Nitrogen vacancy defects in single-particle nanodiamonds sense paramagnetic transition metal spin noise from nanoparticles on a transmission electron microscopy grid

Abstract

Spin-active nanomaterials play a vital role in current and upcoming quantum technologies, such as spintronics, data storage and computing. To advance the design and application of these materials, methods to link size, shape, structure, and chemical composition with functional magnetic properties at the nanoscale level are needed. In this work, we combine the power of two local probes, namely, Nitrogen Vacancy (NV) spin-active defects in diamond and an electron beam, within experimental platforms used in electron microscopy. Negatively charged NVs within fluorescent nanodiamond (FND) particles are used to sense the local paramagnetic environment of Rb_0.5Co_1.3[Fe(CN)₆]·3.7H₂O nanoparticles (NPs), a Prussian blue analogue (PBA), as a function of FND-PBA distance (order of 10 nm) and local PBA concentration. We demonstrate perturbation of NV spins by proximal electron spins of transition metals within NPs, as detected by changes in the photoluminescence (PL) of NVs. Workflows are reported and demonstrated that employ a Transmission Electron Microscope (TEM) finder grid to spatially correlate functional and structural features of the same unique NP studied using NV sensing, based on a combination of Optically Detected Magnetic Resonance (ODMR) and Magnetic Modulation (MM) of NV PL, within TEM imaging modalities. Significantly, spin-spin dipole interactions were detected between NVs in a single FND and paramagnetic metal centre spin fluctuations in NPs through a carbon film barrier of 13 nm thickness, evidenced by TEM tilt series imaging and Electron Energy-Loss Spectroscopy (EELS), opening new avenues to sense magnetic materials encapsulated in or between thin-layered nanostructures. The measurement strategies reported herein provide a pathway towards solid-state quantitative NV sensing with atomic-scale theoretical spatial resolution, critical to the development of quantum technologies, such as memory storage and molecular switching nanodevices.

Fig. 1. (A) A scheme of experimental set-up showing FNDs on a glass coverslip or TEM grid within a custom-made PCB with a wire antenna for microwave delivery. A magnet is also shown which is placed directly above the dried FNDs for the delivery of a magnetic field. Light illumination and detection pathways have also been shown. Schematic Jablonski diagrams showing the excitation and decay pathways of the NV⁻ centre without (B) and with an off axis magnetic field (C). Transitions between ground and excited triplet states (solid arrows) as well as microwave stimulation are shown with non-radiative singlet state pathways, weak (-•-) and strong (--), also highlighted. The off axis magnetic field defines the quantization axis, causes spin mixing in the ground and excited states, and leads to all non-radiative transition rates being likely to be non-zero.

Fig. 2. Thermal variation of the value χ_mT for PBA (A). The blue dotted line shows as temperature increases χ_mT approaches the high temperature spin-only limit of 2.25 emu K mol⁻¹ for a non-interacting S = 3/2 and S = 1/2 pair, consistent with Co(ii)-HS and Fe(iii)-LS states. Curie constant 2.14 K emu mol⁻¹ and Weiss constant −6.39 K confirming the presence of antiferromagnetic interactions at temperatures below the Curie point (14.5 K). Isothermal magnetisation M(H) measured at 2 K is shown in the ESI (Fig. S3). High resolution XPS spectra of PBA in the Fe 2p_3/2 (B) and Co 2p_3/2 (C) regions. The raw and fitted data is shown along with the peaks for oxidation state and satellite evaluation (1–6). Peaks 1, 2 and 3 correspond to a satellite, Fe(iii) and Fe(ii) respectively. Peaks 4, 5 and 6 correspond to a satellite, Co(ii) and Co(iii) respectively.

Fig. 3. (A) MM and ODMR (B) traces for a series of PBA additions to a cluster of FNDs. The solid black line (FNDs) shows the optical response for the FND cluster before addition. The blue, magenta and red lines represent the 1st, 2nd and 3rd addition of the PBA material respectively. The 3rd addition of PBA shows a substantially larger PL contrast reduction compared to 1st and 2nd additions; this is due to more PBA material landing in the field of view in close proximity to FND cluster 2 upon addition (ESI File and Fig. S7 for optical images). Frame number here and throughout indicates an arbitrary time axis, this can be related to real time by multiplying by exposure times used (see Experimental section).

Fig. 4. IL-CLEM NV⁻ sensing of PBA NPs. (A) PL image of drop cast FNDs on a TEM grid. (B) Zoomed-in area of (A) showing the FND of interest. (C) and (D) TEM and STEM images respectively indicating the same area of interest highlighted in (B). Here, we see the exact FND-PBA interacting particles which give the target localised paramagnetic PL response. (C) indicates the FND of interest by a white arrow. (E–G) STEM-EDX mapping of carbon, cobalt and iron respectively. (H) Background subtracted MM trace and (I) ODMR spectrum of the interacting FND-PBA particles. The black trace shows the contrast from the isolated FND particle before PBA addition and the red trace shows the contrast after PBA NPs are added. For MM, raw data is shown alongside averaged data for each ‘on’ and ‘off’ application of the external magnetic field. Error bars for the averaged data are shown and were calculated using simple sum standard deviation analysis. For both MM and ODMR, the quantitative difference between the FND and FND + PBA signal is shown in blue (%) – the method for calculating this difference is in Fig. S21 of the ESI.

Fig. 5. (A) BF-TEM tilt series image (−50°, top right, around the x-axis of the stage holder) of the interacting single FND and PBA NPs. The single plate-like FND particle is on the opposite side of the carbon film to the spherical PBA NPs. Videos of tilt series both in BF-TEM and DF-STEM are the ESI Section. (note that surfaces of NPs became coated with an amorphous deposit after SEM measurements, taken after all NV sensing measurements, details are in ESI File and Fig. S17†). Inset in A, a zoomed-in image of the spin-active interacting particles (area shown using a white star) highlighting the theoretical dipole–dipole interactions between randomly oriented paramagnetic Fe^III and Co^II unpaired electrons (white arrows) and NV⁻ centres through the carbon film. (B) DF-STEM image (−25°, top right, around the x-axis of the stage holder) showing selected areas of EELS thickness analysis. Red boxes indicate the area in which STEM-EELS analysis was taken. EELS measurements were taken at 0° tilt but here a −25° image is used to better illustrate areas of analysis. Carbon film thickness measured in area 1 11 nm, area 2 12 nm, area 3 13 nm and area 4 13 nm. Averaging over these areas (and the other areas away from target as shown in ESI File and Fig. S20 – 13 nm and 18 nm) gives 13 ± 5 nm (x̄ ± 2σ).

Fig. 6. Change in FND PL contrast (%) as a function of shortest FND-PBA interparticle edge-to-edge separation, d(FND-PBA). Inset shows a plot of continuous x-axis. As NV sensing protocol used herein is a surface sensitive technique, distance is measured in 2D from TEM images between shortest edges of the FND and PBA (except for data point 3 which is measured from EELS), PL is measured from the whole FND(s). A table of extracted numerical values of contrast change for both NV sensing schemes, MM and ODMR, can be found in ESI, Table S3. Individual ODMR spectra and MM traces can also be found in the ESI, Fig. S21. TEM micrographs of each area, 1–8, show positions and orientations of paramagnetically interacting or non-interacting FND-PBA particles. Location 4 had traces of silica contamination on the carbon support as evidenced by EDX spectroscopy (ESI File and Fig. S23 for more information). For justification of distance measurements and error see ESI, Fig. S24.

Fig. 7. Local EDX analysis of two locations TEM micrographs (A) and (E) with FND-PBA particles in direct contact (distance between FND and PBA is assumed to be 0 nm). NV sensing response for both MM and ODMR, (B) and (C) for location (A), (F) and (G) for location (E) respectively is shown. (D) and (H), local EDX spectra of locations (A) and (E) respectively. EDX atomic percentage ratios for spin active Co and Fe is shown in inset of (A) and (E), ratio normalised in both from Fe (1.0) in spectra (H). EDX spectra (spot size of electron beam during measurement) were obtained roughly over the entire areas shown in (A and E).