🔬 optics

High-sensitivity silicon nitride microring resonator opto-fluidic sensor

This paper demonstrates a scalable, high-sensitivity silicon nitride microring resonator opto-fluidic sensor fabricated via foundry processes, achieving a mean refractive index sensitivity of 579 nm/RIU in the C-band for potential environmental and bio-sensing applications.

Original authors: Davey O. Armstrong, Sherif Ibrahim, Shirin Naserikarimvand, Simon Whelan, Owen J. Guy, Anthony J. Bennett, John P. Hadden

Published 2026-01-28

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Davey O. Armstrong, Sherif Ibrahim, Shirin Naserikarimvand, Simon Whelan, Owen J. Guy, Anthony J. Bennett, John P. Hadden

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a tiny, invisible drum made of glass (specifically, silicon nitride) that is so small it fits on a computer chip. This drum is designed to vibrate only when hit by a very specific color of light. In the world of physics, this is called a microring resonator.

Here is how the paper explains this technology, broken down into simple concepts:

1. The Problem: Finding "Needles in Haystacks"

Currently, checking water for dangerous germs or chemicals usually requires sending a sample to a big, expensive laboratory. It takes a long time, and you have to wait for results. The researchers want to build a tiny, fast sensor that can do this job right at the source (like a river or a tap) without needing a lab.

2. The Solution: The "Light Drum"

The team built a sensor using a ring of silicon nitride (a type of glass). They shine a laser light into this ring.

The Analogy: Think of the ring like a race track. The light is a runner. The runner can only stay on the track if they run at a perfect speed that matches the length of the track. If the speed is slightly off, the runner falls off.
The Magic: When the ring is empty (filled with air), the light runs at a specific speed. But if you pour a liquid over the ring, the "friction" of the air changes to the "friction" of the liquid. This changes the speed the light needs to run to stay on the track.

3. How It Detects Contaminants

The researchers tested this by pouring different mixtures of water and rubbing alcohol (isopropyl alcohol) over the chip.

The Shift: As they added more alcohol, the "perfect speed" for the light changed. This caused the color of the light that successfully ran the track to shift slightly.
The Measurement: They measured exactly how much the color shifted. They found that for every tiny change in the liquid's density (refractive index), the light color shifted by a measurable amount.
The Result: Their sensor was incredibly sensitive. It could detect changes in the liquid's properties with a sensitivity of 579 nanometers per unit of density change. To put that in perspective, this is a very high level of sensitivity, comparable to or better than many other similar sensors currently in use.

4. Why This Material?

They chose silicon nitride (SiN) because it is like the "Swiss Army Knife" of optical materials:

Transparent: It lets light pass through easily without losing much energy.
Stable: It doesn't react wildly to temperature changes (unlike some other materials that might give false readings if the room gets warm).
Scalable: It can be mass-produced in factories (foundries) just like computer chips, making it cheap to build in large numbers.

5. What They Actually Claim (and What They Don't)

What they did: They built a working chip, tested it with alcohol-water mixtures, and proved it can detect tiny changes in liquid density very accurately. They also showed that the device is stable even when the temperature fluctuates slightly.
What they don't claim yet: They did not say they have already detected viruses, heavy metals, or specific pollutants in real water. They did not say this device is ready for hospitals or drinking water systems today.
The Future Plan: The paper mentions that to make this useful for real-world pollution detection, they would need to coat the sensor with special chemicals (like "molecular Velcro") that grab onto specific bad guys (like heavy metal ions). They also plan to make the fluid delivery system better so it's easier to swap out water samples.

Summary

The researchers created a tiny, factory-made glass ring that acts like a super-sensitive musical instrument. When a liquid touches it, the "note" the instrument plays changes. By listening to that change, they can tell exactly how dense the liquid is. This proves that we can build cheap, tiny sensors that might one day help us monitor our environment instantly, though more work is needed to teach them to recognize specific pollutants.

Technical Summary: High-sensitivity silicon nitride microring resonator opto-fluidic sensor

Problem Statement
There is a growing demand for low-cost, highly deployable biochemical sensors capable of operating outside centralized laboratory settings, particularly for environmental monitoring and personal healthcare. Current monitoring approaches for water contaminants (pathogens, viruses, toxic chemicals) are often slow, expensive, and reliant on complex sample processing. While photonic integrated circuits (PICs) offer a solution for real-time, sensitive analysis, there is a need to balance high sensitivity with manufacturability. Specifically, high-quality (Q) factor microcavity sensors often suffer from reduced interaction between the cavity mode and the analyte due to field confinement, whereas devices with lower Q-factors may offer higher refractive index sensitivity through increased field overlap. The challenge lies in developing a scalable, foundry-fabricated sensor that achieves high sensitivity without the complexity of specialized geometries or feedback loops found in other high-sensitivity demonstrations.

Methodology
The authors investigated the refractive index sensitivity of a silicon nitride (SiN) microring resonator (MRR) based photonic integrated circuit. The devices were fabricated using a 300 nm SiN multi-project-wafer service at the CORNERSTONE foundry. The fabrication process involved a 300 nm Si3N4 epilayer grown via low-pressure chemical vapour deposition (LPCVD) on a 3 µm buried oxide layer, clad with a 2 µm SiO2 layer via plasma-enhanced chemical vapour deposition (PECVD). A channel was etched near the resonator to allow controlled exposure to fluid samples.

The specific device characterized was a 40 µm radius MRR in an 'add-drop' configuration. Optical characterization was performed using a C-band tunable continuous-wave laser (1528 nm to 1568 nm) coupled via a 1D grating coupler. To ensure thermal stability during liquid measurements, the sample was mounted on a heating block connected to a thermoelectric controller (TEC), as the thermo-optic effect can cause resonance shifts.

Sensitivity was evaluated by exposing the sensor to varying concentrations of isopropyl alcohol (IPA) solutions (1% to 5% by volume) in deionized water. Resonance shifts were measured via transmission spectroscopy. Additionally, 2D finite-difference time-domain (FDTD) simulations using Lumerical's varFDTD solver were conducted to model the sensor's behavior, varying the microring radius (40–80 µm) and cladding refractive index (1.33–1.38).

Key Contributions and Results

Fabrication and Characterization: The study successfully demonstrated a foundry-fabricated SiN MRR sensor. The device exhibited a mean free spectral range (FSR) of 4.83 nm. In air, the sensor achieved a mean loaded Q-factor ( $Q_L$ ) of 23,800 and a mean intrinsic Q-factor ( $Q_I$ ) of 26,300.
Sensitivity Performance: Upon exposure to IPA solutions, the resonant wavelengths exhibited a steady redshift. The sensor achieved a mean sensitivity of 579 nm/RIU (nanometres per refractive index unit). Individual mode sensitivities were calculated as 585 ± 3 nm/RIU, 578 ± 3 nm/RIU, and 573 ± 3 nm/RIU. This measured sensitivity exceeded the values predicted by simulation, a discrepancy the authors attribute to the underestimation of evanescent field interaction by the variational FDTD solver.
Limit of Detection (LOD): Based on the calculated sensitivity and Q-factor, the theoretical limit of detection for the resonant mode at 1539.6 nm was determined to be $2.55 \times 10^{-4}$ % IPA concentration. The authors note that current performance is limited by noise and wavelength calibration rather than the fundamental physics of the device.
Simulation Insights: FDTD modeling revealed that sensitivity decreases as the microring radius increases. This suggests that for single-mode MRRs, the bulk refractive index sensitivity is more dependent on the field-analyte overlap than on the increase in sampling length provided by a larger radius. Smaller radii push the optical mode further into the outer edge, increasing overlap with the sensing region.
Thermal Stability: The device demonstrated thermal stability with a resonance shift of only 13 ± 2 pm/K over a 10 K temperature range.

Significance and Claims
The paper claims that this work demonstrates a scalable, high-sensitivity opto-fluidic sensor compatible with recognition marker surface functionalization. By utilizing a standard foundry process (SiN-on-insulator) and a simple geometry, the authors achieved a sensitivity (579 nm/RIU) that is competitive with, and in some cases exceeds, other reported SiN and SOI-based sensors, while avoiding the complex geometries (such as feedback-coupled rings) required for ultra-high sensitivity in other studies.

The authors state that the on-chip nature and CMOS compatibility of these devices pave the way for rapid, low-cost, large-scale fabrication, which is essential for commercial viability. They conclude that this technology opens the door to applications in environmental and bio-sensing, provided that future work addresses surface functionalization (e.g., silanisation and thiolated-aptamer binding) to target specific pollutants and integrates light sources and detectors to create a fully independent sensor system. The paper remains modest regarding its current limitations, acknowledging that noise and calibration currently constrain the practical limit of detection.