Introduction
Numerical simulation method
Implementation and Results
A. Optimization of Organic Tandem Solar Cells
B. Optimization of Perovskite-Silicon Concatemers
- Introduction
- Numerical simulation method
- result
Introduction
Designing requires optimization of numerous structural and compositional parameters, such as the bandgap and layer thickness of the component materials, as well as the tandem interlayer design in the more industrially relevant monolithic tandem devices. Numerical device simulations can provide instrumental insights for determining optimal multilayer configurations. In organic tandem devices, while optical simulations of thin-film layer stacks are routinely used, simulations of complete optoelectronic devices including recombination junctions formed in the interlayer region are uncommon. In the case of , the widespread use of silicon heterojunction technology is characterized by the combination of large-scale textures and thin-film contact layers, but the specificity of perovskite materials in terms of ion transport effects also gives this Optical and electrical simulation of multijunction devices presents challenges.
We address the above challenges using an integrated optoelectronic device simulation framework designed for numerical optimization of organic solar cells and light-emitting devices. Electrical excitation of all-organic tandem devices is achieved by a . On the other hand, full-textured peroxide-silicon tandem (optical) optimization is achieved by combining light scattering models at the silicn texture and wave propagation models in thin-film silicon and peroxide layers into a dedicated multiscale simulation framework realized.
Figure 1. (a) Layer stack and energy level structure of a high-efficiency organic tandem solar cell. (b) SEM image of a hybrid tandem solar cell structure with a peroxide top absorber layer conformally deposited on both side textured silicon heterojunction solar cells.
Numerical simulation method
For the simulation of complex tandem solar cell devices, transport models of charge carriers and excitons need to be combined with a multiscale framework for optical simulations of quasi-one-dimensional structures. To this end, we use a combination of several modules of The charge carriers due to light absorption in the active layer are simulated using the , which is coupled to the Figure 2(a) shows a multiscale optics approach in which a three-dimensional ray-traced simulation of the coherently stacked optics in the form of a transition matrix and scattering of large-scale textures is coupled with a net radiation model of incoherent light propagation. Coupling is done by extracting transmission and reflection coefficients from transmission matrices and ray tracing models for net radiation formalism. For the simulation of charge carrier and exciton transport, the was used, the generation rates of which were inferred from optical simulations. The electronic model takes into account the special properties of organic semiconductors in terms of different mobility models reflecting disorder and local presence. To describe the charge transport at the heterointerface, which is suppressed in the drift-diffusion diagram, a conforming to the Miller-Abrahams theory of thermally activated tunneling is employed [Fig. 2(b)]. Finally, optimization of PV device performance can be achieved in Setfos by sifting through the relevant configuration parameter space ("sweep"), or by defining objectives for local or global optimization algorithms.
Figure 2. Model for tandem solar cell simulation. (a) Multiscale optical model for large-scale texturing of conformal coatings as in peroxide-silicon tandem [OptEx]. (b) Miller-Abrahams hopping model [JAP] of charge transfer at organic-organic heterointerfaces associated with tandem OPVs.
Implementation and Results
A. Optimization of Organic Tandem Solar Cells
For the simulation of organic tandem solar cells, we consider Li et al.The described high-efficiency device structure includes a DR3TSBDT:PC71BM top cell and a DPPEZnP-TBO:PC61BM bottom cell with a reconstituted interlayer composed of ZnO nanoparticles and PEDOT:PSS combination. Thin layers of CuSCN and PFN were used as anode and cathode buffer layers, respectively. The electrodes are formed of ITO and Al. The layer structure and energy level arrangement of the experimental model system used for the simulations are shown in Fig. 1. Correspondingly, the HOMO-LUMO levels were taken from the experimental reference, and the electron and hole transport levels of the individual in vitro heterojunctions of the subunits were determined as Fig. 1(a). The initial layer stack and energy level arrangement—including the heterointerface—is shown in Figure 3.
Figure 3. Implementation of SETFOS (a) a layer stack with optical and electrical analog domains, and (b) an energy level arrangement including hopping interfaces that identify connecting subunits.
To validate the simulation method and extract unknown material parameters for consistent modeling of single-junction and tandem cells, the experimental JV curves were fitted using a The starting values of the optical and electrical parameters were taken from the literature. As in the experimental work, single-junction solar cells were realized with two absorber materials sandwiched between PEDOT:PSS and PFN/Al contacts as HTL. For the tandem simulations, the parameter values for the single junction material were taken as starting values and kept at the same order of magnitude when optimizing new parameters for the attempted frequency of CuSCN HTL, ZnO ETL/interlayer and LUMO-HOMO transfer at the hopping interface. In this way, all JV curves can achieve a good fit with a consistent set of parameters, as shown in Figure 4.
Figure 4: Model validation and parameter extraction by consistent fitting of the experimental current-voltage characteristics of Setfos' single-junction and tamdem cells using a global optimization algorithm.
Using the material parameters obtained in this way, we further optimize the structure (ie layer thickness) by optical simulation of the layer stack for normal incident illumination in the AM1.5g spectrum. The left side of Figure 5(a) shows the limiting photocurrent as a function of the thickness of the top and bottom absorbers, obtained by a "brute force" double sweep of the thickness parameter. The black cross marks the result of the global optimization of the thickness of the active layer, and the remaining layers are kept fixed, which is exactly the same as the frequency sweep result of the optimal configuration. In the next step, the number of layers in the global optimization is increased by subsequent optimization of the thicknesses of the CuSCN, ITO, ZnO and PEDOT layers. The right side of Fig. 5(a) shows the corresponding adjustment of the layer thickness. Starting from the optimal configuration in terms of optics (grey horizontal line), a comprehensive optoelectronic device optimization is performed by maximizing the electrical output power of the device. Likewise, the optimal configuration (black cross) determined by global optimization is in perfect agreement with the scan results of the active layer thickness, while the computation is an order of magnitude faster. Figure 5(b) shows the variation of the maximum efficiency with increasing number of layers in the global optimization, showing the potential of this method to yield device efficiencies that greatly exceed the results obtained by optical optimization alone.
Figure 5. (a) Left panel: Optical simulation of the limiting photocurrent as a function of the thickness of the top and bottom absorbers. The black cross on the current matching ridge marks the result of the global optimization. Right image. The optical optimization results of the short-circuit current density increasing with the number of layers, CuSCN, ITO, ZnO and PEDOT:PSS layer thicknesses were sequentially added in the configuration parameters allowed to vary. (b) Left: optoelectronic coupling simulation of the limiting photocurrent as a function of top and bottom absorber thickness. The black intersections again mark the results of the global optimization, while the grey intersections represent solutions configured using optical optimization. Right image. Changes in power conversion efficiency after structural optimization using a global optimization algorithm and full-drift diffusion device simulation.
B. Optimization of Perovskite-Silicon Concatemers
Recently, the above numerical optimization method was applied to the optical design of , in which the silicon device is a double-sided textured silicon heterojunction solar cell with perovskite conformally coated on top of the textured silicon [OptEx ]. As shown in Figure 6, the optical multiscale simulation method was successfully validated for three different device structures: A--flat cell, B--flat cell with textured anti-reflective foil on front-side contact, and C--backside textured Flat interface of silicon tandem with perovskite. For all devices, the simulated and experimental characteristics achieved close agreement.
Figure 6. Left: Cascade for validation. A - planar structure, B - planar stack with front textured foil, C - planar top cell and back textured crystalline silicon wafer. Right image. Comparison of measured and simulated EQE of a monolithic peroxidase-silicon tandem solar cell, whose structures are defined on the left.
In the first optimization step, for the nip and pin configurations, a rear textured cell with a planar top cell and a textured foil for the front contact is simulated [LHS of Fig. 7(a) and (b)]. The topography of the textured foil, measured by AFM (atomic force microscopy), was directly imported into the simulation. During the optimization process, in the case of the nip configuration, the thicknesses of the perovskite and solenoid layers were adjusted, while in the case of the pin, only the thickness of the peroxide layer was changed. The EQE was then fitted by simulating the absorbance, assuming an IQE of 95% for peroxide and 100% for silicon. This yielded label photocurrents—from 17.0 mA/cm2 to 18.9 mA/cm2—derived from the higher absorption in the perovskite and silicon layers of the latter configuration [RHS Figures 7(a) and (b) )], as a similar structure was observed in experiments.
Figure 7. Left: Device configuration used in numerical optimization. (a) Back textured crystalline silicon with flat top cell and front textured foil--nip structure, (b) same as (a) but with pin structure, (c) double-sided textured crystalline silicon wafer and calcium The pin configuration on the titanium layer. Right image. Comparison of layer-resolved absorptivity for the configurations shown on the left.
In a further step, the effect of the two-sided texture of the silicon wafer on the photocurrent was investigated. To this end, the conformal structure shown in Fig. 7(c) is optimized by tuning the MgF and perovskite layers with full consideration of the conformal texture. From the layer deabsorptivity shown in Fig. 7(c), it can be inferred that there is an additional photocurrent gain associated with the reduction of reflection losses at long wavelengths, resulting in a Jsc of 19.6 mA/cm2. The beneficial effect of double-sided texture was recently confirmed by the results of the experimental realization of a conformal perovskite deposition method [Nat. Mater].
In conclusion, we present a comprehensive numerical simulation approach suitable for the design, optimization and investigation of complex multijunction device structures with important applications in organic and hybrid photovoltaics. The combination of coherent and incoherent optics and carrier hopping at the drift diffusion and charge recombination interfaces enables simulation-based optimization of state-of-the-art tandem solar cell devices beyond standard estimates, evaluating optical and electronic losses at appropriate operating points. This enables the overall optoelectronic design of solar cells to be optimized under realistic operating conditions.