Project results
To answer these questions, we first had to choose a host material (ie: a crystal) to dope with Erbium ions. Our choice was Sheelite; with a chemical CaWO4 its also known as calcium-tungstate (see picture attached). Calcium-tungstate chosen because it has a very low concentration of nuclear spins. These nuclear spins are attributed to the tungsten (W) the calcium-tungstate crystal, and they destroy the fragile quantum coherence of the erbium ions’ magnetic electron-spin transition because of the magnetic noise they create. Indeed, before the SMERC project, the state-of-the-art for erbium electron-spin coherence time was only 50 microseconds [1], which is not very long compared to other commonly used systems, such as bismuth or phosphorus ions implanted in Silicon. To compete with these other systems, our goal was to measure erbium electron-spin coherence at the extremely low temperature of 10mK. This temperature is about 300 times colder than deep space (!!) and can only be reached using a special type of refrigerator known as a “dilution” refrigerator.
To perform these coherence measurements, we used a technique that was developed by the Quantronics Group at CEA Saclay several years ago [2]. A superconducting micro-resonator (fabricated on the crystal surface) is used to perform electron paramagnetic resonance (EPR). A diagram illustrating this spectroscopy is shown attached. These micro-resonators require very low power because their small size; typically, just a few nanowatts. This is critical to making EPR spectroscopy work at 10mK, because even a microwatt of heat can drastically raise the temperature of the crystal. To detect signals with such low-power we had to use a special type of quantum-limited microwave amplifier, known as a Josephson Parametric Amplifier (JPA) that operates with very little power.
For project SMERC, we applied this technique for the first time to a rare-earth doped crystal. We developed a process for making micro-resonators using a very thin sheet (about 100nm thick) of superconducting niobium deposited on the surface of the calcium-tungstate crystal that was doped with erbium. For these measurements we also had to achieve large “quality factors” with the niobium resonators, as the quality-factor directly impacts the detection sensitivity in EPR measurements. During this project we were able to demonstrate quality-factors of 50,000. This is comparable to state-of-the-art results for niobium resonators fabricated on other materials such as silicon and diamond.
In this way, we were able to measure a spin coherence time of 20ms in a CaWO4 crystal doped with 10 part-per-billion (ppb) erbium. This is a 400-fold improvement over the previous record for Erbium spin coherence in calcium-tungstate and consists the first major result of the project SMERC. We also performed a detailed study of erbium spin decoherence in a crystal with a higher erbium concentration of 18ppm. In this crystal we showed the quenching of spin decoherence due to the polarization of other paramagnetic impurities in the crystal, such as ytterbium and manganese.
The second important result of the project was to strongly-couple the erbium ions to the microwave photons in the resonator. This was realised by using the small volume in which the electromagnetic field of the resonator is concentrated. Combined with the large quality factors of our niobium resonators, we demonstrated strong coupling of an ensemble of erbium ions to a superconducting resonator [3].
