Finite slab first passage statistics of Henyey Greenstein scattering

Technical Summary: Finite-slab first-passage statistics of Henyey–Greenstein scattering

Problem Statement
The paper addresses the classical radiative transport problem of light reflection, transmission, and absorption by a finite, plane-parallel scattering slab. The medium is characterized by anisotropic scattering described by the Henyey–Greenstein (HG) phase function, where step lengths between collisions follow an exponential distribution. The central objective is to express macroscopic slab properties—reflectance ($R$), transmittance ($T$), absorptance ($A$), and emergent angular distributions—strictly in terms of the first-passage statistics of the underlying random walk. The authors frame the problem by viewing reflection and transmission as first-passage events where a photon enters the slab, performs a persistent random walk, and eventually escapes through one of the two boundaries or is absorbed.

Methodology
The authors employ two independent, complementary approaches to solve the problem, demonstrating their agreement within numerical and statistical uncertainties:

1. Radiative-Transfer (RT) Approach:

   • This is a deterministic method that treats the evolution of angular intensity within the slab.
   • By averaging the HG phase function over the azimuthal angle, the problem is reduced to a Markov process defined by the state pair $(z, \mu)$, where $z$ is the depth and $\mu$ is the direction cosine.
   • The slab is discretized into thin sub-layers of thickness $\Delta\tau$. Scattering within each layer is treated to first order.
   • A transfer operator is constructed using a composition rule (adding-doubling method) that re-sums the infinite series of back-and-forth reflections between layers.
   • Absorption is introduced via a survival probability $a$ per collision, allowing the decomposition of reflectance and transmittance into scattering-order expansions. This approach explicitly links the finite-slab geometry to the first-return statistics of a semi-infinite medium.

2. Monte Carlo (MC) Approach:

   • This method generates extremely long random walks in 3D space independent of any specific boundaries.
   • Step lengths are sampled from an exponential distribution, and scattering angles are sampled using the HG phase function.
   • The method relies on the memoryless property of the exponential distribution: the portion of a step crossing a boundary retains the same statistical distribution as the full step.
   • By intersecting these long walks with a collection of overlapping slabs (objects) of varying thicknesses and positions, the algorithm generates an ensemble of "excursions" (segments of the walk contained within a slab).
   • Statistics such as first-passage probabilities, scattering orders, and angular crossing laws are extracted directly from this database of excursions. Absorption is applied post-hoc by weighting excursions of order $n$ by $a^n$.

Key Contributions and Results

• First-Passage Formulation: The paper establishes that $R$, $T$, $A$, and angular distributions are direct consequences of first-passage statistics. Specifically, the reflectance is expressed as a sum over scattering orders involving the first-return probability in a semi-infinite medium ($P_\infty(n)$) and a depth-survival factor ($S(n, \tau)$) representing the probability that an $n$-step trajectory remains below the upper boundary.
• Conservative Limit and Universal Tails: In the conservative limit (no absorption, $a \to 1$), the paper analyzes the approach of reflectance to unity. It demonstrates that the deficit $1 - R(a)$ follows a square-root edge behavior ($\propto \sqrt{1-a}$) with a non-analytic cusp at $a=1$. This behavior is governed by the universal $n^{-3/2}$ tail of the first-return law for random walks (Sparre–Andersen exponent).
• Thick-Slab Recovery: As slab thickness $\tau \to \infty$, the order-resolved reflection probability $P_R(n, \tau)$ converges term-by-term to the semi-infinite first-return probability $P_\infty(n)$, recovering the universal tail.
• Angular Distributions: The paper derives the exit-angle distributions for reflection and transmission, showing they are weighted by the equilibrium crossing law $p(\mu) = 2|\mu|$, independent of slab thickness and asymmetry parameter $g$.
• Validation: The two methods agree to within $\le 1.6 \times 10^{-3}$ across a wide range of parameters ($g \in [0, 0.95]$, $\tau \in [0.5, 16]$, $a \in [0.5, 1]$). The RT operator is further validated against independent deterministic solvers (adding-doubling and successive-orders) and the Chandrasekhar H-function, with mutual agreement better than $1.1 \times 10^{-4}$.

Significance and Claims
The paper claims that the primary significance of this work lies in the explicit connecting identities between the stationary random walk and the slab transport operator. By separating the geometry of the first passage from the absorption mechanism, the authors provide a unified framework where:

1. The reciprocity between entry and exit statistics is a direct consequence of the reversibility of the HG walk.
2. The specific form of the absorption edge ($\sqrt{1-a}$) is fixed solely by the $n^{-3/2}$ first-return tail of the random walk.
3. The "depth-survival factor" $S(n, \tau)$ encapsulates all thickness dependence, linking the finite slab problem to Brownian excursion statistics in the diffusive regime.

The authors note that while the current work focuses on the $m=0$ (azimuthally averaged) mode, the framework allows for the extension to higher azimuthal modes to obtain full bidirectional distributions. They also identify open directions, including finding a closed form for the leading eigenvalue of the confined operator and addressing the strongly forward-peaked regime ($g \to 1$) relevant to tissue and clouds. The work does not propose new experimental setups but rather offers a rigorous statistical and computational foundation for existing radiative transfer problems.