---
title: 'Beam Divergence: Concepts and Applications'
url: https://www.emergentmind.com/topics/beam-diverging-effect-3d3505c7-c590-4e01-887e-e79435bd7721
type: topic
---

# Beam Divergence: Concepts and Applications

Searching arXiv for the cited paper and closely related beam-divergence literature to ground the article in the provided sources.
“Beam diverging effect” is not a single, field-invariant term. Across the cited literature, it denotes several distinct but structurally related phenomena: ordinary transverse broadening under propagation; lower bounds on divergence imposed by orbital angular momentum (OAM); resonance-induced reshaping, narrowing, or splitting of reflected beams; defocusing by negative-lens media or graded-index landscapes; active-particle transport in spatially diverging optical fields; scattering- or interaction-driven spreading in charged-particle beams; and deliberately synthesized wide beams for near-field beam training in extremely large arrays [1609.01497], [1601.02350], [2509.16035]. The common theme is a redistribution of transverse energy, momentum, or trajectory support, but the operative observable varies by domain: rms width, encircled-energy angle, centroid shift, beam footprint, halo population, or sectorwise coverage.

## 1. Terminological scope and conceptual distinctions

In the optical-propagation literature, “beam divergence” often refers to monotonic broadening quantified by a width law or a far-field angle. Several papers here explicitly warn that many experimentally important “diverging” effects are not of that type. For finite-energy Fresnel diffraction patterns, the salient observables are the trajectory of the main intensity maximum, the suppression and reappearance of oscillatory fringes, and a critical propagation distance separating deceleration from acceleration; there is no divergence angle, no \(w(z)\) law, and no second-moment width analysis [1312.3477]. Near critical incidence in a dielectric prism, the asymmetric Goos–Hänchen regime likewise produces a dynamical shift, asymmetric interference, and multiple peaks, rather than a replacement of the ordinary Gaussian diffraction law [1311.4773]. Near quasi-bound states in the continuum, the reflected beam can undergo distortion, asymmetric broadening, and splitting into a main spot plus a satellite spot; this is best understood as resonant angular-spectrum reshaping, not as simple free-space broadening [2303.06196].

A precise distinction therefore runs through the modern usage. One must separate width growth from trajectory bending, centroid displacement, spectral filtering, lobe splitting, and profile deformation. In some papers, the “diverging effect” is literal defocusing; in others it is a convenient descriptor for any process by which beam support spreads, reorganizes, or becomes less localized in its nominal state space. This suggests that the term is best treated as a family resemblance concept rather than a single transport law.

## 2. Diffraction, OAM, and finite-energy propagation

For paraxial optical beams carrying OAM, divergence is constrained by a rigorous lower bound. The general theorem of “A general theorem on the divergence of vortex beams” proves that the mean absolute OAM content raises the minimum possible rms divergence through
\[
\mathcal M_{\rm rms}^2 \equiv k\,\theta_{\rm rms}\,\sigma_m \ge 1+\langle |\ell| \rangle,
\]
or equivalently
\[
\sigma_r\,\sigma_k \ge 1+\langle |\ell| \rangle.
\]
The bound is tight and is saturated by arbitrary superpositions of Laguerre–Gauss modes with radial index \(n=0\) [1601.02350]. In that sense, OAM does not merely correlate with divergence; it imposes a universal lower floor on confinement and focusing.

A complementary practical clarification is given for \(p=0\) Laguerre–Gaussian modes. “Divergence of an orbital-angular-momentum-carrying beam upon propagation” resolves the apparent conflict between \(\sqrt{|\ell|}\) and \(|\ell|\) divergence scaling by showing that both apply, depending on what is held fixed. If the Gaussian waist \(w_0\) is fixed, the far-field divergence scales as \(\alpha_\ell \propto \sqrt{|\ell|+1}\). If instead the rms launch radius is fixed, then \(\alpha_\ell \propto |\ell|+1\) [1410.8722]. The discrepancy is therefore not physical inconsistency but a change of constraint manifold.

Bessel-beam optics provides another corrective to oversimplified language. “Engineering the axial intensity of Bessel beams” emphasizes that a real Bessel beam is a finite-energy, quasi-Bessel beam: its transverse profile is approximately preserved only over a finite depth of focus,
\[
DOF = \frac{R}{(n-1)\alpha},
\]
after which it gradually transforms into a ring of constant width and increasing radius [1711.00576]. The work’s central concern is not elimination of divergence in the strict sense, but control of axial intensity redistribution within that finite nondiffracting interval.

Finite-energy Fresnel accelerating beams furnish a different non-Gaussian divergence picture. Their dominant lobe follows \(x^2\propto z\), with an initial deceleration branch \(-x^2\propto z\), a critical distance where transverse motion stops, and self-smoothing of fringes before oscillations reappear [1312.3477]. The relevant transport observable is therefore a turning-point dynamics of structured lobes, not envelope broadening.

## 3. Resonant and interfacial beam reshaping

Resonant structures can reverse the usual intuition that reflection increases beam width. In “Leveraging beam deformation to improve the detection of resonances,” the reflected beam width change is quantified by the centered second-moment variation
\[
\Delta=\frac{\int (x-\delta)^2 |E_r|^2 dx}{\int |E_r|^2 dx}-\frac{\int x^2 |E_i|^2 dx}{\int |E_i|^2 dx},
\]
with the asymptotic wide-beam limit
\[
\Delta_{\infty}=\frac{\rho'^2-\rho\,\rho''}{2\rho^2}.
\]
At a reflection minimum, \(\rho'=0\) and \(\rho''>0\), so \(\Delta_\infty<0\): the reflected beam becomes narrower than the incident beam. The paper reports narrowing by about \(10\%\) over a very narrow angular interval and uses the resulting sharper signature to improve surface-plasmon-resonance detection resolution by a factor three [1609.08473]. Diverging behavior is thus not universal even in highly resonant reflection; amplitude curvature can produce transient narrowing.

Near critical incidence in a right-angle prism, the outgoing angular spectrum becomes strongly asymmetric in \(k_y\), producing what the authors call an asymmetric Goos–Hänchen effect. The beam acquires a nonzero \(\langle k_y\rangle\), and the maximum drifts according to relations such as
\[
\Delta y=-\frac{\langle k_y\rangle}{k}\Delta z
\quad\text{or}\quad
\Delta z=\frac{\langle k_y\rangle}{k}\Delta y,
\]
depending on geometry [1311.4773]. The transmitted profile may also develop asymmetric interference and multiple peaks. The apparent “spreading” is therefore a consequence of spectrum filtering near criticality, not a new free-propagation divergence law.

A stronger version of this resonant reshaping appears near quasi-BICs. In a symmetric SiO\(_2\)-air-SiO\(_2\) multilayer, one of the coupled Berreman modes evolves toward a bound state in the continuum, generating an ultra-narrow reflection dip and a sharp phase jump. For a finite \(p\)-polarized Gaussian beam, the reflected angular spectrum splits into left and right lobes, and the real-space beam becomes distorted, with a main spot plus a satellite spot. The origin is destructive interference between the two spectral halves [2303.06196]. In this regime, even centroid-based Goos–Hänchen shifts cease to characterize the full deformation, because the reflected beam is no longer close to a translated Gaussian.

## 4. Diverging fields as transport and lensing mechanisms

In active colloids, a diverging optical field can qualitatively reorganize particle dynamics. “Trochoidal trajectories of self-propelled Janus particles in a diverging laser beam” studies hematite–TPM Janus colloids illuminated by a defocused optical tweezer whose focal point lies approximately \(L \approx \SI{10}{\micro m}\) below the lower glass surface. In the plane of motion, the measured intensity is
\[
I(\mathbf r)= I_0 \exp\!\left(-\frac{r^2}{2\sigma^2}\right),
\]
with \(\sigma=\SI{7.3(2)}{\micro m}\). The diverging beam simultaneously sets the self-thermophoretic propulsion speed,
\[
\mathbf v(\mathbf r)= \alpha\, I(\mathbf r)\,\hat n,
\]
and generates a weak outward radiation-pressure component on the off-center absorbing inclusion, creating an optical torque. The coupled translation-rotation dynamics yields bounded rosette or hypotrochoid-like trajectories instead of ballistic escape or passive trapping [1609.01497]. Here the beam-diverging effect is not beam broadening but the appearance of an outward in-plane force component that exists because the beam is defocused.

Microhydrodynamic optics supplies a literal defocusing analogue. “Theory of microdroplet and microbubble deformation by Gaussian laser beam” shows that a water microdroplet in air acts as a positive lens, whereas an air microbubble in water acts as a negative lens. The bubble therefore spreads light out and away from its interior, strongly suppressing rear-side intensity enhancement. The resulting surface stress,
\[
\sigma(\theta)=\frac{\varepsilon_0n_2^2}{4} (n^2-1)(n^2|E^w_r|^2+|E^w_\theta|^2+|E^w_\phi|^2),
\]
and the induced static deformation amplitudes,
\[
h_l(\infty) = \frac{\sigma_l a^2}{\gamma}\frac1{l^2+l-2},
\]
differ radically between droplets and bubbles; the paper concludes that a laser which significantly disfigures a water droplet will typically hardly change a bubble at all [1212.2756]. The diverging effect here is the optical self-defocusing of a lower-index sphere.

A magnonic counterpart appears in graded-index spin-wave optics. In a \(10\) nm-thick out-of-plane magnetized YIG film, a region where the static magnetic field increases gradually acts as a diverging lens for spin waves, because the local spin-wave refractive index decreases and the group-velocity direction rotates continuously across the beam. The ray model uses
\[
\cot\varphi(y)=\frac{\sqrt{k^{2}\left(\omega,H\left(y\right)\right)-k_{x}^{2}}}{k_{x}},
\]
with \(k_x\) conserved. When the increase of \(H(y)\) is slow and remains below the total-reflection threshold, the transmitted spin-wave beam is spread rather than reflected [1707.09768]. Divergence is therefore realized as graded refraction in a structured magnetic landscape.

## 5. Charged-particle beams: scattering, pinching, and beam-beam blow-up

In charged-particle systems, “beam diverging effect” often denotes emittance growth, halo, or beam blow-up rather than geometric diffraction. A clear material-scattering example is the proton beam window in spallation sources. The proton beam window is the first material obstacle seen by the high-power beam, and the paper identifies multiple scattering as the dominant process degrading target beam quality. In the CSNS-I simulations, increasing the aluminum-window thickness from \(0\) to \(4\) mm raised beam loss outside the target from \(5.0\) W to \(1362\) W, while increasing the PBW-to-target distance from \(0.5\) m to \(2.5\) m at \(t=1.0\) mm raised that loss from \(308\) W to \(430.2\) W [1012.5460]. The operative mechanism is thin-scatterer angular diffusion followed by drift-space conversion into footprint broadening and halo.

Plasma wakefield acceleration provides a contrasting case in which divergence is suppressed rather than enhanced. “Effect of Transverse Beam Size on the Wakefields and Driver Beam Dynamics in Electron Beam Driven Plasma Wakefield Acceleration” explicitly states that the self-consistent transverse evolution of a relativistic electron driver in a cold homogeneous plasma is dominated by transverse pinching. The relevant transverse force is
\[
F_x = -e\,(E_x - v_{bz} B_y),
\]
and in the cases studied its sign is net focusing over the beam core. The paper concludes that the beam propagating inside the plasma undergoes transverse pinching much earlier than longitudinal modification, so the main transverse effect is not net divergence but early focusing-induced deformation [1909.09874]. This is an important negative case: finite transverse size introduces new transverse dynamics, but those dynamics need not be diverging.

In colliders, Herr and Pieloni describe beam–beam interaction as one of the most severe limitations of high-intensity facilities. Each opposing bunch acts as a nonlinear electromagnetic lens. For round beams, the head-on kick is
\[
\Delta r'~~=~~-\frac{2N r_{0}}{\gamma } \cdot\frac{1}{r}\cdot\left [  1 - {\mathrm{exp} \left (  -\frac{r^{2}}{2\sigma^{2}}\right ) \right ],
\]
and the linearized beam–beam parameter is
\[
\xi~=~\frac{N r_{0} \beta^{*}}{4\pi\gamma\sigma^{2}}.
\]
The resulting tune spread, resonance excitation, dynamic-aperture reduction, tail growth, and halo formation constitute the collider analogue of a diverging effect [1601.05235]. What appears operationally as beam spreading is thus the phase-space consequence of nonlinear kicks, not a free-propagation envelope law.

## 6. Engineered beam divergence in near-field communications

Near-field beam training for extremely large arrays introduces a deliberately synthesized use of beam divergence. “Near-Field Beam Training Through Beam Diverging” argues that sharply focused near-field codewords are ill-suited to coarse search because they are highly sensitive to mismatch in angle and distance. The paper defines a diverging codeword for a virtual focal point behind the array by
\[
\bm c(\theta_v,r_v)\triangleq \overline{\bm f(\theta_v,r_v)},
\]
and observes that the normalized received power is substantially higher when the user lies in the sector induced by that virtual point than when it lies outside. The resulting diverging polar-domain codebook supports hierarchical angular localization using only
\[
2\log_2(N)
\]
pilots, followed by near-field refinement [2509.16035]. The key shift is conceptual: divergence is not a parasitic imperfection but a phase-only, single-RF-chain design primitive for robust wide-beam coverage.

The 3D UPA extension generalizes this idea from sectors to volumetric frusta. In “3D Near-Field Beam Training for Uniform Planar Arrays through Beam Diverging,” a diverging codeword associated with a virtual focal point behind the array produces, on each plane \(y_U=\hat y\), a rectangular high-power region given by a projection of the aperture, and over depth this becomes an unbounded rectangular pyramidal frustum [2509.16055]. The coarse search is therefore organized hierarchically in 3D, after which a sampled near-field focusing codebook performs refinement. The proposed two-phase method uses \(214\) pilots in the reported \(64\times 64\) UPA setting, versus \(1383\) for ULA-based DFT sweeping and \(381\) for UPA partitioning, while retaining one-RF-chain implementation [2509.16055].

These communication papers invert the conventional meaning of divergence. A wide beam is no longer a failure of alignment but a controlled means of reducing overhead, enlarging capture probability, and stabilizing hierarchical search in the Fresnel regime.

## 7. Unifying interpretation

Across these domains, the beam diverging effect can be organized by the quantity that is spreading. In paraxial OAM optics, the spread is a lower-bounded far-field width set by \(\langle|\ell|\rangle\) [1601.02350]. In resonant reflection and critical-angle transport, the spread is spectral or modal, and may even coexist with transient narrowing [1609.08473]. In defocused optical tweezers, graded magnonic media, and microbubble optics, the spread is imposed by geometry: an outward beam component, a decreasing spin-wave refractive index, or a negative-lens inclusion reorganizes trajectories or stress maps [1609.01497], [1212.2756]. In accelerator settings, the spread is phase-space dilution, halo, or material-scattering broadening [1012.5460]. In near-field communications, it is an intentionally designed coverage region [2509.16035].

A persistent misconception is therefore that “beam diverging effect” must mean monotonic Gaussian broadening. The cited literature shows a much richer landscape: divergence may be bounded, suppressed, reversed, resonantly sharpened, converted into torque, realized as a lensing function, or engineered as a codebook-level resource. The broader implication is that beam divergence is best understood not as a single observable but as a geometry-dependent redistribution of transverse support in real space, reciprocal space, or phase space.

Source: https://www.emergentmind.com/topics/beam-diverging-effect-3d3505c7-c590-4e01-887e-e79435bd7721