10.25819/ubsi/10474
Open Access Version

Abstract:
Superconducting radio frequency cavities are a key technology in modern accelerators, and, over the past years, their performance improved such that additional losses from trapped magnetic fux are a limiting factor in their performance. This is especially important for accelerators operating in continuous wave mode where high losses in the cavity make operation too energy consuming. For this reason there are many experiments investigating how trapped fux can be reduced. It is investigated how diferent materials and their treatments infuence trapped fux, and how it is afected by cooldown parameters during the transition from the normal to superconducting state. These experiments are often done using cavities as samples. This makes changing material parameters expensive and time consuming. Additionally, the tests themselves are very time consuming so that the number of obtainable data points are often limited. Within the scope of this thesis a new experimental setup is designed which uses fat, rectangular samples to investigate trapped fux. Using these samples has the advantage that diferent materials and treatments can be tested more easily. Additionally, geometric efects during transition are easier to model, and understand. Besides the easier sample preparation the new setup allows for more cooldowns in a shorter period of time so that around 300 thermal cycles can be performed in one day. This is roughly two orders of magnitude more than what is achieved with cavities. With the new setup cooldown parameters like the temperature gradient across the sample, the cooldown rate, and the external magnetic feld can be independently controlled so systematic investigations how each parameter infuences trapped fux can be performed. Measurements conducted with diferent niobium samples confrm efects reported from other experiments. For example a decrease in trapped fux for increasing temperature gradient is observed as well as a linear increase of trapped fux with external magnetic feld under certain conditions. But the ability to record more data points and a relative large parameter space also revealed unexpected results: For large grain niobium it is observed that when a sample is cooled down with a temperature gradient across the sample fux gets only trapped when the external feld is larger than a certain threshold feld which depends on the temperature gradient. Additionally, it is noticed that very fast cooldowns lead to high trapped fux magnitudes almost independent of the temperature gradient. Besides these newly discovered efects the measured dependence of trapped fux on temperature gradient during cooldown does not agree with an existing model. For this reason a new phenomenological model is developed in cooperation with Prof. T. Kubo.