Scaling Optical Interaction Length in Chipscale Evanescent Sensors

16 concepts10 papers1 proposal

On-chip optical sensors offer a powerful framework for label-free chemical and biological detection. By guiding light through a microscopic channel and exposing it to a surrounding sample, we can detect trace amounts of molecules via changes in light absorption or refractive index.

However, integrated photonic sensors face a fundamental geometric dilemma. In a standard waveguide, light is tightly confined inside a high-refractive-index core. Only a tiny fraction of the electromagnetic field—the evanescent field—leaks out into the surrounding environment. Because this sensing tail is weak, light must travel across a long physical distance to accumulate a measurable signal from dilute analytes. As Guan et al. note, traditional transmission-based optical sensors are inherently limited by this physical distance constraint. If you simply stretch out a straight waveguide to increase sensitivity, the device footprint expands, destroying the core advantage of a compact "lab-on-a-chip."

To resolve this conflict, modern integrated photonics alters either the path the light travels or the cross-sectional geometry of the waveguide itself. These strategies allow researchers to scale the effective light-matter interaction length by orders of magnitude while keeping the physical device microscopic.


Compressing Space: Spirals and Resonant Cavities

The most direct way to increase the interaction length without expanding a chip's footprint is to fold the optical path.

Spiral Geometries

Instead of routing light in a straight line, a waveguide can be tightly wound into a concentric loop. Su et al. utilize a compact spiral waveguide geometry to maximize the distance light travels through a gas sample. By packing a long optical path into a minimal spatial area, the authors successfully measured the mid-infrared absorption attenuation of methane without increasing the total sensor size.

Cavity Enhancement

An even more compact approach is to reuse the same physical space repeatedly. By bending a waveguide into a closed loop—such as a micro-ring or micro-disk resonator—light can be trapped via total internal reflection.

Straight Waveguide:  [======================>]  (Short interaction length)

Resonant Cavity:     ( O ) <--- Light circulates thousands of times
                                (Centimeters of interaction in micrometers of space)

When light couples into these resonant structures, it circulates thousands of times before escaping. The effective interaction length scales with the cavity's quality factor ($Q$), which measures how long photons are trapped inside.

  • Musgraves et al. and Agarwal et al. demonstrate this principle using high-$Q$ microdisk resonators fabricated from chalcogenide glass. Agarwal et al. achieved a 40-fold reduction in physical device length compared to conventional waveguides while simultaneously boosting detection sensitivity.
  • Chen et al. show that micro-ring resonators can fold an effective interaction length of several centimeters into a physical footprint of only tens of micrometers, dramatically improving the signal-to-noise ratio for trace chemical detection.

This concept of using optical cavities to multiply effective path length mirrors large-scale physics experiments like the LIGO and Virgo gravitational wave interferometers (Acernese et al.), but miniaturized down to the scale of a silicon chip.


Engineering the Mode: Slot and Grating Waveguides

While cavities increase the time light spends near the analyte, another approach modifies the waveguide's cross-section to increase the strength of the evanescent field itself.

Slot Waveguides

In a conventional strip waveguide, light stays inside the high-index core. Slot waveguides turn this paradigm inside out. By placing two high-index rails incredibly close to each other, separated by a narrow, low-index gap (the slot), they exploit electromagnetic boundary conditions.

Maxwell's equations dictate that the normal component of the electric displacement field must be continuous across a dielectric interface ($D_{\text{high}} = D_{\text{low}}$). Because $D = \epsilon E$, the electric field ($E$) must undergo a massive, discontinuous jump when transitioning from the high-index material ($\epsilon_{\text{high}}$) to the low-index slot ($\epsilon_{\text{low}}$).

$$\frac{E_{\text{low}}}{E_{\text{high}}} = \frac{\epsilon_{\text{high}}}{\epsilon_{\text{low}}}$$

This forces a huge concentration of optical power directly into the empty slot—precisely where the target analytes are located. Ranno et al. leverage this slot architecture to maximize the overlap between the optical field and surface-bound ions, enabling parts-per-billion sensitivity. Taking this concept further, Rahman introduces a cross-slot geometry that confines nearly 40% of the total optical power within the sensing region, nearly doubling the sensitivity of single-slot designs and allowing for even shorter sensing arms.

Strip Waveguide Profile:          Slot Waveguide Profile:
     |  Core  |                        |Rail| Slot |Rail|
     |        |                        |    |      |    |
  ___/________\___                  ___/____|  /\  |____\___
     | Field  |                        |    | /  \ |    |
     | Confinement                     |    |/    \|    |
                                            | Field Peak|

Subwavelength Gratings (SWG)

Instead of continuous rails, subwavelength grating waveguides replace the solid core with a periodic arrangement of isolated silicon pillars. If the period of these pillars is significantly smaller than the wavelength of the propagating light, the light cannot resolve the individual structures. Instead, it treats the segmented track as a uniform metamaterial with an engineered, intermediate refractive index.

Ning et al. demonstrate that this structural modification allows the optical mode to expand outward, drastically increasing the overlap between the evanescent field and the surrounding sample area compared to a solid strip waveguide. This greater light penetration enables the detection of extremely low molecular concentrations, pushing detection limits down to the femtomolar range.


Maximizing Interface Sensitivity

By combining these two paradigms—using cavities or spirals to extend interaction time, and using slot or grating cross-sections to maximize field overlap—modern chipscale sensors overcome the physical constraints of optical path length. When coupled with sensitive phase-interrogation techniques like total internal reflection weak measurements (Xiong et al.), these engineered architectures transform tiny refractive index shifts at the waveguide interface into highly readable, macroscopic signals.

Key Foundational Work

Development of chipscale chalcogenide glass based infrared chemical sensors (2010) — contributed 3 concept elements to this direction.
Towards on-chip, integrated chalcogenide glass based biochemical sensors (2010) — contributed 1 concept elements to this direction.
LIGO and VIRGO: large interferometers searching for gravitational waves (2006) — contributed 1 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_evanescent_field_sensing_f27020a9,
  title = {evanescent field sensing},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/evanescent-field-sensing/rdr_f27020a962c143dda4c7636a6fe22dfe},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_evanescent_field_sensing_f27020a9_v1,
  title = {evanescent field sensing},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/evanescent-field-sensing/rdr_f27020a962c143dda4c7636a6fe22dfe/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake