Predicting Trends in $V_{OC}$ Through Rapid, Multimodal Characterization of State-of-the-Art p-i-n Perovskite Devices

Amy E. Louks; Anthony T. Troupe; Axel F. Palmstrom; Brandon T. Motes; Dane W. deQuilettes; Joseph J. Berry

REVIEW 4 major objections 5 minor 6 references

Predicting Trends in $V_{OC}$ Through Rapid, Multimodal Characterization of State-of-the-Art p-i-n Perovskite Devices

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Multimodal optical measurements taken before contacts are deposited can predict the open-circuit voltage trends of finished p-i-n perovskite solar cells, with a consistent ~170 mV offset appearing after top-layer deposition.

desk verdict Useful physics-based screening workflow for perovskite semi-fabricates, but the load-bearing PLQE model is undisclosed; deserves peer review only after the authors open that box. read the letter →

arxiv 2508.21037 v1 pith:5MCVTMPA submitted 2025-08-28 cond-mat.mtrl-sci physics.app-ph

Amy E. Louks , Brandon T. Motes , Anthony T. Troupe , Axel F. Palmstrom , Joseph J. Berry , Dane W. deQuilettes This is my paper · ORCID

classification cond-mat.mtrl-sciphysics.app-ph

keywords perovskitesolarcellsopen-circuitvoltagepredictiontime-resolvedphotoluminescencespectrally-resolvedUrbachtailnon-radiativerecombinationmultimodalmetrologyp-i-ndevicestack

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper aims to show that unfinished perovskite solar cells can be screened for eventual voltage performance using only quick, non-contact optical measurements. By combining transmission, spectrally resolved photoluminescence, and time-resolved photoluminescence on 120 device areas across five processing batches, the authors reconstruct the radiative voltage limit from the absorption edge and band tail, then subtract the non-radiative loss estimated from the long-lived photoluminescence decay. The predicted open-circuit voltages rank the batches in the same order as the voltages measured on completed devices, with a consistent ~170 mV gap attributed to the C60/BCP/Ag top contact. If this holds, researchers and manufacturers could make go/no-go decisions before depositing expensive contacts, and could isolate whether voltage loss comes from the absorber, the hole-transport interface, or the electron-transport interface.

What carries the argument

The load-bearing identity is the detailed-balance decomposition of open-circuit voltage into a radiative limit and a non-radiative penalty: VOC = VOC,rad − (kT/q)|ln(PLQE)|. VOC,rad is obtained from the absorptivity spectrum, with the sub-bandgap Urbach tail reconstructed from the photoluminescence spectrum through the reciprocity relationship rather than from a completed-device external quantum efficiency. The non-radiative penalty is anchored by a single experimental input: the slower decay component of the low-fluence time-resolved photoluminescence trace on the half-stack, converted into a non-radiative recombination rate and fed into coupled differential equations to simulate the steady

What would settle it

Measure the external photoluminescence quantum efficiency of the semi-fabricates directly with an integrating sphere and compare it with the simulated value; a divergence beyond uncertainty would invalidate the TRPL-to-PLQE conversion. Separately, complete a batch with three different C60 thicknesses or a different electron-transport material: if the predicted-versus-measured VOC offset changes systematically, the constant ~170 mV ETL-loss attribution is wrong.

Watch

Extended reading notes

Core claim

The central discovery is that trends in the open-circuit voltage of completed p-i-n perovskite devices can be predicted from semi-fabricates before any contact layers are deposited. The radiative contribution to VOC is computed by stitching a transmission-derived absorption spectrum to an Urbach tail extracted from the low-energy photoluminescence shoulder via the reciprocity relation. The non-radiative contribution is obtained by taking the slower time-resolved photoluminescence decay component as the non-radiative recombination rate, simulating the steady-state photoluminescence quantum efficiency from coupled differential equations, and using VOC = VOC,rad − (kT/q)|ln(PLQE)|. Across the f

Load-bearing premise

The load-bearing premise is that the slower component of the time-resolved photoluminescence decay, measured before contacts are added, can be converted by an undisclosed model into the non-radiative recombination rate that determines the steady-state photoluminescence efficiency; if that conversion or its model parameters are inaccurate, the predicted voltages and their trends would be systematically wrong.

Editorial extensions

If this is right

If the offset is truly set by C60/BCP/Ag deposition, the same semi-fabricate measurement can assign a device's voltage losses to the absorber/hole-transport side versus the electron-transport/contact side without building full devices.
Fabrication batches with poor hole-transport/perovskite interfaces can be rejected before contact evaporation, saving time, materials, and processing labor.
The workflow's 360 automated optical measurements on 120 device pads demonstrate that the approach scales to statistically meaningful screening, with 24 samples per condition.
Predictions are currently for VOC trends only, but the same measurements contain absorption and recombination information that the authors expect can be extended to short-circuit current density, fill factor, and efficiency.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the ~170 mV contact-induced offset is a stable stack constant, a single calibration measurement per device architecture would turn this trend predictor into an absolute VOC predictor for manufacturing.
A direct test would be to vary C60 thickness or substitute a different electron-transport material and check whether the offset moves; if it stays fixed, the attribution to the electron-transport interface would need revision.
Because the coupled equations and fitting parameters are not specified in the paper, an independent lab cannot reproduce the photoluminescence-efficiency simulation without contacting the authors; publishing that model would make the method portable.
Combining this physics-based pipeline with a purely learned optical-to-performance model could reveal where the TRPL-to-PLQE conversion is misspecified, for instance if both agree on batch ranking but disagree on the size of the hole-transport-layer effect.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

This paper is worth reading. It shows that an automated multimodal optical setup on half-stacked p-i-n perovskite devices—SRPL, TRPL, transmission—can rank VOC trends across a small set of process variations, using detailed balance plus a simulated PLQE. That combination, non-contact screening before top contacts are deposited, is new relative to the ML approach in ref 23, and it gives you physical interpretability instead of a black box. If the core claim holds, it shortens materials optimization loops for single junctions and tandems, which is genuinely useful within the perovskite PV community.

The work is careful in several places. Twin samples shipped to Optigon and kept at NREL show similar JV trends. They measured 120 device pads, and the absolute offset between predicted and measured VOC, about 170 mV, matches published voltage losses after C60/BCP/Ag deposition. The radiative limit calculation, stitching the PL-derived Urbach tail to transmission data, is standard and sensible. They explicitly note that they ignore reflectance and scattering, which is honest.

The soft spot is not the physics; it is the disclosure. Equation 2 links VOC to PLQE, but the PLQE is never measured. It is simulated from the second TRPL decay component using coupled differential equations that are only cited to refs 38 and 39—no equations, no radiative or trap parameters, no carrier-density calibration, no fitting procedure, no code or data. The stress-test note is right: a factor of two error in PLQE changes predicted VOC by about 18 mV, an order of magnitude by about 60 mV, and the differences you are ranking are on the order of 150 mV. The offset attributed to the ETL could be a residue of model misspecification rather than a clean measurement. Also, the longer TRPL decay component in a half-stack is not purely non-radiative; trapping, radiative recombination, and interface contributions can be present and may vary with HTL chemistry, which could confound the batch ranking. The abstract says "quantitative trends," but the paper only demonstrates qualitative trend agreement, and there are no error bars on predicted VOC anywhere. Five batches with n=24 per condition is a reasonable start, not a comprehensive validation.

That said, these are addressable gaps, not fatal flaws. The authors' own conclusion says "good predictions in qualitative VOC trends," which is honest. The 170 mV offset is presented as an observation, not a fitted constant, and it is consistent with literature. The missing model specification is a reproducibility requirement, not a stylistic preference. I would not accept the paper as-is, but I would send it to peer review with a clear request: disclose the equations, parameters, and fitting procedure, add error propagation from optical fits to VOC, and soften the abstract's "quantitative" claim to match what is shown. If the authors provide those, this becomes a solid methods paper for perovskite PV labs and metrology developers. Send it to review.

Referee Report

4 major / 5 minor

Summary. The paper presents a multimodal optical metrology workflow (spectrally resolved PL, time-resolved PL, and transmission) applied to semi-fabricate p-i-n perovskite solar cells before top-contact deposition. From these measurements the authors compute a radiative VOC using detailed balance and a non-radiative correction from a simulated PLQE, yielding a predicted VOC for each device pad. They apply this workflow to 120 device areas across five batches varying the hole-transport layer and perovskite annealing time, then complete the devices and compare predicted VOC with JV-measured VOC. The central claim is that the predicted VOC trends correlate with measured VOC trends, with a median offset of ~170 mV attributed to losses introduced by C60/BCP/Ag deposition, and that the dominant differences among batches are governed by the perovskite/HTL interface.

Significance. If the workflow is validated and made reproducible, it would be a practically useful, rapid, non-destructive screening tool for perovskite device fabrication, with a physically grounded basis in detailed balance and reciprocity. The study has genuine strengths: a relatively large dataset (360 optical measurements, 120 device areas), a direct comparison of predictions on semi-fabricates with electrical measurements on completed devices, and a loss-budgeting framework that separates absorber/HTL losses from ETL/contact losses. The ~170 mV offset is consistent with published ranges for C60-related losses, and I see no circularity in the main workflow because the measured JV VOC is not used to calibrate the optical model. However, the load-bearing PLQE simulation is not specified, no uncertainty propagation is provided, and the absolute offset is inferred without correcting for reflectance/scattering. These gaps currently prevent the paper from fully supporting its quantitative claims.

major comments (4)

[Figure 3d / Eq. (2)] The predicted VOC rests on a PLQE value that is simulated, not measured, from the second TRPL decay component using 'a set of coupled differential equations consistent with previous work' (refs 38,39). The paper does not state those equations, the radiative/Auger/trap rate constants, the carrier-density calibration, the generation rate used for the AM1.5 dotted line, or the fitting procedure. This is load-bearing: from Eq. (2), a factor-of-2 error in PLQE changes VOC by kT ln2 ~ 18 meV, and an order-of-magnitude error changes it by ~59 meV—comparable to the ~150 mV HTL trend and the ~170 mV offset attributed to C60/BCP/Ag. Without the model and parameter values, the offset is not distinguishable from model residue. Please provide the full model, all parameters, and a sensitivity analysis.
[Figure 3c and accompanying text] The text states that 'we use the second decay component to represent the non-radiative recombination rate.' On a half-stack, a low-fluence TRPL decay is not a direct measure of the non-radiative lifetime; it can include radiative recombination, trapping/detrapping, and interfacial effects, and the 'second component' depends on the chosen fit function, fluence, and instrument response. If these extra contributions vary with HTL chemistry (NiOx vs NiOx/PTAA vs Me-2PACz), the predicted PLQE ranking is confounded even if the model is internally consistent. Please validate the conversion against directly measured PLQE on the same semi-fabricates and report the TRPL fit form and residuals.
[Figure 4 and '~170 mV' offset] No uncertainty propagation is shown from the optical fits (Urbach tail, transmission, TRPL) into the predicted VOC, and no quantitative agreement metric is reported—there is no R², RMSE, or confidence interval on the 170 mV offset. The box plots report n=24 per batch, but those 24 values are not independent (4 samples × 6 pads), so the effective statistical weight is smaller and the qualitative 'correlate well' claim is not yet quantitatively supported. Please provide error bars on predicted VOC, cluster-aware statistics, and a direct predicted-vs-measured parity plot.
[Radiative VOC calculation / 'We note that we do not account for reflectance or scattering'] The absolute VOC,rad computed via Eq. (1) depends on the magnitude of the absorptivity spectrum. Omitting reflectance and scattering can shift VOC,rad by tens of meV, and the ~170 mV ETL-loss attribution is an absolute comparison between predicted and measured VOC. This missing correction therefore directly affects a central quantitative claim. Please bound this error with a simple optical correction (e.g., 1−R) or demonstrate that it cannot alter the conclusions by more than ~20 meV.

minor comments (5)

[Introduction] Typo: 'tool that that autonomously measures' should read 'tool that autonomously measures'.
[Figure 4] The legend order and plotting style are unclear: clarify which color corresponds to 'Radiative VOC', 'Predicted VOC', and 'Measured VOC', and define the open-circle outliers. Also state explicitly whether the six device pads per sample are treated as independent measurements.
[References] Ref. 15 is given only as 'arXiv 2024' without an identifier or DOI; please provide the full citation. Ref. 31 is cited in a context about masking but no mask is described in the JV methods; please clarify.
[Supporting Information] The SI would benefit from example TRPL fit residuals, the full parameter set used in the PLQE simulation, and a description of the automated fitting/quality-control criteria implemented in the 'automated analysis pipeline'.
[Nomenclature] The SAM material is called '2PACz' in the main text but 'Me-2PACz' in the SI; use one consistent name throughout.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: predicted VOC is computed from independent optical measurements via standard reciprocity; the 170 mV offset is a post-hoc comparison, not a fitted input.

full rationale

The paper's derivation chain is: (i) measure transmission, SRPL, and TRPL on semi-fabricates; (ii) construct absorptivity from transmission and the Urbach tail from PL; (iii) compute VOC,rad via detailed balance (Eq. 1); (iv) take the second TRPL decay component as the non-radiative recombination rate; (v) simulate steady-state PLQE with coupled differential equations cited to refs 38/39; (vi) compute predicted VOC by Eq. 2, VOC = VOC,rad - kT ln(PLQE). The measured final JV VOC appears only after the prediction for validation and does not enter any model parameter. The 170 mV offset is described as an observed 'absolute offset ... across all sets of devices' (Figure S3) and is attributed to C60/BCP/Ag deposition; it is not a constant added to align predictions. The main load-bearing assumption—using the long TRPL component as the non-radiative rate—is a physical proxy, and the paper does not disclose the PLQE model equations/parameters; that is a reproducibility/validity concern, not circularity. Ref. 33 (deQuilettes et al.) is self-cited for the external radiative efficiency framework, but detailed balance and the PLQE-VOC relation are standard and independently established, so the citation is not load-bearing. The conflict-of-interest note about Optigon is a financial disclosure and does not constitute circular reasoning. No equation defines predicted VOC in terms of measured VOC, and no fitted target is renamed as a prediction.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central prediction relies on standard detailed balance plus two fitted or undisclosed modeling steps: the TRPL-derived non-radiative lifetime and the undisclosed PLQE simulation parameters. No new physical entities are introduced. The claims are therefore conditional on the validity of those published but under-specified models.

free parameters (2)

non-radiative lifetime (τ_nr) from second TRPL decay component = not reported; sample-dependent
Extracted from low-fluence TRPL fits and used as the recombination input for the PLQE simulation; the predicted VOC scales with this fitted lifetime.
PLQE simulation parameters (radiative coefficient, recombination constants, densities) = not disclosed
The coupled differential equations from refs 38/39 require material parameters; none are given in the paper, so the simulated PLQE (Figure 3d) is conditional on undisclosed parameters.

assumptions (6)

standard math Detailed balance and reciprocity between absorption and photoluminescence
Used to derive VOC,rad (Eq 1) from absorptivity and to extract the Urbach tail from the PL spectrum.
domain assumption The PL low-energy tail reflects the absorption band tail via reciprocity
Assumed to construct the band-edge absorptivity; stitched to transmission data in Figure 3a/b.
domain assumption The second (long) TRPL decay component represents the total non-radiative recombination rate of the half-stack
Stated near Figure 3c; combines bulk, surface, and interfacial recombination but is not independently verified.
domain assumption The coupled differential-equation model from refs 38/39 simulates steady-state PLQE with the fitted non-radiative rate
Referenced but not derived or parameterized in the paper; the PLQE curve in Figure 3d rests on this.
domain assumption Reflectance and scattering can be neglected in the absorptivity approximation
Acknowledged in the text; the absorption coefficient is approximated from transmission and PL only, which affects absolute VOC,rad.
domain assumption The ~170 mV voltage loss after C60/BCP/Ag deposition is uniform across all sample batches
Used to interpret the absolute offset between predicted and measured VOC (Figure S3); not used to fit predictions but underpins the loss attribution.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Predicting Trends in $V_{OC}$ Through Rapid, Multimodal Characterization of State-of-the-Art p-i-n Perovskite Devices." pith.science (2026). https://pith.science/paper/5MCVTMPA

@misc{pith2026250821037,
  author       = {Pith},
  title        = {Pith review of: Predicting Trends in $V_OC$ Through Rapid, Multimodal Characterization of State-of-the-Art p-i-n Perovskite Devices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5MCVTMPA}},
  note         = {Machine review of arXiv:2508.21037}
}

read the original abstract

Perovskite photovoltaic technologies are approaching commercial deployment, yet single junction and tandem architectures both still have significant room to improve power conversion efficiency and stability. The ability to perform rapid screening of material quality after altering processing conditions is critical to accelerating the optimization and commercialization of perovskite-based technologies. Currently, researchers utilize a wide range of stand-alone metrology tools to isolate sources of power loss throughout a device stack, which can be slow and labor intensive. Here, we demonstrate the use of a multimodal metrology approach to rapidly determine the maximum achievable and predicted open circuit voltages of > 100 perovskite devices during fabrication. Acquisition of these different data are facilitated by combining them into a single integrated measurement platform. We show that these data and automated analysis can be used to rapidly understand and ultimately predict quantitative trends in open circuit voltages of state-of-the-art devices architectures. The data and automated analysis workflow presented provides a reliable approach to quickly identify absorber and charge transport layer combinations that can lead to improved open circuit voltages.

Figures

Figures reproduced from arXiv: 2508.21037 by the authors.

**Figure 1.** shows an overview of the feedback cycle demonstrated in this study. First, we explore a set of different processing conditions for device fabrication, which are separated into 5 unique batches. Next, we characterize semi-fabricate samples midway through device completion through automated characterization. We use these data as inputs into an automated analysis pipeline to predict metrics relevant to device performan… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

6 extracted references · 6 canonical work pages

[2]

We observe only small changes in the predicted VOC for active layers annealed at different conditions

Notably, changing the HTL to a bilayer with PTAA or a SAM-like layer results in a ~150 mV increase in the predicted VOC relative to the control. We observe only small changes in the predicted VOC for active layers annealed at different conditions. For the NiOx/PTAA bilayer compared to the SAM-like layer, we predict a slightly lower predicted VOC for the S...

work page 2016
[13]

V.; Dimitrov, S.; De Castro, C

(19) Péan, E. V.; Dimitrov, S.; De Castro, C. S.; Davies, M. L. Interpreting Time-Resolved Photoluminescence of Perovskite Materials. Phys. Chem. Chem. Phys. 2020, 22 (48), 28345-28358, 10.1039/D0CP04950F. DOI: 10.1039/D0CP04950F. (20) Yu, S.; Liu, P.; Xiao, S. A Review of Main Characterization Methods for Identifying Two-Dimensional Organic–Inorganic Hal...

work page doi:10.1039/d0cp04950f 2020
[1276]

(41) Stolterfoht, M.; Wolff, C

DOI: 10.1038/s41467-024-45512-8. (41) Stolterfoht, M.; Wolff, C. M.; Márquez, J. A.; Zhang, S.; Hages, C. J.; Rothhardt, D.; Albrecht, S.; Burn, P. L.; Meredith, P.; Unold, T.; et al. Visualization and Suppression of Interfacial Recombination for High-Efficiency Large-Area PIN Perovskite Solar Cells. Nat. Energy. 2018, 3 (10), 847-854. DOI: 10.1038/s41560...

work page doi:10.1038/s41467-024-45512-8 2018
[2024]

Quantifying Recombination and Charge Carrier Extraction in Halide Perovskites via Hyperspectral Time-Resolved Photoluminescence Imaging

(16) Phirke, H.; Gharabeiki, S.; Singh, A.; Krishna, A.; Siebentritt, S.; Redinger, A. Quantifying Recombination and Charge Carrier Extraction in Halide Perovskites via Hyperspectral Time-Resolved Photoluminescence Imaging. APL Energy 2024, 2 (1), 016111. DOI: 10.1063/5.0188166. (17) Mahon, N. S.; Korolik, O. V.; Khenkin, M. V.; Arnaoutakis, G. E.; Galaga...

work page doi:10.1063/5.0188166 2024
[2025]

W.; Bulović, V

https://www.nrel.gov/pv/cell-efficiency.html (3) Brenes, R.; Laitz, M.; Jean, J.; deQuilettes, D. W.; Bulović, V. Benefit from Photon Recycling at the Maximum-Power Point of State-of-the-Art Perovskite Solar Cells. Phys. Rev. Appl. 2019, 12 (1), 014017. DOI: 10.1103/PhysRevApplied.12.014017. (4) Hörantner, M. T.; Leijtens, T.; Ziffer, M. E.; Eperon, G. E....

work page doi:10.1103/physrevapplied.12.014017 2019
[4654]

(12) Cakan, D

DOI: 10.1038/s41467-024-48768-2. (12) Cakan, D. N.; Kumar, R. E.; Oberholtz, E.; Kodur, M.; Palmer, J. R.; Gupta, A.; Kaushal, K.; Vossler, H. M.; Fenning, D. P. PASCAL: The Perovskite Automated Spin Coat Assembly Line Accelerates Composition Screening in Triple-Halide Perovskite Alloys. Digital Discovery 2024, 3 (6), 1236-1246, 10.1039/D4DD00075G. DOI: 1...

work page doi:10.1038/s41467-024-48768-2 2024

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [2]

We observe only small changes in the predicted VOC for active layers annealed at different conditions

Notably, changing the HTL to a bilayer with PTAA or a SAM-like layer results in a ~150 mV increase in the predicted VOC relative to the control. We observe only small changes in the predicted VOC for active layers annealed at different conditions. For the NiOx/PTAA bilayer compared to the SAM-like layer, we predict a slightly lower predicted VOC for the S...

work page 2016

[2] [13]

V.; Dimitrov, S.; De Castro, C

(19) Péan, E. V.; Dimitrov, S.; De Castro, C. S.; Davies, M. L. Interpreting Time-Resolved Photoluminescence of Perovskite Materials. Phys. Chem. Chem. Phys. 2020, 22 (48), 28345-28358, 10.1039/D0CP04950F. DOI: 10.1039/D0CP04950F. (20) Yu, S.; Liu, P.; Xiao, S. A Review of Main Characterization Methods for Identifying Two-Dimensional Organic–Inorganic Hal...

work page doi:10.1039/d0cp04950f 2020

[3] [1276]

(41) Stolterfoht, M.; Wolff, C

DOI: 10.1038/s41467-024-45512-8. (41) Stolterfoht, M.; Wolff, C. M.; Márquez, J. A.; Zhang, S.; Hages, C. J.; Rothhardt, D.; Albrecht, S.; Burn, P. L.; Meredith, P.; Unold, T.; et al. Visualization and Suppression of Interfacial Recombination for High-Efficiency Large-Area PIN Perovskite Solar Cells. Nat. Energy. 2018, 3 (10), 847-854. DOI: 10.1038/s41560...

work page doi:10.1038/s41467-024-45512-8 2018

[4] [2024]

Quantifying Recombination and Charge Carrier Extraction in Halide Perovskites via Hyperspectral Time-Resolved Photoluminescence Imaging

(16) Phirke, H.; Gharabeiki, S.; Singh, A.; Krishna, A.; Siebentritt, S.; Redinger, A. Quantifying Recombination and Charge Carrier Extraction in Halide Perovskites via Hyperspectral Time-Resolved Photoluminescence Imaging. APL Energy 2024, 2 (1), 016111. DOI: 10.1063/5.0188166. (17) Mahon, N. S.; Korolik, O. V.; Khenkin, M. V.; Arnaoutakis, G. E.; Galaga...

work page doi:10.1063/5.0188166 2024

[5] [2025]

W.; Bulović, V

https://www.nrel.gov/pv/cell-efficiency.html (3) Brenes, R.; Laitz, M.; Jean, J.; deQuilettes, D. W.; Bulović, V. Benefit from Photon Recycling at the Maximum-Power Point of State-of-the-Art Perovskite Solar Cells. Phys. Rev. Appl. 2019, 12 (1), 014017. DOI: 10.1103/PhysRevApplied.12.014017. (4) Hörantner, M. T.; Leijtens, T.; Ziffer, M. E.; Eperon, G. E....

work page doi:10.1103/physrevapplied.12.014017 2019

[6] [4654]

(12) Cakan, D

DOI: 10.1038/s41467-024-48768-2. (12) Cakan, D. N.; Kumar, R. E.; Oberholtz, E.; Kodur, M.; Palmer, J. R.; Gupta, A.; Kaushal, K.; Vossler, H. M.; Fenning, D. P. PASCAL: The Perovskite Automated Spin Coat Assembly Line Accelerates Composition Screening in Triple-Halide Perovskite Alloys. Digital Discovery 2024, 3 (6), 1236-1246, 10.1039/D4DD00075G. DOI: 1...

work page doi:10.1038/s41467-024-48768-2 2024