TECHNOLOGY EXPLAINER
Conversion Efficiency and Performance Metrics
— before the percentage, read how it was measured
The efficiency of a perovskite solar cell changes with the way it is measured. Values shift depending on whether voltage is swept up or down (hysteresis), and in one inter-laboratory comparison the lab-to-lab spread for slow-responding cells was about ten times that of a silicon reference cell. This article sorts out PCE, Voc, Jsc, FF, EQE, MPPT, stabilised output and certified measurement from a materials engineer's point of view.
- What conversion efficiency is, in three points
- Reading the current-voltage curve: Voc, Jsc, FF and PCE
- Our calculation: multiply three numbers and you get the efficiency
- EQE (external quantum efficiency): a report card by wavelength
- Hysteresis: efficiency depends on scan direction
- A materials engineer's view (1): hysteresis is a sign of slow material response
- Stabilised output and MPPT: measuring the value once it settles
- Different labs, different numbers: results of an inter-comparison
- Certified measurement: who measures, and on what basis
- A materials engineer's view (2): how you measure area alone can shift efficiency
- Where the standards stand
- Open problems, and what this article could not confirm
- Glossary / References / Claim-to-source audit
Sourced = stated in published material, a peer-reviewed paper or the public summary of a standard (link given)
Our calculation = a value this article derived from assumptions it states
Not yet confirmed = a plan, target or outlook with no confirmed track record
Structural readings and interpretations for materials and evaluation are marked separately as Commentary.
Because the full texts of the standards are paid, this article limits itself to the titles and scope descriptions published on the IEC webstore.
1. What conversion efficiency is, in three points
- Definition: power conversion efficiency (PCE) is the maximum power a device can deliver divided by the power of the incoming light. The paper by Dunbar et al. describes efficiency as the ratio of the device's maximum power output to the input powerSourced
- Standardising the conditions: because solar cell performance changes with irradiance, temperature, spectrum and angle of incidence, the convention is to report under standard test conditions (STC): a device temperature of 25 °C, the reference spectrum known as AM1.5G, and a total irradiance of 1,000 W/m²Sourced
- What makes perovskites difficult: an IEC technical report (IEC TR 63228:2019) gives as reasons why perovskites and other emerging solar cells are hard to measure accurately with the IEC 60904 series, which was developed with silicon in mind, performance that changes over time, unusual spectral responsivity, small areas, transient responses to external stimuli, optical interference and non-linearity of current with irradiance, among othersSourced
When you read a perovskite efficiency, check before the number itself: cell or module, area, certified or not, and scanned or stabilised output. In a study where two accredited institutions, CSIRO and NREL, and eight laboratories measured the same cells, the efficiency spread for a slow-responding perovskite cell was about ten times that of a silicon reference cell, and the biggest cause was reported to be not the equipment but the choice of measurement methodSourced.
2. Reading the current-voltage curve: Voc, Jsc, FF and PCE
The central measure of solar cell performance is the current-voltage (I-V) curve under illumination. Dunbar et al. explain that this curve represents the current that flows, and the power that can be extracted, as a function of the potential difference (voltage) between the electrodesSourced. When plotted as current density (J), current divided by area, it is called a J-V curve.
| Metric | Meaning | Where materials mainly have an effect (our summary) |
|---|---|---|
| Jsc (short-circuit current density) | The current that can be drawn with no applied voltage: how much light was absorbed and how much charge was collected | Bandgap and light absorption, film thickness, reflection and parasitic absorption |
| Voc (open-circuit voltage) | The voltage when no current is drawn | Bandgap, and losses from non-radiative recombination at defects and interfaces |
| FF (fill factor) | How much of the Jsc × Voc rectangle the maximum power fills | Series resistance (electrodes, transport layers, interfaces) and shunts (short-circuit paths) |
| PCE (power conversion efficiency) | Maximum power ÷ power of the incident light | The product of the three above |
This table is commentary: this article's summary of the general definitions and relationships used for solar cells. Bandgap and composition are covered in our explainer on bandgap and composition design, and interface losses in our explainer on passivation.
On the theoretical ceiling, in 1961 Shockley and Queisser calculated the "detailed balance limit of efficiency" for an ideal p-n junction in which only radiative recombination occurs, showing that, assuming the sun and the cell are blackbodies at 6000 K and 300 K respectively, efficiency reaches a maximum of about 30% at a bandgap of 1.1 eVSourced. This approach is the starting point for thinking about the efficiency ceiling of materials with different bandgaps (our commentary).
3. Our calculation: multiply three numbers and you get the efficiency
A paper published in Nature in 2026 (on reducing interface losses through stereoelectronic design of ligands) reports, for a p-i-n cell with an area of 0.074 cm², a Voc of 1.218 V, a Jsc of 26.13 mA/cm² and an FF of 86.69%, giving a PCE of 27.58% for the cell under the same conditionsSourced.
Assumption: incident light at STC of 1,000 W/m² (= 100 mW/cm²)Our calculation
- 1.218 V × 26.13 mA/cm² × 0.8669 = 27.59 mW/cm²
- 27.59 mW/cm² ÷ 100 mW/cm² = about 27.6%
This agrees with the paper's PCE (27.58%) within rounding. Assumptions and limits: this value is based on measurements within the paper (reverse scan), and is a separate number from the value the same paper obtained through third-party certification (Section 5).
4. EQE (external quantum efficiency): a report card by wavelength
External quantum efficiency (EQE) is the fraction of photons of a given wavelength that result in an electron delivered to the external circuit. Measure the EQE at each wavelength, multiply by the photon count of the reference spectrum and integrate, and you can estimate the Jsc under that spectrum (our commentary).
- Used as a cross-check on Jsc: the 2026 Nature paper states that the current integrated from the EQE matched the J-V measurement (a difference of less than 3.2%)Sourced
- Standard: JET states that it measures the spectral responsivity of solar cell devices in accordance with IEC 60904-8, for bare cells, laminated cells and modulesSourced
- In use since the first report: in 2009, in the first report using CH3NH3PbI3 and CH3NH3PbBr3 as sensitisers on TiO2, Kojima et al. obtained 3.8% for the PbI3-based cell, and a Voc of 0.96 V with an external quantum conversion efficiency of 65% for the PbBr3-based cellSourced
5. Hysteresis: efficiency depends on scan direction
In 2014 Snaith and colleagues at the University of Oxford pointed out "anomalous hysteresis" in the current-voltage curves of perovskite solar cells: the curve measured while raising the voltage (forward) does not match the one measured while lowering it (reverse). The paper gave examples of factors that strengthen or weaken hysteresis and proposed that it is useful to report a "stabilised power output" under operating conditions, alongside the efficiency obtained from scansSourced.
The paper sent its best cell to an accredited photovoltaic metrology laboratory (the National Institute of Metrology, China, NIM) and obtained, through third-party certification, 27.41% by reverse scan and 26.85% stabilised output; the abstract also gives 26.35% by forward scanSourced.
Assumption: here we compare using (reverse-scan efficiency − forward-scan efficiency) ÷ reverse-scan efficiencyOur calculation
- Difference: 27.41 − 26.35 = 1.06 percentage points
- Ratio: 1.06 ÷ 27.41 = about 3.9%
- Difference between stabilised output and reverse scan: 27.41 − 26.85 = 0.56 percentage points
Assumptions and limits: the same paper gives a "hysteresis index of 2.1%" for an in-house-measured cell, but this article has not checked how it is defined. Definitions of the hysteresis index can differ from paper to paper, so the 3.9% derived here is not a value to compare with that paper's 2.1%.
6. A materials engineer's view (1): hysteresis is a sign of slow material response
Dunbar et al. attribute the difficulty of measuring perovskite solar cells to complex, dynamic responses to changes in voltage and in test conditions (especially the move from ambient conditions to STC), writing that the timescales of the main dynamic processes range from nanoseconds to seconds, minutes and even hours. They also point out that the response depends strongly on film morphology and device architectureSourced.
For comparison, the same paper also summarises other kinds of solar cellSourced.
- Dye-sensitised solar cells: with responses of milliseconds to seconds, sweeps of 5 seconds or longer (scan rates below 150 mV/s) can be needed
- CIGS, CdTe and amorphous silicon: can be measured with fast scans, but because of long metastable behaviour they need light soaking (preconditioning) before measurement
- High-efficiency crystalline silicon: because of its large capacitance, sweeps of 100 ms or longer can be needed
In other words, the awkwardness of measuring perovskites is not a problem with the equipment; it is a sign that something inside the material is moving slowly. The leading candidate for that "something" is ion migration, and the fact that the size of the hysteresis changes with the design of composition, grain boundaries and interfaces means that hysteresis can also serve as a diagnostic for materials development (our commentary). The relationship between ion migration and degradation is covered in detail in our explainer on ion migration and degradation.
7. Stabilised output and MPPT: measuring the value once it settles
If the value depends on scan direction, how should it be measured? Dunbar et al. sort the methods used for perovskite efficiency measurement into four groupsSourced.
- MPPT (maximum power point tracking): keeps updating the load resistance, for example by "perturb and observe", to find the stable maximum power point. The paper notes that whereas MPPT in a real power system is there to follow changes in weather, MPPT in efficiency measurement is for finding the steady-state maximum power point that establishes itself slowly under unchanging conditionsSourced
- Stabilised current at fixed voltage (SCFV): keeps measuring at one voltage until the current settles. If the voltage is correctly chosen at the maximum power point, this gives the steady-state efficiencySourced
- Dynamic I-V: the same as an ordinary stepped scan, but holding at each voltage until the current settles. The paper says it is considered the most reliable method; in CSIRO's implementation it waited until the rate of change in current over the last 4 minutes fell below 0.2% per minute, and a single reverse-and-forward cycle took more than 3 hoursSourced
Of the method of illuminating at open circuit and then scanning quickly (OC soak), the paper shows an example where even though hysteresis looked small, the result did not reflect the steady state and overestimated the current, noting that this is called "TEBBing" (a temporary enhancement caused by biasing) in the literatureSourced. No hysteresis does not mean a correct value (our commentary).
8. Different labs, different numbers: results of an inter-comparison
In the inter-comparison by Dunbar et al. (April to October 2016), two institutions accredited for PV performance measurement (CSIRO and NREL) and eight laboratories measured the same three kinds of cell: a slow-responding perovskite cell A, a fast-responding perovskite cell B, and a silicon reference cell CSourced.
| Cell | Spread in repeat measurements at the same institution (1σ) | Spread across non-accredited labs (1σ) |
|---|---|---|
| C: silicon (reference) | under 1% | 3.7% (forward), 3.8% (reverse) |
| B: fast-responding perovskite | 3.2 to 3.3% | 4.5 to 4.9% |
| A: slow-responding perovskite | 6.4% (forward), 1.3% (reverse) | 34.8% and 37.0% |
All Sourced (Dunbar et al. 2017 [Reference 5]). The spread in repeat measurements of cell B is attributed to degradation over a period of days.
The paper states that for slow-responding perovskite cells, the spread between laboratories can be as much as about ten times that for the silicon reference cell, and concludes that the biggest cause was the choice of measurement method rather than the equipment. It then gives recommendations for choosing an appropriate method according to how a cell stabilises and degradesSourced.
9. Certified measurement: who measures, and on what basis
Dunbar et al. explain that when the highest accuracy is required, such as for independent confirmation of breakthrough performance, the practice is to rely on test laboratories externally accredited for technical competence under international standards (IEC 60904-1 and ASTM E948)Sourced.
| Institution | What the official page states |
|---|---|
| Fraunhofer ISE CalLab (Germany) | CalLab PV Cells has worked closely with the Physikalisch-Technische Bundesanstalt (PTB) since its founding and has been accredited as a calibration laboratory since 2008. CalLab PV Modules had its calibration competence confirmed under DIN EN ISO/IEC 17025:2005 by the German accreditation body (DAkkS) on 24 March 2015. It takes part regularly in international inter-laboratory comparisons. As of June 2026, it had certified world records through accredited testing 152 times in total |
| AIST (National Institute of Advanced Industrial Science and Technology, Japan) | Provides a calibration service for primary reference solar cells, the basis for solar cell performance measurement |
| JET (Japan Electrical Safety & Environment Technology Laboratories, Japan) | Offers, under IEC standards, a service measuring the electrical output of various solar cell devices, including perovskites. Calibrates secondary reference cells and modules in accordance with IEC 60904-2 using primary reference solar cells. Spectral responsivity measurement follows IEC 60904-8 |
| CSIRO (Australia) and NREL (US) | Took part in the Dunbar inter-comparison as "institutions accredited for PV performance measurement" |
| National Institute of Metrology, NIM (China) | Commissioned for third-party certification as "an accredited photovoltaic metrology laboratory" by the 2026 Nature paper |
All Sourced (Fraunhofer ISE [Reference 8], AIST [Reference 9], JET [Reference 10], Dunbar et al. [Reference 5], the 2026 Nature paper [Reference 7]).
Record efficiencies are compiled in NREL's "Best Research-Cell Efficiency Chart" and in the "Solar cell efficiency tables". For the latest edition of the latter (Version 68, Green et al., Joule 10, 102494, 2026), the publications repository of the European Commission's Joint Research Centre (JRC) describes it as consolidated tables giving an extensive listing of the highest independently confirmed efficiencies for solar cells and modulesSourced. Japan's Next-Generation Solar Cell Strategy (November 2024), citing the NREL chart, states that perovskite cell efficiency had risen to 26.7% as of November 2024Sourced.
Because the NREL chart and the text of the latest efficiency tables (Version 68) could not be accessed from this article's working environment, the official records as of September 2026 are not given here. Several peer-reviewed papers in 2026 report third-party certified values above 27% for single-junction cells (such as the 27.41% (reverse scan) in the Nature paper in Section 5, and a steady-state efficiency of 27.12% in a Nature Communications paper)Sourced. Whether these are listed as records in the efficiency tables, however, has not been confirmedNot yet confirmed.
10. A materials engineer's view (2): how you measure area alone can shift efficiency
Efficiency is a per-area quantity, so any error in area goes straight into the efficiency. A 2026 Nature Communications paper states that it defined the device area with a black metal aperture mask, and that during certification the certifying body re-measured the active areaSourced. IEC 60904-1:2020 is also described as having added informative annexes on area measurement, capacitive devices and the effects of spatial non-uniformity of irradianceSourced.
Our calculation: treating a 0.074 cm² cell as a square gives a side of about 2.72 mm. If the mask opening were 20 µm wider than designed on each of its four sides, the area would be larger by (2.72 + 0.04)² ÷ 2.72², or about 3.0%. The same offset on a module with an aperture area of 655.2 cm² (a side of about 256 mm) amounts to only about 0.03%Our calculation.
Assumptions and limits: how the area is defined (by mask or by electrode) and how stray light is handled vary with measurement conditions; the calculation is meant to show the order of magnitude by which an area error affects efficiency.
The smaller the cell, the more the materials and machining precision of the fixtures, such as the dimensional accuracy of the mask, the shape of its edge and how little it reflects, feed directly into the efficiency figure. Conversely, for modules the effect of area error shrinks, while the effects of series resistance and shunts from Section 2 grow (our commentary). The same 2026 paper reports a certified 27.12% (steady-state efficiency) for a cell, and a certified 22.25% for a module with an aperture area of 655.2 cm²Sourced. The relationship between area and efficiency was covered in our explainer on roll-to-roll processing and scale-up.
11. Where the standards stand
| Standard | Title (as given on the IEC webstore) | Summary of scope |
|---|---|---|
| IEC 60904-1:2020 | Measurement of photovoltaic current-voltage characteristics | Procedures for measuring the I-V curves of PV devices in natural or simulated sunlight. Applies to single cells, sub-assemblies of cells and modules. The third edition added annexes on area measurement, capacitive devices, dark I-V and spatial non-uniformity of irradiance |
| IEC 60904-3:2019 | Measurement principles for terrestrial PV solar devices with reference spectral irradiance data | Measurement principles for relating the performance of PV devices to a common reference spectrum. The reference spectrum is used in classifying solar simulators under IEC 60904-9 |
| IEC 60904-9:2020 | Classification of solar simulator characteristics | Classifies solar simulators as A+, A, B or C on spectral match, non-uniformity of irradiance on the test plane, and temporal instability of irradiance. A+ was introduced in the third edition |
| IEC TR 63228:2019 | Measurement protocols for photovoltaic devices based on organic, dye-sensitized or perovskite materials | A technical report setting out the current view on evaluating the performance of organic, dye-sensitised and perovskite devices. It states that because the IEC 60904 series was developed with silicon in mind, these devices present measurement challenges |
All Sourced (the IEC webstore page for each standard [References 11 to 14]). Because the full texts of the standards are paid, this article stays within the published titles and scope summaries. Public information on perovskite-specific technical specifications (such as IEC TS 63624) could not be confirmed by this article.
12. Open problems, and what this article could not confirm
(1) "Efficiency" and "usable energy" are different things
The scope description of IEC TR 63228 says that, because of measurement challenges, output in the laboratory can differ greatly from the output observed in real applicationsSourced. Efficiency at STC is a yardstick for comparison; outdoor energy yield changes with temperature, irradiance, spectrum and degradation. This is covered in our explainer on outdoor testing and durability assessment.
(2) The latest official records
As noted in Section 9, this article has not been able to check the NREL chart or the text of the latest efficiency tables. When reading a company announcement or news report claiming a "world record", check whether it gives the certifying body, the area, how the area is defined, and the measurement method (scan or stabilised output) (our commentary).
(3) Defining the hysteresis index
The definition of the hysteresis index may differ between papers, and this article has not confirmed any primary source giving a unified definition.
- Efficiencies are compared under STC (25 °C, AM1.5G, 1,000 W/m²)Sourced
- PCE = Voc × Jsc × FF ÷ incident light. 1.218 V × 26.13 mA/cm² × 86.69% gives about 27.6%Our calculation
- The same cell gave 27.41% by reverse scan, 26.85% stabilised and 26.35% by forward scan: about 1 point apart depending on the methodSourced
- For slow-responding cells, the spread between labs is about ten times that of silicon. The main cause is the choice of measurement methodSourced
- The measurements to trust are those by institutions accredited under standards such as ISO/IEC 17025, including Fraunhofer ISE, AIST, JET, NREL and CSIROSourced
- On a 0.074 cm² cell, a mask edge offset of just 20 µm changes the area by about 3%Our calculation
13. Glossary
- PCE (power conversion efficiency)
- Maximum power ÷ power of the incident light. Compared using values at STC.
- Jsc (short-circuit current density)
- The current per unit area that flows at zero voltage.
- Voc (open-circuit voltage)
- The voltage at zero current.
- FF (fill factor)
- Maximum power ÷ (Jsc × Voc). How "square" the curve is.
- STC (standard test conditions)
- A device temperature of 25 °C, the AM1.5G spectrum and 1,000 W/m².
- AM1.5G
- A reference spectrum representing terrestrial sunlight, defined in IEC 60904-3.
- EQE (external quantum efficiency)
- The number of electrons delivered externally per incident photon, at each wavelength.
- Hysteresis
- A shift in the J-V curve depending on the direction or speed of the voltage scan.
- Stabilised output
- The value once the output near the maximum power point has settled.
- MPPT
- Maximum power point tracking: continually adjusting the load to follow the maximum power point.
- ISO/IEC 17025
- The international standard for the competence of testing and calibration laboratories, and the basis for accreditation.
- Primary reference solar cell
- A calibrated solar cell that serves as the reference for measuring irradiance.
14. References (primary sources)
- Snaith HJ et al. “Anomalous Hysteresis in Perovskite Solar Cells”, The Journal of Physical Chemistry Letters 5, 1511–1515 (2014) https://doi.org/10.1021/jz500113x
- Shockley W, Queisser HJ “Detailed Balance Limit of Efficiency of p-n Junction Solar Cells”, Journal of Applied Physics 32, 510–519 (1961) https://doi.org/10.1063/1.1736034
- Kojima A, Teshima K, Shirai Y, Miyasaka T “Organometal Halide Perovskites as Visible-Light Sensitizers for Photovoltaic Cells”, Journal of the American Chemical Society 131, 6050–6051 (2009) https://doi.org/10.1021/ja809598r
- Public-Private Council for Expanding the Deployment of Next-Generation Solar Cells and Strengthening Industrial Competitiveness (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
- Dunbar RB et al. (CSIRO, NREL and others) “How reliable are efficiency measurements of perovskite solar cells? The first inter-comparison, between two accredited and eight non-accredited laboratories”, Journal of Materials Chemistry A 5, 22542–22558 (2017) (open-access version) https://eprints.qut.edu.au/112795/9/c7ta05609e.pdf
- Yang R, Ding J, Hou T et al. “Soft supermolecule stabilized buried interface for high-performance inverted perovskite solar cells and modules”, Nature Communications 17, 8560 (2026) (full text on PMC) https://pmc.ncbi.nlm.nih.gov/articles/PMC13482877/
- Yang T, Zhao E, Wu N et al. “Stereoelectronic manipulation of ligands for perovskite solar cells”, Nature 654, 660–667 (2026) (full text on PMC) https://pmc.ncbi.nlm.nih.gov/articles/PMC13275290/
- Fraunhofer ISE “CalLab PV Cells / CalLab PV Modules”, official page https://www.ise.fraunhofer.de/en/rd-infrastructure/accredited-labs/callab.html
- AIST “Services: AIST photovoltaic technology development” (in Japanese) https://unit.aist.go.jp/rpd-envene/PV/ja/service/index.html
- Japan Electrical Safety & Environment Technology Laboratories (JET) “Electrical output measurement service for solar cell devices (PV testing)” (in Japanese) https://www.jet.or.jp/renewable/photovoltaic/pv/
- IEC “IEC 60904-1:2020 Photovoltaic devices - Part 1: Measurement of photovoltaic current-voltage characteristics”, IEC webstore https://webstore.iec.ch/en/publication/32004
- IEC “IEC 60904-3:2019 Photovoltaic devices - Part 3: Measurement principles for terrestrial photovoltaic (PV) solar devices with reference spectral irradiance data”, IEC webstore https://webstore.iec.ch/en/publication/61084
- IEC “IEC 60904-9:2020 Photovoltaic devices - Part 9: Classification of solar simulator characteristics”, IEC webstore https://webstore.iec.ch/en/publication/28973
- IEC “IEC TR 63228:2019 Measurement protocols for photovoltaic devices based on organic, dye-sensitized or perovskite materials”, IEC webstore https://webstore.iec.ch/en/publication/64040
- European Commission Joint Research Centre (JRC) “Solar cell efficiency tables: Version 68” (Green MA, Dunlop ED, Yoshita M et al., Cell Press, 31 July 2026), publications repository entry https://publications.jrc.ec.europa.eu/repository/handle/JRC146451
15. Claim-to-source audit
| Claim in the text | Basis | Label |
|---|---|---|
| That the I-V curves of perovskite solar cells show anomalous hysteresis, with examples of contributing factors. The proposal to report stabilised power output under operating conditions alongside scan efficiencies. That the authors are affiliated with the University of Oxford | Reference 1 https://doi.org/10.1021/jz500113x | Sourced |
| The detailed balance limit for an ideal p-n junction with only radiative recombination. A maximum of about 30% at 1.1 eV, assuming blackbodies at 6000 K for the sun and 300 K for the cell | Reference 2 https://doi.org/10.1063/1.1736034 | Sourced |
| CH3NH3PbI3 and CH3NH3PbBr3 as sensitisers on TiO2, with 3.8% for the PbI3-based cell and a Voc of 0.96 V and external quantum conversion efficiency of 65% for the PbBr3-based cell | Reference 3 https://doi.org/10.1021/ja809598r | Sourced |
| The statement, citing the NREL chart, that perovskite cell efficiency had risen to 26.7% as of November 2024 | Reference 4 https://www.meti.go.jp/shingikai/energy_environment/perovskite_solar_cell/pdf/20241128_1.pdf | Sourced |
| The definition of efficiency (maximum power output ÷ input power). STC (25 °C, AM1.5G, 1,000 W/m²). The description of the I-V curve. That where the highest accuracy is needed, accredited laboratories working to IEC 60904-1 and ASTM E948 are relied on. That response times span nanoseconds to hours and depend on morphology and architecture. The comparison with dye-sensitised, CIGS, CdTe, a-Si and crystalline silicon cells. The four measurement methods (standard scan, MPPT, SCFV, dynamic I-V) and their descriptions; the note on the purpose of MPPT in efficiency measurement; that dynamic I-V is considered most reliable, with 0.2%/min, 4 minutes and over 3 hours at CSIRO; NREL's Asymptotic Pmax. OC soak and TEBBing. The participants, period and three cells of the inter-comparison; each spread value; about ten times; that the main cause was the choice of method; and that recommendations were given | Reference 5 https://eprints.qut.edu.au/112795/9/c7ta05609e.pdf | Sourced |
| That the area was defined with a black metal aperture mask, and the certifying body re-measured the active area during certification. A certified steady-state efficiency of 27.12% for the cell, and a certified 22.25% for a module with an aperture area of 655.2 cm² | Reference 6 https://pmc.ncbi.nlm.nih.gov/articles/PMC13482877/ | Sourced |
| A 0.074 cm² p-i-n cell with Voc 1.218 V, Jsc 26.13 mA/cm², FF 86.69% and PCE 27.58%. An in-house hysteresis index of 2.1%. Third-party certification by NIM's photovoltaic metrology laboratory of 27.41% reverse and 26.85% stabilised output, with 26.35% forward in the abstract. A difference of under 3.2% between the EQE-integrated current and J-V | Reference 7 https://pmc.ncbi.nlm.nih.gov/articles/PMC13275290/ | Sourced |
| That CalLab PV Cells has worked with PTB and been accredited as a calibration laboratory since 2008; that CalLab PV Modules had its calibration competence under DIN EN ISO/IEC 17025:2005 confirmed by DAkkS on 24 March 2015; participation in international comparisons; 152 world-record certifications as of June 2026 | Reference 8 https://www.ise.fraunhofer.de/en/rd-infrastructure/accredited-labs/callab.html | Sourced |
| That AIST provides a calibration service for primary reference solar cells | Reference 9 https://unit.aist.go.jp/rpd-envene/PV/ja/service/index.html | Sourced |
| That JET offers electrical output measurement under IEC standards for various devices including perovskites, calibration to IEC 60904-2 using primary reference cells, and spectral responsivity measurement to IEC 60904-8 | Reference 10 https://www.jet.or.jp/renewable/photovoltaic/pv/ | Sourced |
| The scope of IEC 60904-1:2020 and the annexes added in the third edition | Reference 11 https://webstore.iec.ch/en/publication/32004 | Sourced |
| The scope of IEC 60904-3:2019 (the reference spectrum, and its use in the classification under 60904-9) | Reference 12 https://webstore.iec.ch/en/publication/61084 | Sourced |
| That IEC 60904-9:2020 classifies solar simulators as A+, A, B and C on three criteria, with A+ newly introduced | Reference 13 https://webstore.iec.ch/en/publication/28973 | Sourced |
| That IEC TR 63228:2019 sets out the view on evaluating OPV, DSC and PSC performance; that because IEC 60904 was developed with silicon in mind there are challenges (time dependence, spectral responsivity, small area and temperature measurement, transient response, optical interference, non-linearity with irradiance); and that laboratory output can differ greatly from that in real use | Reference 14 https://webstore.iec.ch/en/publication/64040 | Sourced |
| The bibliographic details of the Solar cell efficiency tables (Version 68) (Green, Dunlop, Yoshita et al., Cell Press, 31 July 2026) and the description as consolidated tables of the highest independently confirmed efficiencies. That the journal is Joule volume 10, 102494 (confirmed from Crossref metadata) | Reference 15 https://publications.jrc.ec.europa.eu/repository/handle/JRC146451 | Sourced |
| 1.218 × 26.13 × 0.8669 = 27.59 mW/cm², about 27.6%. 27.41 − 26.35 = 1.06 points, 1.06 ÷ 27.41 = about 3.9%, 27.41 − 26.85 = 0.56 points. A side of about 2.72 mm for a 0.074 cm² square; area increases of about 1.5%, 3.0% and 7.5% for 10, 20 and 50 µm on each side; about 0.03% for 20 µm on 655.2 cm² | Our calculation. Incident light of 100 mW/cm², the definition of the hysteresis ratio, and treating the aperture as a square are this article's assumptions | Our calculation |
| Whether the certified values above 27% reported in 2026 are listed as records in the efficiency tables | Not confirmed, because the NREL chart and the text of efficiency tables Version 68 could not be accessed from this article's working environment (commentary) | Not yet confirmed |
| The official records as of September 2026, public information on perovskite-specific technical specifications such as IEC TS 63624, a unified definition of the hysteresis index, the general explanation of EQE, and the table of where materials have an effect | Not stated because they could not be confirmed, or commentary in which this article summarises general definitions | Commentary |
| Reading hysteresis as a sign of slow material response and as a diagnostic. The point that no hysteresis does not mean a correct value. The reading that fixture materials and machining precision matter more for smaller cells. The cautions on reading records | This article's own framing and commentary based on published content. Not views expressed by the institutions or authors | Commentary |
| That Figs. 1, 2, 3 and 5 are explanatory drawings rather than real measured data, and that the hero image and Fig. 4 are AI-generated images | Our note (commentary) | Commentary |
Last updated 25 September 2026. Sources are limited to primary material (peer-reviewed papers, IEC webstore standard pages, the official pages of accredited calibration laboratories, METI's strategy document and the JRC publications repository). Because the full texts of the standards are paid, the article stays within the published titles and scope summaries. The NREL chart and the text of the latest efficiency tables (Version 68) could not be accessed from this article's working environment, so official records as of September 2026 are not given. The efficiencies quoted here differ in area, certification and measurement method, and cannot be compared directly. All figures are explanatory concept graphics. Figs. 1, 2, 3 and 5 are vector drawings; the hero image and Fig. 4 are AI-generated images, and none of them shows real measurement equipment or measured data.