Direct observation of hydrides formation in cavity-grade niobium

Abstract

Niobium is an important technological superconductor used to make radio frequency cavities for particle accelerators. Using laser confocal microscopy we have directly investigated hydride precipitates formation in cavity-grade niobium at 77 and 140 K. We have found that large hydrides were usually formed after chemical or mechanical treatments, which are known to lead to a strong degradation of the quality factor known as Q disease. From our experiments we can conclude that hydrides causing Q disease are islands with a characteristic thickness of * 100 nm and in-plane dimensions 1-10 m. Our results show that mechanical polishing uploads a lot of hydrogen into bulk niobium while electropolishing leads to a mild contamination. Vacuum treatments at 600-800 C are demonstrated to preclude large hydride formation in line with the absence of Q disease in similarly treated cavities.

Precipitation of hydrides in high purity niobium after different treatments

“…Hydrides appeared at 140 K in the form of "lens-like" structures parallel to each other and grew along the longest dimension as temperature decreased down to $100 K. With further temperature decrease both dimensions and a number of hydrides remained constant. Thermal hysteresis has been observed and upon warming up hydrides disappeared at a higher temperature of about 200 K. Surface relief (dents), which look similar to what we observed in our previous studies, 12 remained after hydrides disappeared. After samples were baked at 800 C for 3 h, no observable hydrides could be found during cooldown, and also none were found after any of the subsequent treatments, which included 40 lm EP and 120 C baking for 48 h. A summary of the fast cooldown test results is presented in Table II.…”

“…After several hours at 140-160 K typical heights of hydrides were about 3 lm, which is much larger than the depth of the surface relief left by hydrides dissociation reported in Ref. 12, where dents had a characteristic height of 150 nm. We believe the reason is that surface relief is "proportional" to the hydride-induced deformation or in other words to the difference in volumes of hydride and niobium phases, and this difference is much smaller than the volume of the hydride phase.…”

“…The discussion of what defects can serve as nucleation centers is still open but we think that vacancies and hydrogen-vacancies complexes are among likely candidates. 5,12 We start with the assumption that hydrogen is uniformly distributed in the sample volume, and hydride nuclei are spaced by the distance L. The schematic used for this model is shown in Fig. 11.…”

“…Observed heights are much larger than our previous "lower bound" estimate ($100 nm) made from the observations of surface relief due to hydrides formation. 12 Hydride shape appeared to be dependent upon the crystallographic orientation in each particular grain. This effect was especially pronounced in the case of a single grain material [see Figs.…”

“…11 It is therefore important to systematically investigate for cavity-grade niobium and surface treatments associated with cavities the details of how hydride form, their morphology, and effect of trapping centers. In our previous work, 12 we reported observation of surface relief left by the formation of hydrides in a temperature range 77-150 K, from which we estimated spatial dimensions of the hydrides. It had a disadvantage that all the measurements were taken at room temperature when hydrides had already disappeared and consequently it provided only a "lower bound" on the hydride precipitation.…”