivity of 1.3 Ω cm and even 0.9%abs for a wafer resistivity of 80 Ω cm.

The maximum achievable silicon single junction solar cell efficiency is limited by intrinsic recombination and by its limited capability of absorbing sun light. For Lambertian light trapping the maximum theoretical solar cell efficiency is around 29.5%. Recently a new approach for light trapping has been proposed for silicon photovoltaics. Highly regular structures with a size in the range of the wavelengths of the incident light act as so-called photonic crystals. Such structures allow wave-interference light trapping beyond the Lambertian limit. Applying these photonic crystals to silicon solar cells can help to reduce the absorber thickness and thus to minimizing the unavoidable intrinsic recombination. From a simulation study, we can conclude that 31.6% is the maximum possible single junction solar cell efficiency for a 15 μm-thin substrate. Furthermore, we present a process flow for the preparation of regular inverted pyramid structure, that acts as photonic crystal. Finally, regular inverted pyramid structures are prepared on polished and shiny-etched, i. e. on surfaces with a certain roughness, substrates. Surface passivation of these structured surfaces shows as good lifetimes as on conventional randomly pyramid textured surface. Excellent total saturation current densities on asymmetric samples of 4 ± 2 fA/cm2 for n-type and of 4.5 ± 2.2 fA/cm2 on p-type substrates are obtained.

The widely accepted limiting efficiency for crystalline silicon solar cells with Lambertian light trapping under 1 sun was previously calculated to be 29.43% for a 110-μm-thick device by using the commonly applied weak absorption approximation for light trapping. However, the short-circuit current density increases by 0.17 mA/cm2 when modeling the optical absorptance of an ideal Lambertian light trapping scheme exactly. The resulting new 1-sun efficiency limit is 29.56% and holds for a cell that is 98.1 μm in thickness.

We present a systematic study on the benefit of the implementation of poly-Si on oxide (POLO) or related junctions into p-type industrial Si solar cells as compared with the benchmark of Passivated Emitter and Rear Cell (PERC). We assess three aspects: (a) the simulated efficiency potential of representative structures with POLO junctions for none (=PERC+), one, and for two polarities; (b) possible lean process flows for their fabrication; and (c) experimental results on major building blocks. Synergistic efficiency gain analysis reveals that the exclusive suppression of the contact recombination for one polarity by POLO only yields moderate efficiency improvements between 0.23%abs and 0.41%abs as compared with PERC+ because of the remaining recombination paths. This problem is solved in a structure that includes POLO junctions for both polarities (POLO2), for whose realization we propose a lean process flow, and for which we experimentally demonstrate the most important building blocks. However, two experimental challenges—alignment tolerances and screen-print metallization of p+ poly-Si—are unsolved so far and reduced the efficiency of the “real” POLO2 cell as compared with an idealized scenario. As an intermediate step, we therefore work on a POLO IBC cell with POLO junctions for one polarity. It avoids the abovementioned challenges of the POLO2 structure, can be realized within a lean process flow, and has an efficiency benefit of 1.59%abs as compared with PERC—because not only contact recombination is suppressed but also the entire phosphorus emitter is replaced by an n+ POLO junction.

We report on the transfer of our lab-type POLO2-IBC process with POLO contacts for both polarities towards an industrial level. Here we demonstrate a shortened cell fabrication process that uses p-type wafers and keeps the Al-back surface field of the PERC process but substitutes the phosphorous diffusion by a n-type poly-Si deposition. The resulting POLO-IBC process is similarly short as the PERC process. A high lifetime with the Cz material and highly selective POLO junctions require a reduce thermal budget and a reduced thickness of the interfacial oxide compared to our previous lab cells that used FZ silicon wafers. Our POLO-IBC cells have an efficiency potential of 24.5 % as deduced from simulations. We measure an efficiency of 21.8 % after finishing the first cell batch. For a cell from our second cell batch with improved passivation we measure an implied pseudo efficiency of 25.2 % before laser contact openings.

In recent years, poly-Si based passivated contacts elevated the conversion efficiencies of crystalline Si solar cells to levels of 26%abs due to their outstanding electrical surface passivation performance and current transport characteristics. A major associated challenge, however, is the large parasitic light absorption within the doped poly-Si, regardless if the contacts are applied on the front and/or on the rear side of the solar cell. It, therefore, might be beneficial to confine the passivated contacts to local regions underneath the metal contacts. We present an effective and flexible laser-based approach to structure the poly-Si layer after its full-area deposition. Laser pulses with a pulse duration of 9 ps and a wavelength of 355 nm trigger an amorphization of the poly-Si surface. The minimum threshold fluence for amorphization is between 89 and 129 mJ/cm2. The a-Si layer, which is laterally homogeneous and up to (33 ± 4) nm in thickness, works as an etch barrier in an alkaline solution. The most robust barrier corresponding to the maximum thickness of the a-Si layer is found for a fluence of (270 ± 30) mJ/cm2. Besides the impact of the laser fluence on the etch resistiveness of the modified poly-Si layer, we study the role of oxygen during the laser process. We find that oxygen becomes incorporated into the material for certain laser fluences, which results in a more robust etch barrier. The amount of oxygen incorporated is below 3 wt. %. Eventually, we present a phenomenological model of our findings.

Perimeter recombination is a relevant loss mechanism, in particular for cells with a large perimeter-to-area ratio and with poorly passivated edges, e.g., cut or cleaved solar cells for shingled modules. We experimentally demonstrate that cut edges can be well passivated during front-end processing. The resulting cells have an efficiency of 26%. The designated cell area of our lab-type highly efficient cells is smaller than the total area of the wafer. This causes recombination losses in the masked perimeter region. We separate the active cell area from the wafer on two sides of the cell by slits to reduce the transport of carriers into the perimeter region. We apply a diffusion model to describe impact of the slits on the perimeter recombination. The slits have an effective surface recombination velocity of down to 9 cm/s, depending on the resistivity of the base. For a base resistivity of 80 Ωcm, the average cell efficiency increases by 0.7 % abs as compared to embedded cells and by 2.3 % abs as compared to laser-cut cells due to the passivated slits.

Highly efficient solar cells require minimized recombination and maximized optical absorption. We apply Si nanopowder with a median particle size of 500 nm to the rear side of poly-Si on oxide (POLO) passivated Si wafers that have a planar front side. The enhanced optical absorption consists of a useful component from the wafer and useless absorption by the Si pigments and the poly-Si layer. We derive and successfully apply an analytical model that accounts for both contributions and for the light trapping that is caused by light scattering at the nanopowder layer. We measure and model that this rear side increases the photogenerated current density by 1.3 mA/cm2 for a 140 μm-thick planar cell. We compare the performance of the Si-pigmented diffuse rear side reflectors (PDR) with reflectors using random pyramids (RPs) and POLO junctions. We find that for full surface coverage by Si nanopowder, the better surface passivation compensates for an inferior optical performance of a PDR when compared to RP.

We demonstrate an independently confirmed 25.0%-efficient interdigitated back contact silicon solar cell with passivating polycrystalline silicon (poly-Si) on oxide (POLO) contacts that enable a high open circuit voltage of 723 mV. We use n-type POLO contacts with a measured saturation current density of J 0n = 4 fA cm?2 and p-type POLO contacts with J 0p = 10 fA cm?2. The textured front side and the gaps between the POLO contacts on the rear are passivated by aluminum oxide (AlO x ) with J 0AlO x = 6 fA cm?2 as measured after deposition. We analyze the recombination characteristics of our solar cells at different process steps using spatially resolved injection-dependent carrier lifetimes measured by infrared lifetime mapping. The implied pseudo-efficiency of the unmasked cell, i.e., cell and perimeter region are illuminated during measurement, is 26.2% before contact opening, 26.0% after contact opening and 25.7% for the finished cell. This reduction is due to an increase in the saturation current density of the AlO x passivation during chemical etching of the contact openings and of the rear side metallization. The difference between the implied pseudo-efficiency and the actual efficiency of 25.0% as determined by designated-area light current?voltage (I?V) measurements is due to series resistance and diffusion of excess carriers into the non-illuminated perimeter region.

Kerf-free techniques for subdividing a single thick crystalline Si wafer into a multitude of thin Si layers have a large potential for cost reductions. In this paper, we explore pore formation in Si for separating many 18 μm-thick surface-textured layers from a thick wafer with a single etching process. We demonstrate the fabrication and separation of four macroporous Si layers in a single etching step. Generating many instead of single macroporous layers per etching step improves the economics of the macroporous Si process. We present our etching process that maintains the pore pattern defined by photolithography even after etching many absorber and separation layers. Federal Ministry for Environment, Nature Conservation, and Nuclear Safety/FKZ 0325147