In situ electron paramagnetic resonance spectroscopy using single nanodiamond sensors

Electron paramagnetic resonance (EPR) spectroscopy is a powerful technique used to analyze molecules containing unpaired electrons. Its applications span diverse scientific fields, particularly in studying dynamic processes like redox reactions and molecular motions. A major research focus is adapting EPR for single-cell analysis within living cells. This necessitates developing EPR sensors with high spin sensitivity and biocompatibility. Conventional EPR sensors (macroscopic resonant microwave cavities) have limited spin sensitivity. While microscopic sensors like magnetic resonance force microscopy, scanning tunneling microscopy, and superconducting microresonators have improved sensitivity, they require cryogenic temperatures and high vacuum. Nitrogen-vacancy (NV) centers in diamond offer an alternative: single-spin sensitivity even at ambient conditions. Further, these NV centers can be incorporated into nanodiamonds (NDs) for flexibility as *in situ* sensors (e.g., magnetometry, relaxometry, thermometry). However, utilizing flexible NDs as EPR sensors presents a significant challenge.

Previous studies have demonstrated the use of NV centers in diamond for various sensing applications. Grinolds et al. (2013) showed nanoscale magnetic imaging of a single electron spin under ambient conditions. Shi et al. (2015, 2018) achieved single-protein and single-DNA electron spin resonance spectroscopy in aqueous solutions. Other researchers have explored the use of NDs in different biological environments, including polymers, lipid bilayers, and living cells for magnetometry, relaxometry, and thermometry respectively. However, the inherent flexibility of NDs introduces uncertainty in their orientation, which affects the anisotropic response of NV centers to magnetic fields due to the NV center’s principal axis along the N-V axis. This orientation variability hinders current EPR detection schemes such as double electron-electron resonance (DEER) and cross-relaxation, both requiring precise NV spin state control. Existing strategies to overcome this challenge include actively manipulating ND orientation (e.g., optical tweezers, orientation tracking) or optimizing control pulses, or passively developing orientation-insensitive detection schemes like zero-field EPR, which removes orientation dependence from the resonance frequency of the target molecules, but still shows dependence on the NV sensor’s orientation. This paper addresses the limitation of existing zero-field techniques by introducing a new methodology to make the method robust to the orientation of both the target molecule and the sensor.

This study introduces a generalized zero-field EPR technique robust to both the target and sensor orientations. The key innovation lies in applying amplitude modulation to the continuous driving microwave field (B1 cos ft cos ωt). This modulation creates a series of Floquet sidebands with splitting determined by the modulation frequency (f), rather than the field strength. The researchers initially demonstrate the technique's robustness using fixed NDs, observing P1 centers (another type of defect in diamond). The resonance condition becomes f = ω, independent of the angle (θ) between the microwave field and the N-V axis, thus mitigating the effects of ND tumbling. The signal strength, however, remains dependent on the relative driving index (κ = Ω/f). To further test this robustness, the researchers embed single NDs in an aqueous glycerol solution of vanadyl sulfate. The NDs are tethered to a coverslip using polyethylene glycol (PEG) molecules to restrict translational motion while allowing nearly unperturbed rotational motion. The vanadyl ions ([VO(H2O)5]2+), with an electron spin (S = 1/2) and nuclear spin (I = 7/2), exhibit hyperfine interaction at zero magnetic field. Their zero-field EPR spectrum was simulated, showing multiple potential peaks due to transitions between energy levels. Experimental measurements focused on the middle frequency range to maximize signal clarity and avoid interference from strong background peaks. The experimental setup involved a home-built confocal microscope for optical excitation and detection, and an arbitrary waveform generator, microwave amplifier, and coplanar waveguide for microwave control. The preparation of the tumbling NDs involved surface modification of a coverslip with amino groups, followed by attachment of biotinylated PEG and mPEG, and finally, streptavidin-mediated attachment of biotinylated NDs. Preparation of the vanadyl ion solution included deoxygenation of MilliQ water and glycerol to prevent oxidation, and acidification with sulfuric acid. The experimental EPR spectra were obtained by scanning the amplitude-modulation frequency (f), while adjusting the driving amplitude (B1) to maintain a constant relative driving index (κ). Data analysis focused on fitting the experimental EPR spectra with a Lorentzian function to extract peak positions and hyperfine constants.

The researchers experimentally validated the robustness of their amplitude-modulated zero-field EPR technique. Using fixed NDs, they observed the expected EPR peaks for P1 centers, with peak positions remaining independent of the driving amplitude (B1), thus confirming the insensitivity of the method to ND orientation variations. When the method was tested in a solution of vanadyl ions, a clear EPR spectrum was observed, even with the tumbling motion of the NDs and vanadyl ions. The extracted hyperfine constants (A|| = 195 ± 2 MHz and A⊥ = 579 ± 8 MHz) from the spectrum were slightly different from previous studies with conventional EPR techniques, likely due to changes in the local ligand environment caused by the diamond surface. Repeated measurements using different NDs showed variations in the hyperfine constants, confirming that the hyperfine constant depends on the specific ND. The linewidth of the measured EPR spectrum was consistent with theoretical estimations based on relaxation and diffusion rates of vanadyl ions, considering intrinsic relaxation, dipole-dipole interaction, and rotational diffusion. These findings suggest that the observed signal primarily originates from the vanadyl ions in an adsorption layer on the ND surface rather than freely diffusing ions. The signal contrast obtained in the experimental measurement of vanadyl ions could not be explained by the freely diffusing ions, pointing towards an adsorption layer of vanadyl ions on the ND surface that is responsible for the main signal.

The successful acquisition of the vanadyl ion EPR spectrum in a tumbling ND system demonstrates the practical applicability of the amplitude-modulated zero-field EPR technique. The method's robustness to sensor and target orientation significantly simplifies the experimental setup, eliminating the need for complex orientation control. The observed difference in hyperfine constants compared to previous results highlights the potential of this method for studying the local environment around paramagnetic molecules. The fact that the peak positions are solely determined by the intrinsic interaction makes the measurement robust to the presence of other ions. This opens up the possibility of using this technique in complex biological systems, where the presence of multiple paramagnetic molecules can typically affect the EPR spectra obtained from the conventional EPR spectroscopy. The minor discrepancies in hyperfine constants are attributed to interactions with the diamond surface. Future work should address potential limitations such as optimizing the signal-to-noise ratio, reducing measurement time, and improving spatial resolution.

This study successfully demonstrated a robust method for in situ EPR spectroscopy using single nanodiamond sensors. By employing amplitude-modulated driving fields, the researchers eliminated the dependence of EPR spectra on the orientation of both the sensor and the target molecule, overcoming a key challenge in ND-based EPR. This is a substantial step towards enabling nanoscale EPR measurements in complex biological environments, particularly *in vivo* EPR within single cells. Future research directions include improving the spectral resolution, reducing measurement time and broadening the scope of applications to other types of radicals. Addressing challenges such as microwave heating and ND cellular uptake will be crucial for realizing the full potential of this technique for biological applications.

The current implementation is still time-consuming, partly due to the low signal contrast and measurement efficiency. The observed signal is primarily from vanadyl ions adsorbed onto the ND surface, potentially limiting the information about freely diffusing ions. The ND size (40 nm) might be too large for some single-cell applications. While the technique successfully addresses the orientation issue, challenges remain in optimizing microwave power to avoid cell damage through heating, and improving cellular uptake of NDs for in vivo applications.