---
title: Single-Plane Approximation in Science
url: https://www.emergentmind.com/topics/single-plane-approximation
type: topic
---

# Single-Plane Approximation in Science

Searching arXiv for recent papers and context on "Single-Plane Approximation".
Single-plane approximation denotes a family of domain-specific reductions in which a transformation, field, or geometric object is represented, interrogated, or approximated on one plane, one mask, or one \(xy\)-monotone surface rather than through a multilayer, volumetric, or fully discrete construction. In the recent literature, the phrase is not attached to a single universal formalism. Instead, it appears in Fourier-optical mode sorting, high-frequency Helmholtz discretization, holographic phase retrieval, crystal scattering, tangent-plane estimation, and arrangement geometry, with each use specifying a different object of approximation and a different validity regime [2604.14119], [2107.09797], [2509.22048], [2312.00941], [1404.1823], [1601.04755].

## 1. Conceptual scope and recurring structure

Across the cited works, the single-plane construction serves one of three roles. First, it can be a **device architecture**, as in a single optical mask that performs mode-to-spot mapping. Second, it can be a **measurement geometry**, as in recovering a complex field from intensity measured on one hyperplane. Third, it can be a **reduced model or surrogate surface**, as in local plane-wave ansatzes for Helmholtz solutions, continuous-potential models for a crystalline plane, tangent-plane recovery from inscribed triangles, or a single-surface approximation of a level in a 3D plane arrangement [2604.14119], [2509.22048], [2107.09797], [2312.00941], [1404.1823], [1601.04755].

| Domain | Single-plane object | Target quantity |
|---|---|---|
| Fourier optics | One mask \(S(x,y)\) | Mode sorting/generation |
| Helmholtz numerics | Local directional ansatz | High-frequency approximation |
| Holography | One measurement hyperplane \(X\) | Complex field recovery |
| Crystal scattering | One crystalline plane with continuous potential | Differential cross section |
| Surface geometry | Tangent plane from balanced triangles | Tangent bivector and area |
| Plane arrangements | One \(xy\)-monotone terrain | Approximate \(k\)-level |

A common structural motif is the replacement of a globally complicated transformation by a one-plane representation whose correctness is certified only relative to a specific observable: orthogonality at detector locations, \(L^2\) approximation on an element, asymptotic far-field recovery, eikonal scattering amplitudes, tangent bivector limits, or level-range guarantees. This suggests that “single-plane approximation” is best treated as a methodological pattern rather than a single theory.

## 2. Single optical plane as a mode sorter

In Fourier optics, a mode sorter separates a set of \(M\) orthogonal spatial modes \(\{f_m(x,y)\}_{m=1}^M\) in a shared input channel into \(M\) different output channels. The central single-plane construction in "Single Plane Spatial Mode Sorter" is an analytic mask \(S(x,y)\) placed in one optical plane, followed by far-field detection at positions \((\alpha_m,\beta_m)\). The field at far-field position \((\xi,\eta)\) is
\[
E_m(\xi,\eta)=\iint f_m(x,y)S(x,y)e^{-i\frac{2\pi}{\lambda z}(\xi x+\eta y)}\,dxdy,
\]
and the detector-sampled amplitude is
\[
E_{m,\mu}=\iint f_m(x,y)S(x,y)e^{-i\frac{2\pi}{\lambda z}(\alpha_\mu x+\beta_\mu y)}\,dxdy.
\]
The analytic mask is given by
\[
S(x,y)=\frac{1}{\sqrt{M}\sum_{m'=1}^M f_{m'}^*(x,y)\,e^{i\frac{2\pi}{\lambda z}(\alpha_{m'}x+\beta_{m'}y)}}.
\]
For orthonormal modes, the sampled matrix is driven toward a diagonal form, yielding \(I_{m,\mu}\approx \frac{1}{M}\delta_{m,\mu}\), and the paper reports experimental validation with almost no cross-talk for Hermite-Gaussian, Laguerre-Gaussian, and Bessel-Gaussian families [2604.14119].

The construction is family-agnostic as long as the modes are orthogonal. For Orbital Angular Momentum modes with zero radial index, the design reproduces the well-known Fork grating configuration. By taking the limit \(M\to\infty\), the mask becomes a generating function over the mode family, and closed analytical forms are given for several families. The device can also be operated in reverse to generate arbitrary modes by illuminating the mask with a Gaussian beam [2604.14119].

The principal limitation is energetic rather than algebraic. The power transmission coefficient per mode scales as \(1/M\), and the paper provides a proof that this is optimal for any typical arrangement of detector positions. In the \(M\to\infty\) limit, power per sorted mode therefore vanishes, while the mask becomes extremely high-frequency and is limited by optical or pixel resolution. The same work studies sensitivity to wavelength and random phase noise, and reports that phase-only implementations on SLMs introduce negligible error for commonly used mode sets [2604.14119].

## 3. Helmholtz approximation via a single local wave direction

In high-frequency Helmholtz computation, the phrase “single-plane approximation” appears in relation to basis functions of the form \(p(\mathbf r)e^{i\kappa \mathbf d\cdot\mathbf r}\), where a single local direction \(\mathbf d\) captures the dominant phase gradient. "A plane wave method based on approximate wave directions for two dimensional Helmholtz equations with large wave numbers" develops an adaptive plane wave space that extends this idea by computing approximate local wave directions from the geometrical optics ansatz
\[
u(\omega,\mathbf r)=\sum_{n=1}^{N(\mathbf r)}A_n(\mathbf r)e^{i\omega\phi_n(\mathbf r)},
\qquad |\nabla \phi_n(\mathbf r)|^2=\xi(\mathbf r).
\]
On an element \(K_0\) with barycenter \(\mathbf r_0\), each detected direction yields
\[
\tau_{h,n}(\mathbf r)=\sqrt{\xi(\mathbf r_0)}\,\mathbf d_{h,n}\cdot(\mathbf r-\mathbf r_0),
\]
and the local adaptive space
\[
V_{ray}(K_0)=\mathrm{span}\left\{\varphi_{n,j}(\omega,\mathbf r)\mid n=1,\dots,N(\mathbf r_0),\ j=1,2\right\},
\]
with
\[
\varphi_{n,j}(\omega,\mathbf r)=p_j^{\tau_{h,n}(\mathbf r)}e^{i\omega\tau_{h,n}(\mathbf r)}.
\]
The direction computation combines a low-frequency solve, NMLA, and nonlinear least-squares post-processing, improving the direction error to \(\mathcal O(\omega^{-1})\) [2107.09797].

The resulting best local approximation estimate is
\[
\|u-\pi_h u\|_{L^2(K_0)}\le C\left(\omega^{-1}+h+\omega h^2\right).
\]
When \(h\sim \omega^{-1}\), this becomes \(\|u-\pi_h u\|_{L^2(K_0)}\le C\omega^{-1}=Ch\), i.e. first-order convergence in \(h\). The paper explicitly identifies the method as an extension of the “single-plane approximation” used in earlier works, with two principal changes: higher-accuracy directional recovery and a reduction to two basis functions per direction, which lowers the local dimension relative to earlier linear-polynomial enrichments [2107.09797].

A related but distinct line of Helmholtz research shows that plane-wave-only approximation can be numerically unstable even when the approximation space is formally rich. "Stable approximation of Helmholtz solutions in the disk by evanescent plane waves" proves that propagative plane waves alone can require exponentially large coefficients regardless of orientations and number of plane waves, whereas continuous superpositions of evanescent plane waves give bounded representations and a continuous frame for the solution space [2202.05658]. This indicates that a single-direction or purely propagative local ansatz is not, by itself, a guarantee of stable numerical realization.

## 4. One measurement plane in holographic phase recovery

In inverse scattering and holography, the single-plane approximation is a measurement geometry rather than a basis reduction. "A two-point phase recovering from holographic data on a single plane" considers the total field \(\psi_0+\psi_1\) for the Helmholtz equation and seeks to recover the unknown radiation solution \(\psi_1\) on a hyperplane \(X\) from intensity alone:
\[
I(x)=|\psi_0(x)+\psi_1(x)|^2,\qquad x\in X.
\]
For a plane \(X\) at large distance \(s\) from the origin, the paper derives a two-point local formula for approximate recovery of the far-field pattern \(f_1(\theta)\) from intensity values at two nearby points \(x\) and \(y=x+\zeta\):
\[
f_1(\theta)\approx f_{1,1}(\theta)=\frac{1}{D}\left[e^{i(ky-|k||y|)}a(x,k)-e^{i(kx-|k||x|)}a(y,k)\right],
\]
where
\[
a(x,k)=|x|^{(d-1)/2}\left(|\psi_0(x)+\psi_1(x)|^2-1\right),
\qquad
D=2i\sin\big(k\zeta+|k||x|-|k||y|\big).
\]
The reconstruction error is
\[
\frac{1}{D}\left(\mathcal O\left(\frac{1}{|x|^\sigma}\right)+\mathcal O\left(\frac{|\zeta|}{|x|}\right)\right),
\quad
\sigma=
\begin{cases}
1/2,& d=2,\\
1,& d\ge 3.
\end{cases}
\]
The same asymptotic substitution yields recovery of \(\psi_1(x)\) itself on the plane [2509.22048].

The significance of the construction is that it uses only intensity data on a single plane and does not require spatial integration. Its accuracy improves with increasing plane distance \(s\) and decreasing separation \(|\zeta|\). The method is nevertheless not translation invariant, and it has unavoidable singular directions: the denominator \(D\) can vanish, and the direction \(\theta=k/\kappa\) is always singular in the two-point formula. The numerical implementation reported for \(d=3\) shows a central artifact near this singular direction and improved accuracy as \(s\) increases [2509.22048].

## 5. Distinction from plane-wave approximation

A recurrent source of confusion is the proximity of the phrases **single-plane approximation** and **plane wave approximation**. The cited literature treats them as different notions. Single-plane approximation concerns the number of planes, masks, or surfaces used in the model or measurement; plane-wave approximation concerns the choice of basis or continuum state.

This distinction is explicit in molecular orbital tomography. "Molecular orbital tomography beyond the plane wave approximation" revisits the standard HHG reconstruction method in which the continuum electron is modeled as
\[
\psi_{\mathbf k}^{\mathrm{PW}}(\mathbf r)=(2\pi)^{-3/2}e^{i\mathbf k\cdot \mathbf r},
\]
an assumption that neglects the parent ion’s Coulomb field. The paper states that this can produce artifacts and unphysical features, and identifies a long-standing controversy about the validity of plane-wave-based orbital tomography. Its replacement is a continuum-wave formalism in momentum space with a mapping matrix
\[
\mathbb S_{ij}=\varphi_{\mathbf k_i}(\mathbf k'_j),
\]
so that reconstruction proceeds by inverting \(\mathbb S\). Using a Two-Center Coulomb wave, the method reconstructs the \(3\sigma_g\) HOMO of \(\mathrm{N_2}\) with good agreement with the \emph{ab initio} orbital, whereas plane-wave-based reconstructions are reported as inferior [1401.4050].

An analogous issue appears in neutrino–nucleus pion production, though there the approximation is RPWIA rather than a single-plane construction. In "Pion production within the hybrid relativistic plane wave impulse approximation model at MiniBooNE and MINERvA kinematics," outgoing hadrons are treated as plane waves and final-state interactions are neglected. The paper emphasizes that this is a first step and a baseline calculation, but also reports that hybrid-RPWIA predictions largely underestimate the MiniBooNE data and lie below several MINERvA channels [1710.08374]. These results do not invalidate single-plane reductions in other settings; they instead show that plane-wave simplifications can fail when neglected interactions are dynamically important.

## 6. Tangent planes, continuous crystalline planes, and single-surface surrogates

In geometric approximation, one use of the term concerns literal tangent-plane recovery. "Algorithms, unaffected by the Schwarz paradox, approximating tangent planes and area of smooth surfaces via inscribed triangular polyhedra" shows that naive secant-plane limits may fail, even locally, because of the Schwarz paradox. The remedy is a balanced construction using a mirror vertex \(a'\) associated with an inscribed triangle \([a,b,c]\). For a smooth surface \(s:\Omega\to\mathbb E_n\), the balanced mean bivector satisfies
\[
\lim_{(a,b,c)\to(x,x,x)}
\frac{\langle s(a);s(b);s(c)\rangle-\langle s(a');s(b);s(c)\rangle}
{\langle a;b;c\rangle-\langle a';b;c\rangle}
=
\partial_{\ell_1}s(x)\wedge \partial_{\ell_2}s(x),
\]
and the associated area approximation is
\[
\lim_{\|\Pi\|\to 0}\frac14\sum_{[a_i,b_i,c_i]\in\Pi}
\left|[s(a_i')-s(a_i)]\wedge [s(c_i)-s(b_i)]\right|
=
\int_P |\partial_{\ell_1}s(x)\wedge \partial_{\ell_2}s(x)|\,dx.
\]
The single-plane object here is the tangent plane encoded by the limiting bivector, and correctness depends on the balanced-triangle construction rather than on arbitrary secant planes [1404.1823].

A different geometric use appears in arrangements of planes in \(\mathbb R^3\). "Approximating the \(k\)-Level in Three-Dimensional Plane Arrangements" constructs an \(xy\)-monotone polyhedral terrain that approximates the \(k\)-level. The terrain has \(O(r/\epsilon^3)\) triangular faces, lies entirely between levels \((1\pm\epsilon)n/r\), and yields a shallow cutting by taking the vertical prisms beneath its triangles. In the paper’s language, this is a single-surface approximation of a complicated level set, with each prism crossed by at most \(O(n/r)\) planes [1601.04755].

Single-plane reduction also appears in scattering from a crystalline plane. "The quantum rainbow scattering effect on a single crystalline plane in approximation of continuous potential" replaces the discrete atomic lattice by a continuous potential depending only on the distance \(x\) normal to the plane and uses the eikonal approximation of quantum electrodynamics. The reduced amplitude is
\[
\tilde a=\int_{-\infty}^{\infty}dx\,
e^{\frac{i}{\hbar}[q_xx+N\bar\chi_0(x)]},
\]
with differential cross section
\[
\frac{d\sigma}{dq_x}=\frac{L_y}{2\pi}|\tilde a|^2.
\]
Near the rainbow point, the cubic expansion of the phase yields an Airy-function formula for the cross section. Here the single-plane approximation is a continuous-potential model of one crystallographic plane, valid in the eikonal, high-energy, small-angle regime [2312.00941].

Taken together, these geometric and scattering examples show that one-plane reduction can mean a tangent object extracted from local samples, a terrain approximating a level set, or a continuous-potential replacement for a discrete atomic plane. The recurring gain is analytical tractability or algorithmic compression; the recurring cost is a sharply delimited regime of validity. In the surveyed literature, such limitations are explicit: balancedness is required to avoid the Schwarz paradox, terrain approximations are controlled only between neighboring levels, eikonal formulas require appropriate high-energy conditions, and optical single-plane sorters are constrained by the \(1/M\) power bound.

Source: https://www.emergentmind.com/topics/single-plane-approximation