Mars Desert Research Station

Transmission electron microscopy (TEM) and focused ion beam (FIB) are proven tools to produce site-specific samples in which to study devices from initial processing to causes for failure, as well as investigating the quality, defects, interface layers, etc. However, the use of polymer substrates presents new challenges, in the preparation of suitable site-specific TEM samples, which include sample warping, heating, charging, and melting. In addition to current options that address some of these problems such as cryo FIB, we add an alternative method and FIB sample geometry that address these challenges and produce viable samples suitable for TEM elemental analysis. The key feature to this approach is a larger than usual lift-out block into which small viewing windows are thinned. Significant largely unthinned regions of the block are left between and at the base of the thinned windows. These large unthinned regions supply structural support and thermal reservoirs during the thinning process. As proof-of-concept of this sample preparation method, we also present TEM elemental analysis of various thin metallic films deposited on patterned polycarbonate, lacquer, and poly-di-methyl-siloxane substrates where the pattern (from low-to high-aspect ratio) is preserved.

We show the controllable patterning of palladium nanoparticles in both one and two dimensions using electron-beam lithography and reactive ion etching of a thin film of poly(acrylic acid) (PAA). After the initial. patterning of the PAA, a monolayer of polystyrene-b-poly-2-vinylpyridine micelles is spun cast onto the surface. A short reactive ion etch is then used to transfer the micelle pattern into the patterned poly(acrylic acid). Finally, PdCl2 is loaded from solution into the patterned poly(acrylic acid) features, and a reactive-ion etching process is used to remove the remaining polymer and form Pd nanoparticles. This method yields location-controlled patches of nanoparticles, including single- and double-file lines and nanoparticle accuracy of 9 nm or less in one direction was achieved by optimizing the size of the PAA features.

The authors report the ozonation of patterned, vertically aligned carbon nanotube (CNT) forests as a method of priming them for subsequent pseudo atomic layer deposition (psi-ALD) (alternating layer deposition) of silica to produce microfabricated, CNT-templated thin layer chromatography (TLC) plates. Gas phase ozonation simplifies our deposition scheme by replacing two steps in our previous fabrication process: chemical vapor deposition of carbon and ALD of Al2O3, with this much more straightforward priming step. As shown by x-ray photoelectron spectroscopy (XPS), ozonation appears to prime/increase the number of nucleation sites on the CNTs by oxidizing them, thereby facilitating conformal growth of silica by psi-ALD, where some form of priming appears to be necessary for this growth. (As shown previously, psi-ALD of SiO2 onto unprimed CNTs is ineffective and leads to poor quality depositions.) In conjunction with a discussion of the challenges of good peak fitting of complex C 1s XPS narrow scans, the authors present an analysis of their C 1s data that suggests an increase in oxidized carbon, particularly the C=O group, with increasing oxygen content of the CNT forests. After coating with SiO2, the CNTs are removed by elevated temperature air oxidation, the SiO2 is rehydrated, and the plates are coated with 3-aminopropyltriethoxysilane (APTES). The resulting APTES-coated plates separate various fluorescent dyes giving results that are generally at least as good as those the authors reported previously with their more complicated fabrication/priming scheme. TLC plates with different geometries are microfabricated, where plates with narrower channels show longer run times (lower mobile phase velocities) and plates with narrower features appear to give higher efficiencies. (C) 2013 American Vacuum Society.

The effects of iron catalyst thickness on the fabrication and performance of microfabricated, binder-free, carbon nanotube (CNT)-templated, thin layer chromatography (TLC) plates are demonstrated. The iron catalyst was deposited at thicknesses ranging from 4 to 18 nm in increments of 2 nm. Its thickness plays a key role in governing the integrity and separation capabilities of microfabricated TLC plates, as determined using a test dye mixture. Atomic force microscopy and scanning electron microscopy show that smaller and more numerous catalyst nanoparticles are formed from thinner Fe layers, which in turn govern the diameters and densities of the CNTs. The average diameter of the Fe nanoparticles, D-p, is approximately six times the initial Fe film thickness, t(Fe): D-p approximate to t(Fe). After deposition of relatively thick silicon layers on CNTs made with different Fe thicknesses, followed by oxidation, all of the resulting CNT-templated SiO2 wires had nearly the same diameter. Consequently, their surface areas were very similar, although their areal densities on the TLC plates were not because thinner catalyst layers produce denser CNT forests. For t(Fe)=6 nm, nanotube growth appears to be base growth, not tip growth. Best TLC separations of a test dye mixture were obtained with plates prepared with 6 or 4 nm of catalyst. Calculations suggest a loss of surface area for TLC plates made with thicker Fe layers as a result of fewer, thicker CNTs, where the density of silica nanotubes (device surface area) goes approximately as 1/t(Fe)(2). While the focus of this paper is toward a greater understanding of the processing conditions that lead to the best TLC plates, a baseline separation of three analgesics (caffeine, phenacetine, and propyphenazone) is shown on a normal phase TLC plate grown with 6 nm of iron. (C) 2013 American Vacuum Society.

We have fabricated nanoscale fuses from CVD graphene sheets with a 'bow tie' geometry for write-once-read-many data storage applications. The fuses are programmed using thermal oxidation driven by Joule heating. Fuses that were 250 nm wide with 2.5 mu m between contact pads were programmed with average voltages and powers of 4.9 V and 2.1 mW, respectively. The required voltages and powers decrease with decreasing fuse sizes. Graphene shows extreme chemical and electronic stability; fuses require temperatures of about 400 degrees C for oxidation, indicating that they are excellent candidates for permanent data storage. To further demonstrate this stability, fuses were subjected to applied biases in excess of typical read voltages; stable currents were observed when a voltage of 10 V was applied to the devices in the off state and 1 V in the on state for 90 h each.