Investigation of photon emitters in Ce-implanted hexagonal boron nitride

Abstract

Color centers in hexagonal boron nitride (hBN) are presently attracting broad interest as a novel platform for nanoscale sensing and quantum information processing. Unfortunately, their atomic structures remain largely elusive and only a small percentage of the emitters studied thus far has the properties required to serve as optically addressable spin qubits. Here, we use confocal fluorescence microscopy at variable temperature to study a new class of point defects produced via cerium ion implantation in thin hBN flakes. We find that, to a significant fraction, emitters show bright room-temperature emission, and good optical stability suggesting the formation of Ce-based point defects. Using density functional theory (DFT) we calculate the emission properties of candidate emitters, and single out the CeVB center — formed by an interlayer Ce atom adjacent to a boron vacancy — as one possible microscopic model. Our results suggest an intriguing route to defect engineering that simultaneously exploits the singular properties of rare-earth ions and the versatility of two-dimensional material hosts. Solid-state quantum registers formed by interacting electron and nuclear spins amenable to high-fidelity state manipulation and readout provide a promising architecture for quantum technologies, including local information processing and storage, error correction, and longdistance state transmission via photon emission [1,2]. Pioneering work during the last decade has propelled color centers in diamond, silicon carbide, and other wide-bandgap semiconductors as reference platforms in the search for optimal chip-integrated solid-state quantum processors [3–5]. Adding to this list, recent effort has been devoted to the investigation of photon emitters in two-dimensional (2D) materials [6,7], where everimproving ‘pick-and-place’ techniques promise opportunities for integration with electronic or photonic structures [8–10]. From among a growing set of options, hexagonal boron nitride (hBN) is attracting broad interest as the single wide-bandgap van der Waals host of color centers with narrow emission linewidths [11,12], high spectral-tunability [13,14], and strain activation [15]. Moreover, recent work showed the emission from some of these defects exhibits a magneto-optical response [16], ultimately exploited to demonstrate optically-detected magnetic resonance (ODMR) [17,18]. Unfortunately, it is yet unclear whether these intrinsic defects — whose exact physical nature is only now being uncovered [19] — can be more controllably engineered, and whether the strong hyperfine interactions with surrounding nuclei — all of which are spinactive — are compatible with efficient spin-qubit manipulation protocols. Here we explore an alternative family of defects in hBN derived from low-energy implantation of rare earth ions. The latter are attractive in that their partly filled 4f electron orbitals are screened from the outside by outer-lying electrons, which makes their response less sensitive to the host crystal. Further, 4f electrons are often unpaired and hence exhibit magnetic moments that can be polarized and probed. For the present work, we focus on Ce, a rare-earth atom studied in other wide bandgap hosts down to the level of single emitters [20]. Cerium has already been incorporated into hBN as a way to modify the crystal electronic structure and photoluminescence (PL) properties [21– 24], although the doping concentrations (102-103 ppm) are greater than those explored herein (2–6 ppm). We study Ce-implanted flakes of hBN on a SiO2 substrate featuring an array of 40-μmdeep pits. From white-light optical measurements [25,26], we determine the thickness to range from ~50 to 100 nm, consistent with prior observations relying on the same dry-transfer protocols [27–29]; Raman microscopy measurements indicate the crystal quality is high [30– 32] (see Supplemental Material, Section S1 for details). Room-temperature confocal microscopy reveals patchy areas of PL equally scattered across supported and suspended hBN regions (dashed white square in Fig. 1a), hence pointing to in-flake emitters minimally impacted by the substrate. Since flakes are virtually emitter-free prior to implantation, we conclude the observed color centers stem from Ce bombardment. This process, however, is likely to produce a broad class of defects hosting (or not) a Ce ion, a notion our experimental results seem to confirm. Specifically, Fig. 1b shows the spectral PL response from representative sites across the flake, exhibiting either broad emission (Type I), or narrow fluorescence lines (Type II); ‘mixed’ spectra — where both types coexist — are also common. Further, Type I emitters are found to be remarkably photostable under continuous excitation, while Type II are found to blink persistently and ultimately bleach after prolonged laser exposure. While unambiguous assignment is difficult, we associate Type II spectra with intrinsic hBN color centers created during ion implantation, likely vacancyor antisite-based complexes analogous to those obtained from electron irradiation and bombardment with other species, such as carbon or oxygen [19,33,34]. Type I spectra, on the other hand, are more intriguing as they seem to originate from a different class of color center we tentatively associate to Ce-based defects. To interpret our observations, we begin with the Dieke-Carnall energy diagram of Ce3+, a characteristic charge state used here as an initial guide (Fig. 1c) [35]. With only one electron in its 4f orbital, the 2F ground state breaks into a manifold of seven Kramers doublets. Since the separation between them is relatively large, selective excitation from (to) the lowest energy doublet in the 4f (5d) set is possible with a laser tuned at (or near) 489 nm (blue arrow in Fig. 1c) [20,35]. Note that because 5d levels correspond to outer lying orbitals, excited state lifetimes in Ce3+ are expected to be short. Therefore, unlike heavier rare-earth ions — featuring low-photon-yield, intra-4f transitions — Ce3+ defects are expected to be bright. Admittedly, however, the connection between Ce3+ ions and Type I emission is not immediate, as Type I spectra do not display a visible zero-phonon line (ZPL) near 489 nm, and their peak fluorescence wavelengths are heterogeneously distributed over a broad range. On the other hand, both observations are not unexpected as the Huang-Rhys factor of Ce3+ defects in common hosts is high (due to electronic-confinement mismatch following 4f↔5d atomic transitions) [20,35,36], whereas strainand/or local-charge-induced spectral shifts (also affecting Type II emitters, Fig. 1d) are known to be large [13,34,37]. Despite the present limitations, valuable clues emerge from a more extensive optical characterization. Fig. 2 zeroes in on a group of Type I sites, A through C, with nearly identical emission spectra and comparable excited state lifetimes. We monitor the fluorescence from one of these sites (A in Fig. 2a) as we vary the excitation wavelength using a tunable laser (Fig. 2b). Except for an overall scaling, we find the emission spectrum remains virtually unchanged, gradually fading as the excitation wavelength exceeds ~ 490 nm. This response is consistent with that anticipated for a (hypothetical) Ce3+-based defect, as anti-Stokes-assisted photon absorption must decrease below the 4f ↔ 5d transition energy [38]. Note that presumably identical sites (e.g., B, C in Fig. 2a), show different photoluminescence excitation (PLE) spectra (Fig. 2c), thus hinting at several subclasses of Type I emitters (the most likely scenario following an implantation process). Although these results suggest the ‘atom-like’ properties of Ce ions are somewhat preserved, the very notion of a Ce-based defect in hBN is a priori unclear, especially given the large atomic number mismatch between the implanted ions and the atoms of the crystal host. To gain some understanding, we resort to density-functional theory (DFT) in the generalizedgradient approximation (GGA) including corrections for van der Waals interactions [39– 47]. Throughout our calculations, we use the projector augmented wave (PAW) method as implemented in the Vienna Ab-initio Simulation Package [48], with an orthorhombic supercell [49,50] of 6 layers, each containing 112 atoms (Supplemental Material, Section S.II.a). Fig. 3 summarizes the DFT results from one of the considered defects, where cerium associates with a boron vacancy in one of the adjacent atomic planes (Fig. 3a). In the following, we denote this system as CeVB instead of CeB due to the inter-layer position of the Ce atom. Unlike the case with a nitrogen vacancy, CeVN (Supplemental Material, Section S.II.b), we find reduced deformation of the atomic planes surrounding the defect. Calculation of the charge stability transition energies point to the positive, neutral, and negative charge states as the three most likely, with one or the other becoming dominant depending on the system’s Fermi energy (Fig. 3b). The latter can be modified locally via strain, which, in turn, can lead to pockets of fluorescent emitters heterogeneously distributed across the hBN flake; we suspect this mechanism underlies the formation of the bright spots seen in Figs. 1a and 2a. Inspection of the calculated density-of-states (DOS) plots in Fig. 3c-e reveals intraFig. 1. Steady-state spectroscopy of Ce-implanted hBN flakes. (a) Confocal microscopy images of a Ce-implanted hBN flake under 460 nm laser excitation at room-temperature (left) and 5 K (right) revealing a high density of emitting sites across the flake. Emission stemming from a suspended region of the flake (dashed white square) indicates emitters are localized within hBN, not at the interface with the substrate. (b) Representative PL spectra collected from different sites throughout the hBN flake, revealing the simultaneous presence of Ce-derived defects (Type I) and intrinsic hBN defects