Michael Gorman, Ph.D., Professor, Physics, UH
The laboratory experienced developments in four major areas of enterprise this past year in new discoveries, in the reconstruction of the experiment, in training of personnel to analyze data, and in computer-aided data analysis. Funding from the Institute for Space Systems Operations played a crucial role in the progress of this experimentation.
Results recently published in the Physical Review Letter focused upon the ratcheting motion of concentric rings in cellular flames. In these states one or both of the two rings of cell slowly rotate, speeding up and slowing down in a characteristic manner. Prior to this year we had observed four such states. In 1995, we found eighteen additional examples of this motion. The motions of these states are determined by tracking all of the cells using computer animation and visualization.
The experiment was rebuilt in the beginning of the year with a new CCD camera which allowed much higher contrast in the videos of the various flame motions. We now record flame motions using an industrial SVHS video recorder which offers a significantly greater range of capabilities than commercial VHS machines. Using this new equipment, we have re-recorded many of the flame motions so that my collaborator in San Antonio, Kay Robbins, and her students can analyze very long data runs (10 minutes of videotape fills up 2 Gbytes of disk) using image processing techniques.
A number of undergraduates have worked in the lab since the beginning of last year and have proven themselves important to the day-to-day operations of the research effort. Richard Jackson is in charge of computer maintainance. Aaron Jackson is in charge of image processing and video production. Emmanual Convert is in charge of data analysis of the heteroclinic connections. These students are getting firsthand experience on a world-class research project.
ISSO played a very important part in the maintainance of our computer software capabilities. We now have a full complement of software packages for: computer-to-computer communications (we can run demos from San Antonio), image processing, data analysis, and computer-generated videos. We have established a home page on the World Wide Web, which receives hundreds of hits per week. It can be found at: vip.cs.utsa.edu/flames/overview.html.
Back to Table of Contents
Thomas J. Hebert, Ph.D., Associate Professor, Electrical and Computer Engineering, UH
VISUAL RECOGNITION--Results of an automated algorithm for recognizing pliers on a
tabletop. (a) Camera image from the mobile robot. (b) Segmentation into "table
top" and "not tabletop". (c) Segmentation of "tabletop" region
into "objects on the tabletop". (d) Backprojection B of these regions onto a
plane oriented at the camera tilt angle. (e) Identification of objects correponding to
"pliers" using shape descriptors computed from (d) and a predetermined lookup
table for known objects. (f) Forward projection F of identical "pliers" and
zeroing out of "not pliers" regions.
The focus of this project is the development of a mobile robot's capability to visually identify a workspace region that contains tools to be used by astronauts for construction and repair tasks. The robot is expected to identify particular tools within that workspace visually.
Let y denote a row-ordered camera image. A vector map representation x of y consists of polygons, points, lines, and text that represent labeled regions within y. By overlaying smaller regions within larger regions, a multiresolution description of y is possible, whereby x is of dramatically reduced dimension over that of y.
Where the nonlinear operator s(y) segments an image y into a set of polygonal regions, nonlinear operator f(s,y) extracts features of each polygonal region, and y(f) labels each region, x = [ st|yt ]t. Operator y(◊,◊) can be defined through the use of shape descriptors d(◊) that are functions of the polygonal outlines of each labeled region. Were these shape descriptors to be extracted from each region in the 2D camera images, they would be dependent upon view angle. However, shape descriptors extracted from 3D representations of objects are robust as long as an unambiguous object coordinate system can be determined for each object. Therefore, we have defined a backprojection matrix B and a forward projection matrix F that project regions s in the camera image, respectively, onto and from regions B in a plane oriented at the tilt angle of the robot camera. Shape descriptors computed in this projection space are independent of view angle under certain limiting conditions. Therefore
f(s, y) = d(Bs)
and y(f) can be defined to identify objects in y using a lookup table of precomputed d values for a set of known objects.
We designed a short experiment whereby ima-ges from a pair of stereo cameras mounted upon a mobile robot were collected. A small table was set up with a number of tools upon the table top. Images were captured at a sequence of view angles. For the robot to interact with the tools, it must first recognize the table top within the images, then recognize one or more objects on the table.
The future direction of this re-search intends to address the case of obscured objects. Projection space B provides a natural domain for combining all information from a sequence of camera images. This technology will allow obscured regions to be resolved first within the projection space and, next, within the camera images themselves.
Back to Table of Contents
David M Hoffman, Ph.D., Associate Professor, Chemistry, UH
Aluminum (AlN), gallium (GaN) and indium (InN) ni-tride have direct band gaps of 6.2, 3.4 and 1.9 eV, respectively.1-4 They form a continuous alloy system that could, in principle, be fabricated into optical devices that are active at wavelengths ranging from the red to the ultraviolet region of the spectrum. Semicon-ductor devices based on other materials are currently available to operate from the IR to the green wavelengths. Extending the range into the blue region would greatly advance display and imaging technologies. A-nother particularly important band is the 240-280 nm range (~4.75 eV).1 These wavelengths are absorbed by earth's ozone layer. Operating in this range, satellite-to-satellite communications would be shielded from earth monitoring, and imaging array detectors located on the dark side of the earth could provide for a very sensitive surveillance of objects coming up out of the earth's atmosphere.1
A considerable effort has been made in recent years to grow and characterize group 13 nitride thin films, the most useful form of the materials for applications.1-4 Significant difficulties in preparing high quality GaN and InN have been encountered, particularly with high background electron concentrations presumably resulting from nitrogen site vacancies. Nitrogen vacancies are thought to result from the high deposition temperatures required by the film precursors currently in use. Another problem with nitride depositions is a lack of lattice matching substrates.
We recently established that tris(dimethylamido)aluminum and gallium, M(N(CH3)2)3 (M = Al, Ga), and ammonia are excellent precursors for the low temperature (<450 °ree;C) chemical vapor deposition (CVD) of MN films (equation 1).5,6 The amido compounds are moderately air sensitive, volatile solids that are easy to prepare and purify.7-9 Having developed this successful low temperature route to aluminum and gallium nitride, we sought to establish whether the analogous approach of
M(N(CH3)2)3 + NH3 Ĉ MN + 3HN(CH3)2 (1)
M = Al, GA
using tris(amido)indium compounds as precursors to indium nitride would work.
The only previously prepared tris(amido)indiumcomplex, In(N(SiMe3)2)3, is not volatile enough to be practical as a CVD precursor.10 Attempts to synthesize simpler tris(dial-kyl)amido complexes of indium, such as, In(N-(CH3)2)3 or In(N(CH2CH3)2)3, have failed.11 We therefore turned to the synthesis of compounds with hybrid amido ligands having trimethylsilyl and alkyl substituents (equation 2), a strategy that has proved successful. Thus far, we have synthesized In(N(Si(CH3)3)R)3 derivatives where R = CH3, C(CH3)3 and C6H5. A single crystal X-ray structure determination of the ether adduct In(N(Si(CH3)3)(C6H5))3(ether) has been carried out to confirm our formulations (Fig. 1).
InCl3 + 3LiN(Si(CH3)3)R Ĉ In(N(Si(CH3)3)R)3 + 3LiCl
(2) R = CH3, C(CH3)3 and C6H5
Figure 1. Thermal ellipsoid plot of In(N(Si(CH3)3)(C6H5))3(ether)
from a single crystal X-ray diffraction study.
To realize our goal of finding a low temperature method for the preparation of InN films, we need to carry out additional synthetic work, studies on the physical properties of the new precursors, and chemical vapor deposition experiments. Specifically, we plan to carry out the following tasks:
Chemical vapor deposition experiments will be carried out with low pressure and atmospheric pressure apparatus already in my labs. Characterization of the films will be carried out in the Department of Chemistry (IR, UVVIS) and through existing collaboration with Dr. Wei-Kan Chu at the Texas Center for Superconductivity (located on the UH campus) and David Smith's group at the Los Alamos National Laboratory. Synthesis and characterization (IR, NMR and X-ray) of the new precursors will be done entirely in the Department of Chemistry.
References
1Strite, S. and H. Morkoç. "GaN, AlN, and InN: A Review." J. Vac. Sci. Technol.
B 10 (1992): 1237ff.
2Strite, S., M. E. Lin, H. Morkoc. "Progress and Prospects for GaN and the III-V
Nitride Semiconductors." Thin Solid Films 231 (1993): 197.
3Edgar, J. H. "Prospects for Device Implemenation of Wide Band Gap Seminconductor.s.
J. Mater. Res. 7.1 (1992): 235.
4Matsuoka, T. "Current Status of GaN and Related Compounds as Wide-Gap
Semiconductors." J. Cryst. Growth 124 (1992): 433-38.
5Gordon, R. G., D. M. Hoffman, and U. Riaz. "Atmospheric Pressure Chemical Vapor
Deposition of Gallium Nitride Thin Films." Mat. Res. Soc. Symp. Proc. 204 (1991): 95.
6.Gordon, R. G., D. M. Hoffman, and U. Riaz. "Low Temperature Preparation of Gallium
Nitride Thin Films." Mat. Res. Soc. Symp. Proc. 242 (1992): 445-50.
7Waggoner, K. M., M. M. Olmstead, and P. P. Power. "Structural and Spectroscopic
Characterization of the Compounds [Al(NMe2)3]2, [Ga(NMe2)3]2, [(Me2N)2Al{m-N(H)1-Ad}]2
(1-Ad=1-ADAMANTANYL) and [{Me(m-NPh2)al}2NPh(m-C6H4)].Polyhedron 9 (1990): 257-63.
8Nöth, H. and P Konra. "Preparation, Structure, and Some Reactions of
Trisdimethylaminogallane."Z. Naturforsh. B 10 (1955): 234.
9Lappert, M. F., P. P. Power, A. R. Sanger, and R. C. Srivastava. Metal and Metalloid
Amides Chichester: Ellis Horwood, Ltd., and N.Y. et al., John Wiley & Sons, 1980.
10Bürger, H., J. Cichon, U. Goetze, U. Wannagat, and H. J. J. Wismar. "Beitrage zu
Chemie der Silicium-Stickstoff-Verbindungen CVII, Darstellung, Schwingungsspektren und
Normalkoordinatenanalyse von Disilylamiden der Gruppe: M[N(SiMe3)2]3 mit M = al, Ga, und
In."Organomet. Chem. 33 (1971): 1-12.
11Rossetto, G., N. Brianese, A. Camporese, M. Porchia, P. Zanella, R. Bertoncello.
"Synthesis and Characterization of Hexakis(diethylamido)diindium(III) and bis
cyclopentadienyl (diethylamido)indium(III)." Main GroupMet. Chem. 14 (1991): 113.
Contents
ISSO -- Institute for Space Systems Operations
1994-1995 Annual Report
|
|
