E58 · Surface spectromicroscopy

PEEM: Thresholds & Magnetic Domains

Can PEEM tell magnetic contrast apart from brightness caused by uneven illumination? Yes, once each image is corrected for the incoming flux. A region can also look bright because of its chemistry or work function. Normalize every image to its monitored flux, then compare images taken with opposite circular polarization. Only contrast that reverses when the polarization flips is magnetic.

Light the whole field of a synthetic surface. Compare UV work-function thresholds, absorption spectra and images taken with opposite helicities. Pull spectra out of the stored image stacks, and correct for the incident flux before you interpret any magnetic contrast.

Selected-region recorded counts $\text{—}$

Selected-region helicity asymmetry $\text{—}$

Conditional counting standard error $\text{—}$

Negative / positive flux monitor $\text{—}$

More measurements and target

Active contrast $\text{—}$

Selected photon energy $\text{—}$

Pixels in selected region $\text{—}$

Total two-helicity stack exposure $\text{—}$

Experiment target $\text{—}$

Simulated examples; each plot identifies its data source.

More examples

Contrast mode: 1 UV, 2 absorption, 3 XMCD

3

Negative / positive incident flux

1.5

Use incident monitors: 0 off, 1 on

0

Magnetization angle to photon projection

\alpha

0\,\mathrm{^\circ}

Advanced settings

Stored X-ray photon energy

h\nu

707.8\,\mathrm{eV}

Stored UV photon energy

h\nu_{UV}

4.7\,\mathrm{eV}

Exposure per energy and helicity

t

0.4\,\mathrm{s}

Assumed image FWHM

w

0.18\,\mathrm{\mu m}

Negative-helicity image displacement

\delta x

0\,\mathrm{\mu m}

Stored region center: horizontal

x_s

-1.5\,\mathrm{\mu m}

Stored region center: vertical

y_s

1.5\,\mathrm{\mu m}

Stored region radius

r_s

0.6\,\mathrm{\mu m}

Learning goal and model limits

Predict, then pick the signal that answers it

Can PEEM tell magnetic contrast apart from brightness caused by uneven illumination?

Target: at the phase-A resonance, use XMCD mode with monitor normalization. Select the upper-left domain without crossing its edge. Keep the image displacement at or below 0.05 micrometre and the image width at or below 0.3 micrometre, and collect more than 1,000 region counts with a counting error below 0.035. Align the magnetization with the photon projection. Flux correction cannot repair drift.

Model assumptions

This is a teaching model of UV-PEEM and of secondary-yield XAS/XMCD-PEEM, with an ideal 18 keV cathode image relay. The Lab does not do energy-filtered core-level XPS. The 8 micrometre square surface has two invented absorption resonances and work functions, and two opposite magnetic projections. A collection-yield ridge shared by both phases imitates topographic brightness. The UV yield follows a quadratic threshold at zero temperature. The X-ray spectra are assumed Gaussian resonances, not computed atomic cross sections. Both helicity stacks hold sampled detector counts after a Gaussian spatial blur, and their flux monitors read exact teaching values. Drift of the negative-helicity image misregisters the two images. Changing the region or the normalization reuses the stored counts. Gray pixels in the asymmetry map have fewer than five weighted counts. Each image has its own logarithmic color scale; the asymmetry scale is linear. The 0.2 micrometre pixels limit spatial sampling. The Lab does not solve the electrostatic field and does not include material calibration, damage, space charge or reconstruction of the magnetic vector.

Physics tutorial

PEEM: Thresholds & Magnetic Domains

BackgroundLight falls on the whole field of a synthetic surface at once. You compare UV work-function thresholds, absorption spectra and images taken with opposite helicities, pull spectra out of the stored image stacks and correct for the incident flux before interpreting any magnetic contrast.

Why it mattersThe task is to measure the magnetic contrast of one domain. Use XMCD mode with monitor normalization at the phase-A resonance, and keep the selected region inside the upper-left domain without crossing an edge. Hold the image displacement to 0.05 micrometre or less and the image width to 0.3 micrometre or less, collect more than 1,000 counts in the region and bring the counting error below 0.035. Align the magnetization with the photon projection. Flux correction cannot repair drift.

Start with the essentials

Focus question: Can PEEM tell magnetic contrast apart from brightness caused by uneven illumination?
One-sentence intuition: Chemical, threshold and magnetic contrast each depend on different experimental settings. Correct for flux and look at local spectra before you decide why a region is bright.

Core mathematical model

Photoemission yield near the UV threshold

Y_{UV}\propto(h\nu-\Phi)_+^2

This near-threshold yield assumes zero temperature. Image brightness also depends on how well the electrons are collected.

Intensities divided by the flux monitor

I_\pm=\frac{C_\pm}{F_\pm}

Dividing the counts by the incident exposure removes the flux imbalance. The two images still have unequal counting precision.

Asymmetry between the two helicities

A=\frac{I_+-I_-}{I_++I_-}

Under the assumed response, this ratio measures the dichroism along the beam projection. It does not give the full magnetic vector.

Common difficulties

Brightness needs a model

Typical misconceptionIf a region looks bright, I can tell directly what makes it bright.

Better mental modelChemical, threshold and magnetic contrast each depend on different experimental settings. Correct for flux and inspect the local spectra before you assign a bright region to a physical cause.

Run the experiment

01
Predict the false contrast
Look at the unequal-flux preset, then turn on monitor normalization.
What to observe: Unequal illumination shifts the raw ratio. After normalization, the upper and lower domains show opposite signs again.
02
Separate chemistry from threshold
Compare the absorption and UV presets. Move the region across the chemical boundary.
What to observe: The two phases have different resonance positions and emission thresholds. The bright ridge they share is not a map of chemical concentration.
03
Test the magnetic interpretation
Try low counts, drift and transverse magnetization before checking the target.
What to observe: Noise, misregistered images or a small projection on the beam can each make the target fail.