Last updated: 12 November 2026 — Includes the latest perovskite and silicon solar cell efficiency records, independent verification details and technical analysis.
What Is the Highest Efficiency Perovskite Solar Cell in 2026?
The highest single-junction perovskite efficiency in the authoritative records checked for this update is 28.0%, achieved by Hainan University. The NLR research-cell dataset records an NPVM measurement in February 2026, using a 0.051 cm² designated illumination area. NLR research-cell data
Solar Cell Efficiency Tables: Version 68 reports 28.0 ± 0.6% and includes this small device among its notable exceptions, rather than the main class records. Device size and area definition matter when comparing efficiencies. Green et al., Version 68
For perovskite–silicon tandems, LONGi announced 35.5% in July 2026 and reported certification by the European Solar Test Installation (ESTI). This newer announcement is distinguished here from the 35.2% result in the downloaded NLR dataset. LONGi announcement, NLR data
| Cell type | Efficiency | Measured area | Institution | Evidence and date |
|---|---|---|---|---|
| Single-junction perovskite | 28.0% | 0.051 cm², designated illumination | Hainan University | NPVM, February 2026; NLR dataset and Efficiency Tables, Version 68 . Small-area notable exception. |
| Perovskite–silicon tandem | 35.2% | 0.9994 cm², designated illumination | LONGi | ESTI, February 2026; NLR dataset and Efficiency Tables, Version 68 . |
| Perovskite–silicon tandem — newer announcement | 35.5% | Not stated in the cited announcement | LONGi | Announced July 2026; LONGi reports ESTI certification . |
| All-perovskite tandem — NLR-listed benchmark | 30.1% | 0.04934 cm², designated illumination | Nanjing University / Renshine | JET; NLR dataset entry dated June 2024 . |
| All-perovskite tandem — newer published result | 30.3% | Not verified for this update | NIMTE, Chinese Academy of Sciences, and collaborators | Paper published 27 April 2026 reports certified 30.3%. |
Sources checked: 12 November 2026. These are research-cell results, not commercial module efficiencies. Device areas, area definitions and evidence sources differ.
Version 68 reports uncertainties of ±0.6 percentage points for the 28.0% result and ±1.1 percentage points for the 35.2% result. The underlying certificates for the newer 35.5% and 30.3% results were not inspected for this update. Certification is attributed to the cited company announcement or research paper.
Sources: NLR original dataset, Version 68, LONGi, and Tian et al., Nature Nanotechnology.
The 30.3% paper means 30.1% should no longer be described without qualification as the highest reported certified all-perovskite efficiency. The paper’s supplementary certificate was not available for inspection in this update, so its test centre, certified area and measurement protocol are not supplied here. The earlier device’s area must not be transferred to the newer result.
These benchmarks use different device sizes and evidence sources. Inclusion in a record compilation, certification reported in a research paper, and certification reported by a manufacturer are identified separately. A high initial efficiency also does not establish long-term operational stability.
What Is the Record Efficiency for Perovskite–Silicon Tandem Solar Cells?
LONGi’s latest announcement found in this review reports 35.5%. The company announced the result on 14 July 2026, with its English news article published on 15 July, and attributes certification to ESTI. The announcement does not specify the device area or measurement uncertainty. LONGi, July 2026
The corresponding benchmark in Solar Cell Efficiency Tables: Version 68 is 35.2 ± 1.1%, measured by ESTI in February 2026 on 0.9994 cm². The downloaded NLR dataset also ends at 35.2% for this category. The two figures therefore represent different stages of reporting, rather than conflicting measurements of the same device. Version 68, NLR data
Want to explore what drives these tandem gains yourself? You can model perovskite–silicon tandem stacks — optical coupling, current matching and JV behaviour — in Setfos, and try it free for a month on your own device data.
Fig. 1. Selected research-cell efficiency milestones, 1978–2026. Filled markers use measurements from the NLR dataset; the silicon series includes heterojunction devices. Open markers show newer LONGi announcements, plotted by announcement date: 35.5% for perovskite–silicon tandems and 28.29% for crystalline silicon. LONGi reports ESTI and ISFH verification, respectively. Lines guide the eye and do not represent additional measurements. Devices differ in area and area definition. Sources checked on 12 November 2026.
Model your own perovskite–silicon tandem in Setfos
Simulate optical coupling, current matching and JV behaviour in tandem stacks — free for one month.
Start your free Setfos trial →So why is breaking this limit important? Due to optical, thermal and electronic losses, it is highly unlikely that a commercially viable single junction silicon solar cell will ever have an efficiency close to the S-Q limit. Continuous improvements in silicon solar cell technology (like passivation techniques, advanced light trapping, etc.) are pushing their efficiencies higher, but surpassing the S-Q limit will not be possible.
This article will explore the underpinnings of multi-junction solar cells, and explain how optical and electrical simulations can be utilized in understanding how to optimize the performance of tandem solar cells.
Why Are Silicon Solar Cells Limited by Efficiency?
Fig. 2. Schematic illustration of optical (green), thermal (yellow) and electronic (blue) losses in a solar cell. Their relative contributions depend on the absorber bandgap and the assumptions used in the calculation.
Single-junction silicon solar cells cannot convert all incoming sunlight into electricity. Their performance is constrained by the solar spectrum, the silicon bandgap and fundamental recombination processes.
The Shockley–Queisser limit describes an ideal single-junction solar cell and reaches approximately 33.7% at the optimum bandgap under standard one-sun illumination. This is not a silicon-specific limit. For crystalline silicon, calculations that include intrinsic radiative and Auger recombination give a limiting efficiency of approximately 29.4%, under the modelled conditions. Polman et al., Science; Niewelt et al., Solar Energy Materials and Solar Cells.
Silicon’s indirect bandgap means it absorbs light relatively weakly near its absorption edge. Efficient silicon cells therefore require suitable absorber thicknesses and effective light trapping. However, the indirect bandgap alone does not explain their efficiency limit.
Three broad categories help explain the losses:
Optical losses: Photons with energies below the bandgap cannot generate electron–hole pairs through band-to-band absorption. In practical devices, reflection and absorption in inactive layers also reduce the light available for electricity generation.
Thermal losses: When an absorbed photon has more energy than the bandgap, the excited carriers generally relax towards the band edges. Their excess energy is released as heat through a process called thermalisation.
Electronic losses: Recombination reduces the number of carriers available for extraction and limits the achievable voltage. Radiative recombination emits photons, while non-radiative processes—including Auger and defect-assisted recombination—transfer energy through other pathways. The operating voltage also lies below the bandgap expressed as a voltage.
These contributions should not be assigned universal percentages: their magnitudes depend on the material, device structure, illumination and modelling assumptions.
Commercial modules introduce further losses through interconnections, inactive areas, optical interfaces and electrical mismatch. Their efficiencies must therefore be distinguished from research-cell records; 20% is not a fundamental ceiling for silicon modules.
How Can We Overcome the Shockley–Queisser Limit?
The Shockley–Queisser limit reaches approximately 33.7% for an ideal single-junction solar cell with an optimal bandgap under standard one-sun illumination. In real-world conditions, commercial silicon modules top out at around 25% due to optical, thermal, and electronic losses. Tandem solar cells—particularly perovskite/silicon tandems—offer a way to exceed this limit by using materials that absorb different portions of the solar spectrum. Here, we list only a few solutions the community has proposed to reduce losses in silicon solar cells. We invited the reader to search for more detailed analyses if interested in this topic:
Hot carrier solar cells extract high-energy charges before the excess energy turns into heat.
Intermediate bandgap solar cells absorb photons at energies smaller than a bandgap without reducing the voltage. The Intermediate band (IB) energy level is between the conduction and valence bands. The IB allows photons with energy smaller than the bandgap to excite a charge from the valence band to the conduction.
Multiple exciton generations in a solar cell create more than one charge carrier from a single high-energy photon.
Despite the proposed solutions, theoretically, should improve the efficiency of a silicon solar cell, they are rarely employed in commercial photovoltaics.
Fig. 3. Three advanced solar cell mechanisms: hot carrier absorption, intermediate band transitions, and multiple exciton generation, each enhancing electron excitation and energy conversion efficiency.
Fig. 4. Energy band diagrams for a photovoltaic cell, showing the generation of an open-circuit voltage from absorbed photons with increasing energies and the resultant movement of electrons from the valence to the conduction band
What Are Multi-Junction and Tandem Solar Cells?
Conventional solar cells have a single PN Junction, a combination of p and n-doped semiconductor layers. The opposite type of doping creates an internal electric field that promotes the extraction of photogenerated charges. A multi-junction solar cell has two or more junctions, and each one absorbs at a different wavelength. The more junctions, the greater the portion of the solar spectrum that is absorbed, resulting in higher theoretical efficiency.
Solar cells with infinite multi-junctions have a theoretical efficiency of up to 86.8 %. -Currently, the most efficient solar cell in the world has an efficiency of 47.6 % under concentrating illumination, and it is made up of 4 junctions. The record efficiency under standard Global Spectrum instead amounts to 39.5 % for a triple junction solar cell. Both of these record solar cells have light-absorbing materials constituted by III-V semiconductors. These advanced solar cells are massively more expensive than silicon solar cells. Their use is mostly limited to special applications such as space applications where cost is less of an issue than other aspects such as weight.
Fig. 5. Various tandem solar cell configurations and their corresponding electrical circuit diagrams, demonstrating how light management and connectivity affect the overall efficiency and performance of the solar cells.
Tandem Cell Architectures: 2-Terminal, 4-Terminal, and 3-Terminal
A multi-junction solar cell with only two absorbing materials is a so-called tandem solar cell, and it is the most interesting option for “standard” commercial uses. They can be fabricated with four or two electrodes, also known as four-terminal and two-terminal tandem configurations. The three-terminal tandem is the third contact variation that is less known but is catching up. In a four-terminal tandem solar cell the two sub-cells are fabricated independently and are electrically isolated from each other, the sub-cells are either mechanically stacked or coupled with a spectral splitter which is essentially a set of lenses to direct light to the appropriate sub-cell. However, this configuration has considerable parasitic absorption and reflection from the inactive layers and a higher overall material cost due to multiple substrates, complex assembly, and wiring.
Fig. 6. A solar energy system where a holographic spectrum splitter divides incident sunlight into high and low energy photons to be absorbed by corresponding high and low band-gap solar cells, enhancing the photovoltaic efficiency.
In two-terminal tandem solar cells, the subcells are electrically connected in series and must carry the same current. A mismatch between their generated photocurrents can limit the tandem’s power output, so current matching is an important design consideration. This configuration is obtained by stacking the sub-cells mechanically or, most effectively, by adding a transparent conducting layer, also known as the recombination layer, which is essential to form an ohmic contact between the sub-cells. This configuration requires less wiring and is simpler to install, but the fabrication process is challenging since the top cell must be deposited without damaging the bottom cell. Additionally, with daily and seasonal variations of the solar spectrum, the current matching requirement is not always satisfied.
Three terminal tandems have a middle contact between the two sub-cells avoiding the need for the tandem junction and current matching as in two terminal tandems. However, these devices are less explored than 2-terminal and 4-terminal solar cells. We will not discuss it here.
Fig. 7. Contour plots comparing the performance of tandem solar cells, with axes representing the bandgaps of top and bottom cells and contours indicating efficiency percentages. Source: R. K. Kothandaraman, Y. Jiang, T. Feurer, A. N. Tiwari, F. Fu, Near-Infrared-Transparent Perovskite Solar Cells and Perovskite-Based Tandem Photovoltaics, Small Methods 2020, 4, 2000395. https://doi.org/10.1002/smtd.202000395
The highest theoretical efficiencies under standard light intensities and AM1.5 G are 46% and 45.7% for four-terminal and two-terminal tandem solar cells, respectively. Two-terminal tandem PVs have fewer bandgap combinations to reach the highest efficiency because they require both electrical and optical matching, whereas 4-terminal tandem PVs need only optical matching, allowing a larger variation in the band gap of the top cell. Which of these two solutions could be the winner is still unclear. Both of them could have several benefits, but they have issues to be resolved before we can see them in the market.
So far, we assumed that the illumination impinges exclusively on the top subcell. It is also worth mentioning bifacial tandem solar cells. These interesting solutions collect light irradiation impinging on the rear side of the solar cells, increasing the overall amount of photons that are absorbed by the active layer.
Fig. 8. A bi-facial tandem solar cell setup where sunlight is captured by both the top and bottom cells, with the bottom cell also utilizing reflected light (albedo) for energy generation.
How Does Light Management Improve Efficiency?
The most widely studied tandem solar cells are silicon/perovskite, where silicon is the bottom sub-cell. Silicon has a band gap of 1.12 eV. To maximize the efficiency, the ideal band gap of the top cells should be between 1.67 and 1.75 eV. Perovskites are great solutions for the top cell as they can be engineered to have a bandgap between 1.15 and 3.1 eV. They can be deposited at low temperatures with solution process methods like spin coating, avoiding any damage to the bottom cell during the fabrication of the two-terminal solar cell. Additionally, perovskites have a direct band gap and a strong light absorption coefficient. Therefore, thicknesses of around 500 nm are sufficient for the perovskite sub-cell. As a comparison, the silicon layer is almost a thousand times thicker than the perovskite due to its indirect band gap and low absorption coefficient.
Fig. 9. A cross-sectional diagram of a tandem solar cell, showing the layering from the superstrate to the encapsulation. The perovskite and silicon sub-cells are separated by transparent contacts and coupled optically.
Several strategies have been developed to increase the light absorption of the silicon sub-cell. The usage of a textured surface (Fig. 10) is one of the most commonly employed techniques. However, determining the optimal roughness for such surfaces is a complex process that cannot be achieved through experimental trial and error alone. That is where computer simulations come in handy. From an optical simulation, we can appreciate, for example, how much current can be recovered from a two-terminal silicon perovskite tandem solar cell by adding texture interfaces. This simulation is shown in Figure 11 and was performed with the software Setfos.
Fig. 10. An electron microscopy view of textured silicon surfaces with pyramid-like structures intended to enhance light absorption within a silicon sub-cell, illustrating the complexity of optimizing surface roughness for solar applications.
Fig. 11. The interaction of light with a solar cell, showing both wave and ray optical phenomena, with rays being reflected and absorbed while waves are diffracted within the cell's structure.
We used an optical simulation model that incorporates wave-optics to compute the reflection and transmission of coherent thin-film components, ray-optics to evaluate the angular scattering properties of the texture interfaces, and a net-radiation algorithm that uses this information to quantify the light propagation in the entire layer stack.
Fig. 12. A detailed layer stack of a perovskite/silicon tandem solar cell, specifying the materials and thicknesses of each layer from the air interface down to the substrate.
Both stand-alone optical simulation and combined optoelectronic simulation of perovskite/Si tandem solar cells are extremely important for the development of this innovative PV technology. The simulation could also help to understand the relationship between optical properties and electrical properties of the tandem PV. From the simulated absorptance of the full stack, we can calculate the photocurrent of the devices. Starting from a planar device with no texture interfaces, the addition of an anti-reflected coating improves the absorption of both the perovskite and silicon absorbers at wavelengths between 500 and 1000 nm. The texturing of both sides of the silicon sub-cell brings a clear improvement in the light absorption at wavelengths larger than 750 nm. The current increases by almost 23% compared to the planar device. It is silicon that especially benefits from light management strategies whereas perovskite does not require such treatments thanks to the high absorption coefficient.
Fig. 13. A cross-sectional scanning electron microscopy (SEM) of the front and rear side of a planar perovskites silicon tandem solar cell. Source: © D. Türkay (EPFL), C. Wolff (EPFL), F. Sahli (CSEM), Q. Jeangros (CSEM). Calculated absorptance of a full perovskite/Si tandem solar cell.
However, spin-coating perovskite on a textured surface leads to uneven deposition and the formation of holes. Possible solutions include reducing the texture size and depositing a thicker perovskite layer or depositing the perovskite conformally to the texture of the silicon sub-cell.
These two approaches led to the highest efficiency records for silicon perovskite tandem solar cells in 2022. The record efficiency of 32.5 % was obtained for a planarized tandem solar cell with a nano texture between the two sub-cells, improving light management and the deposition quality of the perovskite absorber. For a tandem cell with a micrometric texture and conformally deposited perovskite, the highest certified efficiency amounts to 31.3 %.
Fig. 14. The left graph plots current-voltage characteristics for different defect densities in a solar cell, and the right graph shows the increase in perovskite layer thickness with increasing defect density, highlighting the optical optimum at the lowest defect concentration.
What Causes Recombination Losses in Tandem Cells?
Getting the most out of solar cells requires good light management and limiting losses by recombination at the interfaces. Non-radiative recombination at the interfaces between the perovskite and the charge of transporting layers is a typical source of performance loss.
Electrical simulations with SETFOS (Fig. 1) show that with increasing defect density at the perovskite interfaces, the open-circuit voltage (Voc) decreases considerably. To compensate for the loss of carriers, the perovskite thickness should be increased by up to 30 nm from the optimum thickness value inferred from Optical simulations (Fig. 15).
Future Research Directions in Tandem Solar Technology
In concluding this broad overview of tandem solar cells, it is evident that the field is ripe with innovation and opportunities for further exploration. Here is a quick look at some of the pivotal areas of ongoing research:
Three-Terminal Tandem Solar Cells: The introduction of the three-terminal tandem design marks a significant leap forward, overcoming the limitations of the top sub-cell bandgap and promising higher energy yields with strong economic prospects and reduced environmental impact.
Innovations to Reduce Parasitic Absorption: Addressing parasitic absorption is crucial. Recent shifts in local selective contact materials aim to minimize these losses, enhancing the overall efficiency of tandem cells.
Enhancing Light Absorption: Tuning the bandgap of perovskite materials can significantly boost light absorption and photoelectric conversion efficiency, making it a promising candidate for future low-cost solar cells.
Advances in Perovskite Absorber Material: Halide perovskite materials have been a game-changer, leading to more efficient tandem solar cells with tunable and strong light-absorbing capabilities. Research is still ongoing and we are expecting several new developments that could continue increasing the efficiency of the solar cells but, most importantly, increase their operational stability.
These developments represent just the beginning of what is possible with tandem solar cells. As research continues to advance, we can expect to see even more innovative solutions and improvements in solar cell technology, driving us toward a more sustainable and efficient energy future.
Model your own perovskite–silicon tandem in Setfos
Simulate optical coupling, current matching and JV behaviour in tandem stacks — free for one month.
Start your free Setfos trial →Frequently Asked Questions
As of November 2026, the highest independently verified efficiency for a single-junction perovskite solar cell is 28.0%, achieved by Hainan University, China. The result was measured by the National PV Industry Measurement and Testing Center (NPVM) on a 0.051 cm² device and is listed as a notable exception in Solar Cell Efficiency Tables: Version 68 due to its small area.
Which company holds the record for perovskite-silicon tandem efficiency?LONGi Solar announced a record efficiency of 35.5% in July 2026 for a crystalline silicon-perovskite tandem solar cell, reporting independent certification by the European Solar Test Installation (ESTI). The published Solar Cell Efficiency Tables: Version 68 lists LONGi's earlier 35.2% result.
Why are tandem solar cells more efficient?Tandem solar cells combine absorber materials with different bandgaps to capture different parts of the solar spectrum. This reduces transmission and thermalisation losses compared with single-junction cells, enabling higher theoretical power conversion efficiencies.
The Shockley–Queisser limit describes the maximum theoretical power conversion efficiency of an ideal single-junction solar cell, accounting for transmission, thermalisation and radiative recombination losses. Under standard one-sun AM1.5G illumination, its maximum is approximately 33.7% at an optimal bandgap of around 1.34 eV. The exact value depends on the spectrum and modelling assumptions. Tandem solar cells can exceed this single-junction limit by combining absorbers with different bandgaps, but remain subject to their own theoretical efficiency limits.
For accurate optical and electrical simulation of tandem architectures, researchers use Setfos, Fluxim’s simulation platform for thin-film devices. It's widely applied in tandem optimization workflows, from EQE matching to current density prediction.
Related Perovskite Research Blogs and Papers
How is perovskite stability assessed?
Accurate benchmarking relies on standardized ISOS stability protocols for perovskite devices, which define test conditions for reliability and lifetime.