Elliptic Microwave Polarization

• Elliptic microwave polarization is the state in which the transverse electric field of a wave traces an ellipse, defined by amplitude and phase relationships.
• It encompasses geometric (Stokes and Poincaré) descriptions, polarization-angle conventions, and challenges in instrument calibration and systematics control.
• Recent experimental approaches, including antenna synthesis and in situ atomic spectroscopy, achieve high-fidelity polarization control and mitigate cross-polar leakage.

Searching arXiv for recent and relevant papers on elliptic microwave polarization, polarization-angle conventions, microwave beam polarization, and CMB polarization systematics. Elliptic microwave polarization is the general polarization state of a microwave electromagnetic wave in which the transverse electric field traces an ellipse in time at a fixed point in space. In astronomical and microwave polarimetry, the relevant ellipse is defined in the plane orthogonal to the line of sight, and the reported polarization angle is the position angle of the major axis of that polarization ellipse (Alighieri, 2016). Across microwave science, the topic spans at least three interconnected domains: the geometric and Stokes-parameter description of polarization states; instrumental generation, control, and measurement of nearly linear, circular, or arbitrary superposed states; and propagation or optical-system effects that rotate, distort, or partially convert polarization, thereby producing effective ellipticity in the observed or detected field (Yuan et al., 2023, Kurdak et al., 9 Aug 2025, Ren et al., 7 Jan 2026).

1. Geometric definition and polarization-state description

For an electromagnetic wave propagating along a chosen axis, the transverse field components can be written in standard form as

$$E_x(t) = a_x \cos(\omega t), \quad E_y(t) = a_y \cos(\omega t + \delta),$$

with amplitudes $a_x, a_y$ and relative phase $\delta$. The tip of the vector $(E_x(t),E_y(t))$ then traces an ellipse. Linear polarization is recovered when the ellipse collapses to a line, and circular polarization is the special case of equal orthogonal amplitudes with a quadrature phase relation (Alighieri, 2016, Yuan et al., 2023).

In astronomical usage, the polarization angle, often denoted $\psi$, is defined by the orientation of the major axis of the polarization ellipse rather than by the instantaneous electric-field direction. The convention emphasized in microwave astronomy is that “in the more general case of elliptical polarization, PA is the position angle of the major axis of the polarization ellipse” (Alighieri, 2016). Because the major axis is an unoriented line, the angle is defined modulo $180^\circ$.

The Stokes formalism provides the standard state description. In the usual convention,

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$

where $I$ is total intensity, $p$ is the fractional linear polarization, and $\psi$ is the polarization position angle (Alighieri, 2016). Equivalently,

$a_x, a_y$0

with appropriate quadrant handling and mapping to $a_x, a_y$1 (Alighieri, 2016). In this representation, $a_x, a_y$2 and $a_x, a_y$3 encode the linear part of the ellipse, while $a_x, a_y$4 encodes circular polarization or handedness. A direct implication is that two waves with the same $a_x, a_y$5 but opposite $a_x, a_y$6 have the same major-axis orientation and therefore the same polarization angle, while differing in the sense of rotation of the field about that axis (Alighieri, 2016).

A complementary representation uses the Poincaré sphere, on which a fully polarized state is specified by a position angle $a_x, a_y$7 and an ellipticity angle $a_x, a_y$8. In the statistical treatment of polarization-angle estimation, the ellipticity angle is written

$a_x, a_y$9

with $\delta$0, so that $\delta$1 corresponds to purely linear polarization and $\delta$2 to purely circular polarization (McKinnon, 25 Aug 2025). The axis ratio of the polarization ellipse satisfies

$\delta$3

where $\delta$4 and $\delta$5 are the semi-major and semi-minor axes (McKinnon, 25 Aug 2025).

2. Polarization-angle conventions in microwave astronomy

The modern microwave literature distinguishes sharply between the official IAU convention and the opposite convention introduced in WMAP and propagated through much of CMB polarization analysis. The IAU convention, enforced in 1973, measures the position angle from North toward East, increasing counter-clockwise when looking at the source (Alighieri, 2016). In this convention, the sky-plane coordinate frame is right-handed with $\delta$6 along declination, $\delta$7 along right ascension, and $\delta$8 along the direction of propagation from the source to the observer (Alighieri, 2016).

The WMAP/CMB convention adopts the opposite geometrical sense: the polarization angle is measured starting from South and increasing clockwise when looking at the source (Alighieri, 2016). This corresponds to a frame in which the $\delta$9-axis points toward South and the $(E_x(t),E_y(t))$0-axis is reversed relative to the IAU propagation convention. The practical consequence is a sign change in one of the linear Stokes parameters: $(E_x(t),E_y(t))$1 Thus the two conventions differ by a reversal of the sign of $(E_x(t),E_y(t))$2, and polarization angles are correspondingly related by $(E_x(t),E_y(t))$3 (Alighieri, 2016).

This distinction is not a semantic matter. Mixed use of the two conventions can mirror vector orientations on polarization maps and can mislead interpretation of derived quantities. The discussion in the conventions paper explicitly notes that mixing measurements made under different conventions produced “the wrong impression that most measurements are negative” in an early figure on cosmic polarization rotation (Alighieri, 2016). The same issue propagates into any analysis that depends on angle-consistent $(E_x(t),E_y(t))$4 combinations, including E/B decomposition and interpretation of foreground polarization geometry.

The practical transformation is simple. To convert WMAP/CMB-style data to the IAU/IEEE convention, no full software rewrite is required; one changes the sign of the reduced-data $(E_x(t),E_y(t))$5 parameter: $(E_x(t),E_y(t))$6 while $(E_x(t),E_y(t))$7 and $(E_x(t),E_y(t))$8 are unchanged (Alighieri, 2016). This leaves $(E_x(t),E_y(t))$9 invariant and therefore preserves the linear polarization fraction while changing the reported orientation to the official astronomical convention.

3. Instrumental realization: co-polar, cross-polar, and effective beam ellipticity

In microwave instrumentation, ellipticity is not only a property of the incoming wave. It also appears as an effective property of the receiving or transmitting system when a nominally linear device responds imperfectly to the orthogonal polarization. A concrete example is provided by the polarized beam characterization of the PIXIE multi-moded concentrator (Kogut et al., 2018).

That work distinguishes co-polar and cross-polar beam patterns. The co-polar beam is the response when incident linear polarization matches the detector’s nominal polarization; the cross-polar beam is the response to the orthogonal incident polarization (Kogut et al., 2018). In an ideal linear polarimeter, the cross-polar beam would vanish everywhere. In the measured multi-moded system, the cross-polar response is nonzero, implying partial polarization mixing and an effective departure from pure linear response.

The paper quantifies this behavior through the polarization efficiency

$\psi$0

measured to be $\psi$1 at $\psi$2 GHz, $\psi$3 at $\psi$4 GHz, and $\psi$5 at $\psi$6 GHz (Kogut et al., 2018). This corresponds to about $\psi$7–$\psi$8 leakage into the orthogonal polarization channel (Kogut et al., 2018). The same study describes the co-polar and cross-polar patterns as flat-topped beams with a few-dB tilt along one instrument axis due to $\psi$9 off-axis illumination, and it identifies low-azimuthal-order structure in the cross-polar beam, especially a dominant spin-1 component (Kogut et al., 2018).

This instrumental ellipticity has direct consequences for CMB polarization systematics. Convolving the measured cross-polar beam with simulated CMB polarization maps, the uncorrected spin-2 systematic is found to have

$180^\circ$0

corresponding to an apparent inflationary $180^\circ$1-mode amplitude

$180^\circ$2

(Kogut et al., 2018). The same paper shows that a template formed by convolving the measured cross-polar beam with CMB E-mode maps reduces the systematic to negligible levels (Kogut et al., 2018). This suggests that modest beam ellipticity is tolerable if the response is well characterized and the dominant contamination mode is separable from the spin-2 cosmological signal.

4. Generation and control of elliptic microwave polarization in quantum platforms

Microwave polarization control is a central experimental capability in atomic and molecular physics, where transitions are polarization selective. Two complementary architectures illustrate how nearly pure circular states and arbitrary superpositions are engineered and diagnosed.

A planar cloverleaf antenna designed at $180^\circ$3 GHz realizes circularly polarized microwaves by superposing two orthogonal linear fields with controlled relative phase (Yuan et al., 2023). The antenna comprises four one-wavelength elliptical loop antennas in a planar cloverleaf arrangement. Loop pair A–C generates a horizontal component $180^\circ$4, loop pair B–D a vertical component $180^\circ$5, and the overall polarization is tuned by controlling relative feed phases. The familiar synthesis conditions apply: equal amplitudes and a phase difference of $180^\circ$6 yield circular polarization; deviations in amplitude balance or phase shift yield elliptical polarization (Yuan et al., 2023).

Using ultracold NaCs molecules as a polarization-selective quantum sensor, the study reports a Rabi frequency of $180^\circ$7 MHz on a $180^\circ$8 transition, corresponding to an electric field amplitude of $180^\circ$9 V/cm at $Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$0 mm from the antenna (Yuan et al., 2023). The measured lab-frame field ratios are

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$1

from which the inferred microwave-frame polarization ellipticity is

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$2

corresponding to a $Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$3 dB suppression of the undesired circular component (Yuan et al., 2023). The paper adopts a handedness convention in which left-handed polarization is counterclockwise as observed from the source, and it defines ellipticity in the microwave frame by

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$4

(Yuan et al., 2023).

A more general control architecture appears in a Rydberg-ensemble experiment that treats the microwave field at the atoms as a complex vector in the spherical basis $Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$5 (Kurdak et al., 9 Aug 2025). There the field is written

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$6

and three in-vacuum electrodes act as independently driven microwave sources. Their complex drive amplitudes $Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$7 are related to the field components through a calibrated chamber matrix

$Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$8

so that arbitrary target polarization vectors can be synthesized by solving $Q = I\,p \cos 2\psi, \quad U = I\,p \sin 2\psi,$9 (Kurdak et al., 9 Aug 2025). The work demonstrates purified $I$0, $I$1, and $I$2 fields with fidelities $I$3, $I$4, and $I$5, respectively, and explicitly notes that arbitrary superpositions follow by linearity (Kurdak et al., 9 Aug 2025). This implies direct synthesis of arbitrary elliptical microwave polarization once the basis vectors are under high-fidelity control.

These experiments also illustrate a broader methodological point: in strongly reflective microwave environments, polarization is often more reliably characterized in situ by atomic or molecular spectroscopy than by geometric field modeling alone (Yuan et al., 2023, Kurdak et al., 9 Aug 2025).

5. Measurement and inference: from linear angle readout to full ellipse statistics

Microwave polarimetry encompasses both direct state measurement and inference of polarization geometry from calibrated observables. In a Rydberg atom-based mixer, polarization direction can be determined in a single measurement by using two orthogonally polarized local microwave fields (Yin et al., 2024). In that system, the signal field is decomposed onto two orthogonal linear local oscillators, and the beat-note amplitudes satisfy

$I$6

so that with matched local fields

$I$7

for a linearly polarized signal (Yin et al., 2024). A weak static magnetic field then uses Zeeman-induced asymmetry to distinguish $I$8 from $I$9, enabling real-time determination of polarization direction over $p$0–$p$1 (Yin et al., 2024). The paper treats linear polarization, but its own discussion states that the phase-sensitive mixer structure could in principle be extended to full elliptic polarization analysis by determining amplitudes and relative phase of orthogonal components (Yin et al., 2024).

At the instrument level, a microwave polarimeter with near-infrared up-conversion is calibrated by fitting angle and polarization-efficiency errors with sinusoidal functions (Casas et al., 2022). Operating from $p$2 to $p$3 GHz, the demonstrator was calibrated with a linearly polarized source of variable angle, and after using five fitting terms it achieved maximum residual polarization-angle errors of about $p$4–$p$5 depending on output and maximum residual polarization-percentage errors of about $p$6–$p$7 (Casas et al., 2022). Although the calibration source was linear, the Stokes-based architecture is directly relevant to arbitrary elliptical states because ellipse orientation and ellipticity are encoded in the same calibrated observables (Casas et al., 2022).

A separate statistical issue concerns estimation of the ellipticity angle itself. For a fixed polarization amplitude with Gaussian noise in $p$8, the ellipticity angle is

$p$9

and its measured distribution is generally asymmetric, unlike the better-known position-angle distribution (McKinnon, 25 Aug 2025). The same study derives that the standard deviation and confidence limits of the ellipticity angle vary inversely with polarization SNR, but depend on the intrinsic ellipticity angle, and that the measured ellipticity angle is noise-biased, particularly at low SNR and large intrinsic $\psi$0 (McKinnon, 25 Aug 2025). A plausible implication is that in weakly polarized microwave sources or circularly dominated states, reporting a single symmetric error bar on ellipticity can be misleading unless the full asymmetric statistics are used.

6. Propagation, optics, and induced ellipticity in microwave astronomy

Microwave polarization can be modified not only at the source or by the receiver but also during propagation through matter or through the optical train of an experiment. In refractive CMB telescopes, differential transmission in anti-reflection coatings produces polarization-dependent amplitude and phase shifts. A hybrid physical-optics method that incorporates full-wave electromagnetic simulations of AR coatings shows that such differential transmission rotates the far-field polarization angle by $\psi$1–$\psi$2 across the focal plane, with $\psi$3 near the design frequency of a single-layer quarter-wave coating but up to $\psi$4 toward band edges and off-axis feeds (Ren et al., 7 Jan 2026).

The relevant local description is a Jones matrix

$\psi$5

where $\psi$6 and $\psi$7 are complex transmission coefficients for the two orthogonal polarization states (Ren et al., 7 Jan 2026). Differential amplitude rotates the effective major axis after aperture integration, while differential phase implies a nonzero phase difference between orthogonal components and therefore, in general, a weakly elliptical output state. The paper emphasizes rotation and $\psi$8 mixing rather than explicit ellipticity maps, but it states that if ignored in analysis the effect can produce temperature-to-polarization leakage and Stokes $\psi$9 mixing (Ren et al., 7 Jan 2026). For an off-axis feed with $a_x, a_y$00 polarization offset at $a_x, a_y$01 GHz, the beam-averaged temperature-to-polarization leakage is $a_x, a_y$02 (Ren et al., 7 Jan 2026).

An astrophysical propagation mechanism appears in the study of one-loop vacuum polarization by milli-charged fermions in cosmic magnetic fields (Ejlli, 2017). Starting from density-matrix evolution for the Stokes vector, that work shows that linearly polarized CMB radiation at decoupling can acquire circular polarization and rotated linear polarization during propagation, thereby becoming elliptically polarized (Ejlli, 2017). In the best scenario, the acquired degree of circular polarization can be

$a_x, a_y$03

depending on $a_x, a_y$04 and observing frequency (Ejlli, 2017). The same mechanism can generate polarization even when the CMB is initially in thermal equilibrium (Ejlli, 2017). This suggests that in cosmological contexts, elliptic microwave polarization can be both an instrumental systematic and a potential signature of beyond-standard-model propagation physics.

7. Far-from-circular fields and polarization tolerances in molecular shielding

A distinct use of elliptic microwave polarization appears in ultracold polar-molecule shielding. Conventional microwave shielding works best with nearly pure circular polarization, but one study shows that effective shielding can still be achieved with elliptical polarization that is around $a_x, a_y$05 circular (Karman et al., 2019). In that analysis, the field in the circular basis is parameterized as

$a_x, a_y$06

so that $a_x, a_y$07 is pure circular and $a_x, a_y$08 is linear polarization (Karman et al., 2019). Coupled-channels calculations for RbCs+RbCs and CaF+CaF collisions show that shielding remains effective for $a_x, a_y$09, corresponding to a power ratio around $a_x, a_y$10 dB between the desired and unwanted circular components (Karman et al., 2019).

A related and more general proposal reframes a far-from-circular field as a perfectly circular component in one plane plus a linear component along a perpendicular direction (Karman, 2019). By applying a modest static electric field along that perpendicular axis, the problematic linear-coupled molecular state is Stark-shifted out of resonance, and the effective shielding Hamiltonian reduces to the circularly polarized case (Karman, 2019). The work reports that with $a_x, a_y$11 kV/cm, effective shielding is possible up to $a_x, a_y$12, and for RbCs at $a_x, a_y$13K loss rates around $a_x, a_y$14 are achievable near $a_x, a_y$15 MHz and $a_x, a_y$16 kV/cm (Karman, 2019). This suggests that in strongly interacting molecular systems, ellipticity can be mitigated not only by improving polarization purity but by redesigning the dressed-state structure around the actual polarization ellipse.

A common misconception is that any significant deviation from circular polarization necessarily destroys microwave shielding. The later far-from-circular analysis does not support that blanket statement; instead it identifies a modified geometry in which highly eccentric fields remain usable if the static field is aligned with the axis singled out by the polarization decomposition (Karman, 2019). The earlier requirement for approximately $a_x, a_y$17 circular polarization therefore applies to the original shielding architecture, not to all microwave-shielding schemes (Karman et al., 2019, Karman, 2019).

8. Broader significance and unresolved issues

Elliptic microwave polarization is a unifying concept across microwave astronomy, quantum control, molecular collision engineering, and polarimetric metrology. In each domain, the same mathematical object—the polarization ellipse or its Stokes/Poincaré representation—appears, but the operational meaning differs. In astronomical data analysis, the critical issue is often the orientation of the major axis and the convention used to report it (Alighieri, 2016). In instrument design, the practical concern is how beam asymmetry, cross-polar response, and optical coatings produce effective angle rotation or partial conversion between nominal polarization channels (Kogut et al., 2018, Ren et al., 7 Jan 2026). In quantum platforms, the same formalism becomes an overview and control problem in the spherical basis $a_x, a_y$18 (Kurdak et al., 9 Aug 2025, Yuan et al., 2023).

Several technical themes recur. One is the separation between orientation and handedness: $a_x, a_y$19 determine the major-axis geometry, while $a_x, a_y$20 determines the sense of rotation and therefore the ellipticity angle (Alighieri, 2016, McKinnon, 25 Aug 2025). Another is that small instrumental nonidealities can be cosmologically important; beam cross-polarization at the few-percent level can already correspond to $a_x, a_y$21 if uncorrected (Kogut et al., 2018), and polarization-angle rotations of order $a_x, a_y$22–$a_x, a_y$23 can generate non-negligible $a_x, a_y$24 mixing and temperature leakage in refractive CMB optics (Ren et al., 7 Jan 2026). A third is that polarization purity is often easier to define than to realize. Reflection-dominated environments, finite-mode optical systems, and basis-convention mismatches all act to obscure the intended polarization state unless the full system response is calibrated in situ (Alighieri, 2016, Kogut et al., 2018, Kurdak et al., 9 Aug 2025).

This suggests a useful synthesis. “Elliptic microwave polarization” is not only the general free-space polarization state of a microwave field; it is also the effective state created by system-level transformations that alter amplitude and phase relationships between orthogonal components. In that broader sense, the subject includes the geometry of the polarization ellipse, its measurement conventions, its statistical inference under noise, its deliberate synthesis in controlled experiments, and its inadvertent generation by propagation media and optical hardware.