REVIEW 5 major objections 5 minor 8 references
Spectroscopic ellipsometry of CsPbCl${_3}$ perovskite thin films
T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash
Pith's one-line read Thermally evaporated CsPbCl3 thin films have a sharp absorption edge at 411 nm and an optical bandgap of about 2.99 eV, with n and k spectra now available for device simulation.
desk verdict Genuinely new n/k data for CsPbCl3, but the oscillator model as printed is unreproducible and the bandgap agreement is partly circular. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the Tauc-Lorentz oscillator for the imaginary part of the dielectric function, parameterised by band gap Eg, centre energy Eo, broadening Br, and amplitude A, with the real part obtained through Kramers-Kronig consistency; two harmonic oscillators are added to capture higher-energy transitions. The extraction pipeline starts from a Cauchy layer in the transparent region to fix thickness, then replaces it with a Kramers-Kronig constrained B-spline to capture the full dielectric function, and finally re-parametrises that function with the Tauc-Lorentz-plus-oscillators model. A Bruggeman effective-medium approximation layer with a fixed 50% void fraction represents surface roughness. The central work of this machinery is to convert measured Psi and $\Delta$ spectra into the real and imaginary parts of the dielectric function, hence n and k, without assuming the material's optical behaviour in advance.
What would settle it
Measure the same films with a second independent method, for example by inverting variable-angle ellipsometry data at each wavelength without any oscillator model, or by measuring transmittance and reflectance separately; if the extracted absorption edge moves away from 411 nm or the Tauc-plot bandgap shifts by more than about 0.05 eV, the reported n and k depend on the chosen oscillator model rather than being uniquely determined by the data.
Extended reading notes
Core claim
The central claim is that a simple optical dispersion model, one Tauc-Lorentz oscillator combined with two harmonic oscillators on a Si/SiO2 substrate stack with a 50%-void effective-medium roughness layer, describes the polarization response of evaporated CsPbCl3 films from 300 to 900 nm with mean squared errors of 3.15 and 6.59 for the two thicknesses studied. From that fit the paper extracts thicknesses of about 92.7 nm and 187.2 nm, consistent with the intended 90 nm and 180 nm depositions, and obtains n(lambda) and k(lambda) spectra whose extinction coefficient rises sharply at 411 nm and is essentially zero above roughly 500 nm. Converting k to an absorption coefficient and using a Tauc plot gives a direct optical bandgap of approximately 2.99 eV for both samples, matching the bandgap obtained from UV-Vis absorption spectra. The authors take this cross-check as validation of the model and present the n and k values as usable inputs for optoelectronic device design.
Load-bearing premise
The result stands on the assumption that the chosen optical model, one Tauc-Lorentz oscillator, two harmonic oscillators, and a roughness layer fixed at 50% air, is the right representation of how CsPbCl3 films respond to light, and the paper does not test alternative models or report uncertainties in the fitted constants.
Editorial extensions
If this is right
- Device modellers can use the reported n(lambda) and k(lambda) spectra to simulate CsPbCl3-based UV photodetectors, violet/blue LEDs, and lasers without assuming values from related perovskites.
- The same ellipsometric model provides a non-destructive thickness measurement, with fitted values of 92.7 nm and 187.2 nm close to the intended 90 nm and 180 nm depositions.
- The consistent 2.99 eV gap across two thicknesses indicates that the optical bandgap of evaporated CsPbCl3 films is thickness-independent in the 90 to 180 nm range.
- Transparency above roughly 500 nm and a sharp edge at 411 nm define the practical operating window for CsPbCl3-based photonic devices.
- The validated workflow can be transferred to other all-inorganic halide perovskites whose optical constants are missing.
Reading between the lines
- A model-free cross-check, such as fitting the same data with a standalone Kramers-Kronig-constrained B-spline, would test whether the sharp edge at 411 nm and the 2.99 eV gap are properties of the data rather than artifacts of the Tauc-Lorentz plus harmonic-oscillator parameterisation.
- Because the roughness layer fixes the void fraction at 50%, films with different surface morphologies may require re-fitting that parameter; a natural extension is to correlate the void fraction with SEM images.
- The near-zero broadening allowed on one harmonic oscillator in sample B hints that the fit may be absorbing minor spectral features; constraining broadening to positive values would show how much of the result is oscillator choice.
- Mapping the same extraction procedure across a CsPbCl3-CsPbBr3 composition series would test how the 2.99 eV gap and 411 nm edge shift with halide ratio, directly supporting bandgap-tunable device design.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript reports spectroscopic ellipsometry measurements of thermally evaporated CsPbCl3 thin films (two thicknesses, nominally 90 nm and 180 nm) on Si/SiO2 substrates. The authors construct a multilayer optical model and parameterize the perovskite layer with a Tauc-Lorentz oscillator plus two harmonic oscillators, obtaining thicknesses of 92.73 nm and 187.18 nm, mean squared errors of 3.15 and 6.59, and wavelength-dependent n and k over 300-900 nm. A sharp extinction edge is found at 411 nm, and optical bandgaps of 2.99 eV are reported from both UV-Vis Tauc plots and Tauc plots made from the SE-derived extinction coefficient.
Significance. Reliable complex optical constants for CsPbCl3 would be a useful addition to the literature, because this wide-bandgap perovskite is relevant for UV/violet optoelectronics and existing data are sparse. The paper has genuine strengths: two film thicknesses are measured, the SE results are cross-checked against SEM and UV-Vis absorption, and low MSE values are reported. However, the manuscript as written does not yet establish the reliability of the extracted n/k spectra. The oscillator model contains entries that appear unphysical or duplicated, the SE-based bandgap validation is partly circular, and important modeling choices are fixed without sensitivity analysis. These issues affect the central claim of the paper and require substantive revision.
major comments (5)
- [Tables I and II and the optical model description] The reported oscillator model is not reproducible as printed. Sample A's first harmonic oscillator has Eo = 1.06e-8 eV with A = 1.74e-7 eV, which is not a physically meaningful oscillator near the CsPbCl3 band edge; sample B's second harmonic oscillator has Br = 0.00 eV, i.e., zero broadening; and sample A's second harmonic oscillator (A = 430.43, Br = 0.091, Eo = 2.99) is identical to sample B's Tauc-Lorentz parameters. In addition, the functional form of the 'harmonic oscillator' is never given, so Eq. (5) and the text do not fully specify the model. If any of these values are typographical errors, they must be corrected; if they are real fit results, their physical role must be explained. As printed, the model cannot be independently reconstructed, and consequently the n/k spectra in Fig. 4 cannot be assessed for artifacts.
- [Figure 5 and the abstract's agreement claim] The claim that the SE-derived 'simulated' bandgap (2.99 eV) agrees with experiment is partly circular. The k spectrum used for the Tauc plot in Figs. 5(c) and 5(d) is generated from the Tauc-Lorentz model whose fitted Eg values are 2.97 eV (sample A) and 2.95 eV (sample B) in Table I; a Tauc-type analysis of that k will return approximately the fitted Eg by construction. The only independent validation is the UV-Vis Tauc plot, which does give 2.99 eV. I ask the authors to remove the SE-derived Tauc plot as a validation, or replace it with a bandgap estimate from the unparameterized B-spline dielectric function, and to explicitly reconcile the Table I Eg values with the reported 2.99 eV.
- [Measurement configuration and parameter correlation] The SE data were acquired at a single angle of 75 degrees. For a multilayer stack that includes SiO2, the perovskite layer, and an EMA roughness layer, thickness and optical constants are strongly correlated in single-angle data. Low MSE values do not by themselves establish uniqueness. The authors should report multiple-angle SE measurements, or add an independent constraint such as spectrophotometric transmittance, or provide confidence limits and parameter correlation information from the fitting software to support the uniqueness of the extracted n/k spectra.
- [Roughness layer and substrate assumptions] The roughness layer is modeled as a Bruggeman EMA with a fixed 50% void fraction, with no justification or sensitivity test. Since the roughness layer sits directly above the perovskite layer, this fixed assumption can bias both the film thickness and the extracted n/k values. I request that the void fraction be allowed to vary, or that the authors show explicitly that the n/k results are insensitive to it. The same applies to the fixed parameters of the substrate and interface layers.
- [B-spline intermediate step] The manuscript states that the Cauchy layer was converted to a Kramers-Kronig-constrained B-spline before the final oscillator parameterization, but the B-spline dielectric function is never shown or compared with the final Tauc-Lorentz-plus-harmonic-oscillator result. Without that comparison, it is impossible to tell whether the two harmonic oscillators are required by the data or are fitting artifacts. Please include the B-spline n/k or complex dielectric function and the difference between the B-spline and oscillator model over the fitted range.
minor comments (5)
- [Eq. (5)] The printed Tauc-Lorentz formula is garbled: the denominator appears as '(E-E_on^2)^2' and the trailing '1/E' is misplaced. It should read e.g. epsilon_2 = A E_o B_r (E - E_g)^2 / [(E^2 - E_o^2)^2 + B_r^2 E^2] * 1/E for E > E_g. Please correct the typesetting.
- [Eq. (2)] The MSE expression is written with '1p-q'; the standard definition uses '2p-q' because each wavelength point has two measured quantities (Psi and Delta). Please correct the denominator and define p and q explicitly.
- [Figure 1 caption] The caption of Fig. 1 is incomplete; it ends with 'Normalized UV-Vis absorption spectra of the CsPbCl3' and lacks a period and the words 'thin films' (or similar).
- [References] References 33 and 35 are the same paper (Yim et al., Appl. Phys. Lett. 104, 103114 (2014)); one should be removed and the citations renumbered.
- [Equation numbering] The absorption coefficient formula is referred to as 'Equation 9', but no Eqs. (6)-(8) are present in the text. Please renumber or remove the label.
Circularity Check
The SE-derived 'simulated' bandgap is partly a restatement of the fitted Tauc-Lorentz Eg, but the core 2.99 eV claim is independently supported by UV-Vis, so the circularity is partial.
-
fitted input called prediction
[Section on bandgap determination, around Eq. (9) and Fig. 5, page 10 of the manuscript.]
"To validate the results, the simulation data of the extinction coefficient determined from the optical model were also used to extract the optical bandgap of both samples. Equation 9 was used to calculate the absorption coefficient (α) from the extinction coefficient obtained from SE measurements: α = 4πk/λ (9). Figures 5c and d show that the bandgap values of 2.99eV derived from the simulated results closely align with the calculated values for samples A and B, respectively. The optical bandgap values also align well with the literature."
The 'simulated' extinction coefficient is generated by the Tauc-Lorentz oscillator model, whose fitted parameters in Table I include the band-gap parameter Eg (2.97 eV and 2.95 eV for samples A and B). Equation (5) forces ε2 = 0 for E ≤ Eg and imposes a Tauc-like (E-Eg)² energy dependence above Eg. Converting this model k to α and re-extracting a Tauc bandgap therefore cannot be an independent validation: the resulting ~2.99 eV is essentially a reflection of the fitted Eg parameter, slightly shifted by the Tauc-plot extrapolation. Presenting this as agreement between 'experimental values' and 'simulation results' is thus partly by construction. The genuinely independent anchor for 2.99 eV is the UV-Vis Tauc plot, not the SE-based 'simulation'.
full rationale
The paper's main derivation chain is a standard ellipsometry inversion: Cauchy thickness estimate, B-spline dielectric function, then parameterization by a Tauc-Lorentz oscillator plus two harmonic oscillators, yielding n and k. That procedure is not circular in itself. The circular step is the 'validation' of the model by comparing the UV-Vis Tauc gap to a Tauc gap extracted from the simulated k: the simulated k is produced by a model that already contains Eg as a fitted parameter, so the agreement of the simulated Tauc gap with the fitted Eg is forced by the model form. This is a genuine fitted-input-called-prediction step. However, the central 2.99 eV bandgap claim does not rest solely on that step: the paper reports independent UV-Vis absorption Tauc plots giving 2.99 eV for both samples, and the SE k edge at 411 nm is consistent with the measured absorption feature. Those external benchmarks keep the paper from being wholly circular. The apparent irregularities in Table II (unphysical oscillator amplitudes, zero broadening, and duplicated parameters) are serious reproducibility and model-reliability concerns, but they are not circularity, so they do not further raise the circularity score; they belong in a correctness/reproducibility review.
Assumptions & free parameters
free parameters (8)
- Tauc-Lorentz Eg sample A =
2.97 eV
- Tauc-Lorentz Eg sample B =
2.95 eV
- Tauc-Lorentz A, Br, Eo, Einf for samples A and B =
Values in Table I
- Harmonic oscillator parameters for samples A and B =
Values in Table II
- Film thicknesses =
92.73 nm (A), 187.18 nm (B)
- Roughness EMA layer thickness =
not reported
- Substrate layer thicknesses (Si, interface, SiO2) =
not reported
- Cauchy model coefficients A, B, C =
not reported
assumptions (5)
- domain assumption The Tauc-Lorentz model plus two harmonic oscillators adequately represents the dielectric function of CsPbCl3 over 300-900 nm.
- domain assumption The surface roughness can be modeled by a 50% air/50% film Bruggeman EMA layer.
- domain assumption The underlying Si/SiO2 substrate optical properties are accurately known from a prior three-layer fit.
- domain assumption The films are homogeneous, isotropic, and have abrupt parallel interfaces.
- domain assumption The Tauc plot method with (αhν)^2 versus hν is valid for extracting the direct bandgap.
Cite this review
Pith. "Pith review of Spectroscopic ellipsometry of CsPbCl${_3}$ perovskite thin films." pith.science (2026). https://pith.science/paper/R3VVZLCU
@misc{pith2026241205618,
author = {Pith},
title = {Pith review of: Spectroscopic ellipsometry of CsPbCl$_3$ perovskite thin films},
year = {2026},
howpublished = {\url{https://pith.science/paper/R3VVZLCU}},
note = {Machine review of arXiv:2412.05618}
}
abstract
Designing optoelectronic devices based on cesium lead chloride (CsPbCl${_3}$) perovskites requires accurate values of their optical constants. Unfortunately, experimental data for this material is very limited thus far. Therefore, here, we applied spectroscopic ellipsometry (SE) to measure the complex optical constants of thermally evaporated CsPbCl${_3}$ thin films with different thicknesses on Si/SiO${_2}$ substrates. The data were corroborated with scanning electron microscopy (SEM) images and absorption spectroscopy. An optical dispersion model was developed to derive the complex optical constants and film thicknesses. The Tauc-Lorentz model, in conjunction with two harmonic oscillators, was used to extract the required parameters. The extinction coefficient spectrum exhibited a sharp absorption edge at 411 nm, consistent with the absorption spectrum. In addition, the optical bandgap of the film was calculated from the absorption spectra and SE data. The experimental values agree well with the simulation results, with values of $\sim$ 2.99 eV for different film thicknesses. This work provides fundamental information for designing and modeling CsPbCl3-based optoelectronic devices.
Figures
Figures from the paper (2 more)
Reference graph
Works this paper leans on
-
[1]
Gao, M. Kraft, J.A. Steele, E. Debroye, J. Hofkens, and R. Ameloot, “Flexible Metal Halide Perovskite Photodetector Arrays via Photolithography and Dry Lift‐Off Patterning,” Adv Eng Mater 24(1), 2100930 (2022). 2 P.J. Cegielski, A.L. Giesecke, S. Neutzner, C. Porschatis, M. Gandini, D. Schall, C.A.R
work page 2022
-
[2]
Perini, J. Bolten, S. Suckow, S. Kataria, B. Chmielak, T. Wahlbrink, A. Petrozza, and M.C. Lemme, “Monolithically Integrated Perovskite Semiconductor Lasers on Silicon Photonic Chips by Scalable Top-Down Fabrication,” Nano Lett. 18(11), 6915–6923 (2018). 3 P.J. Cegielski, S. Neutzner, C. Porschatis, H. Lerch, J. Bolten, S. Suckow, A.R.S. Kandada, B. Chmie...
work page 2018
-
[3]
Perovskite light- emitting diodes with external quantum efficiency exceeding 20 per cent,
Yan, W. Li, X. Liu, Y . Lu, J. Kirman, E.H. Sargent, Q. Xiong, and Z. Wei, “Perovskite light- emitting diodes with external quantum efficiency exceeding 20 per cent,” Nature 562(7726), 245–248 (2018). 7 Z. Chen, B. Turedi, A.Y . Alsalloum, C. Yang, X. Zheng, I. Gereige, A. AlSaggaf, O.F. Mohammed, and O.M. Bakr, “Single-Crystal MAPbI 3 Perovskite Solar Ce...
work page 2018
-
[4]
Recent Progress on Electrical and Optical Manipulations of Perovskite Photodetectors,
Wang, J. Miao, F. Chen, J. Wang, X. Chen, A. Pan, C. Shan, L. Liao, and W. Hu, “Recent Progress on Electrical and Optical Manipulations of Perovskite Photodetectors,” Advanced Science 8(14), 2100569 (2021). 9 C. Liu, Q. Zeng, H. Wei, Y . Yu, Y . Zhao, T. Feng, and B. Yang, “Metal Halide Perovskite Nanocrystal Solar Cells: Progress and Challenges,” Small M...
work page 2021
-
[5]
Kowalsky, H.-C. Scheer, P. Görrn, and T. Riedl, “Room-Temperature Stimulated Emission and Lasing in Recrystallized Cesium Lead Bromide Perovskite Thin Films,” Advanced Materials 31(39), 1903717 (2019). 11 M. Gong, R. Sakidja, R. Goul, D. Ewing, M. Casper, A. Stramel, A. Elliot, and J.Z. Wu, “High-Performance All-Inorganic CsPbCl 3 Perovskite Nanocrystal P...
work page 2019
-
[6]
Unruh, N. Li, and C.J. Brabec, “Revealing the strain-associated physical mechanisms impacting the performance and stability of perovskite solar cells,” Joule 6(2), 458–475 (2022). 19 H. Li, C. Cui, X. Xu, S. Bian, C. Ngaojampa, P. Ruankham, and A.P. Jaroenjittchai, “A review of characterization of perovskite film in solar cells by spectroscopic ellipsomet...
work page 2022
-
[7]
Publishing, Cham, 2018). 21 P. Löper, M. Stuckelberger, B. Niesen, J. Werner, M. Filipič, S.-J. Moon, J.-H. Yum, M. Topič, S. De Wolf, and C. Ballif, “Complex Refractive Index Spectra of CH 3 NH 3 PbI 3 Perovskite Thin Films Determined by Spectroscopic Ellipsometry and Spectrophotometry,” J. Phys. Chem. Lett. 6(1), 66–71 (2015). 22 M.S. Alias, I. Dursun, ...
work page 2015
-
[8]
Growth, Carrier Transport Properties, and Self‐Powered Photodetection,” Advanced Optical Materials 5(2), 1600704 (2017)
work page 2017
Reviewed August 11, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.