Near-field optical microscopy can reveal how light behaves around features too small to resolve with conventional far-field imaging. For industrial teams developing nanoscale photonic and plasmonic devices, however, collecting a high-resolution image is only part of the metrology challenge. Measurements must also be interpreted in a way that can be compared with electromagnetic simulations and applied to as-fabricated devices. Work on scattering scanning near-field optical microscopy (sSNOM) for plasmonic antennas in heat-assisted magnetic recording (HAMR) demonstrates how modeling the measurement process can help close that gap.
Why characterize the near field?
Plasmonic devices use structures that confine optical energy to regions smaller than the wavelength of the incident light. In a HAMR write head, a waveguide-coupled plasmonic antenna concentrates light into an intense, rapidly decaying near field. That energy heats a small area of magnetic recording media, enabling its magnetization to be changed under an applied magnetic field. The size and distribution of the optical hot spot are therefore important characteristics of the device.
Electromagnetic simulation can predict a device’s optical field, but a manufactured part may vary from its design. Experimental characterization is needed to check predictions and investigate the behavior of as-fabricated structures. This is especially significant when devices are produced in large quantities and measurement is intended to inform process quality control. The same basic need arises in other applications where device operation depends on light near nanoscale features, including photonics and sensing.
How sSNOM measures plasmonic devices
In sSNOM, an atomic force microscopy (AFM) tip scans above the sample while interacting with its optical near field. The tip scatters light into the far field, where it can be collected and detected. Unlike aperture-based near-field microscopy, this approach uses the sharpness of the AFM tip to set spatial resolution rather than relying on a small aperture to transmit light. The study used uncoated silicon tips with typical radii of approximately 5–15 nm.
The detected signal includes background light as well as light associated with the tip-sample near-field interaction. To distinguish the latter, the measurement system detects signal components at the fundamental frequency and higher harmonics of the oscillating AFM cantilever. Higher-order harmonics have a stronger dependence on near-field strength and can provide greater near-field contrast. Consequently, the signal at each harmonic is not simply a direct picture of the electric field at the sample surface.
Modeling the measurement, not just the device
To compare measurements with simulation, the researchers developed a model that translates simulated electric-field data into expected sSNOM harmonic maps. A simple starting approximation treated the tip motion as sinusoidal, the field as decaying exponentially away from the surface, and scattering as proportional to intensity. This offered a basic picture of why the signal changes during each cantilever oscillation, but it did not account for the real antenna field or the probe’s interaction with the sample.
The more detailed model incorporated the simulated field and a scattering parameter based on the interaction between a spherical probe and an image dipole. The calculation also accounted for the material beneath each part of the probe, distinguishing among gold, oxide, and magnetic material. To represent the AFM tip’s shape, the modeled field was convolved with separate descriptions of the conical shaft and the tip end.
This approach matters because the probe is part of the measurement system. Its radius, shape, distance from the sample, and interaction with the surface affect the signal that reaches the detector. Comparing a raw simulated field directly with an sSNOM image would overlook those effects and could lead to an incorrect interpretation of the measured spatial pattern.
Tip geometry and distance affect the result
The model showed sensitivity to both tip radius and distance of closest approach (DCA). A sharper tip can provide a more localized measurement, while a blunter tip changes how the field is sampled. The researchers evaluated radii from 5 to 40 nm and found that the modeled harmonic maps varied with tip size. They also examined DCA, which sets how close the probe comes to the sample during its oscillation.
For comparison with the experimental data, a 15 nm tip radius and an 8 nm DCA gave the best fit under the model assumptions. This was not a universal setting: it was a choice for representing the reported experiment. The work also used AFM mechanical modeling to assess the probe motion and considered material buildup on tips and effects associated with a water meniscus. These details illustrate why metrology teams need to characterize probe conditions alongside the sample.
In the measured and modeled maps, higher harmonics—particularly the fifth and sixth—were more representative of the near field in the region relevant to the recording medium than lower-order harmonics. The comparison also identified a feature in the measured fourth harmonic that the model did not capture. The authors associated this behavior with possible non-sinusoidal tip motion during close interaction with the sample, an effect not included in their model.
Practical implications for process and design teams
For engineers and system integrators, the central lesson is that nanoscale optical characterization is most useful when the measurement physics is included in the comparison with simulation. A modeled measurement can help teams interpret harmonic images, assess whether a near-field feature is resolved, and identify where assumptions about probe motion or geometry may limit the comparison.
- For design teams: use measurement-aware predictions when evaluating whether simulated optical confinement matches observed behavior.
- For metrology teams: document probe geometry and operating conditions that influence the measured signal.
- For manufacturing teams: treat the relationship between simulated and measured fields as part of validating as-fabricated devices, rather than assuming an image directly represents the underlying field.
The study focuses on HAMR plasmonic antennas, but its modeling approach may be applicable to other plasmonic structures. It does not remove the need to understand the specific probe, sample, and operating conditions of a measurement. Instead, it provides a framework for making those influences explicit.
Conclusion
sSNOM offers a way to characterize optical fields at dimensions below the diffraction limit, but interpreting its output requires more than inspecting a high-resolution image. Modeling the probe-sample interaction and the cantilever’s harmonic response helps connect simulated fields to measurements on manufactured devices. For industrial teams developing nanoscale optical components, that connection supports more meaningful device validation and provides a clearer basis for investigating process variation.