Hysteretic photochromic switching (HPS) of europium-magnesium defects in gallium nitride: a potential route to a new solid-state qubit

Doping is the incorporation of chosen atomic impurities to make a material behave better or differently. When Shuji Nakamura developed a method of producing electrically conducting GaN by activating magnesium (Mg) atoms, he continued a tradition fundamental to all modern electronic devices. Mg doping of GaN allowed production of p-n junctions for today's ubiquitous 'white' light-emitting diodes (LED) and won Nakamura a share in the 2014 Physics Nobel prize. In the same way, europium (Eu) doping of oxide phosphors provided the necessary red optical emission in the 'fluorescent' lamps of a previous lighting revolution. We now propose to take the science of Eu-doped GaN beyond the limited goal of improving red III-nitride LEDs. We aim to explore the potential of hysteretic photochromic switching (HPS), recently discovered by us in GaN co-doped with Eu and Mg, to form the basis of a new solid state qubit or quantum bit.
First trials of rare earth (RE-) doped semiconductors, carried out in the late 1980's, suggested that materials with a wider band gap would show better high-temperature performance, thus favouring II-VI materials and III-nitrides over conventional semiconductors like silicon. However it was not until the present century that III-N semiconductors, grown as high-quality epitaxial thin films on sapphire, were good enough to test this conjecture; another decade passed before Fujiwara demonstrated an LED based on GaN doped with Eu during growth (2010).
Extensive comparative studies of Eu doping methods by the proposer and coworkers in the decade 2001-2011 established that, while such thick GaN:Eu samples could produce brighter overall emission, material produced by ion implantation, followed by annealing, was actually more efficient per dopant ion, by up to 400 times at low temperatures. We also showed that the defect responsible for the GaN:Eu red LED emission was the 'prime' defect, Eu2, consisting of an isolated Eu ion on a Ga lattice site. The commoner Eu1 defect has a more complex emission spectrum, suggesting a Eu atom perturbed by a lattice defect, such as a vacancy or interstitial atom. The total number of such complex centres reported in the GaN:Eu literature is larger than 10.
While attempting to improve the light emission advantage further by implanting Eu in p-type or n-type GaN templates, we discovered hysteretic photochromic switching (HPS) in GaN(Mg):Eu: p-type, Mg-doped GaN samples implanted with Eu ions and annealed. The HPS shows itself in the temperature dependence of the photoluminescence spectrum. At room temperature, the dominant emission, due to the centre Eu0, shows a sharp line at 619 nm. For comparison, Eu1 has a peak at 622 nm and Eu2 at 621 nm. On cooling the sample, the Eu0 intensity increases, as expected, until about 230 K, when it appears to saturate. Below 30 K, we observe a surprising rapid decline of Eu0 as the temperature decreases towards the base temperature of the cooling system. At the same time, an Eu1-like spectrum emerges and effectively replaces Eu0 at 11 K. We deduce that Eu0 somehow switches to Eu1 on cooling over a narrow temperature range. This switching does not reverse if the temperature is then increased from 11 K through 30 K. In fact, Eu1 fades rather slowly, allowing Eu0 to reappear only above ~ 100 K; this is hysteresis. Sample emission is maximum at about 200 K and then fades, reversibly, between 230 K and room temperature. The occurrence of photochromic switching near 20 K on cooldown followed by luminescence hysteresis on warming is given the acronym HPS (hysteretic photochromic switching).
The surprises continue: for samples cooled in the dark, switching from Eu0 to Eu1 can be seen in the time domain; and a resonance line appears at an intermediate wavelength between Eu0 and Eu1. The proposed project aims to determine if the resonance is an actual superposition of Eu0 and Eu1, promising a novel and simple solid state qubit based on Mg acceptor defects.

Current EPSRC-designated research areas on which this project will impact include Photonic Materials and Quantum Technologies. The national importance of III-N semiconductors is acknowledged through the role of the GaN white LED as an actual leading light of Advanced Materials (one of the Eight Great Technologies), and prominent from viewpoints of commerce ($13Bn revenue in 2013) and science (2014 Physics Nobel Prize). It is central to the activities of the United Kingdom Nitrides Consortium (UKNC), now 200-strong, of which Strathclyde University was a founder member. Quantum Information Devices feature in the National III-V centre roadmap.
Our research will benefit III-nitride communities in general, including materials scientists and companies involved in producing commercial devices. A flagship project in RE doping of III-N materials would also have shaping capability for work on RE and other magnetic ions in related research communities. We will encourage involvement with this wider community by opening and pursuing avenues of collaborative research wherever they can be found.
The potential impact of the project touches upon a number of different critical areas:
The most speculative of the impact points is the prospect that the defects Eu0 and Eu1 together form a qubit, a quantum system that can exist in a superposition state, as the 'resonance' emission suggests (see Case for a detailed description). At the least, appearance of the resonance in PL at certain points of the Eu0/Eu1 hysteresis cycle introduces a new class of solid state defect, which is intermediate in configuration between two defects that are stable in different circumstances. Present candidate defects for qubit operation in the solid state are complicated and difficult to control; the behaviour of Eu0/Eu1 is straightforward by comparison.
The incorporation of rare earth (RE) ions in solids impinges upon a whole raft of unsolved problems, both experimental and theoretical. Since pioneering work by Favennec and others, the promise of exploiting sharp f-f optical transitions, which are forbidden in the free ions, in solid state lasers and light emitting devices, has only partly been realised. Favennec's (1989) prediction of 100% quantum efficiency for Erbium and other ions in widegap semiconductor hosts has proven to be a mirage. The central problem here is to unify conventional descriptions of semiconductor band structure with the excitation processes of inner-shell electrons, which produce the observed luminescence. An 'excitonic' description of RE luminescence has been sketched by the proposers, but remains to be verified. At the same-time, our detection of photo-induced magnetism in the excitation cycle of Eu3+ ions has opened up a new experimental approach to the study of 4f electronic states of trapped ions in solids.
An abiding point of controversy in semiconductors is the nature of the Mg acceptor which is responsible for the p-type activation of all the commercial III-nitride devices made to date. Our project offers the opportunity to throw light on this problem because the defect responsible for the Eu0 luminescence necessitates selective association of a Mg acceptor with substitutional Eu. In fact our current model of the photochromic switching between Eu1 and Eu0 further develops ideas of the metastability of the "accidentally shallow' Mg acceptor that have been predicted theoretically. A clear understanding of the acceptor action in widegap semiconductors would have a huge impact on the field and lead the way to the development of deep-UV devices in materials with even larger bandgaps than GaN.
Ion implantation allows control of doping level which simplifies the spectroscopic problem by reducing the number of defect species and defects to a minimum. We aspire to address individual qubits optically, a step in the development of what may prove to be the most important technological device, ever, the quantum computer.