For three decades, researchers building nanoscale quantum sensors have faced a hard physical constraint: their best sensors were atomic-scale defects trapped inside a solid diamond crystal, and they could not be moved to where they needed to go. A new paper from the University of Waterloo, published in Physical Review X on July 28, now breaks that constraint — demonstrating for the first time that a single organic molecule can function as a precision quantum sensor capable of detecting the magnetic signals from just a handful of atomic nuclei. The advance is the most direct step yet toward the 30-year grand challenge of nanoscale magnetic resonance imaging: mapping the three-dimensional structure of an individual protein molecule.

That goal has implications that extend far beyond physics. Every drug that works does so by fitting, at atomic-scale precision, into the three-dimensional geometry of a protein. The structural tools researchers use to understand those geometries — X-ray crystallography and cryo-electron microscopy — are powerful, but both require averaging over vast numbers of identical molecules, which hides the conformational variation that can exist between individual copies of the same protein. Rare structural states, partially disordered regions, and misfolding intermediates directly relevant to Alzheimer's disease, Parkinson's disease, and cancer are averaged out. A sensor that could image a single protein would, for the first time, reveal that hidden structural diversity.

Diamond Sensors Dominate Nanoscale Quantum Sensing: But Their Crystal Prison Limits How Close They Can Get

The reigning approach to nanoscale quantum sensing uses nitrogen-vacancy (NV) centers: atomic-scale defects deliberately engineered into synthetic diamond crystals. NV centers behave as quantum antennas, responding to nearby magnetic fields with shifts in their spin state that can be read out optically using laser light. The technique is elegant and has enabled magnetic field sensitivity at the picotesla level at room temperature, with applications ranging from magnetometry to single-nuclear-spin imaging — the Taminiau group at Delft demonstrated atomic-scale imaging of a 27-nuclear-spin cluster using an NV center sensor as recently as 2019.

But the NV center's strength is also its limitation. The sensor is locked inside a crystal lattice, which means it cannot be positioned arbitrarily close to a biological target. Every additional nanometer of distance between sensor and target costs sensitivity, because the magnetic interactions being detected fall off sharply with distance. For protein-level imaging, that proximity gap has proven stubborn.

The Waterloo team's paper explains that diamond-based sensors are structurally prevented from approaching a target molecule closely enough for high-sensitivity molecular sensing.

Trityl-OX063: When the Sensor Is the Molecule

The Waterloo team's solution was to replace the diamond defect with a free-standing organic molecule. The molecule they chose — trityl-OX063, a synthetic organic radical — is roughly the size of a small protein domain, which means it can be positioned directly adjacent to a biological target. It carries a single unpaired electron whose quantum spin state is exquisitely sensitive to its local magnetic environment.

The molecular architecture protects that sensitivity. Three sulfur-substituted phenyl rings arranged around the central carbon (trityl) core isolate the unpaired electron from unwanted magnetic noise in the surrounding environment, preserving its quantum coherence while still allowing it to detect the tiny magnetic fields generated by nearby atomic nuclei. The molecule can also be chemically functionalized — tethered to specific sites on a protein of interest — giving the sensor not just proximity but targeting capability.

The platform the Waterloo team built around OX063 has a name: SQUINT, for Spin-based QUantum Integrated Nanomechanical Transduction. The acronym signals the key difference from existing platforms: the readout is mechanical, not optical.

Force-Based Readout: How SQUINT Detects Spin

Where diamond NV-center sensors read out their spin state using laser light, SQUINT uses custom-fabricated silicon nanowire (SiNW) probes — 100 nm (about 0.000004 inches, roughly the diameter of a virus) in diameter and 20 µm (about 0.0008 inches, approximately the width of a human hair) long — as ultrasensitive mechanical oscillators. When the OX063 electron spin changes state in response to a nearby nuclear spin, it exerts an infinitesimally small force on the nanowire. The nanowire physically deflects. Those deflections are measured to reconstruct the spin signal.

The approach is conceptually related to magnetic resonance force microscopy (MRFM), a technique that has been building toward single-protein imaging since John Sidles first proposed combining MRI with atomic force microscopy in the early 1990s. Professor Raffi Budakian, the IQC professor who led the Waterloo team, is one of the field's founders — he co-authored the 2004 Nature paper that first detected the magnetic resonance signal of a single electron spin using an MRFM cantilever, for which the World Technology Network awarded him its 2005 materials prize. SQUINT extends that lineage by replacing the conventional cantilever with silicon nanowires and the ferromagnetic tip with a molecular spin sensor.

The platform additionally incorporates a strong, time-dependent magnetic field gradient that allows spatial encoding of the detected spins. This opens the door to Fourier-based magnetic resonance imaging at the molecular scale — essentially, nanoscale MRI.

Budakian has said that force-based spin detection opens an entirely new family of approaches to mapping single protein structures.

How Does Quantum Sensing Work at the Molecular Level?

To understand why the Waterloo result matters, it helps to understand what quantum sensing is measuring. Every atomic nucleus has an intrinsic quantum property called spin — a tiny magnetic moment. When two nuclei are close enough together, their magnetic fields interact, shifting each other's energy levels slightly. A quantum sensor with a spin state of its own can detect these shifts, because the nearby nuclear spins change how the sensor's own spin state evolves over time. By measuring that evolution, researchers can infer the positions and identities of the surrounding nuclei — exactly the information needed to reconstruct a three-dimensional molecular structure.

The challenge is signal strength. Nuclear spin interactions at the distances relevant to molecular imaging generate magnetic fields in the nanotesla range — a trillion times weaker than the Earth's magnetic field. Detecting them requires a sensor that can remain in a coherent quantum state long enough to accumulate a meaningful signal.

This is where quantum coherence time becomes the critical engineering parameter. The longer a sensor maintains its coherent quantum state before environmental noise scrambles it, the more signal it can accumulate from weak magnetic sources.

The Coherence Breakthrough: XYXY Decoupling and the 60-Fold Gain

The central technical achievement of the Waterloo paper is a dramatic extension of OX063's coherence time using a quantum control technique called dynamical decoupling — a method that applies rapid timed pulse sequences to cancel out unwanted environmental interactions that would otherwise destroy the coherent state.

Standard spin-echo pulse sequences — the conventional method — produced only modest coherence extension in the OX063 system. Sahand Tabatabaei, the PhD candidate who serves as the paper's first author, developed a modified XYXY dipolar decoupling sequence specifically tailored to OX063 — one that suppresses electron-electron dipolar interactions across a broad distribution of control field strengths. The result: a coherence time of approximately 400 microseconds in an attoliter-scale droplet containing around 100 OX063 radicals.

Tabatabaei said the core challenge was keeping the OX063 spins coherent long enough to be useful, and that the new control sequence extended coherence to 400 microseconds — roughly 60 times longer than standard spin-echo techniques in the same system.

The 60-fold improvement is not the result of better materials or a different molecule — it is the result of better quantum control engineering. That distinction matters: it means further advances in decoupling sequences could extend OX063 coherence further still, without requiring a new sensor species.

Where SQUINT Stands: Current Capability and the Path to Single-Spin Sensitivity

In its current form, the SQUINT platform can detect the collective magnetic state of approximately 10 nuclear spins simultaneously. That is a significant advance in the MRFM lineage — for comparison, the minimum sample size for conventional MRI-based microscopy is on the order of a trillion nuclear spins — but the team's target is single-spin sensitivity, which is what single-protein structural imaging actually requires.

Budakian is direct about the remaining distance. More milestones lie ahead before the sensor reaches single-molecule sensitivity, he has said, but the team knows how to get there and regards the current result as a landmark advance that puts the final goal within reach.

The team has identified specific upgrade paths. Silicon nanowire force sensors could be replaced by state-of-the-art membrane or string mechanical resonators, which offer better force sensitivity. The OX063 radical itself could potentially be replaced by a purpose-engineered molecular qubit with even longer coherence times — designed by tuning the local nuclear spin environment around the radical center. The platform is not committed to either component.

The research was supported by the Canada First Research Excellence Fund through the Transformative Quantum Technologies program at the University of Waterloo.

Why Drug Discovery Needs Single-Protein Imaging

Current structural biology tools are transformative but fundamentally limited by their reliance on ensemble averaging. X-ray crystallography requires the protein to be crystallized — a process that locks it into a single static conformation and is impossible for many clinically important proteins that resist crystallization. Cryo-electron microscopy captures proteins flash-frozen in solution and then computationally averages the images of thousands of identical copies to reconstruct a three-dimensional structure.

Both methods reveal the population average. They cannot show what happens when a single copy of a protein misfolds, adopts a rare but disease-relevant conformation, or partially unfolds in a way that exposes a drug-binding site normally hidden by the average structure. These are precisely the states that matter most for understanding diseases driven by protein aggregation — Alzheimer's, Parkinson's, ALS — and for designing drugs that exploit transient or rare structural features.

A sensor at the SQUINT platform's theoretical limit would allow researchers to watch a single protein change shape in real time — folding, unfolding, binding a drug candidate — rather than inferring those dynamics from the static ensemble average. Budakian has noted that this capability could deepen understanding of how diseases develop and how drugs interact with their targets, pointing toward more effective treatments.

Frequently Asked Questions

What is a molecular quantum sensor and how does SQUINT differ from diamond-based sensors?

A quantum sensor is a device that uses a quantum property — typically the spin state of an electron — to detect extraordinarily weak magnetic fields. Diamond-based quantum sensors embed their sensing element (a nitrogen-vacancy defect) inside a solid crystal lattice, which limits how close the sensor can be placed to a biological target. SQUINT uses a free-standing organic molecule — trityl-OX063 — as the sensor instead. Because it is not locked inside a crystal, the OX063 molecule can be positioned directly adjacent to a protein or other biological target, improving sensitivity. It also reads out its signal mechanically (via silicon nanowires that physically deflect in response to spin-induced forces) rather than optically, which opens new engineering possibilities.

How could this technology change drug discovery?

Drug design depends on understanding the precise three-dimensional shape of the protein a drug needs to bind. Current structural tools — X-ray crystallography and cryo-electron microscopy — require averaging over thousands or millions of identical protein copies, which hides the structural variation that exists between individual molecules. Rare protein conformations, partially disordered regions, and disease-related misfolding events are invisible to these ensemble methods. A molecular quantum sensor operating at single-spin sensitivity could image individual proteins without averaging, revealing the structural diversity that current tools miss and potentially identifying drug-binding sites that have never been seen before.

What does 'coherence time' mean and why does the 60-fold improvement matter?

Coherence time is how long a quantum sensor maintains its quantum state before environmental noise scrambles it. A longer coherence time means the sensor can accumulate signal from weak magnetic sources — like the nuclear spins inside a protein molecule — before losing track of what it measured. Standard pulse sequences gave the OX063 molecular sensor a coherence time of roughly a few microseconds. The modified XYXY decoupling sequence developed by Sahand Tabatabaei extended that to 400 microseconds — 60 times longer — which is what makes the platform practically useful for detecting nuclear-spin clusters at the nanoscale.

How far is SQUINT from imaging a single protein, and what comes next?

The current platform can detect the magnetic state of roughly 10 nuclear spins simultaneously, compared to the single-spin sensitivity that protein imaging ultimately requires. The Waterloo team has identified two primary upgrade paths: replacing the silicon nanowire force sensors with membrane or string resonators that offer greater mechanical sensitivity, and replacing or engineering the OX063 molecular sensor with a purpose-designed molecular qubit that has even longer coherence. Neither is a fundamental barrier — they are engineering milestones with known solutions — which is why Budakian characterizes the result as "not too far away" from the single-molecule goal.