Optical coupling to defect qubits

Because of the high refractive index of diamond, most of the photons emitted by the defect qubits are total internally reflected at the diamond-air interface. Even for a high numerical-aperture objective (NA=0.9) only 4% of emitted photons can be collected into a free-space mode.

This is exacerbated by the fact that for some defect qubits such as NV centers, only a small percent (~3%, zero-phonon-transition) of all emitted photons are useful for quantum information.

Photonic architecture for entanglement generation

Gallium phosphide (GaP) with a higher refractive index than diamond (3.3 vs 2.4) and large bandgap (2.24 eV) is well suited for photonic coupling to qubit emission in the VIS/NIR band. A unique wet-liftoff process enables transfer of thin (~200 nm) single-crystal GaP layer to ultra-pure diamond samples (containing our defect qubits).

Nanophotonic integration enables efficient optical coupling to qubits and provides a scalable architecture for photon routing, interaction and detection.

Challenges for scalable photonic integration

It is necessary to demonstrate important metrics such as qubit coherence time, optical behavior, fidelity of quantum operation, etc. in a device with favorable characteristics, but more importantly it must be possible to reliably reproduce these characteristics.

The primary factor behind device variability is the high sensitivity of defect qubits to their local environment. This sensitivity is particularly problematic for QIP schemes that rely on remote two-photon qubit entanglement. Here, the indistinguishable nature of emitted photons is essential, requiring spectrally stable NV centers. Our studies on the optical stability of NV defects qubits for photonic integration is published here: Phys. Rev. B 104, 085425 (2021)

We demonstrate three approaches to photonic integration, each with their own advantages and disadvantages.

Qubits integrated with disk resonators

By coupling near surface NV defect qubits to the cavity mode of a nanophotonic gallium phosphide(GaP)-on-diamond disk-resonator, we can efficiently collect defect photon emission into a guided mode. Moreover, defect interaction with a high quality-factor (Q~30,000) cavity greatly enhances the emission in the useful NV zero-phonon-transition via Purcell enhancement. An enhancement factor of 30 was demonstrated.

This work is published at J. Opt. Soc. Am. B 33, B35-B42 (2016) and Nano Letters 18, 2, 1175–1179 (2018)

The advantages of this approach are the ease of design and fabrication, and the high collection efficiency. The disadvantage are manifested as increased spectral instability of the coupled defect qubits. To define the disk-resonators, the diamond host substrate has to be etched via plasma processing. The defect qubits are perturbed by the proximity to etched surfaces.

Qubits integrated with inverse-designed photon extractors

A inverse-design optimization framework is utilized to reconcile a wide range of design constraints and generate planar dielectric GaP-on-diamond photonic structure. A key feature here is the ability to demonstrate NV photon collection enhancement while avoiding etching into the diamond substrate. Fabricated devices show up to 14-fold enhancement of free-space photon collection for device-coupled single NV defect qubits.

This work is published here: Optica 7, 1805-1811 (2020)

The photon-extractors provide better NV spectral stability than disk-resonators while providing moderate enhancement of photon collection. However with this geometry we are limited to free space collection, and the spectral stability although improved, is not yet comparable to native grown-in NV centers. Exposure to chlorine plasma during the GaP plasma etch process deteriorates the defect spectral stability.

Qubits integrated with 1-D photonic crystal cavities

Photonic crystal cavities provide mode confinement to a small volume on the order of a cubic wavelength, enabling operation of the cavity-qubit system well within the strong-coupling regime. Further, as seen above, fabrication directly on diamond is typically detrimental to the spectral stability of qubits. This can be overcome by independent fabrication of photonics and integration via gentle stamp-transfer process.

Starting with wet-transferred GaP-on-oxide material, the 1-D cavities are patterned in GaP by electron beam lithography followed by reactive-ion plasma etching. After initial optical testing of the 1-D cavities, promising devices are picked up and transferred to diamond with pre-characterized qubits by polymer stamp-transfer. Thus, the qubits hosted by the diamond substrate are protected from harsh fabrication processes.

Ancillary projects

An important step towards qubit-device integration is the capability for on demand qubit creation. NV centers can be created on demand with high precision by targeted nitrogen ion-implantation and high-temperature vacuum annealing. This implantation process results in damage to the diamond lattice and results in concomitant creation of other defects. Full control over NV center formation requires a detailed understanding of all the underlying defect kinetics. We utilize a new tool, the tracking of thousands of individual NV centers, to provide insight into the formation, quenching, and orientation kinetics of a quantum defect in an ultra-pure diamond host.
This work is published here: Phys. Rev. Materials 4, 023402 (2020)

For long-distance fiber transmission of single visible or NIR photons emitted by defect qubits, quantum frequency conversion to telecom-band is essential. Multi-resonant frequency conversion of photons to/from the telecom-band has been demonstrated in nanophotonic structures fabricated from non-linear materials such as gallium phosphide (GaP). However, tight fabrication tolerances for such devices typically cause resonant wavelengths to differ significantly from their designed values. We introduce a promising technique for post-fabrication nanoscale control of photonic resonances to address this challenge via electron-beam modification of hydrogen silsesquioxane (HSQ) cladding.
This work has been submitted for publication: arXiv:2110.03612 (2021)

Automated large-scale photonics testing

Typically reliable nanophotonic qubit coupling is difficult to achieve due to unintentional fabrication variation. Hence we make large arrays (>1k devices) with systematic sweep of geometry parameters.

We designed and built a fully automated custom confocal microscope with glavo scanning mirrors and micrometer + piezo stages. Automation scripts written in MATLAB are used to optimize optical excitation\collection coupling for individual devices and survey thousands of devices in a few hours.

To the right you can see microscope snapshots of automated measurements performed on a series of NV defect coupled disk-resonator devices. Wide-field images and image processing are utilized to move between devices and correct for drift during measurements.