OTHER TECHNOLOGY EFFORTS
NOVA has been involved in several different technology arenas in the past 13 years. Successful programs in semiconductor processing, MALDI/TOF spectrometry, and X-RAY/VUV optics have been completed. Below is an overview of some of the technology involved in those programs.
X-Ray Lithography
In its first sponsored program NOVA designed and constructed a collimator for X-ray lithography through support of the Defense Advanced Research Projects Agency (DARPA). X-ray lithography is one of several options for making much denser, smaller and faster microchips for computers. The line spacing in semiconductor processing is largely defined by the wavelength of photons used to expose the masks lying above the silicon. The processing technology has moved from exposure with visible to ultraviolet to extreme ultraviolet. X-rays have even shorter wavelengths than that of the EUV photons, making them suitable for patterning features below 100 nm. In this application X-rays emanating from a point source (plasma or laser bombarded metal tape) are redirected into a very uniform and broad parallel beam roughly 40 mm. in diameter (Shown above). Collimators will be a critical component to widespread application of X-ray point sources which, in turn, will permit local end use and dramatic economic enhancement for X-ray lithography.
The NOVA collimator developed has a highly complex cross section and precision radius as designed by computer modeling. The model is employed to define a highly uniform field intensity and divergence.
The X-ray photons undergo only one or two reflections at the pore interface. These devices therefore obey the thin lens formulas, acting exactly like simple glass lenses, faithfully relaying the X-ray images. The surface smoothness of the pore surfaces is measured to be an average of 20Å. The collimator was extensively measured for uniformity, local divergence, and global divergence with soft 1 keV X-rays in both a synchrotron beam and in the laboratory.
X-Ray Astronomy
The collimator concept can also be adapted for use in astronomy by simply operating it in reverse: very faint X-ray signals coming from distant galaxies essentially constitute a parallel beam and can, therefore, be focused down onto an imaging detector, providing astronomers with improved detection limits and the ability to probe much deeper into space.
For these applications, highly sophisticated Monte-Carlo computer modeling has shown square channel optics exhibit improved focusing as compared with round channels. NOVA developed proprietary techniques to ensure highly square pores having 10-200 µm pore sizes (Image of square pore optic above). Accurate spherical radiusing allows for magnification as well as 'lobster eye' formatting , where the optic focuses incoming rays to a point. Metallization of the channel surfaces enlarge the critical angle acceptance and enhance single bounce reflectivity. Below is an experimental cross-shaped point-to-point focused X-ray image, measured using a repetitive laser-plasma X-ray source at the Rutherford Appleton Laboratory in England. The focusing lens is comprised of square channels. Without the lens, the X-ray field would be diffuse.
These soft X-ray focusing components will become the centerpiece of a ESA supported "all-sky monitor" orbiting telescope operating at 1 keV. This system has many such lenses, each staring at different points in the sky. The square-pore structure mimics the actual eye of the lobster. For a more detailed discussion of the 'Lobster Eye' program, please follow this link to the site maintained by our colleagues at the University of Leicester, Leicester England.
MALDI-TOF Spectrometry
NOVA evaluated new techniques for a high-mass detector for MALDI time-of-flight mass spectrometry (schematic below) under the sponsorship of the National Cancer Institute, a part of the National Institutes of Health. This program achieved improved detection sensitivity for biomolecules having very large masses (10,000-1,000,000 amu). The detection and analysis of large biomolecules is a critical gateway to improved analytical capabilities for medical researchers working on proteins, carbohydrates, new drug synthesis, cancer research, DNA sequencing and potentially, viruses.
Detection of large ions has historically been complicated by their low momentum transfer and the multiplicity of charges. In particular, for samples desorbed by laser as in MALDI, large numbers of very small ions from the matrix substrate arrive at the detector first, causing saturation. To better understand these phenomena, NOVA worked with a respected spectroscopist, Dr. James Hill of MIT, and constructed a sophisticated MALDI test bed. This work was reported at a meeting of American Society for Mass Spectrometry. ("Development of a Time-of-Flight Spectrometer for Detector Evaluation", ASMS 1998, Orlando, FL.) The research focused on modifying microchannel plate detectors to enhance the sensitivity for the higher mass ions, and resulted in a US patent.
Miniaturized TOF Sensor
NOVA earlier developed a concept for sensor miniaturization, demonstrating a proof-of-principle operation of a flashlight-battery-sized, time-of-flight (TOF) detector for NASA to carry out measurements of the solar and interstellar winds in long-duration solar system probes.
The concept, shown in the accompanying diagram, employs a extremely thin carbon stripping foil for incoming ions and neutrals. Reaction products can include both negative ions and electrons. The electrons are used to produce the start pulse for the time-of-flight measurement, and the arrival time for the ion can then be measured. This system has the advantage of having the starting and stopping pulses measured by the same detector. It also takes advantage of extremely fast electronics. This concept is limited to ion masses of 40 amu and below due to time resolution. Most typical spectra are normally broken down into daughter ions in this range. Miniaturization of sensors can be an important method of providing small, in-situ monitoring control of critical manufacturing processes, improving yields and reducing costs. Small sensors can be effective in monitoring many sites in parallel with low unit cost.
High Dynamic Range Detectors
NOVA developed a technique to build a detector with very high dynamic range supported by the National Science Foundation. Often, sensitive scientific analyses can be complicated by the simultaneous presence of both very weak as well as very strong signals. Weak signals can be unresolved because of noise limitations or poor quantum detection. Intense fluxes can saturate the detector and prevent accurate measurement of data. Count rate capabilities of 1011 counts/cm2/sec were demonstrated using an ultraviolet photon input. Prototypes demonstrated operational lifetimes and the ability to be repeatedly cycled. Detectors with an extended signal range can provide expanded capabilities for researchers in areas such as mass spectrometry, high energy physics, UV, X-ray and neutron imaging, signal processing, and solar astronomy - in general, all applications requiring detection of very high signal fluxes.
Semiconductor Processing
NOVA earlier carried out a program on maskless lithography using electron beams through the co-sponsorship of two agencies of the Department of Defense. E-beam lithography is a candidate for next generation lithographic patterning of silicon wafers for feature sizes below 100 nm competing with extreme ultraviolet (EUV) optical systems.
Electron beams saw early commercial application in electron microscopes and other analytical equipment. A properly focused electron beam can write features as small as 5 nm and exhibit a very large depth of field. E-beam technology is also a well established mask writing process for creating the photomasks used in semiconductors today. Error sources for present masks are typically controlled in the 10-100 nm range.
When used to pattern sensitive resists on silicon, the electron beam has a significant advantage in that the beam can directly expose the resist without the additional requirement of a mask. By comparison, optical systems and X-ray will require the added expense and processing of a proximity mask.
A major drawback to introducing e-beam technology into large-scale silicon manufacturing has been the low wafer throughput. In the typical electron microscope application, the beam is scanned across the sample, an inherently slow, serial process. This is the same process now employed for constructing photomasks. The writing speed limitation translates into a very low throughput for patterning complex integrated circuits on silicon wafers. Whereas present optical systems can expose roughly tens of wafers per hour, a single electron beam would have substantially lower throughput.
Because of this important limitation, NOVA proposed and evaluated a highly innovative approach to assembling a multiplicity of beams in an array and turning the individual beams off and on. This concept provides for large-scale patterning at significantly higher throughput rates, with a very small excursion of the silicon wafer beneath the beams of less than 250 µm in either direction.
NOVA integrated a number of precision silicon components into a working prototype "source cartridge" as shown in the accompanying schematic diagram. This prototype featured 10,000 individual electron beams precisely aligned on 250 µm centers. A successful demonstration of producing sub-100 nm features in 40 nm PMMA photoresist on silicon substrates was carried out with this system.
|
