Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 29 of the 29 references it cites were found in a registry or an index; 6 of the 21 sources its claims cite were checked for support.

Unveiling the Origin of Spurious Features in THz-TDS of Powder Compacts

Open access
Gorecki, Jon, Murphy, Keir N., Markl, Daniel, Burnett, Andrew D., Naftaly, Mira · 2025 · IEEE Transactions on Terahertz Science and Technology

Abstract

Terahertz time-domain spectroscopy (THz-TDS) is a vital tool for scientific and industrial analysis, however, many commonly analyzed products, such as those found in pharmaceutical, agriculture, and mining sectors, are produced as powders or granular materials, and these sample morphologies have been reported to produce anomalous spectral features which can often obscure known material resonances. The cause of these anomalous features has been poorly understood, making it difficult to predict their presence and limiting the applicability of THz–TDS for such materials. Here, we systematically study how the sample morphology of granular compacts produces anomalous spectral features by performing extensive experimental measurements on two-part powder compacts with varying microsphere size and concentration. Further, we employ ray-tracing simulations to identify the physical mechanism whereby these spectral features arise owing to variations in optical path length within the heterogeneous sample. We believe this is the first time that the physical cause of spurious spectral features within powder samples has been adequately explained and that a robust method has been presented for modeling this effect. By understanding these features, we propose that instead of being seen as a parasitic effect, their presence can be utilised to extract morphological properties of the samples, thereby enhancing the utility of THz-TDS for granular materials.

Overview

CoChat Summary

TL;DR

The paper explains spurious loss features in THz-TDS of granular compacts as arising from optical path-length differences: rays group in time by number of microsphere scattering events, producing temporal peaks separated by Δt = ΔnD/c and loss peaks at f0 = 1/(2Δt). Ray-tracing simulations and experiments on Borofloat microspheres in PTFE validate this mechanism and show the features depend on microsphere size and concentration, with potential to use them for morphology estimation.

Key findings

  • Spurious loss peaks are caused by rays becoming grouped in time according to the number of Borofloat microspheres they pass through; the temporal peak separation depends on microsphere diameter and refractive-index difference, not concentration. For 98 µm microspheres the separation was 0.21 ps, giving f0 = 2.4 THz; for 196 µm microspheres it was 0.42 ps, giving f0 = 1.2 THz.
  • Increasing microsphere concentration increases the number of temporal peaks and produces a spectral splitting effect around f0, with multiple loss peaks appearing symmetrically around the central frequency; intra-batch variation also increases, especially for smaller microspheres and at 20–30 % concentration.
  • Simulations show that reducing microsphere diameter below 40 µm removes sharp temporal peaks and the associated loss peaks; experimentally, 41 µm microspheres gave a much smoother loss spectrum than 98 µm microspheres, though simulations and experiments diverged below 100 µm.
  • The model explains spectral features in region 1 (between 1–3.5 THz), but features in regions 2 (~3.7–3.8 THz) and 3 (~4.2 THz) remain unexplained and may require effective medium theories.

Methods

  • Borofloat glass microspheres of several size distributions (38–45, 90–106, 125–150, 150–180, 180–212 µm diameter; 2.2 g/cm³) were blended with PTFE powder (mean particle size 1 µm; 2.2 g/cm³), compacted at 400 MPa into 9 mm diameter, 1 mm thick disks, and allowed one week to relax; pure PTFE was used as a control. THz-TDS transmission measurements used a Teraflash Pro (Toptica) spectrometer purged with dry air, extracting effective refractive index and loss coefficient.
  • A bespoke 2D ray-tracing simulation in Matlab modeled a 1 mm deep PTFE slab with randomly distributed Borofloat microspheres (5 µm pixels). One million rays started at random x positions with incidence angles uniformly distributed between ±20°, scattered at microspheres using an angular distribution derived from COMSOL for a 200 µm microsphere, and were binned into 600 bins over 0–20 ps before Fourier transformation.
  • Simulated and experimental temporal profiles, loss spectra, and refractive-index spectra were compared across microsphere sizes and concentrations; an analytical expression f0 = 1/(2Δt) = c/(2ΔnD) was used to predict loss-peak frequencies, and experimental loss-peak frequencies at 5 % concentration were compared with the model.

Limitations

  • Stated: simulations and experiments diverge for microspheres below 100 µm; simulations suggest loss increases as microsphere size decreases, while experiments show the opposite. The ray-tracing method may become invalid when microsphere size is far smaller than the incident wavelengths, where diffraction may dominate, and the scattering-angle distribution is identical regardless of microsphere size, which becomes invalid for smaller objects.
  • Stated: features in spectral regions 2 and 3 remain unexplained; bifurcations in refractive index are of unknown cause, possibly phase-unwrapping errors or a physical mechanism. Experimental refractive index below 1.44 at low concentrations is erroneous and possibly due to porosity, and other sample preparation methods may produce varying results because of porosity.
  • Evident from the text: the experimental frequency range is limited to 4 THz by the dynamic range of the THz-TDS instrumentation; the analytical validation uses limited experimental data points; the study is restricted to the Borofloat/PTFE system; and the 2D simulation approximates a focused Gaussian beam with a uniform ±20° incidence-angle distribution.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1109/tthz.2025.3549944

Cites

29 of 29 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 (a) Loss coefficient of solid Borofloat glass and a PTFE powder compact, both of 1 mm thickness. (b) Loss coefficient spectra of 20 % v/v compacts with varied microsphere size revealing several anomalous features. The maximum measurable loss is depicted as a dashed line. (c) Real refractive index for Borofloat glass sample, and PTFE powder compact sample, revealing smooth flat spectra. (d) Real refractive index of 20 % v/v compacts with varied microsphere size revealing several anomalous features.
Fig. 1 (a) Loss coefficient of solid Borofloat glass and a PTFE powder compact, both of 1 mm thickness. (b) Loss coefficient spectra of 20 % v/v compacts with varied microsphere size revealing several anomalous features. The maximum measurable loss is depicted as a dashed line. (c) Real refractive index for Borofloat glass sample, and PTFE powder compact sample, revealing smooth flat spectra. (d) Real refractive index of 20 % v/v compacts with varied microsphere size revealing several anomalous features.From the PDF
Fig. 2 THz-TDS loss spectra of powder compacts of Borofloat microspheres in PTFE. The maximum measurable loss is depicted as a dashed line. Three spectral regions are highlighted in grey boxes, labelled 1, 2 and 3, grouping together spectral features of similar visual appearance to aid discussion. (a & b) Varying microsphere size, at 5 % and 20 % concentration, respectively. (c & d) Varying microsphere concentration, at sizes of 90 - 106 µm and 180 212 µm , respectively.
Fig. 2 THz-TDS loss spectra of powder compacts of Borofloat microspheres in PTFE. The maximum measurable loss is depicted as a dashed line. Three spectral regions are highlighted in grey boxes, labelled 1, 2 and 3, grouping together spectral features of similar visual appearance to aid discussion. (a & b) Varying microsphere size, at 5 % and 20 % concentration, respectively. (c & d) Varying microsphere concentration, at sizes of 90 - 106 µm and 180 212 µm , respectively.From the PDF
Fig. 3 THz-TDS loss spectra of Borofloat microspheres at 10 % v/v and 90 - 106 µm for each compact. The maximum measurable loss is plotted in a dashed line. We highlight three spectral regions (numbered 1, 2, 3) in which we group together spectral features of similar visual appearance to aid discussion.
Fig. 3 THz-TDS loss spectra of Borofloat microspheres at 10 % v/v and 90 - 106 µm for each compact. The maximum measurable loss is plotted in a dashed line. We highlight three spectral regions (numbered 1, 2, 3) in which we group together spectral features of similar visual appearance to aid discussion.From the PDF
Fig. 4 Schematic diagram of geometry for the ray-tracing simulation. The sample consists of a slab of PTFE with randomly placed Borofloat microspheres. Rays begin at randomised locations on the top surface of the slab and are traced through, slowing down and deflecting when they travel through the Borofloat. The exit time and intensity of the rays are recorded to generate the temporal profile.
Fig. 4 Schematic diagram of geometry for the ray-tracing simulation. The sample consists of a slab of PTFE with randomly placed Borofloat microspheres. Rays begin at randomised locations on the top surface of the slab and are traced through, slowing down and deflecting when they travel through the Borofloat. The exit time and intensity of the rays are recorded to generate the temporal profile.From the PDF
Fig. 5 Flow chart depicting each stage of the ray tracing simulation and data analysis pipeline, highlighting the steps involved to produce subsequent figures.
Fig. 5 Flow chart depicting each stage of the ray tracing simulation and data analysis pipeline, highlighting the steps involved to produce subsequent figures.From the PDF
Fig. 6 (a) Simulated ray-counts and (b) experimental system response for varying microsphere concentrations at 98 µm diameter. (c) Simulated ray-counts and (d) experimental system response for varying microsphere concentrations at 196 µm diameter.
Fig. 6 (a) Simulated ray-counts and (b) experimental system response for varying microsphere concentrations at 98 µm diameter. (c) Simulated ray-counts and (d) experimental system response for varying microsphere concentrations at 196 µm diameter.From the PDF
Fig. 7 (a) Simulated and (b) experimental loss spectra for varying microsphere concentrations at 98 µm diameter. (c) Simulated and (d) experimental loss spectra for varying microsphere concentrations at 196 µm diameter.
Fig. 7 (a) Simulated and (b) experimental loss spectra for varying microsphere concentrations at 98 µm diameter. (c) Simulated and (d) experimental loss spectra for varying microsphere concentrations at 196 µm diameter.From the PDF
Fig. 8 (a) Simulated and (b) experimental refractive index spectra for varying microsphere concentrations at 98 µm diameter. (c) Simulated and (d) experimental refractive index spectra for varying microsphere concentrations at 196 µm diameter.
Fig. 8 (a) Simulated and (b) experimental refractive index spectra for varying microsphere concentrations at 98 µm diameter. (c) Simulated and (d) experimental refractive index spectra for varying microsphere concentrations at 196 µm diameter.From the PDF
Fig. 9 The modified ray-tracing simulation records the number of scattering events S that each ray undergoes. Temporal profiles are then produced for each sub-set of rays according to scattering number. Microspheres simulated with 20 % concentration at (a) 98 µm diameter exhibit several sharp peaks, while (b) 12 µm diameter microspheres exhibit a smooth broad distribution. For the 12 µm microspheres the scattering events are plotted in integer steps from 0 to 15. The modified ray-tracing simulation does not include absorption effects.
Fig. 9 The modified ray-tracing simulation records the number of scattering events S that each ray undergoes. Temporal profiles are then produced for each sub-set of rays according to scattering number. Microspheres simulated with 20 % concentration at (a) 98 µm diameter exhibit several sharp peaks, while (b) 12 µm diameter microspheres exhibit a smooth broad distribution. For the 12 µm microspheres the scattering events are plotted in integer steps from 0 to 15. The modified ray-tracing simulation does not include absorption effects.From the PDF
Fig. 10 (a) Time domain and (b) frequency domain of signals containing increasing number of Dirac delta functions at 0.21 ps separation. For two Dirac delta functions the amplitude spectra displays a dip at a frequency f 0 = 1 / 2∆ t = 0.24 THz, which repeats with a period of 2 f 0 . As the number of Dirac delta functions increases the amplitude dips split symmetrically around the central frequency f 0 , as highlighted by the curved arrows, where the number of splitting features increases with the number of Dirac delta functions.
Fig. 10 (a) Time domain and (b) frequency domain of signals containing increasing number of Dirac delta functions at 0.21 ps separation. For two Dirac delta functions the amplitude spectra displays a dip at a frequency f 0 = 1 / 2∆ t = 0.24 THz, which repeats with a period of 2 f 0 . As the number of Dirac delta functions increases the amplitude dips split symmetrically around the central frequency f 0 , as highlighted by the curved arrows, where the number of splitting features increases with the number of Dirac delta functions.From the PDF
Fig. 11 (a) Time domain, (b) loss, and (c) refractive index spectra for microspheres simulated at 5 % concentration with varying microsphere diameter. Experimentally measured (c) loss and (d) refractive index spectra for microspheres at 98 and 41 µm diameter.
Fig. 11 (a) Time domain, (b) loss, and (c) refractive index spectra for microspheres simulated at 5 % concentration with varying microsphere diameter. Experimentally measured (c) loss and (d) refractive index spectra for microspheres at 98 and 41 µm diameter.From the PDF
Fig. 12 Refractive index difference between matrix and microsphere ∆ n is plotted against spectral frequency f 0 and microsphere diameter utilising Equation 4. The upper and lower values of ∆ n plotted in black corresponds to the case of Borofloat microspheres surrounded by PTFE matrix. Experimental loss-peak frequencies for samples at 5 % concentration of plotted in red squares.
Fig. 12 Refractive index difference between matrix and microsphere ∆ n is plotted against spectral frequency f 0 and microsphere diameter utilising Equation 4. The upper and lower values of ∆ n plotted in black corresponds to the case of Borofloat microspheres surrounded by PTFE matrix. Experimental loss-peak frequencies for samples at 5 % concentration of plotted in red squares.From the PDF
Image 13
From the PDF
Image 14
From the PDF
Image 15
From the PDF

Verification

Claims the paper makes

6 Valid · 2 Not valid · 7 Not checked · 128 Own results and interpretations

72 more claims

Create a free account to see all 143 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 13
  • It may state two separate claims as one. 59

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

4 of 4

Citations

Cites

Show all 27 references

Cited by

See all 4

Metadata

Authors
Gorecki, J., Murphy, K. N., Markl, D., Burnett, A. D., & Naftaly, M.
Affiliations
  • Imperial College London (GB)
  • University of Manchester (GB)
  • University of Strathclyde (GB)
  • University of Leeds (GB)
  • National Physical Laboratory (GB)
Venue
IEEE Transactions on Terahertz Science and Technology
Published
2025-05
Volume
15
Issue
3
Pages
418-430
ISSN
  • 2156-342X (ISSN-L)
  • 2156-3446
OpenAlex
W4408281341
Cited by
4
References
27
Citations by year
2026: 4
FWCI
0.90
Type
journal article
Language
en
Locations
  • IEEE Transactions on Terahertz Science and Technology (publishedVersion)
  • White Rose Research Online (University of Leeds, The University of Sheffield, University of York) (acceptedVersion)
  • Strathprints: The University of Strathclyde institutional repository (University of Strathclyde) (acceptedVersion, cc-by)
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
powder compacts, granular materials, sample morphology, microsphere size, optical path length, ray tracing simulation, terahertz spectroscopy, morphological characterization, heterogeneous samples

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 2026-10-12, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-12 (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: Unveiling the Origin of Spurious Features in THz-TDS of Powder Compacts. https://app.cochat.ai/papers/10.1109/tthz.2025.3549944
The paper
Gorecki, J., Murphy, K. N., Markl, D., Burnett, A. D., & Naftaly, M. (2025). Unveiling the Origin of Spurious Features in THz-TDS of Powder Compacts. IEEE Transactions on Terahertz Science and Technology. https://doi.org/10.1109/tthz.2025.3549944

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.