🔬 applied physics

Toward a CMOS-integrated quantum diamond biosensor based on NV centers

This paper presents the design and performance analysis of a scalable, CMOS-integrated quantum diamond biosensor utilizing NV centers and a 40 nm SPAD array, which achieves an estimated magnetic field sensitivity of 90 nT/ $\sqrt{\mathrm{Hz}}$ per pixel to enable compact, quantitative magnetic imaging of biological samples.

Original authors: Ioannis Varveris, Gianni D. Aliberti, Felix J. Barzilaij, Zhi Jin, Samantha A. van Rijs, Qiangrui Dong, Daan Brinks, Salahuddin Nur, Ryoichi Ishihara

Published 2026-02-25

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ioannis Varveris, Gianni D. Aliberti, Felix J. Barzilaij, Zhi Jin, Samantha A. van Rijs, Qiangrui Dong, Daan Brinks, Salahuddin Nur, Ryoichi Ishihara

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 want to listen to a whisper in a crowded, noisy stadium. That's essentially what scientists are trying to do when they study tiny magnetic fields inside living cells. For decades, the tools to do this were like trying to hear that whisper using a giant, room-sized microphone system filled with heavy wires and expensive lasers. It worked, but it was bulky, expensive, and hard to move.

This paper describes a breakthrough: shrinking that entire stadium-sized system down to the size of a postage stamp by putting it on a computer chip.

Here is the story of how they did it, explained simply:

1. The Super-Sensitive "Diamond Ear"

At the heart of this invention is a special type of diamond. But not just any diamond—this one has tiny defects inside called Nitrogen-Vacancy (NV) centers.

Think of these defects as tiny, atomic-sized compasses.

Normally, these compasses spin in a specific way.
When you shine a green laser light on them, they glow red.
The Magic Trick: If a magnetic field (like the one from a cell) is nearby, the compass spins change slightly. This change makes the red glow get a little dimmer or shift its color.
By measuring that tiny change in the glow, scientists can "see" the magnetic field.

2. The Problem: The "Whisper" is Too Faint

The problem is that the magnetic signals from biological cells (like a heart cell or a cancer cell) are incredibly weak. To hear them, you need to catch every single photon (particle of light) that the diamond emits.

Old systems used big lenses and cameras to catch this light. But the authors wanted to make it small and cheap. So, they asked: What if we put the camera right underneath the diamond?

3. The Solution: The "Digital Net" (CMOS SPADs)

They built a custom computer chip (using the same technology as your smartphone processor) that acts as a super-fast, super-sensitive light catcher.

The SPADs: Imagine a grid of 256 tiny "ears" (called Single-Photon Avalanche Diodes) sitting directly under the diamond. Each ear is so sensitive it can hear a single photon.
The Speed: These ears don't just listen; they listen fast. They can reset themselves in a blink of an eye (nanoseconds) to catch the next photon. This allows them to count millions of photons per second without getting tired.
The Integration: Instead of having a diamond sitting on a table with a camera miles away, they glued the diamond directly onto the chip. It's like putting the microphone right inside the ear of the person whispering.

4. How It Works: The "Traffic Light" System

To make these diamond compasses work, the system needs three things working together perfectly:

The Green Laser (The Wake-Up Call): A laser shines green light into the side of the diamond. It bounces around inside (like light in a mirror maze) until it wakes up all the diamond compasses.
The Microwave (The Tuner): A tiny antenna on the chip sends out microwave signals (like a radio station) to tune the compasses. When the microwave frequency matches the magnetic field, the diamond's glow changes.
The Chip (The Counter): The chip underneath catches the red glow, counts the photons, and sends the data to a computer.

5. The Real-World Test: Listening to Cells

To prove it works, the team tested it on HEK293T cells (a common type of human cell used in research).

They tagged these cells with tiny magnetic nanoparticles (like giving the cells a tiny magnetic badge).
They placed the cells on their diamond chip.
The Result: The chip successfully detected the tiny magnetic field generated by the cells. It was like hearing a whisper in a stadium, but now the microphone is the size of a coin.

Why Does This Matter?

Imagine if you could take a medical scanner out of the hospital and put it in a doctor's pocket.

Current MRI machines are huge, cost millions, and can only see details about the size of a grain of sand (60 microns).
This new chip is tiny, cheap to make, and can potentially see details the size of a single cell (microns).

The Analogy:
If traditional MRI is like looking at a forest from a helicopter (you see the trees, but not the leaves), this new technology is like having a drone that can hover over a single leaf and count the veins on it.

The Bottom Line

This paper is a blueprint for the future of medical imaging. It shows how to combine quantum physics (diamonds), computer engineering (chips), and biology (cells) to create a device that is small enough to hold in your hand but powerful enough to map the magnetic secrets of life. It's a giant leap from "lab equipment" to "portable diagnostic tool."

1. Problem Statement

Traditional Quantum Diamond Microscopes (QDMs) utilizing Nitrogen-Vacancy (NV) centers for magnetic sensing are currently limited by their reliance on bulky free-space optics, external microwave hardware, and discrete detection electronics. These limitations result in large system footprints, high costs, and complex alignment requirements, hindering their deployment for practical, portable biological imaging. Furthermore, conventional optical imaging methods suffer from autofluorescence and scattering in complex biological samples, whereas magnetic sensing is immune to these issues but requires highly sensitive, scalable detectors to resolve weak magnetic signals from biological sources (e.g., cells labeled with nanoparticles).

2. Methodology

The authors propose a fully integrated, scalable biosensing platform that combines a diamond substrate containing NV centers with a custom-designed CMOS Single-Photon Avalanche Diode (SPAD) array. The methodology involves several key subsystems:

CMOS SPAD Array Design:
- Technology: Fabricated in TSMC 40 nm CMOS technology.
- Architecture: A $16 \times 16$ pixel array (approx. $1 \text{ mm}^2$ ) with on-chip active quenching, recharge, and digital hold-off circuits.
- Performance Targets: Designed for high count rates ( $>3 \text{ Mcps}$ ), low afterpulsing, and picosecond timing resolution.
- Optical Integration: The chip includes a metal grating filter to suppress the 532 nm excitation laser while transmitting the 600–800 nm NV fluorescence, eliminating the need for bulky external filters.
- Readout: A hierarchical architecture using Parallel-In-Serial-Out (PISO) circuits to serialize data from 256 pixels into 16 channels, read by an FPGA, processed by an Arduino, and visualized on a PC.
Optical Excitation and Collection:
- Excitation: Utilizes Total Internal Reflection (TIR) via side-coupled optical fibers or free-space beams to create a uniform evanescent excitation sheet across the NV layer.
- Collection: Analyzed three geometric designs: a simple diamond slab, quartz micropillars, and a diamond-quartz-diamond stack. The authors selected the "simple design" (diamond slab directly on SPADs) for its fabrication simplicity, noting it provides sufficient photon counts ( $\sim 0.75\%$ peak collection efficiency) to saturate the SPADs without complex micro-optics.
Microwave Delivery:
- Designed compact, microfabricated loop antennas or wire grids positioned close to the diamond surface to generate homogeneous $B_1$ fields at the NV resonance frequency ( $\sim 2.87 \text{ GHz}$ ), minimizing phase accumulation and ohmic losses.
Sensitivity Modeling:
- Developed a theoretical model for Continuous Wave Optically Detected Magnetic Resonance (CW-ODMR) sensitivity, accounting for geometric collection efficiency, transmission losses, SPAD Photon Detection Efficiency (PDE), and NV ensemble properties (density, coherence time $T_2^*$ ).

3. Key Contributions

System Integration: Demonstrated a complete architecture for a CMOS-integrated NV sensor, moving from bulky lab setups to a chip-scale device.
Custom SPAD Array: Designed and simulated a 40 nm CMOS SPAD array with active quenching and programmable dead-time control, specifically optimized for the NV emission spectrum.
Efficiency Analysis: Conducted a rigorous comparison of optical stack designs (slab vs. pillars vs. microlenses), concluding that a simple diamond slab is sufficient for ensemble sensing while simplifying fabrication.
Sensitivity Estimation: Provided a detailed sensitivity budget calculation, estimating a magnetic field sensitivity of $\approx 90 \text{ nT}/\sqrt{\text{Hz}}$ per pixel for the proposed architecture.
Biological Use Case Validation: Modeled the detection of Superparamagnetic Iron Oxide Nanoparticle (SPION)-labeled HEK293T cells, demonstrating that the system can resolve the sub-µT magnetic fields generated by biocompatible SPION concentrations.

4. Results

Sensor Specifications:
- Pixel Pitch: $55 \text{ µm}$ (limited by current CMOS design; smaller pitches are possible with 3D integration).
- Field of View: $880 \times 880 \text{ µm}^2$ .
- Sensitivity: Estimated at $90 \text{ nT}/\sqrt{\text{Hz}}$ .
- Count Rate: Capable of handling $\sim 1 \text{ Mcps}$ per pixel (optimal for shot-noise limited detection) with a cycle time of $\sim 20 \text{ ns}$ .
Optical Performance:
- The simple diamond slab design collects light equivalent to an effective collection surface of $4.6 \text{ µm}^2$ , which is sufficient to drive the SPADs into the optimal operating regime.
- Simulations show that TIR excitation creates a uniform intensity profile across the diamond after an initial $750 \text{ µm}$ spread zone.
Biological Application:
- Modeling of SPION-labeled cells indicates that at safe concentrations ( $40 \text{ µg/mL}$ ), cells generate magnetic fields of $0.4 - 0.8 \text{ µT}$ .
- This corresponds to an ODMR peak splitting of $\sim 0.01 \text{ MHz}$ , which is resolvable against the typical ensemble linewidth of $\sim 0.15 \text{ MHz}$ given the system's sensitivity.

5. Significance

This work outlines a practical pathway toward portable, quantitative magnetic biosensors for complex biological environments. By integrating the detector, optical filtering, and readout electronics onto a single CMOS chip, the authors address the primary bottlenecks of size, cost, and complexity associated with current QDMs.

Scalability: The 2D array architecture allows for arbitrary scaling of the sensor area, limited only by bandwidth, enabling wide-field imaging.
Biomedical Impact: The demonstrated sensitivity is sufficient for detecting magnetic signatures from labeled cells, opening doors for applications in cancer detection (cellular level imaging), cardiography, and neuron imaging without the spatial resolution limits of traditional MRI.
Future Outlook: The paper establishes a foundation for future iterations involving smaller pixel pitches (for single-cell resolution), 3D integration, and time-gated ODMR protocols, positioning CMOS-integrated NV sensors as a transformative technology for real-time, quantitative bioimaging.