Scanning Polar MOKE Microscopy

Updated 17 April 2026

Scanning polar MOKE microscopy is a high-sensitivity optical imaging method that maps out-of-plane magnetization in thin films, heterostructures, and nanomaterials.
The technique utilizes precise polarization optics and detection schemes (confocal, Sagnac, pump–probe, and near-field) to achieve sub-micrometer spatial resolution and picosecond temporal resolution.
Advancements in cryogenic setups, plasmonic enhancements, and interferometric methods are expanding its application in spintronics, topological magnetism, and ultrafast spin dynamics.

Scanning polar magneto-optical Kerr effect (MOKE) microscopy is a high-sensitivity, spatially-resolved optical technique for imaging out-of-plane magnetization structures in thin films, heterostructures, and nanomaterials. It exploits the polar Kerr effect—rotation and ellipticity induced in linearly polarized light reflected from a magnetized surface perpendicular to the plane—using raster-scanned focused probes or near-field tips, with implementations spanning cryogenic–room temperature, picosecond time scales, and sub-micrometer spatial resolution. The methodology is central to research in spintronics, topological magnetism, ultrafast spin dynamics, and correlated electron systems.

1. Fundamental Principles of Polar MOKE Microscopy

In polar MOKE microscopy, the local magnetization $M_z$ normal to the sample induces off-diagonal elements in the dielectric tensor, modifying the reflection coefficients for s- and p-polarized light. The fundamental MOKE observable is the Kerr rotation $\theta_K$ , given at near-normal incidence by

$\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$

where $r_{pp}$ and $r_{ps}$ are the respective Fresnel reflection coefficients. For small $\theta_K$ , the reflected polarization is rotated and partially elliptic, yielding a differential intensity at the detector after an analyzer:

$I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$

The signal is directly proportional to $M_z(x,y)$ , so spatially resolved detection reconstructs local magnetization patterns (Lange et al., 2017).

A defining feature is the immunity to stray fields and $M_{xy}$ crosstalk in pure polar geometry, enabling clean mapping of $M_z$ in PMA (perpendicular magnetic anisotropy) systems, chiral magnets, and compensated antiferromagnets (Farhang et al., 13 Jul 2025). The detection physics is valid across continuous-wave, pulsed, and interferometric probe schemes.

2. Optical Instrumentation and Detection Architectures

Scanning polar MOKE platforms can be categorized into confocal/focused beam systems, Sagnac interferometers, pump–probe modalities, and near-field (aperture-tip/cantilever) instruments.

2.1 Confocal and Widefield Polarizing Microscopes

High-resolution setups employ a fiber-coupled diode laser ( $\theta_K$ 0 nm, $\theta_K$ 1 mW), precision polarization optics (Glan–Thompson polarizer, λ/4 and λ/2 plates), and a high-numerical-aperture ( $\theta_K$ 2) objective in a confocal design. Beam scanning is achieved via a fast-steering mirror with $\theta_K$ 3 range and $\theta_K$ 4rad step (telecentric image formation, $\theta_K$ 5m $\theta_K$ 6 field) (Lange et al., 2017).

A Wollaston prism splits the reflection into orthogonal polarizations, measured with a four-quadrant photodiode, digitized and demodulated (lock-in detected) at modulation frequencies up to 1 MHz. Balanced detection strongly rejects common-mode noise and enables sensitivities $\theta_K$ 7 rad/ $\theta_K$ 8 (confocal) and $\theta_K$ 9 rad/ $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 0 (widefield). Full-field acquisition is possible with sCMOS or CCD cameras.

2.2 Sagnac Interferometer-based Microscopes

The scanning Sagnac interferometer architecture employs an all-fiber loop (PM fibers, thermal isolation, electro-optic modulation) to encode the polar Kerr rotation as a differential phase between counter-propagating beams. With $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 1 nm or $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 2 nm lasers, the zero-area loop ensures only time-reversal symmetry breaking signals (i.e., true Kerr rotation) survive detection (Fried et al., 2014, Farhang et al., 13 Jul 2025).

The detected interference at the output photodiode is

$\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 3

where $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 4. Modulation and demodulation at first and second harmonics of the EOM drive yield $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 5 and $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 6, from which

$\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 7

With sub- $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 8rad noise floors and minimal drift ( $\theta_K \simeq \mathrm{Re}\left(\frac{r_{ps}}{r_{pp}}\right), \qquad \psi_K \simeq \mathrm{Im}\left(\frac{r_{ps}}{r_{pp}}\right)$ 9rad per 84 hours), Sagnac configurations enable $r_{pp}$ 00.01 µrad/ $r_{pp}$ 1 shot-noise-limited Kerr angle sensitivity and near-theoretical spatial resolution $r_{pp}$ 2m with $r_{pp}$ 3 (Farhang et al., 13 Jul 2025).

2.3 Time-Resolved and Pump–Probe MOKE

Ultrafast applications use supercontinuum fiber-laser sources (400–1600 nm, sub-ps pulses, repetition rates $r_{pp}$ 4 MHz) for two-color pump–probe microscopy. Spectral and spatial filtering permit independent tuning of pump and probe arms; scanning is performed either by moving the probe focus or sample. Balanced bridge detection, combined with lock-in demodulation (e.g., PEM-modulated at $r_{pp}$ 5 kHz), delivers picosecond time and $r_{pp}$ 6m spatial resolution across 8–300 K (Henn et al., 2013).

2.4 Near-field Scanning Kerr Microscopy

Sub-diffraction-limited imaging is achieved by integrating a metallic AFM tip with a FIB-milled nanoscale aperture ( $r_{pp}$ 7 nm). The tip guides focused optical pulses to the sample, producing a near-field spot with $r_{pp}$ 8 nm FWHM (approaching the aperture limit, not NA/diffraction) (Keatley et al., 2017). The system preserves $r_{pp}$ 91 mdeg polar Kerr signal for $r_{ps}$ 0 tip–sample spacings. Finite-element simulations indicate localized field concentration and possible plasmonic enhancement of the near-field Kerr effect.

3. Cryogenic, Field, and Scanning Environments

Most state-of-the-art scanning polar MOKE systems operate in cryogenic vacuum cryostats ( $r_{ps}$ 1 mbar, $r_{ps}$ 2–300 K), with integration to $r_{ps}$ 3He flow or liquid-He environments (Lange et al., 2017, Henn et al., 2013). Sample mounts accommodate translation and piezoelectric nanopositioners for $r_{ps}$ 4m scan ranges and $r_{ps}$ 5 nm step resolution. Magnetic field control is realized via rotatable electromagnets (up to $r_{ps}$ 6 mT in-plane, $r_{ps}$ 7 mT out-of-plane), water-cooled pole pieces, and Helmholtz coils allowing precise angular field sweeps and pulsed field protocols (Lange et al., 2017, Karim et al., 5 Feb 2026).

System stability is ensured by battery or temperature-stabilized laser/electronics, differential reference measurements, and field/temperature calibration. Closed-loop feedback and synchronization of scan, camera, and field drive are essential for multi-modal and time-dependent studies (Karim et al., 5 Feb 2026).

4. Performance Metrics and Image Analysis

Spatial resolution ( $r_{ps}$ 8) is dictated by optical NA or near-field aperture:

Diffraction-limited: $r_{ps}$ 9 (widefield), $\theta_K$ 0 (confocal).
Experimentally verified $\theta_K$ 1 nm at $\theta_K$ 2 nm, NA $\theta_K$ 3; near-field FWHM $\theta_K$ 4 nm with $\theta_K$ 5 nm aperture (Lange et al., 2017, Keatley et al., 2017).

Kerr sensitivity:

Confocal: $\theta_K$ 6 rad/ $\theta_K$ 7
Sagnac (fiber): $\theta_K$ 8 µrad/ $\theta_K$ 9 (practical, $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 0 µrad/ $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 1) (Farhang et al., 13 Jul 2025, Fried et al., 2014).
Near-field: $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 2 mdeg for $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 3 ms integration per point (Keatley et al., 2017).

Imaging speed and dwell time: $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 4– $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 5 Hz full frames (confocal/piezo scan), $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 6 ms dwell times allow high throughput; up to $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 7 fps in sCMOS widefield mode.

Data analysis includes:

Flat-field correction and background subtraction ( $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 8 reference, dark count, lock-in baseline)
Intensity normalization $I(\beta) = I_0 \sin^2(\beta + \theta_K) \approx I_0[\sin^2\beta + 2\sin\beta \cos\beta\,\theta_K]$ 9 for quantitative $M_z(x,y)$ 0 mapping
Extraction of $M_z(x,y)$ 1 via material Kerr constants
Domain, wall, and switching analysis (cross-correlation, centroid tracking, Sobel operators)
Magnetization dynamics from $M_z(x,y)$ 2 traces (damped sinusoids), spatial profiles (diffusion equations), and domain-wall creep models (Karim et al., 5 Feb 2026, Henn et al., 2013).

5. Applications in Magnetic Materials and Spin Systems

Scanning polar MOKE microscopy uniquely resolves $M_z(x,y)$ 3 in heterogeneous, nanoscale, and ultrafast magnet systems:

PMA ferromagnets (e.g., Pt/CoFeB/Ru): imaging domain nucleation, wall propagation, DMI-stabilized chiral Néel walls, angular-dependent switching, and wall creep (Karim et al., 5 Feb 2026).
Topological antiferromagnets: direct spatial mapping of domains with quantized scalar spin chirality in zero net-moment, SOC-free Co $M_z(x,y)$ 4TaS $M_z(x,y)$ 5. Sagnac-based imaging revealed $M_z(x,y)$ 6– $M_z(x,y)$ 7 µrad, resolving mesoscale chirality domains and switching under applied field (Farhang et al., 13 Jul 2025).
Superconductors: Meissner and vortex imaging with MO indicator films; beam-induced voltage mapping in high-T $M_z(x,y)$ 8 and 2DEG systems (Lange et al., 2017).
Ultrafast spin dynamics: pump–probe MOKE resolves $M_z(x,y)$ 91 ps spin dephasing, ballistic/diffusive transport, and precessional magnetization in semiconductors and 2D materials (Henn et al., 2013).
Sub-diffraction magnetization mapping: near-field MOKE provides $M_{xy}$ 0 nm spatial resolution in confined micro/nanomagnets (Keatley et al., 2017).

6. Limitations, Technical Trade-offs, and Future Directions

The spatial resolution is fundamentally limited by the probe wavelength and NA, or near-field aperture geometry. Near-field schemes yield resolution below the diffraction limit but at the expense of reduced throughput (Bethe’s $M_{xy}$ 1 scaling) and alignment sensitivity. Plasmonic enhancement at the aperture vicinity could offer partial compensation (Keatley et al., 2017).

Sagnac interferometry virtually eliminates reciprocal background, with stability limited by fiber thermal drift, residual amplitude modulation in EOMs, and detector noise. Polarizing microscope platforms must correct for birefringence, stress-optic effects, and analyzer misalignments to realize quantitative $M_{xy}$ 2 mapping.

Emerging directions include THz-probed Kerr microscopy (for quantized topological MOKE), adaptation to shorter wavelengths for higher resolution, integration with ultrafast pulse sequences for study of non-equilibrium magnetization phenomena, and extension to quantum materials with compensated or exotic spin order (Farhang et al., 13 Jul 2025).

A plausible implication is that further advances in tip engineering, detector sensitivity, and field/temperature control will permit true nanoscale, single-spin, and coherent quantum-state resolved MOKE imaging.

7. Summary Table: Core System Performance Parameters

Instrument Type	λ (nm)	Spatial Res. (µm)	Kerr Sensitivity ( $M_{xy}$ 3)	Scan Range / Speed	Reference
Confocal polarizing	405	0.24	$M_{xy}$ 4	$M_{xy}$ 5m @ 1 Hz	(Lange et al., 2017)
Sagnac interferometer	820/1550	1.5–2.0	$M_{xy}$ 6rad	$M_{xy}$ 7m @ 0.1–1 Hz	(Fried et al., 2014, Farhang et al., 13 Jul 2025)
Pump–probe (ultrafast)	400–1600	2	$M_{xy}$ 8 (lock-in)	$M_{xy}$ 9m @ 1–10 Hz	(Henn et al., 2013)
Near-field aperture	800	0.55	$M_z$ 0 mdeg ( $M_z$ 1)	$M_z$ 2m @ 0.002 Hz	(Keatley et al., 2017)
Widefield camera	630	1.8	$M_z$ 3	$M_z$ 4m @ 30 Hz	(Karim et al., 5 Feb 2026)

All metrics correspond to representative best values attained in the cited works; practical performance depends on system optimization and operating parameters.

Scanning polar MOKE microscopy is thus established as a pivotal tool for quantitative, high-resolution mapping of out-of-plane magnetization, domain structures, and ultrafast spin phenomena across a broad range of condensed-matter systems. Continued innovations in optical design, detection, and environmental control are expanding its power and versatility in the study of emergent magnetic phenomena.