Limitations on the use of scanning probe microscopy for the measurement of field emission from copper surfaces

Abstract

High electric-field breakdown in cm-wave x-band radio frequency (rf) accelerating structures is an important phenomenon limiting attainable accelerating gradients. Linear collider development requires an accelerating gradient of, at least, 70MeV/m (150MV/m surface electric field). The observed breakdown sequence usually consists of field emission (FE) from an electrically-conducting surface feature which heats the point of emission, thereby releasing gas from the surface and nearby bulk. The ionized gas makes a plasma that regeneratively heats the metal and releases more gas and electrons (via secondary emission) until a flashover occurs. The FE, however, occurs at a much lower surface electric field in experiments than predicted by theory. Experimental measurements of FE from macroscopic surface areas require the use of a geometrical field-enhancement factor, {beta}, to fit best the data to a Fowler-Nordheim (F-N) emission model with reasonable physical parameters. A newer, microscopic interpretation of the results proposes the existence of quasi-filamentary electrically-conducting channels between the metal bulk and the surface. These channels connect to emitting features near or on the surface, acting as microscopically field-enhanced electron emission sources. Analysis of emitter behavior suggesting that the macroscopic emission should primarily be due to classical geometrical-enhancement is, at least in some cases, due tomore » the presence of these conducting nano-structures in the surface region. We describe an accelerator materials-related study of high electric field breakdown from polished copper (and other) surfaces, in dry-nitrogen gas ambient, using an atomic force microscope (AFM) modified to make F-N FE measurements. In early periods of this work, the results showed that the instrument might be capable of making F-N measurements on small surface areas of mechanically-polished and natively-oxidized high-quality copper surfaces to the GV/m level. The conductive AFM is a topography-imaging instrument capable of extremely high magnification, making it possible to image the actual point of field emission. The technique appeared to be a powerful tool for in-situ measurement of the F-N current combined with imaging of the emitter itself. Following initially promising results, we began a systematic program of characterizing FE from natively-oxidized and oxide-coated polished copper surfaces. The results obtained were, however, not consistently repeatable and we believed that the difficulty was due to a lack of a mechanically-robust and electrically-conductive coating for the AFM tip surface. Similar problems have been recently reported by O'Shea et a1 for similar measurements. The metal-coated tips have been found to wear rapidly during image scanning and measuring current such that the conductive layer on the very end of the tip becomes insulating after minimal use. This difficulty is magnified by working in ambient because the sample and/or tip becomes quickly contaminated, producing chemical changes to the surface barrier heights in the high electric field region. Even in vacuum, wear of tip is found to affect tip-sample adhesion. In general, metal-coated tips are not reliable for obtaining repeatable FE data, although the problem is not completely resolved. We present our data for conductive-AFM using Pt/Ir-coated Si tips generating FE from natively-oxidized Cu, and films of Au on mica, Al{sub 2}O{sub 3} on Pt, Mg on Cu, Hf and W on Cu and, also, Hf ion-implanted into Cu, all in dry nitrogen atmosphere. Hardened surfaces, such as Hf-implanted, and oxide coatings on metals have historically exhibited higher breakdown thresholds and were chosen as sample surfaces for our investigation. F-N behavior was observed in all cases; however, the difficulties pointed out by O'Shea et a1 were readily observable.« less