pith. sign in

arxiv: 2605.15973 · v1 · pith:E4VVQZ4Inew · submitted 2026-05-15 · 🧮 math.AP

On the rate of convergence to steady state in a linear chromatography model

Joaqu\'in Menacho , Marta Pellicer , J. Sol\`a-Morales This is my paper

Pith reviewed 2026-05-20 16:26 UTC · model grok-4.3

classification 🧮 math.AP
keywords modelconvergencelinearrateanalyticalcasechromatographyequations
0
0 comments X

The pith

Proves the convergence rate to steady state in the linear TMB chromatography model is set by the dominant eigenvalue of the coupled hyperbolic system, with explicit characteristic function and application to omeprazole enantiomer separation.

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

Chromatography separates chemicals by how fast they move through a column with a moving solid and liquid phase. The True Moving Bed model uses eight linear partial differential equations that track concentrations in different zones. The authors show that after a long time the concentrations approach a steady pattern at a speed determined by the largest eigenvalue of the system. They prove this eigenvalue exists using a theorem about positive operators and build a characteristic equation whose roots give all the eigenvalues. For very large times they describe the shape of the solution using both math and computer calculations. They then plug in numbers for separating two forms of the drug omeprazole.

Core claim

We prove that the rate of convergence is given by a dominant eigenvalue, whose existence we prove by means of the Krein-Rutman Theorem, and by comparison arguments.

Load-bearing premise

The coupled system of eight hyperbolic PDEs with the given boundary conditions defines a positive operator to which the Krein-Rutman theorem applies directly, yielding a simple dominant eigenvalue that controls the decay rate.

Figures

Figures reproduced from arXiv: 2605.15973 by Joaqu\'in Menacho, J. Sol\`a-Morales, Marta Pellicer.

Figure 1
Figure 1. Figure 1: Left: scheme of an SMB. Right: scheme of a TMB model. The circles represent the less adsorbable component, while the squares represent the more adsorbable one. The arrows on the left figure represent the movement of the ports at regular time intervals, and in the right figure, the direction of circulation of the liquid and solid phases. 1.1 The equations The simplest model of countercurrent linear chromato… view at source ↗
Figure 2
Figure 2. Figure 2: Scheme of the situation described by (1.4), (1.5), and (1.6). Remark 1.1 In the limit case of the inequalities (1.6), namely when vi = v > 0, i = 1, 2, 3, 4, the contour conditions (1.5) become simply continuity conditions for c and q at all the boundaries of each interval Ii. This is a case of not much practical interest, but it is the unique case where we have been able to find an explicit and complete p… view at source ↗
Figure 3
Figure 3. Figure 3: Left: partial plots of the graph of the function ∆(λ), for R = 9, 18, 27, 36 (vi for i = 1, . . . 4, and P as in the case study of Section 7), at quite different scales. Right: complete plots of the graph of the function (2/π) arctan(∆(λ)), for the same values of R, at the same scale. Observe that the graphs are very steep near their largest real zero (see [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Plot of first eigenvalues λ + (in blue-star) and λ − (in red-circle), k = −80, . . . 80, for equal velocities v = 1.275, and R = 18. Left: for P = 1.03 (> 1); right: P = 0.5 (< 1). 7 A case study: steady state, λ0, dominant eigenfunction and spectrum, and sensitivity analysis This section presents a case study illustrating the applicability of the results developed in this work. The practical example corre… view at source ↗
Figure 5
Figure 5. Figure 5: Steady state for (v1, v2, v3, v4, R, P) = (1.53, 1.12, 1.43, 1.02, 18, 1.03) and f0 = 1. In blue (solid line), c; in red (dash–dot line), q. Now we use an initial value solver adapted to this type of problem (see J. Menacho et al. [13, 15, 14]) to see the convergence to the steady state. Because of the linearity of the equations, we solve (1.4)-(1.5) with f0 = 0, and then we see the convergence to zero. Th… view at source ↗
Figure 6
Figure 6. Figure 6: Two views of the solution of the initial value problem with f0 = 0, and initial conditions (1, P), in the case study. First, contour plots of c(x, t), q(x, t), respectively, at different times. Second, a view of the evolution of the solution, in terms of t (and x). 22 [PITH_FULL_IMAGE:figures/full_fig_p022_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Left: time evolution of the root-mean-square difference between the solution and the mean￾scaled eigenfunction, showing the convergence of the solution toward the eigenfunction profile. Right: same graphic, at logarithmic scale. Second, we want to compute λ0, the dominant eigenvalue, in this case study. To numerically implement it, we proceed according to the following scheme: 1. We construct the matrix C(… view at source ↗
Figure 8
Figure 8. Figure 8: Eigenfunction of A (left) and of A∗ (right) for the dominant eigenvalue λ0. Blue-solid line: c(x), c∗ (x); red dash-dot line: q(x), q∗ (x). The complete calculations can be found in Appendix A.2. and ∂λ0 ∂R = 0.007208020694512, ∂λ0 ∂P = −4.091507558661688, (7.2) which have a natural interpretation, both in terms of the sign and the values. In particular, the fact λ0 is increasing in R can be observed in [… view at source ↗
Figure 9
Figure 9. Figure 9: Partial view of the spectrum of A in the case study. In blue (circles): N = 30 (low resolution); in red (crosses): N = 45 (high resolution). A Appendix: complete calculations of Mi, eigenfunctions coef￾ficients and sensitivity analysis In this section, we present the detailed calculations used in several parts of the previous sections. For clarity, they are placed in a separate Appendix. The formulas below… view at source ↗
read the original abstract

We study the rate of convergence to the steady state in the True Moving Bed model of linear chromatography, as a function of the six parameters that appear in the model. The model is a system of eight linear partial differential equations of hyperbolic type, coupled through the equations themselves and also through boundary conditions. We prove that the rate of convergence is given by a dominant eigenvalue, whose existence we prove by means of the Krein-Rutman Theorem, and by comparison arguments. We show how to construct a (not at all simple) characteristic function, whose roots are the eigenvalues. We also study the asymptotic profile of the solutions for large times, although this part is not purely analytical, but a combination of analytical and numerical techniques. Beyond the theoretical results, these models also offer explicit quantitative information: we apply all our results to a Case Study, namely the separation of omeprazole enantiomers. Finally, we consider a simpler limit case, where all the calculations become explicit.

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 paper studies the rate of convergence to steady state for the True Moving Bed (TMB) model of linear chromatography, a system of eight coupled linear hyperbolic PDEs with boundary coupling. It proves that the decay rate is controlled by a simple dominant eigenvalue whose existence follows from the Krein-Rutman theorem together with comparison arguments, constructs an explicit (though complicated) characteristic function whose roots are the eigenvalues, analyzes the large-time asymptotic profile via a combination of analysis and numerics, applies the results to the separation of omeprazole enantiomers, and treats a simplified limit case in which all quantities become explicit.

Significance. If the positivity and compactness hypotheses needed for Krein-Rutman are fully verified, the work supplies a rigorous spectral characterization of transients in a practically important class of hyperbolic systems with nonlocal boundary conditions. The explicit characteristic function and the concrete case study provide quantitative information that could be useful for parameter tuning in chromatography. The combination of analytic and numerical techniques for the asymptotic profile is a reasonable compromise, though the purely analytic content would be strengthened by a fully rigorous description of the dominant eigenfunction.

major comments (2)
  1. [§3] §3 (spectral analysis): the manuscript invokes the Krein-Rutman theorem for the generator A of the eight-component hyperbolic system, but does not supply an explicit proof that the resolvent R(λ,A) is compact for Re λ large or that the semigroup is strongly positive (or at least irreducible) on the positive cone of the product space. For hyperbolic transport operators on bounded intervals, compactness is not automatic and must be recovered from the specific inlet/outlet boundary maps; without this verification the existence of a simple, strictly dominant eigenvalue is not yet guaranteed.
  2. [§4] §4 (comparison arguments): the comparison principle used to show that the dominant eigenvalue is strictly larger in modulus than all other spectrum points relies on the same positivity property. The argument should be written out in detail for the coupled boundary conditions of the TMB model, including an explicit estimate showing that no other eigenvalue can lie on the imaginary axis.
minor comments (2)
  1. [§5] The numerical procedure for extracting the asymptotic profile (eigenfunction shape and decay rate) is described only at a high level; adding a short paragraph on the discretization scheme, mesh size, and convergence check would improve reproducibility.
  2. [§6] In the omeprazole case study, the six model parameters are taken from the literature; a single sentence citing the precise source for each value would help readers replicate the numerical example.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thorough review and valuable suggestions. We will revise the manuscript to address the concerns regarding the verification of the hypotheses for the Krein-Rutman theorem and the detailed comparison arguments.

read point-by-point responses

  1. Referee: [§3] §3 (spectral analysis): the manuscript invokes the Krein-Rutman theorem for the generator A of the eight-component hyperbolic system, but does not supply an explicit proof that the resolvent R(λ,A) is compact for Re λ large or that the semigroup is strongly positive (or at least irreducible) on the positive cone of the product space. For hyperbolic transport operators on bounded intervals, compactness is not automatic and must be recovered from the specific inlet/outlet boundary maps; without this verification the existence of a simple, strictly dominant eigenvalue is not yet guaranteed.

    Authors: We appreciate the referee's observation. The manuscript applies the Krein-Rutman theorem assuming the standard positivity and compactness properties for such transport systems, but we concur that an explicit verification is warranted given the complexity of the eight-component coupled system. In the revised version, we will include a new subsection in §3 that proves the compactness of the resolvent R(λ, A) for Re(λ) large enough. This will be done by solving the resolvent equation explicitly along the characteristic lines and showing that the boundary conditions lead to a compact perturbation. Furthermore, we will establish that the semigroup is irreducible (and in fact strongly positive) by demonstrating that any positive initial condition leads to a strictly positive solution after a finite time determined by the transport speeds, using the specific inlet/outlet maps of the TMB model. revision: yes

  2. Referee: [§4] §4 (comparison arguments): the comparison principle used to show that the dominant eigenvalue is strictly larger in modulus than all other spectrum points relies on the same positivity property. The argument should be written out in detail for the coupled boundary conditions of the TMB model, including an explicit estimate showing that no other eigenvalue can lie on the imaginary axis.

    Authors: We agree with the referee that the comparison arguments require more detail to be fully rigorous for the coupled system. We will revise §4 to provide a complete, self-contained proof of the comparison principle adapted to the eight coupled hyperbolic equations with the TMB boundary conditions. This will include showing how positivity is preserved and propagated through the periodic-like couplings at the boundaries. Moreover, we will add an explicit estimate proving that no non-dominant eigenvalue can lie on the imaginary axis: assuming λ = iω with ω real and nonzero, we derive a contradiction by integrating the equations against suitable test functions or using a maximum principle argument that exploits the strict positivity and the fact that the dominant eigenvalue is real and simple by Krein-Rutman. revision: yes

Circularity Check

0 steps flagged

No circularity: existence of dominant eigenvalue follows from external Krein-Rutman theorem plus explicit characteristic equation

full rationale

The derivation applies the standard Krein-Rutman theorem to the positive linear operator generated by the eight coupled hyperbolic PDEs with the given boundary conditions, then constructs an explicit (though complicated) characteristic function whose roots locate the eigenvalues. Neither step reduces to a self-definition, a fitted parameter renamed as prediction, nor a load-bearing self-citation; the theorem is an independent classical result and the characteristic function is derived directly from the system without presupposing the decay rate. The paper is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the applicability of the Krein-Rutman theorem to the specific linear operator generated by the eight coupled hyperbolic equations and boundary conditions; no free parameters or new entities are introduced.

axioms (1)
  • standard math The linear operator defined by the eight PDEs with the given boundary conditions is positive and compact in an appropriate function space, allowing application of the Krein-Rutman theorem.
    Invoked to guarantee existence of a dominant eigenvalue that determines the convergence rate.

pith-pipeline@v0.9.0 · 5702 in / 1102 out tokens · 50413 ms · 2026-05-20T16:26:14.580622+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [1]

    Simulated Moving Bed Strategies and Designs: From Established Systems to the Latest Developments

    J. P. Aniceto, C. M. Silva, “Simulated Moving Bed Strategies and Designs: From Established Systems to the Latest Developments”, Separation and Purification Reviews 44 (1) (2015), 41-73

    work page 2015

  2. [2]

    Aris, and N.R

    R. Aris, and N.R. Amundson, Mathematical Methods in Chemical Engineering. 2. Fist-Order Partial Differential Equations with Applications , Prentice-Hall, Englewood Cliffs (NJ) 1973. 31

    work page 1973

  3. [3]

    Arrieta, C

    I. Arrieta, C. Gaudet, C. Haskell, S. Peacock, O. Tzschatzsch, B.-C. Schulze, Industrial SMB chro- matography as an alternative to the conventional refining of a raw cane sugar, Sugar Industry 149(10) (2024), 669-677. DOI: 10.36961/si32325

    work page doi:10.36961/si32325 2024

  4. [4]

    Ching, K.H

    C.B. Ching, K.H. Chu, K. Hidajat, M.S. Uddin, Experimental and modeling studies on the transient behavior of a simulated countercurrent adsorber. Journal of Chemical Engineering of Japan 24 (5) (1991) 614-621

    work page 1991

  5. [5]

    K. J. Engel, R. Nagel, A short course on operator semigroups , Springer-Verlag New York, 2006

    work page 2006

  6. [6]

    Grosfils, C

    V. Grosfils, C. Levrie, M. Kinnaert, A. Vande Wouwer, On simplified modelling approaches to SMB processes, Computers and Chemical Engineering 31 (2007), 196-205

    work page 2007

  7. [7]

    Guiochon, B

    G. Guiochon, B. Lin, Modeling for Preparative Chromatography . Academic Press, San Diego (CA), 2003

    work page 2003

  8. [8]

    Iooss and D

    G. Iooss and D. D. Joseph, Elementary stability and bifurcation theory, second edition, Undergraduate Texts in Mathematics, Springer, New York, 1990

    work page 1990

  9. [9]

    Jeong, N

    W. Jeong, N. Jang, J. H. Lee, Bayesian optimization for quick determination of operating variables of simulated moving bed chromatography, Computers and Chemical Engineering , 192 (2025), 108872. DOI: 10.1016/j.compchemeng.2024.108872

    work page doi:10.1016/j.compchemeng.2024.108872 2025

  10. [10]

    H.-J. Kang, H. Woo, S. Mun, Optimal design of a low-pressure SMB process for continuous separation of 3,6-anhydrogalactose from galactose with high productivity and high product concentration, Sep- aration and Purification Technology, 357, Part A (2025), 130040. Doi: 10.1016/j.seppur.2024.130040

    work page doi:10.1016/j.seppur.2024.130040 2025

  11. [11]

    Krein, M.A

    M.G. Krein, M.A. Rutman, (1948) Linear operators leaving invariant a cone in a Banach space. Uspekhi Mat. Nauk. New Series (in Russian). 3 (1(23))(1948): 1–95. English translation: Amer. Math. Soc. Transl. (1950) (26)

    work page 1948

  12. [12]

    Ma, N.-H

    Z. Ma, N.-H. L. Wang, Standing Wave Analysis of SMB Chromatography: Linear Systems, AIChE Journal 43(10) (1997), 2488-2508

    work page 1997

  13. [13]

    Menacho, O

    J. Menacho, O. Pou, E. Serra, R. Nomen, X. Tomás, J. Sempere, A method for simulating adsorption columns, Afinidad 68 (2011), 86-94

    work page 2011

  14. [14]

    Menacho, Simulación numérica y optimización de los parámetros de funcionamiento de un SMB, Ph.D

    J. Menacho, Simulación numérica y optimización de los parámetros de funcionamiento de un SMB, Ph.D. thesis, IQS–Universitat Ramon Llull, Barcelona, 2012

    work page 2012

  15. [15]

    Menacho, O

    J. Menacho, O. Pou, X. Tomás, E. Serra, R. Nomen, J. Sempere, Efficient simulation of a separation column with axial diffusion and mass transfer resistance, Computers and Chemical Engineering 53 (2013), 143-152. Doi: 10.1016/j.compchemeng.2013.03.008

    work page doi:10.1016/j.compchemeng.2013.03.008 2013

  16. [16]

    Menacho, M

    J. Menacho, M. Pellicer, J. Solà-Morales, Long-time behaviour of the correlated random walk system, Evol. Equ. Control Theory , 14 (4) (2025), 841–867

    work page 2025

  17. [17]

    Menacho, J

    J. Menacho, J. Solà-Morales, Convergence to steady-state and boundary layer profiles in a linear chromatography system, SIAM Journal on Applied Mathematics , 75 (2) (2015), 745-761

    work page 2015

  18. [18]

    Meyer-Nieberg, Banach Lattices, Universitext, Springer-Verlag, Berlin, Heidelberg, 1991

    P. Meyer-Nieberg, Banach Lattices, Universitext, Springer-Verlag, Berlin, Heidelberg, 1991

    work page 1991

  19. [19]

    Nagel (Ed.), One-parameter semigroups of positive operators LNM vol

    R. Nagel (Ed.), One-parameter semigroups of positive operators LNM vol. 1184, Springer-Verlag, Berlin, Heidelberg, New York, Tokyo (1986)

    work page 1986

  20. [20]

    Nogueira, A.M

    I.B.R. Nogueira, A.M. Ribeiro, A.E. Rodrigues, J.M. Loureiro, Dynamic response to process distur- bances - A comparison between TMB/SMB models in transient regime. Computers and Chemical Engineering, 99 (2017) 230-244. Doi: 10.1016/j.compchemeng.2017.01.026

    work page doi:10.1016/j.compchemeng.2017.01.026 2017

  21. [21]

    Oh, S.-K

    T.H. Oh, S.-K. Oh, H. Kim, K. Lee, J.M. Lee, Transition model for simulated moving bed under nonideal conditions. Ind. Eng. Chem. Res. 58 (2019) 21625-21640. DOI: 10.1021/acs.iecr.9b04447

    work page doi:10.1021/acs.iecr.9b04447 2019

  22. [22]

    Pazy, Semigroups of Linear Operators and Applications to Partial Differential Equations, Springer- Verlag, New York, 1983

    A. Pazy, Semigroups of Linear Operators and Applications to Partial Differential Equations, Springer- Verlag, New York, 1983. 32

    work page 1983

  23. [23]

    W. H. Press, S. A. Teukolsky, W. T. Vetterling, B. P. Flannery, Numerical Recipes in Fortran, Cambridge University Press 1992

    work page 1992

  24. [24]

    Rajendran, G

    A. Rajendran, G. Paredes, M. Mazzotti, Simulated moving bed chromatography for the separation of enantiomers, Journal of Chromatography A , 1216 (4), 2009, 709-738. DOI: 10.1016/j.chroma.2008.10.075

    work page doi:10.1016/j.chroma.2008.10.075 2009

  25. [25]

    Rodrigues, J

    H.M. Rodrigues, J. Solà-Morales, A note on the relationship between spectral radius and norms of bounded linear operators, Matemática Contemporânea, 36 (2009), 131-136

    work page 2009

  26. [26]

    Process Concepts

    M. Schulte, K. Wekenborg, W. Wewers, “Process Concepts”, in H. Schmidt-Traub (Ed.), Preparative Chromatography, Wiley-VCH, Weinheim, 2005, pp.173-214

    work page 2005

  27. [27]

    Storti, M

    G. Storti, M. Mazzotti, M. Morbidelli, and S. Carrà, Robust design of binary countercurrent adsorp- tion separation processes. AIChE Journal 39(3) (1993), 471–492. Doi: 10.1002/aic.690390310

    work page doi:10.1002/aic.690390310 1993

  28. [28]

    Model Based Design and Optimization

    A. Susanto, K. Wekenborg, A. Epping, A. Jupke, “Model Based Design and Optimization”, in H. Schmidt-Traub (Ed.), Preparative Chromatography, Wiley-VCH, Weinheim, 2005, pp. 313-370

    work page 2005

  29. [29]

    F. Wei, B. Shen, M. Chen, From Analytical Chromatography to Simulated Moving Bed Chromatog- raphy: Resolution of Omeprazole Enantiomers, Ind. Eng. Chem. Res. 45 (2006), 1420-1425. 33

    work page 2006