Skip to content

Bio-sensing with NV-centers in diamond

FieldValue
JournalJournal of Physics Photonics
Published2026-02-17
AuthorsΜαρία Νιώρα, Aldona Mzyk, Kirstine Berg‐Sørensen
DOI10.1088/2515-7647/ae46fe

Two routes now make nitrogen vacancy centers in diamond usable for biosensing on cells, tissue and small organisms. In the first route, a bulk diamond crystal carries a shallow layer of NV centers with the biological specimen resting on top. In the second route, fluorescent nanodiamonds with NV centers are internalized into the specimen and read out from inside the cell. Niora, Mzyk and Berg-Sørensen review both routes in Journal of Physics: Photonics, covering the materials, the sensing protocols and the numbers reported in the literature.

Approach I keeps the diamond and the specimen physically separated. The crystal is reused between experiments, and its NV centers can be characterized before the sample arrives. The cost is that the sensor stays outside the biological object, and high laser power is needed for good resolution or for single-center work.

Approach II internalizes fluorescent nanodiamonds, so the measurement happens at a defined location inside the specimen. The cost is that the particles vary from batch to batch, their coherence properties differ, and co-localization with cellular structures becomes difficult.

ApproachStrengthsChallenges
IDiamond reused, NV centers pre-characterizedExternal to the biological object, high laser power
IIFNDs internalized, localized sensingVariation in properties, coherence properties of FNDs, complexity of co-localization

A variation of approach I fixes fluorescent nanodiamonds at a defined position on a glass slide and places the specimen on top, which keeps the sensor outside the cell while allowing pre-characterization.

Four modalities build on the same level structure. The negatively charged state NV- is the useful one, because it is the only charge state with an optically addressable spin-1 to spin-1 transition and a shelving state that permits optical polarization. The ground state zero field splitting is 2.87 GHz and the excited state splitting is 1.42 GHz.

ModalityQuantity readNumber and condition
Fluorescence spectrometryNV- to NV0 charge ratioEmission in 550 nm to 620 nm compared with 655 nm to 750 nm
ODMRMagnetic field and temperatureGround state splitting 2.87 GHz, excited state 1.42 GHz
T1 relaxometryParamagnetic species and free radicalsAbout 14 electron spins in 10 s, single NV in a 10 nm nanodiamond
NV NMRProton chemical shifts1 Hz resolution in about 10 pL, using a 13 µm NV layer

Optically detected magnetic resonance reads the ground state splitting. At zero field there is one fluorescence dip near 2.87 GHz. A static magnetic field splits it into two dips, equidistant from 2.87 GHz when the field aligns with the NV axis.

Fluorescence spectrometry reads the charge state distribution instead. NV0 emits from about 550 nm to 800 nm with a zero-phonon line at 575 nm, while NV- emits in the red from about 637 nm to 800 nm with a zero-phonon line at 638 nm.

T1 relaxometry measures longitudinal relaxation of the NV electron spin and registers fluctuating magnetic fields from paramagnetic species. The first demonstration functionalized the particle surface with Gd3+ ions and reached a detection threshold of about 14 electron spins within a 10 s acquisition.

NMR with NV centers remains the most demanding modality. Coherently averaged synchronized readout resolved a proton spectrum of several organic compounds in a volume of about 10 pL using a bulk crystal with a 13 µm thick NV layer, at 1 Hz resolution.

The first recording of action potential propagation used excised axons from the marine fanworm Myxicola infundibulum and inshore squid Loligo pealeii, plus live fanworms. The sensor was a single-crystal diamond chip with a uniform 13 µm layer of high-density NV centers. Green light of up to 4.5 W entered from below at an angle giving total internal reflection, which kept absorption in the specimen low. Live fanworms showed no adverse effects over more than 24 h in the setup, and the reported signal was averaged over 1650 stimulations.

Later work used leg muscle extensor digitorum longus from genetically modified mice, where blue light replaced inserted electrodes. Samples were held alive in physiological fluid with continuous carbogen bubbling and lasted 8 h to 20 h with stimulation every 2 s. The diamond was a [100] oriented electronic-grade single crystal, 2 × 2 × 0.5 mm, overgrown by chemical vapor deposition to add a 20 µm layer with a final NV density of about 5 ppm. Green excitation of up to 2 W was delivered by total internal reflection, and Kapton tape insulated against heat.

Action potential propagation was then measured in a mouse brain slice showing a bundle of axons between the hemispheres, the corpus callosum. Slices stayed alive up to 4 h in carbogenated artificial cerebrospinal fluid. A detectable signal needed at least 300 stimulations, and results matched numerical models when averaged over 60 min, which corresponds to 7200 stimulations.

Dynamic imaging and thermometry with bulk diamond

Section titled “Dynamic imaging and thermometry with bulk diamond”

Dynamic imaging demonstrations rely on a lock-in camera to separate the small magnetic signal from laser noise. Signals of 20 µT amplitude were resolved with 1 ms temporal resolution, and a minimum resolvable signal of 104 nT per √Hz was reached in a single frame. Another setup resolved 5 µT fields with sub-second temporal resolution. A Ramsey detection scheme reached 4.1 nT per √Hz per pixel with sub-millisecond temporal resolution and spatial resolution below 10 µm.

Those amplitudes come from microcoils and microcircuits rather than from physiological nerve signals, which are weaker. The review expects improved cameras and pulse protocols to close that gap.

Thermometry reuses the same physics because the transition frequency also depends on temperature. One example placed a dipalmitoylphosphatidylcholine bilayer on bulk diamond with shallow NV centers, 10 nm below the surface in the [111] direction, and combined NMR analysis with thermometry to follow the ordered to disordered phase transition.

Fluorescent nanodiamonds are made by high-pressure high-temperature synthesis, then milled or ground into nanoscale particles and cleaned of residues. The suspension is fractionated by centrifugation or size-exclusion chromatography to narrow the size distribution. The carbon lattice carries no harmful ions, and bare particles show good biocompatibility in eukaryotic cells, tumor spheroids, flies, worms, zebrafish and rodents.

Dose matters. For 40 nm, 70 nm and 120 nm particles in sperm cells, a concentration below 10 µg/ml was recommended. In phagocytic monocytes, 120 nm particles showed no significant difference in cell viability between 10 and 50 µg/ml. Long term tracking works because particles eventually leave the cell by exocytosis, with no effect on proliferation or differentiation observed.

Signal from the particles has to be separated from cellular autofluorescence. Magnetic modulation of nanodiamond fluorescence enabled background free detection in tissue, and fewer than five short blue light exposures photobleached autofluorescence in a simpler approach.

Two figures set the scale of intracellular performance. Relaxometry detects paramagnetic species and free radicals down to nanomolar concentrations in biologically relevant solutions, although sensitivity worsens in protein rich media because a protein corona interferes with the measurement. A silica shell raised spin relaxation times up to 1 ms, and a single step high temperature plastic deformation route produced 50 nm nanodiamonds with T1 of about 1 ms and a fivefold increase in optical Rabi contrast.

Surface chemistry decides the charge state

Section titled “Surface chemistry decides the charge state”

The termination of the diamond surface decides how many centers stay in the negatively charged state. The NV- to NV0 ratio decreases in the order oxygen, nitrogen, hydrogen, so oxygen termination gives the best charge stability. Hydrogen termination gives negative electron affinity and upward band bending, which destabilizes NV- through a p-type conductive surface layer. Fluorination of hydrogen terminated diamond raises the population of negatively charged centers, and chlorine termination does the same more weakly because of its lower electronegativity.

The environment also matters. Adsorption of oxygen from air and of water from solution raises the abundance of NV-. Diamagnetic ions increase relaxation times, with ab initio work suggesting that they trap mobile electrons at lower energy sites. Surface fluorescence changes strongly in the pH ranges 2 to 5 and 9 to 12.5, where protonation and deprotonation occur, and stays constant from pH 5 to 9.

Spin noise comes from dangling bonds on the surface. Annealing in oxygen or vacuum, or terminating with nitrogen, reduces it. Non-thermal plasma raises the ratio of C-H and O-H bonds to C=O bonds on H and O terminated surfaces, which increased both T1 and T2. Borane reduction of oxygen terminated surfaces doubled the spin relaxation time compared with thermal oxidation. Adsorbing a 2 nm alumina layer stabilized the charge state of NV-.

Both routes are limited by predictability of the material rather than by the sensing physics. For bulk crystals, the NV centers form stochastically because a substitutional nitrogen next to a vacancy is needed. Orientation, depth below the surface, layer thickness and concentration therefore vary between crystals.

Orientation still matters. The field component along the NV axis has the strongest influence, nanoscale NMR depends on alignment between the NV axis and the oscillating field, and electric field sensing is direction sensitive through the Stark effect. T1 relaxometry and ensemble ODMR do not depend on orientation. Carbon-13 abundance matters for NMR, and bulk crystals are often electronic grade with more than 99% 13C.

Nanodiamonds have their own bottleneck. Heterogeneity in size distribution, surface chemistry and aggregation in physiological media complicate the sensing properties, and the coherence time of commercially available aqueous nanodiamond solutions tends to degrade over time. Most biofunctionalization also increases particle size, which pushes NV centers farther from the surface and weakens the sensing response.

Protocol limits remain. T1 relaxometry cannot separate different free radicals, so specificity has to come from co-localization, pharmacology and control assays. Charge state transitions during the dark interval of a pulse sequence can corrupt the readout when a particle holds only a few NV centers. Laser and microwave excitation can heat the sample, which matters for long intracellular runs.

Measurements are averages. Intracellular nanodiamond measurements typically take minutes, while bulk diamond recordings can need many hours of repetitive stimulation on living tissue. Single neuron measurements are harder still because a single action potential produces a very low field. The review therefore calls for reference materials with certified NV density, orientation, size distribution, surface chemistry and T1 and T2 times, exchanged between laboratories.

Approach I needs bulk single-crystal plates with a polished surface for specimen placement and total internal reflection illumination. 6CCVD grows single-crystal diamond by MPCVD and offers plates and wafers up to 125 mm with thickness from 0.1 to 500 µm, polished to Ra below 1 nm for single-crystal material. Its metallization with Au, Pt, Pd, Ti, W or Cu suits the microwave antenna that the review places close to the biological sample. Available formats are listed in the diamond materials catalog, and custom fabrication requests go through engineering support.

View Original Abstract

Abstract Color centers in diamond, most frequently the negatively charged nitrogen vacancy (NV − ) center, have demonstrated their applicability as optically actuated quantum sensors of e.g. temperature, magnetic field, or electrical fields. Further, the promise of NV − as a nano- or microscale nuclear magnetic resonance spectrometer has created attention. In this review, the focus is on sensing capabilities offered by NV-centers in particular for investigations of biological systems. Two different approaches, using bulk diamond or nanodiamonds with NV-centers, the so-called fluorescent nanodiamonds, respectively, is discussed, with each their strengths and challenges.

  1. 2018 - Quantum technologies with optically interfaced solid-state spins [Crossref]
  2. 2020 - Hexagonal boron nitride: a review of the emerging material platform for single-photon sources and the spin-photon interface [Crossref]
  3. 2020 - Silicon carbide color centers for quantum applications [Crossref]
  4. 2025 - Quantum sensing with spin defects beyond diamond [Crossref]
  5. 2014 - Nitrogen-vacancy centers in diamond: nanoscale sensors for physics and biology [Crossref]
  6. 2021 - Toward quantitative bio-sensing with nitrogen-vacancy center in diamond [Crossref]
  7. 2022 - Quantum nanodiamonds for sensing of biological quantities: angle, temperature and thermal conductivity [Crossref]
  8. 2023 - Nanoscale quantum sensing with nitrogen-vacancy centers in nanodiamonds - a magnetic resonance perspective [Crossref]
  9. 2023 - Quantum sensors for biomedical applications [Crossref]
  10. 2024 - Super-resolution enabled widefield quantum diamond microscopy [Crossref]