Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

A variational quantum algorithm based on a weak formulation solves the 1D diffusion equation with a piecewise-constant diffusivity and interface jumps; simulations indicate pronounced hydroxide gradients appear only when the layer diffusivi

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 · deepseek-v4-flash

2026-08-03 19:16 UTC pith:LKUOT723

load-bearing objection Solid proof-of-concept for a VQA with piecewise-constant diffusivity and interface jump conditions, but the abstract's "ratio ~50" threshold claim is unsupported and should be qualified or cut. the 3 major comments →

arxiv 2512.01120 v2 pith:LKUOT723 submitted 2025-11-30 physics.flu-dyn quant-ph

Variational quantum algorithm for anion exchange across electrolyzer membrane

Timur Gubaev , Philipp Pfeffer , Christian Dre{\ss}ler , J\"org Schumacher This is my paper
classification physics.flu-dyn quant-ph
keywords variational quantum algorithmdiffusion equationanion exchange membranepiecewise-constant diffusivityweak formulationinterface jump conditionsexpressibilityhydroxide transport
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 paper aims to show that a variational quantum algorithm—a hybrid quantum-classical optimizer—can solve a physically realistic transport problem: hydroxide diffusion through a two-layer anion-exchange membrane, where the diffusion coefficient jumps at the layer interface. It reformulates the diffusion equation in a weak variational form, splits the field into steady and transient parts, and encodes the transient part in a parameterized quantum circuit; an analytical solution of the same problem provides a benchmark. In ideal statevector simulations with 4 to 6 qubits (16 to 64 grid points), the algorithm tracks the finite-difference benchmark, with errors influenced by the circuit's expressibility and by a shortage of variational parameters at 6 qubits. The physically striking result is that the relaxation time and concentration gradients depend sharply on the diffusivity ratio, and pronounced gradients—interpreted as precursors of chemical instability—occur only for ratios above about 50. This matters because it gives membrane designers a concrete quantitative target and demonstrates a path toward quantum simulation of layered transport systems.

Core claim

On the paper's own terms, the central claim is that a variational quantum algorithm built on a weak variational formulation solves the one-dimensional diffusion equation with piecewise-constant diffusivity, fixed boundary concentrations, and interface flux-continuity jumps. The algorithm decomposes the concentration into a time-independent steady state plus a transient, discretizes the transient equation in time with a backward difference, and turns each time step into a quadratic minimization whose terms are evaluated by ancilla-assisted inner-product estimation circuits on a real-amplitude parameterized quantum circuit. A successive-bisection state-preparation routine encodes the piecewise

What carries the argument

The load-bearing object is the weak-form cost function whose terms are quadratic and linear forms over the discrete transient concentration; each term is evaluated with an ancilla-assisted estimation circuit, and the normalized transient profile is generated by a real-valued parameterized quantum circuit (RPQC) with reverse linear entanglement. Classical optimizers—a quasi-Newton method, a simplex-based method, and a surrogate-based method—minimize the nonconvex cost. The piecewise-constant coefficient states are prepared by a successive-bisection routine that recursively splits constant segments and applies controlled rotation gates, giving O(p n^2) gate complexity for p constant pieces. Th

Load-bearing premise

The fixed-depth parameterized quantum circuit is assumed to be expressive enough to represent the discrete transient solution at every time step, and the nonconvex classical optimization is assumed to reach a good global minimum of the cost function; the paper's own 6-qubit results show more grid points than variational parameters, and expressibility is measured only statistically.

What would settle it

Using the paper's analytical solution, compute the maximum transient concentration gradient for diffusivity ratios finely spaced around 50 at several fixed times; if pronounced gradients appear for ratios clearly below 50, or fail to appear above 50, the threshold claim is false. Alternatively, run the same algorithm on real noisy hardware with shot noise at 6 qubits: if the mean squared error no longer tracks the ideal-statevector result beyond sampling error, the proof-of-concept does not carry over to noisy devices.

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

If this is right

  • For 16, 32, and 64 grid points, the VQA reproduces the finite-difference benchmark in ideal statevector simulations; its error is bounded below by the finite-difference error, so the quantum algorithm's accuracy is at most as good as the classical discretization it mimics.
  • The VQA can handle non-periodic boundary conditions and interface jump conditions in the diffusivity, extending earlier quantum-algorithm formulations that assumed periodic or constant-diffusivity settings.
  • The relaxation time to steady state in a two-layer membrane is governed by the smallest separation constant, and the simulations identify a diffusivity ratio of about 50 as the threshold above which pronounced hydroxide gradients—and thus potential chemical instabilities—appear.
  • A quasi-Newton optimizer is the most reliable classical optimizer for this cost function among those tested; the surrogate-based method trails by orders of magnitude in mean squared error for equal function-evaluation budgets.
  • At 6 qubits the number of variational parameters is smaller than the number of grid points, so the discrete profile is underdetermined and spatial oscillations appear; this shows that expressibility alone does not guarantee representability of the solution.

Where Pith is reading between the lines

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

  • An extension the authors leave implicit: the threshold of roughly 50 is established for one two-layer geometry with specific boundary concentrations; the paper's own analytical solution could be used to map how the threshold shifts with layer thickness, number of layers, and boundary values.
  • The successive-bisection state-preparation algorithm is not limited to this diffusion problem; any transport or wave problem with piecewise-constant coefficients in layered media could reuse the same polynomial-cost preparation, which is testable on near-term quantum hardware.
  • The underdetermination at 6 qubits suggests that a problem-specific ansatz with a parameter count matching the grid resolution—rather than a generic parameterized circuit—is likely needed for larger systems; this is an editorial inference, not a claim proven by the paper's simulations.
  • Because the model omits concentration-dependent diffusivity and water transport, the diffusivity-ratio threshold should be read as a conservative design guideline; including nonlinear feedback could plausibly make instability appear at lower ratios.

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

3 major / 4 minor

Summary. The paper presents a variational quantum algorithm (VQA) based on a weak formulation of the one-dimensional diffusion equation with a piecewise-constant diffusion coefficient, targeting hydroxide-ion transport through a two-layer anion-exchange membrane. The authors derive an analytical series solution for the initial-boundary-value problem, use it and a conservative finite-difference method as benchmarks, and report ideal statevector VQA simulations for 4, 5, and 6 qubits (16, 32, and 64 grid points) over time steps. For 4 and 5 qubits the VQA reproduces the reference solutions with MSE around 1e-5 to 1e-4; the 6-qubit results are explicitly reported as underdetermined and show spatial oscillations. The abstract additionally claims that pronounced concentration gradients and chemical instabilities can occur only for a membrane diffusivity ratio above approximately 50.

Significance. If scoped to the 4- and 5-qubit results, the paper is a sound proof-of-concept: it demonstrates that a weak-form VQA can handle interface jump conditions and non-periodic Dirichlet boundaries, and it does so with an honest comparison against an independently derived analytical solution and a conservative finite-difference scheme. The piecewise-constant state-preparation algorithm with O(p n^2) gate complexity is a useful practical contribution. However, the advertised 6-qubit / 64-point demonstration is not actually successful, the physical diffusivity-ratio threshold in the abstract has no evidentiary basis, and the abstract overstates the simulations performed. These issues are fixable but currently prevent the results from being accepted as stated.

major comments (3)
  1. [Abstract; Sec. IV C] The abstract's claim that 'pronounced hydroxide ion concentration gradients, and thus chemical instabilities, can occur only when the ratio of diffusivity in both layers of the membrane exceeds approximately 50' is not supported by the manuscript. The parameter sweep in Sec. IV C covers D2 = 0.75, 0.50, 0.25, 0.01 with D1 = 1, i.e., ratios 1.33, 2, 4, and 100. No criterion for 'pronounced' gradients or 'chemical instability' is defined, and the relaxation analysis in Sec. II B (tau_s = 1/lambda_1) concerns the slowest decay mode, not gradient sharpness. The threshold appears only in the abstract and never in the derivation, results, or conclusion. Please either remove it or substantiate it with a quantitative measure of gradient steepness and a finer scan that actually brackets a threshold.
  2. [Sec. IV C 2; Fig. 10] The paper claims in the abstract and introduction that the algorithm is demonstrated for 16 to 64 grid points (n = 4 to 6). However, Sec. IV C 2 and Fig. 10(b) show that the 6-qubit VQA has fewer variational parameters (42 for d = 6) than the 64-dimensional solution space, is underdetermined, and produces oscillations not present in the FDM or analytical solutions; its MSE is also higher than the 4-qubit MSE in Fig. 8. The text honestly reports this, but the 'demonstrate applicability' claim should be limited to 16 and 32 grid points unless a more expressive, regularized, or otherwise corrected 6-qubit construction is provided. As written, the 64-point simulation does not constitute a successful demonstration.
  3. [Abstract; Sec. I, IV, V] The abstract states that the simulations include 'ideal statevector and shot-based quantum simulations implemented using Qiskit,' but the paper reports only ideal statevector simulations (Sec. IV A-C). Shot-based and noisy simulations are explicitly listed as future work in the Conclusion ('noisy quantum circuit and shot-based simulations have to be conducted'). Please correct the abstract to reflect the actual simulations performed, or add the missing shot-based results.
minor comments (4)
  1. [Sec. IV A] The expressibility analysis measures how well the RPQC approximates the Haar-random fidelity distribution, but it does not guarantee representability of the specific transient solution. This is acknowledged indirectly by the 6-qubit underdetermination; a sentence clarifying that expressibility is necessary but not sufficient would help readers interpret Fig. 6.
  2. [Sec. II A] The analytical solution derivation is thorough, but the notation using x-tilde for interfaces and x for grid points is easy to confuse. Consider using a different symbol set, for example y_j for interfaces.
  3. [Fig. 3] The units in Fig. 3(b) are printed as tau_s in ell^2/D and 1/lambda_1, but it is not immediately clear what the abscissa represents; please label both axes explicitly with units and variables.
  4. [Sec. IV B] In Eq. (74), the empirical parameter gamma is introduced, and later gamma = 1 is chosen. A brief sentence justifying this choice or stating its sensitivity would improve reproducibility.

Circularity Check

0 steps flagged

No significant circularity: the VQA solution is validated against an independently derived analytical solution and a conservative FDM; the abstract's diffusivity-ratio threshold is poorly supported but is not circular.

full rationale

The central derivation is self-contained against independent benchmarks. The VQA minimizes the weak-form cost function (Eq. 63) whose discretized diffusion terms are evaluated by Hadamard-test circuits, and the resulting transient profiles are compared with the analytical series solution (Eq. 43) and a conservative finite-difference scheme (Eq. 46). The analytical solution is derived in the paper via separation of variables and a numerically solved eigencondition (Eqs. 38-41), not fitted to the VQA output, and the FDM is a standard independent discretization. No parameter is fitted to the benchmark and then renamed as a prediction. The paper's use of prior work—notably the weak formulation of Bengoechea et al. [26] and the boundary-circuit treatment of Over et al. [27]—is methodological citation rather than a self-citation chain, and the core benchmark does not rest on those citations. The authors' own prior work [25,34] is mentioned for context and comparison, not as the load-bearing justification. The paper honestly discloses its main weaknesses: the RPQC expressibility is chosen by a KL-divergence heuristic rather than proven, and the 6-qubit case is underdetermined because the number of variational parameters is less than the number of grid points. These are correctness/robustness limitations, not circular reasoning. The abstract's claim that pronounced gradients require a diffusivity ratio above approximately 50 is not derived in the text: the D2 sweep in Sec. IV C covers ratios 1.33-100, and no quantitative criterion for 'pronounced' gradients is given. That is an evidentiary gap, but it is not an input defined in terms of the output or a fitted quantity presented as a prediction. Accordingly, no circular step is identified.

Axiom & Free-Parameter Ledger

1 free parameters · 5 axioms · 0 invented entities

The central VQA result rests on the weak-formulation framework from prior work, the physical diffusion model, and an unproven expressibility/optimization assumption. No free parameters are fitted to the analytical benchmark or to the physical threshold; the only hand-set constant is an optimizer heuristic. No new physical entities are introduced.

free parameters (1)
  • γ (HPS heuristic) = 1
    Empirical parameter in the surrogate-based optimization (HPS variant) chosen to double the anticipated distance between variational parameters; affects optimizer convergence only, not the PDE solution or the physical threshold claim.
axioms (5)
  • standard math Separation-of-variables eigenfunction expansion is complete for the 1D diffusion operator with piecewise-constant D and flux-continuity interface conditions.
    Used to construct the analytical solution in Sec. II A 2; standard Sturm-Liouville result but not restated as a theorem.
  • standard math The conservative finite-difference scheme with staggered coefficients is stable and convergent under the CFL condition (51).
    Relied on for the FDM baseline; the paper cites ref. [40] for sufficiency, but does not re-derive stability.
  • domain assumption The weak formulation of Bengoechea et al. [26] with backward Euler and midpoint quadrature is equivalent to the original PDE.
    Adopted from prior work and used as the basis of the cost function (Eqs. 58-63); assumed without independent verification in this paper.
  • domain assumption OH- transport across the AEM is macroscopically modeled by Fickian diffusion with space-dependent D, with continuity of concentration and flux at layer interfaces.
    Physical model described in the introduction and encoded in interface conditions (6)-(7); the two-layer reduction in Fig. 1(b) is also assumed.
  • ad hoc to paper The RPQC ansatz with depth d is sufficiently expressive and the classical optimizer finds a good global minimum of the nonconvex cost function.
    Needed for VQA accuracy; expressibility is only measured statistically (Sec. IV A), and the 6-qubit underdetermination shows the assumption is not guaranteed.

pith-pipeline@v1.3.0-alltime-deepseek · 170 in / 9375 out tokens · 152875 ms · 2026-08-03T19:16:54.737097+00:00 · methodology

0 comments
read the original abstract

We present a variational quantum algorithm that solves the one-dimensional diffusion problem with a space-dependent diffusion constant $D(x)$. This problem is relevant for the exchange of hydroxide ions across a two-layer membrane in an alkaline electrolyzer, where the concentration of OH$^-$ ion determines the chemical stability for longer time periods. We use $16$ to $64$ grid points across the membrane, resulting from $n=4$ to 6 data qubits for the ideal statevector and shot-based quantum simulations implemented using Qiskit. For these qubit numbers, the depth of the parametric quantum circuit has been chosen to ensure sufficient expressibility. The state preparation requires particular attention since the diffusivity $D$ is piecewise constant in the different layers with discontinuities at the interface. Furthermore, we compare different classical optimization schemes with respect to their convergence in the VQA method. We demonstrate the applicability of the quantum algorithm to a problem with non-trivial boundary conditions and jump conditions of the diffusion constant and outline possible extensions of the proof-of-concept application case of quantum computing. Our simulations show that pronounced hydroxide ion concentration gradients, and thus chemical instabilities, can occur only when the ratio of diffusivity in both layers of the membrane exceeds approximately 50.

Figures

Figures reproduced from arXiv: 2512.01120 by Christian Dre{\ss}ler, J\"org Schumacher, Philipp Pfeffer, Timur Gubaev.

Figure 1
Figure 1. Figure 1: FIG. 1: Sketch of the anion-exchange membrane (AEM). (a) The 4-layer AEM electrolysis system. (b) The [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2: Analytical solution compared at different time instants together with the finite-difference solution. Panel (a) [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3: Relaxation rate analysis of the problem. (a) The function [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4: Quantum circuits for computing individual terms of the cost function. Circuit (a) evaluates the terms [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5: Sketch of the principle of the state preparation algorithm shown in panels (a)-(d) and a general quantum [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6: Real-valued parametric quantum circuit (RPQC) and its expressibility in terms of the Kullback-Leibler [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7: Sketch of the maximum distance max [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8: Mean squared error (MSE) versus time step [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9: Comparison between the time evolutions shown in panels (a)-(c) for [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10: Concentration profile for [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗
Figure 12
Figure 12. Figure 12: FIG. 12: Comparison of various classical optimization [PITH_FULL_IMAGE:figures/full_fig_p017_12.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11: Number of iterations of the BFGS [PITH_FULL_IMAGE:figures/full_fig_p017_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.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Quantum-enhanced Markov chain Monte Carlo sampling to model Lagrangian tracer dispersion in turbulent boundary layer

    physics.flu-dyn 2026-06 unverdicted novelty 6.0

    QE-MCMC with quantum-circuit proposals samples height-dependent turbulent accelerations for Lagrangian tracers, producing pair-dispersion scaling laws that match classical MCMC and stochastic models while showing larg...

Reference graph

Works this paper leans on

57 extracted references · 1 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    ,˜xm−1 between the layers where discontinuities arise

    Steady-state analytical solution To obtain the steady-state solution in the presence of discontinuities, the right-hand side of (1) is equated to zero inmindividual regions (˜x j−1,˜xj) with each region corresponding to thej-th layer, and, addition- ally, interface conditions are imposed at the interfaces ˜x1,˜x2, . . . ,˜xm−1 between the layers where dis...

  2. [2]

    Ajk exp i λkp Dj ˜xj−1 ! +iB jk exp i λkp Dj ˜xj−1 ! # , (25) ˜Bjk = 1 2

    Full time-dependent analytical solution Similarly to the heat equation with uniform thermal diffusivity, eq.(1) is solved by the method of separation of variables. We use the ansatz c(x, t) =K(x)T(t).(18) In each region (˜xj−1,˜xj), (1) simplifies as: ∂c ∂t =D j ∂2c ∂x2 in Ω T = (˜xi−1,˜xi)×(0, T].(19) Substituting (18) into (19) and rearranging terms giv...

  3. [3]

    0 0 0

    0 0 0 0r 3 2 s2 r 5 2 . . .0 0 0 ... ... . . . . . . . . . ... ... ... 0 0 0 0. . . r N− 1 2 sN rN+ 1 2 0 0 0 0. . .0 0 1   .(50) The first and the last rows ofAserve to incorporate the inhomogeneous Dirichlet boundary conditions located at cl 0 andc l N+1 . To ensure that this scheme is stable, the generalization of the Courant–Friedrichs–Lewy (...

  4. [4]

    In each iteration, the process ensures that the approximation resulting from the bisection can be progressively refined in subsequent iterations

    Successive bisection The underlying principle of the algorithm employs a recursive dichotomy process: starting with an initial con- stant vector, the procedure iteratively bisects it into con- stant subsegments using a finite number of iterations until the constructed vector precisely matches the de- sired target state. In each iteration, the process ensu...

  5. [5]

    From the algorithmic description of the procedure, it is clear that there is one single-qubit 11 (e) (a) (b) (c) (d) FIG

    Gate complexity estimate A circuit generated by Algorithm 1 requires three types of components, namely, single-qubitR Y , HadamardH, and MCRY gates. From the algorithmic description of the procedure, it is clear that there is one single-qubit 11 (e) (a) (b) (c) (d) FIG. 5: Sketch of the principle of the state preparation algorithm shown in panels (a)-(d) ...

  6. [6]

    E. Delp, R. Mishra, and E. Wagner, Evaluation of com- plex multi-physics phenomena at gas diffusion electrodes during high-pressure water electrolysis with aems, Sci. Rep.15, 24866 (2025)

    2025

  7. [7]

    The inter- face position ˜x1 is chosen to be 10/11ℓas in Fig

    Quantum simulation parameters The diffusion equation (1) was solved approximately for 100 time steps with an integration step width ∆t= ∆x2/2 using the FDM and VQA methods withD 1 = 1ℓ 2/τD,D 2 = 0.01,0.25,0.50,0.75ℓ 2/τD andN= 16,32,64 with ideal state-vector simulations. The inter- face position ˜x1 is chosen to be 10/11ℓas in Fig. 1(b), with ˜x0 = 0ℓan...

  8. [8]

    Time evolution and dependence on strength of discontinuity of diffusion constant Inspecting the graphs, one can deduce that, in terms of MSE, the FDM always performs better than the VQA. The reason is that while the VQA utilizes the same fi- nite differences as the FDM, it uses an ansatz circuit to express the solution, which inevitably introduces error a...

  9. [9]

    11(a)-(b)

    Results for different optimization schemes The number of iterations of the BFGS algorithm used in the VQA for various numbers of qubits and different values ofD 2 is depicted in Figs. 11(a)-(b). As shown, whenD 2 = 0.01ℓ 2/τD, the BFGS algorithm requires the fewest iterations compared to other cases. For example, Fig. 11(a) illustrates that after approxim...

    1953

  10. [10]

    Z. Yan, J. L. Hitt, J. A. Turner, and T. E. Mallouk, Re- newable electricity storage using electrolysis, Proc. Natl. Acad. Sci. USA117, 12558 (2020)

    2020

  11. [11]

    El-Shafie, Hydrogen production by water electroly- sis technologies: A review, Results Engin.20, 101426 (2025)

    M. El-Shafie, Hydrogen production by water electroly- sis technologies: A review, Results Engin.20, 101426 (2025)

    2025

  12. [12]

    H. I. Lee, M. Mehdi, S. K. Kim, H. S. Cho, M. J. Kim, W. C. Cho, Y. W. Rhee, and C. H. Kim, Advanced zirfon- type porous separator for a high-rate alkaline electrolyser operating in a dynamic mode, J. Membr. Sci.616, 118541 (2020)

    2020

  13. [13]

    Schalenbach, W

    M. Schalenbach, W. Lueke, and D. Stolten, Hydrogen diffusivity and electrolyte permeability of the zirfon perl separator for alkaline water electrolysis, J. Electrochem. Soc.163, F1480 (2016)

    2016

  14. [14]

    Brauns, J

    J. Brauns, J. Sch¨ onebeck, M. R. Kraglund, D. Aili, J. Hn´ at, J.ˇZitka, W. Mues, J. O. Jensen, K. Bouzek, and T. Turek, Evaluation of diaphragms and membranes as separators for alkaline water electrolysis, J. Electrochem. Soc.168, 014510 (2021)

    2021

  15. [15]

    Nielsen and I

    M. Nielsen and I. Chuang,Quantum Computation and Quantum Information: 10th Anniversary Edition(Cam- bridge University Press, Cambridge, UK, 2010)

    2010

  16. [16]

    N. Du, C. Roy, R. Peach, M. Turnbull, S. Thiele, and C. Bock, Anion-exchange membrane water electrolyzers, Chem. Rev.122, 11830 (2022)

    2022

  17. [17]

    Henkensmeier, W

    D. Henkensmeier, W. C. Cho, P. Jannasch, J. Stojadi- novic, Q. Li, D. Aili, and J. O. Jensen, Separators and membranes for advanced alkaline water electrolysis, Chem. Rev.124, 6393 (2024)

    2024

  18. [18]

    Sugawara, S

    Y. Sugawara, S. Sankar, S. Miyanishi, R. Illathvalappil, P. K. Gangadharan, H. Kuroki, G. M. Anilkumar, and T. Yamaguchi, Anion exchange membrane water elec- trolyzers: An overview, J. Chem. Eng. Jpn.56, 2210195 (2023)

    2023

  19. [19]

    M. G. Marino, J. P. Melchior, A. Wohlfarth, and K. D. Kreuer, Hydroxide, halide and water transport in a model anion exchange membrane, J. Membr. Sci.464, 61 (2018)

    2018

  20. [20]

    Zelovich, L

    T. Zelovich, L. Vogt-Maranto, M. A. Hickner, S. J. Pad- dison, C. Bae, D. R. Dekel, and M. E. Tuckerman, Hy- droxide ion diffusion in anion-exchange membranes at low hydration: Insights from ab initio molecular dynamics, Chem. Mater.31, 5778 (2019)

    2019

  21. [21]

    Chen, Y.-L

    C. Chen, Y.-L. S. Tse, G. E. Lindberg, C. Knight, and G. A. Voth, Hydroxide solvation and transport in an- ion exchange membranes, J. Am. Chem. Soc.138, 991 (2016)

    2016

  22. [22]

    Al Otmi, P

    M. Al Otmi, P. Lin, W. Schertzer, C. M. Colina, R. Ram- prasad, and J. Sampath, Investigating correlations in hy- droxide ion transport in anion exchange membranes from atomistic molecular dynamics simulations, ACS Appl. Polym. Mater.6, 11270 (2024)

    2024

  23. [23]

    A. J. Bard and L. R. Faulkner,Electrochemical Methods – Fundamentals and Applications(John Wiley & Sons, New York, 2001)

    2001

  24. [24]

    Jaksch, P

    D. Jaksch, P. Givi, A. J. Daley, and T. Rung, Variational quantum algorithms for computational fluid dynamics, AIAA J.61, 1885 (2023)

    2023

  25. [25]

    Preskill, Quantum computing in the NISQ era and beyond, Quantum2, 79 (2018)

    J. Preskill, Quantum computing in the NISQ era and beyond, Quantum2, 79 (2018)

    2018

  26. [26]

    S. S. Bharadwaj and K. R. Sreenivasan, Quantum com- putation of fluid dynamics, Indian Academy of Sciences Conference Series3, 77 (2020). 19

    2020

  27. [27]

    Succi, W

    S. Succi, W. Itani, K. R. Sreenivasan, and R. Steijl, Quantum computing for fluids: Where do we stand?, Eu- rophys. Lett.144, 10001 (2023)

    2023

  28. [28]

    Tennie, S

    F. Tennie, S. Laizet, S. Lloyd, and L. Magri, Quantum computing for nonlinear differential equations and tur- bulence, Nat. Rev. Phys.7, 1 (2025)

    2025

  29. [29]

    Cerezo, A

    M. Cerezo, A. Arrasmith, R. Babbush, S. C. Benjamin, S. Endo, K. Fujii, J. R. McClean, K. Mitarai, X. Yuan, L. Cincio, and P. J. Coles, Variational quantum algo- rithms, Nat. Rev. Phys.3, 625 (2021)

    2021

  30. [30]

    Demirdjian, D

    R. Demirdjian, D. Gunlycke, C. A. Reynolds, J. D. Doyle, and S. Tafur, Variational quantum solutions to the advec- tion–diffusion equation for applications in fluid dynamics, Quant. Inf. Process.21, 322 (2020)

    2020

  31. [31]

    F. Y. Leong, W.-B. Ewe, and D. E. Koh, Variational quantum evolution equation solver, Sci. Rep.12, 10817 (2022)

    2022

  32. [32]

    F. Y. Leong, D. E. Koh, W.-B. Ewe, and J. F. Kong, Vari- ational quantum simulation of partial differential equa- tions: Applications in colloidal transport, Int. J. Num. Meth. Heat Fluid Flow33, 3669 (2023)

    2023

  33. [33]

    A. J. Pool, A. D. Somoza, C. McKeever, M. Lubasch, and B. Horstmann, Nonlinear dynamics as a ground-state so- lution on quantum computers, Phys. Rev. Res.6, 033257 (2024)

    2024

  34. [34]

    Ingelmann, S

    J. Ingelmann, S. Bharadwaj, P. Pfeffer, K. Sreenivasan, and J. Schumacher, Two quantum algorithms for solving the one-dimensional advection-diffusion equation, Com- put. Fluids281, 106369 (2024)

    2024

  35. [35]

    Bengoechea, P

    S. Bengoechea, P. Over, D. Jaksch, and T. Rung, Toward variational quantum algorithms for generalized linear and nonlinear transport phenomena, AIAA J. , 1 (2025)

    2025

  36. [36]

    P. Over, S. Bengoechea, T. Rung, F. Clerici, L. Scan- durra, E. de Villiers, and D. Jaksch, Boundary treatment for variational quantum simulations of partial differential equations on quantum computers, Comput. Fluids288, 106508 (2025)

    2025

  37. [37]

    N. M. Guseynov, A. A. Zhukov, W. V. Pogosov, and A. V. Lebedev, Depth analysis of variational quantum al- gorithms for the heat equation, Phys. Rev. A107, 052422 (2023)

    2023

  38. [38]

    Y. Y. Liu, Z. Chen, C. Shu, S. C. Chew, B. C. Khoo, X. Zhao, and Y. D. Cui, Application of a variational hybrid quantum-classical algorithm to heat conduction equation and analysis of time complexity, Phys. Fluids 34, 117121 (2023)

    2023

  39. [39]

    Lubasch, J

    M. Lubasch, J. Joo, P. Moinier, M. Kiffner, and D. Jaksch, Variational quantum algorithms for nonlinear problems, Phys. Rev. A101, 010301(R) (2020)

    2020

  40. [40]

    K¨ ocher, H

    N. K¨ ocher, H. Rose, S. S. Bharadwaj, J. Schumacher, and S. Schumacher, Numerical solution of nonlin- ear Schr¨ odinger equation by a hybrid pseudospectral- variational quantum algorithm, Sci. Rep.15, 23478 (2025)

    2025

  41. [41]

    A. J. Pool, A. D. Somoza, M. Lubasch, and B. Horstmann, Solving partial differential equations us- ing a quantum computer, 2022 IEEE International Con- ference on Quantum Computing and Engineering , 864 (2022)

    2022

  42. [42]

    Javadi-Abhari, M

    A. Javadi-Abhari, M. Treinish, K. Krsulich, C. J. Wood, J. Lishman, J. Gacon, S. Martiel, P. D. Nation, L. S. Bishop, A. W. Cross, B. R. Johnson, and J. M. Gambetta, Quantum com- puting with Qiskit, preprint arXiv:2405.08810v3 https://doi.org/10.48550/arXiv.2405.08810 (2024)

    Pith/arXiv arXiv doi:10.48550/arxiv.2405.08810 2024

  43. [43]

    isP Haar(F) = (N−1)(1−F) N−2 withN= 2 n. Therefore, the measure of expressibility is: DKL = Z 1 0 PRPQC(F;λ) log PRPQC(F;λ) PHaar(F) dF(70) Since Haar random states uniformly sample the Hilbert space of pure states, the smaller theD KL, the better the states generated by the RPQC approximate this uniform distribution over the space, and therefore, the hig...

  44. [44]

    Pfeffer, P

    P. Pfeffer, P. Brearley, S. Laizet, and J. Schumacher, Spectral quantum algorithm for passive scalar transport in shear flows, Sci. Rep.15, 41172 (2025)

    2025

  45. [45]

    Shaffer, L

    R. Shaffer, L. Kocia, and M. Sarovar, Surrogate-based optimization for variational quantum algorithms, Phys. Rev. A107, 032415 (2023)

    2023

  46. [46]

    Holmes, K

    Z. Holmes, K. Sharma, M. Cerezo, and P. J. Coles, Con- necting Ansatz Expressibility to Gradient Magnitudes and Barren Plateaus, PRX Quantum3, 010313 (2022)

    2022

  47. [47]

    S. Sim, P. D. Johnson, and A. Aspuru-Guzik, Expressibil- ity and entangling capability of parameterized quantum circuits for hybrid quantum-classical algorithms, Adv. Quantum Technol.2, 1900070 (2019)

    2019

  48. [48]

    Bressloff, Accumulation time of diffusion in a two- dimensional singularly perturbed domain, Proc

    P. Bressloff, Accumulation time of diffusion in a two- dimensional singularly perturbed domain, Proc. R. Soc. A478, 20210847 (2022)

    2022

  49. [49]

    A. G. Neelan, F. S. Don Bosco, N. S. Jarugumalli, and S. B. Vedarethinam, Physics-informed neural net- works: Bridging the divide between conservative and non-conservative equations, preprint arXiv:2506.22413 https://doi.org/10.48550/arXiv.2506.22413 (2025)

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2506.22413 2025

  50. [50]

    J. Lee,Stability of finite difference schemes on the dif- fusion equation with discontinuous coefficients, Master’s thesis, Department of Mathematics, Massachusetts Insti- tute of Technology (2017)

    2017

  51. [51]

    Zindorf and S

    B. Zindorf and S. Bose, Efficient implementation of multi- controlled quantum gates, Phys. Rev. Applied24, 044030 (2025)

    2025

  52. [52]

    ˙Zyczkowski and H.-J

    K. ˙Zyczkowski and H.-J. Sommers, Average fidelity be- tween random quantum states, Phys. Rev. A71, 032313 (2005)

    2005

  53. [53]

    J. A. Nelder and R. Mead, A simplex method for function minimization, Comput. J.7, 308 (1965)

    1965

  54. [54]

    Liu, S.-X

    S. Liu, S.-X. Zhang, S.-K. Jian, and H. Yao, Training variational quantum algorithms with random gate acti- vation, Phys. Rev. Res.5, L032040 (2023)

    2023

  55. [55]

    Arrasmith, M

    A. Arrasmith, M. Cerezo, P. Czarnik, L. Cincio, and P. J. Coles, Effect of barren plateaus on gradient-free optimization, Quantum5, 558 (2021)

    2021

  56. [56]

    Yassin, I

    K. Yassin, I. G. Rasin, S. Brandon, and D. R. Dekel, Quantifying the critical effect of water diffusivity in anion exchange membranes for fuel cell applications, J. Membr. Sci.608, 118206 (2020)

    2020

  57. [57]

    D. R. Dekel, I. G. Rasin, M. Page, and S. Brandon, Steady state and transient simulation of anion exchange membrane fuel cells, J. Power Sources375, 191–204 (2018)

    2018