Strain Modulation of Si Vacancy Emission from SiC Micro- and Nanoparticles

Single-photon emitting point defects in semiconductors have emerged as strong candidates for future quantum technology devices. In the present work, we exploit crystalline particles to investigate relevant defect localizations, emission shifting, and waveguiding. Specifically, emission from 6H-SiC micro- and nanoparticles ranging from 100 nm to 5 μm in size is collected using cathodoluminescence (CL), and we monitor signals attributed to the Si vacancy (VSi) as a function of its location. Clear shifts in the emission wavelength are found for emitters localized in the particle center and at the edges. By comparing spatial CL maps with strain analysis carried out in transmission electron microscopy, we attribute the emission shifts to compressive strain of 2–3% along the particle a-direction. Thus, embedding VSi qubit defects within SiC nanoparticles offers an interesting and versatile opportunity to tune single-photon emission energies while simultaneously ensuring ease of addressability via a self-assembled SiC nanoparticle matrix.


Supporting Methods
SiC micro-and nanoparticles were provided by Washington Mills Co. Crystals having an average size of 2 and 5 µm were deposited and mechanically dispersed on flat Si substrates. To intentionally create the V Si defects, a sample with SiC particles was irradiated with 1.8 MeV protons to a fluence of 8×10 13 cm -2 and at an angle 8° off the surface normal to reduce channelling effects. The protons have a projected range of around 27 µm, that is, the Bragg peak of implanted hydrogen is situated far beyond the micron-sized SiC particles. Therefore, we anticipate no involvement of hydrogen in the CL measurements. The irradiation was followed by thermal annealing on a hot plate at 300 ºC for 30 min in air to clear out interstitial defects, similar to that performed in Ref.
[7] of the main text.
The samples were studied with a JEOL JSM-IT300 SEM equipped with a Delmic SPARC Cathodoluminescence (CL) System. CL spectra were measured at room temperature (RT) and 80 K with acceleration voltages from 5 to 10 kV and probing currents of 100-200 pA. It is important to mention that the measurements were stable and reproducible, thus the amount of V Si defects studied was not altered by electron irradiation from the SEM. Hyperspectral data was collected with an Andor Shamrock 303i spectrometer with a 300 l/mm grating and a charged couple device (CCD) Andor Newton DU940P-BU2 detector. Long pass filters with cut-off wavelengths of 550 nm were used to block second order diffractions to record the NIR region. Note that the CL spectra have been filtered in order to minimize the etaloning effect caused by optical interference on the back-illuminated CCD detectors in the range 700-1000 nm.
The micro-Raman analysis was performed at room temperature using a Horiba Jobin-Yvon LabRAM HR800 confocal microscope equipped with a 325-nm He-Cd laser source. The laser power density of about 8×10 4 W·cm -2 was attenuated by using a neutral filter to avoid thermal heating effects.
(Scanning) Transmission Electron Microscopy (S)TEM investigations were conducted on a FEI Titan G2 60-300 kV equipped with a CEOS DCOR probe-corrector and Super-X EDX detectors. Observations were performed at 300 kV with a probe convergence angle of 24 mrad. The camera length was set at 77 mm and simultaneous STEM imaging was conducted with three detectors: high-angle annular dark-field (HAADF) with collection angles of 98.7-200 mrad, annular dark-field (ADF) with collection angles of 21.5-98.7 mrad, and annular bright-field (ABF) with collection angles of 10.6 -21.5 mrad. The resulting spatial resolution achieved was approximately 0.08 nm. Geometric Phase Analysis (GPA) was applied on high-resolution images for nanoscale strain measurements in the GMS Gatan software suite. Energy-Dispersive X-ray Spectroscopy (EDX) was applied to identify the chemical composition of the particles. The powder samples were mixed with Epofix glue, and were prepared by mechanical grinding and polishing (Allied MultiPrep). Final thinning was performed by Ar ion milling with a Fishione Model 1010, and plasma cleaning was applied before the STEM investigations with a Fishione Model 1020.

Supporting Analyses
Particle polytype identification Figure 1 of the main text details the microcrystals employed for the present work, by highlighting the large variation in the particle size and shape distribution (see Figure  1a of the main text). Furthermore, Fig. 1 (main text) reveals that the microcrystal sample set is composed of both the 4H, 6H and 15R polytypes, by using a combination of CL and micro-Raman spectroscopy to identify and determine the main spectral signatures of the different SiC polytypes that are present in the initial powder. The spectra in Figure  1(b) (main text) correspond to Raman measurements from individual or small agglomerates of particles (few micron size) acquired at room temperature with an UV laser source of 325 nm. We identify the 6H, 15R and 4H polytypes with reported active Raman modes at 149, 172 and 204 cm -1 , respectively, with 6H-SiC being the dominant polytype over all measured particles.
The near band edge emission as collected by CL spectroscopy also provides important information regarding the polytype of each particle. Figure 1(c) of the main text shows a SEM micrograph of select particles (top) and their corresponding CL false colour maps (bottom) displaying the spectral range from 1.6 to 3.3 eV. The hyperspectral CL measurements were recorded at 80 K and the spectra of the marked particles in Fig.  1(c) (main text) are shown in Figure 1(d) (main text). Vertical bars at 2.39, 2.98, 3.02 and 3.26 eV in Fig. 1(d) of the main text illustrate the excitonic energy gap (E gX ) reported in the literature for the 3C, 15R, 6H and 4H polytypes, respectively. In general, 3C or 4H polytypes can be easily identified because of the dominant luminescence band centred at around 2.0 or 3.0 eV, respectively, and extending up to or close to their corresponding E gX . In contrast, since 15R and 6H polytypes possess similar E gX , an assignment based on the luminescence band extension would be imprecise. However, we observe that 15R particles generally show a narrow peak at 2.98 eV consistent with its E gX .

Supporting Results and discussion
Particle crystal structure Figure S1. (a) High resolution ADF-STEM image from the edge of the microparticle and (b) fast Fourier transform diffractogram from a larger area, verifying that the particle is of the 6H polytype without other polytype inclusions.
With 6H being identified as the dominant SiC polytype over all studied particles, we now turn to an in-depth investigation of the crystal structure of a representative 6H-SiC microcrystal. Figure S1 illustrates the stacking sequence of a microparticle as collected using scanning transmission electron microscopy, and demonstrates that the particle contains only one polytype. That is, a particle of 6H-SiC polytype will not exhibit polytype inclusions. The characteristic ABCACB stacking sequence of the hexagonal 6H-SiC polytype along the c-axis ([0001] direction) is clearly resolved in Fig. S1. The microparticles were single crystals, and no polytype transformations or extended defects were observed. The fast Fourier transform diffractogram in the inset of Fig. S1 also verifies that the particle is of the 6H polytype without other polytype inclusions.  Figure S2. Normalized CL spectra acquired at different localizations within a particle. Figure S2 shows a normalized version of the CL spectra shown in Fig. 2(e) of the main text from an implanted 6H-SiC particle. V1 is more intense in spot P2 while V2 dominates in P2 over all V Si centers, indicating local changes in the emission efficiency.  Figure S3. (a) SEM micrograph of a proton-irradiated square particle and its corresponding (b) UV-Visible and (c) NIR intensity maps. The dashed lines in (b) and (c) contour the particle edge.

Distribution of V Si defects after proton irradiation and annealing
(d) Total CL spectrum at 80 K from the particle in (a). The NIR region has been enlarged in the inset of (d).
Next, let us consider the portion of CL emission from the particles that is related to Si vacancies in the 6H SiC polytype. Figure S3(a) shows the SEM micrograph of an irradiated square particle. The particle contour has been marked with a dashed line (note that there is a slight horizontal drift during CL acquisition) in the UV-Vis and NIR CL maps of the Figures S3(b) and S3(c), respectively. The NIR maps present significantly rougher contrast with respect to the UV-Vis map of Fig. S3(b). The CL spectrum acquired from the entire particle is shown in Figure S3(d), with similar characteristics for 6H-SiC as observed the Figure 2(a) of the main text. Details of the NIR region can be observed in the inset of Fig. S3(d), with the V1, V2 and V3 lines peaking at 1.435, 1.399 and 1.359 eV, respectively. Note that the V2 line dominates over V1, and to a lesser extent over V3. Interestingly, the full width at half maximum (FWHM) of V2 and V3 is around 13-14 meV, similar to the FWHM observed averaging over a great number of particles ( Figure  2(a) in manuscript). The false colour maps in Figure S4 show the distribution of the different V Si centres embedded within the particle. Note that V3 shows similar distribution as compared to V2.