Alternative approaches for surface treatment of Nb superconducting cavities

With improvements in fabrication and ultra cleanliness techniques, the limitation on SC cavity performance now seems to be the surface state generated by the etching process. The influences of various surface treatments are now being studied at several laboratories (KEK, DESY, CERN...). This paper presents the results obtained at Saclay, mainly on cavities, along with surface studies conducted on Niobium samples (morphology, surface composition). 1 CAVITIES PERFORMANCES. The following issues can briefly summarize general results on cavity performances: • Cavities treated with standard BCP, (hereafter called "FNP", for hydroFluoric, Nitric, orthoPhosphoric acids), commonly reach 25-30 MV/m. However, at about 20-25 MV/m, they exhibit a decrease in the Q factor. This degradation can be attributed to a general heating of the cavity by 5 to 10 mK [1]. (This effect can mainly be observed on heat treated or high RRR cavities since their earlier performances were not high enough to allow this slope to be observed.) • Electro-polished cavities also exhibit a slope in the Q = f(Eacc) curves, even on non heat-treated cavities [2]. • Baking leads to significant improvement of the Q slope [3], whatever the former treatment, but the effect is dramatic on electro-polished cavities. Even non heat-treated cavities can reach accelerating gradients as high as 35-40 MV/m [2]. • We could observe that other surface treatments, like "FNS" (a chemical polishing with different reactants, described in §2), could modify the Qslope. • This surface effect is reversible: for instance improved electro-polished cavities where again degraded by further FNP and vice versa; and there is a kind of “memory effect”: a minimum of 100 μm seems to be necessary to reach a full result. • It is not clear yet to what extent those surface treatments also influence quench field, but some differences where observed. Looking for alternative chemistry appears to be a way to experiment with different cavity surface states. Many chemistry approaches have been tested on samples as described in §2, but only a few could be applied to cavities for practical reasons. That’s why we are also conducting a complete surface study of samples in order to determine which parameters are modified by the different treatments, and which ones are the most preponderant. Preliminary results are detailed in § 3. Three main factors can be explored at first sight: • Influence of surface morphology (roughness at high and low scale). • Surface chemical composition. • Nb lattice strain induced by a superficial oxide layer. This last point somehow ensues from the previous one as chemical composition of the surface can influence the structure of the oxide layer. We will see in the discussion of § 4 that this hypothesis is fairly probable. 2 ALTERNATIVE CHEMICAL TREATMENTS. There are many “recipes” in the literature for etching niobium, see for example [4] or [5]. In all these mixture we can find the same basic constituents: • A niobium complexant, i.e. a chemical species that reacts with ions Nb 5+ and forms a compound soluble in water. • An oxidant, which reacts with metallic niobium Nb 0 and turns it into the oxidized form Nb , which in its turn will be soluble. • Additional compounds like buffer or brightening agents. Note that in the case of electro-polishing, there is no need for an oxidant since a high potential (= oxidant) is already imposed on the niobium by the means of the electrodes and the solution. The drawback of BCP as commonly applied is that it etches rather than polishes the niobium surface, After heavy etching, BCP tends to etch preferentially at grain boundaries, leaving some crevices, which are difficult to rinse correctly and which enhance the surface roughness. Moreover, ortho-phosphoric acid is known to incorporate into the oxide layer in the form of POx (x~2) ions [6], and is regularly found at the metal-oxide interface during chemical analysis [7]. Even before testing on cavities, developing alternative chemical polishing from all the possibilities was difficult. What aspects should be prioritized: roughness, brilliance, etching speed, or safety? One of us (A.A.) made a systematic chemical study of FNS (hydroFluoric, Nitric and Sulfuric acids in various proportions) for several reasons: • It gives rise to very shiny surfaces, • There is no preferential etching at grain boundaries, • It does not contain H3PO4. But it has several drawbacks, among them is the difficulty in preparation and handling. Nevertheless, this surface treatment is now under systematic study on cavities in order to understand what impact it has on cavity performance. After a bibliographic study [4, 5, 8, 9, 10,...] alternative chemical baths like HF-H 2O2, and alkaline media were also tested, but with less success. They will also be briefly described.