Reflectionless Modes in Wave Scattering

• Reflectionless modes are wave solutions engineered to cancel backscattering in designated channels while still allowing transmission, conversion, or absorption.
• They appear across various platforms such as SNAP fibers, photonic lattices, and microwave cavities, utilizing boundary conditions, symmetry, or specially designed potentials.
• Spectral operator methods, coherent multi-channel interference, and adiabatic mode conversion form the core mechanisms that enable these modes in practical wave systems.

Searching arXiv for the cited works on reflectionless modes and closely related formulations. Reflectionless modes are wave solutions or scattering states engineered so that no reflected amplitude appears in a designated set of channels, despite nontrivial propagation through an inhomogeneous medium or structure. Across optics, acoustics, quantum-wave systems, photonic lattices, microwave cavities, and waveguides, the term denotes related but not identical objects: whispering gallery modes whose axial envelope experiences a reflectionless potential in a SNAP fiber (Suchkov et al., 2015); propagating lattice scattering states rendered reflectionless by fast drift in a discrete photonic lattice (Longhi, 2017); reflectionless scattering modes defined by zeros of a filtered reflection operator in multiport cavities (Sol et al., 2022); and eigenfunctions of a non-selfadjoint spectral problem in open waveguides (Dhia et al., 2018). In most formulations, the unifying criterion is not the absence of all scattering, but the cancellation of backscattering into chosen input channels, with transmission, conversion, absorption, trapping, or delay still allowed. This suggests that “reflectionless modes” are best understood as a broad scattering-theoretic class organized by boundary conditions, symmetry constraints, and the analytic structure of reflection operators rather than by a single physical platform.

1. Definitions and conceptual variants

A common formal definition appears in multiport scattering theory. If outgoing amplitudes $b$ and incoming amplitudes $a$ are related by $b=S(\omega)a$, and if only a subset of ports is designated as input ports, then the relevant object is the submatrix $R_{\rm in}$ mapping those inputs back into themselves. A reflectionless condition is the existence of a nonzero $a_{\rm in}$ such that

$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$

Equivalently, $\det R_{\rm in}(\omega_0)=0$, or $R_{\rm in}(\omega_0)$ has a zero eigenvalue. The corresponding complex-frequency solutions are $R$-zeros, and when an $R$-zero lies on the real axis the resulting real-frequency state is a reflectionless scattering mode (Sol et al., 2022). In this formulation, coherent perfect absorption is the special case $a$0, so reflectionless modes generalize CPA rather than coincide with it.

In one-dimensional or quasi-one-dimensional scattering, the definition is often expressed more directly through the reflection coefficient. In SNAP optical fibers, the axial whispering-gallery-mode envelope propagates in an effective one-dimensional potential, and a mode is reflectionless when the corresponding reflection coefficient vanishes, $a$1, even though the potential still induces phase shift, resonance structure, localization, and delay (Suchkov et al., 2015). In stratified optics, Kay–Moses-type dielectric profiles are designed so that incoming waves are transmitted with zero reflection, and the optical field can then be interpreted as a reflectionless scattering state of the corresponding one-dimensional Schrödinger problem (Biswas et al., 2024).

Several papers distinguish exact complex-frequency reflectionless modes from real-frequency approximations. In asymmetric Damour–Solodukhin wormholes, quasi-reflectionless scattering modes are defined on the real axis as minima of $a$2, whereas reflectionless modes are their exact analytic continuation into the complex-frequency plane (Qian et al., 1 Nov 2025). In finite photonic structures, complex-frequency reflection zeros are the primary spectral objects, and a physical reflectionless scattering mode appears only when one of those zeros is tuned to the real axis (Stone et al., 2020). In open waveguides with obstacles, a reflectionless mode is a total field that is ingoing in the input lead and outgoing in the output lead, with only evanescent tails in the reflected lead, and such modes can be extracted as real eigenvalues of a non-selfadjoint operator (Dhia et al., 2018).

A recurring misconception is that reflectionless means invisible. Several of the cited works explicitly reject that identification. A drifting defect in a discrete photonic lattice can become reflectionless for all propagating channels when the drift exceeds the lattice light-cone velocity, yet the transmitted packet can still be distorted, delayed, or advanced (Longhi, 2017). In SNAP fibers, reflectionless axial wells alter transmission amplitude, phase, and time delay while suppressing backscattering (Suchkov et al., 2015). Reflectionless acoustic cloaking by liner surface modes suppresses backscattering but still exhibits slow-sound delay and phase distortion (Farooqui et al., 2018). Thus reflectionless propagation is generally weaker than invisibility.

2. Spectral and operator formulations

One major line of work treats reflectionless modes as zeros of analytically continued reflection operators. In arbitrary finite photonic structures, the generalized reflection matrix $a$3 defines the condition

$a$4

which identifies complex-frequency R-zeros. A reflectionless scattering mode is obtained when such an R-zero lies on the real axis (Stone et al., 2020). In the multichannel case, the reflectionless input is the null vector of $a$5, so the mode is wavefront-specific rather than channel-independent. This suggests that in generic systems reflectionlessness is a codimension-one condition in parameter space, because one must generally tune $a$6 to zero.

A closely related but distinct operator formulation appears for waveguides with obstacles. There, the standard Helmholtz operator is combined with a complex scaling that uses opposite signs of the imaginary part in the two leads. This produces a non-selfadjoint operator $a$7 whose real isolated eigenvalues correspond either to trapped modes or to reflectionless modes (Dhia et al., 2018). The decisive distinction is whether the field projected onto propagating modes on the input side vanishes. If it does, the eigenstate is a trapped mode; if it does not, the eigenstate is reflectionless. This is significant because it turns a scattering-zero problem into an eigenvalue problem, permitting direct spectral computation of frequencies with zero reflection.

In wormhole scattering, the same split between transfer-matrix zeros and Green-function poles appears in another form. Reflectionless modes satisfy $a$8 for the total transfer matrix, with $a$9, whereas echo modes satisfy $b=S(\omega)a$0 with $b=S(\omega)a$1 (Qian et al., 1 Nov 2025). In the modified Green-function picture, reflectionless modes are poles of a resolvent constructed with reflectionless boundary conditions. The paper argues that these two viewpoints are equivalent descriptions of the same spectral condition.

Floquet systems introduce a further extension. In a time-periodically driven cavity, reflectionless scattering is encoded in an auxiliary non-Hermitian operator defined on the synthetic frequency lattice. The corresponding synthetic reflection zeros satisfy $b=S(\omega)a$2, and real zeros define synthetic reflectionless modes (Tuxbury et al., 5 Aug 2025). Here the “mode” is not merely a spatial cavity field but an eigenstate of an auxiliary operator in harmonic space, reflecting the synthetic-dimension structure of the Floquet problem.

3. Symmetry, $b=S(\omega)a$3 symmetry, and exceptional degeneracies

Symmetry is one of the most systematic mechanisms for producing real reflectionless modes without explicit tuning. In arbitrary photonic structures, systems with parity and time-reversal symmetry or with $b=S(\omega)a$4 symmetry generically possess subsets of real R-zeros, so reflectionless scattering modes can exist without structural tuning (Stone et al., 2020). This differs from generic asymmetric systems, where one normally must tune at least one parameter to bring an R-zero to the real axis.

In open waveguides with mirror-symmetric obstacles, the non-selfadjoint operator $b=S(\omega)a$5 obeys $b=S(\omega)a$6, and its spectrum is symmetric with respect to complex conjugation (Dhia et al., 2018). Real eigenvalues may then persist until a broken-$b=S(\omega)a$7-symmetry transition occurs, after which they split into complex-conjugate pairs. In numerical examples, the real eigenvalues correspond to exact trapped or reflectionless modes, while nearby complex eigenvalues indicate weak-reflection regimes.

A different $b=S(\omega)a$8-symmetric construction appears in two-cavity optical limiters. There, the relevant operator is not the resonance Hamiltonian but an auxiliary operator

$b=S(\omega)a$9

whose real eigenvalues are reflection zeros (Riboli et al., 2023). Under

$R_{\rm in}$0

the R-zero eigenvalues are

$R_{\rm in}$1

At $R_{\rm in}$2 they coalesce at an exceptional point of degeneracy, and the transmission adopts the quartically flat form

$R_{\rm in}$3

The resulting flat-top passband is then destroyed by cavity detuning, which breaks the $R_{\rm in}$4-symmetric reflectionless-mode spectrum and turns the device reflective (Riboli et al., 2023).

Floquet-driven systems display an analogous but synthetic version. After reduction to a resonant subspace, the auxiliary reflection-zero operator becomes a local $R_{\rm in}$5-symmetric dimer,

$R_{\rm in}$6

with exceptional degeneracy at $R_{\rm in}$7 (Tuxbury et al., 5 Aug 2025). Near this SRM-EPD, the reflection amplitude behaves as

$R_{\rm in}$8

so the reflectance obeys

$R_{\rm in}$9

This quartic law is the Floquet counterpart of the flattened reflection minimum found in static $a_{\rm in}$0-symmetric reflectionless-mode degeneracies (Tuxbury et al., 5 Aug 2025).

4. Canonical physical mechanisms

One large class of reflectionless modes is produced by deliberately engineered one-dimensional potentials. In SNAP optical fibers, nanoscale axial variation of the effective fiber radius generates an axial potential for whispering gallery modes. For the tuned Pöschl–Teller/Kay–Moses form

$a_{\rm in}$1

the potential is reflectionless, and the local Green’s function becomes non-periodic in the coupler position $a_{\rm in}$2 (Suchkov et al., 2015). By contrast, conventional wells generate nonzero $a_{\rm in}$3 and thus periodic oscillations in transmission and delay. Reflectionlessness therefore suppresses oscillatory tails and cross-talk between nearby elements (Suchkov et al., 2015).

Inverse-scattering-designed stratified dielectric media provide an optical analog of the same mechanism. Starting from the Kay–Moses determinant formula

$a_{\rm in}$4

one obtains an index profile

$a_{\rm in}$5

whose ideal TE plane-wave scattering is exactly reflectionless (Biswas et al., 2024). For finite Gaussian and Laguerre–Gaussian beams, the profile remains near reflectionless when analyzed through angular-spectrum decomposition, typically yielding less than $a_{\rm in}$6 reflection in most scenarios and preserving beam shape far better than a conventional $a_{\rm in}$7 antireflection coating (Biswas et al., 2024).

A second mechanism is adiabatic mode conversion. In acoustic ducts lined with a smoothly varying resonant admittance $a_{\rm in}$8, the propagating plane wave is continuously deformed into a surface-confined mode localized near one wall. Because the liner varies slowly, this conversion occurs with negligible reflection, creating a silent zone near the opposite wall that can cloak obstacles (Farooqui et al., 2018). Here “reflectionless” is practical rather than exact: the paper reports, for example, reductions such as $a_{\rm in}$9 for triangular obstacles and $R_{\rm in}(\omega_0)\,a_{\rm in}=0.$0 for a rectangular obstacle at $R_{\rm in}(\omega_0)\,a_{\rm in}=0.$1 (Farooqui et al., 2018).

A third mechanism is kinematic channel suppression in lattices. In a discrete photonic lattice with a drifting localized potential, the moving-frame dispersion

$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$2

becomes strictly monotonic when

$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$3

Then each conserved energy has only one real propagating channel, so backward scattering is impossible and any localized potential becomes reflectionless regardless of shape (Longhi, 2017). This mechanism has no continuum analog because it relies on bounded group velocity and the failure of Galilean invariance on the lattice (Longhi, 2017).

A fourth mechanism is coherent multichannel cancellation. In coupled resonator optical waveguides with intra-resonator CW–CCW mixing, exact zero reflection occurs under

$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$4

yielding

$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$5

The reflectionless states are specific coherent CW/CCW superpositions, and the output chirality is selected by the input coefficients (2207.14453). This suggests a reflectionless mechanism based not on suppressing intermodal coupling but on balancing it for destructive interference.

5. Representative realizations across wave systems

The literature spans a broad range of physical platforms. The following table summarizes the main realizations described in the cited works.

Platform	Reflectionless object	Defining mechanism
SNAP optical fibers	Axial WGM envelope	Tuned $R_{\rm in}(\omega_0)\,a_{\rm in}=0.$6 reflectionless potential (Suchkov et al., 2015)
Discrete photonic lattice	Propagating lattice scattering states	Fast transverse drift $R_{\rm in}(\omega_0)\,a_{\rm in}=0.$7 (Longhi, 2017)
Microwave chaotic cavity	Reflectionless scattering modes	Real zero eigenvalue of $R_{\rm in}(\omega_0)\,a_{\rm in}=0.$8 (Sol et al., 2022)
Acoustic lined duct	Plane-wave to surface-mode conversion	Adiabatic liner-induced surface mode (Farooqui et al., 2018)
Open waveguide with obstacle	Reflectionless eigenstates	Opposite-sign complex scaling and non-selfadjoint spectrum (Dhia et al., 2018)
Stratified dielectric film	Plane-wave and beam scattering states	Kay–Moses inverse-scattering profile (Biswas et al., 2024)
Two-cavity optical limiter	Real reflection zeros	$R_{\rm in}(\omega_0)\,a_{\rm in}=0.$9-symmetric R-zero spectrum (Riboli et al., 2023)
Floquet-driven cavity	Synthetic reflectionless modes	Real zeros of auxiliary synthetic operator (Tuxbury et al., 5 Aug 2025)
Multiterminal Josephson junction	Zero-energy ABS-generating normal modes	Zero-energy reflectionless mode of $\det R_{\rm in}(\omega_0)=0$0 (Ohnmacht et al., 13 Mar 2025)

A distinct subfamily concerns programmable routing. In a highly overdamped four-port microwave cavity with 304 programmable metasurface elements, reflectionless scattering modes satisfy

$\det R_{\rm in}(\omega_0)=0$1

and can be functionalized for wavelength demultiplexing and multiport routing (Sol et al., 2022). In this context, “reflectionless” means no echo back into chosen launch ports rather than zero total scattering. The experiments report reflection suppression of at least $\det R_{\rm in}(\omega_0)=0$2 dB at two operating frequencies with undesired transmission suppression of at least $\det R_{\rm in}(\omega_0)=0$3 dB in a reflectionless demultiplexing task (Sol et al., 2022).

Another specialized realization appears in reciprocal photonic topological insulators. A rotating magnetic dipole source in a bianisotropic metawaveguide selectively excites one member of a reciprocal pair of topological edge states so strongly that the launched propagation is effectively unidirectional, with suppression of the undesired launched direction down to about one part in $\det R_{\rm in}(\omega_0)=0$4 (Xiao et al., 2016). This is source-selective reflectionless excitation rather than a globally nonreciprocal bulk phenomenon.

6. Topology, nonlinear effects, and current directions

Several recent works connect reflectionless modes to topology. In multiterminal Josephson junctions, a zero-energy reflectionless mode of the normal scattering matrix, defined by

$\det R_{\rm in}(\omega_0)=0$5

implies a unity transmission eigenvalue and generates a zero-energy Andreev bound state when the superconducting phases impose an effective $\det R_{\rm in}(\omega_0)=0$6-shift between channel sectors (Ohnmacht et al., 13 Mar 2025). These zero crossings form topological phase boundaries in ABS spectra and, in the four-terminal case, appear as Weyl nodes in the three-dimensional superconducting phase space (Ohnmacht et al., 13 Mar 2025). This suggests that normal-state reflectionless scattering provides a physically transparent origin for at least some topological singularities in multiterminal superconducting devices.

A complementary topological perspective is provided by Dirac-like formulations of Maxwell’s equations. For non-Hermitian stratified media with complex permittivity

$\det R_{\rm in}(\omega_0)=0$7

the field

$\det R_{\rm in}(\omega_0)=0$8

is an exact constant-amplitude scattering solution (Horsley, 2019). The asymptotic degree of $\det R_{\rm in}(\omega_0)=0$9 classifies the profile as reflectionless, coherent-perfect-absorbing, or lasing: $R_{\rm in}(\omega_0)$0 (Horsley, 2019). This suggests a topological organization of scattering behavior by asymptotic sign structure rather than by local impedance matching.

Nonlinearity complicates the picture. In scattering of one-dimensional quantum droplets by a linearly reflectionless Pöschl–Teller well,

$R_{\rm in}(\omega_0)$1

the nonlinear droplet can still undergo a sharp transition between full reflection and full transmission at a critical speed, despite the linear potential being reflectionless (Hu et al., 2023). The key distinction is that linear reflectionlessness does not automatically carry over to nonlinear self-bound objects. The same theme appears in cubic–quintic NLSE scattering, where “reflectionless” refers not to $R_{\rm in}(\omega_0)$2 but to negligible radiation loss; the outgoing state may still be fully transmitted or fully quantum reflected, with resonant switching between the two (Sakkaf et al., 2022). These results guard against a common overgeneralization: reflectionless linear-wave design is not equivalent to universal reflectionless nonlinear dynamics.

Current directions include dense multi-element architectures in SNAP fibers enabled by suppressed cross-talk (Suchkov et al., 2015), programmable reflectionless routing in strongly overlapping multimode cavities (Sol et al., 2022), quartically flattened frequency-conversion responses at synthetic reflectionless exceptional degeneracies (Tuxbury et al., 5 Aug 2025), and broader operator-theoretic frameworks in which reflectionless modes, trapped modes, and near-reflectionless complex eigenvalues are handled within a common non-selfadjoint spectral formalism (Dhia et al., 2018). A plausible implication is that reflectionless modes increasingly serve as a unifying design language across wave physics: not merely as isolated zero-reflection points, but as spectrally structured states linking scattering zeros, symmetry, topology, and functional wave control.