The DOI matches its Crossref record; 50 of the 50 references it cites were found in a registry or an index; 21 of the 29 sources its claims cite were checked for support.
A New Metric for the Comparison of Permittivity Models in Terahertz Time-Domain Spectroscopy
Open accessAbstract
Wepresent a robust method, as well as a new metric, for the comparison of permittivity models in terahertz time-domain spectroscopy (THz-TDS). In this article, we show that the signal is perturbed by a deterministic part that has to be either modeled or removed. Therefore, we perform an extensive perturbation/noise analysis of a THz-TDS system, remove and model the unwanted deterministic perturbation/noises and implement them into our fitting process. This new method has now been implemented in our open-source software, Fit@TDS. This article is the first step toward the derivation of uncertainties, and therefore the use of error bars. We hope that this will lead to performing analytical analysis with THz-TDS, as results obtained from different setups will be comparable. Finally, we apply this protocol to the study of a$\alpha$-lactose monohydrate pellet in order to give more insight on the molecular dynamics behind the absorption peaks. The comparison with simulation results is made easier thanks to the probabilities derived from the metric.
Overview
CoChat Summary
TL;DR
The paper introduces a time-domain fitting protocol and a new metric, based on the Akaike criterion, to compare permittivity models in THz-TDS while accounting for deterministic and correlated noise. It applies the method to an α-lactose monohydrate pellet, showing that a Voigt model is preferred for the first absorption peak and that doublet Voigt models are more probable for some higher-frequency peaks.
Key findings
- The authors show that much of the experimental noise in their THz-TDS setup is deterministic or artifact-noise and must be removed or modeled before building a noise correlation matrix; dark noise decreased from 5·10⁻⁶ µV/√Hz at 50 averages to 5·10⁻⁷ µV/√Hz at 5000 averages.
- A high-pass filter with a 200 GHz cutoff removed the low-frequency deterministic THz-dark artifact-noise, after which its contribution was of the same order as dark noise.
- For α-lactose monohydrate, the final Akaike criterion was 2.028·10⁶, strongly lower than the base criterion of 6.1·10⁸; the first peak was best fit by a Voigt model with Δε = 0.046, ω₀ = 0.531 THz, γ = 23.17 GHz, σ = 4.280 GHz.
- The 1.37 THz peak was better represented by a doublet Voigt model, and the probability derived from the Akaike criterion confirmed the doublet model as more probable; the final fit also retrieved ε∞ = 2.739 and pellet thickness 599.5 µm.
Methods
- The study uses a commercial Menlosystems Terasmart THz-TDS setup in a nitrogen-purged box, with a 100 MHz laser repetition rate, to measure a pressed α-lactose monohydrate pellet (13 mm diameter, 600 µm thickness, no dilution) and a reference without pellet.
- The protocol pre-processes time traces by removing deterministic noise, applying a 200 GHz high-pass filter, correcting delay drift, and fitting leftover noise phenomenologically; two consecutive measurements of reference and sample are used to build noise correlation matrices.
- Fitting is performed in the time domain with Fit@TDS using a dual annealing algorithm and a noise-weighted error function; models are compared peak by peak using the Akaike criterion, testing Lorentz, Voigt, doublet Voigt, and a continuum model for high-frequency losses.
Limitations
- The paper states that only five measurements were taken, that the acquisition software does not allow recording spectra one by one, and that the leftover noise is fitted with a phenomenological model that needs further investigation.
- The authors state the final fit is not perfect, that above 2.6 THz the fitted curve strays from experimental data, and that the Lorentz and Voigt models may no longer describe the dynamics correctly there.
- The paper states it did not perform direct statistical comparisons between models and that the distinctions drawn are not definitive; it also notes that scattering, non-perpendicular incidence, and pellet fabrication effects are not taken into account.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
20 Valid · 4 Not valid · 4 Not checked · 48 Own results and interpretations
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“As a result, such algorithm struggles to distinguish between accurate and inaccurate models, as explained in [30], [31].”
What the sources say
- Supportedcites [30] source
Section 3 states that the ETFE fitting procedure offers poor discrimination between good and bad models, even when it produces accurate parameter estimates. This directly supports the attributed claim about distinguishing accurate from inaccurate models. (LLM, conf=0.94)
- Not checked — no full text of the sourcecites [31] source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This starts by the fact that, so far, the experimental noise is not correctly taken into account during the fitting procedure as raised by [30].”
What the sources say
- Partly supportedcites [30] sourceExperimental — accuracy not yet measured
Section 3 (page unavailable) says least-squares fits implicitly restrict noise to the output variable, although input noise is always present in THz-TDS, and that constant-weight fits can bias parameter estimates. This supports the criticism of those fitting procedures, but the claim is broader than the paper’s finding: the paper also presents a maximum-likelihood procedure intended to address the problem. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Still, the results obtained match the data available in the literature [40], [41].”
What the sources say
- Supportedcites [40] source
Section 2.2 (Explosives) says that spectral features measured by different groups generally agree and compares RDX and C4 datasets. Table 1 also compiles reported absorbance peaks, providing literature data for the comparison. The cited paper cannot independently verify the citing authors’ results, but it supports the attributed point that comparable data are available in the literature. (LLM, conf=0.84)
- Not checked — no full text of the sourcecites [41] source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Thus, when the signal includes any noise correlated to the signal, the algorithm mistakenly incorporates it into the analysis, potentially confusing it with genuine signal attributes.”
What the sources say
- Partly supportedcites [30] sourceExperimental — accuracy not yet measured
Section 3 reports that correlated noise can affect frequencies within the signal bandwidth and produce fitted residuals that could be mistaken for physical features. Section 6 shows a conventional ETFE fit favoring extra parameters that erroneously suggest a real physical effect. The claim overgeneralizes this problem to any signal-correlated noise and an unspecified algorithm; the paper finds its TLS method more robust against such overfitting. (LLM, conf=0.86)
- Not checked — no full text of the sourcecites [31] source
hedgedstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“When the input noise is convoluted by the transfer function (see equation 1), its contribution is found in all of the Fabry-Perot echoes in the sample time-trace.”
What the sources say
- Partly supportedcites [30] sourceExperimental — accuracy not yet measured
Section 3, “Transfer function estimation in THz-TDS” (page unavailable), states that the impulse response is convolved with the measured input and explicitly identifies a convolved input-noise term. Our copy does not establish the more specific claim that this noise appears in all Fabry–Perot echoes of a sample time-trace. (LLM, conf=0.83)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
15 more claims · 1 with their cited sources checked
Create a free account to see all 76 claims and every source verdict.
- Probably not a substantial claim. 32
- It may state two separate claims as one. 32
- The quote only partly states this claim. 2
Citations
Cites
Far-infrared time-domain spectroscopy with terahertz beams of dielectrics and semiconductors
Grischkowsky, D., Keiding, S., van Exter, M., & Fattinger, C. (1990) · Journal of the Optical Society of America B · doi:10.1364/josab.7.002006 · Cited by 2,239 · cited 1×
“…Spectroscopy Melanie Lavancier , Nabil Vindas-Yassine, Juliette Vlieghe, Theo Hannotte , Jean-François Lampin, François Orieux , and Romain Peretti Abstract —Wepresent a robust method, as well as a new metric, forthecomparisonofpermittivitymodelsinterahertztime-domain spectroscopy(THz-TDS).”
MethodTHz imaging and sensing for security applications—explosives, weapons and drugs
Federici, J. F., Schulkin, B., Huang, F., Gary, D., Barat, R., Oliveira, F., & Zimdars, D. (2005) · Semiconductor Science and Technology · doi:10.1088/0268-1242/20/7/018 · Cited by 1,756 · cited 2×
“Indeed, [40] find that the peaks are located at 0.54, 1.20, 1.38, 1.82, 2.54, 2.87, and 3.29 THz.”
InfluentialKindt, J. T., & Schmuttenmaer, C. A. (1996) · The Journal of Physical Chemistry · doi:10.1021/jp960141g · Cited by 687 · cited 1×
“Most of the fits are done in the frequency domain using a least-square algorithm on the retrievedpermittivitycurvesfromexperimentaltime-traces[20], [21],[22],[23],[24],[25].”
MethodA model for the complex permittivity of water at frequencies below 1 THz
Liebe, H. J., Hufford, G. A., & Manabe, T. (1991) · International Journal of Infrared and Millimeter Waves · doi:10.1007/bf01008897 · Cited by 686 · cited 1×
“Most of the fits are done in the frequency domain using a least-square algorithm on the retrievedpermittivitycurvesfromexperimentaltime-traces[20], [21],[22],[23],[24],[25].”
MethodRo/nne, C., Thrane, L., Åstrand, P.-O., Wallqvist, A., Mikkelsen, K. V., & Keiding, S. R. (1997) · The Journal of Chemical Physics · doi:10.1063/1.474242 · Cited by 587 · cited 1×
“Index Terms —Electromagnetic modeling, refractive index, spectroscopy, terahertz (THz) materials, terahertz metamaterials.”
BackgroundBaxter, J. B., & Guglietta, G. W. (2011) · Analytical Chemistry · doi:10.1021/ac200907z · Cited by 477 · cited 1×
“T ERAHERTZ (THz) spectroscopy of materials, ranging from 0.1 to 10 THz (i.e., 3–333 cm − 1 , i.e., 0.1–10 ps), witnesses many chemicophysical processes in solids, liquids, and gases [1], [2].”
BackgroundInfluentialTerahertz Spectroscopy and Imaging
(2013) · Springer Series in Optical Sciences · doi:10.1007/978-3-642-29564-5 · Cited by 323
Conti Nibali, V., & Havenith, M. (2014) · Journal of the American Chemical Society · doi:10.1021/ja504441h · Cited by 314 · cited 1×
“Therefore,weperformanextensiveperturbation/noiseanalysisofaTHz-TDSsystem,removeandmodeltheunwanted deterministicperturbation/noisesandimplementthemintoourfittingprocess.”
Yada, H., Nagai, M., & Tanaka, K. (2008) · Chemical Physics Letters · doi:10.1016/j.cplett.2008.09.015 · Cited by 283
Dynamic conductivity and coherence peak inYBa2Cu3O7superconductors
Nuss, M. C., Mankiewich, P. M., O’Malley, M. L., Westerwick, E. H., & Littlewood, P. B. (1991) · Physical Review Letters · doi:10.1103/physrevlett.66.3305 · Cited by 259 · cited 1×
“D materials [6], probing the gap of superconductors [7], [8], [9], [10], and interrogating the phonon dynamics in molecular crystals [11].”
Background
Cited by
2025 photonics for agrifood roadmap: towards a sustainable and healthier planet
Smeesters, L., Venturini, F., Paulus, S., Mahlein, A., Perpetuini, D., Cardone, D., Merla, A., Gac, B., Perrin, S., Trégoat, D., et al. (2025) · Journal of Physics Photonics · doi:10.1088/2515-7647/adbea9 · Cited by 14
Self-citationSignal Estimation and Uncertainties Extraction in Terahertz Time-Domain Spectroscopy
Denakpo, E., Hannotte, T., Osseiran, N., Orieux, F., & Peretti, R. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3554287 · Cited by 6
“Peretti et al. [26] and Lavancier et al. [27], [28] proposed a software called Fit@TDS, which implements a method where the permittivity models are directly fit from the data in the time domain and includes the noise correlation matrix.”
Self-citationTowards the elimination of water in THz-TDS data
Osseiran, N., Bichon, J., Eliet, S., & Peretti, R. (2024) · International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz) · doi:10.1109/irmmw-thz60956.2024.10697539 · Cited by 3
Self-citationOsseiran, N., Bichon, J., Hannotte, T., Abdel-Kader, M., Witt, M., Eliet, S., & Peretti, R. (2024) · International Conference on Infrared, Millimeter, and Terahertz Waves (IRMMW-THz) · doi:10.1109/irmmw-thz60956.2024.10697647 · Cited by 1
Self-citationLi, G., Ge, H., Jin, X., et al. (2026) · Journal of Food Composition and Analysis · doi:10.1016/j.jfca.2026.109486
Metadata
- Authors
- Lavancier, M., Vindas-Yassine, N., Vlieghe, J., Hannotte, T., Lampin, J.-F., Orieux, F., & Peretti, R.
- Affiliations
- Univ. Lille, CNRS, Univ. Polytechnique Hauts-de-France, UMR 8520 IEMN - Institut d'Electronique de Microélectronique et de Nanotechnologie, Lille, France
- Centre National de la Recherche Scientifique (FR)
- ORCID
- Venue
- IEEE Transactions on Terahertz Science and Technology
- Published
- 2024-09
- Volume
- 14
- Issue
- 5
- Pages
- 725-733
- ISSN
- 2156-342X (ISSN-L)
- 2156-3446
- OpenAlex
- W4392616957
- Semantic Scholar
- 55cab2caf6f92be7b438f876fe6364155e9416e2
- Cited by
- 5
- References
- 44
- Citations by year
- 2026: 1 · 2025: 2 · 2024: 2
- FWCI
- 0.52
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE Transactions on Terahertz Science and Technology (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- HAL (submittedVersion, other-oa)
- HAL (submittedVersion, cc-by)
- arXiv (Cornell University)
- Funders
- I-site ULNE
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Terahertz technology and applications
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Microwave and Dielectric Measurement Techniques
- Physical Sciences › Chemistry › Spectroscopy › Spectroscopy and Laser Applications
- Keywords
- deterministic perturbations, noise analysis, uncertainty quantification, error bars, alpha-lactose monohydrate, terahertz absorption, molecular dynamics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: A New Metric for the Comparison of Permittivity Models in Terahertz Time-Domain Spectroscopy. https://app.cochat.ai/papers/10.1109/tthz.2024.3431202
- The paper
- Lavancier, M., Vindas-Yassine, N., Vlieghe, J., Hannotte, T., Lampin, J.-F., Orieux, F., & Peretti, R. (2024). A New Metric for the Comparison of Permittivity Models in Terahertz Time-Domain Spectroscopy. IEEE Transactions on Terahertz Science and Technology. https://doi.org/10.1109/tthz.2024.3431202
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.