Pith. sign in

REVIEW 2 major objections 4 minor 38 references

Treating the noisy backtracking step of the Y-procedure as a regularized discrete inverse problem recovers reaction-rate versus concentration curves more accurately than Fourier filtering, especially for nonlinear kinetics.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.5

2026-07-11 13:40 UTC pith:AKIL5IFT

load-bearing objection Clean inverse-problem fix for the Y-procedure that beats Fourier filtration on synthetic linear and nonlinear TAP data, with an objective cutoff rule for state-defining pulses. the 2 major comments →

arxiv 2607.04776 v1 pith:AKIL5IFT submitted 2026-07-06 physics.chem-ph physics.data-anphysics.flu-dyn

Resolving the inverse problem in pulse response analysis of TAP reactors

Anjali Aleria , Evgeniy Redekop , A. K. Suresh , Jason R. Picardo This is my paper
classification physics.chem-ph physics.data-anphysics.flu-dyn
keywords TAP reactorY-procedureinverse problemtruncated SVDtransient kineticsthin-zone reactorregularizationpulse response
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The Y-procedure reconstructs the rate-concentration relationship of a catalyst from TAP-reactor outlet-flux measurements without assuming a kinetic model, but the required reverse diffusion step amplifies noise and has previously been handled by subjective Fourier smoothing. This paper reformulates that step as a discrete inverse problem using a basis of localized square pulses, solves it by truncated singular-value decomposition, and chooses the truncation automatically by minimizing a multivaluedness measure for state-defining pulses. On synthetic data for linear, quadratic and cubic irreversible reactions the new reconstructions stay closer to the true R(C) curves than optimally filtered Fourier results. The approach therefore removes human tuning, supports automation of pulse-response analysis, and opens the door to other standard inverse-problem tools.

Core claim

Formulating the ill-posed backtracking of diffusion in the outlet zone of a thin-zone TAP reactor as the discrete linear system Aw = b, with w the coefficients of a square-pulse basis for the inlet flux, and regularizing by truncated SVD whose cutoff is selected by minimizing the multivaluedness measure Ψ of the reconstructed R(C), yields more accurate rate-concentration curves than Fourier filtration (with λ likewise chosen by minimizing Ψ), especially when the kinetics are nonlinear.

What carries the argument

The discrete inverse problem Aw = b built from Galerkin projection onto square-pulse basis functions, solved by truncated singular-value decomposition whose cutoff mode is chosen either from the Picard plot or by minimizing the multivaluedness integral Ψ of the reconstructed R(C) for state-defining experiments.

Load-bearing premise

The synthetic Gaussian noise model and the idealized thin-zone Knudsen diffusion equations are representative enough that the superiority of TSVD and the reliability of the Ψ-minimum cutoff will hold for real mass-spectrometer TAP pulses.

What would settle it

Apply both the TSVD+Ψ and Fourier+Ψ pipelines to the same set of experimental TAP outlet pulses whose true R(C) is independently known (or tightly constrained by independent kinetic measurements) and check whether the TSVD reconstructions remain systematically closer to the reference curve, especially for nonlinear cases.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Pulse-response analysis of state-defining TAP experiments can be fully automated because the regularization parameter is chosen by a computable scalar minimum rather than by visual inspection.
  • Other discrete inverse-problem methods (Tikhonov regularization, machine-learning inverses) become directly applicable once the square-pulse matrix A is formed.
  • Nonlinear rate functions, previously harder to recover because of asymmetric flux profiles, can be reconstructed with less oversmoothing and fewer spurious oscillations.
  • The same multivaluedness criterion can be used to tune any future regularizer for state-defining experiments.

Where Pith is reading between the lines

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

  • If the square-pulse basis is replaced by a wavelet or learned dictionary adapted to typical TAP pulse shapes, the condition number of A may drop and fewer modes may be needed for the same accuracy.
  • Extending the Ψ idea to a measure of petal-area or loop-consistency could allow objective regularization for state-altering pulses where R(C) is genuinely multi-valued.
  • Because the method works on synthetic data with both broad and narrow inlet pulses, it is likely to remain useful when experimental pulse widths vary across instruments.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 4 minor

Summary. The paper reformulates the noise-sensitive first step of the Y-procedure for thin-zone TAP reactors as a discrete linear inverse problem Aw = b. The unknown inlet flux to zone 3 is expanded in a basis of localized square pulses; the resulting matrix A is regularized by truncated SVD. Cutoff selection is guided by Picard plots and, for state-defining experiments, by minimizing a multivaluedness measure Ψ of the reconstructed R(C) curve. On synthetic outlet fluxes generated from linear, quadratic and cubic irreversible kinetics (with additive heteroscedastic Gaussian noise), the TSVD+Ψ reconstructions of R versus C are shown to be more accurate than those obtained by Fourier filtration with λ likewise chosen by minimizing Ψ, especially for the nonlinear cases. The formulation is presented as enabling automation and the future use of other discrete inverse-problem techniques.

Significance. If the reported superiority of TSVD+Ψ holds under the stated synthetic conditions, the work supplies a practical, objective route to regularizing the classic Y-procedure and thereby removes a long-standing barrier to automated, model-free extraction of R(C) curves from TAP pulses. The square-pulse Galerkin construction, the explicit use of Picard plots, and the definition of Ψ are clean contributions that open the door to Tikhonov, total-variation or learned regularizers. The comparisons for nonlinear kinetics (Figs. 8–9) are particularly valuable because they expose the limitations of generic Fourier bases on asymmetric flux profiles. The absence of experimental validation is a clear limitation of scope, but the mathematics and the synthetic evidence are self-contained and reproducible.

major comments (2)
  1. [§2.2, Eq. (7); §§6–7] The central claim of superior R(C) reconstructions (Abstract; §§6–7; Figs. 7–9) is demonstrated exclusively on synthetic data generated from the idealized thin-zone Knudsen model (Eqs. 1–5) and the specific noise model of Eq. (7). While the paper never asserts experimental transfer, the abstract and concluding remarks frame the method as facilitating automation of pulse-response analyses in general. A short sensitivity study (varying the coefficients in Eq. (7) or replacing Gaussian noise by a realistic mass-spectrometer noise model) or a single experimental TAP pulse would make the practical advantage of TSVD+Ψ over Fourier filtration far more convincing; without it the transferability of both the ranking and the reliability of the Ψ-minimum cutoff remains an untested assumption.
  2. [§6; §8] The objective cutoff strategy based on Ψ (Eq. 25) is derived and validated only for state-defining experiments in which R is known a priori to be single-valued. The paper correctly notes that the strategy does not apply to state-altering pulses (petal plots). Because many practical TAP campaigns deliberately use larger pulses, the manuscript should either (i) supply a provisional selection rule for multivalued R(C) (e.g., based on Picard transition alone) or (ii) state more explicitly that the fully automated pipeline is presently restricted to the state-defining regime.
minor comments (4)
  1. [§3–4] Typographical error: “the procure is outlined” should read “the procedure is outlined” (end of §3 / start of §4).
  2. [Figs. 8–9] Figure captions for the nonlinear cases (Figs. 8 and 9) omit the exact values of the rate constants k2 and k3; stating them (or the normalization kn Q^{n-1}=k) would aid reproducibility.
  3. [Appendix] The appendix demonstrates a narrower inlet pulse but does not quantify how the location of the Ψ minimum or the reconstruction error changes with pulse width; a one-sentence summary of that dependence would be useful.
  4. [throughout] Notation for the measured outlet flux switches between F*_out and F^*_out; a single consistent superscript would improve readability.

Circularity Check

0 steps flagged

No significant circularity: reconstructions are scored against independently generated forward-simulation ground truth and a definitional single-valuedness property of state-defining experiments.

full rationale

The paper formulates the first step of the Y-procedure as a discrete inverse problem Aw = b by expanding the unknown zone-3 inlet flux in a square-pulse basis (Eqs. 10–11), constructing the Galerkin matrix A from the known linear diffusion operator in zone 3 (Eqs. 12–19), and regularizing via TSVD truncation (Eq. 24). The cutoff m_cut is chosen either by inspection of the Picard plot of the SVD components of the noisy measurement b or by minimizing the multivaluedness measure Ψ (Eq. 25). Ψ is not a fit to any particular kinetic form; it simply quantifies departure from the single-valued R(C) property that, by definition of a state-defining experiment, must hold when surface coverage is constant. Ground-truth R(C) curves are generated independently by solving the forward thin-zone model (Eqs. 1–6) with prescribed linear, quadratic or cubic kinetics and then adding the synthetic noise of Eq. 7; reconstructions are compared to that independent truth (Figs. 6–9). Citations to the original Y-procedure papers supply background and the transfer functions used for the Fourier baseline; they are not load-bearing uniqueness theorems that force the present TSVD or Ψ constructions. Consequently the claimed superiority of TSVD+Ψ over Fourier filtration on the synthetic data is an empirical numerical result, not a tautology obtained by construction or by self-citation.

Axiom & Free-Parameter Ledger

4 free parameters · 4 axioms · 1 invented entities

The central claim rests on the standard thin-zone TAP diffusion model, the definition of state-defining single-valued R(C), a conventional additive noise model, and the new but simple multivaluedness functional Ψ used only for cutoff selection. No new physical entities are postulated; free parameters are the usual regularization cutoffs and discretization size.

free parameters (4)
  • m_cut (TSVD truncation index)
    Regularization parameter selected by minimizing Ψ (or guided by Picard plot); optimal values reported as 28 (linear), 29 (quadratic), 30 (cubic), 54 (narrow pulse).
  • λ (Fourier filtration parameter)
    Baseline method’s smoothing parameter, likewise chosen by minimizing Ψ for fair comparison (e.g. λ=13, 11, 14).
  • M (number of square-pulse basis functions)
    Discretization size fixed at M=500 after convergence checks; affects resolution of retained SVD modes.
  • Noise amplitude coefficients 0.004 and 0.05 in Eq. (7)
    Hand-chosen synthetic noise model following Yablonsky et al.; not fitted to real QMS data in this work.
axioms (4)
  • domain assumption Gas transport in inert zones is pure Knudsen diffusion (Eq. 1) with vacuum Dirichlet exit and continuous concentration / discontinuous flux at the thin catalyst zone (Eq. 2).
    Standard TZTR model taken from Yablonsky et al. 2007; load-bearing for both forward synthetic data and inverse reconstruction.
  • domain assumption In state-defining experiments R is a single-valued function of gas-phase C, so the reconstructed R–C curve should collapse to a single branch.
    Used to define and minimize Ψ for objective cutoff selection (§6); cited from prior TAP literature [21].
  • ad hoc to paper Measurement noise is adequately represented by the additive heteroscedastic Gaussian model of Eq. (7).
    Adopted from earlier synthetic studies; never replaced by experimental noise statistics.
  • standard math Linear superposition and Galerkin projection onto square pulses yield a well-defined discrete inverse problem Aw=b whose SVD modes separate signal from noise.
    Standard discrete inverse-problem theory (Hansen); applied in §5.
invented entities (1)
  • Multivaluedness measure Ψ = ∫[R+(c)−R−(c)]² dc no independent evidence
    purpose: Scalar score of departure from single-valued R(C) used to select m_cut or λ automatically for state-defining pulses.
    Defined in Eq. (25); not previously standard in Y-procedure literature. Independent evidence is only the synthetic minima that coincide with good reconstructions; no external experimental validation.

pith-pipeline@v1.1.0-grok45 · 23461 in / 3181 out tokens · 28565 ms · 2026-07-11T13:40:24.419823+00:00 · methodology

0 comments
read the original abstract

Pulse experiments in the temporal analysis of products (TAP) reactor are one of the most important methods for studying transient kinetics of gas-solid catalytic reactions. The Y-procedure (Yablonsky et al., Chem. Eng. Sci. 62, 6754, 2007) is a model-free analysis framework for inferring the relationship between the reaction-rate $R$ and the reactant concentration $C$ from measurements of the outlet flux of gas. While elegant in conception, its application is hindered by the amplification of measurement noise that results from having to backtrack diffusive transport from the outlet to the reaction zone. Here, we explicitly recognize the inverse problem inherent in the Y-procedure and treat it using well-developed tools from the field of inverse problems. While previous implementations of the Y-procedure used Fourier-based filtering, we do not pre-process the measurements with an ad hoc noise-filter. Instead, we use a basis of localized square pulses to formulate a discrete inverse problem, whose regularized solution is obtained via the truncated singular value decomposition (TSVD) method. This method requires one to select a cutoff mode number; while we show how the choice of this regularization parameter can be guided by a Picard plot, we also develop an objective selection strategy for state defining experiments, for which $R(C)$ is a single-valued function. We apply our proposed inverse-problem approach to synthetic data corresponding to linear and nonlinear reactions and compare the results with the Fourier-filtration method. The former produces better reconstructions of the $R$ vs $C$ relationship, especially for nonlinear reactions. Our work facilitates the automation of pulse response analyses and enables the application of other discrete inverse-problem techniques, such as Tikhonov regularization or machine-learning methods.

Figures

Figures reproduced from arXiv: 2607.04776 by A. K. Suresh, Anjali Aleria, Evgeniy Redekop, Jason R. Picardo.

Figure 1
Figure 1. Figure 1: Schematic of a thin-zone TAP reactor. A pulse of gas molecules di [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Outlet flux profile emerging at the exit of zone 3 after undergoing [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Y-procedure reconstruction using Fourier-filtration. The top row presents the inlet flux entering zone 3 (with the concentration in the thin reaction zone [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Illustration of (a) the first pulse function [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Picard plot showing the singular values σi along with the magnitudes of the corresponding SVD components of the measured vector b, |u T i b|, and the solution vector w, |u T i b|/σi (see the legend in panel (b)). The plot in (a) is restricted to the first 250 modes out of the total M = 500 modes, while the plot in (b) zooms into the transition region where noise first begins to dominate. (a) (b) (c) (d) (e… view at source ↗
Figure 6
Figure 6. Figure 6: Y-procedure reconstruction using the inverse problem formulation and the TSVD regularized solution. The top row presents the inlet flux entering zone 3 [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Selection of the regularization parameter for the TSVD method (top row) and the Fourier-filtration method (bottom row). (a) Variation of the degree [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Y-procedure reconstruction for a quadratic reaction using the TSVD method and Fourier-filtration, with the corresponding optimal regularization parameter [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Y-procedure reconstruction for a cubic reaction using the TSVD method and Fourier-filtration, with the corresponding optimal regularization parameter [PITH_FULL_IMAGE:figures/full_fig_p011_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Outlet flux at the exit of zone 3 corresponding to a narrower inlet [PITH_FULL_IMAGE:figures/full_fig_p013_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Analogue of Fig. 7 for a narrower pulse (µ [PITH_FULL_IMAGE:figures/full_fig_p014_11.png] view at source ↗

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

38 extracted references · 9 canonical work pages

  1. [1]

    García-Serna, R

    J. García-Serna, R. Piñero-Hernanz, D. Durán-Martín, In- spirational perspectives and principles on the use of cat- alysts to create sustainability, Catal. Today 387 (2022) 237–243. 100 years of CASALE SA: a scientific perspec- tive on catalytic processes

    2022

  2. [2]

    T. S. Kim, C. R. O’Connor, C. Reece, The new breed of cutting-edge catalysts, Nature 537 (2016) 156–158. doi:10.1038/537156a

    work page doi:10.1038/537156a 2016

  3. [3]

    C. M. Friend, B. Xu, Heterogeneous catalysis: A cen- tral science for a sustainable future, Acc. Chem. Res. 50 (2017) 517–521. doi:10.1021/acs.accounts.6b00510

    work page doi:10.1021/acs.accounts.6b00510 2017

  4. [4]

    Morgan, N

    K. Morgan, N. Maguire, R. Fushimi, J. T. Gleaves, A. Goguet, M. P. Harold, E. V . Kondratenko, U. Menon, Y . Schuurman, G. S. Yablonsky, Forty years of temporal analysis of products, Catal. Sci. Technol. 7 (2017) 2416– 2439

    2017

  5. [5]

    E. A. Redekop, G. S. Yablonsky, J. T. Gleaves, Truth is, we all are transients: A perspective on the time-dependent nature of reactions and those who study them, Catal. Today 417 (2023) 113761. doi:https://doi.org/10.1016/j.cattod.2022.05.026, tran- sient Kinetics Seminar

    work page doi:10.1016/j.cattod.2022.05.026 2023

  6. [6]

    J. T. Gleaves, G. Yablonsky, X. Zheng, R. Fushimi, P. L. Mills, Temporal analysis of products (tap)—recent advances in technology for kinetic analysis of multi- component catalysts, J. Mol. Catal. A: Chem. 315 (2010) 108–134. In memory of M.I. Temkin

    2010

  7. [7]

    T. S. Kim, C. R. O’Connor, C. Reece, Interrogating site dependent kinetics over sio2-supported pt nanoparticles, Nat. Commun. 15 (2024) 2074. doi:10.1038/s41467-024- 46496-1

    doi:10.1038/s41467-024- 2024

  8. [8]

    Yonge, G

    A. Yonge, G. S. Gusmão, R. Fushimi, A. J. Medford, Model-based design of experiments for temporal analy- sis of products (tap): A simulated case study in oxida- tive propane dehydrogenation, Ind. Eng. Chem. Res. 63 (2024) 4756–4770. doi:10.1021/acs.iecr.3c03418

    work page doi:10.1021/acs.iecr.3c03418 2024

  9. [9]

    J. T. Gleaves, J. R. Ebner, T. C. Kuechler, Temporal analy- sis of products (tap)—a unique catalyst evaluation system with submillisecond time resolution, Cat. Rev. 30 (1988) 49–116

    1988

  10. [10]

    Brandão, E

    L. Brandão, E. A. High, T.-S. Kim, C. Reece, Simplifying the temporal analysis of products reactor, Chem. Eng. J. 478 (2023) 147489

    2023

  11. [11]

    Constales, G

    D. Constales, G. Yablonsky, G. Marin, J. Gleaves, Multi- zone tap-reactors theory and application: I. the global transfer matrix equation, Chem. Eng. Sci. 56 (2001) 133– 149

    2001

  12. [12]

    Constales, S

    D. Constales, S. Shekhtman, G. Yablonsky, G. Marin, J. Gleaves, Multi-zone tap-reactors theory and applica- tion iv. ideal and non-ideal boundary conditions, Chem. Eng. Sci. 61 (2006) 1878–1891

    2006

  13. [13]

    G. Y . G.B. Marin, D. Constales, Temporal Analysis of Products: Principles, Applications, and Theory, 2019, pp. 307–381

    2019

  14. [14]

    G. S. Yablonskii, S. O. Shekhtman, S. Chen, J. T. Gleaves, Moment-based analysis of transient response catalytic studies (tap experiment), Ind. Eng. Chem. Res. 37 (1998) 2193–2202. 13 (a) TSVD (b) TSVD (mcut =54) (c) Fourier-filtration (d) (e) Fourier-filtration (λ=34) (f) Figure 11: Analogue of Fig. 7 for a narrower pulse (µ=1/12.5 s andσ=1/40 s). (a) Va...

    1998

  15. [15]

    state defining

    J. G. S.O. Shekhtman, G.S. Yablonsky, R. Fushimi, "state defining" experiment in chemical kinetics—primary char- acterization of catalyst activity in a tap experiment, Chem. Eng. Sci. 58 (2003) 4843–4859

    2003

  16. [16]

    E. V . Kondratenko, J. Pérez-Ramírez, Mechanism and kinetics of direct n2o decomposition over fe-mfi zeolites with different iron speciation from temporal analysis of products, J. Phys. Chem. B 110 (2006) 22586–22595

    2006

  17. [17]

    Prasad, A

    V . Prasad, A. M. Karim, A. Arya, D. G. Vlachos, As- sessment of overall rate expressions and multiscale, mi- crokinetic model uniqueness via experimental data injec- tion: Ammonia decomposition on Ru/γ-Al 2O3 for hydro- gen production, Ind. Eng. Chem. Res. 48 (2009) 5255–

    2009

  18. [18]

    doi:10.1021/ie900144x

    work page doi:10.1021/ie900144x

  19. [19]

    Kumar, X

    A. Kumar, X. Zheng, M. P. Harold, V . Balakota- iah, Microkinetic modeling of the no+h2 sys- tem on pt/al2o3 catalyst using temporal analy- sis of products, J. Catal. 279 (2011) 12–26. doi:https://doi.org/10.1016/j.jcat.2010.12.006

    work page doi:10.1016/j.jcat.2010.12.006 2011

  20. [20]

    Yablonsky, D

    G. Yablonsky, D. Constales, S. Shekhtman, J. Gleaves, The y-procedure: How to extract the chemical transforma- tion rate from reaction–diffusion data with no assumption on the kinetic model, Chem. Eng. Sci. 62 (2007) 6754– 6767

    2007

  21. [21]

    Shekhtman, G

    S. Shekhtman, G. Yablonsky, S. Chen, J. Gleaves, Thin- zone tap-reactor – theory and application, Chem. Eng. Sci. 54 (1999) 4371–4378

    1999

  22. [22]

    E. A. Redekop, G. S. Yablonsky, D. Constales, P. A. Ra- machandran, C. Pherigo, J. T. Gleaves, The y-procedure methodology for the interpretation of transient kinetic data: Analysis of irreversible adsorption, Chem. Eng. Sci. 66 (2011) 6441–6452

    2011

  23. [23]

    Ross Kunz, T

    M. Ross Kunz, T. Borders, E. Redekop, G. S. Yablon- sky, D. Constales, L. Wang, R. Fushimi, Pulse response analysis using the y-procedure: A data science approach, Chem. Eng. Sci. 192 (2018) 46–60

    2018

  24. [24]

    Hadamard, Sur les problèmes aux dérivées partielles et leur signification physique, Princeton University Bulletin, Princeton, NJ, (1902) 49–52

    J. Hadamard, Sur les problèmes aux dérivées partielles et leur signification physique, Princeton University Bulletin, Princeton, NJ, (1902) 49–52

    1902

  25. [25]

    J. L. Mueller, S. Siltanen, Linear and Nonlinear Inverse Problems with Practical Applications, Society for Indus- trial and Applied Mathematics, Philadelphia, PA, 2012. doi:10.1137/1.9781611972344

    work page doi:10.1137/1.9781611972344 2012

  26. [26]

    M. R. Kunz, R. Batchu, Y . Wang, Z. Fang, G. Yablon- sky, D. Constales, J. Pittman, R. Fushimi, Probability the- ory for inverse diffusion: Extracting the transport/kinetic time-dependence from transient experiments, Chem. Eng. J. 402 (2020) 125985

    2020

  27. [27]

    P. C. Hansen, Discrete Inverse Problems: Insight and Al- gorithms, Society for Industrial and Applied Mathemat- ics, USA, 2010. 14

    2010

  28. [28]

    K. L. Gering, C. Baroi, R. R. Fushimi, Transport modeling and mapping of pulsed reactor dynamics near and beyond the onset of viscid flow, Chem. Eng. Sci. 192 (2018) 576–

    2018

  29. [29]

    doi:https://doi.org/10.1016/j.ces.2018.07.060

    work page doi:10.1016/j.ces.2018.07.060 2018

  30. [30]

    Smith, Numerical Solution of Partial Differential Equa- tions (Finite Difference Methods), 1985, p

    G. Smith, Numerical Solution of Partial Differential Equa- tions (Finite Difference Methods), 1985, p. 3rd edition

    1985

  31. [31]

    S. C. van der Linde, T. Nijhuis, F. Dekker, F. Kapteijn, J. A. Moulijn, Mathematical treatment of transient kinetic data: Combination of parameter estimation with solving the related partial differential equations, Appl. Catal., A: Gen. 151 (1997) 27–57. Transient Kinetics

    1997

  32. [32]

    W. E. Schiesser, The numerical method of lines: integra- tion of partial differential equations, Elsevier, 2012

    2012

  33. [33]

    Virtanen, R

    P. Virtanen, R. Gommers, T. E. Oliphant, M. Haber- land, T. Reddy, D. Cournapeau, E. Burovski, P. Peterson, W. Weckesser, J. Bright, et al., Scipy 1.0: fundamental algorithms for scientific computing in python, Nat. Meth- ods 17 (2020) 261–272

    2020

  34. [34]

    A. J. Roberts, Linear Algebra for the 21st Century, Oxford University Press, Oxford, United Kingdom, 2021

    2021

  35. [35]

    Hintermüller, A

    M. Hintermüller, A. Hauptmann, B. Jin, C. Schönlieb (Eds.), Machine Learning Solutions for Inverse Problems: Part A, volume 26 ofHandbook of Numerical Analysis, Elsevier, 2025

    2025

  36. [36]

    Kirisits, B

    C. Kirisits, B. Mejri, S. Pereverzev, O. Scherzer, C. Shi, Regularization of nonlinear inverse problems – from functional analysis to data-driven approaches, 2025. arXiv:2506.17465

    Pith/arXiv arXiv 2025

  37. [37]

    M. T. McCann, M. Unser, Biomedical image recon- struction: From the foundations to deep neural net- works, Found. Trends Signal Process. 13 (2019) 283–359. doi:10.1561/2000000101

    work page doi:10.1561/2000000101 2019

  38. [38]

    Roelant, D

    R. Roelant, D. Constales, G. S. Yablonsky, R. V . Keer, M. A. Rude, G. B. Marin, Noise in temporal analysis of products (tap) pulse responses, Catalysis Today 121 (2007) 269–281. The TAP Reactor in Catalysis: Recent Advances in Theory and Practise. 15

    2007