The global Copper Indium Gallium Selenide (CIGS) photovoltaic market has grown strongly over the last decade [1]. Among various novel semiconductors, CIGS is doubtless one of the most promising materials, performing exceptionally well in terms of carrier lifetime and processability while ensuring high conversion efficiency [2,3,4,5,6,7,8,9]. Numerical modeling and simulation have been employed since the early days of this research field to correlate particular device and material parameters with cell output characteristics [10,11,12,13,14]. Modern simulation tools, such as Atlas of Silvaco, are based on digital-resolution 2D transport equations that govern conduction mechanisms in semiconductor devices [1]. Another widely used tool is the Solar Cell Capacitance Simulator (SCAPS-1D), a well-established numerical tool for modeling thin-film solar cells by solving continuity and electrostatic potential equations in a steady-state environment [15].

CIGS is a direct bandgap chalcopyrite semiconductor with a polycrystalline nature, chemically represented as Cu(In_1-x, Ga_x)Se₂, where x can vary from 0 to 1 [16,17]. This tunability allows the bandgap to be easily changed from 1.0 eV to 1.7 eV by increasing the gallium-to-indium ratio of the absorber [18,19,20]. Effective bandgap engineering can reduce the recombination rate via electric field enhancement, thereby increasing the electron-hole pair production rate [21]. Recent electrical CIGS simulation baseline models have incorporated post-deposition treatment (PDT) effects, such as the use of heavy alkali fluoride compounds like KF, RbF, and CsF, resulting in significant performance improvements [22]. Figure 1, illustrates an updated SCAPS model schematic consisting of the CIGS absorber, a surface defect layer (SDL), a CdS buffer, and window layers of Zn_1-xMg_xO (ZMO) and Al:ZnO (AZO).

To enhance performance, several critical layers and their properties must be optimized. The window layer must be highly transmittable in the visible spectrum while maintaining high electrical conductivity to aid in the collection of photogenerated carriers [23]. The n-type conductivity of materials like ZnO enables effective electron transit, and doping with aluminum (ZnO:Al) increases conductivity to maintain low resistive losses [23]. In addition to traditional window layers, the role of front contacts like Indium Tin Oxide (ITO) has been thoroughly explored using SILVACO ATLAS to achieve ultra-high efficiencies [24]. The buffer layer, such as CdS, contributes to p–n-junction formation and enhances the lattice match between the absorber and window layers [18,25,26]. However, to avoid the toxic effects of cadmium, simulations have investigated alternative buffer layers like Zn(O,S) [1] or less toxic Indium Phosphide (InP) [27].

Further performance enhancements are achieved through the implementation of a back surface field (BSF) and rear-surface passivation. Including a BSF layer, such as thin CuInSe, in thin-film cells has a significant impact compared to conventional structures [28]. The introduction of a BSF layer aims to minimize recombination losses by optimizing the band alignment and electric field profile within the device [29]. Recent studies have reported the use of Sb₂S₃ as a back-surface field (BSF) layer, achieving power conversion efficiencies (PCE) above 31% in numerical simulations under idealized conditions, in which recombination losses are significantly suppressed and interface properties are optimized [30]. Passivation layers, such as Al₂O₃ at the rear interface, can effectively reduce carrier recombination and improve the open-circuit voltage (V_OC) [22].

Advanced architectures, such as tandem solar cells, integrate CIGS with other materials to exceed the efficiency threshold of single-junction PV cells [31]. Perovskite-CIGS tandem cells have demonstrated promise in overcoming these constraints, utilizing a top cell for high-energy photons and a CIGS bottom cell for low-energy photons [31]. Furthermore, innovative designs using double or triple-junction absorber layers, such as CZO/CIGS/CZTS configurations, have increased efficiency from 18.11% to 24.5% due to optimized light absorption and improved charge-carrier dynamics [32].

This section reviews one-dimensional numerical modeling of CIGS solar cells using SCAPS-1D. Its steady-state, one-dimensional formulation makes SCAPS particularly well-suited for systematic analysis of absorber properties, defect density, and layer thickness optimization. However, SCAPS-1D cannot capture lateral inhomogeneities, grain-boundary effects, or complex interface geometries; these phenomena require multidimensional modeling approaches discussed in Section 3.

Numerical analysis provides insight into PV device operation and enables systematic optimization of parameters such as thickness, doping density, and defect concentration [36]. SCAPS-1D, developed at Ghent University [37], solves the Poisson and carrier continuity equations under steady-state conditions and is widely used for modeling thin-film solar cells [23,36,37,38].

This section addresses multidimensional device modeling using TCAD tools such as Silvaco Atlas and Synopsys Sentaurus. Unlike SCAPS-1D, TCAD enables 2D and 3D simulations of spatially resolved physical processes, including band-alignment variations, grain-boundary recombination, and module-level effects such as interconnect scribing. This makes TCAD essential for capturing complex device physics beyond the limitations of one-dimensional models.

Advanced Technology Computer-Aided Design (TCAD) tools, such as Silvaco Atlas and Synopsys Sentaurus, provide multi-dimensional insight into device physics that go beyond 1D modeling [1,16,21]. These simulators are essential for modeling conduction-band offsets (CBO), interface states, and spatial inhomogeneities [21,33]. The optimal CBO of +0.5 eV is found to be significant for heterojunction performance [1]. This optimal value corresponds to a spike-like conduction band alignment at the CIGS/CdS interface. This band alignment is illustrated schematically in Figure 4. A small positive (spike) CBO suppresses interface recombination by introducing a barrier for minority carriers (holes), while still allowing efficient electron transport across the junction. In contrast, a negative (cliff-like) CBO enhances interface recombination, leading to reduced open-circuit voltage and overall device performance.

Therefore, precise control of band alignment at the buffer/absorber interface is critical for minimizing recombination losses and maximizing device efficiency.

The band alignment illustrated in Figure 4 governs the primary recombination pathway at the buffer/absorber interface and forms the physical basis for CBO optimization studies reviewed throughout this section.

Multi-junction designs simulated in Silvaco TCAD demonstrate significant PCE enhancements: from 18.11% for single-junction CIGS to 21% for double CZO/CIGS, and further to 24.5% for triple CZO/CIGS/CZTS configurations (Table 3) [33]. This multi-layered approach employs the structures shown in Figure 5 to optimize light absorption and charge dynamics. In parallel, the combination of CZTSe and CIGS double absorbers can improve efficiency up to 29.22% while minimizing the cost of expensive indium [42].

2D simulations are critical for understanding the behavior of interconnected modules, particularly the impact of laser-patterning scribes (P1, P2, P3) [43]. The P1 scribe separates back electrodes and can introduce a shunt resembling a junction field-effect transistor, in which expanding the space charge region (SCR) into the P1 trench reduces shunt conductance. This behavior is detailed in Figure 6 [44].

The P1-P3 interconnect analysis illustrated in Figure 6 underscores that module-level efficiency losses are intrinsically linked to process precision and shunt geometry, motivating the need for fully calibrated device models. For high-efficiency record cells, such updated baseline models must incorporate post-deposition treatment (PDT) with alkali fluorides, which passivate grain boundaries (GBs) and increase the net free-carrier concentration [22]. The progression of these model updates is depicted in Figure 7, which shows how successive calibration steps — from a basic absorber model to one incorporating SDL, PDT, and electromagnetic optics — progressively close the gap between simulated and certified experimental performance [22]. The N1 signal observed in thermal admittance spectroscopy (TAS) of CIGS devices — historically attributed to deep bulk defects — has been reinterpreted through TCAD back-contact modeling as arising from a transport barrier at the rear Mo/CIGS interface [45].

In the realm of path-finding, co-optimization of composition gradients and buffer layers is critical. Simulations show that thin CdS (10 nm) is preferred for accumulated Ga-profiles, while thicker CdS enhances collection in flatter profiles [46]. Furthermore, the introduction of ITO as a front contact can lead to efficiencies of 23.074% [24].

Bifacial CIGS architectures require transparent rear contacts such as ITO, though these often present an electrical barrier that limits rear-side efficiency [47]. Novel designs like the PMoSe–P architecture-alternating lines of MoSe₂ and dielectric-allow for conversion efficiencies surpassing 20% and 15% for front and rear side illumination, respectively [47].

Comprehensive models must also reproduce thermal admittance spectra (TAS) and illumination-dependent I-V-T curves to be considered reliable [48]. As summarized in the following table, experimental and simulated data must align to validate the multi-physics approach.

Finally, the inclusion of time-resolved photoluminescence (TRPL) simulations, as applied to a 1D CIGSe absorber model with linear bandgap grading and shallow defects, enhances the assessment of material quality during production [49].

Despite significant progress in numerical modeling of CIGS solar cells, several limitations remain that constrain the predictive accuracy of simulation results. First, the polycrystalline nature of CIGS introduces complex grain-boundary (GB) effects that act as both recombination centers and carrier transport pathways. These effects cannot be accurately captured in one-dimensional SCAPS models and require computationally intensive multidimensional approaches.

Second, defect distributions in real devices are inherently non-uniform, whereas most simulations assume spatially homogeneous defect densities. This simplification can lead to discrepancies between simulated and experimental performance. Third, advanced TCAD simulations, particularly in 3D, require substantial computational resources, often limiting the feasibility of large-scale parametric studies.

Furthermore, current models generally neglect coupled thermal and mechanical stress effects, which can influence long-term device stability and degradation. Finally, accurate experimental validation remains challenging due to uncertainties in key input parameters, including defect densities, interface recombination velocities, and material properties. These limitations highlight the need for more physically realistic and experimentally validated modeling frameworks.

This study consolidates recent advances in the numerical modeling and structural optimization of CIGS thin-film solar cells. One-dimensional SCAPS-1D simulations effectively capture the influence of absorber thickness, doping density, defect states, and bandgap grading on key photovoltaic parameters. At the same time, multidimensional TCAD analyses provide deeper insights into band alignment, conduction band offsets, grain-boundary effects, and module-level phenomena.

Performance improvements are primarily driven by three coordinated strategies: (i) absorber engineering through compositional grading and defect control, (ii) interface optimization via buffer selection, conduction band alignment, and passivation treatments, and (iii) carrier-selective contact design including advanced window layers and BSF structures. These approaches collectively suppress recombination losses and enhance device performance.

Emerging architectures—including double absorber layers, tandem configurations, and bifacial designs—demonstrate the potential of CIGS technology to surpass conventional single-junction limits. However, efficiencies exceeding 30% are currently achieved only in numerical simulations under idealized conditions, while experimentally certified values remain around ~23.35%, highlighting a significant gap between theoretical predictions and practical realization.

Future progress will likely rely on integrating perovskite/CIGS tandem structures, bifacial thin-film architectures, and AI-assisted multi-parameter optimization frameworks. In parallel, advances in grain-boundary passivation, alkali-treatment strategies, and scalable fabrication of carrier-selective contacts will be essential for translating simulation insights into experimentally viable devices.

Key open challenges remain in accurately modeling defect physics, grain boundary behavior, and interface recombination, as well as in achieving reliable experimental validation of model parameters. Addressing these challenges is critical for bridging the gap between simulation and experiment and for enabling the next generation of high-efficiency, scalable CIGS solar technologies.

Overall, this review highlights that the primary bottleneck in CIGS performance is not optical absorption. Still, the control of recombination mechanisms at interfaces and defects underscores the central role of physically realistic modeling in future device design.