Pith. sign in

REVIEW 3 major objections 4 minor 20 references

Dynamics of local photoconductivity in GaAs and InP investigated by THz SNOM

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read THz scanning near-field microscopy can tell whether photoexcited electrons recombine or are driven away from the surface.

desk verdict The InP band-bending claim is undercut by a diffusion modeling omission; the methodology is still worth engaging. read the letter →

arxiv 2501.11615 v1 pith:7YJEVPTL submitted 2025-01-20 cond-mat.mes-hall

Tinkara Troha , Arvind Singh , Petr Kužel , Hynek Němec This is my paper · ORCID
classification cond-mat.mes-hall
keywords THz-SNOMnear-fieldmicroscopyphotoconductivityGaAsInPbandbendingcarrierdiffusiontime-resolvedTHzspectroscopy
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 measures ultrafast THz conductivity spectra of GaAs and InP with a scattering-type THz scanning near-field microscope, following the local response for hundreds of picoseconds after a laser pulse. In GaAs the near-field spectra retrieve an electron density that decays exponentially with a time constant of about 46 ps and an electron scattering time that decreases with density. In InP the near-surface electron density decays within about 600 ps while the total electron population seen by far-field THz transmission stays essentially constant, which the authors take as evidence that electrons are diffusing away from the photoexcited region and being repelled from the surface by band bending. The paper also derives general strategies for analyzing THz-SNOM signals and maps the parameter regions where carrier density and photoexcited-layer thickness can be retrieved independently.

What carries the argument

The central object is the normalized scattered signal $S^{(m)} = s_{\rm exc}^{(m)}/s_{\rm gnd}^{(m)}$, modelled by the finite-dipole expression $s^{(m)} \propto \alpha^{(m)}(1+r)^2$, where $\alpha^{(m)}$ is the effective tip polarizability and $r$ the far-field p-polarized reflectance of the sample; the ratio removes the incident pulse and instrumental response. The depth-profiling mechanism is harmonic demodulation: signals demodulated at higher harmonics of the tip tapping frequency respond to thinner photoexcited layers, so measuring the second and third harmonics gives some depth resolution. For InP, the authors use a two-layer model (a thin top layer and a deep stationary layer) to fit the spectra and extract both densities and layer thicknesses, with the far-field reflectance interference term providing the sensitivity needed to separate the two parameters.

What would settle it

Measure the InP near-surface electron density with THz-SNOM after removing the surface band bending, for example by chemical passivation or a different surface termination: if the fast drop in top-layer density persists, the band-bending drift explanation is wrong, while if it disappears the explanation is supported.

Watch

Extended reading notes

Core claim

The central claim, stated in the abstract, is that in InP the THz photoconductivity decay is controlled mainly by electron diffusion away from the photoexcited area and by drift due to band bending at the surface, whereas in GaAs the decay is a simpler recombination-limited process. The evidence is a two-layer fit to the normalized near-field spectra: a thin top layer of photoexcited electrons whose density collapses to nearly zero within 10 ns, over a deeper, long-lived population at $3$ to $5\times10^{16}\,\mathrm{cm^{-3}}$ extending about $6\,\mu\mathrm{m}$ into the crystal. Because far-field transmission shows the total electron density barely decaying, surface recombination is ruled out, leaving outward drift as the explanation. Alongside this, the paper shows that the depth selectivity of higher tapping harmonics can be used to distinguish surface-near conductivity from bulk conductivity, and that the independent retrieval of density and thickness is possible mainly where the far-field reflectance interference term contributes significantly.

Load-bearing premise

The conclusions rest on the finite-dipole model being accurate enough that its fitted effective tip length is a harmless stand-in for the real tip; the same physical tip gives different fitted lengths for different harmonics and for the two semiconductors, so this assumption is not fully secure.

Editorial extensions

If this is right

  • THz-SNOM spectra can expose carrier transport along the surface normal that far-field THz spectroscopy averages over.
  • Higher tapping harmonics can act as a crude depth scan, since each harmonic is sensitive to a different thickness of the photoexcited layer.
  • In InP, band-bending drift can dominate the near-surface THz response within tens of picoseconds, so surface treatment must be considered when interpreting near-field photoconductivity.
  • For thick photoexcited layers, the interference of the THz beam inside the layer enables simultaneous retrieval of carrier density and layer thickness, while for thin layers only the sheet conductivity can be reliably obtained.
  • The harmonic-ratio signal $X^{(2,3)}$ cancels the reflectance and can be used on structured surfaces, but with about four times worse accuracy than the normalized single-harmonic signal.

Reading between the lines

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

  • If the band-bending interpretation is right, passivating or otherwise changing the InP surface should suppress the fast collapse of the top-layer electron density while leaving the far-field decay nearly unchanged; this is a directly testable prediction.
  • The fitted effective tip length differs between harmonics and between GaAs and InP, which suggests the finite-dipole model is absorbing some unmodeled near-field physics; a more physical tip model could alter the quantitative drift rates, though not necessarily the qualitative InP conclusion.
  • The harmonic-depth-selectivity idea could be applied to other layered or nanostructured semiconductors, such as heterostructures with buried two-dimensional electron gases, to extract depth-resolved carrier dynamics without sectioning the sample.
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.

Referee Report

3 major / 4 minor

Summary. The paper reports THz scattering-type SNOM measurements of photoexcited GaAs and InP wafers, combined with a finite-dipole model to extract complex photoconductivity spectra at ultrafast pump-probe delays. It develops three normalization strategies (S(m), X(m,m'), and a self-referenced harmonic ratio), analyzes their sensitivity and parameter-retrieval accuracy via error ellipses, and applies them to the two semiconductors. For GaAs, the retrieved electron density and scattering time are consistent between the 2nd and 3rd harmonics. For InP, the near-field signal S(2) decays faster than the far-field transmittance, which the authors interpret as electron diffusion away from the surface combined with drift due to surface band-bending, supported by a two-layer model with a long-lived carrier population.

Significance. If the InP interpretation is correct, the paper demonstrates that THz-SNOM spectra can access depth-dependent carrier transport that far-field THz spectroscopy averages over, and it provides a useful general framework for analyzing such measurements. The sensitivity analysis with error ellipses and the critical discussion of effective tip length are valuable methodological contributions. The GaAs results are robust and give confidence in the experimental and fitting procedures. However, the central InP claim is not yet quantitatively established, because the simplified two-layer model does not exclude ordinary ambipolar diffusion as the sole cause of the faster near-field decay, and the model parameters show inconsistencies across harmonics and samples.

major comments (3)
  1. [Sec. 5, Fig. 12] The two-layer model fixes the top-layer thickness at dexc = 350 nm for all delays shorter than 1 ns and only allows it to expand to 1.4 µm at 10 ns. With the stated ambipolar diffusion coefficient D = 10 cm^2/s [19], the diffusion length sqrt(2Dt) reaches about 1.1 µm at 600 ps, already three times the assumed top-layer thickness. A continuously diffusing shallow profile would reduce the near-surface density and make S(2) decay faster than the far-field integrated density even in the complete absence of band bending; the faster decay of S(2) compared with the far-field transmittance (Fig. 13) is therefore not by itself strong evidence of a force pushing electrons away from the surface. The authors should either include continuous ambipolar diffusion in the model or explicitly demonstrate that the early-time decay cannot be explained by diffusion alone.
  2. [Sec. 5, p. 21] The retrieved long-lived density differs by a factor of 1.6 between the two harmonics (n_long = 3.3×10^16 cm^-3 for m=2 and 5.2×10^16 cm^-3 for m=3), and the effective tip length L differs between harmonics (0.87 µm vs 1.9 µm in GaAs) and between samples (0.37 µm in InP). The paper acknowledges that the two-layer homogeneous-density model cannot reproduce the complex spatial profile after diffusion and drift (p. 21). Because the top-layer density decay is the direct evidence for the drift claim, the inconsistency of the model parameters across harmonics and samples leaves the quantitative support for band-bending-driven drift unconvincing; a more physical model with shared parameters or an explicit treatment of the spatial profile is needed.
  3. [Sec. 4.1 vs Sec. 5] For GaAs the authors verify that increasing the assumed photoexcited layer thickness from 750 nm to 1000 nm changes the fitted material parameters by less than 5%, which justifies neglecting diffusion there. No such verification is reported for InP, where the absorption depth is only ~350 nm and the ambipolar diffusion coefficient is 10 cm^2/s, so the diffusion length is large relative to the layer thickness. The absence of this control makes the InP interpretation fragile: the faster near-field decay could be an artifact of the fixed-thickness approximation rather than a signature of drift. A sensitivity test varying dexc with delay, or a direct diffusion-model fit, should be provided.
minor comments (4)
  1. [Throughout] The manuscript contains many diacritic and spelling artifacts (e.g., 'Cžech Academy of Sciences', 'Terahertž', 'polarižability'); these should be corrected for a polished final version.
  2. [Figure captions, e.g., Fig. 8] The caption of Fig. 8 refers to 'Eq. (5)' for the normalized scattered signal S(2), but the definition is given in Eq. (2); please check all cross-references.
  3. [Reference [13]] Reference [13] is cited as 'Optica 15, 8550 (2007)', but the correct source is Optics Express 15, 8550 (2007); please correct the journal name.
  4. [Sec. 6] In the final paragraph, the phrase 'was observed was observed' is duplicated; please remove the repetition.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: the central InP claim rests on an independent measured near-field vs far-field comparison; the only self-citation is interpretive and not load-bearing.

full rationale

The paper's derivation chain is not circular in the senses defined above. The quantitative analysis uses the finite-dipole model as an external, established framework [13,14], and the fitted quantities (carrier density, scattering time, effective tip length) are not renamed as predictions of the central claim. The key InP observation is a measured comparison: the decay of the near-field signal S(2) is faster than the decay of the measured far-field transmittance (Sec. 5, Fig. 13). Neither quantity is constructed from the other, and the faster near-field decay is not forced by the fitting model; it is an experimental input. The two-layer model is a fitting approximation, and the retrieved top-layer density is a fit output, not a predicted quantity. The interpretation of the faster decay as band-bending-driven drift is supported by external references [15,20] and by the separate observation that the far-field total electron density does not decay correspondingly, which rules out surface recombination. There is a minor self-citation: Ref. [15] includes two of the present authors and is cited when attributing band-bending transport and when noting prior observation of rapid electron escape. However, that citation is contextual and interpretive rather than the load-bearing derivation: the independent far-field comparison and the sensitivity analysis carry the argument, and the same claim is also attributed to an external reference [20]. The effective-tip-length inconsistency and the neglect of continuous diffusion for sub-nanosecond delays in InP are model-accuracy or correctness concerns, not circularity: they do not reduce the paper's conclusion to its own inputs by definition. Overall, no specific equation or fitted parameter is demonstrably equivalent to the claimed result, so the paper is not significantly circular.

Assumptions & free parameters 8 free parameters · 6 assumptions · 0 invented entities

The analysis rests on the finite-dipole forward model with two semi-empirical tip parameters (L fitted, g fixed), a Drude conductivity model, and, for InP, an ad hoc two-layer carrier profile. The paper introduces no new physical entities. The effective tip length L is the most consequential free parameter: its fitted value varies by a factor of 2 to 5 across harmonics and samples, meaning the absolute carrier densities and decay times carry model uncertainty. The InP two-layer approximation is explicitly acknowledged as imperfect.

free parameters (8)
  • Effective tip length L (GaAs, m=2) = 0.87 +/- 0.07 um
    Fitted independently for each harmonic and experiment; used in the finite-dipole model to convert spectra to carrier parameters. Its value differs between harmonics and samples, indicating model uncertainty.
  • Effective tip length L (GaAs, m=3) = 1.9 +/- 0.3 um
    Same as above; inconsistency with the m=2 value is acknowledged by the authors.
  • Effective tip length L (InP) = about 0.37 um
    Fitted for InP spectra; differs from the GaAs value, reinforcing that L is an unphysical calibration parameter.
  • Electron scattering time tau_s (GaAs) = about 20 to 100 fs depending on delay
    Drude-model parameter extracted from the spectral line shape; used to infer density dynamics.
  • Photoexcited electron density n_exc (GaAs) = initial about 5.5e16 cm^-3, decaying with about 46 ps time constant
    Central extracted quantity; comparison with fluence estimate gives a factor-of-five discrepancy.
  • Long-lived electron density n_long (InP) = 3.3e16 cm^-3 (m=2), 5.2e16 cm^-3 (m=3)
    Second-layer carrier density in the two-layer model; harmonic discrepancy is acknowledged.
  • Long-lived layer depth d_long (InP) = 6.0 um
    Depth to which long-lived electrons extend; fitted from the 10 ns spectrum.
  • Expanded top-layer thickness d_10ns (InP) = 1.4 um
    Thickness of the top photoexcited layer at the 10 ns delay, fitted from the negative-delay spectrum.
assumptions (6)
  • domain assumption The scattered THz signal is proportional to the effective tip polarizability times the far-field reflectance factor (1+r)^2 (Eq. 1).
    Basis of all quantitative extraction; inherited from the finite-dipole model of Cvitkovic et al. and Hauer et al., not independently validated here.
  • domain assumption The finite-dipole model with tip parameters (radius 40 nm, minimal distance 1 nm, tapping amplitude 150 nm, g=0.7e0.06i, effective length L) quantitatively describes the near-field interaction.
    Used throughout Sections 4 and 5 for fits; the effective length L must be fitted and varies between harmonics, so the model is only approximately valid.
  • domain assumption The photoexcited permittivity follows a Drude model with a single electron scattering time and effective mass (Eq. 5).
    Used to convert fitted permittivity changes into conductivity and carrier density; holes are neglected in GaAs because of much lower mobility.
  • ad hoc to paper In InP, the carrier depth profile can be represented by two homogeneous layers with abrupt boundaries and two densities n_exc and n_long.
    The model is introduced specifically for InP analysis; authors concede it cannot reproduce the real complex profile, and n_long differs between harmonics.
  • domain assumption Electron redistribution by ambipolar diffusion is negligible in GaAs on the measured time scale, so a single 750 nm layer suffices.
    The authors verify that assuming 1000 nm instead of 750 nm changes fitted parameters by less than 5 percent; acceptable but still an approximation.
  • domain assumption Normalizing the excited signal by the unexcited signal cancels the unknown instrument response and tip illumination factors (Eq. 2).
    Central to all three analysis strategies; any harmonic-dependent systematic error would propagate into the ratios.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamics of local photoconductivity in GaAs and InP investigated by THz SNOM." pith.science (2026). https://pith.science/paper/7YJEVPTL

@misc{pith2026250111615,
  author       = {Pith},
  title        = {Pith review of: Dynamics of local photoconductivity in GaAs and InP investigated by THz SNOM},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7YJEVPTL}},
  note         = {Machine review of arXiv:2501.11615}
}
read the original abstract

Terahertz scanning near-field optical microscope (THz-SNOM) is employed to measure ultrafast evolution of THz conductivity spectra after photoexcitation of GaAs and InP wafers using ultrashort laser pulses. Unlike in GaAs, the THz photoconductivity decay in InP is controlled mainly by the diffusion of electrons away from the photoexcited area, and also by the drift due to band-bending at the surface of the semiconductor. We propose and discuss several general strategies of the analysis of signals measured using THz-SNOM, and we estimate the accuracy of the obtained near-field photoconductivity spectra.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

  1. [19]

    Y. Song, L. Cao, B. D. Peng, G. Z. Song, Z. Q. Yue, J. M. Ma, L. Sheng, B. K. Li, and H. X. Wang, Investigation of an InP-based image converter with optical excitation, Rev. Sci. Instrum. 88, 033109 (2017)

    work page 2017

  2. [1]

    Ulbricht, E

    R. Ulbricht, E. Hendry, J. Shan, T . F . Heinž, and M. Bonn, Carrier dynamics in semiconductors studied with time-resolved terahertž spectroscopy, Rev. Mod. Phys. 83, 543 (2011)

    work page 2011

  3. [2]

    Lloyd-Hughes and T .-I

    J. Lloyd-Hughes and T .-I. Jeon, A review of the terahertž conductivity of bulk and nano- materials, Int. J. Infrared Millimeter Waves 33, 871 (2012)

    work page 2012

  4. [3]

    Kuž el and H

    P . Kuž el and H. Ne mec, Terahertž spectroscopy of nanomaterials: a close look at charge- carrier transport, Adv. Opt. Mater. 8, 1900623 (2020)

    work page 2020

  5. [4]

    C. J. Docherty and M. B. Johnston, Terahertž properties of graphene, J. Infrared. Milli. Terahž. Waves 33, 797 (2012)

    work page 2012

  6. [5]

    C. A. Schmuttenmaer , Exploring dynamics in the far -infrared with terahertž spectroscopy, Chem. Rev. 104, 1759 (2004)

    work page 2004

  7. [6]

    T . L. Cocker, V. Jelic, R. Hillenbrand, and F . A. Hegmann, Nanoscale terahertž scanning probe microscopy, Nature Photonics 15, 558 (2021)

    work page 2021

  8. [7]

    Jelic, S

    V. Jelic, S. Adams, M. Hassan, K. Cleland-Host, S. E. Ammerman, and T . L. Cocker, Atomic- scale terahertž time-domain spectroscopy, Nature Photonics 18, 898 (2024). 24

    work page 2024

Show all 20 references
  1. [8]

    Zenhausern, M

    F . Zenhausern, M. P . O’Boyle, and H. K. Wickramasinghe, Apertureless near-field optical microscope, Appl. Phys. Lett. 65, 1623 (1994)

    1994

  2. [9]

    Keilmann and R

    F . Keilmann and R. Hillenbrand, Near-field microscopy by elastic light scattering from a tip, Phil. Trans. R. Soc. Lond. A 362, 787 (2004)

    2004

  3. [10]

    A. J. Huber, F . Keilmann, J. Wittborn, J. Aižpurua, and R. Hillenbrand, Terahertž near-field nanoscopy of mobile carriers in single semiconductor nanodevices, Nano Lett. 8, 3766 (2008)

    2008

  4. [11]

    Eisele, T

    M. Eisele, T . L. Cocker, M. A. Huber, M. Plankl, L. Viti, D. Ercolani, L. Sorba, M. S. Vitiello, and R. Huber, Ultrafast multi -terahertž nano -spectroscopy with sub -cycle temporal resolution, Nature Photonics 8, 841 (2014)

    2014

  5. [12]

    Knoll and F

    B. Knoll and F . Keilmann, Enhanced dielectric contrast in scattering-type scanning near- field optical microscopy, Opt. Commun. 182, 321 (2000)

    2000

  6. [13]

    Cvitkovic, N

    A. Cvitkovic, N. Ocelic, and R. Hillenbrand; Analytical model for quantitative prediction of material contrasts in scattering-type near-field optical microscopy, Optica 15, 8550 (2007)

    2007

  7. [14]

    Hauer, A

    B. Hauer, A. P . Engelhardt, and T . Taubner, Quasi-analytical model for scattering infrared near-field microscopy on layered systems, Opt. Express 20, 13173 (2012)

    2012

  8. [15]

    Pushkarev, H

    V. Pushkarev, H. Ne mec, V. C. Paingad, J. Man a k, V. Jurka, V. Nova k, T . Ostatnicky , and P . Kuž el, Charge transport in single -crystalline GaAs nanobars: Impact of band bending revealed by terahertž spectroscopy, Adv. Funct. Mater. 32, 2107403 (2022)

    2022

  9. [16]

    D. E. Aspnes and A. A. Studna, Dielectric functions and optical parameters of Si, Ge, GaP , GaAs, GaSb, InP , InAs, and InSb from 1.5 to 6.0 eV, Phys. Rev. B 27, 985 (1983)

    1983

  10. [17]

    D. L. Rode and S. Knight, Electron Transport in GaAs, Phys. Rev. B 3, 2534 (1971)

    1971

  11. [18]

    B. A. Ružicka, L. K. Werake, H. Samassekou, and H. Zhao, Ambipolar diffusion of photoexcited carriers in bulk GaAs, Appl. Phys. Lett. 97, 262119 (2010)

    2010

  12. [20]

    Lloyd-Hughes, S

    J. Lloyd-Hughes, S. K. E. Merchant, L. Sirbu, I. M. Tiginyanu, and M. B. Johnston, Terahertž photoconductivity of mobile electrons in nanoporous InP honeycombs, Phys. Rev. B 78, 085320 (2008)

    2008

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.