MENU
Advanced Materials Intelligence
  1. Home
  2. Ion Migration and Degradation Explained | Perovskite Solar Cells

Ion Migration and Degradation Explained | Perovskite Solar Cells

Back to Technology, Institution & Company Guides

TECHNOLOGY EXPLAINER

Ion Migration and Degradation
— iodine is walking around inside the semiconductor

A perovskite is a mixed conductor, in which ions move as well as electrons. Calculations and measurements indicate that iodide ions hop between vacancies with an activation energy of about 0.6 eV. This article sets out, from the materials side, how that ease of ion movement leads to hysteresis, halide segregation, and decomposition by heat and moisture.

Built from primary sources: peer-reviewed papers (Nature Communications, Energy & Environmental Science, Chemistry of Materials and others) and material from METI / Last updated September 2026

Conceptual image of a uniform dark brown thin film on a dark surface, sitting under a transparent cover plate with a few small water droplets resting only on the top of the cover
AI-generated concept image. An impression of keeping outside moisture from reaching the film. It does not represent any real product, encapsulation structure or state of degradation.
What this article covers
  1. Ion migration and degradation, in three points
  2. What moves: iodide ions hopping through vacancies
  3. Our calculation: how far does an iodide ion wander through the film?
  4. Hysteresis: the efficiency depends on how you measure it
  5. Halides that separate under light
  6. Heat, moisture and electrodes: three routes to decomposition
  7. A materials engineer's view (1): design on the premise of a “soft crystal”
  8. Our calculation: how much water does it take to hydrate 1 m² of film?
  9. A materials engineer's view (2): translating barrier performance into days
  10. Strengths, weaknesses and open issues
  11. Glossary / References / Claim-to-source audit
How claims are labelled in this article

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. Ion migration and degradation, in three points

  • Ions move: Eames and colleagues reported that in CH3NH3PbI3 iodide ions migrate readily via vacancies, with an activation energy of about 0.6 eV, and that measurement and calculation agreed well. They describe halide perovskites as mixed ionic–electronic conductorsSourced
  • The efficiency depends on how it is measured: Snaith and colleagues drew attention to anomalous hysteresis in current–voltage curves and proposed that stabilised power output under operating conditions be reported alongside efficiencies from voltage sweepsSourced
  • It decomposes easily: Conings and colleagues showed that MAPbI3 decomposes substantially on annealing at 85 °C even in an inert atmosphere, which they say supports the view of the material as “soft matter” with a low formation energySourced
The single most important line in this article

Much of perovskite degradation branches out from a single property: the ions can move. Under an electric field they pile up at interfaces (hysteresis); under light the composition separates (segregation); under heat or moisture they leave the lattice (decomposition). So the countermeasures also run in three directions: make the ions harder to move, let them return if they do move, and keep outside stimuli out (our commentary).

Which stimulus triggers which phenomenon (our summary) Left = outside stimulus / centre = what happens in the material / right = what shows up in the device Bias (in use, testing) Light Heat Moisture Iodide vacancies drift to interfaces Halides separate Lattice breaks down, gases escape Electrode metal can migrate too Hydrates form Liquid water gives PbI2 J–V hysteresis Voc hits a ceiling Irreversible power loss Current collapses May recover on drying Note: rows follow Eames et al. [Ref. 1], Hoke et al. [Ref. 4], Refs. 5 to 7 (heat, electrodes) and Leguy et al. [Ref. 8]. Note: the split into four rows is our framing. In reality several stimuli act at once and influence one another.
Fig. 1 Concept diagram (vector drawing). Each row follows the reports in the papers [Refs. 1 and 4 to 8]. Pairing each stimulus with one phenomenon is this article's own framing; in practice several stimuli act together.

2. What moves: iodide ions hopping through vacancies

Using first-principles calculations, Eames and colleagues derived activation energies for three kinds of ion migrating via vacancies in CH3NH3PbI3Sourced.

What moves (vacancy)Calculated activation energyThe paper's assessment
Iodide ion (I⁻)0.58 eVThe lowest; vacancy-mediated diffusion occurs readily. Agrees well with the 0.60 to 0.68 eV activation energy measured from hysteresis
Methylammonium (MA⁺)0.84 eVHas to pass through a “narrow window” formed by four iodide ions
Lead (Pb²⁺)2.31 eVSignificant migration is unlikely on the temperatures and timescales of measurements

All Sourced (Eames et al. [Ref. 1], Table 1 and text). Assuming an attempt frequency of 10¹² Hz at 320 K, the same paper estimates diffusion coefficients of about 10⁻¹² cm²/s for I⁻ and about 10⁻¹⁶ cm²/s for MA⁺.

The paper also cites an earlier prediction that Schottky-type vacancies (sets of I⁻, Pb²⁺ and MA⁺ vacancies) form easily, with vacancy concentrations above 0.4% at room temperatureSourced. In other words, the lattice comes with empty seats for ions to move into from the start (our commentary).

Iodide moves most easily: activation energy and the effect of temperature Left = the paper's calculated values (eV) / right = speed-up of migration from 25 °C to 85 °C (our calculation) I⁻ vacancy 0.58 eV MA⁺ vacancy 0.84 eV Pb²⁺ vacancy 2.31 eV Speed-up from 25 °C to 85 °C Ratio of Arrhenius terms exp(−Ea/kT) about 44× for Ea = 0.58 eV (I⁻) about 240× for Ea = 0.84 eV (MA⁺) Note: activation energies calculated by Eames et al. [Ref. 1]. Bar length is proportional to eV (1 eV = 100 px). Note: ratios are our calculation, assuming a temperature-independent prefactor. They do not show how decomposition speeds up. Note: 85 °C was chosen because Conings et al. [Ref. 5] reported decomposition there. It is not a test-standard condition.
Fig. 2 Drawing that includes our calculation (vector drawing). The activation energies are calculated values from Eames et al. [Ref. 1]. The ratios on the right (about 44 times and about 240 times) were calculated by this article with the Arrhenius equation and are not published figures. This is a simple estimate that treats the pre-exponential factor as constant.

3. Our calculation: how far does an iodide ion wander through the film?

Our calculation: diffusion length L ≈ √(D × t)
  • Assumption: for the diffusion coefficient D, use the I⁻ value estimated by Eames and colleagues at 320 K (about 47 °C): 1×10⁻¹² cm²/s
  • Assumption: take the film thickness as 400 nm, the absorber thickness METI uses in its own calculations
  • 10 seconds: √(10⁻¹² × 10) = 3.2×10⁻⁶ cm = about 32 nm (Eames and colleagues also give about 30 nmSourced)
  • 1 hour: √(10⁻¹² × 3,600) = about 600 nm
  • Time to cross a 400 nm film: t ≈ L² ÷ D = (4×10⁻⁵)² ÷ 10⁻¹² = 1,600 s = about 27 minutes

All of these are Our calculation. Assumptions and limits: D is an estimate from calculation, and in real films it can change by orders of magnitude with grain boundaries, composition and temperature. Under an electric field ions move faster than by diffusion (Eames and colleagues estimate a drift of about 80 nm in 10 seconds with 1 V across 500 nmSourced). The calculation exists to show the order of magnitude: ions can cross the film thickness on a timescale of minutes to hours.

4. Hysteresis: the efficiency depends on how you measure it

A solar cell's efficiency is obtained from a “J–V curve”, measuring current while sweeping the voltage. In perovskites the curve can differ depending on whether the voltage is swept up or down. Snaith and colleagues called this “anomalous hysteresis” and proposed that stabilised power output at the operating point be reported in addition to efficiencies from sweepsSourced. Eames and colleagues showed that the activation energy for relaxation of this hysteresis is 0.60 to 0.68 eV, close to the calculated 0.58 eV for iodide vacancy migrationSourced.

The curve shifts with scan direction: J–V hysteresis (schematic) Vertical = current density, horizontal = voltage. No values are given Voltage → Current density Reverse scan Forward scan Why it shifts (one hypothesis) 1 Under bias, ions (vacancies) drift to the interfaces and pile up 2 The stored charge alters the internal field and charge extraction 3 Ions move in seconds to minutes, so scan speed and direction matter Note: hysteresis and stabilised output from Snaith et al. [Ref. 2]; the link to ion migration from Eames et al. [Ref. 1]. Note: curve shapes and the direction and size of the offset are schematic and vary with device structure and scan conditions.
Fig. 3 Concept diagram (vector drawing). The existence of hysteresis and the proposal to report stabilised output follow Snaith et al. [Ref. 2]; the link to ion migration follows Eames et al. [Ref. 1]. The curves are schematic, not measured data. The direction and size of the offset differ from device to device. Snaith and colleagues put forward three hypotheses for the cause; the box on the right is this article's summary of the explanation based on ion migration.

Pitfalls in measuring efficiency are covered in our explainer on conversion efficiency and performance metrics.

5. Halides that separate under light

In films that mix iodine and bromine, light alone is enough to separate the composition. Hoke and colleagues reported that illuminating MAPb(BrxI1−x)3 with 0.2 < x < 1 produced an emission peak at about 1.68 eV within a minute, forming minority iodide-rich regions (x ≈ 0.2), and that the films returned to their original state after a few minutes in the darkSourced. The authors propose that photogenerated holes gathering in the iodide-rich regions could be the driving force for segregationSourced. The link with composition design is covered in detail in our explainer on bandgap and composition design.

6. Heat, moisture and electrodes: three routes to decomposition

RouteWhat has been reportedSource
HeatMAPbI3 decomposes substantially on annealing at 85 °C even in an inert atmosphere. Supports the view of it as “soft matter” with a low formation energyConings et al.
HeatGases released by thermal decomposition of MAPbI3 were analysed by coupled thermogravimetry–mass spectrometry, and decomposition into NH3 and CH3I was observedJuarez-Perez et al. (Okinawa Institute of Science and Technology, OIST)
Heat (electrode)At 70 °C, gold from the gold electrode migrates through the hole-transport layer (spiro-MeOTAD) into the perovskite and severely impairs performance. Decomposition of the perovskite layer was not the main cause. A Cr interlayer avoided the irreversible lossDomanski et al.
Moisture (vapour)Water vapour forms the monohydrate CH3NH3PbI3·H2O, and over longer exposure (CH3NH3)4PbI6·2H2O. Drying reverses this completely. On hydration the film turns from dark brown to transparent (the monohydrate has a bandgap of 3.1 eV)Leguy et al.
Moisture (liquid)Liquid water causes irreversible decomposition to PbI2Leguy et al.
Moisture (cell)Exposing unencapsulated cells to water vapour cut the short-circuit current by more than 90% and the open-circuit voltage by about 200 mV. Six hours in dry nitrogen brought full recovery, but hysteresis was much larger afterwardsLeguy et al.

All Sourced (Conings et al. [Ref. 5], Juarez-Perez et al. [Ref. 6], Domanski et al. [Ref. 7], Leguy et al. [Ref. 8]).

Two points in Leguy and colleagues' results stand out: the monohydrate forms regardless of depth in the film, suggesting that water molecules are carried rapidly along grain boundaries, and their proposal that irreversible decomposition by water vapour only accelerates once an entire grain has turned into monohydrateSourced. In other words, hydration itself may be reversible, but past some threshold there is no way back (our commentary).

7. A materials engineer's view (1): design on the premise of a “soft crystal”

A materials engineer's view: treat degradation as the material's character, not a defect

Line up the reports so far and the common thread is that the ions and molecules that make up the lattice move or escape at low energies. Iodide vacancy migration at 0.58 eVSourced; decomposition at 85 °C even in an inert atmosphereSourced; hydrates forming in water vapour at room temperatureSourced — by the standards of oxide ceramics or silicon, all of this is orders of magnitude “softer” behaviour (our commentary).

On that premise, the design questions can be grouped as follows (our commentary).

  • Inside the material: reduce mobile components (such as volatile MA), make vacancies harder to form, treat grain boundaries
  • The neighbouring layers: stop mobile species entering from outside the perovskite (such as electrode metals). Domanski and colleagues' Cr interlayer is an exampleSourced
  • The outside: encapsulation and barriers that keep out moisture and oxygen
  • Evaluation: some degradation is reversible, so results depend on when a test is stopped and measured

The last point is a question of test methods. The ISOS consensus statement by Khenkin and colleagues identifies ion redistribution under electric fields and reversible degradation as properties peculiar to perovskites, and proposes test procedures and reporting items that take them into accountSourced. The statement also says these procedures are not intended to replace existing certification standardsSourced. Evaluation is covered in more detail in our explainer on outdoor testing and durability assessment.

8. Our calculation: how much water does it take to hydrate 1 m² of film?

To look at Leguy and colleagues' result from Section 6 (water vapour forms the monohydrate CH3NH3PbI3·H2O) from the encapsulation side, we calculate the amount of water needed to turn 1 m² of absorber entirely into monohydrate.

Our calculation: the water needed to form the monohydrate
  • Assumption: the absorber is MAPbI3, 400 nm thick, with a density of 3.9 g/cm³ (all values used in METI's own calculations). Per m²: 4×10⁻⁵ cm × 10⁴ cm² × 3.9 g/cm³ = 1.56 g
  • Assumption: formula weight of MAPbI3 620.0 g/mol (from C 12.011, H 1.008, N 14.007, Pb 207.2, I 126.904); water 18.015 g/mol
  • 1.56 g ÷ 620.0 g/mol = 2.52 mmol. The monohydrate is 1:1, so the water needed is also 2.52 mmol = about 45 mg

Result: the water needed to turn 1 m² of absorber entirely into monohydrate is about 45 mgOur calculation. Assumptions and limits: the further step to (CH3NH3)4PbI6·2H2O and decomposition by liquid water are not considered. This is a guide to the minimum amount of water consumed by hydration.

9. A materials engineer's view (2): translating barrier performance into days

The performance of encapsulants and barrier films is expressed as water vapour transmission rate (WVTR, g/m²/day). Dividing the roughly 45 mg from Section 8 by the WVTR gives the number of days it would take to reach the amount for 1 m², if all the water that permeates reached the absorber and all of it went into forming monohydrate.

Days until about 45 mg has arrived, by WVTR (our calculation) Log horizontal axis. Extreme assumption: all permeated water goes into hydrating the absorber 1 day1 year10 years 10⁻¹ about 11 hours 10⁻² about 4.5 days 10⁻³ about 45 days 10⁻⁴ about 1.2 years 10⁻⁵ about 12 years 10⁻⁶ about 124 years WVTR (g/m²/day) Note: about 45 mg/m² is our calculation (MAPbI3, 400 nm and 3.9 g/cm³ as used in METI's calculation [Ref. 9]). Note: in reality only part of the permeated water reaches the absorber, and hydration is reversible. Not a lifetime forecast. Note: monohydrate formation follows Leguy et al. [Ref. 8]. Bar length is proportional to the log of the number of days.
Fig. 4 Drawing that includes our calculation (vector drawing). Monohydrate formation follows Leguy et al. [Ref. 8]; the absorber thickness and density follow METI [Ref. 9]. All the day counts were calculated by this article and are not a prediction of the lifetime of an encapsulated module. The extreme assumption is that all permeated water goes into hydrating the absorber.
A materials engineer's view: the gap between 10⁻⁴ and 10⁻⁵ is the gap between one year and twelve

The calculation in Fig. 4 rests on an extreme assumption, but it conveys the orders of magnitudeOur calculation.

  • At a WVTR of 10⁻² g/m²/day, the roughly 45 mg arrives in about 4.5 days
  • At 10⁻⁴ it takes about 1.2 years, at 10⁻⁵ about 12 years, and at 10⁻⁶ about 124 years

That just 45 mg or so of water could in principle hydrate an entire square metre of absorber means that encapsulating perovskites calls for barrier performance of a different order from food or pharmaceutical packaging (our commentary). Leguy and colleagues also suggest that water molecules are carried rapidly along grain boundariesSourced, so it is not only the film surface but also the pathways inside the film that govern how moisture gets in.

On the other hand, there is also the finding that hydration is reversed by dryingSourced. The barrier requirement should therefore be framed not as “let not a single molecule in” but as “never let the threshold (the point at which a whole grain is hydrated) be crossed” (our commentary). Specific requirement values and encapsulation structures are covered in our explainer on encapsulation and barrier layers.

Conceptual image, from a low angle, of the edge of a uniform dark brown thin film sandwiched between a transparent cover plate and a base plate, with a narrow band of transparent sealant running around the film
Fig. 5 AI-generated concept image. An impression of the encapsulation idea of keeping moisture from entering at the edge of the film. It does not represent any real encapsulation structure, material or dimensions.

10. Strengths, weaknesses and open issues

AspectConfirmed in primary sourcesRemaining issue
Ion migrationMainly I⁻ vacancies, activation energy 0.58 eV (Eames et al.)Ions can cross the film thickness in minutes to hoursOur calculation
MeasurementReporting of stabilised output has been proposed (Snaith et al.)Efficiency varies with sweep conditions
LightSegregation is reversible in the dark (Hoke et al.)In operation, light falls on the cell continuously
HeatGold migration from the electrode avoided with a Cr interlayer (Domanski et al.)MAPbI3 decomposes at 85 °C even in an inert atmosphere (Conings et al.)
MoistureHydration by water vapour is reversible on drying (Leguy et al.)Liquid water gives PbI2 irreversibly (Leguy et al.)
EvaluationConsensus procedures based on ISOS have been proposed (Khenkin et al.)Results change with how reversible degradation is counted

Each item is Sourced (source in brackets), except “can cross the film thickness in minutes to hours”, which is our calculation. The sorting is this article's own framing.

(1) Materials of other compositions will not match the numbers in this article

Most of the numbers presented here are for MAPbI3. A different composition can change the activation energy, the decomposition temperature and the sensitivity to moisture (our commentary), so the numbers in this article cannot be applied directly to other compositions. This article does not compare compositions.

(2) Outdoor lifetime cannot be inferred from this article

The calculations in Sections 8 and 9 are extreme estimates meant to convey the order of magnitude of encapsulation requirements. Real lifetime is set by the combination of composition, encapsulation, installation environment, temperature cycling and more. This article gives no figure for lifetime in years.

The article in summary
  • A perovskite is a mixed ionic–electronic conductor, and the most mobile species is the iodide vacancy (0.58 eV)Sourced
  • From 25 °C to 85 °C, I⁻ migration speeds up about 44 times, and ions can cross a 400 nm film in about 27 minutesOur calculation
  • Hysteresis is a measurement issue, and at the same time a sign of ion migrationSourced
  • MAPbI3 decomposes at 85 °C even in an inert atmosphere, and in one case gold migrated from the electrode at 70 °CSourced
  • Hydration by water vapour is reversible; liquid water is notSourced
  • About 45 mg of water hydrates 1 m² of absorber. Only at a WVTR of 10⁻⁵ does that stretch to about 12 yearsOur calculation

11. Glossary

Mixed conductor
A material in which both electrons (or holes) and ions carry charge.
Vacancy
A defect in which a lattice site that should hold an ion is empty. Migration happens when a neighbouring ion hops into it.
Schottky defect
A defect made of a charge-balanced set of cation and anion vacancies.
Activation energy
The height of the energy barrier an ion must cross to hop to a neighbouring site.
Arrhenius equation
The relationship in which the rate of a reaction or diffusion rises with temperature in proportion to exp(−Ea/kT).
Diffusion coefficient (D)
A measure of how fast diffusion proceeds. The rough distance travelled in time t is √(D×t).
J–V hysteresis
The mismatch of current–voltage curves depending on the direction or speed of the voltage sweep.
Stabilised power output
The output that settles when the cell is held near its maximum-power voltage. A metric that avoids apparent differences caused by sweeping.
Monohydrate
A compound that takes up one water molecule per formula unit. For MAPbI3, CH3NH3PbI3·H2O.
WVTR
Water vapour transmission rate: the mass of water passing through 1 m² per day (g/m²/day).
spiro-MeOTAD
An organic hole-transport material widely used in perovskite solar cells.
ISOS
A set of test procedures for assessing the stability of organic solar cells, agreed at the International Summit on Organic Photovoltaic Stability and later extended to perovskites.

12. References (primary sources)

  1. Eames et al. “Ionic transport in hybrid lead iodide perovskite solar cells”, Nat. Commun. 6, 7497 (2015) https://doi.org/10.1038/ncomms8497
  2. Snaith et al. “Anomalous Hysteresis in Perovskite Solar Cells”, J. Phys. Chem. Lett. 5, 1511 (2014) https://doi.org/10.1021/jz500113x
  3. Khenkin et al. “Consensus statement for stability assessment and reporting for perovskite photovoltaics based on ISOS procedures”, Nat. Energy 5, 35 (2020) https://doi.org/10.1038/s41560-019-0529-5
  4. Hoke et al. “Reversible photo-induced trap formation in mixed-halide hybrid perovskites for photovoltaics”, Chem. Sci. 6, 613 (2015) https://doi.org/10.1039/C4SC03141E
  5. Conings et al. “Intrinsic Thermal Instability of Methylammonium Lead Trihalide Perovskite”, Adv. Energy Mater. 5, 1500477 (2015) https://doi.org/10.1002/aenm.201500477
  6. Juarez-Perez et al. (Okinawa Institute of Science and Technology) “Thermal degradation of CH3NH3PbI3 perovskite into NH3 and CH3I gases observed by coupled thermogravimetry–mass spectrometry analysis”, Energy Environ. Sci. 9, 3406 (2016) https://doi.org/10.1039/C6EE02016J
  7. Domanski et al. “Not All That Glitters Is Gold: Metal-Migration-Induced Degradation in Perovskite Solar Cells”, ACS Nano 10, 6306 (2016) https://doi.org/10.1021/acsnano.6b02613
  8. Leguy et al. “Reversible Hydration of CH3NH3PbI3 in Films, Single Crystals, and Solar Cells”, Chem. Mater. 27, 3397 (2015) https://doi.org/10.1021/acs.chemmater.5b00660
  9. 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 textBasisLabel
That iodide ions migrate readily via vacancies with an activation energy of about 0.6 eV, that measurement and calculation agree, and that halide perovskites are mixed conductors. Calculated vacancy migration energies (I⁻ 0.58 eV, MA⁺ 0.84 eV, Pb²⁺ 2.31 eV). The 0.60 to 0.68 eV activation energy measured from hysteresis. That MA⁺ passes through a window framed by four I⁻. Diffusion coefficients at 320 K with an attempt frequency of 10¹² Hz (I⁻ about 10⁻¹² cm²/s, MA⁺ about 10⁻¹⁶ cm²/s). A diffusion length of about 30 nm in about 10 seconds, and a drift of about 80 nm with 1 V across 500 nm. The cited prediction of Schottky vacancies above 0.4% at room temperatureNat. Commun. paper (abstract, text and Table 1). Reference 1 https://doi.org/10.1038/ncomms8497Sourced
The identification of anomalous hysteresis in current–voltage curves. The proposal to report stabilised power output under operating conditions alongside efficiencies from sweeps. That three hypotheses are put forward for the causeJ. Phys. Chem. Lett. paper (abstract). Reference 2 https://doi.org/10.1021/jz500113xSourced
Consensus procedures for stability assessment based on ISOS. That they identify ion redistribution under electric fields and reversible degradation as properties peculiar to perovskites and propose procedures and reporting items. That they are not intended to replace existing certification standardsNat. Energy paper (abstract). Reference 3 https://doi.org/10.1038/s41560-019-0529-5Sourced
That in MAPb(BrxI1−x)3 with 0.2 < x < 1 an emission peak at about 1.68 eV appears within a minute of illumination and iodide-rich regions (x ≈ 0.2) form. That the films recover after a few minutes in the dark. The hypothesis that holes gathering in iodide-rich regions could drive segregationChem. Sci. paper (full text). Reference 4 https://doi.org/10.1039/C4SC03141ESourced
That MAPbI3 decomposes substantially on annealing at 85 °C even in an inert atmosphere. That this supports the view of it as soft matter with a low formation energyAdv. Energy Mater. paper (abstract). Reference 5 https://doi.org/10.1002/aenm.201500477Sourced
That gases released by thermal decomposition of MAPbI3 were analysed by coupled thermogravimetry–mass spectrometry and decomposition into NH3 and CH3I was observedEnergy Environ. Sci. paper (title and abstract). Reference 6 https://doi.org/10.1039/C6EE02016JSourced
That at 70 °C gold migrates through spiro-MeOTAD into the perovskite and severely impairs performance. That decomposition of the perovskite layer was not the main cause of irreversible degradation. That a Cr interlayer avoided irreversible losses at high temperatureACS Nano paper (abstract). Reference 7 https://doi.org/10.1021/acsnano.6b02613Sourced
That water vapour forms CH3NH3PbI3·H2O, and over longer exposure (CH3NH3)4PbI6·2H2O, reversed completely by drying. That liquid water decomposes it irreversibly to PbI2. That hydration turns the film from dark brown to transparent, with a monohydrate bandgap of 3.1 eV. That the monohydrate forms regardless of depth, suggesting rapid transport of water along grain boundaries. That unencapsulated cells lost more than 90% of short-circuit current and about 200 mV of open-circuit voltage, recovered in 6 hours in dry nitrogen, and showed increased hysteresis afterwards. The proposal that irreversible decomposition accelerates only once whole grains have turned to monohydrateChem. Mater. paper (abstract). Reference 8 https://doi.org/10.1021/acs.chemmater.5b00660Sourced
That MAPbI3, a thickness of 400 nm and a density of 3.9 g/cm³ are used in calculating the absorberMETI, Next-Generation Solar Cell Strategy, page 52. Reference 9 https://www.meti.go.jp/shingikai/energy_environment/perovskite_solar_cell/pdf/20241128_1.pdfSourced
The speed-up of migration from 25 °C to 85 °C (about 44 times for Ea 0.58 eV, about 240 times for 0.84 eV). Diffusion lengths (about 32 nm in 10 seconds, about 600 nm in 1 hour) and the time of about 27 minutes to cross a 400 nm film. 1.56 g and 2.52 mmol per m² of absorber, about 45 mg of water to form the monohydrate, and the number of days for each WVTR (from about 11 hours at 10⁻¹ to about 124 years at 10⁻⁶)Our calculation. A constant pre-exponential factor, representing D by the estimate at 320 K, and all permeated water going into hydration are assumptions set by this article. Includes the numerical parts of Figs. 2 and 4Our calculation
The framing that degradation branches out from the ability of ions to move. The three directions for countermeasures. The table pairing stimuli with phenomena. The design questions on the premise of a “soft crystal”. Framing the barrier requirement as not letting the hydration threshold be crossed. The explanation in the right-hand box of the hysteresis figureOur summary and commentary based on published content. Not views expressed by the authors of the papersCommentary
Comparison of degradation across compositions, outdoor lifetime in years, and specific encapsulation requirement valuesNo generalisable primary source was confirmed within the scope of this article, so nothing is stated (our note)Commentary
That Figs. 1 to 4 are explanatory drawings, not measured data or micrographs. That the hero image and Fig. 5 are AI-generated imagesOur noteCommentary

Last updated 25 September 2026. Sources are limited to primary material (peer-reviewed papers and a government strategy document). Because the article includes a structured reading of degradation and interpretations of encapsulation requirements, those are marked as Commentary and kept separate from sourced fact. Comparisons of degradation across compositions, outdoor lifetime in years, and specific encapsulation requirement values are not given in this article. The calculations in Sections 3 and 8 are order-of-magnitude estimates, not lifetime predictions. The content of papers is limited to what could be confirmed in the abstract or the openly available full text. All figures are explanatory concept graphics. Figs. 1 to 4 are vector drawings; the hero image and Fig. 5 are AI-generated images, and none of them shows real measured data, a degraded sample or a physical product.

🌐 Japanese