Pith. sign in

REVIEW 3 major objections 4 minor 39 references

A positivity-preserving hybrid DDG method for Poisson--Nernst--Planck systems

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A modified numerical flux provably keeps ion concentrations positive in arbitrary high-order discontinuous Galerkin schemes for the Poisson–Nernst–Planck system.

desk verdict Solid 1D positivity theorem; multi-D proof has a Gauss-Lobatto exactness gap that undermines the claimed higher-dimensional result. read the letter →

arxiv 2505.24058 v1 pith:MPHXYMFL submitted 2025-05-29 math.NA cs.NA

Hailiang Liu , Zhongming Wang , Peimeng Yin This is my paper · ORCID
classification math.NAcs.NA MSC 35K4065M6065M1282C31
keywords Poisson-Nernst-PlanckpositivitypreservingdiscontinuousGalerkinmethoddirectDGfluxcellaveragesGauss-Lobattoquadratureenergydissipationhigh-orderaccuracy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to prove that an arbitrary high-order discontinuous Galerkin discretization of the Poisson–Nernst–Planck system can be made positivity-preserving, not just in practice but by theorem. The earlier version of the scheme conserved mass and dissipated free energy, but it could generate negative cell averages of ion concentrations, which then made its own positivity limiter unusable. The fix is a local correction to the numerical flux, added to the logarithmic flux formulation, which makes every term in the Gauss-Lobatto decomposition of the updated cell average non-negative. With a time-step restriction, cell averages stay positive, and a reconstruction limiter converts that into pointwise positivity. The paper extends the proof to rectangular meshes in several dimensions and verifies high-order accuracy, positivity, mass conservation, and energy dissipation numerically.

What carries the argument

The central object is the modified DDG normal-derivative flux $\widehat{\partial_x p_i} = \hat{\partial}_x p_i + \frac{\tilde{\beta}_i}{2}[c_i]$, where $\tilde{\beta}_i = |\hat{\partial}_x p_i|/\{c_i\}$ when the interface average is positive. This local correction makes every boundary term in the Gauss-Lobatto quadrature representation of the updated cell average non-negative, so the average becomes a sum of positive quantities. The companion positivity-preserving reconstruction (3.19) rescales each polynomial toward its cell average so its minimum stays above a small threshold without changing the average.

What would settle it

A single run of scheme (3.15) with $c^n_{ih}>0$ that yields a non-positive cell average $\bar{c}^{n+1}_{ij}$ while the time step obeys (3.18) would refute Theorem 3.1.

Watch

Extended reading notes

Core claim

The central claim is Theorem 3.1: for the fully discrete DDG scheme (3.15), if the current concentration polynomial is pointwise positive and the time step satisfies (3.18), then the next cell average $\bar{c}^{n+1}_{ij}$ is positive. The proof rewrites the cell-average update as a Gauss-Lobatto quadrature and shows that the modified flux $\widehat{\partial_x p_i} = \widehat{\partial_x p_i} + \frac{\tilde{\beta}_i}{2}[c_i]$ with $\tilde{\beta}_i = |\widehat{\partial_x p_i}|/\{c_i\}$ makes all interface coefficients non-negative. The same argument is carried to tensor-product rectangular meshes in Theorem 4.1. To obtain pointwise positivity, the paper applies a positivity-preserving reconstruction limiter that keeps each polynomial above a small threshold $\delta$ without changing its cell average and without degrading accuracy when $\delta < h^{k+1}$. The result is an arbitrary high-order, provably positivity-preserving DDG method for the Poisson–Nernst–Planck system.

Load-bearing premise

The proof assumes the positivity-preserving limiter applied at each step keeps the concentration polynomial pointwise positive and the time-step condition valid, an interaction analyzed only in one dimension.

Editorial extensions

If this is right

  • In one dimension, the modified flux plus forward Euler guarantees positive cell averages for any polynomial degree $k$, as long as the time step obeys (3.18).
  • When combined with the reconstruction limiter, the scheme yields pointwise positive concentrations, keeping the logarithmic term $\log c_i$ well-defined at every step.
  • On rectangular meshes in several dimensions, the same positivity propagation holds for tensor-product $Q^k$ elements under the multi-dimensional CFL condition (4.28).
  • Numerical tests show that the modified flux does not degrade the expected $k+1$ order of accuracy, and that mass conservation and free-energy dissipation are observed in the hybrid scheme.

Reading between the lines

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

  • Beyond the paper: the local, data-dependent choice of $\tilde{\beta}_i$ should transfer to other gradient-flow systems written as $\partial_t c = \nabla\cdot(c\nabla p)$ with $p=\delta E/\delta c$.
  • Beyond the paper: the multi-dimensional theorem certifies only the modified-flux update, so pointwise positivity in 2D and 3D still relies on the limiter interaction the paper does not formalize.
  • Beyond the paper: the numerical evidence for energy dissipation is not accompanied by a proof for the hybrid algorithm; proving dissipation for the modified flux would close the remaining structural gap.
  • Beyond the paper: the $1/d$ factor in the multi-dimensional CFL bound makes high-order SSP-RK stages cheaper but time-step heavy, so an implicit or locally varying variant is a natural practical follow-up.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. The manuscript proposes a hybrid direct discontinuous Galerkin (DDG) method for the Poisson-Nernst-Planck system that enforces positivity of ion concentrations. The main ingredients are: (i) a locally modified numerical flux, defined in (1.3) and (3.16)-(3.17), which makes the cell-average update nonnegative under a forward Euler discretization and a suitable CFL condition; (ii) a positivity-preserving reconstruction limiter, (3.19), applied at each step to obtain pointwise positivity; and (iii) an extension of the scheme to rectangular meshes in multiple dimensions, with a claimed cell-average positivity theorem (Theorem 4.1). Numerical experiments in 1D and 2D demonstrate high-order accuracy, positive cell averages, mass conservation, and free-energy dissipation, although the latter two properties are explicitly stated to be only numerically observed, not proven.

Significance. If the multi-dimensional positivity result were rigorously established, the paper would deliver a genuinely arbitrary-high-order, positivity-preserving DG scheme for PNP systems, which is a valuable contribution given the difficulty of maintaining positivity in high-order discretizations. The 1D argument is elegant and internally consistent: the modified flux is designed so that the Gauss-Lobatto decomposition of the cell average has nonnegative coefficients exactly when the CFL condition (3.18) holds. The numerical evidence for accuracy and positivity is strong, including a comparison showing failure of the unmodified scheme from [24]. However, the proof of the multi-dimensional extension contains a serious quadrature gap, and the hybrid limiter is not analyzed in multiple dimensions, so the central multi-D claim is not yet supported.

major comments (3)
  1. [§4.2, proof of Theorem 4.1] The proof is not complete as written. After deriving the update for the cell average, the authors apply the M-point Gauss-Lobatto quadrature rule in all directions except x^l to an integrand that contains the modified flux (4.25)-(4.26). Because beta-tilde_i^l = |\partial_{x_l} p_i^n| / c_i^n, the integrand involves the absolute value of a polynomial in the transverse variables and is not a polynomial. Gauss-Lobatto quadrature is therefore not exact for that integrand, and a positive quadrature sum does not imply that the exact cell average is positive. The stated CFL condition (4.28) controls |\partial_{x_l} p_i^n| only at the quadrature nodes, but a polynomial can attain a much larger maximum away from those nodes; for example, the transverse profile of a degree-3 element can peak at x=0.3 while the 3-point Gauss-Lobatto nodes are 0, 0.5, and 1. To conclude positivity of the exact integral, one would need the integrand to be nonnegative pointwise, which requires a pointwise (supremum) CFL bound rather than a node-only bound. Thus Theorem 4.1 is not established as stated. The one-dimensional Theorem 3.1 is unaffected by this issue.
  2. [Section 4 / Algorithm 1] The paper claims that the hybrid algorithm provides pointwise positivity in higher dimensions, but the positivity-preserving reconstruction (3.19) and Algorithm 1 are formulated for one-dimensional P^k elements. Section 4 introduces Q^k tensor-product elements and proves, even modulo the quadrature gap, only cell-average positivity for the modified flux; it does not define the reconstruction limiter on Q^k elements or analyze its effect on the multi-D update. Consequently, the pointwise-positive claim for the multi-D hybrid algorithm is not supported, even if the quadrature issue in Theorem 4.1 is repaired.
  3. [§3.3, Algorithm 1 and Remark 3.1] The algorithm as stated does not guarantee that the CFL hypothesis (3.18) holds at every step. Step 3 modifies c_i^n by the limiter, which changes p_i^n through (3.15b) and therefore changes \partial_x p_i^n; the practical time-step rule (3.21) contains an empirical factor gamma and is not proven to imply (3.18) for the reconstructed data. To claim that the hybrid algorithm preserves cell-average positivity, the authors should either prove a bound for the limiter's effect on \partial_x p_i^n, or explicitly state that (3.18) is verified after reconstruction and the time step is reduced when necessary.
minor comments (4)
  1. [Equations (3.18) and (4.28)] The notation |\partial_x p|^{-1} should be accompanied by the convention that this quantity is infinite when |\partial_x p| = 0; this is done implicitly in the proof of Theorem 3.1 for boundary edges but should be stated uniformly for all interior edges and for the multi-D condition (4.28).
  2. [Remark 3.2 and Section 6] Remark 3.2 states that discrete mass conservation and free-energy dissipation are not formally proven. Mass conservation for the modified scheme actually follows by summing (3.15a) over cells because the interface flux is single-valued; adding this one-line argument would strengthen the paper and sharpen the remark.
  3. [Table 5.5] For the case (m, gamma) = (3, 1/4), the reported convergence orders are non-monotone and the error increases from N=30 to N=40; the authors should comment on whether this is an artifact of the time-step rule or an indication that the CFL condition is not being enforced as required.
  4. [Labels in Theorem 2.1 and Lemma 3.1] The displayed labels in Theorem 2.1 and Lemma 3.1 contain corrupted text ('tieacce...lowercase'); these should be cleaned in the final version.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the positivity theorem is a constructive consequence of the deliberately modified flux, and the self-citations to prior work are not load-bearing.

full rationale

I find no significant circularity. The central result, Theorem 3.1, is a direct theorem about the fully discrete scheme (3.15), in which the modified flux parameter is deliberately defined by (3.17) as beta_i = |d_dx p_i^n| / {c_i^n}. The proof uses exactly this choice to rewrite the interface terms as (d_dx p_i^n + beta_i {c_i^n}) c_i and (-d_dx p_i^n + beta_i {c_i^n}) c_i, each of which is nonnegative by the inequality a + |a| >= 0. This is a constructive design choice rather than a fitted parameter later renamed as a prediction; no quantity asserted as an output is fed back as an input. The pointwise positivity of the hybrid algorithm is enforced by the limiter (3.19), while the theorem supplies the cell-average positivity needed to keep the limiter applicable; this is a standard induction on the scheme, not a tautology. The paper cites the authors' earlier framework [24] for the DDG formulation, Theorem 2.1, and Lemma 3.1, but these are independent mathematical statements with stated assumptions that do not include the positivity result, so they do not make the argument circular. I also flag two limitations that are correctness concerns rather than circularity: Remark 3.2 and the concluding section explicitly state that discrete energy dissipation is not formally proven, and the proof of Theorem 4.1 applies Gauss-Lobatto quadrature in the transverse directions to integrands involving beta_i^l c_i with beta_i^l = |d_xl p_i^n| / c_i^n, which is not a polynomial in the transverse variables; thus the multi-D cell-average positivity theorem is not established as written. These observations do not undermine the 1D Theorem 3.1 and are not circular steps.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The scheme introduces no new physical entities and no fitted physical constants. The hand-tuned numerical parameters (gamma, delta, beta0, beta1) are standard implementation choices and do not serve as predictions. The positivity proof relies on quadrature exactness, mesh regularity, and the accuracy-preserving reconstruction lemma carried over from [24].

free parameters (3)
  • gamma (CFL safety factor) = 1/4, 1/2, 1 in Table 5.5
    Equation (3.21) introduces an empirical factor gamma below the theoretical CFL bound; values are selected by numerical experiment, not derived.
  • delta (limiter threshold) = not reported; constrained delta < h^{k+1}
    Equation (3.19) uses a small threshold delta in the positivity limiter; the paper requires delta below the resolution scale but the numerical tests do not state the value.
  • DDG flux coefficients beta0 and beta1 = e.g., (2, -), (4, 1/20), (3, -), (9, 1/12), (19, 1/12), (16, 1/6)
    Chosen by hand from prior guidelines [25,36,37] in each convergence test; they set the baseline flux and are not fitted to target data.
assumptions (4)
  • standard math Gauss-Lobatto quadrature with M >= (k+3)/2 points is exact for the k-th degree polynomial cell average and has positive weights.
    Used in Theorem 3.1 and Theorem 4.1 to decompose cell averages into point values at nodes, Sections 3.1 and 4.2.
  • domain assumption The mesh is regular, h_max <= C h_min, and the DDG flux scales as O(h^{-1}).
    Remark 3.1 uses these to convert the CFL bound (3.18) into the practical O(h^2) time-step condition (3.20).
  • domain assumption The accuracy-preserving reconstruction lemma (Lemma 3.1, quoted from [24]) holds and the limiter preserves cell averages.
    The hybrid Algorithm 1 relies on this limiter to turn positive cell averages into pointwise positivity without losing formal accuracy; the lemma is cited, not reproved.
  • standard math SSP-Runge-Kutta time discretizations are positive linear combinations of forward Euler steps and therefore inherit the positivity bound.
    Remark 3.1.3 extends the forward Euler positivity result to high-order time stepping.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A positivity-preserving hybrid DDG method for Poisson--Nernst--Planck systems." pith.science (2026). https://pith.science/paper/MPHXYMFL

@misc{pith2026250524058,
  author       = {Pith},
  title        = {Pith review of: A positivity-preserving hybrid DDG method for Poisson--Nernst--Planck systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MPHXYMFL}},
  note         = {Machine review of arXiv:2505.24058}
}
read the original abstract

In earlier work [H. Liu and Z. Wang, J. Comput. Phys., 328(2017)], an arbitrary high-order conservative and energy-dissipative direct discontinuous Galerkin (DDG) scheme was developed. Although this scheme enforced solution positivity using cell averages as reference values, it lacked a theoretical guarantee for the positivity of those cell averages. In this study, we develop a novel arbitrary high-order DDG method with rigorously proven positivity-preserving properties. Specifically, the positivity of the cell averages is ensured through a modified numerical flux in combination with forward Euler time discretization. To achieve point-wise positivity of ion concentrations, we introduce a hybrid algorithm that integrates a positivity-preserving limiter. The proposed method is further extended to higher-dimensional problems with rectangular meshes. Numerical results confirm the scheme's high-order accuracy, guaranteed positivity preservation, and consistent discrete energy dissipation.

Figures

Figures reproduced from arXiv: 2505.24058 by the authors.

Figure 5.1
Figure 5.1. Numerical solution versus exact solution at T = 0.05 0 0.2 0.4 0.6 0.8 1 0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.2 0.4 0.6 0.8 1 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0 0.2 0.4 0.6 0.8 1 -7 -6 -5 -4 -3 -2 -1 0 10-3 h=0.025 h=0.05 h=0.1 Exact 5.2. 1D tests on solution properties. We test solution properties such as solution positivity, mass conservation and free energy dissipation. We apply the proposed scheme to the … view at source ↗
Figure 5.2
Figure 5.2. Smallest cell averages of c1, c2. system (1.1) with m = 2 in Ω = [0, 1], defined by the following equations: ∂tc1 = ∂x(∂xc1 + q1c1∂xψ), ∂tc2 = ∂x(∂xc2 + q2c2∂xψ), − ∂ 2 xψ = q1c1 + q2c2, c1(0, x) = c in 1 (x), c2(0, x) = c in 2 (x), ψ(t, 0) = 0, ∂xψ(t, 0) = 0, ∂xψ(t, 1) = 0, ∂xci + qici∂xψ = 0, x = 0, 1. In this system, the ion charges are q1 = 1 and q2 = −1, and the initial concentrations are given by c in 1 (x) = … view at source ↗
Figure 5
Figure 5. displays the evolution of the total masses of [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figures from the paper (8 more)
Figure 5.3
Figure 5.3. Figure 5.3: Conservation of mass and decay of free energy 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 t 0.15 0.155 0.16 0.165 0.17 0.175 mass -0.8 -0.6 -0.4 -0.2 0 0.2 energy total mass of c 1 total mass of c 2 free energy F [PITH_FULL_IMAGE:figures/full_fig_p015_5_3.png]
Figure 5.4
Figure 5.4. Figure 5.4: Numerical solution evolution 0 0.2 0.4 0.6 0.8 1 x 0 0.05 0.1 0.15 0.2 0.25 0.3 c 1 0 0.2 0.4 0.6 0.8 1 x 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 c 2 t=0 t=0.005 t=0.01 t=0.05 t=0.1 0 0.2 0.4 0.6 0.8 1 x -0.045 -0.04 -0.035 -0.03 -0.025 -0.02 -0.015 -0.01 -0.005 0 that par…
Figure 5.5
Figure 5.5. Figure 5.5: Smallest cell averages of c1, c2 [PITH_FULL_IMAGE:figures/full_fig_p016_5_5.png]
Figure 5.7
Figure 5.7. Figure 5.7: Numerical solution evolution 0 0.2 0.4 0.6 0.8 1 x 0 0.05 0.1 0.15 0.2 0.25 0.3 c 1 0 0.2 0.4 0.6 0.8 1 x 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 c 2 t=0 t=0.005 t=0.01 t=0.05 t=0.1 0 0.2 0.4 0.6 0.8 1 x -0.04 -0.035 -0.03 -0.025 -0.02 -0.015 -0.01 -0.005 0 5.3. 2D accurac…
Figure 5.8
Figure 5.8. Figure 5.8: Smallest cell averages of c1, c2. where ∂ΩD = {(x, y) ∈ Ω : ¯ x = 0, x = 1} and ∂ΩN = ∂Ω\∂ΩD, with ion charges q1 = 1 and q2 = −1. For this example, the DG schemes use P 2 polynomials with numerical flux parameters β0 = 16 and β1 = 1 6 . Following the approach in [25…
Figure 5
Figure 5. Figure 5: (a) and 5.8(b) show the evolution of the smallest cell averages of [PITH_FULL_IMAGE:figures/full_fig_p020_5.png]
Figure 5.9
Figure 5.9. Figure 5.9: Conservation of mass and decay of free energy 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 t 0.02 0.04 0.06 0.08 mass Conservation of mass and decay of free energy -0.3 -0.28 -0.26 -0.24 energy total mass of c 1 total mass of c 2 free energy F References [1] J.-D. Benamou…
Figure 5.10
Figure 5.10. Figure 5.10: The contours evolution of c1 − 1/30, c2 − 2/30 and ψ. [13] J. Hu and X. Huang. A fully discrete positivity-preserving and energy-dissipative finite difference scheme for Poisson￾Nernst-Planck equations. Numer. Math., 145:77–115, 2020. [14] J. Jerome. Analysis of Cha…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

39 extracted references · 39 canonical work pages

  1. [24]

    Liu and Z

    H. Liu and Z. Wang. A free energy satisfying discontinuous Galerkin method for one-dimensional Poisson-Nernst-Planck systems.J. Comput. Phys., 328:413–437, 2017

    work page 2017

  2. [25]

    H. Liu, Z. Wang, P. Yin, and H. Yu. Positivity-preserving third order DG schemes for Poisson-Nernst-Planck equations. J. Comput. Phys., 452:110777, 2022

    work page 2022

  3. [3]

    J. A. Carrillo, H. Liu, and H. Yu. Positivity-preserving and energy-dissipating discontinuous Galerkin methods for non- linear nonlocal Fokker-Planck equations.Commun. Appl. Ind. Math., 2025

    work page 2025

  4. [1]

    Benamou and Y

    J.-D. Benamou and Y. Brenier. A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem. Numer. Math., 84(3):375–393, 2000

    work page 2000

  5. [2]

    J. A. Carrillo, A. Chertock, and Y. H. Huang. A finite-volume method for nonlinear nonlocal equations with a gradient flow structure.Commun. Comput. Phys., 17(1):233–258, 2015

    work page 2015

  6. [4]

    H. Chen, H. Liu, and X. Xu. The Onsager principle and structure preserving numerical schemes.J. Comput. Phys., 523:113679, 2025

    work page 2025

  7. [5]

    Datta.Electronic Transport in Mesoscopic Systems

    S. Datta.Electronic Transport in Mesoscopic Systems. Cambridge University Press, 1997

    work page 1997

  8. [6]

    J. Ding, Z. Wang, and S. Zhou. Positivity preserving finite difference methods for Poisson-Nernst-Planck equations with steric interactions: application to slit-shaped nanopore conductance.J. Comput. Phys., 397:108864, 2019

    work page 2019

Show all 39 references
  1. [7]

    J. Ding, Z. Wang, and S. Zhou. Structure-preserving and efficient numerical methods for ion transport.J. Comput. Phys., 418:109597, 2020

    2020

  2. [8]

    Eisenberg and W

    B. Eisenberg and W. Liu. Poisson-Nernst-Planck systems for ion channels with permanent charges.SIAM J. Math. Anal., 38:1932–1966, 2007

    1932

  3. [9]

    Glasstone.An introduction to Electrochemstry

    S. Glasstone.An introduction to Electrochemstry. Van Nostrand Company, Inc., Princeton, NJ, 1942

    1942

  4. [10]

    Gottlieb, C

    S. Gottlieb, C. Shu, and E. Tadmor. Strong stability-preserving high-order time discretization methods.SIAM Rev., 43(1):89–112, 2001

    2001

  5. [11]

    D. He, K. Pan, and X. Yue. A positivity preserving and free energy dissipative difference scheme for the Poisson-Nernst- Planck system.J. Sci. Comput., 81:436–458, 2019

    2019

  6. [12]

    Hille.Ion channels of excitable membranes, volume 507

    B. Hille.Ion channels of excitable membranes, volume 507. Sinauer Sunderland, MA, 2001. 21 Figure 5.10.The contours evolution ofc 1 −1/30,c 2 −2/30 andψ

    2001

  7. [13]

    Hu and X

    J. Hu and X. Huang. A fully discrete positivity-preserving and energy-dissipative finite difference scheme for Poisson- Nernst-Planck equations.Numer. Math., 145:77–115, 2020

    2020

  8. [14]

    Jerome.Analysis of Charge Transport: A Mathematical Study of Semiconductor Devices

    J. Jerome.Analysis of Charge Transport: A Mathematical Study of Semiconductor Devices. Springer, Berlin, 1996

    1996

  9. [15]

    Jordan, D

    R. Jordan, D. Kinderlehrer, and F. Otto. The variational formulation of the Fokker–Planck equation.SIAM J. Math. Anal., 29(1):1–17, 1998

    1998

  10. [16]

    Kinderlehrer, L

    D. Kinderlehrer, L. Monsaingeon, and X. Xu. A wasserstein gradient flow approach to Poisson-Nernst-Planck equations. ESAIM Control Optim. Calc. Var., 23(1):137–164, 2017

    2017

  11. [17]

    Li.Electrokinetics in Microfluidics

    D. Li.Electrokinetics in Microfluidics. Academic Press, 2004

    2004

  12. [18]

    C. Liu, C. Wang, S. M. Wise, X. Yue, and S. Zhou. A positivity-preserving, energy stable and convergent numerical scheme for the Poisson-Nernst-Planck system.Math. Comp., 90:2071–2106, 2021

    2021

  13. [19]

    C. Liu, C. Wang, S. M. Wise, X. Yue, and S. Zhou. A second order accurate, positivity preserving numerical method for the Poisson-Nernst-Planck system and its convergence analysis.J. Sci. Comput., 97(23), 2023

    2023

  14. [20]

    H. Liu. Analysis of direct discontinuous Galerkin methods for multi-dimensional convection–diffusion equations.Numer. Math., 147(4):839–867, 2021

    2021

  15. [21]

    Liu and W

    H. Liu and W. Maimaitiyiming. Unconditional positivity-preserving and energy stable schemes for a reduced Poisson- Nernst-Planck system.Commun. Comput. Phys., 27:1505–1529, 2020

    2020

  16. [22]

    Liu and W

    H. Liu and W. Maimaitiyiming. Efficient, positive, and energy stable schemes for multi-D Poisson-Nernst-Planck systems. J. Sci. Comput., 87:92, 2021

    2021

  17. [23]

    Liu and W

    H. Liu and W. Maimaitiyiming. A dynamic mass transport method for Poisson-Nernst-Planck equations.J. Comput. Phys, 473:111699, 2023. 22

    2023

  18. [26]

    Liu and J

    H. Liu and J. Yan. The direct discontinuous Galerkin (DDG) methods for diffusion problems.SIAM J. Numer. Anal, 47:675–698, 2009

    2009

  19. [27]

    Liu and J

    H. Liu and J. Yan. The direct discontinuous Galerkin (DDG) method for diffusion with interface corrections.Commun. Comput. Phys., 8(3):541–564, 2010

    2010

  20. [28]

    Liu and H

    H. Liu and H. Yu. The entropy satisfying discontinuous Galerkin method for Fokker–Planck equations.J. Sci. Comput., 62:803–830, 2015

    2015

  21. [29]

    Markowich.The Stationary Semiconductor Device Equations

    P. Markowich.The Stationary Semiconductor Device Equations. Springer-Verlag, New York, 1986

    1986

  22. [30]

    P. A. Markowich, C. A. Ringhofer, and C. Schmeiser.Semiconductor Equations. Springer-Verlag Inc, New York, 1990

    1990

  23. [31]

    Metti, J

    M. Metti, J. Xu, and C. Liu. Energetically stable discretizations for charge transport and electrokinetic models.J. Comput. Phys., 306:1–18, 2016

    2016

  24. [32]

    M. S. Mock.Analysis of Mathematical Models of Semiconductor Devices, volume 3. Boole Press, 1983

    1983

  25. [33]

    Shen and J

    J. Shen and J. Xu. Unconditionally positivity preserving and energy dissipative schemes for Poisson–Nernst–Planck equations.Numer. Math., 148:671–697, 2021

    2021

  26. [34]

    Su and H

    S. Su and H. Tang. A positivity-preserving and free energy dissipative hybrid scheme for the poisson-nernst-planck equations on polygonal and polyhedral meshes.Comput. Math. Appl., 108:33–48, 2022

    2022

  27. [35]

    Z. Sun, J. A. Carrillo, and C.-W. Shu. A discontinuous galerkin method for nonlinear parabolic equations and gradient flow problems with interaction potentials.J. Comp. Phys., 352:76–104, 2018

    2018

  28. [36]

    P. Yin, Y. Huang, and H. Liu. An iterative discontinuous Galerkin method for solving the nonlinear Poisson-Boltzmann equation.Commun. Comput. Phys., 16(2):491–515, 2014

    2014

  29. [37]

    P. Yin, Y. Huang, and H. Liu. Error estimates for the iterative discontinuous Galerkin method to the nonlinear Poisson- Boltzmann equation.Commun. Comput. Phys., 23(1):168–197, 2018

    2018

  30. [38]

    X. Zhang. On positivity-preserving high order discontinuous galerkin schemes for compressible Navier-Stokes equations. J. Comput. Phys., 328:301–343, 2017

    2017

  31. [39]

    Zhang and C.-W

    X. Zhang and C.-W. Shu. On maximum-principle-satisfying high order schemes for scalar conservation laws.J. Comput. Phys., 229:3091–3120, 2010. † Mathematics Department, Iowa State University, Ames, IA 50011, USA Email address:hliu@iastate.edu ‡ Department of Mathematics and St...

    2010

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.