BeamDiffusion in Disordered Media

• BeamDiffusion is the study of diffusive light propagation in disordered media, contrasting with classical diffraction governed by the VCZ theorem.
• Experimental setups using EIT in atomic vapors and nanostructured photonic waveguides validate controlled beam shaping, coherence evolution, and diffusive scaling.
• The mathematical framework extends the Van Cittert–Zernike theorem to a diffusion regime, enabling new methods in imaging, wave manipulation, and engineered energy confinement.

DiscDiff, short for "diffusive beam propagation" or "BeamDiffusion," refers to the theoretical and experimental study of light field evolution in disordered or diffusive media, where spatial correlations in an optical beam propagate according to diffusion physics rather than wave diffraction. Unlike ordinary diffraction, which is governed by the paraxial Helmholtz equation and leads to the canonical Van Cittert–Zernike (VCZ) theorem, DiscDiff considers the scenario where a partially coherent optical field undergoes spatial transport due to genuine diffusion—typically realized within a medium of mobile scatterers or via engineered dissipative photonic structures. DiscDiff encompasses a spectrum of applications, from coherent diffusion experiments in atomic vapors to imaging and manipulation of light in disordered photonic waveguides, and intersects with fields such as mesoscopic wave physics and non-classical light transport (Chriki et al., 2018, Yamilov et al., 2013).

1. Physical Systems and Experimental Realizations

DiscDiff is physically realized in platforms where the spatial evolution of optical fields is governed by a true diffusion process. The most extensively studied system is a warm vapor of rubidium-87 ($^{87}$Rb) under conditions of electromagnetically induced transparency (EIT). Here, two strong "control" fields and a weak "probe" field interact with the D$_1$ transition line, forming a four-wave mixing (FWM) geometry (Chriki et al., 2018):

• The probe, imprinted with a spatial speckle pattern (created by a spatial light modulator), couples to the ground-state Zeeman coherence of the vapor.
• The atomic spin coherence, characterized by a diffusion coefficient ($D \approx 9.7 \pm 0.5$ cm²/s), spreads during a variable EIT group delay $\tau$, controlled by two-photon detuning.
• The retrieved "signal" field, generated in the orthogonal direction via FWM, inherits the spatial statistics of the diffused atomic coherence and is imaged on a camera for further analysis.

In disordered photonic waveguides (Yamilov et al., 2013), Disorder and absorption are engineered using nanostructured silicon membranes, with photonic bandgap confinement and randomly positioned air holes. The out-of-plane scattering introduces effective dissipation, and complex spatially-resolved diffusion is measured using near-field and far-field imaging techniques.

2. Mathematical Framework and Governing Equations

The dynamics of DiscDiff are described by the diffusion equation for a complex field $\phi(\mathbf{r}, t)$:

$$\frac{\partial \phi(\mathbf{r}, t)}{\partial t} = D \nabla_{r}^2 \phi(\mathbf{r}, t)$$

The mutual intensity (mutual coherence) is modeled as:

$$W(\mathbf{r}_1, \mathbf{r}_2; t) = \langle \phi^*(\mathbf{r}_1, t)\phi(\mathbf{r}_2, t) \rangle$$

$$\frac{\partial W}{\partial t} = D (\nabla_{1}^2 + \nabla_{2}^2)W$$

Introducing center ($\Bar{\mathbf{r}}$) and difference ($\Delta\mathbf{r}$) coordinates, this becomes:

$_1$0

In transverse Fourier space, the mutual coherence evolves as:

$_1$1

This separability of the propagator underlies the key distinction from diffraction, where center and difference variables become coupled via phase factors.

3. Diffusive Van Cittert–Zernike Theorem and Coherence Evolution

DiscDiff generalizes the VCZ theorem to diffusion. For a quasi-homogeneous source with intensity envelope $_1$2 (scale $_1$3) and initial coherence $_1$4 (scale $_1$5):

• Under diffusion, the mutual intensity propagates as two independent diffusions:

$_1$6

For a Gaussian initial distribution, explicit formulas follow:

• Coherence radius: $_1$7
• Beam envelope expansion: $_1$8
• Number of speckles: $_1$9
• Speckle contrast: $D \approx 9.7 \pm 0.5$0

Coherence and global beam parameters broaden monotonically and irreversibly, without a near/far-field separation.

4. Comparison with Traditional Diffraction and Other Transport Regimes

Diffraction and diffusion govern spatial propagation via fundamentally distinct equations:

• Paraxial diffraction: $D \approx 9.7 \pm 0.5$1
• Diffusion: $D \approx 9.7 \pm 0.5$2

The mutual coherence propagators accordingly exhibit coupling (for diffraction) versus separability (for diffusion) in center/difference variables:

Regime	Propagator	Regimes of Coherence Evolution
Diffraction	$D \approx 9.7 \pm 0.5$3	Deep Fresnel ($D \approx 9.7 \pm 0.5$4): no local change; Far field ($D \approx 9.7 \pm 0.5$5): classical VCZ scaling
Diffusion	$D \approx 9.7 \pm 0.5$6	Single regime: all coherence widths broaden as $D \approx 9.7 \pm 0.5$7 in $D \approx 9.7 \pm 0.5$8

In DiscDiff, both the global envelope and local coherence decay independently and simultaneously, in contrast to diffraction where local coherence is preserved in the deep Fresnel zone and only widens in the far field.

5. Position-Dependent Diffusion and Boundary/Dissipation Effects

In disordered waveguides and similar open systems, DiscDiff reveals position-dependent diffusion due to interference effects and boundary escape. The self-consistent theory yields (Yamilov et al., 2013):

$D \approx 9.7 \pm 0.5$9

$\tau$0

where $\tau$1 is the modal density and $\tau$2 is the diffusion Green's function. Near system boundaries, the effective diffusion constant approaches the bare value $\tau$3, while in the interior, return probability corrections suppress it.

Dissipative processes such as out-of-plane scattering or localized absorption act to limit interference-induced suppression of diffusion, establishing plateaus in the interior and restoring $\tau$4 at edges. The spatial profile of $\tau$5 can be engineered by geometry or dissipation, enabling the on-chip control of beam expansion, energy confinement, and other wave transport phenomena.

6. Experimental Validation and Applications

Experiments with EIT-based atomic vapors confirmed the theoretical scaling of DiscDiff. By varying the group delay $\tau$6, coherent speckle expansion and reduction in speckle contrast were observed, with autocorrelation widths growing linearly in $\tau$7 (slope $\tau$8), matching diffusion coefficient measurements (Chriki et al., 2018). Bessel-beam-based speckles were shown to be invariant under both diffusion and diffraction, verifying analytic predictions.

In photonic waveguides, direct measurement of internal intensity profiles via out-of-plane scattering allowed mapping of position-dependent diffusion. Results were in quantitative agreement with self-consistent theoretical predictions, confirming spatial control of $\tau$9 via waveguide geometry, disorder strength, and dissipation (Yamilov et al., 2013).

Applications include:

• Tailored manipulation of light transport in photonic chips (on-chip beam shaping, wave broadening/focusing)
• Imaging through turbid and scattering media by exploiting or compensating for diffusive blur
• Random laser devices and controlled energy storage via spatially varying diffusion landscapes
• Extension of DiscDiff concepts to other domains (acoustic/microwave/electronic systems) through analogous control of dephasing and dissipation

7. Fundamental Implications and Theoretical Extensions

DiscDiff establishes a rigorous framework for understanding partial coherence propagation under diffusion. Its extension of the VCZ theorem to a diffusive regime connects directly with wave transport theory in disordered media and mesoscopic physics. The factorization of the propagator highlights the irreversibility of coherence decay in real-time diffusion as opposed to phase-conserving diffraction.

The methodology of spatially varying diffusion, validated both analytically and experimentally, opens avenues for engineering "diffusion landscapes," providing new mechanisms for non-resonant and dissipative control of waves. The precision closed-form solutions and direct experimental observability in DiscDiff mark a significant advancement over previous diffusion/transport approaches limited to intensity-only descriptions. The independence of local and global coherence evolution further distinguishes DiscDiff as a robust tool for probing and manipulating mesoscopic optical systems (Chriki et al., 2018, Yamilov et al., 2013).