AVS1996 Session NS-ThM: Nanofabrication II

Thursday, October 17, 1996 8:20 AM in Room 202 A/B

Thursday Morning

Start	Invited?	Item
8:20 AM		NS-ThM-1 The Fabrication of Nanostructured Organic Materials by Means of a Chemically Modified Template H. Sugimura, N. Nakagiri (Nikon Co., Japan) The patterning of organic materials such as biomolecules in minute dimensions is a key technology for their application in future chemical systems or devices. We report here an approach to the nano-scale assembling of organic materials onto solid surfaces through the chemical modification of a predefined template. The template is fabricated by nanolithography using scanning probe microscopy (SPM) and organosilane monolayers. Silicon substrates terminated with an alkylsilane monolayer were used as samples. This surface termination monolayer was removed from the area scanned by the SPM tip due to tip-induced anodic oxidation, i. e., scanning probe anodization [1, 2]. Since the silicon oxide surface beneath emerged in the tip-scanned area, it became reactive toward organosilane molecules, while the unscanned area remained unreactive due to the terminal alkyl groups. Thus, the tip-scanned area selectively reacts with the organosilane molecules, resulting in the formation of a self-assembled monolayer confined to the scanned pattern. By choosing an appropriate precursor organosilane molecule, a nano-patterned chemical reaction site terminated with functional groups other than alkyl groups, in this case, amino groups, was fabricated. Furthermore, we demonstrate that this patterned region terminated with amino groups could serve as a template for the fabrication of organic nanostructures ! through the area-selective fixatio n of nanoparticles or biomolecules. [1] H. Sugimura, T. Uchida, N. Kitamura and H. Masuhara, J. Phys. Chem. 98, 4352 (1994). [2] H. Sugimura and N. Nakagiri, Langmuir 11 3623 (1995).
8:40 AM		NS-ThM-2 Novel Chemical Approaches to Pattern Transfer using Self-assembled Monolayer Resists K. Seshadri, A. Parikh, D. Allara (Pennsylvania State University); M. Lercel, H. Craighead (Cornell University) We have previously investigated the chemistry of electron irradiation on octadecylsiloxane self-assembled monolayers (SAMs) and have demonstrated their viability for use as ultra-high resolution electron beam resists. Current work is directed towards maximizing the pattern transfer selectivity through the use of chemical development steps and etch contrast masks. For SAMs deposited on the native oxide of silicon and patterned with e-beam irradiation, we observe that exposure to photochemically generated ozone (uv-ozone) removes the residual carbonaceous material in the write lines in a highly selective manner compared to attack at the unirradiated regions and results in significantly enhanced pattern transfer into the silicon substrate upon subsequent treatment with conventional wet etches. The photochemical oxidation also serves to activate the exposed resist to allow deposition of inorganic material, e.g.,MnO\sub 2\, along the pattern line, effectively producing an etch mask. We have also investigated the use of other oxidizing agent in the developing step and find that conditions can be found for which peroxysulfuric acid also shows a high degree of selectivity. Finally, the mechanism of uv-ozone treatment has been studied using model surfaces that mimic the resist material before and after exposure in order to provide a fundamental basis for development of new pattern transfer processes.
9:00 AM		NS-ThM-3 Sulfone Containing Self-assembled Monolayers as Electron Beam Resists P. St. John, H. Craighead (Cornell University); T. Burgin, J. Tour (University of South Carolina); D. Allara (Pennsylvania State University) We have demonstrated increased electron beam sensitivity and semiconductor pattern transfer processes for sulfone containing self-assembled monolayer (SAM) resists. Sulfone polymers have long been used as highly sensitive conventional electron beam resists because, chain scission occurs at low electron doses and cross-linking is negligible. We have applied this concept to ultra-thin electron beam resists by forming sulfonated SAMs on SiO\sub 2\ where the resist thickness is ~2 nm. The ultra-thin self-assembled resist layers have advantages for low energy electron lithography, as would be done with proximal or microcolumn electron sources. Critical doses for straight chain hydrocarbon SAMs are >100\mu\C/cm\sup 2\ at 25 keV. We have observed an increased sensitivity to electron beam exposure with the addition of a sulfone and benzyl group in the monolayer chain. We have used atomic force microscopy to observe and compare the patterned SAM resists. With in-situ Auger spectroscopy during electron beam exposure, we have observed the removal of the sulfur as a function of electron beam dose. These materials act as self-developing electron beam resists and electron beam exposed patterns have been etched with a two step chemical etch using HF and KOH to transfer monolayer patterns into the oxidized Si wafer substrates. This work has been supported by ARPA.
9:20 AM		NS-ThM-4 Electron Beam Profiles and Limits to Nanolithography with the STM T. Mayer, D. Adams, B. Swartzentruber (Sandia National Laboratories) We explore the limits to performance of STM-based, low energy electron beam lithography by measurement and simulation. We measure the beam diameter and tip-sample separation using hydrogen desorption from Si(001) as a probe of the current distribution. For sample bias of 5 to 30 V we measure a beam diameter of 4 to 15 nm, with little variation for currents of 0.1 - 10 nA. To simulate tip performance we calculate the potentials, electric field, emission current density at the tip, and trajectories of emitted electrons for a hemispherical model tip. We calculate the beam diameter at the sample as a function of emitter radius, tip-sample bias, emission current, resist thickness, and tip work function. Beam diameter is primarily affected by the tip-sample gap, increasing at larger gaps characteristic of high bias and large tip curvature. Beam diameter is nearly independent of emission current over the range 0.05 - 50 nA. Resist films cause an increase in beam diameter due to increased tip-substrate gap. Beam diameter increases dramatically for low work function tips. Asperities on flat surfaces produce significantly smaller beam diameter due to focusing of electron trajectories. Beam diameter becomes insensitive to tip geometry, however, at low (<15V) bias, converging to approx. 5 nm at 10 V bias. Experimentally measured beam diameter is consistent with emission from an asperity tip. We conclude that practical, sub-10 nm lithography is feasible using very thin resists (< 10 nm), low voltage (< 15 V), with minimal sensitivity to tip geometry. This work is supported by U.S. DOE contract no. DE-AC04-94AL850000.
9:40 AM		NS-ThM-5 Nanolithography with Electrons C. Marrian, E. Dobisz, D. Park, R. Bass (Naval Research Laboratory); K. Rhee (Sachs Freeman Associates); F. Perkins (Naval Research Laboratory) To understand the resolution limits of nanolithography using electrons, we have developed Monte Carlo codes which describe the mechanisms of electron elastic and inelastic scattering processes in resist materials and substrates. The simulations have been compared with experimental results using focused high energy (20-50 kV) electron beams with 4 different resist materials of thicknesses up to 1500 nm. In all cases, minimum resolvable feature sizes were larger than predicted by the Monte Carlo code. However, the sub micron (100-1000 nm) size regime is accurately described if the generation and tracking of fast secondaries (>50 eV) are included. The >1000 nm regime is also well replicated if those primaries which backscatter into the resist layer from the substrate are tracked. As a result, arbitrary patterns with minimum feature sizes down to about 100 nm can be simulated with a high degree of confidence. The accuracy of the simulations of patterns with sub 100 nm features can be improved by increasing the apparent beam size to ~40 nm, which is over 3 times larger than that used in the nanolithography experiments. While this empirical 'fix' is widely employed, the physical justification for it is not clear. We have examined this issue by comparing results from our previous STM exposures of resists with simulations of the beam profile from the STM tip. This suggests that the total (Gr\um u\n) range of low energy (<50 eV) electrons is over 10 times greater than expected. Alternatively, one can consider that these electrons scatter as if they had a higher energy, i.e. 450 eV rather than 50 eV. The corresponding improvement in the nanolithography simulations will be discussed.
10:00 AM		NS-ThM-6 Nanofabrication using Selective Thermal Desorption of SiO\sub 2\/Si Induced by Electron Beams S. Fujita, S. Maruno, H. Watanabe, M. Ichikawa (JRCAT - Angstrom Technology Partnership, Japan) When an oxide film on Si substrate is irradiated by electron beams(EB), it is known that elemental Si is produced on the oxide surface as a result of oxygen desorption. As the temperature is increased by heating after EB irradiation, we have found that volatile SiO is selectively formed at the irradiated area as follows: Si(EB induced production) + SiO\sub 2\ -> 2SiO(gas), and selectively desorbed from the oxide surface; and Si substrate is exposed. We have demonstrated 10 nm scale nanofabrication using the EB induced STD(EB-STD). Oxide films of thickness 0.7nm were thermally formed on Si(111) substrates. Line patterns were written on them by focused EB of diameter 2nm at room temperature. By heating the sample at 1020K in a UHV chamber after EB irradiation, the irradiated region was selectively desorbed and Si clean surface with the 7x7 structures was exposed in the oxide film. The typical width of the formed "open window" was 10nm, and the narrowest one was 7nm. It has also been verified that EB-STD occurs in oxide films on Si(100) substrates. After the deposition of some materials on the formed open windows in the oxide film, a pattern transfer can be performed by the thermal desorption of the oxide film accompanied with the deposited materials. We name this pattern transfer "UHV lift off". The 10-15nm wide wires of Si crystal or Ge crystal on Si substrate were formed by the UHV lift off technique. It is noted that the whole processes from the oxide film formation to the pattern transfer can be performed in UHV condition without exposing the sample to the air. This study is supported by NEDO.
10:20 AM		NS-ThM-7 Energy Distribution Measurement of the e-beam from the Electron Beam Microcolumn Aligned by STM J. Park, H. Choi, Y. Lee, S. Kang, K. Chun, Y. Kuk (Seoul National University, Korea) Electron beam microcolumn aligned by STM has the performance surpassing the conventional electron beam column. The advantages of microcolumn are as followings: lower aberration, higher brightness, low operating voltage, smaller proximity effect and the compactness of the system. The energy distribution of electron beam microcolumn determines its chromatic aberration,which is dependent on the energy distribution at the source and Coulomb repulsion in the microcolumn. Therefore, The e-beam through the microcolumn has necessarily the dispersion related to the geometry of microcolumn.We constructed a e-beam microcolumn system and measured its performance. Tip preparation, heating procedure and xyz approaching method having resolution 0.1\mu\m was achieved. We were able to observe the behaviors of the e-beam while accelerating , aligning and focusing. We could measure energy distribution of e-beam from this microcolumn system with 127\super o\ electrostatic energy analyzer. E-beam through the microcolumn was compared with e-beam through an extractor only. The former reveals the broader energy distribution than the latter.
10:40 AM	Invited	NS-ThM-8 Nanometrology and the Molecular Measuring Machine J. Villarrubia (National Institute of Standards & Technology) The need for dimensional metrology with uncertainties at the nanometer scale is being driven by the ever-tightening tolerances of modern industry, particularly in semiconductors and magnetic recording. An approach to metrology at these dimensions using scanned probe microscopy (SPM) has lead to the development of NIST's Molecular Measuring Machine (M\super 3\). M\super 3\ is unique in its combination of SPM for high resolution, large 50 mm x 50 mm scan area, sub-millidegree temperature control, and high resolution interferometry for traceability. The average pitch of a specimen formed by laser-focused atomic deposition (nominal pitch: 212.78 nm, measured value: 212.71 nm) has been measured over a distance of 1 mm using M\super 3\. The major components of the uncertainty in this measurement are illustrative of factors likely to be generally important for nano-scale dimensional metrology. They include residual Abbe and other geometrical uncertainties, unmeasured motion errors, and thermal drift. In width measurements dilation of specimen features by the tip is also important.
11:20 AM		NS-ThM-10 Measuring a Focused Ga-ion Beam Profile using an Etch-stop Process and Atomic Force Microscopy J. Wang (Academia Sinica, Republic of China); Y. Wang (Academia Sinica & National Taiwan University, Republic of China) A simple and sensitive method for measuring the profile of a focused ion beam (FIB) with sub-100 nm diameter has been developed. This method exploits the etch-stop mechanism that occurs on a Si(100) substrate after Ga-ion implantation [1]. The ion-implanted area becomes insoluble in aqueous NaOH at room temperature and can be used as negative resist for creating features on the substrate. The height profile and diameter of hillocks created by different ion-beam doses ( 10\super 11\ to 10\super 17\ cm\super -2\) are measured by atomic force microscopy, and the results are used to deduce the profile of the FIB. Since the threshold dose for this etch-stop process is only 1x10\super 12\cm\super -2\ for a 25-keV Ga-ion beam, it is possible to determine the ion beam profile down to 5-order of magnitudes below its peak intensity at the center of the beam. The implications of the beam profile measurement to a FIB-based nanofabrication process will also be discussed.1. P. H. La Marche R. Levi-Setti, and Y. L. Wang, J. Vac. Sci. Technol. B 1, 1056-1058 (1983). *This work is supported by the National Science Council (Contract No. 85-2112-M-001-031), Taiwan, Republic of China.
11:40 AM		NS-ThM-11 High Precision Calibration of a Scanning Probe Microscope (SPM) for Manufacturing Applications D. Chernoff, J. Lohr (Advanced Surface Microscopy, Inc.); D. Hansen, M. Lines (Moxtek, Inc.) A general purpose SPM can function as a metrology SPM when used with a new type of calibration standard and new software. We illustrate this process by measuring the nominal 740-nm track pitch on high density compact discs (DVD). We used a 288-nm pitch, 1-dimensional holographic grating (MOXTEK) as the calibration reference. It consists of a Silicon substrate with a patterned photoresist, overcoated with a tungsten thin film. The holographic exposure process assures uniform feature spacing over the entire specimen area, with an expected accuracy of 0.1%. We operated a NanoScope III/Dimension 3000 large sample SPM in contact mode. To gather sufficient data for statistical analysis, we captured 15-micron wide images of the reference and unknown (compact disc) specimens. We analyzed average cross-section profiles using ASM's Calibrator Pro(TM) software, which reports feature locations with subpixel precision and reveals subtle image distortions. We corrected the nonlinear SPM length scale using additional software. Whereas the raw pitch values for the reference standard had std.dev. = 4.4 nm, corrected pitch values had sd = 1.1 nm. One compact disc had sd= 27 nm and failed the DVD specification. A second disc had sd= 6.8 nm and passed. We have demonstrated a new methodology for calibrating SPM images. The key innovations include: the use of a highly uniform, sub-micron pitch standard; the calculation of feature positions with sub-pixel precision; and replacement of the nonlinear raw length scale with a corrected, linear length scale. We have applied this method to a manufacturing problem requiring careful metrology.