Skip to content

Ion Concentration and Voltage Imaging with Fluorescent Nanodiamonds

FieldValue
JournalAdvanced Materials
Published2026-02-15
AuthorsPatrick Voorhoeve, Alessandro Mameli, Hiroshi Abe, Takeshi Ohshima, Qiang Sun
DOI10.1002/adma.74728

Hydrogenated fluorescent nanodiamonds reached a voltage sensitivity of up to 16 mV Hz⁻¹ᐟ². Their ion response reached up to 1.8% ΔPL per mM NaCl. Both signals came from all-optical changes in NV⁰ photoluminescence within aqueous electrochemical cells. Hydrogenation switched fluorescent NV⁰ toward nonfluorescent NV⁺ in sub-30 nm particles, while reoxidation recovered much of the emission. Surface hydrogenation therefore made particle photoluminescence responsive to applied voltage and local salt concentration.

Surface chemistry controls the NV charge state

Section titled “Surface chemistry controls the NV charge state”

Surface oxidation and hydrogenation reversibly switch the fluorescent NV⁰ and nonfluorescent NV⁺ states in sub-30 nm FNDs. Oxidized nanodiamond surfaces have positive electron affinity, which stabilizes NV⁰ and preserves bright photoluminescence. In the authors’ band-bending model, hydrogenated surfaces have negative electron affinity and favor electron depletion near the surface. The depleted near-surface region favors NV⁺, especially when most defects lie close to a particle surface.

Reversibility was measured on 95 small aggregates made from commercial 20 nm FNDs. Confocal imaging used 520 nm excitation at 900 µW and collected PL above 550 nm. After hydrogenation at 800 °C for 1 h in 95% N₂ and 5% H₂, mean PL fell from 87 kcps to 25 kcps. A UV-ozone treatment raised mean PL to 61 kcps, although recovery was incomplete.

Some substrate-contacting facets may remain inaccessible to ozone, which the authors offer as a possible explanation. The measured recovery supports a charge-state switch rather than annealing damage, but it does not establish full oxidation uniformity.

Hydrogenation worked at the 150 mg powder scale for both 18 nm and 120 nm FNDs. The powders received 2 MeV electron irradiation at a fluence of 1 × 10¹⁸ cm⁻². They were then annealed in argon at 900 °C for 2 h and in air at 520 °C for 2.5 h. The hydrogenation step used the same 800 °C, 1 h forming-gas treatment.

FTIR confirmed a change in surface functional groups. The oxidized O-H and C=O absorption peaks at 1626 cm⁻¹ and 1779 cm⁻¹ decreased after treatment. CH-related peaks appeared at 2832 cm⁻¹ and 2932 cm⁻¹. Average zeta potential shifted from below -25 mV to above 25 mV for both nominal sizes. The positive potential at neutral pH also enabled self-assembly onto negatively charged substrates.

Dynamic light scattering suggested typical diameters near 30 nm and 50 nm for the nominal 18 nm material after oxidation and hydrogenation. The larger diameter may reflect aggregation that probe sonication could not reverse. Hydrogenation drove the 18 nm powder below the PL detection threshold. For the 120 nm powder, PL decreased by 39%, while the spectral shape remained unchanged.

The authors’ model used an 8 nm depletion layer and found that all NVs in an 18 nm particle could switch. It assigned 35% of NVs in a 120 nm particle to that near-surface region. After UV-ozone treatment, the 18 nm powder recovered 4.0-fold in PL. This supports reversible charge-state control at the 100 mg scale.

Voltage sensitivity reaches 16 mV Hz⁻¹ᐟ²

Section titled “Voltage sensitivity reaches 16 mV Hz⁻¹ᐟ²”

The strongest hydrogenated FND response reached 16 mV Hz⁻¹ᐟ². The voltage cell used 18 nm FND-Hyd spin-coated onto ITO-coated glass and immersed in 0.17 M NaCl. A platinum counter electrode was separated from the ITO surface by several millimeters. Wide-field imaging used 532 nm excitation, PL collection above 580 nm, and a camera.

During a voltage ramp from +1 V to -1 V at 2 mV s⁻¹, the best FND-Hyd aggregate changed PL by 42% at -1 V. The same particle produced a -23% change at +1 V. Oxidized FND changed by only -2% at -1 V and 1% at +1 V. The slope from -0.2 V to +0.2 V gave a shot-noise-limited sensitivity of 16 mV Hz⁻¹ᐟ².

Negative bias drew positive charge toward the surface and recovered NV⁰ through NV⁺ to NV⁰ conversion. Positive bias increased upward band bending and depleted NV⁰.

During 36 voltage pulses over 25 min, FND-Hyd gave repeatable 36% and -22% responses at -1 V and +1 V. Each pulse lasted 20 s. Standard deviations were 4.2% and 5.3%, respectively. The maximum baseline PL changed by only 4.0% during that measurement.

Most of each transition occurred within about 100 ms, but PL continued drifting throughout the pulse. The residual drift was 8.2% during +1 V and 5.4% during -1 V. A three-dimensional Nernst-Planck model predicted NaCl changes up to 7 mM. Its diffusion dynamics resembled the zero-voltage PL relaxation.

Wide-field imaging resolves particle-level voltage maps

Section titled “Wide-field imaging resolves particle-level voltage maps”

All 98 particles and aggregates in one field of view responded to both polarities. The mean ΔPL was -11% at +1 V and 15% at -1 V. Video frames were acquired at 10 frames per second while voltage was switched over several minutes. Larger aggregates produced larger ΔPL magnitudes.

The previous study by Karaveli et al. reported a similar magnitude for hydrogenated FNDs of comparable size. The earlier work included particles that responded to only one polarity, whereas every particle here showed the same polarity dependence. That uniform response supports voltage mapping across a particle population rather than relying on selected outliers.

Salt gradients produce spatial ion-concentration images

Section titled “Salt gradients produce spatial ion-concentration images”

Voltage-created NaCl gradients produced spatial PL maps with hydrogenated FNDs. Platinum electrodes were 500 µm wide and separated by 160 µm on quartz. Hydrogenated FNDs were self-assembled between the electrodes in 1.7 M NaCl. Wide-field imaging used 532 nm excitation and PL collection above 580 nm.

After +1 V for 10 s, PL decreased by up to 11% on the left and increased by up to 17% on the right. The pattern matched a simulated NaCl gradient across the electrode gap. COMSOL Multiphysics 6.2 modeled NaCl transport with Nernst-Planck equations.

Both PL and calculated concentration approached zero near the electrodes within about 1.5 s after polarity reversal. Both signals then changed sign and followed the new salt distribution. The measured residual field was 2 V m⁻¹ and uniform, with an ionic current of 38 nA. The uniform field cannot explain the spatial gradient, and the authors also argue that the constant current is unlikely to account for it. Salt concentration remains the most likely cause.

Normalized responses peaked at 0.45% mM⁻¹ for +1 V and 0.31% mM⁻¹ for -1 V. Several particles reached about 1.5% mM⁻¹ or higher, consistent with the reported maximum of 1.8% ΔPL per mM. Large aggregates averaged 1.3% mM⁻¹, while smaller particles averaged 0.29% mM⁻¹. Response amplitude also varied at a fixed salt change, so the relationship was nonlinear and particle dependent.

Adding NaCl without voltage increased PL by 4.9% after the first addition and 2.3% after the last. The average was measured over a 156 × 156 µm² coated region. That test covered 0 M to 2 M in 0.5 M steps with 200 s equilibration between additions.

Calibration limits and engineering implications

Section titled “Calibration limits and engineering implications”

Salt sensitivity depended on the measurement geometry, which prevents a universal calibration from these results. In the electrode geometry, the mean salt response was about 0.4% mM⁻¹. In the bulk addition test, the response stayed below 0.01% mM⁻¹ over a similar 1 M to 1.5 M range. The paper leaves this discrepancy unresolved and calls for further mechanistic study. The authors hypothesize that ion-induced changes in the electric double layer control the NV charge state. That mechanism remains unresolved.

Surface treatment also affects colloidal behavior, because hydrogenation increased DLS diameters and may reduce dispersibility. Incomplete oxidation recovery suggests that substrate contact can limit switching uniformity. The voltage response retains diffusion-like dynamics, so electrode geometry and electrolyte transport must be calibrated together.

Particle size also matters, because the 18 nm powder switched more completely than the 120 nm material. For material selection, sub-30 nm particles place most NV centers close to a controllable surface. Hydrogenated particles are more voltage responsive than oxidized particles, even though their baseline PL is lower. For device integration, the coating must limit aggregate formation while preserving access to the particle surface. These results demonstrate response, but they do not provide a universal calibration for arbitrary particle sizes or electrolytes.

For voltage and ion sensing chips, the paper needs sub-30 nm nanodiamond particles, patterned Pt electrodes, and particle coatings stable during imaging. The 6CCVD diamond materials catalog lists Pt metallization, matching the metal used, although the paper deposits it on ITO-coated glass or quartz. Questions about custom fabrication of the electrode and substrate stack can be directed through engineering support.

View Original Abstract

ABSTRACT The nitrogen‐vacancy (NV) center in diamond exists in different charge states with distinct photoluminescence properties, which are sensitive to the nanoscale electrochemical environment. Hence, the NV charge state is emerging as a powerful all‐optical platform for nanoscale sensing and imaging. Although significant progress has been made in engineering near‐surface NV centers in bulk diamond, controlling the NV charge state in fluorescent nanodiamonds (FNDs) has proven challenging, limiting the sensitivity and reliability of FND‐based charge state sensing. Here, we demonstrate reliable, reversible switching between the fluorescent NV 0 and a dark state, likely NV + , in sub‐30 nm FNDs via surface oxidation and hydrogenation, respectively, for single particles and particle powder. In aqueous electrochemical cells, we demonstrate voltage and ion concentration imaging based on the NV charge state in self‐assembled FND layers on transparent substrates. Applied voltages reliably modulate the FND PL with a sensitivity of up to 16 mV Hz −1/2 . FND PL is also modulated by local changes in salt concentration with a sensitivity of up to 1.8% per millimolar NaCl, enabling all‐optical imaging of ion concentration gradients at the microscale. Our results represent a significant step toward realizing fast, stable, and scalable nanoscale charge‐ and voltage‐imaging technologies with sub‐micrometer spatial resolution.