Lorentz Mirror Model Overview

• The Lorentz Mirror Model is a stochastic lattice model on Z² where light rays reflect deterministically off randomly oriented mirrors based on local rules.
• It quantifies escape probabilities, open-path densities, and finite-scale scaling behavior, providing insight into periodic or percolative trajectories.
• The model’s framework also influences studies in relativistic optics and Lorentz-violating electrodynamics, offering practical analogues for reflection and transport phenomena.

Searching arXiv for recent and canonical papers on the Lorentz Mirror Model and closely related usages of the term. The expression Lorentz Mirror Model is used for several distinct constructions rather than a single universally fixed model. In mathematical physics it usually denotes a random lattice model in which a deterministic light ray moves through a quenched configuration of local mirrors on $\mathbb Z^2$ or on finite-width cylinders; in relativistic optics it can denote reflection from a uniformly moving flat mirror; and in boundary quantum field theory it can denote mirror problems in Lorentz-violating electrodynamics or scalar field theory. These usages share a common theme of specular reflection or edge matching, but they differ in geometry, observables, and physical interpretation: escape probabilities, densities of open paths, localization lengths, conductance, relativistic reflection laws, and charge–mirror interaction energies all appear in different parts of the literature (Kozma et al., 2013, Gjurchinovski, 2021, Borges et al., 2022).

1. Lattice model: definition, state space, and observables

In the standard probabilistic usage, the Lorentz mirror model is defined on the square lattice $\mathbb Z^2$. Each vertex $x\in\mathbb Z^2$ is independently designated a double-sided mirror with probability $p$, and otherwise carries no mirror. Conditional on being present, the mirror is a north-west mirror with probability $1/2$ and a north-east mirror with probability $1/2$. A ray of light travels along lattice edges; at a vertex with no mirror it continues in the same direction, while at a mirrored vertex it is reflected according to the mirror orientation. Once the environment and an initial directed edge are fixed, the dynamics are deterministic (Kozma et al., 2013).

Finite-volume formulations introduce additional observables. In an $L\times L$ box there are $L^2$ vertices and $2L^2$ internal lattice edges. Every edge belongs to exactly one trajectory, and trajectories are classified as closed paths, which are periodic orbits entirely contained in the box, or open paths, which escape the box. Writing $C_L(p)$ for the mean total number of edges contained in all closed paths and $\mathbb Z^2$0 for the mean total number of edges contained in all open paths, one has

$\mathbb Z^2$1

The density of open-path edges is therefore $\mathbb Z^2$2, while the density of closed-path edges is $\mathbb Z^2$3 (Kraemer et al., 2014).

A basic large-scale observable is the escape event from the square

$\mathbb Z^2$4

For $\mathbb Z^2$5, the notation $\mathbb Z^2$6 denotes the event that the ray of light starting at the origin reaches $\mathbb Z^2$7. The central qualitative question is whether trajectories are all periodic or whether, for some $\mathbb Z^2$8, there is positive probability that the light travels to infinity. This question is distinct from other uses of the word “Lorentz,” such as the Lorentz model of hyperbolic geometry, which does not involve mirrors (Kozma et al., 2013, Nickel et al., 2018).

2. Escape, open-path density, and planar asymptotics

A sharp finite-scale lower bound was proved for the planar lattice model: for every mirror density $\mathbb Z^2$9,

$x\in\mathbb Z^2$0

Equivalently, the probability that the trajectory from the origin leaves the square $x\in\mathbb Z^2$1 is at least $x\in\mathbb Z^2$2. The proof uses an odd-width cylinder $x\in\mathbb Z^2$3, a parity argument showing that finite trajectories cross a fixed cut an even number of times, and rotational symmetry on the cylinder. At $x\in\mathbb Z^2$4, the infinite-trajectory question is equivalent to existence of an infinite open path in bond percolation on $x\in\mathbb Z^2$5 at parameter $x\in\mathbb Z^2$6; by Harris’s theorem there is no such infinite path. The note emphasizes that “No similar result is known for $x\in\mathbb Z^2$7” (Kozma et al., 2013).

A complementary finite-volume result concerns the density of open trajectories. For every mirror density $x\in\mathbb Z^2$8,

$x\in\mathbb Z^2$9

equivalently,

$p$0

This rules out a phase with a positive-density set of open trajectories, although it is weaker than the conjecture that all infinite-volume trajectories are closed almost surely. The proof compares general $p$1 configurations with the fully occupied case $p$2 via a mirror-removal algorithm, and it uses numerical scaling input at $p$3: $p$4

$p$5

together with $p$6 with $p$7. The inequality $p$8 is the decisive input behind the vanishing-density conclusion (Kraemer et al., 2014).

These planar results sit alongside earlier work cited in the same literature. For $p$9, Grimmett and Bunimovich–Troubetzkoy imply that all trajectories are closed almost surely in the infinite model; Quas showed that if open orbits exist, then either there is only one open orbit or infinitely many; and Kozma–Sidoravicius supplied a rigorous lower bound on finite-scale escape. The resulting picture is that finite-scale delocalization can be quantified even where infinite-volume classification remains incomplete (Kraemer et al., 2014).

3. Cylinders, low-density scaling, and higher-dimensional transport

On finite-width cylinders the model admits sharper quantitative control. For the Lorentz mirror model on the even cylinder $1/2$0, with $1/2$1 even and $1/2$2, the maximal height reached by a path started from the first street is of order $1/2$3. More precisely, if $1/2$4 is the smallest street index at which no path from the first street remains connected upward, then for all $1/2$5,

$1/2$6

and, for any $1/2$7,

$1/2$8

The proof encodes row-to-row connectivity in the Brauer algebra, so that the stopping variable becomes the first time the diagram accumulates $1/2$9 bars and no north–south paths remain (Ryan, 2020).

A different cylinder result gives a polynomial localization length. On

$1/2$0

for large enough $1/2$1, with high probability, any trajectory intersecting the section $1/2$2 is contained in a region $1/2$3. In explicit form, there exists $1/2$4 such that for $1/2$5 and $1/2$6,

$1/2$7

where $1/2$8 is the event that every trajectory hitting $1/2$9 stays in $L\times L$0. The exponent $L\times L$1 is not claimed to be optimal; the paper explicitly remarks that it “may not be optimal” (Li, 2020).

Recent high-dimensional work changes the transport picture. For the Lorentz mirror walk on $L\times L$2 with $L\times L$3 and small mirror density $L\times L$4, trajectories behave diffusively at all polynomial time scales $L\times L$5 with $L\times L$6. The main theorem states that for $L\times L$7,

$L\times L$8

and in particular

$L\times L$9

Here $L^2$0 is the first closure time. This does not prove infinite-time delocalization, but it rules out closure up to very long times and gives strong evidence for the expected high-dimensional regime (Elboim et al., 2 May 2025).

A further development introduces a hierarchical Lorentz mirror model with exact recursion for the conductance distribution. With

$L^2$1

the mean conductance $L^2$2 satisfies, for $L^2$3,

$L^2$4

so the mean conductance scales as cross-section over length at all scales. The same work proposes a universal fluctuation law,

$L^2$5

supported by numerics in the hierarchical model and by numerical evidence in the original Lorentz mirror model in $L^2$6 (Lefevere et al., 8 Feb 2026).

4. Relation to the Lorentz gas and transport geometry

The lattice Lorentz mirror model is related to, but not identical with, the continuous Lorentz gas. In the Lorentz gas a point particle moves freely between collisions with scatterers $L^2$7, obeying the specular reflection law

$L^2$8

The displacement observable is

$L^2$9

and diffusion is encoded by

$2L^2$0

Dettmann’s survey emphasizes that smooth dispersing finite-horizon geometries yield normal diffusion, infinite corridors yield superdiffusion with $2L^2$1, and polygonal or flat mirror-like geometries lead to more delicate behavior, including recurrence without ergodicity and anomalous scaling laws (Dettmann, 2014).

This comparison is useful because the discrete Lorentz mirror model inherits the language of mirror reflection but not the full hyperbolic structure of smooth Sinai billiards. In the continuous setting, geometry controls transport through finite versus infinite horizon, curvature versus flat points, and periodic versus random arrangements. In the discrete model, the corresponding issues appear as escape to large boxes, open-path density, cylinder localization, and conductance. The two literatures therefore share reflection as a local rule while differing in state space, coarse-graining, and the mechanisms behind macroscopic transport (Dettmann, 2014).

5. Special-relativistic moving mirrors

A completely different use of the term concerns a single flat mirror in uniform translational motion. For a plane-polarized light beam incident on an inclined flat mirror moving with speed $2L^2$2 along the positive $2L^2$3-axis, the mirror making angle $2L^2$4 relative to the negative $2L^2$5-direction, a direct Huygens–Fresnel construction yields the reflection law

$2L^2$6

or equivalently

$2L^2$7

Here $2L^2$8 and $2L^2$9 are the incident and reflected angles relative to the mirror normal. For $C_L(p)$0, the usual static law $C_L(p)$1 is recovered, but for a moving mirror one generally has $C_L(p)$2, and reversibility fails: sending the reflected ray back as incident does not in general recover the original angle. The same paper gives an explicit formula for $C_L(p)$3 and treats Einstein’s classic case of a vertical mirror moving normal to itself as a special case (Gjurchinovski, 2021).

The most distinctive claim of that work is a thought experiment showing that the altered reflection law, combined with Einstein’s postulates, implies the usual Lorentz contraction of the mirror along the direction of motion. In the “Einstein’s cat experiment,” the moving-frame geometry leads to

$C_L(p)$4

and, with $C_L(p)$5,

$C_L(p)$6

The paper presents this as evidence that Lorentz contraction is a physical property of moving matter rather than merely a coordinate artifact; whether one agrees philosophically, that is the paper’s explicit interpretive claim (Gjurchinovski, 2021).

A related electrodynamic formulation replaces the ideal flat reflector by a mobile slab. In that model a slab of finite rest thickness, linear, isotropic, non-magnetizable, ohmic, and with zero free charge density when at rest, interacts self-consistently with the electromagnetic field. Using instantaneous Lorentz transformations, the analysis yields a wave equation with damping and slowly varying coefficients, plus motion-induced mixed-derivative and acceleration terms, and the slab equation of motion acquires both a position- and time-dependent effective mass and a velocity-dependent force related to friction or cooling (Castaños et al., 2014).

These relativistic moving-mirror constructions are not the disordered lattice Lorentz mirror model. They concern a single mirror or slab, deterministic specular reflection, and special relativity, rather than random mirrors on $C_L(p)$7 or conductance problems on cylinders (Gjurchinovski, 2021, Castaños et al., 2014).

6. Lorentz-violating mirror electrodynamics and scalar analogues

In Standard-Model Extension boundary problems, “Lorentz mirror” refers to mirrors in Lorentz-violating field theories. For a real scalar field with an aether-like Lorentz-violating kinetic term and a planar $C_L(p)$8-function mirror, the propagator becomes anisotropic through the denominator

$C_L(p)$9

and the source–mirror interaction depends on the decomposition $\mathbb Z^2$00 relative to the plane. In the Dirichlet limit the image method remains valid, which the paper describes as nontrivial, and a torque appears: $\mathbb Z^2$01 with $\mathbb Z^2$02 the angle between the mirror normal and the spatial LV vector. This torque has no counterpart in Lorentz-symmetric mirror systems (Borges et al., 2018).

The CPT-even nonbirefringent SME photon sector near a semi-transparent mirror exhibits the same qualitative structure. In that setting the mirror is represented by a localized boundary term at $\mathbb Z^2$03, controlled by a transparency parameter $\mathbb Z^2$04, and the full propagator is decomposed as a free-space LV propagator plus a mirror correction. For a point charge near the plane, the interaction energy is modified, and a spontaneous torque again emerges. In the perfect-mirror limit the interaction found via the image method is recovered (Borges et al., 2022).

A closely related CPT-even model with sources and mirrors studies a planar semi-transparent defect coupled to the pure-photon SME sector specialized to a single background vector $\mathbb Z^2$05. The charge–mirror interaction energy takes the form

$\mathbb Z^2$06

and the torque is

$\mathbb Z^2$07

The perfect-mirror limit is $\mathbb Z^2$08, finite $\mathbb Z^2$09 gives a semi-transparent mirror, and the new effect is explicitly orientation dependent (Ferrari, 2023).

In the CPT-odd Carroll–Field–Jackiw sector with a perfectly conducting plate, the full propagator is again written as a free-space part plus a mirror correction,

$\mathbb Z^2$10

For a static point charge at distance $\mathbb Z^2$11 from the plate, the interaction energy is

$\mathbb Z^2$12

and the torque is

$\mathbb Z^2$13

A notable distinction from the CPT-even case is that the image method is valid only when the background vector has solely the component perpendicular to the mirror (Soares et al., 2024).

Across these Lorentz-violating mirror theories, the recurring feature is that a fixed background vector or tensor makes the interaction between source and mirror depend on orientation. Static charge–mirror systems can therefore develop torques that ordinary Maxwell electrodynamics forbids (Borges et al., 2018, Borges et al., 2022, Ferrari, 2023, Soares et al., 2024).

7. Lorentz-invariant moving mirrors and Hawking analogues

A further, again distinct, construction is the Lorentz-invariant moving-mirror approach to Hawking radiation. Here the central invariant is the mirror’s proper acceleration $\mathbb Z^2$14, and the proposed power law is

$\mathbb Z^2$15

The mirror is asymptotically static in the past and future but has a long metastable plateau of nearly constant proper acceleration,

$\mathbb Z^2$16

with $\mathbb Z^2$17 a positive even integer. The paper identifies the plateau acceleration with black-hole surface gravity,

$\mathbb Z^2$18

so that the mirror power and Hawking power share the same quadratic invariant structure (Good et al., 2021).

This is not presented as an exact microscopic equivalence to gravity, but as an asymptotic or thermodynamic analogue. The model is designed to be unitary because the mirror is asymptotically static and the entanglement entropy,

$\mathbb Z^2$19

returns to zero as $\mathbb Z^2$20. With the additional identification $\mathbb Z^2$21, the lifetime parameter scales as

$\mathbb Z^2$22

reproducing the black-hole evaporation timescale. The paper emphasizes that the invariant “quantum Larmor power” is the primary observable, not the stress-tensor flux alone (Good et al., 2021).

This Hawking-analogue mirror model shares neither the quenched-lattice structure of the probabilistic Lorentz mirror model nor the SME boundary terms of Lorentz-violating electrodynamics. Its use of “Lorentz” refers instead to the Lorentz-scalar nature of proper acceleration and the corresponding invariant power law (Good et al., 2021).