TECHNOLOGY EXPLAINER
Passivation
— treating the several square metres of surface hidden inside a 1 m² cell
Perovskites are said to be “defect-tolerant”, yet what still eats into the voltage of high-efficiency cells is non-radiative recombination at defects on surfaces, grain boundaries and interfaces. One study estimated a loss of 80 meV per interface. This article treats passivation (rendering defects inactive) as a problem of coordination chemistry and interface design, and uses calculation to check how luminescence efficiency relates to voltage.
- What passivation is, in three points
- Where the defects are: on surfaces rather than inside grains
- Our calculation: how much surface is there in a 1 m² cell?
- What eats the voltage: non-radiative recombination and luminescence efficiency
- Our calculation: how many millivolts does a tenfold rise in luminescence efficiency add?
- The chemistry of passivation: filling in the missing bonds
- A materials engineer's view (1): passivation is close to designing a surface-treatment agent
- Interfaces are the main battleground: breaking down the 80 meV
- A materials engineer's view (2): damp-heat resistance, not just efficiency, is decided at the interface
- Strengths, weaknesses and open issues
- Glossary / References / Claim-to-source audit
Sourced = stated in published material or a peer-reviewed paper (link given)
Our calculation = a value this article derived from assumptions it states
Not yet confirmed = a plan or target with no confirmed track record
Structural readings and materials-design interpretations are marked separately as Commentary.
1. What passivation is, in three points
- What it does: a treatment that binds molecules or ions to defects on crystal surfaces and grain boundaries (such as atoms missing a bonding partner), so that they no longer trap electrons and holes (our commentary)
- What it improves: Noel and colleagues reported that treating the surface with thiophene or pyridine (Lewis bases) greatly reduced non-radiative recombination, extending the photoluminescence lifetime by about an order of magnitude to 2 µs, and raised cell efficiency from 13% to 15.3% and 16.5%Sourced
- Where it stands: Jiang and colleagues treated the surface with phenethylammonium iodide (PEAI) and obtained a certified efficiency of 23.32% and an open-circuit voltage of 1.18 V at an absorption edge of 1.53 eV (94.4% of the Shockley–Queisser open-circuit voltage limit of 1.25 V)Sourced
The effect of passivation can be measured as luminescence, and it translates directly into voltage. The closer a solar cell is to ideal, the more brightly it glows. This is a thermodynamic relationship, and a tenfold rise in luminescence efficiency raises the open-circuit voltage by about 59 mVOur calculation. Passivation is interface chemistry that stops energy that should come out as light from being thrown away as heat (our commentary).
2. Where the defects are: on surfaces rather than inside grains
As described in our explainer on perovskite basics, Yin and colleagues showed from first-principles calculations that the dominant intrinsic defects in MAPbI3 create only shallow levelsSourced. The inside of a grain is relatively “defect-tolerant”. So where are the losses?
- Grain boundaries: combining confocal fluorescence microscopy with electron microscopy, deQuilettes and colleagues showed that photoluminescence intensity and lifetime vary from grain to grain within the same film, and that grain boundaries are dark, with faster non-radiative decay. They also reported that pyridine treatment made previously dark grains light upSourced
- Surfaces and grain boundaries: Shao and colleagues showed that trap states on surfaces and grain boundaries cause photocurrent hysteresis, and that depositing fullerene on the perovskite cut the trap density by two orders of magnitude and doubled the efficiencySourced
- Interfaces with the charge-transport layers: using luminescence imaging to locate the losses, Stolterfoht and colleagues estimated that recombination at the interfaces adds a free-energy loss of 80 meV per interfaceSourced (Section 8)
3. Our calculation: how much surface is there in a 1 m² cell?
- Assumption: absorber thickness t of 400 nm (the value METI uses in its calculations)
- Assumption: grains are square columns of side d running through the film. Grain-boundary area per m² of film is then 2t ÷ d (each side face is shared by two grains)
- Assumption: the interfaces with the upper and lower charge-transport layers total 2 m²
| Grain size d | Grain-boundary area (m² per m² of film) | Total including top and bottom interfaces |
|---|---|---|
| 200 nm | 4.0 | 6.0 m² |
| 500 nm | 1.6 | 3.6 m² |
| 1 µm | 0.8 | 2.8 m² |
| 2 µm | 0.4 | 2.4 m² |
All of these are Our calculation. In other words, a 1 m² cell contains 2.4 to 6 m² of “surface” that needs treating. Assumptions and limits: real grains are not necessarily columnar, and surface roughness makes the true interface area larger still. Grain size also varies widely with processing conditions.
4. What eats the voltage: non-radiative recombination and luminescence efficiency
In 1967 Ross showed that the chemical potential difference obtainable in a photochemical system (in a solar cell, the counterpart of the open-circuit voltage) can be evaluated if the luminescence quantum yield is knownSourced. In 2007 Rau derived a relationship linking a solar cell's open-circuit voltage to its electroluminescence quantum efficiencySourced.
Put simply, the fraction of excited electrons and holes that recombine by emitting light, rather than by giving off heat (non-radiatively), sets the voltage. In an ideal solar cell every recombination event produces light, and the open-circuit voltage reaches its theoretical ceiling (the radiative limit). The more non-radiative recombination there is, the lower the luminescence efficiency, and the voltage falls accordingly (our commentary).
5. Our calculation: how many millivolts does a tenfold rise in luminescence efficiency add?
Following the Ross and Rau relationship, take Voc ≈ Voc,rad + (kT/q) × ln(ηext), where Voc,rad is the open-circuit voltage at the radiative limit and ηext is the external luminescence efficiency.
- Assumption: at 25 °C (298.15 K), kT/q = 25.7 mV
- If ηext rises tenfold: 25.7 mV × ln 10 = about 59 mV more
- Applied to Jiang and colleagues' cell: taking the radiative limit as the SQ limit of 1.25 V that they give, the 70 mV gap to 1.18 V corresponds to ηext = exp(−0.070 ÷ 0.0257) = about 6.6%
All of these are Our calculation. Assumptions and limits: the real radiative limit depends on the shape of the absorption edge and on light outcoupling, and does not equal the SQ-limit value. The 6.6% or so is a guide on that assumption, not a luminescence efficiency measured by Jiang and colleagues.
6. The chemistry of passivation: filling in the missing bonds
On crystal surfaces and grain boundaries, atoms have lost their usual partners and are under-coordinated. Arranging the treatments reported in the literature by which defect they target and what chemistry they use gives the following.
| Target defect | Treatment (example) | How it works | Reported result | Source |
|---|---|---|---|---|
| Under-coordinated Pb | Thiophene, pyridine | Electronically deactivated by a Lewis base (an electron-pair donor) | Photoluminescence lifetime extended by about an order of magnitude to 2 µs. Efficiency from 13% to 15.3% and 16.5% | Noel et al. |
| Under-coordinated iodide ions | Molecules that form halogen bonds | Complexation through supramolecular halogen bonding | Less charge accumulation and recombination loss from surface traps. 15.7%, and a stabilised output above 15% at 0.81 V | Abate et al. |
| Surface and grain-boundary traps | Fullerene (deposited on top) | Deactivates trap states | Trap density down by two orders of magnitude, efficiency doubled, hysteresis eliminated | Shao et al. |
| Surface defects | Phenethylammonium iodide (PEAI) | Forms an organic halide salt on the surface | Suppressed non-radiative recombination. Certified 23.32%, open-circuit voltage 1.18 V (at 1.53 eV) | Jiang et al. |
| Surface on the electron-transport side | Oleylammonium iodide | Forms a 2D perovskite layer of tailored dimensionality at room temperature | 24.3%, retaining over 95% of initial efficiency after more than 1,000 hours at 85 °C and 85% RH | Azmi et al. |
| Interfaces with transport layers | An ultrathin interlayer | Suppresses minority-carrier recombination at the interface | Interface losses on both sides greatly reduced. Certified 19.83% at 1 cm², open-circuit voltage 1.17 V | Stolterfoht et al. |
All Sourced (Noel et al. [Ref. 2], Abate et al. [Ref. 3], Shao et al. [Ref. 4], Jiang et al. [Ref. 6], Azmi et al. [Ref. 8], Stolterfoht et al. [Ref. 7]). These reports differ in composition, structure and date; the efficiencies are not meant to be compared with each other.
7. A materials engineer's view (1): passivation is close to designing a surface-treatment agent
Translate the table in Section 6 into materials terms and passivation starts to look like the same kind of problem as designing surface-treatment agents, coupling agents or dispersants (our commentary).
- Choice of functional group: Lewis bases containing nitrogen or sulfur (pyridine, thiophene) for under-coordinated PbSourced, halogen-bond donors for under-coordinated iodide ionsSourced — chosen according to the acid–base character of the target site
- Designing the tail: once bound to the surface, the phenethyl group of PEAI or the long alkyl chain of oleylammonium can form a separate layer (a 2D phase, for instance) and also act as a hydrophobic barrier (our commentary)
- Layer thickness: a passivation layer that is too thick gets in the way of charge extraction. Stolterfoht and colleagues used an “ultrathin” interlayerSourced. Controlling an insulating material to a thickness at which it works yet still lets charge through is a question of coating and deposition process technology (our commentary)
As deQuilettes and colleagues showed, the change whereby pyridine makes dark grains light upSourced can be checked by eye with photoluminescence imaging. By the relationship in Section 5, brighter emission translates directly into higher voltage. In other words, passivation materials can be screened first on the luminescence of a film, before any cell is built (our commentary). For materials suppliers, this means the entry point of evaluation can move forward from device efficiency to film luminescence.
Transport layers made from molecules that self-assemble on the surface (SAMs) are covered in our explainer on self-assembled monolayers.
8. Interfaces are the main battleground: breaking down the 80 meV
Working with planar pin cells (with dopant-free organic transport layers), Stolterfoht and colleagues used transient and absolute photoluminescence imaging to make every non-radiative recombination pathway visibleSourced.
- The perovskite film itself (the bulk) loses 135 meV of quasi-Fermi level splittingSourced
- Recombination at the interfaces adds a free-energy loss of 80 meV per interfaceSourced
- As a result, the open-circuit voltage of the whole cell was limited to about 1.12 VSourced
- Inserting ultrathin interlayers between the perovskite and the transport layers greatly reduced interface losses on both the p and n sides, giving over 20% at 1 cm² (certified 19.83%), an open-circuit voltage of 1.17 V and a fill factor above 81%Sourced
- From the relationship in Section 5, a voltage loss ΔV corresponds to luminescence efficiency falling by a factor of exp(ΔV ÷ 25.7 mV)
- 80 meV at one interface: exp(80 ÷ 25.7) = about 22 times. In other words, one interface is equivalent to cutting luminescence efficiency to about 1/22
- 135 meV in the bulk: exp(135 ÷ 25.7) = equivalent to about 190 times
All of these are Our calculation. Treating a single interface can change luminescence efficiency by more than an order of magnitude — the numbers show why passivation work has concentrated on interfaces (our commentary).
9. A materials engineer's view (2): damp-heat resistance, not just efficiency, is decided at the interface
Azmi and colleagues describe the damp-heat test at 85 °C and 85% relative humidity as the standard for verifying the stability of commercial solar modules, and say that passing it had long been an obstacle for perovskitesSourced.
They then passivated the surface on the electron-transport side with a 2D perovskite layer of tailored dimensionality formed at room temperature from oleylammonium iodide, reporting 24.3% in an inverted cell and retention of over 95% of initial efficiency after more than 1,000 hours under damp-heat conditionsSourced.
The implication is that a passivation layer not only cuts electrical losses but can also be the front line against heat and moisture (our commentary). As discussed in our explainer on ion migration and degradation, much perovskite degradation starts with ions moving and escaping. Under-coordinated surface sites may also be starting points for decomposition — seen that way, designing molecules to fill defects becomes design for both efficiency and durability (our commentary).
Passing the damp-heat test, on the other hand, does not in itself guarantee long-term reliability outdoors. How to evaluate that is covered in our explainer on outdoor testing and durability assessment.
10. Strengths, weaknesses and open issues
| Aspect | Confirmed in primary sources | Remaining issue |
|---|---|---|
| Voltage | PEAI treatment reached 94.4% of the SQ open-circuit voltage limit (Jiang et al.) | The remaining few tens of millivolts correspond to roughly an order of magnitude in luminescence efficiencyOur calculation |
| Interfaces | Ultrathin interlayers greatly reduced interface losses on both sides (Stolterfoht et al.) | Too thick a layer can hinder charge extraction (our commentary) |
| Grain boundaries | Pyridine treatment makes dark grains brighter (deQuilettes et al.) | Grain-to-grain variation remains (deQuilettes et al.) |
| Hysteresis | Fullerene cut trap density by two orders of magnitude and eliminated hysteresis (Shao et al.) | — |
| Durability | A 2D layer retained over 95% after more than 1,000 hours of damp heat (Azmi et al.) | Long-term outdoor reliability needs separate evaluation |
Each item is Sourced (source in brackets), except that “roughly an order of magnitude in luminescence efficiency” is our calculation and “too thick a layer can hinder” is our commentary. “—” marks items for which this article identifies no remaining issue.
(1) Long-term stability of the passivation layer itself
Whether molecules or 2D layers bound to the surface stay in the same state for years under light, heat and electric fields could not be established from generalisable primary sources within the scope of this articleNot yet confirmed.
(2) Uniformity over large areas
Most of the reports presented here are results on cells from small areas up to 1 cm². Whether a very thin treatment layer can be formed uniformly at module scale remains an issue for coating and deposition processes. See our explainers on solution coating and on roll-to-roll processing and scale-up.
- Losses occur on surfaces rather than inside grains — grain boundaries are dark, and one interface costs 80 meVSourced
- A 1 m² cell contains 2.4 to 6 m² of surface to treatOur calculation
- A tenfold rise in luminescence efficiency raises the open-circuit voltage by about 59 mVOur calculation
- Lewis bases for under-coordinated Pb, halogen bonds for under-coordinated I — the functional group is chosen by the target siteSourced
- PEAI treatment reached 94.4% of the SQ open-circuit voltage limitSourced
- A passivation layer is an electrical layer and also a damp-heat barrier (our commentary)
11. Glossary
- Passivation
- Binding molecules or ions to defects on surfaces and grain boundaries so that they no longer trap charge.
- Non-radiative recombination
- Recombination of an electron and hole that releases heat instead of light. It costs voltage.
- Under-coordination
- The state, for example at a crystal surface, in which an atom has fewer bonding partners than it normally would.
- Lewis base
- A molecule able to donate an electron pair, such as the nitrogen of pyridine or the sulfur of thiophene.
- Halogen bond
- A non-covalent interaction in which a halogen atom is attracted to an electron-rich partner.
- Trap state
- An energy level within the bandgap that captures electrons or holes.
- External luminescence efficiency
- The fraction of recombination that emerges from the device as light. The higher it is, the closer the open-circuit voltage is to ideal.
- Quasi-Fermi level splitting
- The difference between the chemical potentials of electrons and holes under illumination. It sets the upper limit of the open-circuit voltage.
- Radiative limit
- The open-circuit voltage under ideal conditions in which all recombination produces light.
- 2D perovskite
- A layered structure in which large organic cations separate the inorganic layers. Used as a protective surface layer.
- Damp-heat test
- An accelerated ageing test holding samples at 85 °C and 85% relative humidity for a long time.
- PL imaging
- Photographing the luminescence of a film or cell to locate losses position by position.
12. References (primary sources)
- Yin, Shi, Yan “Unusual defect physics in CH3NH3PbI3 perovskite solar cell absorber”, Appl. Phys. Lett. 104, 063903 (2014) https://doi.org/10.1063/1.4864778
- Noel et al. “Enhanced Photoluminescence and Solar Cell Performance via Lewis Base Passivation of Organic–Inorganic Lead Halide Perovskites”, ACS Nano 8, 9815 (2014) https://doi.org/10.1021/nn5036476
- Abate et al. “Supramolecular Halogen Bond Passivation of Organic–Inorganic Halide Perovskite Solar Cells”, Nano Lett. 14, 3247 (2014) https://doi.org/10.1021/nl500627x
- Shao et al. “Origin and elimination of photocurrent hysteresis by fullerene passivation in CH3NH3PbI3 planar heterojunction solar cells”, Nat. Commun. 5, 5784 (2014) https://doi.org/10.1038/ncomms6784
- deQuilettes et al. “Impact of microstructure on local carrier lifetime in perovskite solar cells”, Science 348, 683 (2015) https://doi.org/10.1126/science.aaa5333
- Jiang et al. “Surface passivation of perovskite film for efficient solar cells”, Nat. Photonics 13, 460 (2019) https://doi.org/10.1038/s41566-019-0398-2
- Stolterfoht et al. “Visualization and suppression of interfacial recombination for high-efficiency large-area pin perovskite solar cells”, Nat. Energy 3, 847 (2018) https://doi.org/10.1038/s41560-018-0219-8
- Azmi et al. “Damp heat–stable perovskite solar cells with tailored-dimensionality 2D/3D heterojunctions”, Science 376, 73 (2022) https://doi.org/10.1126/science.abm5784
- Ross “Some Thermodynamics of Photochemical Systems”, J. Chem. Phys. 46, 4590 (1967) https://doi.org/10.1063/1.1840606
- Rau “Reciprocity relation between photovoltaic quantum efficiency and electroluminescent emission of solar cells”, Phys. Rev. B 76, 085303 (2007) https://doi.org/10.1103/PhysRevB.76.085303
- Ministry of Economy, Trade and Industry (METI) “Next-Generation Solar Cell Strategy”, November 2024 (PDF, in Japanese) https://www.meti.go.jp/shingikai/energy_environment/perovskite_solar_cell/pdf/20241128_1.pdf
13. Claim-to-source audit
| Claim in the text | Basis | Label |
|---|---|---|
| That the dominant intrinsic defects in MAPbI3 create only shallow levels | Appl. Phys. Lett. paper (abstract). Reference 1 https://doi.org/10.1063/1.4864778 | Sourced |
| That surface treatment with thiophene or pyridine (Lewis bases) greatly reduced non-radiative recombination, extending the photoluminescence lifetime by about an order of magnitude to 2 µs. That this was proposed to be due to electronic passivation of under-coordinated Pb. That efficiency rose from 13% to 15.3% (thiophene) and 16.5% (pyridine) | ACS Nano paper (abstract). Reference 2 https://doi.org/10.1021/nn5036476 | Sourced |
| That surface trap states cause charge accumulation and recombination losses. That under-coordinated iodide ions are responsible and were passivated by complexation through supramolecular halogen bonding. A maximum efficiency of 15.7% and a stabilised output above 15% at a constant bias of 0.81 V | Nano Lett. paper (abstract). Reference 3 https://doi.org/10.1021/nl500627x | Sourced |
| That trap states on surfaces and grain boundaries cause photocurrent hysteresis. That fullerene on the perovskite lowered trap density by two orders of magnitude, doubled efficiency and eliminated hysteresis | Nat. Commun. paper (abstract). Reference 4 https://doi.org/10.1038/ncomms6784 | Sourced |
| That correlated confocal fluorescence and electron microscopy showed photoluminescence intensity and lifetime varying from grain to grain within one film, with dark grain boundaries showing faster non-radiative decay. That pyridine treatment made previously dark grains light up | Science paper (abstract). Reference 5 https://doi.org/10.1126/science.aaa5333 | Sourced |
| That PEAI was used for surface passivation of mixed FA–MA perovskite films, reducing defects and suppressing non-radiative recombination. A certified efficiency of 23.32% (quasi-steady state). An open-circuit voltage of 1.18 V at an absorption edge of 1.53 eV, 94.4% of the SQ open-circuit voltage limit of 1.25 V | Nat. Photonics paper (abstract). Reference 6 https://doi.org/10.1038/s41566-019-0398-2 | Sourced |
| That transient and absolute photoluminescence imaging visualised the non-radiative recombination pathways in planar pin cells. A loss of 135 meV of quasi-Fermi level splitting in the bulk, an additional 80 meV per interface, and an open-circuit voltage limited to about 1.12 V. That ultrathin interlayers greatly reduced interface losses on both the p and n sides, giving over 20% at 1 cm² (certified 19.83%), an open-circuit voltage of 1.17 V and a fill factor above 81% | Nat. Energy paper (abstract). Reference 7 https://doi.org/10.1038/s41560-018-0219-8 | Sourced |
| That the damp-heat test (85 °C, 85% RH) is the standard for verifying the stability of commercial modules and that passing it had long been an obstacle for perovskites. That the electron-selective surface was passivated with a 2D layer of tailored dimensionality formed at room temperature from oleylammonium iodide. 24.3% in an inverted cell, and over 95% of initial efficiency retained after more than 1,000 hours of damp heat | Science paper (abstract). Reference 8 https://doi.org/10.1126/science.abm5784 | Sourced |
| That the potential difference between the excited and ground states of a photochemical system can be evaluated if the luminescence quantum yield is known | J. Chem. Phys. paper (abstract). Reference 9 https://doi.org/10.1063/1.1840606 | Sourced |
| That a relationship linking a solar cell's open-circuit voltage to its electroluminescence quantum efficiency was derived | Phys. Rev. B paper (abstract). Reference 10 https://doi.org/10.1103/PhysRevB.76.085303 | Sourced |
| The absorber thickness of 400 nm used in the calculation | METI, Next-Generation Solar Cell Strategy, page 52. Reference 11 https://www.meti.go.jp/shingikai/energy_environment/perovskite_solar_cell/pdf/20241128_1.pdf | Sourced |
| The areas of grain boundaries and interfaces (totals of 6.0 m² for 200 nm grains, 3.6 m² for 500 nm, 2.8 m² for 1 µm and 2.4 m² for 2 µm). kT/q = 25.7 mV and about 59 mV per tenfold rise in luminescence efficiency. η of about 6.6% when the radiative limit is taken as 1.25 V. Converting 80 meV to a luminescence efficiency factor of about 1/22 and 135 meV to about 1/190. The 295 meV total loss and the open-circuit voltage difference of about +50 mV | Our calculation. Columnar grains and flat interfaces, 25 °C, a simplified form of the Ross and Rau relationship, and the SQ limit as a proxy for the radiative limit are assumptions set by this article. Includes the numerical parts of Figs. 2, 3 and 5 | Our calculation |
| Long-term stability of the passivation layer itself (whether it keeps its state for years under light, heat and electric fields) | No generalisable primary source could be confirmed within the scope of this article (our note) | Not yet confirmed |
| Framing passivation as coordination chemistry and surface-treatment-agent design. The left–right pairing by Lewis acid and base. That tails can act as hydrophobic barriers, and that too thick a layer can hinder extraction. The reading that materials can be screened first on film luminescence. The reading that under-coordinated surface sites may also be starting points for decomposition. The issue of uniformity over large areas. The sorting into strengths and weaknesses | Our summary and commentary based on published content. Not views expressed by the authors of the papers | Commentary |
| That Figs. 1 to 5 are explanatory drawings, not real cross-sectional images, measured data or molecular structures. That the hero image and Fig. 6 are AI-generated images | Our note | Commentary |
Last updated 25 September 2026. Sources are limited to primary material (peer-reviewed papers and a government strategy document). Because the article includes materials-design interpretations, those are marked as Commentary and kept separate from sourced fact. The long-term stability of the passivation layer itself and uniformity at module scale are not stated, because no generalisable primary source could be confirmed. The calculations in Sections 3, 5 and 8 are guides based on the simplifications stated. The content of papers is limited to what could be confirmed in the abstracts. All figures are explanatory concept graphics. Figs. 1 to 5 are vector drawings; the hero image and Fig. 6 are AI-generated images, and none of them shows a real cross-sectional image, luminescence image or physical product.