pith. sign in

arxiv: 1906.09608 · v1 · pith:SVKOE76Jnew · submitted 2019-06-23 · ⚛️ physics.data-an · cond-mat.mtrl-sci

Second derivative analysis and alternative data filters for multi-dimensional spectroscopies: a Fourier-space perspective

Rongjie Li , Xiaoni Zhang , Lin Miao , Luca Stewart , Erica Kotta , Dong Qian , Konstantine Kaznatcheev , Jerzy T. Sadowski
show 7 more authors
Elio Vescovo Abdullah Alharbi Ting Wu Takashi Taniguchi Kenji Watanabe Davood Shahrjerdi L. Andrew Wray
This is my paper

Pith reviewed 2026-05-25 18:02 UTC · model grok-4.3

classification ⚛️ physics.data-an cond-mat.mtrl-sci
keywords second derivative imageFourier spaceARPESband pass filternoise reductionmulti-dimensional spectroscopydata filteringbackground suppression
0
0 comments X

The pith

Representing the second derivative image method as a multi-band pass filter in Fourier space shows that removing its higher harmonics reduces noise and background in ARPES data.

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

The paper analyzes the second derivative image (SDI) method used to sharpen dispersive features in multi-dimensional spectroscopies such as ARPES. It recasts the SDI function in Fourier space as a multi-band pass filter and examines how this filter interacts with noise and background signals. The central finding is that final image quality improves when the higher Fourier harmonics of the filter are eliminated. The work also outlines how similar band-pass approaches extend to higher-dimensional data and how prior knowledge of spectral features can yield stronger filters.

Core claim

The SDI procedure takes the form of a multi-band pass filter in Fourier space. Eliminating the higher Fourier harmonics of this filter suppresses undesirable noise and background features while preserving the sharpened dispersive signals, resulting in higher-quality processed images for ARPES and related spectroscopies.

What carries the argument

The Fourier-space representation of the SDI filter as a multi-band pass filter, from which higher harmonics can be removed to control noise.

If this is right

  • Image quality in ARPES data sets improves when higher SDI harmonics are dropped.
  • SDI-like band-pass filters extend directly to higher-dimensional data sets.
  • Filters become more effective when designed with a priori knowledge of the expected spectral features.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same Fourier-harmonic selection principle could be tested on other derivative-based sharpening routines used in imaging.
  • The approach suggests a general route for tailoring band-pass filters once the characteristic length scales of signal versus noise are known.
  • Direct application to simulated spectra with controlled noise levels would quantify how much information is retained after harmonic removal.

Load-bearing premise

Higher Fourier harmonics of the SDI filter contain mostly noise and background rather than useful spectral information.

What would settle it

A side-by-side comparison of SDI images processed with and without the higher harmonics, checking whether removal of those harmonics erases real spectral features instead of improving clarity.

Figures

Figures reproduced from arXiv: 1906.09608 by Abdullah Alharbi, Davood Shahrjerdi, Dong Qian, Elio Vescovo, Erica Kotta, Jerzy T. Sadowski, Kenji Watanabe, Konstantine Kaznatcheev, L. Andrew Wray, Lin Miao, Luca Stewart, Rongjie Li, Takashi Taniguchi, Ting Wu, Xiaoni Zhang.

Figure 1
Figure 1. Figure 1: (a) shows the Fourier decomposition of signal and background in a constant-energy slice (termed a mo￾mentum distribution curve, MDC) of ARPES data near the Fermi level of graphene, intersecting one of the Dirac cones. From this spectrum, we can see that the back￾ground dominates the low frequency and high frequency sectors of Fourier space, while the Dirac band (red curve) has greater intensity in an inter… view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (b). Overall, the 2D-filtered image in [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p006_5.png] view at source ↗
read the original abstract

The second derivative image (SDI) method is widely applied to sharpen dispersive data features in multi-dimensional spectroscopies such as angle resolved photoemission spectroscopy (ARPES). Here, the SDI function is represented in Fourier space, where it has the form of a multi-band pass filter. The interplay of the SDI procedure with undesirable noise and background features in ARPES data sets is reviewed, and it is shown that final image quality can be improved by eliminating higher Fourier harmonics of the SDI filter. We then discuss extensions of SDI-like band pass filters to higher dimensional data sets, and how one can create even more effective filters with some a priori knowledge of the spectral features.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript represents the second-derivative image (SDI) operator for sharpening dispersive features in ARPES and similar multi-dimensional spectroscopies as a multi-band-pass filter in Fourier space. It reviews the interaction of this filter with noise and background, claims that final image quality is improved by truncating higher Fourier harmonics of the SDI filter, and sketches extensions of SDI-like filters to higher-dimensional data sets that incorporate a priori spectral knowledge.

Significance. If the truncation step can be shown to preserve dispersive information while suppressing only noise/background, the Fourier perspective supplies a transparent, low-parameter route to improved data filtering that could be adopted across ARPES and related spectroscopies. The work also supplies a conceptual bridge between ad-hoc SDI usage and more general band-pass design.

major comments (2)
  1. [Abstract] Abstract: the central claim that 'final image quality can be improved by eliminating higher Fourier harmonics of the SDI filter' is asserted without any derivation of the multi-band-pass form, without explicit power spectra separating signal from background, and without quantitative before/after metrics on real or synthetic ARPES data sets. This assumption (higher harmonics contain predominantly undesirable features) is load-bearing for the improvement assertion.
  2. [Abstract] Abstract (paragraph on interplay with noise and background): the manuscript does not demonstrate that truncation preserves all relevant high-momentum or sharp dispersive features across typical ARPES data; a controlled test (e.g., synthetic spectra with known dispersion plus controlled noise) is required to quantify net gain versus loss of information.
minor comments (2)
  1. Notation for the Fourier representation of the SDI operator should be introduced explicitly (even if only in an appendix) so that the multi-band-pass property can be verified by the reader.
  2. The discussion of higher-dimensional extensions would benefit from at least one concrete filter kernel or pseudocode example.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive feedback on our manuscript. The comments highlight areas where additional evidence would strengthen the central claims regarding the benefits of truncating higher Fourier harmonics. We address each point below and commit to revisions that provide the requested quantitative support without altering the core Fourier-space analysis.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that 'final image quality can be improved by eliminating higher Fourier harmonics of the SDI filter' is asserted without any derivation of the multi-band-pass form, without explicit power spectra separating signal from background, and without quantitative before/after metrics on real or synthetic ARPES data sets. This assumption (higher harmonics contain predominantly undesirable features) is load-bearing for the improvement assertion.

    Authors: The derivation of the multi-band-pass form is obtained directly by Fourier transforming the second-derivative operator (see main text, Section II). We agree that the manuscript currently supports the improvement claim primarily through qualitative ARPES examples rather than explicit power spectra or quantitative metrics. In revision we will add (i) power spectra of representative data sets with signal/background separation and (ii) quantitative before/after metrics (e.g., peak sharpness and background suppression ratios) on both real and synthetic data. revision: yes

  2. Referee: [Abstract] Abstract (paragraph on interplay with noise and background): the manuscript does not demonstrate that truncation preserves all relevant high-momentum or sharp dispersive features across typical ARPES data; a controlled test (e.g., synthetic spectra with known dispersion plus controlled noise) is required to quantify net gain versus loss of information.

    Authors: The current text reviews the general interplay of SDI with noise and background but does not include a controlled synthetic test. We accept that such a test is necessary to quantify preservation of high-momentum dispersive features. In the revised manuscript we will add a synthetic-data section that injects known dispersions plus controlled noise, applies both full and truncated SDI filters, and reports quantitative measures of feature retention versus noise reduction. revision: yes

Circularity Check

0 steps flagged

No circularity: Fourier representation and filter modification are direct and non-referential

full rationale

The paper's core step is representing the second derivative image (SDI) operator in Fourier space as a multi-band-pass filter, which follows directly from the mathematical definition of the second derivative without reference to data values or fitted quantities. The subsequent claim that image quality improves by truncating higher harmonics is presented as a heuristic recommendation based on the filter's frequency response and its interaction with noise/background, not as a statistical prediction derived from the same dataset. No equations reduce a result to its own inputs by construction, no parameters are fitted and then relabeled as predictions, and no self-citations serve as load-bearing uniqueness theorems. The derivation chain is therefore self-contained and independent of the target ARPES datasets.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that ARPES spectral features are sufficiently band-limited that higher harmonics of the SDI filter can be removed without discarding signal; no free parameters or invented entities are introduced in the abstract.

axioms (1)
  • domain assumption ARPES dispersive features and background occupy distinct Fourier-frequency bands such that higher harmonics of the SDI filter can be dropped without loss of signal
    Invoked when the abstract states that eliminating higher harmonics improves image quality by reducing noise and background.

pith-pipeline@v0.9.0 · 5708 in / 1272 out tokens · 29289 ms · 2026-05-25T18:02:21.192306+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

34 extracted references · 34 canonical work pages

  1. [1]

    We recommend achieving this by setting T (ξ) = 0 for hξ > 2π

    Cutting off high frequency data elements. We recommend achieving this by setting T (ξ) = 0 for hξ > 2π. Alternatively, it is common to apply a secondary low pass filter (smoothing) to the data, however this has the potential drawback of adding a layer of complexity to the Fourier-space picture

    work page

  2. [2]

    The T (ξ) = 1 −cos(hξ) function has the disadvantage that the peak width and maximum are defined by the same variableh

    Choosing a new underlying function. The T (ξ) = 1 −cos(hξ) function has the disadvantage that the peak width and maximum are defined by the same variableh. We recommend defining a Gaussian-like filter instead, so that the peak position and width are associ- ated with independent variables. In this case, one must symmetrize the function about zero, such as by...

    work page

  3. [3]

    In this case the background is weak, and the cross-like Fourier space feature largely vanishes (see Fig. 4(c)). The high intensity border artifacts seen in the graphene data set (Fig. 3(b)) derive from the interplay of the filter with this cross-like Fourier background feature, and are therefore not present along the SDI panel boundary in Fig. 4(b). Overal...

    work page

  4. [4]

    Lu et al

    D.-H. Lu et al. Annu. Rev. Condens. Matter Phys., 3 (2012), 129-167

    work page 2012

  5. [5]

    Eich et al

    S. Eich et al. J. El. Spect. Rel. Phenom., 195 (2014), 231-236

    work page 2014

  6. [6]

    Katoch et al

    J. Katoch et al. Nat. Phys., 14 (2018), 355-359

    work page 2018

  7. [7]

    L. J. P Ament, M. van Veenendaal, T. P. Devereaux, J. P. Hill, J. van den Brink Rev. Mod. Phys., 83 (2011), 705

    work page 2011

  8. [8]

    L. A. Wray et al. Frontiers in Physics, 3 (2015), 32

    work page 2015

  9. [9]

    Rotenberg, A

    E. Rotenberg, A. Bostwick. J. Synch. Rad., 21 (2014), 5

    work page 2014

  10. [10]

    He et al

    Y. He et al. Rev. Sci. Instrum., 87 (2016), 011301

    work page 2016

  11. [11]

    N Gedik, I. Vishik. Nat. Phys., 13 (2017), 1029-1033

    work page 2017

  12. [12]

    Reber, Y

    M.A.R. Reber, Y. Chen, T.K. Allison. Optica, 3 (2016), 311

    work page 2016

  13. [13]

    Popmintchev et al

    D. Popmintchev et al. Science, 350 (2015), 1225

    work page 2015

  14. [14]

    Eriksson et al

    M. Eriksson et al. Nucl. Instr. And Meth., A 587 (2008), 221-226

    work page 2008

  15. [15]

    S. Y. Lee Phys. Rev. E, 54 (1996), 1940-1944

    work page 1996

  16. [16]

    Einfield, M

    D. Einfield, M. Plesko, J. Schaper J. Synchrotron Radiat., 21 (2014), 856-861

    work page 2014

  17. [17]

    K. M. Shen et al. Phys. Rev. Lett., 93 (2004), 267002

    work page 2004

  18. [18]

    Zhang et al

    P. Zhang et al. Rev. Sci. Inst., 82 (2011), 043712

    work page 2011

  19. [19]

    Shirley, Phys

    D.A. Shirley, Phys. Rev., 55 (1972), 4709

    work page 1972

  20. [20]

    D. W. Lynch and C. G. Olson Photoemission studies of high-temperature superconduc- tors. Vol. 5. Cambridge University Press, 2005

    work page 2005

  21. [21]

    Kaminski et al

    A. Kaminski et al. Phys. Rev. B, 69 (2004), 212509

    work page 2004

  22. [22]

    M. P. Seah, W.A. Dench Surf. Interface. Anal., 1 (1979), 2

    work page 1979

  23. [23]

    Wu et al

    T. Wu et al. Appl. Phys. Lett., 113 (2018), 193502

    work page 2018

  24. [24]

    Ronning et al

    F. Ronning et al. Phys. Rev. B, 71 (2005), 094518

    work page 2005

  25. [25]

    Graf et al

    J. Graf et al. Phys. Rev. Lett., 98 (2007), 067004

    work page 2007

  26. [26]

    B. P. Xie et al. Phys. Rev. Lett., 98 (2007), 147001

    work page 2007

  27. [27]

    Valla et al

    T. Valla et al. Phys. Rev. Lett., 98 (2007), 167003

    work page 2007

  28. [28]

    Meevasana et al

    W. Meevasana et al. Phys. Rev. B, 75 (2007), 174506

    work page 2007

  29. [29]

    Chang et al

    J. Chang et al. Phys. Rev. B, 75 (2007), 224508

    work page 2007

  30. [30]

    D. S. Inosov et al. Phys. Rev. Lett., 99 (2007), 237002

    work page 2007

  31. [31]

    Zhang et al

    W. Zhang et al. Phys. Rev. Lett., 101 (2008), 017002

    work page 2008

  32. [32]

    Moritz et al

    B. Moritz et al. New J. Phys., 11 (2009), 093020

    work page 2009

  33. [33]

    Ikeda et al

    M. Ikeda et al. Phys. Rev. B, 80 (2009), 184506

    work page 2009

  34. [34]

    Liu et al

    Y. Liu et al. Sci. Rep., 5 (2015), 13036

    work page 2015