An arXiv preprint reports that a carefully selected liquid–vapor phase-transition state can remain stable in a one-dimensional model of a compressible van der Waals fluid. When the model starts with sufficiently small disturbances that meet the authors’ restrictions, its solution exists for all later times and converges uniformly to the chosen steady state.
The result concerns a double-interface arrangement in which liquid-like and vapor-like regions meet at phase boundaries. The specific volume—the space occupied per unit mass—jumps across each boundary, while velocity and pressure remain continuous. The volume jump is non-small and does not change with time.
The document is identified as the preprint arXiv:2608.20060v1, dated 20 August 2026. Its evidence is a mathematical theorem about weak solutions and their long-term behavior, rather than a measurement from an experiment or an observed fluid.
A return to a selected state
The paper studies a 2L-periodic boundary-value problem for the isothermal compressible Navier–Stokes equations of a van der Waals fluid, written in Lagrangian coordinates. It asks whether an admissible steady state associated with liquid–vapor phase transition remains asymptotically stable after a small permitted change in the initial profile.
The target state is selected when the average initial specific volume, written by the authors as v̄, lies inside the Maxwell region, v̄ ∈ (α0, β0). In the paper’s construction, that region is where the phase-transition steady state is admissible. The authors interpret their theorem as nonlinear stability of the double-interface solution under sufficiently small initial perturbations.
That conclusion does not apply to arbitrary starting conditions. The global theorem assumes that the initial specific volume satisfies v0 ∈ [b + 2v, v̂ − v] almost everywhere, that regional mean perturbations vanish for i = 1, 2, 3, and that the initial size measure obeys C0 ≤ δ0. These conditions define the narrow class of disturbances covered by the proof.
Within those assumptions, the theorem gives the global estimate A(t) + B(t) + F(t) ≤ CC0. The three terms are analytic quantities used to control the evolving solution; the bound links them to the size of the initial perturbation. The result is therefore a conditional stability statement, not a claim that every disturbance remains controlled.
The analysis also tracks the model’s phase regions. Specific volume remains in the prescribed interval [b + v, v̂] for 0 ≤ t ≤ τ. For t ≥ τ, it lies in the neighborhood Dα on I1 ∪ I3 and in Dβ on I2, as required by the theorem’s phase-region description.
Keeping a sharp boundary
For local existence, the authors construct a semi-discrete staggered-grid difference scheme and derive uniform time-weighted energy estimates. The estimates are designed to handle the initial singularity while the scheme is used to study a solution that may contain discontinuities at the phase boundaries.
The local weak solution has one-sided limits for the specific volume and the spatial derivative of velocity at x = ±l, and it satisfies the required jump condition. This lets the proof preserve the interface compatibility while following the solution on either side of the boundary.
The main theorem then establishes a global weak solution for the periodic problem. It also states that the solution converges to the admissible steady state in the uniform L∞ norm as t tends to infinity. Here, uniform L∞ convergence means that the largest pointwise difference between the evolving solution and the target state tends to zero across the modeled domain.
A result with a narrow reach
The setting is limited to one spatial dimension, periodic boundaries and an isothermal model. The supplied analysis does not establish that the same conclusion holds for multidimensional, nonperiodic or non-isothermal flows. The initial-data restrictions also mean that stability for arbitrary disturbances has not been shown.
The detailed argument focuses on the two-smooth-interface configuration with N = 1, although the paper mentions configurations with 2N jumps. Those broader arrangements are not developed in the main theorem with the same level of detail.
The proof’s techniques do not identify the precise locations of the phase interfaces. They establish the structure of the selected state and its convergence properties, but not a general rule for locating the boundaries within the periodic domain.
The authors also state that when the average initial specific volume lies outside the Maxwell region, the solution remains in a single-phase state. This is presented as an additional remark, while the central theorem addresses the admissible phase-transition case inside that region.
The preprint reports no empirical, experimental or physical-fluid validation of the theorem. Its evidence consists of mathematical existence, regularity, energy and convergence statements for the specified model. Whether the argument can be extended to the broader settings mentioned by the authors, and whether uniqueness of the global weak solution can be established, remain open questions in the supplied analysis.
Paper data and sources
Original title: Stability of admissible solutions for coexisting phase transitions for one-dimensional compressible van der Waals fluids
Authors: Yazhou Chen, Qiaolin He, Dongjuan Niu et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text