Silicon's PV Successor? Perovskite Tandem Solar and the Physics of What Comes Next

Silicon solar cells dominate the market, accounting for over 95% of new global photovoltaic capacity annually. But they face a power conversion efficiency ceiling of approximately 29.4%, the detailed-balance theoretical limit on single-junction devices [1]. The best commercial silicon modules today reach 24-26%, proving silicon a mature technology. As electricity demand outpaces land available for panels, grid interconnection queues stretch into years, and the economics of solar hinge on squeezing more watts from each unit of area, silicon's stagnation simply can't keep up with growth. The solution? Tandem solar cells.

'Tandem' solar cells rest on a simple principle: two is better than one. The technology stacks two photovoltaic absorbing materials so the pair can exceed the efficiency either could reach alone. Each layer harnesses a different portion of the solar spectrum, capturing previously unproductive wavelengths that otherwise bled away as heat.

In September 2024, UK-based Oxford PV completed the world's first commercial sale of perovskite-silicon tandem modules to a U.S. utility-scale installation [2]. In April 2025, China's LONGi demonstrated a certified world-record 34.85% efficiency for a perovskite-silicon tandem cell, smashing the silicon single-junction limit and setting the groundwork to chase the 43% theoretical ceiling for two-junction tandems [3].

First-generation products are shipping and Chinese manufacturers are building gigawatt-scale lines. But before tandems can cut into silicon's market share, the technology still faces hard questions on operational stability, manufacturing yield, and bankability.

Perovskite materials: structure and why they work

The word 'perovskite' refers to a class of materials with an ABX₃ crystal structure: a large cation (A) and a smaller metal cation (B) arranged in an octahedral cage of halide anions (X), as shown in Figure 1 below.

Solar cells are limited by their optical bandgap, the minimum energy a photon must carry to be absorbed. Perovskites are uniquely powerful because their material flexibility gives them a tunable bandgap. By adjusting the halide composition (the ratio of iodide, bromide, and chloride), the bandgap can be varied continuously across a range starting around 1.1 eV [4].

That tunability is critical for tandems. The top cell is engineered to absorb the high-energy portion of the spectrum (blue and green light) and let lower-energy photons pass through to the silicon bottom cell, which absorbs red and near-infrared wavelengths.

The same flexibility gives perovskites a high defect tolerance, part of why lab efficiencies have risen so quickly. Silicon demands extraordinarily pure material and extreme surface passivation to hit high efficiencies; perovskites hold high efficiencies despite relatively large grain boundaries and moderate defect densities.

Finally, and perhaps most important for commercialization, perovskites are solution-processable, which lowers capital intensity through high-throughput coating (spin-, blade-, or slot-die coating). That is a sharp contrast to the high-vacuum, high-temperature processes silicon requires.

Tandem architecture: two-terminal vs. four-terminal

Tandem cells come in two principal configurations, each with distinct performance and manufacturability tradeoffs.

Two-terminal (2T) monolithic tandems integrate the perovskite top cell and silicon bottom cell into a single device during fabrication. The advantage is simplicity: one set of electrical leads and standard inverter compatibility. But because the subcells sit in series, the weaker cell throttles the whole device's current, demanding 'current matching.' Efficiency drops with current mismatch, which grows as the spectrum shifts with the sun across the day.

Four-terminal (4T) stacked configurations treat the two subcells as electrically independent, with the perovskite cell on a transparent substrate placed over the silicon cell. Each has its own contacts, so both are individually optimizable and the current-matching constraint disappears, at the cost of greater manufacturing complexity and some optical loss at the extra interface.

The 2T monolithic architecture is generally seen as the most promising commercial path. It is more directly drop-in compatible with existing silicon module manufacturing and adds no separate optical element. The catch: depositing a perovskite layer directly onto a silicon heterojunction (SHJ) or tunnel-oxide passivated contact (TOPCon) bottom cell without inducing material damage and efficiency loss remains formidable.

Stability: the main roadblock to commercialization

Perovskites have one critical weakness: they are thermodynamically unstable under light. The effect is well documented but still poorly understood, and it has to be solved before perovskites can sell, either stand-alone or inside tandems.

The very halide atoms behind perovskite's best qualities, its tunable bandgap and conductivity, also drive rapid degradation under heat and illumination. This 'ion migration' produces iodide-rich and bromide-rich regions that shift the bandgap and cut photovoltage [5].

No composition to date has proven immune. Standard solar project financing assumes a 20-year debt structure with performance guarantees over that lifetime, and perovskites have yet to clear that bar. The active layer is also moisture-sensitive, but that is handled relatively easily with encapsulation. The main barrier to market is stability: degradation under both light and heat.

The reality of lead toxicity

High-efficiency perovskite cells almost universally use lead iodide at the B-site, a legitimate regulatory and environmental concern the industry has not fully resolved.

Lead is immobile in a functioning, encapsulated cell, but rainfall can leach it from a cracked or degraded module into soil and groundwater. The lead per module is modest next to common products like lead-acid batteries, and quantities at a single site are low. At projected gigawatt deployment scale, though, the aggregate exposure from module failures, end-of-life handling, and inadequate encapsulation is nontrivial and will draw regulatory headwinds.

Tin-based perovskites are a lower-toxicity alternative, but their efficiencies lag badly thanks to simple oxidation chemistry.

Manufacturing at scale: problems and promise

Closing the gap between a record lab cell and a manufacturable product is where most photovoltaic technologies have historically stumbled.

The core challenge is coating uniformity at scale. Record efficiencies come from small-area cells (typically under 1 cm²), where uniformity is easy to control. Commercial silicon modules use wafers of 200 cm² or larger, and laying down a spectrally and morphologically uniform perovskite film across that area is far harder. Blade-coating, slot-die coating, and vapor-phase deposition all show promise, but each trades off throughput, uniformity, and the moisture control needed during deposition.

For tandems specifically, there is a second issue: integrating perovskite processing into existing silicon cell lines. Adding a perovskite solution-deposition step to a mature silicon line introduces contamination risk, thermal constraints, and yield challenges. Converting existing lines rather than building new factories is commercially sensible but technically demanding.

Chinese crystalline silicon is selling below $0.10 per watt. The efficiency premium of tandems delivers real LCOE compression on constrained land, but the raw module-cost advantage that enthusiasts champion rests entirely on scaling assumptions not yet validated at commercial volume.