• crystal growth
• fluoride crystals
• optical refrigeration
• rare earths
• spectroscopy

Optical refrigeration is a phenomenon based on anti-Stokes fluorescence which has been observed in many materials, but has shown the best results in fluoride crystals doped with rare earth ions. Rare earth elements are the elements belonging to the Lanthanide series, along with Scandium and Yttrium, which have an atomic number Z ranging from 58 (Cerium) to 71 (Lutetium). They are very important because they have very interesting optical properties, which derive from their particular atomic structure. Their valence electrons belong to the 4f shell, which is deeply embedded inside the 5s, 5p, and 6s shells that shield it from the surrounding environment. When embedded inside a host, like a fluoride crystal, these elements act as trivalent ions with optical transitions with a high quantum efficiency. This property makes them appealing for optical refrigeration.
Fluoride crystals are ideal hosts for optical refrigeration because they have a large energy gap (about 11 eV), which makes them transparent in the visible and most of the infrared range, they have large interstitial sites that can host atoms with high atomic numbers, and they have low phonon energy (in the 350-500 〖cm〗^(-1) range). During this project, we grew and worked with a particular crystal, 〖LiYF〗_4 (sometimes called YLF), doped with Ytterbium (Yb) and, sometimes, with Thulium. Yb is one of the most promising elements for optical refrigeration because it has only one hole in its 4f shell, which leads to a very simple energy structure composed of only two manifolds. This means that an excited ion can only decay by returning to its ground state, either by radiation emission or by other means.
To achieve optical refrigeration we need a crystal host with a dopant ion which has an overlap between its absorption and emission spectra. If we pump the ion with a wavelength that is longer than its mean emission wavelength, the ion will absorb the incoming radiation, then the excited state will absorb more energy from the vibrational modes of the host lattice before losing its energy by fluorescence emission. Since the mean fluorescence has a longer wavelength than the incoming radiation, it means that this process will result in the system losing energy, which was taken from the vibrational modes, resulting in a temperature decrease. We can define a cooling efficiency η_c=(P_out-P_in)/P_in , where P_out is the energy of the emitted light, and P_in is the energy of the pump. This cooling efficiency can be broken down in a few different components: η_c=η_abs (λ) η_ext λ/λ_f -1. η_abs (λ) is the absorption efficiency, which represents the probability that an incoming photon will be absorbed by the dopant ion. η_ext is the external quantum efficiency, which represents the probability that an excited dopant ion will lose its energy by a radiative decay, and also that this emitted energy will exit the crystal without being trapped or reabsorbed.
After it was first demonstrated in the 1990s, optical refrigeration has made great advances. By focusing on fluoride crystals, in particular YLF doped with Yb, and improving the crystal growth process to ensure high purity of the crystal, researchers were able to reach temperatures as low as 87 K on a crystal that started off at room temperature. In the last years, there has been interest in a Yb-Tm co-doping, since it has been proved that small amounts of Tm can increase the cooling efficiency. These crystals are being developed with the aim of creating an all-solid-state cryocooler, which would be a vibrationless, very compact device ideal for applications such as IR or X-ray detectors, which need to be cooled without any vibrations that might disturb the acquisition. So far, optical refrigeration devices appear very promising for space satellites, since they are light, compact and very reliable, having no moving mechanical parts.
This project was conducted with the aim of growing a few crystals for optical refrigeration, and then to investigate how impurities and other external factors can reduce the optical refrigeration efficiency of a crystal. The crystals we used were grown by us at the New Materials for Laser Applications in the Department of Physics of the University of Pisa. Starting from high-purity powders (5N certified) of LiF , 〖YF〗_3, 〖YbF〗_3, and 〖TmF〗_3, we grew two crystal boules using our home-made Czochralski furnace. The Czochralski technique is one of the most used for crystal growth, since it yields crystals with high purity and very little lattice defects. We grew a boule of YLF:10% Yb and one of YLF:10% Yb - 0.004% Tm. From each one of them, two samples were cut and polished for optical refrigeration measurements.
To verify whether the samples cooled, we put them in a vacuum chamber designed to reduce to a minimum any heat transfer between the sample and the surrounding environment. We then pumped the samples with a VECSEL tuned at 1024 nm, a wavelength ideal to pump the optical transition between the two manifolds of Yb. While the laser pumped the crystal, we controlled the sample temperature with a thermal camera connected to a computer. All of the samples we grew showed no sign of optical refrigeration, since they all heated up during the trials.
To understand what might cause this behavior, we used a pulsed Ti:Sa laser and a monochromator to measure the mean lifetime of the Yb excited manifold, and we compared the result for each of our samples with a sample of the same chemical composition that previously achieved optical refrigeration. Our samples showed a systematically lower lifetimes, meaning there were some impurities responsible for this quenching.
To further investigate the effects that impurities can have on the cooling efficiency, and in particular on the absorption efficiency, we used a YLF crystal doped with Yb which is capable of optical refrigeration. We used our usual setup for cooling measurements, with the addition of a second laser beam incident on the sample. We confronted the temperature the sample could reach when pumped with only the VECSEL at 1024 nm with the one reached when we pumped it with both the VECSEL and the blue laser at 473 nm. We were able to record a variation of the sample temperature. By also recording a spectrum of the emitted fluorescence in both configurations, we were able to identify certain optical transitions activated by the blue laser that produced extra heating phonons, providing an extra thermal load on the system which caused the temperature difference we observed.
Finally, to investigate how reflection and reabsorption of the emitted fluorescence can be detrimental to the cooling efficiency, we put two YLF:Yb samples, both capable of cooling down when pumped with the VECSEL, inside the vacuum chamber. We kept them a few mm apart, and by using a beam splitter we separated the VECSEL beam in two. One of these beams was sent on one of the samples, while the second one was sent on the other one. We recorded the temperature change in one of the samples with the thermal camera, both when the VECSEL was pumping only the sample under analysis and when the VECSEL was pumping both of the samples inside the vacuum chamber. We were able to see that its cooling efficiency results decreased if the second sample, next to it, is cooling as well, emitting its own fluorescence onto the sample under analysis.