Impact of surface and laser-induced noise on the spectral stability of implanted nitrogen-vacancy centers in diamond

Srivatsa Chakravarthi, Christian Pederson, Zeeshawn Kazi, Andrew Ivanov, and Kai-Mei C. Fu

Phys. Rev. B 104, 085425 – Published 24 August 2021

Abstract

Scalable realizations of quantum network technologies utilizing the nitrogen-vacancy (NV) center in diamond require creation of optically coherent NV centers in close proximity to a surface for coupling to optical structures. We create single NV centers by ${}^{15}N$ ion implantation and high-temperature vacuum annealing. The origin of the NV centers is established by optically detected magnetic resonance spectroscopy for nitrogen isotope identification. Near-lifetime-limited optical linewidths $(< 60 MHz)$ are observed for the majority of the normal-implant ( $7^{\circ}$ , $\approx 100$ nm deep) ${}^{15}{NV}$ centers. Long-term stability of the ${NV}^{-}$ charge state and emission frequency is demonstrated. The effect of NV-surface interaction is investigated by varying the implantation angle for a fixed ion energy, and thus lattice damage profile. In contrast to the normal-implant condition, NVs from an oblique implant ( $85^{\circ}$ , $\approx 20$ nm deep) exhibit substantially reduced optical coherence. Our results imply that the surface is a larger source of perturbation than implantation damage for shallow implanted NVs. This work supports the viability of ion implantation for formation of optically stable NV centers. However, careful surface preparation will be necessary for scalable defect engineering.

Received 22 May 2021
Accepted 10 August 2021

DOI:https://doi.org/10.1103/PhysRevB.104.085425

Physics Subject Headings (PhySH)

NV centers Quantum information with solid state qubits

Diamond Nitrogen vacancy centers in diamond

Ion implantation Optically detected magnetic resonance Photoluminescence Spectroscopy

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

Srivatsa Chakravarthi ¹, Christian Pederson², Zeeshawn Kazi ², Andrew Ivanov², and Kai-Mei C. Fu^1,2

¹Electrical and Computer Engineering Department, University of Washington, Seattle, Washington 98105, USA
²Physics Department, University of Washington, Seattle, Washington 98105, USA

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 104, Iss. 8 — 15 August 2021

Reuse & Permissions

Access Options

Article Available via CHORUS

Download Accepted Manuscript

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
Diamond surface preparation and simulations of implantation conditions: (a) Morphology of the diamond surface measured by atomic force microscopy. The RMS surface roughness for sample A (B) was measured to be 0.63 (0.43) nm before implantation and 0.56 (0.30) nm after implantation and annealing. (b) A simulation of the implantation profile obtained by srim for sample A (red, $7^{\circ}$ ) and sample B (blue, $85^{\circ}$ ) showing the damage trails and final positions of implanted ${}^{15}N$ atoms. The green circle represents our excitation laser spot. Note that the implantation yield is low ( $< 5$ %); hence within an excitation spot there is a small probability that each trail results in a NV center upon annealing. Inset i: Illustration of the implantation geometry. Inset ii: A cross section of the simulation showing the ion damage trails for the two implant angles. The total number of vacancies generated per ion is similar for both the implant angles. However, on sample B, some ${}^{15}N$ atoms are lost due to ion scatter out of the surface. Here, dia., diameter; V, vacancy.
Reuse & Permissions
Figure 2
Correlated NV ODMR (RT) and PLE ( $T < 12$ K) measurements on sample A. (a) Pulsed ODMR scheme utilized to identify the N isotope. Laser and radio-frequency (RF) pulses are generated by an acousto-optical modulator and RF switch, respectively (laser power, 0.8–1.0 mW; spot diameter, $\sim 800$ nm). Pulses and photon collection are triggered by a programmable pulse generator. Det, detection; Init, Initialization pulse. (b) Confocal PL map of the implanted region with the measured NV centers indicated by their isotope (green, ${}^{15}N$ ; red, ${}^{14}N$ ). (c) ODMR spectra for the marked NV incorporating implanted ${}^{15}N$ . (d) ODMR spectra for the marked NV incorporating grown-in ${}^{14}N$ . (e) Resonant excitation (PLE) scheme utilized for characterizing the optical coherence of the marked NV centers. Sidebands (SBs) at 2.9 GHz are added (using an electro-optic modulator) to the scanning resonant laser to counteract optical spin pumping. Upon detection of an ${NV}^{-}$ to ${NV}^{0}$ ionization event (indicated by lack of ${NV}^{-}$ PL) a 50 ms off-resonant green repump is used to reset the NV charge state. (f) Time traces of PLE scans measured at 10.5 K for three NV centers. Off-resonant repumps between scans are indicated by green squares along the right column; this induces large spectral jumps in many ${}^{15}{NV}$ centers (e.g., NV 1). The “ $∖ ∖$ ” markers along the scan axis indicate discarded scans where no NV PL is observed. The PLE traces for ${}^{15}{NV}$ centers typically show long periods of spectral stability between repump pulses.
Reuse & Permissions
Figure 3
Photoluminescence excitation characteristics of measured NV centers on samples A and B. (a) The fitted Lorentzian full width at half maximum (FWHM) distributions for each observed NV. The colored boxes and black markers represent the interquartile range and median linewidth, respectively. The total number of scans and measurement duration comprising each distribution are recorded in the accompanying table. The PLE traces for NV centers 1–3 are also shown in Fig. 2. (b) The distributions of the scan-to-scan change in the center frequency of the fits, representing spectral variation. Off-resonant repump pulses are only applied when NV ionization is detected. On average, there are six repump events over a measurement duration of 10 min. The blue markers indicate large spectral jumps after repump pulses.
Reuse & Permissions
Figure 4
Automated ODMR and PLE measurements are performed on four additional samples (samples C–F). We are not able to track the NV centers between ODMR and PLE with the automated protocol. The ODMR ${}^{15}{NV}$ -to- ${}^{14}{NV}$ ratio suggests predominance of ${}^{15}{NVs}$ across all implanted samples. Similarly to the previous data set (Fig. 3), the average (Avg.) linewidth for each individual NV is computed as the mean FWHM obtained by fitting each of the 30 frequency scans to a Lorentzian. Automated PLE scans performed on deep grown-in NVs (25 $μ m$ from surface, sample F) are used as a reference.
Reuse & Permissions
Figure 5
The NV ZPL inhomogeneous distribution. The histogram is generated by recording the center wavelength of all peaks observed with 532-nm excitation from locations within an implantation region (except the grown-in reference NV centers) on the respective samples. Note that no distinction is made between the different transitions associated with the NV excited state spin sublevels. The bin size is 0.025 nm, and the spectrometer resolution is 0.021 nm. The histogram data are fitted to a Gaussian to extract the FWHM of the distribution. Sample B shows an obvious deviation from other samples.
Reuse & Permissions
Figure 6
NV charge state ratio as a function of the 532-nm excitation intensity.
Reuse & Permissions

Physical Review B

covering condensed matter and materials physics