A Review on Perovskite/Silicon Tandem Solar Cells: Current Status and Future Challenges

Abstract

Perovskite/Si tandem solar cells (PSTSCs) have emerged as a leading candidate for surpassing the Shockley–Queisser (SQ) efficiency limit inherent to single-junction silicon solar cells. Following their inaugural demonstration in 2015, perovskite/Si tandem solar cells have experienced remarkable technological progression, reaching a certified power conversion efficiency of 34.9% by 2025. To elucidate pathways for realizing the full potential of perovskite/Si tandem solar cells, this review commences with an examination of fundamental operational mechanisms in multi-junction photovoltaic architectures. Subsequent sections systematically analyze technological breakthroughs across three critical PSTSC components organized by an optical path sequence: (1) innovations in perovskite photoactive layers through component engineering, additive optimization, and interfacial modification strategies; (2) developments in charge transport and recombination management via advanced interconnecting layers; and (3) silicon subcell architectures. The review concludes with a critical analysis of persistent challenges in device stability, scalability, structural optimization and fabrication method, proposing strategic research directions to accelerate the transition from laboratory-scale achievements to commercially viable photovoltaic solutions.

1. Introduction

To date, silicon solar cells, distinguished by their exceptional optoelectronic properties and backed by a sophisticated and mature supply chain with well-established manufacturing processes, have come to dominate the photovoltaics market, accounting for more than 90% of the total share [1,2,3,4]. Single-junction silicon solar cells have achieved a remarkable record power conversion efficiency (PCE) of 27.8% [5], progressively converging towards the theoretical SQ efficiency limit of 29.4% [6]. In the pursuit of overcoming this limit, the most promising strategy identified within the academic community is the development of solar cells featuring a tandem structure, which promises enhanced performance and efficiency.

Tandem solar cells (TSCs) achieve spectral efficiency through vertically stacked sub-cells featuring distinct bandgap energies, where incident photons propagate through the device in a sequential top-to-bottom manner. The configuration follows established optoelectronic principles: wide-bandgap top cells preferentially absorb high-energy photons from short-wavelength regions, while transmitted low-energy photons in long-wavelength regions undergo subsequent absorption by narrow-bandgap bottom sub-cells. This cascade absorption mechanism fundamentally reduces hot carrier thermalization losses, enabling TSCs to achieve near-optimal utilization of the solar spectrum through complementary photon harvesting across different spectral bands.

Perovskite solar cells (PSCs) have garnered substantial research interest in photovoltaic technology owing to their exceptional optoelectronic characteristics, solution-processable fabrication, and remarkable commercialization potential [7,8,9,10]. The PCE of state-of-the-art single-junction PSCs has achieved a certified value of 27.3% through advanced device engineering [5,11,12]. This superior performance originates from the unique ABX₃-type ionic crystal structure of perovskite semiconductors, where A-site monovalent cations (e.g., MA⁺ ([CH₃NH₃]⁺), FA⁺ ([CN(NH₂)₂]⁺), Cs⁺, Rb⁺), B-site divalent metal cations (Pb²⁺, Sn²⁺, Ge²⁺), and X-site halide anions (I⁻, Br⁻, Cl⁻, SCN⁻) form a three-dimensional coordination framework. Systematic compositional engineering enables precise bandgap tuning from 1.17 to 3.10 eV through ion substitution [8], making these materials particularly suitable as top-cell absorbers in TSC architectures. Lead-based perovskite compounds demonstrate strong visible light absorption up to 750 nm wavelength while maintaining spectral transparency in the near-infrared (NIR) region. This optical characteristic synergizes effectively with silicon’s superior NIR absorption capacity, enabling comprehensive solar spectrum utilization through PSTSC configurations. Since the initial demonstration of PSTSCs with 13.7% efficiency in 2015 [13], the certified PCE has progressively increased to 34.9% [5], surpassing the performance limits of single-junction perovskite or silicon photovoltaics. Notably, current achievements remain below the theoretical efficiency limit of 41% predicted for ideal tandem configurations [14], suggesting substantial opportunities for further performance enhancement through innovative material design, interfacial engineering, and device architecture optimization.

This comprehensive review critically examines the transformative potential of PSTSCs in advancing next-generation photovoltaic technology. The analysis commences with an in-depth examination of fundamental working mechanisms in tandem photovoltaic devices and architectural configurations specific to PSTSCs. The discussion systematically investigates three core system components: perovskite-based top cells, functional interconnection layers, and silicon bottom cells, incorporating recent progress in stoichiometric engineering, defect-passivation additives, and charge-transport interface optimization. Particular emphasis is placed on critical challenges hindering commercial deployment, including thermodynamic stability limitations under operational conditions, interfacial compatibility with textured silicon substrates, and manufacturing scalability constraints. Focused primarily on two-terminal (2T) monolithic configurations, this review systematically examines contemporary research advancements in tandem photovoltaics, with the aim of proposing evidence-based strategies to achieve the theoretical performance ceilings of PSTSCs within practical deployment contexts. Through critical analysis of current technological paradigms, we delineate pathways for bridging the persistent gap between laboratory-scale achievements and industrial implementation requirements.

2. Fundamental Mechanisms of the Tandem Photovoltaic Device

The working mechanism of perovskite solar cells is shown in Figure 1a, including charge generation, separation, transport and collection. As direct-bandgap semiconductors, perovskite materials demonstrate photon-induced electron excitation from valence to conduction bands, generating excitons whose binding energies critically determine charge separation efficiency [15]. When the incident photon energy reaches or exceeds the bandgap width of the perovskite layer, it triggers the process of a valence-band electron-directed conduction-band jump, resulting in the formation of electron-hole pairs in the bound state. Thanks to the characteristic low exciton binding energy of perovskite, most of the excitons can be efficiently dissociated into free carriers, thus guaranteeing the effective separation and transport of photogenerated electrons and holes [16]. Then, under the setting of specific layers and interfaces, the dissociated free electrons and holes migrate directionally to the corresponding transport layer, respectively. Eventually, the photogenerated carriers migrate directionally through the transport layer to the positive and negative electrode regions, and the electrons flow through the external circuit, generating an electric current and completing the circuit.

Silicon solar cells work similarly to perovskite solar cells, characterized by silicon photovoltaic devices using precisely doped p-n junctions and a built-in electric field to achieve charge separation through a drift mechanism. Modern silicon solar cell technologies are systematically classified into four architectural paradigms based on passivation and contact configurations: back surface field (BSF), passivated-emitter rear-cell (PERC), tunnel oxide passivating contact (TOPCon), and silicon heterojunction (SHJ) designs. In addition to the above, there are also some integrated variants like TOPerc, which implements the overall approach of the TOPCon and PERC architectures [17]. The silicon bottom cells will be comprehensively analyzed in subsequent sections.

The first PSTSCs were experimentally validated by Mailoa et al. in 2015 [13]. This tandem architecture utilizes wide bandgap perovskite solar cells (WBG-PSCs) as the top subcell for high-energy photon absorption (λ < 800 nm), while narrow bandgap silicon subcells harvest low-energy photons (λ < 1100 nm), achieving enhanced solar spectrum utilization through complementary photonic management (Figure 1b). According to the stacking method of cells, PSTSCs predominantly manifest three structural configurations: 2T series-connected monolithic devices, four-terminal (4T) electrically isolated systems, and three-terminal (3T) hybrid architectures with independent current extraction capabili-ties (Figure 1c) [18]. The 4T configuration enables independent electrical operation and optimization of subcells without current-matching constraints, thereby relaxing bandgap selection limitations and enhancing spectral stability. Nevertheless, this architecture requires multiple electrodes that induce parasitic absorption and increase material complexity, ultimately elevating the levelized cost of electricity (LCOE) [19,20]. The emerging 3T architecture combines reduced electrode complexity with current-matching elimination, though its fabrication challenges have hindered technological progression [21].

Conversely, the 2T monolithic configuration employs an interconnecting layer (ICL) for series integration, eliminating supplementary substrates and transparent electrodes while minimizing parasitic losses. This integration simplifies material requirements and reduces manufacturing costs [22,23]. A critical design constraint in 2T devices involves maintaining current matching, which confines top-cell bandgaps to 1.65–1.75 eV and increases spectral sensitivity. The structure of monolithic PSTSCs is similar to that of conventional PSCs, which are further categorized into n-i-p and p-i-n structures depending on the display order of the transport layer (Figure 1d). Theoretical calculations predict maximum efficiencies approaching 46% for 2T configurations, surpassing single-junction photovoltaic limitations through synergistic spectrum utilization (Figure 1e). Among these architectures, monolithic 2T perovskite/Si tandem cells have emerged as particularly promising for practical photovoltaic applications due to their balance of efficiency potential and manufacturability (Table 1). This review focuses specifically on the technical developments and challenges associated with 2T perovskite/Si tandem configurations.

Figure 1. (a) Working mechanism of PSCs [24]. (b) Schematic diagram of the working principle of PSTSCs [25]. (c) The device configuration of 4-T and 3-T PSTSCs. (d) The n-i-p and p-i-n 2-T PSTSCs. (e) Theoretical efficiency limit for 2T tandems [26].

Figure 1. (a) Working mechanism of PSCs [24]. (b) Schematic diagram of the working principle of PSTSCs [25]. (c) The device configuration of 4-T and 3-T PSTSCs. (d) The n-i-p and p-i-n 2-T PSTSCs. (e) Theoretical efficiency limit for 2T tandems [26].

Table 1. Summary of two-terminal perovskite/silicon tandem solar cells.

Year	Device Structure	Absorber	Si-Type	Eg (eV)	VOC (V)	JSC (mAcm−2)	PCE (%)	Ref
2016	n-i-p	FAxMA1−xPbyI3−yBr	SHJ	none	1.78	14	18	[27]
2016	n-i-p	MAPbI3	SHJ	1.6	1.703	15.8	21.4	[28]
2017	p-i-n	Cs0.17FA0.83Pb(Br0.17I0.83)3 (CsFA)	SHJ	1.63	1.65	18.1	23.6	[29]
2018	n-i-p	MAPbI3	TOPcon	1.6	1.68	16.2	21	[30]
2018	n-i-p	Cs0.05Rb0.05FA0.765MA0.135PbI2.55Br0.45	SHJ	1.63	1.763	17.8	24.5	[31]
2018	p-i-n	CsxFA1−xPb(Br1−yIy)	SHJ	1.6	1.78	19.5	25.2	[32]
2019	p-i-n	(FAPbI3)0.8(MAPbBr3)0.2	BSF	1.64	1.645	16.13	21.19	[33]
2020	p-i-n	Cs0.1MA0.9Pb(I0.9Br0.1)3	SHJ	none	1.82	19.2	26	[34]
2020	p-i-n	Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3	SHJ	1.68	1.92	19.26	29.15	[9]
2020	p-i-n	Peroviskite	SHJ	1.67	1.886	19.12	27.04	[35]
2020	p-i-n	Cs0.05MA0.15FA0.8PbBr0.75I2.25	SHJ	1.68	1.7	19.8	25.7	[36]
2021	p-i-n	Cs0.05(MA0.23FA0.77)0.95Pb(Br0.23I0.77)3	SHJ	1.68	1.89	19.13	28.2	[37]
2021	p-i-n	Cs0.1FA0.9PbBr0.13I2.8	SHJ	none	1.808	19.79	27.48	[38]
2022	p-i-n	Cs0.22FA0.78Pb(Br0.15I0.85)3	TOPcon	1.68	1.794	19.68	27.6	[39]
2022	p-i-n	Cs0.05(FA0.79MA0.21)0.95Pb(I0.79Br0.21)3	SHJ	1.66	1.9	19.45	29.83	[40]
2022	p-i-n	Cs0.1FA0.1MA0.7Pb(Br0.15I0.85)3	SHJ	1.65	1.92	18.95	28.6	[41]
2022	n-i-p	CsPb(IxBr1−x)3	SHJ	1.8	2.04	14.34	22.95	[42]
2022	p-i-n	Cs0.05FA0.8MA0.15Pb(Br0.255I0.755)3	SHJ	1.69	1.92	19.8	29.3	[43]
2022	n-i-p	Cs0.05FA0.8MA0.15PbI2.5Br0.45	TOPcon	1.62	1.783	14.4	17.3	[44]
2022	p-i-n	Cs0.22FA0.78Pb(Cl0.03Br0.15I0.85)3	TOPcon	1.68	1.79	19.68	27.63	[45]
2022	p-i-n	FA0.78Cs0.22Pb(I0.85Br0.15)3	PERC	1.68	1.91	19.29	28.81	[46]
2022	p-i-n	Cs0.05(FA0.83MA0.17)0.95Pb(Br0.17I0.83)3	TOPcon	1.63	1.8	19.2	28.2	[47]
2023	p-i-n	CsPbBr0.15I2.85	SHJ	1.71	1.84	17.95	25.31	[48]
2023	p-i-n	Cs0.22FA0.78Pb(I0.85Br0.15)3 + 5%MAPbI3	SHJ	1.68	2	20.24	32.5	[49]
2023	p-i-n	Peroviskite	SHJ	none	1.82	20.4	28.4	[50]
2023	n-i-p	(FAPbI3)0.83(MAPbBr3)0.17	SHJ	none	1.82	18.1	27.2	[51]
2023	p-i-n	Cs0.05(FA0.83MA0.17)0.95Pb(I0.83Br0.17)3	TOPcon	none	1.83	19.7	29.2	[52]
2023	p-i-n	CsxFA1−xPbI3−yBry	SHJ	none	1.68	19.47	23.22	[53]
2023	p-i-n	Cs0.05MA0.14FA0.81Pb(I0.8Br0.2)3	SHJ	1.68	1.85	19.7	28.4	[54]
2024	p-i-n	Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3	SHJ	1.67	1.941	20.1	29.6	[55]
2024	p-i-n	Peroviskite	SHJ	1.69	1.98	20.86	33.89	[56]
2024	p-i-n	Cs0.05FA0.95−xMAxPb(I1−yBry)3	SHJ	1.67	1.869	20.32	30.52	[57]
2024	n-i-p	Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3	TOPCon	none	1.9	18.82	28.2	[58]
2024	p-i-n	Cs0.05(FA0.9MA0.1)0.95Pb(I0.8Br0.2)3	SHJ	none	1.954	18.92	30.22	[59]
2025	p-i-n	Perovskite	TOPcon	1.68	1.931	19.89	31.32	[60]
2025	p-i-n	Cs0.17FA0.83Pb(I0.83Br0.17)3	TOPCon	1.68	1.977	19.41	30.36	[61]
2025	p-i-n	Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3	SHJ	1.68	1.92	20.3	31.5	[62]
2025	p-i-n	perovskite	SHJ	1.68	1.912	19.92	32.19	[63]

Table 1. Summary of two-terminal perovskite/silicon tandem solar cells.

Year	Device Structure	Absorber	Si-Type	Eg (eV)	VOC (V)	JSC (mAcm−2)	PCE (%)	Ref
2016	n-i-p	FAxMA1−xPbyI3−yBr	SHJ	none	1.78	14	18	[27]
2016	n-i-p	MAPbI3	SHJ	1.6	1.703	15.8	21.4	[28]
2017	p-i-n	Cs0.17FA0.83Pb(Br0.17I0.83)3 (CsFA)	SHJ	1.63	1.65	18.1	23.6	[29]
2018	n-i-p	MAPbI3	TOPcon	1.6	1.68	16.2	21	[30]
2018	n-i-p	Cs0.05Rb0.05FA0.765MA0.135PbI2.55Br0.45	SHJ	1.63	1.763	17.8	24.5	[31]
2018	p-i-n	CsxFA1−xPb(Br1−yIy)	SHJ	1.6	1.78	19.5	25.2	[32]
2019	p-i-n	(FAPbI3)0.8(MAPbBr3)0.2	BSF	1.64	1.645	16.13	21.19	[33]
2020	p-i-n	Cs0.1MA0.9Pb(I0.9Br0.1)3	SHJ	none	1.82	19.2	26	[34]
2020	p-i-n	Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3	SHJ	1.68	1.92	19.26	29.15	[9]
2020	p-i-n	Peroviskite	SHJ	1.67	1.886	19.12	27.04	[35]
2020	p-i-n	Cs0.05MA0.15FA0.8PbBr0.75I2.25	SHJ	1.68	1.7	19.8	25.7	[36]
2021	p-i-n	Cs0.05(MA0.23FA0.77)0.95Pb(Br0.23I0.77)3	SHJ	1.68	1.89	19.13	28.2	[37]
2021	p-i-n	Cs0.1FA0.9PbBr0.13I2.8	SHJ	none	1.808	19.79	27.48	[38]
2022	p-i-n	Cs0.22FA0.78Pb(Br0.15I0.85)3	TOPcon	1.68	1.794	19.68	27.6	[39]
2022	p-i-n	Cs0.05(FA0.79MA0.21)0.95Pb(I0.79Br0.21)3	SHJ	1.66	1.9	19.45	29.83	[40]
2022	p-i-n	Cs0.1FA0.1MA0.7Pb(Br0.15I0.85)3	SHJ	1.65	1.92	18.95	28.6	[41]
2022	n-i-p	CsPb(IxBr1−x)3	SHJ	1.8	2.04	14.34	22.95	[42]
2022	p-i-n	Cs0.05FA0.8MA0.15Pb(Br0.255I0.755)3	SHJ	1.69	1.92	19.8	29.3	[43]
2022	n-i-p	Cs0.05FA0.8MA0.15PbI2.5Br0.45	TOPcon	1.62	1.783	14.4	17.3	[44]
2022	p-i-n	Cs0.22FA0.78Pb(Cl0.03Br0.15I0.85)3	TOPcon	1.68	1.79	19.68	27.63	[45]
2022	p-i-n	FA0.78Cs0.22Pb(I0.85Br0.15)3	PERC	1.68	1.91	19.29	28.81	[46]
2022	p-i-n	Cs0.05(FA0.83MA0.17)0.95Pb(Br0.17I0.83)3	TOPcon	1.63	1.8	19.2	28.2	[47]
2023	p-i-n	CsPbBr0.15I2.85	SHJ	1.71	1.84	17.95	25.31	[48]
2023	p-i-n	Cs0.22FA0.78Pb(I0.85Br0.15)3 + 5%MAPbI3	SHJ	1.68	2	20.24	32.5	[49]
2023	p-i-n	Peroviskite	SHJ	none	1.82	20.4	28.4	[50]
2023	n-i-p	(FAPbI3)0.83(MAPbBr3)0.17	SHJ	none	1.82	18.1	27.2	[51]
2023	p-i-n	Cs0.05(FA0.83MA0.17)0.95Pb(I0.83Br0.17)3	TOPcon	none	1.83	19.7	29.2	[52]
2023	p-i-n	CsxFA1−xPbI3−yBry	SHJ	none	1.68	19.47	23.22	[53]
2023	p-i-n	Cs0.05MA0.14FA0.81Pb(I0.8Br0.2)3	SHJ	1.68	1.85	19.7	28.4	[54]
2024	p-i-n	Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3	SHJ	1.67	1.941	20.1	29.6	[55]
2024	p-i-n	Peroviskite	SHJ	1.69	1.98	20.86	33.89	[56]
2024	p-i-n	Cs0.05FA0.95−xMAxPb(I1−yBry)3	SHJ	1.67	1.869	20.32	30.52	[57]
2024	n-i-p	Cs0.05FA0.8MA0.15Pb(I0.75Br0.25)3	TOPCon	none	1.9	18.82	28.2	[58]
2024	p-i-n	Cs0.05(FA0.9MA0.1)0.95Pb(I0.8Br0.2)3	SHJ	none	1.954	18.92	30.22	[59]
2025	p-i-n	Perovskite	TOPcon	1.68	1.931	19.89	31.32	[60]
2025	p-i-n	Cs0.17FA0.83Pb(I0.83Br0.17)3	TOPCon	1.68	1.977	19.41	30.36	[61]
2025	p-i-n	Cs0.05(FA0.77MA0.23)0.95Pb(I0.77Br0.23)3	SHJ	1.68	1.92	20.3	31.5	[62]
2025	p-i-n	perovskite	SHJ	1.68	1.912	19.92	32.19	[63]

3. Present Status of Monolithic Perovskite/Silicon TSCs

Having established a fundamental comprehension of perovskite/silicon tandem solar cell architectures, this section systematically reviews recent advancements in perovskite/silicon TSCs research. Our analysis follows a top-down structural framework, specifically examining three critical components: the perovskite top cell, interconnection layer, and silicon bottom cell.

3.1. Perovskite Top Cell

Compositional engineering. Although the initial demonstration of PSTSCs utilized single-cation MAPbI₃ [13], its suboptimal bandgap (approximately 1.56 eV) restricts optimal spectral complementarity with silicon bottom cells. As illustrated in Figure 1e [26], the theoretical efficiency limits for tandem architectures suggest that a 1.73 eV WBG perovskite top cell paired with a 1.1 eV silicon bottom cell represents a near-ideal configuration for maximizing PCE. The bandgap tunability and defect formation mechanisms in perovskite materials are fundamentally governed by the coordinated interactions among A-site cations, B-site metal ions, and X-site halide anions, which collectively determine structural stability and electronic properties. Consequently, compositional engineering serves as a strategic approach to precisely tailor the perovskite absorber’s bandgap while simultaneously enhancing the optoelectronic quality of thin-film layers through defect passivation and crystallization optimization.

Notably, dual-cation perovskite compositions have demonstrated superior optoelectronic stability compared to single-cation counterparts, driving accelerated research progress in this domain. In 2015, McMeekin et al. pioneered cation engineering by substituting a fraction of FA with Cs in perovskite films [64]. The optimized FA_0.83Cs_0.17Pb(I_0.6Br_0.4)₃ composition exhibited remarkable operational stability under continuous illumination while maintaining spectral fidelity, as evidenced by the absence of photoluminescence peak shifts. This device achieved a PCE of 17.1% with a 1.74 eV bandgap, though WBG-PSCs still suffered from substantial open-circuit voltage (V_OC) losses relative to narrow-bandgap devices [65]. Subsequent investigations systematically refined cation engineering strategies. Bush et al. (2017) strategically optimized A-site composition through Cs⁺ enrichment rather than excessive X-site bromination, effectively mitigating phase segregation while enhancing V_OC, ultimately achieving a 23.6% PCE in PTSCs [29]. Chen et al. (2020) advanced this approach by integrating Cs_0.1MA_0.9Pb(I_0.9Br_0.1)₃ PSCs with textured silicon substrates, attaining 26% tandem efficiency [34]. Recent progress by Wu et al. (2022) demonstrated 27.6% efficiency through precise I/Br ratio optimization in Cs_0.22FA_0.78Pb(Br_0.15I_0.85)₃ absorbers combined with TOPCon bottom cells [39].

Triple-cation perovskites generally surpass dual-cation systems in film quality, exhibiting enhanced compactness and reduced pinhole density, making them predominant in high-efficiency PSTSCs. Al-Ashouri et al. [9] achieved a landmark PCE of 29.15% using a Cs_0.05(FA_0.77MA_0.23)_0.95Pb(I_0.77Br_0.23)₃ absorber (1.68 eV), demonstrating exceptional interfacial hole extraction and suppressed non-radiative recombination that enabled a record V_OC of 1.92 V. Remarkably, unencapsulated devices retained 95% initial efficiency after 300-h ambient operation. Building upon this foundation, Tockhorn et al. optimized the composition to Cs_0.05(FA_0.79MA_0.21)_0.95Pb(I_0.79Br_0.21)₃ and implemented dielectric-buffered reflectors for improved current matching, culminating in a certified PCE of 29.8% through enhanced spectral utilization efficiency [40].

Additive engineering. The WBG perovskites in the reported PSTSCs typically exhibit bromine content exceeding 20%, which significantly accelerates crystallization kinetics, induces undesirable phase segregation, and substantially constrains the enhancement of V_OC alongside compromising photostability [66,67,68]. To mitigate these challenges, researchers commonly incorporate one or more functional additives into perovskite precursor solutions. These additives function through distinct physicochemical mechanisms to precisely regulate crystallization processes and passivate interfacial defects, thereby simultaneously improving both the PCE and operational stability of PSCs.

In organic–inorganic hybrid WBG-PSCs, uncoordinated organic cations (MA⁺, FA⁺) and Pb²⁺ have been established as predominant defect formation sites, with halogen anion vacancies exacerbating photoinduced phase segregation. Initial research efforts established that Pb(SCN)₂ additives significantly enhance perovskite grain growth kinetics [69], while MA⁺ cation incorporation in FACs-based WBG perovskite precursors effectively mitigates deep-level trap states [70]. Subsequent studies revealed chloride anions’ capacity to improve crystalline quality through grain boundary defect passivation mechanisms [71]. Yang et al. [41] achieved a breakthrough in 2022 through blade-coated deposition of 1-µm-thick WBG perovskite layers using tetrabutylammonium tribromide (TBABr₃) additives, which suppressed iodide interstitial formation and reduced charge recombination losses. This methodology yielded a certified PCE of 21.9% for WBG-PSCs and 28.6% for PSTSCs. Recent advancements in 2023 demonstrated that cadmium acetate (CdAc₂) incorporation in PbI₂ precursors enhances organic salt diffusion dynamics, thereby optimizing crystal orientation and suppressing non-radiative recombination [72].

Certain small organic molecules with tailored functional groups have shown exceptional capability in passivating deep-level charged defects at grain boundaries. Isikgor et al. [73] pioneered in 2021 the integration of phenformin hydrochloride (PhenHCl, Figure 2a) into precursor formulations, exploiting its amphiphilic structure containing electronegative amine/imine groups and electropositive ammonium moieties for dual passivation of interfacial and bulk defects. Concurrently, Wang et al. [60] developed a multifunctional ammonium salt (thioacetylacetamide hydrochloride, TAACl) that modulates perovskite nucleation and crystallization processes through multivalent interactions with precursor constituents (Figure 2b). The optimized TAACl-treated devices demonstrated a stabilized PCE of 31.32% in tandem configurations with TOPCon silicon subcells, retaining 95.43% initial efficiency following a 1000 h continuous operation under a nitrogen atmosphere.

Yang’s research group further advanced this field in 2023 through the introduction of 2-aminoethanesulfonamide hydrochloride (AESCl) [61], a molecular additive featuring multipoint chelation sites and bridging functionalities. This compound effectively suppressed light-induced halide segregation and non-radiative recombination through strong coordination with undercoordinated Pb²⁺ and halide ions. The resultant AESCl-treated perovskite/TOPCon tandem devices achieved a champion PCE of 30.36% (Figure 2c), with encapsulated modules maintaining 96% initial efficiency after 1068 h of continuous AM 1.5G illumination. Beyond single-component additives, recent developments in multi-additive synergistic systems have demonstrated enhanced control over crystallization dynamics and substantial improvements in WBG perovskite film quality through complementary defect-passivation mechanisms [74,75].

An alternative strategy to mitigate halide segregation involves the implementation of all-inorganic perovskite. All-inorganic perovskites exemplified by CsPb(I_1−xBr_x)₃ demonstrate the capacity to achieve a 1.7 eV bandgap with bromine incorporation below a 20% stoichiometric ratio, contrasting with organic cation-containing counterparts requiring 25–30% bromine content for equivalent bandgap attainment [76]. Nevertheless, the photovoltaic performance of p-i-n PSTSCs has been constrained by elevated defect densities at the surface of all-inorganic perovskite films and the inherent thermodynamic instability of black α-phase inorganic perovskites, which undergo spontaneous phase transition to non-photoactive yellow δ-phase under ambient conditions. Recent advancements by Wang et al. (2022) demonstrated NiI₂-mediated surface passivation of 1.80 eV all-inorganic perovskite (Figure 2d), which optimized band alignment with hole transport layers and resulted in single-junction WBG-PSCs exhibiting 19.53% PCE with a V_OC of 1.36 V [42]. Subsequent integration with rear-textured silicon bottom cells enhanced device performance to 22.95% PCE and 2.04 V V_OC. Building upon this foundation, Wang et al. (2023) implemented surface reconfiguration of CsPbI_2.85Br_0.15 films through 2-amino-5-bromobenzamide (ABA) molecular modification [48], establishing a breakthrough in p-i-n monolithic inorganic PSTSCs with certified PCE reaching 25.31%, representing the highest reported efficiency for this architecture to date (Figure 2e).

Interface engineering. Beyond perovskite precursor engineering, interfacial engineering emerges as a critical determinant in device optimization. Multifunctional interface modifications facilitate synergistic defect passivation, precise energy band alignment, and enhanced photon management, thereby comprehensively mitigating carrier recombination losses and addressing charge transport limitations inherent in tandem architectures.

Nonradiative recombination and carrier transport losses at the perovskite/upper interface (e.g., C₆₀) significantly constrain further efficiency improvements. Stable inorganic compounds such as MgF_x and LiF have demonstrated efficacy in enhancing interfacial contact between perovskite and charge transport layers. In 2022, Liu et al. implemented a 1 nm MgF_x interlayer at the perovskite/C₆₀ interface, effectively modifying surface energy levels and suppressing nonradiative recombination (Figure 3a), thereby achieving a certified PCE of 29.3% for 1 cm² devices with improved operational stability [43]. Subsequent advances in molecular engineering have yielded notable progress. Mariotti and coworkers (2023) employed piperazinium iodide modifiers to optimize band alignment and charge extraction processes, fabricating PSTSCs with a certified PCE of 32.5% [49] (Figure 3b). Concurrent research demonstrated that ethylenediamine diiodide (EDAI₂)/4-fluoro-phenethylammonium chloride (4F-PEACl) bilayer modification effectively reduced conduction band offsets at perovskite/C₆₀ interfaces, enabling WBG devices (1.67 eV) to achieve 21.8% PCE with a remarkable V_OC of 1.262 V [55].

Recent developments in interfacial engineering have further advanced device performance. Liu and colleagues developed a dual-passivation strategy incorporating nanoscale lithium fluoride layers followed by diammonium diiodide molecules, achieving an independently certified stable PCE of 33.89% for PSTSCs [56] (Figure 3c). Pei et al. implemented reactive passivation through lead halide interaction, forming low-dimensional phases that simultaneously suppressed interfacial recombination and enhanced carrier transport [57]. This approach enabled WBG perovskite devices with polymer hole transport layers to attain a high V_OC of 1.25 V and certified stable PCE of 30.52% for monolithic PSTSCs (Figure 3d). Ye’s research group systematically engineered halide-substituted piperazine salts for surface modification, revealing that piperazine chloride induces asymmetric bidirectional ion redistribution [62]. This process concentrates large piperazine cations at the perovskite surface while facilitating downward migration of chloride ions to buried interfaces, achieving concurrent bi-interfacial defect passivation and energy band modulation (Figure 3e). The optimized PCl-treated single-junction devices demonstrated a champion PCE of 22.3% with an unprecedented V_OC × FF product reaching 84.4% of the SQ limit.

The accumulation of microscopic defects and component inhomogeneities at buried interfaces exerts substantial influence on perovskite film growth dynamics. Nevertheless, significant challenges persist in fabricating solution-processed heterostructural configurations due to inherent solvent and textural incompatibilities. Recent advancements (2023) demonstrate an innovative sequential deposition methodology combining solution-processed 4-vinylbenzylammonium iodide (VBAI) with vacuum-deposited PbI₂, enabling the formation of a coherent two-dimensional perovskite interfacial layer beneath three-dimensional perovskite frameworks [50]. This engineered interphase simultaneously induces preferential crystallographic orientation, suppresses defect-mediated interfacial degradation, and enhances charge carrier transport efficiency. The optimized architecture culminated in a certified PCE of 28.4% for 1.0 cm² devices, demonstrating exceptional operational stability with 89% initial PCE retention after 1000 h of continuous illumination exposure.

Overall, for the WBG perovskite active layer in PSTSCs, the primary focus is on overcoming challenges related to phase separation and defect density through precursor engineering and interface engineering. Further research should focus on reducing the impact of these factors, refining the treatment strategies for perovskite films and developing new materials with suitable energy levels and passivation functions, which aim to simultaneously improve the stability and performance of PSTSCs.

3.2. Interconnecting Layer

The interconnecting layer serves dual critical functions in 2T PSTSCs, enabling both optical transmission and electrical interconnection between the perovskite top cell and silicon bottom cell. Material selection for ICL components must satisfy two fundamental criteria [77]: (1) Electrical connection to ensure unimpeded charge carrier extraction and transport efficiency while maintaining ideal V_OC and FF characteristics; (2) optimal optical properties to minimize interfacial reflection losses and suppress both external and internal parasitic absorption. Current research primarily focuses on three primary interconnection strategies: transparent conductive oxide (TCO) recombination layers, silicon-based tunnel junctions, and direct interlayer-free contacts. This section comprehensively examines structural optimization methods for each ICL component layer, emphasizing the correlation between interfacial engineering approaches and the systematic optimization of photovoltaic parameters through sophisticated engineering of layer architectures and interfacial characteristics.

TCO Recombination Layer. TCO materials such as indium-doped tin oxide (ITO), indium-doped zinc oxide (IZO), and zinc tin oxide (ZTO) have been demonstrated as highly effective recombination layers in PSTSCs, owing to their exceptional electrical conductivity and optimal optical transmittance characteristics [9,35,74].

TCO has become an essential component in SHJ solar cells, having been optimized for both low contact resistivity and non-damaging processing characteristics. Albrecht et al. [27] pioneered the implementation of an ITO recombination layer in perovskite/SHJ TSCs, achieving a notable V_OC of 1.78 V and a stabilized PCE of 18.1% (Figure 4a). Subsequent advancements by Chen et al. [74] demonstrated PSTSCs utilizing ITO recombination layers that attained a V_OC of 1.8 V with a stabilized PCE of 25.4%. Nevertheless, ITO exhibits significant limitations in near-infrared applications due to pronounced parasitic absorption caused by free carrier absorption. This necessitates either thickness reduction in ITO ICL or adoption of alternative TCO materials with superior carrier mobility and enhanced infrared transparency.

Recent developments have focused on ultrathin TCO configurations to mitigate these limitations. A 1.7 nm ultra-thin ITO layer was successfully implemented as ICL, effectively minimizing parasitic resistance while maintaining efficient trap states for carrier recombination [51]. Comparative studies have identified IZO as a promising alternative, exhibiting higher carrier mobility and requiring lower thermal budgets compared to conventional ITO, with demonstrated applications in both SHJ and silicon homojunction cells [28,78]. Werner et al. [79] further advanced this field through the implementation of ZTO ICL in mesoporous perovskite/homojunction silicon tandem cells, achieving remarkable thermal stability even under high-temperature annealing conditions (Figure 4b).

The lateral conductivity of TCO-based ICL presents additional challenges in device architecture, particularly regarding potential shunting pathways in perovskite top cells when deposited on textured silicon substrates. This issue is compounded by the difficulty in achieving energetically homogeneous self-assembled monolayer (SAM) coverage on textured surfaces. Aydin et al. [78] recently addressed these challenges through innovative use of 5 nm IZO ICL, which demonstrates superior surface-potential homogeneity due to increased SAM anchoring site density and absence of crystalline grain boundaries. This breakthrough enabled the realization of a certified PCE reaching 32.5% (Figure 4c), representing a significant milestone in TSCs’ development.

Interlayer-Free Contact. This strategy circumvents the inherent complexities and performance limitations associated with traditional ICL (e.g., ITO or IZO) through direct integration of adjacent carrier transport layers. In 2018, Zheng’s research team pioneered a streamlined fabrication method involving spin-coated SnO₂ deposition directly onto the P⁺⁺-doped emitter region of homojunction silicon subcells, thereby obviating the requirement for supplementary interfacial modifications [30] (Figure 4d). Subsequent investigations by Shen et al. [31] demonstrated an innovative interlayer-free configuration for perovskite/TOPerc tandem devices through atomic-layer-deposited TiO_x layers directly fabricated on polycrystalline silicon (p⁺) substrates (Figure 4e). Mechanistic analysis revealed that intragap electronic states play a critical role in mitigating energy level mismatches between adjacent carrier transport layers within the tandem architecture. This interlayer-free paradigm was further substantiated through implementation of chemical bath-deposited SnO_x coatings on p⁺-type silicon emitters, yielding comparable performance enhancements in perovskite/TOPerc tandem photovoltaic systems [44,80].

Figure 4. (a) Schematic of a 2T perovskite/SHJ with an ITO interconnection layer [27]. (b) The transmittance and absorptance of a 150 nm thick ZTO layer on glass before and after annealing at 500 °C in air [79]. (c) KPFM surface-potential mapping of 20-nm-thick TCO ICL before and after 2PACz coverage [78]. (d) Schematic device design of interface-layer-free perovskite/silicon homojunction solar cells [30]. (e) Simulated band diagram of the TiO_x/p⁺-Si at equilibrium [31].

Figure 4. (a) Schematic of a 2T perovskite/SHJ with an ITO interconnection layer [27]. (b) The transmittance and absorptance of a 150 nm thick ZTO layer on glass before and after annealing at 500 °C in air [79]. (c) KPFM surface-potential mapping of 20-nm-thick TCO ICL before and after 2PACz coverage [78]. (d) Schematic device design of interface-layer-free perovskite/silicon homojunction solar cells [30]. (e) Simulated band diagram of the TiO_x/p⁺-Si at equilibrium [31].

Silicon-Based Tunnel Junction. Silicon-based tunnel junctions are primarily categorized into two distinct configurations: (1) homogeneous silicon structures and (2) heterointerface systems combining silicon with metal oxides (SnO₂ or TiO₂). Within perovskite/silicon architectures, n⁺⁺/p⁺⁺ silicon-based tunneling junctions have demonstrated exceptional performance as interconnection layers. The advancement of tunnel junction technologies has substantially propelled the development of PSTSCs. Mailoa et al. [13] pioneered this field in 2015 by implementing n⁺⁺-doped hydrogenated amorphous silicon (n⁺⁺-a-Si:H) as a tunnel recombination layer on p⁺⁺-type crystalline silicon emitters (Figure 5a), achieving a breakthrough PCE of 13.7%. Comprehensive characterization verified that the p⁺⁺/n⁺⁺-a-Si:H heterostructure facilitated efficient electron-hole recombination, thereby establishing the foundational architecture for tunnel recombination-based tandem devices. However, the requisite high-temperature annealing process (680 °C) imposed significant limitations for compatibility with SHJ bottom cell technologies. Subsequent research efforts focused on low-temperature fabrication methodologies. Sahli et al. [32] demonstrated a plasma-enhanced chemical vapor deposition (PECVD) technique operating below 200 °C to synthesize hydrogenated nanocrystalline silicon (nc-Si:H) tunnel layers for perovskite/SHJ configurations (Figure 5b). The engineered nc-Si:H structure exhibited densely packed grain boundaries that effectively suppressed lateral carrier transport, thereby minimizing leakage currents and shunt effects in the upper subcell. This innovation yielded a certified PCE enhancement to 25.2%. Photonic management at interconnection interfaces remains critical for tandem performance optimization. Alternative tunnel junction configurations incorporating silicon with interfacial materials, such as SnO_x/p⁺⁺ silicon heterostructures, have shown comparable efficacy. In such architectures, the SnO_x interlayer serves dual functionalities: (1) as an ETL for the perovskite top subcell, and (2) forming low-resistance recombination contacts with p⁺⁺ silicon. Mercaldo et al. [81] advanced this concept through a multilayered nc-SiO_x stack with alternating refractive indices between subcells (Figure 5c). This configuration achieved spectrally selective photon management, demonstrating wavelength-dependent reflection (500–750 nm) and transmission (750–850 nm) characteristics. The optimized photon redistribution strategy enhanced short-circuit current density (J_SC) through a significant reduction in parasitic absorption losses. Emerging material systems for tunnel junction fabrication include polycrystalline silicon (poly-Si), particularly advantageous in TOPCon-based architectures [82]. Zheng et al. [52] recently achieved a breakthrough in poly-Si (p⁺⁺/n⁺⁺) tunnel junction quality through dopant compensation mitigation and interdiffusion suppression. Complementary finite element simulations provided mechanistic insights into carrier transport and quantum tunneling phenomena within the poly-Si-based junction (Figure 5d). The resultant perovskite/TOPCon TSCs demonstrated exceptional performance metrics, achieving a maximum PCE of 29.2% (independently certified at 28.76%), attributable to enhanced carrier transport dynamics and superior surface passivation.

In summary, the ICL functions as the critical interfacial component that electrically couples the top and bottom subcells in PSTSCs. Consequently, establishing and maintaining optimal ohmic contact between the ICL and adjacent active layers becomes paramount for effectively suppressing non-radiative recombination at the heterojunction interface. To address parasitic absorption phenomena within the ICL architecture, systematic optimization strategies should focus on two complementary approaches: precision engineering of layer thickness through advanced deposition techniques and developing alternative organic or inorganic materials characterized by minimal parasitic absorption coefficients while maintaining requisite electrical properties.

3.3. Si Bottom Cell

Based on the doping elements in silicon films, crystalline silicon solar cells can be classified into two primary categories: p-type silicon cells (boron-doped) and n-type silicon cells (phosphorus-doped). The specific contact architecture further determines cell typology, with p-type configurations encompassing BSF and PERC designs, while n-type variants include SHJ and TOPCon technologies. PERC technology constitutes the cornerstone of contemporary mass production, whereas TOPCon and SHJ emerge as predominant technical pathways demonstrating significant potential for advancing silicon photovoltaic performance [83].

BSF. The conventional BSF solar cell configuration comprises a characteristic n⁺pp⁺ architecture featuring a screen-printed silver front grid contact, a hydrogenated silicon nitride anti-reflective coating, a phosphorus-doped n⁺ front emitter, a p-type silicon substrate, and a full-area screen-printed aluminum rear contact (Figure 6a). BSF solar cells have sustained industrial dominance in silicon photovoltaics for multiple decades, attributed to their structural simplicity, exceptional operational stability, continuous advancements in PCE, and streamlined manufacturing protocols. The inaugural demonstration of PSTSCs employing Al-BSF bottom cells was documented by Mailoa et al. [13], achieving a preliminary efficiency of 13.7%. Contemporary research indicates that all reported BSF-based PSTSCs exhibit constrained efficiencies below 23.5% [33,84,85], primarily limited by substantial carrier recombination losses at the rear contact interface.

PERC. PERC technology features a unique structure on the backside, utilizing a dielectric passivation layer to improve efficiency. This layer, typically made of silicon dioxide or aluminum oxide, is selectively removed in certain areas to create local aluminum contacts, enhancing light absorption and electron flow (Figure 6b). This architectural innovation effectively mitigates rear surface recombination while enhancing internal light reflection through back surface passivation, thereby synergistically improving both V_OC and J_SC. When implemented as the bottom cell in PSTSCs by Wu et al., the PERC architecture achieved a certified efficiency of 22.5% [86]. Recent advancements by Phung et al. employing near-industrial PERC configurations in p-i-n perovskite tandem cells demonstrated stabilized PCE reaching 23.6% (Figure 6c) [87]. Although surpassing conventional Al-BSF in efficiency potential, PERC cells remain constrained by surface recombination at localized metal contacts and carrier loss mechanisms including bandgap narrowing and Auger recombination in heavily doped junctions. These inherent limitations currently restrict industrial-scale PERC efficiencies to approximately 24% [88].

TOPCon. TOPCon represents an advanced silicon photovoltaic architecture comprising an ultrathin tunnel SiO_x layer and a heavily doped poly-Si layer. Figure 6d demonstrates that the implementation of rear local contacts in PERC configurations with TOPCon technology substantially reduces recombination current density from 36 fA cm⁻² to 6 fA cm⁻² [89]. Through precise doping optimization, poly-Si layers can be engineered as either p-type (p-TOPCon) or n-type (n-TOPCon) to selectively facilitate hole or electron transport, respectively. Current technological advancements have enabled the successful implementation of p-TOPCon and n-TOPCon as substitutes for conventional p⁺-Si emitters in n-type PERT solar cells [31] and p-type PERL solar cells [45]. Notably, n-TOPCon has demonstrated efficacy in replacing rear local contacts for n-type PERC bottom cells [46]. A significant development involves dual-side TOPCon integration, which eliminates a direct metal–semiconductor interface while minimizing carrier recombination and collection losses at both surfaces. Ying et al. achieved enhanced passivation in n-TOPCon through surface nanostructuring, realizing perovskite/double-sided TOPCon TSCs with a stabilized PCE of 28.2% [47]. Subsequent optimization of p-type poly-SiC_x by Walter et al. yielded 4 cm² perovskite/double-sided TOPCon TSCs exhibiting 28.3% PCE [90]. Singh et al. further demonstrated perovskite/double-sided TOPCon TSCs with 23.2% PCE using poly-SiO_x-based TOPCon architecture [91]. Recent breakthroughs by Ding et al. developed highly passivated p-TOPCon structures achieving a remarkable implied V_OC of 715 mV in double-sided configurations, with corresponding 1 cm² perovskite/TOPCon TSCs reaching 28.2% PCE (Figure 6e) [58].

SHJ. SHJ technology represents an advanced passivated contact approach for crystalline silicon solar cells, alongside TOPCon, employing intrinsic and doped hydrogenated amorphous silicon (a-Si:H) layers for surface passivation (Figure 6f). The doped a-Si:H layers establish a unified framework for contact formation through their inherent band structure properties, enabling efficient charge separation and selective carrier transport. Specifically, p-type a-Si:H layers function as hole-collecting contacts while n-type a-Si:H layers serve as electron-collecting contacts. This inherent carrier-selectivity mechanism contributes to SHJ’s superior efficiency, particularly notable in PSTSCs where SHJ bottom cells demonstrate exceptional efficacy in near-infrared photon harvesting and optical loss minimization within tandem architectures [92]. The foundational work on perovskite/SHJ tandem devices was pioneered by Albrecht et al. in 2015, achieving 18% initial efficiencies through optimized interfacial engineering [27]. Building upon established SHJ fabrication protocols that deliver exceptional V_OC and FF in single-junction devices, current optimization strategies for SHJ bottom cells primarily focus on photon management enhancements. Ren et al. [93] demonstrated significant near-infrared responsivity improvements in n-i-p configured perovskite/SHJ tandems through substitution of doped a-Si:H with hydrogenated nanocrystalline SiO_x (nc-SiO_x:H) interfacial layers. The planar SHJ solar cell achieved a certified efficiency of 19.1%, and when integrated into a perovskite/SHJ tandem device, it reached 20.1% and 18.88% under reverse and forward scans, respectively. After this, the design was adopted as a baseline in several record-efficiency (>29%) perovskite/silicon tandem solar cells [9,49]. In addition to nc-SiO_x:H, subsequent studies have validated comparable performance enhancements using alternative light-coupling materials including μc-SiO_x [53] and nc-Si:H [54]. Recent advancements by Harter et al. introduced molecular engineering of hole transport layers through phosphoric acid-functionalized self-assembled monolayers (Me-4PACz), enabling perovskite absorber optimization. The champion perovskite–silicon tandem device, benefiting from this interfacial modification, delivered a V_OC of up to 1.954 V, a remarkable cumulative J_SC of 40.2 mA cm⁻², and a certified PCE exceeding 30% under standard test conditions (Figure 6g) [59].

Figure 6. (a) Schematic of BSF solar cell. (b) Schematic of PERC solar cell. (c) Performance of PSTSCs with a PERC bottom cell [87]. (d) Schematic of TOPCon solar cell. (e) Structure (left) and cross-sectional SEM images (right) of PSTSCs with a double-sided TOPCon bottom cell [58]. (f) Schematic of SHJ solar cell. (g) Structure (left) and performance (right) of PSTSCs with a SHJ bottom cell [59].

Figure 6. (a) Schematic of BSF solar cell. (b) Schematic of PERC solar cell. (c) Performance of PSTSCs with a PERC bottom cell [87]. (d) Schematic of TOPCon solar cell. (e) Structure (left) and cross-sectional SEM images (right) of PSTSCs with a double-sided TOPCon bottom cell [58]. (f) Schematic of SHJ solar cell. (g) Structure (left) and performance (right) of PSTSCs with a SHJ bottom cell [59].

4. Challenges of PSTSCs

4.1. Stability of Perovskite and PSTSCs

While 2T PSTSCs have achieved certified PCEs exceeding 34%, stability issues in PSTSCs remain a fundamental challenge. A timeline of perovskite/silicon tandem cell stability over the last decade is shown in Figure 7. Current stability assessments primarily follow International Summit on Organic Photovoltaic Stability (ISOS) guidelines, as conventional International Electrotechnical Commission (IEC) standards fail to adequately address halide perovskite-specific ion migration mechanisms. A certified 25.9%-efficient single-junction device maintains 94.8% initial PCE after 10,488 h dark storage (25 ± 2 °C, N₂ environment) but conversely exhibits 10% power output degradation within 100 h of continuous illumination (AM 1.5G, 1 sun, 25 ± 2 °C, N₂ environment) [38].

Key issues concerning stability include moisture-induced degradation [95], photoinduced phase segregation [96,97], thermal instability [98], and ion migration [99,100] within the perovskite layer. Moisture sensitivity primarily stems from the hygroscopic nature of organic cations (e.g., MA⁺) and the hydrolytic instability of Pb-I bonds, leading to irreversible decomposition into PbI₂ and volatile byproducts (Figure 8a). Advanced encapsulation techniques [101] and hydrophobic interfacial layers [102] (e.g., fluorinated graphene) have been explored to mitigate this. However, some encapsulation materials will undergo photodegradation when used outdoors for a long time, and there is currently no ideal encapsulation material to meet the commercial demands of PSTSCs [101]. Photoinduced phase segregation under continuous illumination causes halide ion redistribution, generating I-rich and Br-rich domains that create charge-trapping centers and reduce V_OC [103]. Alloying with cesium/rubidium [104,105] and employing light-soaking preconditioning strategies shows promise in suppressing this phenomenon. Thermal instability arises from the volatile organic components and lattice expansion at elevated temperatures (>85 °C), exacerbated by interfacial reactions with charge transport layers. The degradation process of typical perovskite under light or thermal conditions is shown in Figure 8b [106]. Inorganic perovskite variants (e.g., CsPbI₃) and stabilized α-FAPbI₃ through strain engineering demonstrate improved thermal resilience. Recently, Li et al. integrated a polymer poly (methylmethacrylate, PMMA)-coupled monolayer graphene interface with a perovskite absorber layer to limit photolattice expansion and reduce the deformation ratio from 0.31% to 0.08%, minimizing structural damage due to dynamic lattice evolution, as shown in Figure 8c [107]. Under full-spectrum AM 1.5 G sunlight irradiation, the solar cell maintained more than 97% of the initial photovoltaic conversion efficiency after more than 3670 h of maximum power point tracking at 90 °C. Ion migration, particularly of iodide vacancies and interstitial defects, induces hysteresis effects and accelerates electrode corrosion. Grain boundary passivation via zwitterionic molecules [108] and the integration of 2D/3D heterostructures [109,110,111,112,113,114] have effectively suppressed ion diffusion while maintaining charge transport efficiency. Sun et al. isolated ion migration within the perovskite thin film by reconfiguring the perovskite surface into a 0D structure, in which this modification effectively prevents iodide ions from migrating into the charge transport layer [110]. The solar modules from the study show improved stability, with an estimated T80 lifetime of 2478 cycles, which is equivalent to more than 6.7 years of operation at 25 °C in bright−dark conditions (Figure 8d). Additionally, interfacial reactions between perovskite and charge transport layers (e.g., HTL/ETL) accelerate performance decay by promoting chemical degradation pathways, such as ion migration, interfacial defect formation, and redox processes at heterojunctions, while mismatched thermal expansion coefficients in the tandem structure introduce mechanical stress through repeated thermal cycling, leading to delamination, microcrack propagation, and irreversible lattice strain accumulation at layer interfaces (Figure 8e,f). Another critical challenge lies in achieving optimal interfacial engineering between the perovskite top cell and silicon bottom cell. The inherent lattice mismatch at the heterojunction creates defective states that disrupt carrier transport dynamics [115], while parasitic absorption phenomena in non-active layers disproportionately affect high-energy photons, collectively exacerbating non-radiative recombination pathways. Furthermore, the requirement for photocurrent continuity across tandem interfaces imposes stringent constraints on interfacial charge extraction layers, necessitating atomic-level precision in energy band alignment to mitigate recombination losses. The spectral sensitivity challenge is amplified by the interdependent bandgap tuning requirements—perovskite layers (1.6–1.8 eV) must maintain optimal light harvesting in the visible spectrum while allowing sufficient near-infrared transmission to the silicon bottom cell (1.1 eV). This delicate balance demands real-time adaptive optimization of carrier collection efficiency under varying solar spectra, requiring advanced photon management strategies to reconcile the fundamentally different absorption characteristics and carrier diffusion lengths of the two photovoltaic materials.

Overall, although perovskite-silicon tandem solar cells have made significant progress in terms of efficiency, there are challenges in terms of stability. At present, the stability has been improved through material selection, interface engineering suppression and device encapsulation, but it has not reached a satisfactory level. Therefore, it is highly necessary to be committed to developing new encapsulation materials and advanced photon management strategies. In addition, future work should also focus on the synergistic interaction of environmental stress factors (light, heat, humidity) and the degradation of perovskites and reveal the exact degradation mechanism within perovskites and at the device interfaces.

4.2. Conformal Coating of Perovskite on Textured Silicon

The prevailing method for fabricating high-efficiency perovskite solar cells predominantly employs solution-based spin-coating techniques, which typically require a planar or nano-scale texture of silicon substrates. This approach conflicts with standard industrial practices for single-junction crystalline silicon cells, where alkaline wet chemical etching is routinely implemented to create micro-scale textured surfaces, a critical feature for photon management through enhanced light trapping and reduced reflectance. The pursuit of complete substrate coverage necessitates perovskite films with micrometer-scale thickness [34,36,118], yet elevated precursor concentrations in solution processing induce crystallization-induced stress accumulation during film formation [119,120,121]. These residual stresses concentrate preferentially at inverted pyramid vertices, manifesting as microcracks and interfacial voids that critically constrain the efficiency enhancement and stability of PSTSCs. The hybrid evaporation-solution (HES) process represents a synergistic integration of conventional solution processing and vapor phase deposition methodologies [23]. In this approach, inorganic precursors are initially deposited via thermal evaporation onto textured silicon substrates, followed by solution-phase treatment with organic ammonium salts. This sequential deposition strategy enables complete substrate coverage through conformal coating formation over a micro-scale textured silicon surface. Compared to co-evaporation techniques, the HES process demonstrates enhanced flexibility through the utilization of multifunctional additives that enable efficient perovskite engineering. Crucially, these additives, which are typically non-volatile or require sub-stoichiometric quantities for optimal performance, cannot be effectively incorporated through thermal evaporation approaches due to inherent limitations in vapor pressure and precise dosage control during vapor phase deposition processes. However, the HES method also has its own problems to be solved, for instance, interfacial engineering between perovskite and textured silicon substrates. The three-dimensional topography of micro-sized pyramids creates localized variations in surface energy and charge transport dynamics, potentially inducing non-uniform carrier extraction across the device architecture [122]. Recent studies reveal that incomplete passivation of silicon surface states beneath the perovskite layer can generate recombination spots at pyramid facets, particularly under forward bias conditions. This phenomenon disproportionately affects large-area devices where current collection homogeneity becomes paramount. Another critical challenge lies in maintaining stoichiometric control during the HES process’s sequential deposition stages. While the evaporation step ensures precise inorganic precursor distribution, subsequent solution-phase organic salt infiltration must overcome kinetic barriers imposed by textured surfaces. The wetting behavior of organic solutions on faceted silicon microstructures frequently leads to localized composition deviations, particularly at pyramid vertices where precursor solution accumulation alters the Pb:I ratio. Compared with spin-coated films, those HES processes typically exhibit smaller grains, poorer crystallinity, and diminished optoelectronic performance [123]. Thermal management during post-deposition annealing introduces further complications, as differential thermal expansion coefficients between perovskite and silicon generate shear stresses at the interface [124,125]. These thermo-mechanical stresses synergize with existing crystallization-induced stresses, potentially exacerbating delamination risks during thermal cycling. To address these issues, Er-raji et al. introduced urea in the solution step to accelerate the crystallization kinetics of perovskite and achieved the formation of large grain size at low annealing temperature (100 °C) [126]. Meanwhile, the residual urea locally replaced C₆₀ on the perovskite surface, reducing the non-radiative recombination at the interface and the champion PSTSCs device provided a stable PCE of 30%.

Overall, despite advances in HES processes, challenges remain in achieving electrochemically uniform interfaces on micro-textured silicon and the trade-off between optical gain and electrical loss remains unresolved. Future research should prioritize solvent engineering and interface engineering to reconcile photon management with carrier extraction efficiency.

4.3. Simplified ETL Architecture

The most widely adopted multilayer ETL-TCO architecture in PSTSCs, comprising sequentially deposited functional layers of LiF, C₆₀, SnO_x and ITO, presents inherent compromises between performance optimization and structural sophistication. In this configuration, the LiF layer functions as an interfacial modification layer to passivate defects at the C₆₀/perovskite interface. While both C₆₀ and SnO_x demonstrate electron-selective capabilities in single-junction perovskite devices, their combined integration becomes imperative in PSTSCs. The C₆₀ layer necessitates subsequent atomic layer deposition (ALD) of SnO_x overlayers due to its inadequate mechanical resilience against the sputter damage during TCO deposition [43,127,128,129,130]. Conversely, direct ALD application of SnO_x on perovskite proves infeasible, as the aqueous oxidants and elevated temperatures inherent to ALD processes would induce decomposition of the moisture-sensitive perovskite layer. Although these multilayers exhibit effective charge transport and structural integrity, their rigid architectural configuration imposes substantial device complexity and parasitic absorption loss [131]. If a single layer could substitute the entire multilayer structure or alternatively enable the removal of either the C₆₀ or SnO_x component, this modification would significantly streamline the fabrication process while simultaneously enhancing photocurrent generation and thereby improving the PCE. Recent efforts to simplify ETL configurations face fundamental material compatibility constraints. Hybrid interfacial engineering approaches combining phosphonic acid self-assembled monolayers with ALD-grown metal oxides show promise in replacing conventional C₆₀/SnO_x stacks [63], yet achieving simultaneous optimization of band alignment and chemical stability remains problematic, and the slow deposition rate of the ALD process and the costly system inevitably affects the scale-up of the industrial process. Recent research has focused on a simple process method for preparing ETL without ALD in PSTSCs. Recently, Li et al. replaced ALD-SnO_x with thermal evaporation of 8 nm Bathocuproine (BCP) and combined it with soft-sputter deposition IZO transparent electrodes, achieving a certified PCE of 29.91% on industrial PERX/TOPCon silicon-bottom cells [132]. This strategy not only omits the ALD step but also increases the V_OC and simultaneously reduces the parasitic absorption of 0.41 mA cm⁻², indicating that the tandem structure can be significantly simplified while maintaining the passivation of the interface. Alternatively, Magliano et al. reported a solution-treated protective layer of aluminum-doped zinc oxide (AZO) nanoparticles: dynamic spin-coating of 20 nm AZO (N-21X) led to the formation of dense, vertically oriented grains on the C₆₀ surface, the final semi-transparent devices were prepared with a PCE of 18.1%, and the PSTSCs achieved 25.3% PCE when using a two-step hybrid perovskite [133]. The process requires no vacuum and is compatible with p-type polished/unpolished hybrid wafers, providing another viable route to low-cost, scalable ALD-free stacks.

The requirement for low-temperature processing (<100 °C) to preserve underlying perovskite integrity severely limits available material combinations and crystallization quality of alternative ETL candidates. Reactive plasma deposition (RPD) presents a promising solution to this challenge, exemplified by the direct deposition of SnO_x on perovskite substrates [134], the deposition of TCO on C₆₀ layers [135] and preparation of W-doped In₂O₃ (IWO) films for photovoltaic devices through RPD processes [136,137]. As a non-thermal, solvent-free technique, RPD enables conformal deposition of metal oxides like SnO_x at near-ambient temperatures, circumventing perovskite decomposition risks associated with conventional ALD [138]. In addition, the system offers many significant advantages such as low ion damage, large area deposition, and high throughput and scalability, making it particularly suitable for applications involving delicate materials [139]. Crucially, RPD allows precise tuning of oxygen vacancies and work function through plasma parameter modulation (e.g., O₂/Ar ratio, bias voltage), facilitating optimal band alignment with both perovskite absorbers and TCO electrodes. This capability addresses the persistent trade-off between charge extraction efficiency and chemical stability in simplified ETL designs. RPD has been studied for the manufacture of high-quality transparent conductive layers, but there are few reports on its application in tandem devices. Recently, Wang et al. demonstrated the potential of preparing ITO as a PSC buffer layer through RPD and paved the way for its integration into tandem solar cells [135]. The PSCs with a bandgap of 1.67ev achieved an impressive V_OC of 1.252 V and perovskite/CIGS tandem solar cells achieved an efficiency of 29.03%, which ranks among the highest reported efficiencies for CIGS/perovskite tandem configuration. This plasma-assisted deposition mechanism further enables direct TCO deposition onto C₆₀, eliminating the need for ALD-SnO_x buffer layers and enhancing interfacial integrity throughout the fabrication process.

Overall, although the multilayer ETL-TCO structure widely adopted in current PSTSCs can ensure charge transport, its complex architecture leads to significant parasitic absorption and preparation bottlenecks. Although technologies such as BCP/IZO and RPD have initially simplified the ETL structure, future work should design a single functional layer that combines band matching, interface passivation and sputtering damage resistance to replace multi-layer stacking.

4.4. Fabricaition of PSTSCs Based on Full-Sized Silicon Cell

While the certified PCE of PSTSCs has recently surpassed 34.85%—exceeding the theoretical limit of single-junction silicon solar cells (29.4%)—these record efficiencies remain confined to lab-scale devices with active areas of approximately 1 cm². This scale limitation poses significant challenges for commercial viability, as industrial silicon solar cells typically require cell dimensions exceeding 10 × 10 cm². The predominant laboratory technique of solution spin-coating, despite its success in small-area device optimization, faces fundamental barriers in achieving uniform perovskite deposition across larger substrates due to inherent limitations in solution viscosity control and centrifugal force distribution [140]. Recent advances in scalable deposition techniques demonstrate promising pathways for industrial adaptation. Blade-coating, inkjet printing and slot-die coating methods have emerged as viable alternatives, enabled by precise engineering of thin film deposition (Figure 9a–d). Critical processes include (1) perovskite precursor ink optimization, (2) coating process optimization, and (3) nitrogen flow drying or vacuum drying optimization for controlled solvent evaporation kinetics. These collectively enable the formation of pinhole-free perovskite layers with tunable thickness from 300 to over 1000 nm and enhanced crystallinity over large area substrates [141,142]. In contrast to large-area PSC modules typically spanning several square meters, large-area PSTSCs exhibit substantially reduced active areas dictated by standardized monocrystalline silicon wafer dimensions, specifically M6 (166 × 166 mm²), M10 (182 × 182 mm²), and G12 (210 × 210 mm²). The fabrication of large-area PSCs presents additional technical challenges, particularly regarding the requirement for extended coating widths that significantly exceed those of PSTSCs. Crucially, complex manufacturing processes involving laser ablation techniques, which are essential for achieving series interconnection in PSC modules, become redundant in PSTSCs production, thereby streamlining the manufacturing workflow and substantially reducing costs.

The design and fabrication of metal grids also constitute a critical challenge in the manufacturing of large-area PSTSCs. Conventional screen-printed silver grids, which rely on high-temperature sintering processes (>200 °C) [143], are incompatible with the thermally sensitive perovskite materials [29], necessitating the development of low-temperature curing of silver pastes [144]. Although significant progress has been made in formulating ultra-low-temperature silver pastes (curing at <130 °C) for perovskite tandem cells, their resistivity remains substantially higher than that of traditional silicon-based counterparts, limiting the FF and overall efficiency of tandem devices [145]. To address this issue, large-area cells require finer grid lines to minimize shading losses while maintaining low series resistance, which demands advancements in high-precision screen-printing technologies or alternative metallization approaches. Busbar-free (BBF) architectures have emerged as a promising solution, reducing silver consumption by >30% and simplifying the metallization process through optimized electrode designs (Figure 9e). However, the implementation of BBF technologies necessitates the development of novel electrode materials (e.g., copper-based composites or carbon-based alternatives) and robust interconnection schemes to ensure long-term stability and reliability under operational stresses such as thermal cycling and humidity exposure [146]. From another perspective, the photocurrent density generated by tandem cells reaches only half of that observed in standalone silicon solar cells, indicating that the electric losses at the TCO/collecting grid electrode can be significantly reduced in a tandem scenario. Consequently, PSTSCs exhibit substantial tolerance for metallization conductivity and grid density compared to conventional silicon solar cells. This inherent advantage facilitates the successful establishment of large-area PSTSCs, even when employing metallization layers with suboptimal conductivity or reduced grid coverage.

Overall, PSTSCs are restricted by the small-area (~1 cm²) rotary-coating process, and its industrialization faces two major bottlenecks: the uniformity of large-area deposition and the design of metal grids. At present, blade coating/slot-die coating can initially solve the problem of large-area film formation, but there is no single coating method that can consistently achieve extremely low non-uniformity in thickness, composition and defects on large-area devices. In the future, further collaborative optimization of deposition processes and more compatible metal grid designs are still needed to break through the industrialization bottleneck.

Figure 9. (a) Blade coating on hot substrate surfactant [147]. (b) Blade coating with N₂ knife [34]. (c) Spray coating [148]. (d) Slot-die coating [149]. (e) The busbar-free electrode pattern [150].

Figure 9. (a) Blade coating on hot substrate surfactant [147]. (b) Blade coating with N₂ knife [34]. (c) Spray coating [148]. (d) Slot-die coating [149]. (e) The busbar-free electrode pattern [150].

5. Conclusions

This review comprehensively examines recent advancements in two-terminal PSTSCs, with particular emphasis on subcell optimization strategies. Recent developments demonstrate that PSTSC efficiencies have consistently exceeded those of both single-junction PSCs and commercial silicon solar cells, indicating substantial potential for industrial implementation. However, significant opportunities remain for enhancing the PCE of tandem architectures. Continuous scientific research efforts have driven performance improvements through structural optimizations addressing optical/electrical losses, parasitic absorption mitigation, and reflection reduction. Material innovations, particularly in perovskite absorber layer composition engineering, have contributed to enhanced device stability. Critical challenges requiring resolution for commercial viability include stability of perovskite and PSTSCs, conformal coating of perovskite on textured silicon, simplified ETL architecture and fabrication of PSTSCs based on a full-sized silicon cell. Successful resolution of these technical issues could enable perovskite/silicon tandem devices to reduce costs while increasing photovoltaic module power output, thereby positioning this technology as a pivotal contributor to global renewable energy transition efforts.