Technology Focus Electronics/Computers - NASA€¦ · 27 Suppressing Ghost Diffraction in...

$: Technology Focus Electronics/Computers - NASA€¦ · 27 Suppressing Ghost Diffraction in E-Beam-Written Gratings 28 Target-Tracking Camera for a Metrology System 29 Polarimetric$
Technology Focus

Electronics/Computers

Software

Materials

Mechanics/Machinery

Manufacturing

Bio-Medical

Physical Sciences

Information Sciences

Books and Reports

05-09 May 2009

https://ntrs.nasa.gov/search.jsp?R=20090020533 2020-07-11T22:24:55+00:00Z

NASA Tech Briefs, May 2009 1

INTRODUCTIONTech Briefs are short announcements of innovations originating from research and develop-

ment activities of the National Aeronautics and Space Administration. They emphasize information considered likely to be transferable across industrial, regional, or disciplinary linesand are issued to encourage commercial application.

Availability of NASA Tech Briefs and TSPsRequests for individual Tech Briefs or for Technical Support Packages (TSPs) announced herein shouldbe addressed to

National Technology Transfer CenterTelephone No. (800) 678-6882 or via World Wide Web at www2.nttc.edu/leads/

Please reference the control numbers appearing at the end of each Tech Brief. Infor mation on NASA’s Innovative Partnerships Program (IPP), its documents, and services is also available at the same facility oron the World Wide Web at http://ipp.nasa.gov.

Innovative Partnerships Offices are located at NASA field centers to provide technology-transfer access toindustrial users. Inquiries can be made by contacting NASA field centers listed below.

Ames Research CenterLisa L. Lockyer(650) [email protected]

Dryden Flight Research CenterGregory Poteat(661) [email protected]

Glenn Research CenterKathy Needham(216) [email protected]

Goddard Space Flight CenterNona Cheeks(301) [email protected]

Jet Propulsion LaboratoryAndrew Gray(818) [email protected]

Johnson Space Centerinformation(281) [email protected]

Kennedy Space CenterDavid R. Makufka(321) [email protected]

Langley Research CenterBrian Beaton(757) [email protected]

Marshall Space Flight CenterJim Dowdy(256) [email protected]

Stennis Space CenterRamona Travis(228) 688-3832 [email protected]

Carl Ray, Program ExecutiveSmall Business Innovation Research (SBIR) & Small Business Technology Transfer (STTR) Programs(202) [email protected]

Doug Comstock, DirectorInnovative Partnerships Program Office(202) [email protected]

NASA Field Centers and Program Offices

2 NASA Tech Briefs, May 2009

5 Technology Focus: Wireless

5 Valve-“Health”-Monitoring System

6 Microstrip Antenna for Remote Sensing of SoilMoisture and Sea Surface Salinity

7 Biomedical Wireless Ambulatory Crew Monitor

7 Wireless Avionics Packet To Support Fault Toler-ance for Flight Applications

8 Aerobot Autonomy Architecture

9 Electronics/Computers

9. Submillimeter Confocal Imaging Active Module

10 Traveling-Wave Maser for 32 GHz

11 System Synchronizes Recordings From SeparatedVideo Cameras

12 Piecewise-Planar Parabolic Reflectarray Antenna

13 Reducing Interference in ATC Voice Communication

15 Software

15 EOS MLS Level 1B Data Processing, Version 2.2

15 Auto-Generated Semantic Processing Services

15 Geospatial Authentication

15 Maneuver Automation Software

16 Event Driven Messaging With Role-Based Subscriptions

16 Estimating Relative Positions of Outer-SpaceStructures

17 Manufacturing & Prototyping

17 Fabricating PFPE Membranes for Capillary Electrophoresis

17 Linear Actuator Has Long Stroke and high Resolution

18 Installing a Test Tap on a Metal Battery Case

18 Fabricating PFPE Membranes for MicrofluidicValves and Pumps

21 Materials

21 Room-Temperature-Cured Copolymers forLithium Battery Gel Electrolytes

22 Catalysts for Efficient Production of Carbon Nanotubes

22 Amorphous Silk Fibroin Membranes for Separation of CO2

23 Mechanics/Machinery

23 “Zero-Mass” Noninvasive Pressure Transducers

24 Radial-Electric-Field Piezoelectric Diaphragm Pumps

25 Ejector-Enhanced, Pulsed, Pressure-Gain Combustor

27 Physical Science

27 Suppressing Ghost Diffraction in E-Beam-Written Gratings

28 Target-Tracking Camera for a Metrology System

29 Polarimetric Imaging Using Two Photoelastic Modulators

30 Miniature Wide-Angle Lens for Small-Pixel Electronic Camera

30 Modal Filters for Infrared Interferometry

31 Mo3Sb7–xTex for Thermoelectric Power Generation

32 Two-Dimensional Quantum Model of a Nanotransistor

33 Scanning Miniature Microscopes Without Lenses

34 Manipulating Neutral Atoms in Chip-Based Magnetic Traps

35 Expansion Compression Contacts for Thermoelectric Legs

37 Information Sciences

37 Processing Electromyographic Signals To Recognize Words

38 Physical Principle for Generation of Randomness

38 DSN Beowulf Cluster-Based VLBI Correlator

39 Hybrid NN/SVM Computational System for Optimizing Designs

41 Books & Reports

41 Criteria for Modeling in LES of MulticomponentFuel Flow

41 Computerized Machine for Cutting Space ShuttleThermal Tiles

05-09 May 2009

41 Orbiting Depot and Reusable Lander for LunarTransportation

41 FPGA-Based Networked Phasemeter for a Heterodyne Interferometer

42 Aquarius Digital Processing Unit

43 Bio-Medical

43 Three-Dimensional Optical Coherence Tomography

44 Benchtop Antigen Detection Technique UsingNanofiltration and Fluorescent Dyes

45 Isolation of Precursor Cells From Waste Solid Fat Tissue

45 Identification of Bacteria and Determination of Biological Indicators

46 Further Development of Scaffolds for Regenera-tion of Nerves

47 Chemically Assisted Photocatalytic Oxidation System

47 Use of Atomic Oxygen for Increased Water Contact Angles of Various Polymers for Biomedical Applications

48 Crashworthy Seats Would Afford Superior Protection

49 Open-Access, Low-Magnetic-Field MRI Systemfor Lung Research

49 Microfluidic Mixing Technology for a UniversalHealth Sensor

50 Microfluidic Extraction of Biomarkers UsingWater as Solvent

51 Microwell Arrays for Studying Many Individual Cells

51 Droplet-Based Production of Liposomes

51 Identifying and Inactivating Bacterial Spores


This document was prepared under the sponsorship of the National Aeronautics and Space Administration. Neither the United States Govern-ment nor any person acting on behalf of the United States Government assumes any liability resulting from the use of the information containedin this document, or warrants that such use will be free from privately owned rights.


Technology Focus: Wireless

Upper Patch (thickness not drawn to scale)

Lower Patch (thickness not drawn to scale)

Ground Plane(thickness not drawn to scale)

Coaxial-Inner Conductor

Coaxial-Outer Conductor

Astro-Quartz (0.02”)(εr=3.2)EX-1453 Adhesive (0.002”)(εr=2.75)

Copper

Kapton (0.002”)(εr=3.5)Korex Honeycomb (0.02”)(εr=1.11) Note: εr=Permittivity

Valve-“Health”-Monitoring SystemThis system is adaptable to diverse long-term sensor-data-monitoring applications.Stennis Space Center, Mississippi

A system that includes sensors and data-acquisition, wireless data-communication,and data-processing subsystems has beendeveloped as a means of both real-timeand historical tracking of information in-dicative of deterioration in the mechani-cal integrity and performance of a high-geared ball valve or a linearly actuatedvalve that operates at a temperature be-tween cryogenic and ambient.

In the original application for which thesystem is specifically designed, the valve isof a type used in controlling the flow of apropellant fluid (e.g., liquid hydrogen)

during testing of a rocket engine. Thedata collected by use of this system areused to build a knowledge base of opera-tional characteristics of valves, for the pur-pose of obtaining insight into operationallifetime and enabling prediction of valveoperational lifetimes. This system canstand alone or can be incorporated into anetwork that is part of a higher-level“health”-management system. This systemis also readily adaptable to commercial ap-plications in which there are requirementsfor long-term monitoring of events associ-ated with quantities that may include, but

are not limited to, torsional strain, biaxialstrain, linear strain, cryogenic tempera-tures, ambient temperatures, tripping oflimit switches, current signals rangingfrom 4 to 20 mA, potential signals rangingfrom 0 to 10 V, and/or magnetic fields.

The data collected in the original appli-cation include the number of cryogenicvalve cycles, the total number of all valvecycles, inlet temperature, outlet tempera-ture, valve-body temperature, torsionalstrain, linear strain, valve preload posi-tion, total rotational or linear valve travel,and total number of directional changes.

The Sensor Units and a Base Station together constitute an instrumentation and data-acquisition system that can function alone or can be connected intoa network that is part of a higher-level "health"-management system.

TransmitterReceiver

Regulator

Battery

TransmitterReceiver

Regulator

Battery

TransmitterReceiver

Regulator

Battery

TransmitterReceiver

Regulator

Battery

Transmitter

Regulator

Battery

Microprocessor(Controller)

Microprocessor(Receiver Controller)

Unused Receiver Slot

Unused Receiver Slot

Torsion Sensor(Full-Bridge

Strain Gauge)

Vibration Detector(Wake Up Unit)

Microprocessor(Controller) Microprocessor

(Base-Station Controller)

InternalMemory Bank


Linear-Strain Sensor(Quarter-BridgeStrain Gauge)

Vibration Detector(Wake Up Unit)



TemperatureSensor



Valve-Cycle Counter(Magnetic Switch)



Valve-Travel Sensor(e.g., Linear

Variable-DifferentialTransformer)

Receiver

FlashModule

LANCommunication

Module

PowerAC and Battery

Backup

IRIG InterfaceBus


Events associated with the aforemen-tioned data are recorded and are time-stamped with sufficient precision to en-able synchronization within a timeincrement of 1 ms. The data are then or-ganized into a text file and stored in acompact flash memory card, from whencethe data can be uploaded.

The system (see figure) includes abase station and several self-contained,microprocessor-controlled sensor unitsthat (1) can be mounted remotely fromthe base station and (2) transmit data tothe base station via low-power, short-range [≤ 35 ft (up to about 10 m)] digi-tal radio communication links in the fre-quency band from 902 to 928 MHz. Eachsensor unit has overall dimensions of 3by 2 1⁄2 by 2 in. (about 7.6 by 6.4 by 5.1cm) — small enough to be mounted inthe confined spaces typically available formounting on valves of the type used inthe original rocket-engine-testing appli-cation. Each sensor unit is potted in aflame-retardant epoxy and designed todraw a current of no more than 0.25 A ata supply potential of 9 V, as required forsafe operation in an atmosphere thatmay contain hydrogen. The base stationis not potted; instead, it is mounted in anenclosure that is purged with nitrogen.

Each sensor unit contains two change-able battery packs and a voltage regula-tor that enables bumpless transfer of theload from one battery pack to the other.The temperatures of the battery packsand the microprocessor are monitoredto safeguard against operation outsidetemperature limits. Each sensor unit in-cludes a display-and-control panelthrough which a human technician caneffect setup and can receive a low-bat-tery indication. An interface port for on-board programming and serial commu-nication is also provided. In the case of astrain-sensor unit, to minimize time-av-erage power demand and thereby pro-long battery life, the microprocessor isdesigned to spend most of the time in alow-power sleep mode, from which it isawakened when any valve movement isdetected by a highly sensitive piezoelec-tric vibration-detection subunit.

The base station includes a receivermodule, for each sensor unit, compris-ing a radio receiver and an associated mi-croprocessor. The base station also in-cludes another microprocessor thatserves as the base-station controller, acompact flash module comprising theaforementioned flash memory card andits controller, a local-area-network (LAN)

communication module, a power supplywith battery backup, and an interface toa source of time-stamp signals that con-form to an Inter Range InstrumentationGroup (IRIG) standard. The base-stationcontroller correlates related data fromthe sensor units and generates data-eventlog entries, which are transferred to thecompact flash module. In addition, if thesystem is connected into a network, theselog entries can be transferred to the LANcommunication module for broadcast-ing over the network. The compact flashmodule can be manually removed to ob-tain access to the data stored therein.

Each receiver module maintains com-munications and time synchronization.It relays, to the base-station controller,information on events correlated withsensory and diagnostic informationfrom its sensor unit. Each receiver mod-ule includes an interface port for on-board programming and serial commu-nication with a mobile computer.

This work was done by Scott L. Jensen ofStennis Space Center and George J. Drouantof Jacobs Technology.

Inquiries concerning this technology shouldbe addressed to the Intellectual Property Man-ager at Stennis Space Center (228) 688-1929. SSC-00247-1

This compact, lightweight, dual-fre-quency antenna feed developed for fu-ture soil moisture and sea surface salin-ity (SSS) missions can benefit future soiland ocean studies by lowering mass, vol-ume, and cost of the antenna system. Italso allows for airborne soil moistureand salinity remote sensors operating onsmall aircraft. While microstrip antennatechnology has been developed forradio communications, it has yet to beapplied to combined radar and ra-diometer for Earth remote sensing.

The antenna feed provides a key instru-ment element enabling high-resolution ra-diometric observations with large, deploy-able antennas. The design is based on themicrostrip stacked-patch array (MSPA)used to feed a large, lightweight, deploy-able, rotating mesh antenna for space-borne L-band (≈1 GHz) passive and active

sensing systems. The array consists ofstacked patches to provide dual-frequencycapability and suitable radiation patterns.The stacked-patch microstrip element was

designed to cover the required L-bandcenter frequencies at 1.26 GHz (lowerpatch) and 1.413 GHz (upper patch), withdual-linear polarization capabilities. The

Microstrip Antenna for Remote Sensing of Soil Moisture andSea Surface SalinityThe microstrip array design enables combined radar and radiometer instrumentation for satel-lite or airborne remote sensing. NASA’s Jet Propulsion Laboratory, Pasadena, California

Upper Patch (thickness not drawn to scale)

Lower Patch (thickness not drawn to scale)

Ground Plane(thickness not drawn to scale)

Coaxial-Inner Conductor

Coaxial-Outer Conductor

Astro-Quartz (0.02”)(εr=3.2)EX-1453 Adhesive (0.002”)(εr=2.75)

Copper

Kapton (0.002”)(εr=3.5)Korex Honeycomb (0.02”)(εr=1.11) Note: εr=Permittivity

The Microstrip Stacked-Patch Array incorporates three layers that function as the upper patch, lowerpatch, and ground plane. The lower radar patches sit on a honeycomb structure above the groundplane. The lower patch is fed through the ground plane, while the upper patch acts as a parasitic patch.


dimension of patches produces the re-quired frequencies.

To achieve excellent polarization iso-lation and control of antenna sidelobesfor the MSPA, the orientation of eachstacked-patch element within the array isoptimized to reduce the cross-polariza-tion. A specialized feed-distribution net-work was designed to achieve the re-quired excitation amplitude and phasefor each stacked-patch element.

The patches are thin copper/Kaptonlayers bonded to Astro-Quartz layers. Asillustrated in the figure, threecopper/Kapton/Astro-Quartz layers are

built to function as the upper patch,lower patch, and ground plane. Thelower radar patches sit on a honeycombdielectric structure above the conductingground plane. The honeycomb is filledmostly with air and, therefore, introducesonly a small loss at L-band frequencies.On the top of the radar patches sits an-other honeycomb dielectric structure tosupport the radiometer patches. All ofthe layers and the honeycombs aredrilled to allow attachment to the feedwires to the lower patch (radar). Thelower patch is fed through the groundplane, while the upper patch acts as a par-

asitic patch to introduce the 1.413 GHz.A seven-element stacked patch array

with elements forming a hexagonal pat-tern is the most suitable for space appli-cations; however, a 16-element array witha 4×4 rectangular configuration is betterfor airborne and ground applications.

This work was done by Yahya Ramhat-Samii, Keerti Kona, and Majid Manteghi ofthe University of California at Los Angeles(UCLA); and Steven Dinardo, Don Hunter,Eni Njoku, William Wilson, and SimonYueh of Caltech for NASA’s Jet PropulsionLaboratory. For more information, [email protected]. NPO-44470

Biomedical Wireless Ambulatory Crew MonitorJohn H. Glenn Research Center, Cleveland, Ohio

A compact, ambulatory biometricdata acquisition system has been devel-oped for space and commercial terres-trial use. BioWATCH (Bio medical Wire-less and Ambulatory Telemetry for CrewHealth) acquires signals from biomed-ical sensors using acquisition modulesattached to a common data and powerbus. Several slots allow the user to con-figure the unit by inserting sensor-spe-cific modules. The data are then sentreal-time from the unit over any com-

mercially implemented wireless networkincluding 802.11b/g, WCDMA, 3G.

This system has a distributed computinghierarchy and has a common data con-troller on each sensor module. This allowsfor the modularity of the device along withthe tailored ability to control the cards usinga relatively small master processor. The dis-tributed nature of this system affords themodularity, size, and power consumptionthat betters the current state of the art inmedical ambulatory data acquisition.

A new company was created to marketthis technology.

This work was done by Alan Chmiel andBrad Humphreys of ZIN Technologies forGlenn Research Center. Further information iscontained in a TSP (see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, InnovativePartnerships Office, Attn: Steve Fedor, MailStop 4–8, 21000 Brookpark Road, Cleveland,Ohio 44135. Refer to LEW-18357-1.

Wireless Avionics Packet To Support Fault Tolerance for FlightApplicationsA simple network interface supports fault detection and autonomous fault recovery.NASA’s Jet Propulsion Laboratory, Pasadena, California

In this protocol and packet format,data traffic is monitored by all networkinterfaces to determine the health oftransmitter and subsystems. When fail-ures are detected, the network interfaceapplies its recovery policies to providecontinued service despite the presenceof faults. The protocol, packet format,and interface are independent of thedata link technology used. The currentdemonstration system supports bothcommercial off-the-shelf wireless con-nections and wired Ethernet connec-tions. Other technologies such as 1553or serial data links can be used for thenetwork backbone.

The Wireless Avionics packet is di-vided into three parts: a header, a datapayload, and a checksum. The header

has the following components: magicnumber, version, quality of service, timeto live, sending transceiver, functioncode, payload length, source Applica-tion Data Interface (ADI) address, desti-nation ADI address, sending node ad-dress, target node address, and asequence number.

The magic number is used to identifyWAV packets, and allows the packet for-mat to be updated in the future. Thequality of service field allows routingdecisions to be made based on thisvalue and can be used to route criticalmanagement data over a dedicatedchannel. The time to live value is usedto discard misrouted packets while thesource transceiver is updated at eachhop. This information is used to moni-

tor the health of each transceiver in thenetwork.

To identify the packet type, the func-tion code is used. Besides having a regu-lar data packet, the system supports diag-nostic packets for fault detection andisolation. The payload length specifiesthe number of data bytes in the payload,and this supports variable-length packetsin the network. The source ADI is the ad-dress of the originating interface. Thiscan be used by the destination applica-tion to identify the originating source ofthe packet where the address consists of asubnet, subsystem class within the subnet,a subsystem unit, and the local ADI num-ber. The destination ADI is used to routethe packet to its ultimate destination. Ateach hop, the sending interface uses the


destination address to determine thenext node for the data.

The sending node is the node addressof the interface that is broadcasting thepacket. This field is used to determinethe health of the subsystem that is send-ing the packet. In the case of a packetthat traverses several intermediatenodes, it may be the node address of theintermediate node. The target node isthe node address of the next hop for thepacket. It may be an intermediate node,or the final destination for the packet.

The sequence number is used to iden-tify duplicate packets. Because each in-

terface has multiple transceivers, thesame packet will appear at both re-ceivers. The sequence number allowsthe interface to correlate the receptionand forward a single, unique packet foradditional processing. The subnet fieldallows data traffic to be partitioned intosegregated local networks to supportlarge networks while keeping each sub-net at a manageable size. This also keepsthe routing table small enough so rout-ing can be done by a simple table look-up in an FPGA device.

The subsystem class identifies mem-bers of a set of redundant subsystems,

and, in a hot standby configuration, allmembers of the subsystem class will re-ceive the data packets. Only the activesubsystem will generate data traffic.Specific units in a class of redundantunits can be identified and, if the hotstandby configuration is not used, pack-ets will be directed to a specific subsys-tem unit.

This work was done by Gary L. Block,William D. Whitaker, James W. Dillon, JamesP. Lux, and Mohammad Ahmad of Caltech forNASA’s Jet Propulsion Laboratory. For more in-formation, contact [email protected]

An architecture for autonomous op-eration of an aerobot (i.e., a roboticblimp) to be used in scientific explo-ration of planets and moons in theSolar system with an atmosphere (suchas Titan and Venus) is undergoing de-velopment. This architecture is also ap-plicable to autonomous airships thatcould be flown in the terrestrial atmos-phere for scientific exploration, mili-tary reconnaissance and surveillance,and as radio-communication relay sta-tions in disaster areas. The architecturewas conceived to satisfy requirements toperform the following functions: • Vehicle safing, that is, ensuring the in-

tegrity of the aerobot during its entiremission, including during extendedcommunication blackouts.

• Accurate and robust autonomousflight control during operation in di-verse modes, including launch, deploy-ment of scientific instruments, longtraverses, hovering or station-keeping,and maneuvers for touch-and-go sur-face sampling.

• Mapping and self-localization in theabsence of a global positioning system.

• Advanced recognition of hazards andtargets in conjunction with tracking of,

and visual servoing toward, targets, all toenable the aerobot to detect and avoidatmospheric and topographic hazardsand to identify, home in on, and hoverover predefined terrain features or othertargets of scientific interest.The architecture is an integrated com-

bination of systems for accurate and ro-bust vehicle and flight trajectory control;estimation of the state of the aerobot;perception-based detection and avoid-ance of hazards; monitoring of the in-tegrity and functionality (“health”) ofthe aerobot; reflexive safing actions;multi-modal localization and mapping;autonomous planning and execution ofscientific observations; and long-rangeplanning and monitoring of the missionof the aerobot. The prototype JPL aero-bot (see figure) has been tested exten-sively in various areas in the CaliforniaMojave desert.

This work was done by Alberto Elfes, JefferyL. Hall, Eric A. Kulczycki, Jonathan M.Cameron, Arin C. Morfopoulos, Daniel S.Clouse, James F. Montgomery, Adnan I.Ansar, and Richard J. Machuzak of JPL forNASA’s Jet Propulsion Laboratory. For moreinformation, contact [email protected]

Testing of a Prototype Aerobot is an essentialpart of the continuing effort to develop an ar-chitecture for autonomous operation of aerialvehicles in the atmosphere of Earth as well as ofremote planets. These photos show the JPL aer-obot during autonomous flight tests conductedat the El Mirage dry lake in the Mojave Desert.The top image shows vehicle liftoff, and the bot-tom image shows the aerobot in autonomousflight mode.

Aerobot Autonomy Architecture Potential applications include scientific exploration, military surveillance, and radio relay. NASA’s Jet Propulsion Laboratory, Pasadena, California


Electronics/Computers

Submillimeter Confocal Imaging Active ModuleThis system could be used to image shorter objects between taller ones.NASA’s Jet Propulsion Laboratory, Pasadena, California

The term “submillimeter confocal im-aging active module” (SCIAM) denotesa proposed airborne coherent imagingradar system that would be suitable foruse in reconnaissance, surveillance, andnavigation. The development of theSCIAM would include utilization and ex-tension of recent achievements inmonolithic microwave integrated cir-cuits capable of operating at frequenciesup to and beyond a nominal radio fre-quency of 340 GHz. Because the SCIAMwould be primarily down-looking (incontradistinction to primarily side-look-ing), it could be useful for imagingshorter objects located between tallerones (for example, objects on streets be-tween buildings). The SCIAM would uti-lize a confocal geometry to obtain highcross-track resolution, and would beamenable to synthetic-aperture process-ing of its output to obtain high along-track resolution.

The SCIAM (see figure) would in-clude multiple (two in the initial ver-sion) antenna apertures, separated fromeach other by a cross-track baseline ofsuitable length (e.g., 1.6 m). These aper-tures would both transmit the illuminat-ing radar pulses and receive the returns.A common reference oscillator wouldgenerate a signal at a controllable fre-quency of (340 GHz + Δf )/N, where Δfis an instantaneous swept frequency dif-ference and N is an integer. The outputof this oscillator would be fed to a fre-quency-multiplier-and-power-amplifiermodule to obtain a signal, at 340 GHz +Δf, that would serve as both the carriersignal for generating the transmittedpulses and a local-oscillator (LO) signalfor a receiver associated with each an-tenna aperture.

Because duplexers in the form ofcirculators or transmit/receive (T/R)switches would be lossy and extremelydifficult to implement, the antennaapertures would be designed accord-ing to a spatial-diplexing scheme, inwhich signals would be coupled in andout via separate, adjacent transmittingand receiving feed horns. This schemewould cause the transmitted and re-

ceived beams to be aimed in slightlydifferent directions, and, hence, tonot overlap fully on the targets on theground. However, a preliminary analy-sis has shown that the loss of overlapwould be small enough that the result-ing loss in signal-to-noise ratio (SNR)would be much less than the SNR lossassociated with the use of a 340-GHzT/R switch.

In each antenna aperture, the receiv-ing and transmitting feed horns wouldface a reflector structure that would bedesigned partly according to both aFresnel-lens principle to minimize an-tenna volume and partly according to adiffraction-grating principle so that thebeam direction in the cross-track planewould become dependent on the signalfrequency. The Fresnel surfaces would

The SCIAM would initially include two antenna apertures, through each of which radar signals wouldbe both transmitted and received. The antenna structures in each aperture would be frequency-dis-persive so that frequency scanning could be used to effect cross-track scanning.

TxLORx TxLORx

AntennaApertures

AntennaApertures

Reference Oscillator,(340 GHz + Δf)/N

Outputs toDigital Signal Processor

Tx = TransmitterRx = ReceiverLO = Local OscillatorADCs = Analog-to-Digital

Converters

SIMPLIFIED BLOCK DIAGRAM

ADCs

Cross-Track Scanning

ALONG-TRACK VIEW SHOWING CROSS-TRACK SCANNING OF BEAMS


be confocal paraboloids having focal-length increments of a half wavelength,and the sizes of the Fresnel steps wouldbe chosen to obtain a desired amount ofangular deviation for a given amount offrequency tuning. For example, accord-ing to one tentative design, sweepingthe radio frequency from 335 to 345GHz would cause the beam to scan across-track ground swath 30 m wide

from a height of 1 km. Through post-de-tection processing of the return signalsreceived via the two apertures, a cross-track image resolution of 27 cm wouldbe obtained; in effect, the 30-m cross-track swath could be divided into 111pixels of 27-cm width. Comparablealong-track resolution would be ob-tained through synthetic-aperture post-processing.

This work was done by John Hong,Imran Mehdi, Peter Siegel, Goutam Chat-topadhyay, and Thomas Cwik of Caltech;Mark Rowell of the University of Califor-nia, Santa Barbara; and John Hacker ofRockwell Scientific Company LLC forNASA’s Jet Propulsion Laboratory. Furtherinformation is contained in a TSP (seepage 1).NPO-42924

This 32-GHz Maser would include an improved slow-wave structure containing alternating ruby-filled and evanescent-wave sections.

About 5.6 cm

About 8.6 cm

Cavity Resonant at Signal and Pump

Frequencies

Resonant Ferrite Isolator

Evanescent-Wave Section

Pump Input Terminal

Ruby-Filled Section

Sapphire-FilledSection

Signal Input Terminal

Slow-Wave Structure

PERSPECTIVE VIEW

PARTLY SCHEMATIC VIEW (NOT TO SCALE)

Signal OutputTerminal

WR-15 Pump InputWaveguide

Slow-WaveStructure

WR-28 InputWaveguide

WR-28 Output Waveguide

The figure depicts a traveling-waveruby maser that has been designed(though not yet implemented in hard-ware) to serve as a low-noise amplifierfor reception of weak radio signals in

the frequency band of 31.8 to 32.3 GHz.The design offers significant improve-ments over previous designs of 32-GHztraveling-wave masers. In addition, rela-tive to prior designs of 32-GHz ampli-

fiers based on high-electron-mobilitytransistors, this design affords higherimmunity to radio-frequency interfer-ence and lower equivalent input noisetemperature.

Traveling-Wave Maser for 32 GHzSignificant improvements over prior 32-GHz low-noise amplifiers are anticipated.NASA’s Jet Propulsion Laboratory, Pasadena, California


In addition to the basic frequency-band and low-noise requirements, theinitial design problem included a re-quirement for capability of operation ina closed-cycle helium refrigerator at atemperature ≤4 K and a requirementthat the design be mechanically simpli-fied, relative to prior designs, in order tominimize the cost of fabrication and as-sembly. Previous attempts to build 32-GHz traveling-wave masers involved theuse of metallic slow-wave structures com-prising coupled transverse electromag-netic (TEM)-mode resonators that weresubject to very tight tolerances and,hence, were expensive to fabricate andassemble. Impedance matching for cou-pling signals into and out of these ear-lier masers was very difficult.

A key feature of the design is a slow-wave structure, the metallic portions ofwhich would be mechanically relativelysimple in that, unlike in prior slow-wavestructures, there would be no internalmetal steps, irises, or posts. The metallicportions of the slow-wave structure wouldconsist only of two rectangular metalwaveguide arms. The arms would contain

sections filled with the active material(ruby) alternating with evanescent-wavesections. This structure would be trans-parent in both the signal-frequency band(the aforementioned range of 31.8 to32.3 GHz) and the pump-frequency band(65.75 to 66.75 GHz), and would imposelarge slowing factors in both frequencybands. Resonant ferrite isolators wouldbe placed in the evanescent-wave sectionsto provide reverse loss needed to sup-press reverse propagation of power at thesignal frequency.

This design is expected to afford alarge gain-bandwidth product at the sig-nal frequency and efficient coupling ofthe pump power into the paramagneticspin resonances of the ruby sections.The more efficiently the pump powercould be thus coupled, the more effi-ciently it could be utilized and the heatload on the refrigerator correspondinglyreduced. To satisfy the requirement foroperation over the 0.5-GHz-wide signal-frequency band, the paramagnetic spinresonances would be broadened by ap-plying a magnetic field having a lineargradient along the slow-wave structure.

The gradients in the two arms are offsetin order to compensate for the gaps ofthe evanescent sections.

The two arms of the slow-wave struc-ture would be connected into a U-shaped assembly. At the base of the U (atthe right end in the figure) there wouldbe a cavity that would be simultaneouslyresonant in the TE301 mode at the signalfrequency and in the TE403 mode at thepump frequency. The pump powerwould be injected into this cavity fromtwo dielectric-filled waveguides thatwould be beyond cut-off at the signal fre-quency. The dielectric-filled waveguideswould be excited from the two outputarms of an electric-field-plane T junc-tion. The signal power would enter andleave via WR-28 waveguides connectedto opposite ends of the U-shaped assem-bly. Impedance matching would be ef-fected by means of dielectric-filled (sap-phire) sections of the waveguide armsnear their input and output ends.

This work was done by James Shell andRobert Clauss of Caltech for NASA’s Jet Propul-sion Laboratory. Further information is con-tained in a TSP (see page 1). NPO-41273

A system of electronic hardware andsoftware for synchronizing recordingsfrom multiple, physically separatedvideo cameras is being developed, pri-marily for use in multiple-look-anglevideo production. The system, the timecode used in the system, and the under-lying method of synchronization uponwhich the design of the system is basedare denoted generally by the term “Geo-TimeCodeTM.” The system is embodiedmostly in compact, lightweight, portableunits (see figure) denoted video time-code units (VTUs) — one VTU for eachvideo camera. The system is scalable inthat any number of camera recordingscan be synchronized. The estimated re-tail price per unit would be about $350(in 2006 dollars).

The need for this or another synchro-nization system external to video cam-eras arises because most video camerasdo not include internal means for main-taining synchronization with other videocameras. Unlike prior video-camera-syn-chronization systems, this system doesnot depend on continuous cable or

radio links between cameras (however, itdoes depend on occasional cable linkslasting a few seconds). Also, whereas thetime codes used in prior video-camera-synchronization systems typically repeat

after 24 hours, the time code used in thissystem does not repeat for slightly morethan 136 years; hence, this system ismuch better suited for long-term deploy-ment of multiple cameras.

System Synchronizes Recordings From Separated Video CamerasA large, immobile timing infrastructure is not needed.Stennis Space Center, Mississippi

This Mockup of a VTU is based on a prototype VTU circuit powered by a standard 9-volt battery. Tothe extent possible, production VTUs would be made of low-power, precision electronic componentsthat are already commercially available.


Each VTU contains a free-running, ex-tremely stable clock, based on a 32,768-Hz (215-Hz) quartz-crystal oscillator. Theclock begins a binary count up fromzero when reset and continues countingup until reset again (or until it automat-ically restarts from zero when the timecode repeats after more than 136 years).Each VTU also contains digital and ana-log audio circuitry required for synchro-nization of video recording.

The GeoTimeCode is a variant of theInter Range Instrumentation Group B(IRIG-B) time code, which is widely usedin the aerospace industry. The GeoT-imeCode can easily be converted toother standard time codes, includingthe Society of Motion Picture and Televi-sion Engineers (SMPTE) time code. TheGeoTimeCode is similar enough to theIRIG-B time code that software can eas-ily be adapted to read either code.

A VTU can be synchronized to a Uni-versal Time source (e.g., an Internet

time server or a radio time signal) or toother, possibly distant VTUs by use of acomputer equipped with the appropri-ate software and ancillary electronichardware. Optionally, without using acomputer, multiple VTUs can be syn-chronized with each other by temporar-ily connecting them together via stan-dard patch cables and pressing a resetbutton. At the instant when synchroniza-tion is performed, the synchronization isaccurate to within less than a millisec-ond. Synchronization can be done ei-ther before or after a video recording ismade; the clock in a VTU is stable andaccurate enough that as long as synchro-nization is performed within about 8hours of recording, timing is accurate towithin 0.033 second (a typical videoframe period).

A portion of the time code is reservedfor a serial number that identifies eachVTU and, hence, the camera fromwhich each recording is taken. Another

portion of the time code is reserved forevent markers, which can be added man-ually during recording by means of apushbutton switch. Each event markerincludes an event number from acounter that is incremented for eachevent. The serial numbers and eventmarkers can be used to identify specificimage sequences during post processingof video images by editing software.

This work was done by William “Bud”Nail, William L. Nail, Jasper M. Nail, andDuong T. Le of Technological Services Co. forStennis Space Center.

Inquiries concerning rights for the commercialuse of this invention should be addressed to:

Technological Services Company100 Street A, Suite BPicayune, MS 39466(601) 799-2403E-mail: [email protected] to SSC-00253, volume and number

of this NASA Tech Briefs issue, and thepage number

Piecewise-Planar Parabolic Reflectarray AntennaPerformance is equalized in horizontal and vertical polarizations.NASA’s Jet Propulsion Laboratory, Pasadena, California

The figure shows a dual-beam, dual-polarization Ku-band antenna, the re-flector of which comprises an assemblyof small reflectarrays arranged in a piece-wise-planar approximation of a parabolicreflector surface. The specific antennadesign is intended to satisfy require-ments for a wide-swath spaceborne radaraltimeter, but the general principle ofpiecewise-planar reflectarray approxima-tion of a parabolic reflector also offersadvantages for other applications inwhich there are requirements for wide-swath antennas that can be stowed com-pactly and that perform equally in bothhorizontal and vertical polarizations.

The main advantages of using flat (e.g.,reflectarray) antenna surfaces instead ofparaboloidal or parabolic surfaces is thatthe flat ones can be fabricated at lowercost and can be stowed and deployedmore easily. Heretofore, reflectarray an-tennas have typically been designed to re-side on single planar surfaces and to em-ulate the focusing properties of, variously,paraboloidal (dish) or parabolic anten-nas. In the present case, one approxi-mates the nominal parabolic shape byconcatenating several flat pieces, whilestill exploiting the principles of the pla-nar reflectarray for each piece.

Prior to the conception of the presentdesign, the use of a single large reflectar-ray was considered, but then abandonedwhen it was found that the directional

and gain properties of the antennawould be noticeably different for thehorizontal and vertical polarizations.The reason for this difference in per-

Five Flat Panels are arranged in a piecewise-planar approximation of a parabolic reflector surface.Each panel is a reflectarray designed to emulate the corresponding part of the parabolic reflector.

Vertical-Polarization Feed

Horizontal-Polarization Feed

ReflectarrayPanels

Projected Aperture2.16 x 0.35 m,Focal Length

1.125 m


formance is related to strong spatial vari-ations in phase, including phase wraps(phase variations in excess of 360°).

By arranging small reflectarrays in apiecewise-planar approximation of aparabola, instead of constructing onelarge reflectarray on a single planar

surface, one minimizes the number ofphase wraps per panel and reduces theangle of incidence at each reflectarraypatch. This makes it possible to simul-taneously maximize the vertical- andhorizontal-polarization gains, to im-prove the radiation pattern, and re-

duce sensitivity to fabrication and ad-justment errors.a

This work was done by Richard Hodgesand Mark Zawadzki of Caltech for NASA’s JetPropulsion Laboratory. For more information,contact [email protected]

Reducing Interference in ATC Voice CommunicationDigital signal processing would be used to suppress unwanted signals.NASA’s Jet Propulsion Laboratory, Pasadena, California

Three methods have been proposedto be developed to enable reduction ofthe types of interference that oftenoccur among voice-communicationradio signals involved in air-traffic con-trol (ATC). For historical reasons andfor compatibility with some navigationsystems, control towers and aircraft useamplitude modulation (AM) for voicecommunication. In the presence of twosimultaneous AM transmissions in thesame frequency channel, what is heardthrough a receiver includes not onlythe audio portions of both transmis-sions but also an audio heterodyne sig-nal at the difference between the car-rier frequencies of the transmissions (asa practical matter, the carrier frequen-cies almost always differ somewhat).The situation is further complicated bymultiple heterodyne signals in the pres-ence of more than two simultaneoustransmissions. Even if one of the trans-missions does not include AM becauseof a transmitter malfunction or becausea transmitter was inadvertently turnedon or left on, the heterodyne signalmakes it difficult to understand theaudio of the other transmission. Theproposed methods would utilize digitalsignal processing to counteract thistype of interference.

In the first of the three methods, apost-detection audio digital signalprocessor (DSP) in a receiver would re-

duce the level of the heterodyne signalsignificantly. The DSP would be a self-contained unit that would be connectedbetween (1) the output terminal of theATC receiver audio circuitry and (2) aloudspeaker and/or headphones. TheDSP would use a well-understood least-mean-square (LMS) algorithm to auto-matically adjust the coefficients of a fi-nite-impulse-response filter in order tominimize the amplitude of such highlycorrelated signals as sine waves (includ-ing audio heterodyne signals). The DSPwould operate without intervention bythe human operator.

In the event that the first method asdescribed thus far did not reduce inter-ference sufficiently, it could be sup-planted or augmented by a variant inwhich a DSP would be added to the lastintermediate-frequency (IF) stage of thereceiver, where it would be possible toeffect improvements through increaseddynamic range and linearity and the op-portunity to shape the IF pass band foroptimum rejection of other types of in-terference.

The second method, involving inde-pendent sideband reception, could beused alone or in combination with thefirst method. This method would ex-ploit the fact that (1) the two simultane-ously transmitted signals would not havethe same carrier frequency and (2) theupper sideband would yield a higher

signal-to-noise ratio (SNR) for thehigher-carrier-frequency signal whilethe lower sideband would yield a higherSNR for the lower-carrier-frequency sig-nal. Some development would be neces-sary to determine the best way to makeuse of the two signals.

In the third method, multiple anten-nas would be used for reception andtheir outputs would be combined by anadaptive beam former that would usethe same LMS algorithm as that of thefirst method. In this case, the weightingbetween the antennas would be ad-justed to minimize the coherent com-ponent of the received signal. Amethod equivalent to this one has beenused in microwave data communica-tion.

This work was done by John O. Battle ofCaltech for NASA’s Jet Propulsion Laboratory.

In accordance with Public Law 96-517,the contractor has elected to retain title to thisinvention. Inquiries concerning rights for itscommercial use should be addressed to:

Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099(818) 354-2240E-mail: [email protected] to NPO-43950, volume and number

of this NASA Tech Briefs issue, and thepage number.


Software

EOS MLS Level 1B DataProcessing, Version 2.2

A computer program performs level-1B processing (the term “1B” is ex-plained below) of data from observationsof the limb of the Earth by the Earth Ob-serving System (EOS) Microwave LimbSounder (MLS), which is an instrumentaboard the Aura spacecraft. This soft-ware accepts, as input, the raw EOS MLSscientific and engineering data and theAura spacecraft ephemeris and attitudedata. Its output consists of calibrated in-strument radiances and associated engi-neering and diagnostic data. [This soft-ware is one of several computerprograms, denoted product generationexecutives (PGEs), for processing EOSMLS data. Starting from level 0 (repre-senting the aforementioned raw data,the PGEs and their data products are de-noted by alphanumeric labels (e.g., 1Band 2) that signify the successive stagesof processing.]

At the time of this reporting, this soft-ware is at version 2.2 and incorporates im-provements over a prior version thatmake the code more robust, improve cal-ibration, provide more diagnostic out-puts, improve the interface with the Level2 PGE, and effect a 15-percent reductionin file sizes by use of data compression.

This program was written by VincentPerun, Robert Jarnot, Herbert Pickett,Richard Cofield, Michael Schwartz, and PaulWagner for Caltech for NASA’s Jet PropulsionLaboratory.

This software is available for commercial li-censing. Please contact Karina Edmonds ofthe California Institute of Technology at(626) 395-2322. Refer to NPO-44841.

Auto-Generated SemanticProcessing Services

Auto-Generated Semantic Processing(AGSP) Services is a suite of softwaretools for automated generation of othercomputer programs, denoted cross-plat-form semantic adapters, that support in-teroperability of computer-based com-munication systems that utilize a varietyof both new and legacy communicationsoftware running in a variety of operat-ing-system/computer-hardware combi-nations. AGSP has numerous potentialuses in military, space-exploration, andother government applications as well as

in commercial telecommunications. Thecross-platform semantic adapters take ad-vantage of common features of com-puter-based communication systems toenforce semantics, messaging protocols,and standards of processing of streams ofbinary data to ensure integrity of dataand consistency of meaning among in-teroperating systems.

The auto-generation aspect of AGSPServices reduces development time andeffort by emphasizing specification andminimizing implementation: In effect,the design, building, and debugging ofsoftware for effecting conversionsamong complex communication proto-cols, custom device mappings, andunique data-manipulation algorithms isreplaced with metadata specificationsthat map to an abstract platform-inde-pendent communications model. AGSPServices is modular and has been shownto be easily integrable into new andlegacy NASA flight and ground commu-nication systems.

This program was written by Rodney Davisand Greg Hupf of Command and ControlTechnologies Corp. for Kennedy Space Center.For further information, contact:

Command and Control Technologies Corp.Rodney Davis1425 Chaffee Drive, Suite 1Titusville, FL 32780-7900Phone No.: (321) [email protected] Inquiries concerning rights for the commer-

cial use of this invention should be addressedto the Kennedy Innovative Partnerships Of-fice at (321) 861-7158. Refer to KSC-13072.

Geospatial AuthenticationA software package that has been de-

signed to allow authentication for deter-mining if the rover(s) is/are within a set ofboundaries or a specific area to access crit-ical geospatial information by using GPSsignal structures as a means to authenti-cate mobile devices into a network wire-lessly and in real-time has been developed.The advantage lies in that the system onlyallows those with designated geospatialboundaries or areas into the server.

The Geospatial Authentication soft-ware has two parts — Server and Client.The server software is a virtual privatenetwork (VPN) developed in Linux oper-ating system using Perl programming lan-guage. The server can be a stand-alone

VPN server or can be combined withother applications and services. Theclient software is a GUI Windows CE soft-ware, or Mobile Graphical Software, thatallows users to authenticate into a net-work. The purpose of the client softwareis to pass the needed satellite informationto the server for authentication.

This work was done by Stacey D. Lyle ofGeospatial Research Innovation Design forNASA’s Stennis Space Center.

Inquiries concerning rights for its commer-cial use should be addressed to:

Dr. Stacey D. Lyle, RLPSConrad Blucher Institute of Surveying and Science 6300 Ocean DriveTexas A&M UniversityCorpus Christi, TX 78412Phone No. : (361) 825-3712 Fax: (361) 825-5848E-mail: [email protected] to SSC-00282, volume and number

of this NASA Tech Briefs issue, and thepage number

Maneuver Automation Software

The Maneuver Automation Software(MAS) automates the process of generat-ing commands for maneuvers to keep thespacecraft of the Cassini-Huygens missionon a predetermined prime mission trajec-tory. Before MAS became available, a teamof approximately 10 members had to workabout two weeks to design, test, and imple-ment each maneuver in a process that in-volved running many maneuver-relatedapplication programs and then seriallyhanding off data products to other parts ofthe team. MAS enables a three-memberteam to design, test, and implement a ma-neuver in about one-half hour after Navi-gation has process-tracking data. MAS ac-cepts more than 60 parameters and 22files as input directly from users.

MAS consists of Practical Extraction andReporting Language (PERL) scripts thatlink, sequence, and execute the maneu-ver-related application programs: “Push-ing a single button” on a graphical user in-terface causes MAS to run navigationprograms that design a maneuver; pro-grams that create sequences of commandsto execute the maneuver on the space-craft; and a program that generates pre-dictions about maneuver performanceand generates reports and other files thatenable users to quickly review and verify


the maneuver design. MAS can also gener-ate presentation materials, initiate elec-tronic command request forms, andarchive all data products for future refer-ence.

This program was written by Hal Uffel-man, Troy Goodson, Michael Pellegrin, LynnStavert, Thomas Burk, David Beach, Joel Sig-norelli, Jeremy Jones, Yungsun Hahn, AhlamAttiyah, and Jeannette Illsley of Caltech forNASA’s Jet Propulsion Laboratory.


Event Driven MessagingWith Role-Based Subscrip-tions

Event Driven Messaging with Role-Based Subscriptions (EDM-RBS) is aframework integrated into the ServiceManagement Database (SMDB) to allowfor role-based and subscription-based de-livery of synchronous and asynchronousmessages over JMS (Java Messaging Serv-ice), SMTP (Simple Mail Transfer Proto-col), or SMS (Short Messaging Service).This allows for 24/7 operation with usersin all parts of the world. The softwareclassifies messages by triggering datatype, application source, owner of datatriggering event (mission), classification,sub-classification and various other sec-ondary classifying tags. Messages arerouted to applications or users based onsubscription rules using a combinationof the above message attributes.

This program provides a frameworkfor identifying connected users andtheir applications for targeted deliveryof messages over JMS to the client appli-

cations the user is logged into. EDM-RBS provides the ability to send notifica-tions over e-mail or pager rather thanhaving to rely on a live human to do it. Itis implemented as an Oracle applicationthat uses Oracle relational databasemanagement system intrinsic functions.It is configurable to use Oracle AQ JMSAPI or an external JMS provider for mes-saging. It fully integrates into the event-logging framework of SMDB (SubnetManagement Database).

EDB-RBS is currently integrated intothe SMDB component of the ServicePreparation Subsystem of the DSN (DeepSpace Network). Users worldwide rely onthis service for notification confirmingacceptance or denial of submitted prod-ucts in support of mission tracking activi-ties. DSN operations rely on this servicefor asynchronous notifications of prob-lems needing human intervention. Thisinnovation is a core service supportingthe concept of lights-out operations andoperations with human intervention byexception, a plan to improve the level ofservice while reducing operations costs.

This work was done by Tung Bui, BachBui, Shantanu Malhotra, Fannie Chen,Rachel Kim, Christopher Allen, Ivy Luong,George Chang, and Silvino Zendejas of Cal-tech and Syed Sadaqathulla of Raytheon forNASA’s Jet Propulsion Laboratory.


Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099E-mail: [email protected]

Refer to NPO-45015, volume and numberof this NASA Tech Briefs issue, and thepage number.

Estimating Relative Posi-tions of Outer-Space Struc-tures

A computer program estimates therelative position and orientation of twostructures from measurements, made byuse of electronic cameras and laserrange finders on one structure, of dis-tances and angular positions of fiducialobjects on the other structure. The pro-gram was written specifically for use indetermining errors in the alignment oflarge structures deployed in outer spacefrom a space shuttle.

The program is based partly on equa-tions for transformations among the var-ious coordinate systems involved in themeasurements and on equations that ac-count for errors in the transformationoperators. It computes a least-squares es-timate of the relative position and orien-tation. Sequential least-squares esti-mates, acquired at a measurement rateof 4 Hz, are averaged by passing themthrough a fourth-order Butterworth fil-ter. The program is executed in a com-puter aboard the space shuttle, and itsposition and orientation estimates aredisplayed to astronauts on a graphicaluser interface.

This program was written by HarryBalian, William Breckenridge, and PaulBrugarolas of Caltech for NASA’s Jet Propul-sion Laboratory.



Manufacturing & Prototyping

Fabricating PFPE Membranes for Capillary ElectrophoresisPrecisely sized and positioned holes are defined by photomasks.NASA’s Jet Propulsion Laboratory, Pasadena, California

A process has been developed for fab-ricating perfluoropolyether (PFPE)membranes that contain microscopicholes of precise sizes at precise locations.The membranes are to be incorporatedinto “laboratory-on-a-chip” microfluidicdevices to be used in performing capil-lary electrophoresis.

The present process is a modified ver-sion of part of the process, described inthe immediately preceding article, thatincludes a step in which a liquid PFPElayer is cured into solid (membrane)form by use of ultraviolet light. In thepresent process, one exploits the factthat by masking some locations to pre-vent exposure to ultraviolet light, one

can prevent curing of the PFPE in thoselocations. The uncured PFPE can bewashed away from those locations in thesubsequent release and cleaning steps.Thus, holes are formed in the mem-brane in those locations.

The most straightforward way to imple-ment the modification is to use, duringthe ultraviolet-curing step, an ultravioletphotomask similar to the photomasksused in fabricating microelectronic de-vices. In lieu of such a photomask, onecould use a mask made of any pattern-able ultraviolet-absorbing material (forexample, an ink or a photoresist).

This work was done by Michael C. Lee,Peter A. Willis, and Frank Greer of Caltech

and Jason Rolland of Liquidia TechnologiesInc. for NASA’s Jet Propulsion Laboratory.


Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099E-mail: [email protected] to NPO-45782, volume and number


Linear Actuator Has Long Stroke and High Resolution There are potential applications in precise measurement and precise fabrication. NASA’s Jet Propulsion Laboratory, Pasadena, California

The term “precision linear actuator,direct drive” (“PLADD”) refers to a ro-bust linear actuator designed to be capa-ble of repeatedly performing, over a life-time of the order of 5 to 10 years,positioning maneuvers that include, var-iously, submicron increments or slews ofthe order of a centimeter. The PLADD iscapable of both long stroke (120 mm)and high resolution (repeatable incre-

ments of 20 nm). Unlike precise linearactuators of prior design, the PLADDcontains no gears, levers, or hydraulicconverters. The PLADD, now at the pro-totype stage of development, is intendedfor original use as a coarse-positioningactuator in a spaceborne interferometer.The PLADD could also be adapted toterrestrial applications in which thereare requirements for long stroke and

high resolution: potential applicationsinclude medical imaging and fabricationof semiconductor devices.

The PLADD (see figure) includes acommercially available ball-screw actua-tor driven directly by a commerciallyavailable three-phase brushless DCmotor. The ball-screw actuator com-prises a spring-preloaded ball nut on aball screw that is restrained against rota-

This Partly Schematic Cross Section depicts some of the components of the PLADD.

Screw

Outer Bellows Bearing Housing Inner Bellows

Angular Contact Bearings

Bearing Preload Spring Carbon Fiber Tube

Motor Rotor/Windings


Fabricating PFPE Membranes for Microfluidic Valves and PumpsThis process contributes to development of “laboratory-on-a-chip” devices.NASA’s Jet Propulsion Laboratory, Pasadena, California

A process has been developed for fab-ricating membranes of a perfluoropoly-ether (PFPE) and integrating them intovalves and pumps in “laboratory-on-a-chip” microfluidic devices. Membranesof poly(tetrafluoroethylene) [PTFE]and poly(dimethylsilane) [PDMS] havebeen considered for this purpose andfound wanting. By making it possible touse PFPE instead of PTFE or PDMS, thepresent process expands the array of op-tions for further development of mi-crofluidic devices for diverse applica-tions that could include detection ofbiochemicals of interest, detection oftoxins and biowarfare agents, synthesis

and analysis of proteins, medical diagno-sis, and synthesis of fuels.

To be most useful, a membrane mate-rial for a microfluidic valve or pumpshould be a chemically inert elastomer.PTFE is highly chemically inert and is athermoplastic and, therefore, subject tocold flow and creep. Also, proceduresfor fabricating PTFE membranes are ex-cessively complex. PDMS is an elastomerthat has been used in prior microfluidicdevices but, undesirably, reacts chemi-cally with some liquids (acetonitrile,acids, and fuels) that might be requiredto be handled by microfluidic devices insome applications. On the other hand,

the PFPE in question has elastomericproperties similar to those of PDMS anda degree of chemical inertness similar tothat of PTFE.

The specific membrane material towhich the present process applies is acommercially available, ultraviolet-cur-able PFPE. A microfluidic device of thetype to which the process applies con-sists mainly of this PFPE sandwiched be-tween two plates of a highly chemicallyresistant, low-thermal-expansion borosil-icate glass manufactured by the floatmethod. Heretofore, there have beentwo obstacles to fabrication of microflu-idic devices from this combination of

tion as described below. The motor iscoupled directly (that is, without an in-tervening gear train) to a drive link that,in turn, is coupled to the ball nut. Byeliminating the gear train, the direct-drive design eliminates the complexity,backlash, and potential for misalign-ment associated with a gear train.

To prevent inadvertent movement,there is a brake that includes flexuredlevers compressed against the drivelink by preload springs. This is apower-off brake: There are also piezo-electric stacks that can be activated tooppose the springs and push the leversaway from the drive link. Hence,power must be applied to the piezo-electric stacks to release the drive linkfrom braking.

To help ensure long operational life,all of the mechanical drive componentsare immersed in an oil bath within her-metically sealed bellows. The outer endof the bellows holds the outer end of theball screw, thereby preventing rotationof the ball screw.

Positioning is controlled by an elec-tronic control system that includes dig-ital and analog subsystems that interactwith the motor and brake and with twosensor/encoder units: a Hall-effect-sen-sor rotation encoder and a linear glass-scale encoder. This system implementsa proportional + integral + derivativecontrol algorithm that results in varia-tion of voltage commands to each ofthe three pairs of windings of thebrushless DC motor. In one of two alter-

native control modes, the voltages areapplied to the windings in a trapezoidalcommutation scheme on the basis oftiming signals obtained from the Hall-effect sensors; this scheme yields rela-tively coarse positioning — 24 steps permotor revolution. The second controlmode involves a sinusoidal commuta-tion scheme in which the output of thelinear glass-scale encoder is transposedto rotational increments to yield muchfiner position feedback — more than400,000 steps per revolution.

This work was done by Brant T. Cook,Donald M. Moore, David F. Braun, John S.Koenig, and Steve M. Hankins of Caltech forNASA’s Jet Propulsion Laboratory. For moreinformation, contact [email protected]

Installing a Test Tap on a Metal Battery CaseLyndon B. Johnson Space Center, Houston, Texas

A mechanical fitting and relativelysimple and safe method of installing iton the metal case of a battery havebeen devised to provide access to theinterior of the battery to perform in-spection and/or to measure such inter-nal conditions as temperature andpressure. A metal boss or stud havingan exterior thread is attached to thecase by capacitor-discharge stud weld-ing (CDSW), which takes only 3 to 6milliseconds and in which the metallur-

gical bond (weld) and the heat-af-fected zone are limited to a depth of afew thousandths of an inch (a few hun-dredths of a millimeter).

These characteristics of CDSW pre-vent distortion of the case and localizedinternal heating that could damage thechemical components inside of the bat-tery. An access hole is then drilledthrough the stud and case, into the in-terior of the battery. A mechanical fit-ting having a matching thread is in-

stalled on the stud and the interior endof the fitting is sealed with a pressure-sealing washer/gasket. The exteriorend of the fitting is configured for at-tachment of whatever instrumentationis required for the selected inspectionor measurement.

This work was done by Daniel R. Mayes ofJohnson Space Center and Daniel J. Rybickiof Lockheed Martin Corp. Further informa-tion is contained in a TSP (see page 1).MSC-23827-1


materials: (1) The lack of chemical reac-tivity between the PFPE and the glassmakes it impossible to form a lastingbond between them; and (2) such con-ventional membrane-fabrication tech-niques as spin coating yield membranesthat are not sufficiently flat and not suf-ficiently resistant to curling upon them-selves. The present process overcomesthese obstacles.

The process consists mainly of the fol-lowing steps:1. A fluorocarbon-based polymer is

formed on the glass plates by meansof a plasma deposition subprocess.

2. The polymer is patterned by use of aphotoresist and conventional photo-lithography.

3. The polymer is removed in the pat-tern by use of an O2/Ar plasma.

4. The remaining polymer surface areasare cleaned and modified by use of alow-energy O2 plasma.

5. The glass plates are spin-coated with a

lift-off material, which is then curedby heating to a temperature of 150 °Cfor 5 minutes.

6. The liquid (uncured) PFPE materialis pressed between the two lift-off-layer-coated glass plates, along with250-µm-thick shims to define the de-sired thickness of the PFPE mem-brane.

7. The liquid PFPE is cured to a solid byexposure to ultraviolet light for 5minutes.

8. The PFPE membrane is releasedfrom the glass plates by submersionin a developer solution and/or ace-tone.

9. The glass plates and the PFPE mem-brane are cleaned and activated forbonding by exposure to an O2 plasma.

10. The glass plates and the membraneare aligned and sandwiched to-gether at a temperature ≤100 °C anda pressure of 3 bar (0.3 MPa) for onehour. This combination of pressure

and temperature is sufficient tocause a chemical reaction that re-sults in bonding of the PFPE mem-brane to the polymer coats on theglass plates.

This work was done by Frank Greer, VictorE. White, Michael C. Lee, Peter A. Willis,and Frank J. Grunthaner of Caltech andJason Rolland and Jake Sprague of LiquidiaTechnologies Inc. for NASA’s Jet PropulsionLaboratory.





Materials

Polyimide-PEO copolymers (“PEO”signifies polyethylene oxide) that havebranched rod-coil molecular structuresand that can be cured into film form atroom temperature have been inventedfor use as gel electrolytes for lithium-ionelectric-power cells. These copolymersoffer an alternative to previouslypatented branched rod-coil polyimidesthat have been considered for use aspolymer electrolytes and that must becured at a temperature of 200 °C. Inorder to obtain sufficient conductivity

for lithium ions in practical applicationsat and below room temperature, it isnecessary to imbibe such a polymer witha suitable carbonate solvent or ionic liq-uid, but the high-temperature curemakes it impossible to incorporate andretain such a liquid within the polymermolecular framework. By eliminatingthe high-temperature cure, the presentinvention makes it possible to incorpo-rate the required liquid.

The curing of a polyimide-PEOcopolymer according to the invention

results in formation of a gel network thatis capable of conducting lithium ions.The PEO molecular segments providethe lithium-ion conductivity, while theimide segments and branching providedimensional stability. The network canhold as much as four times its ownweight of liquid while maintaining ahigh degree of dimensional stability.The liquid can aid significantly in theconduction of lithium ions, and in somecircumstances, can increase cell cyclelife. Electrolytes have been preparedthat contain no volatile components,have a potential stability window of >4.5V, and have exhibited stable Galvanic cy-cling between lithium metal electrodesin a coin cell for over 1,000 hours at 60°C and 0.25 mA/cm2 current density.

The copolymer is synthesized in thefollowing process (see figure): 1. A PEO oligomer that is terminated

with primary aliphatic amines on bothends is reacted with a dianhydride tomake a polyamide-acid prepolymer.The reaction takes place in a solvent,and the stoichiometry of the oligomerand the dianhydride is adjusted sothat the resulting prepolymer is a lin-ear polymer capped with amines onboth ends. The solvent must be care-fully chosen to solubilize the prepoly-mer, have a boiling temperaturepreferably between 150 and 200 °C,and to be inert to lithium metal andother cell ingredients.

2. The polyamide-acid prepolymer isimidized in solution. The water gener-ated in the imidization reaction is re-moved by azeotropic distillation.

3. The appropriate additives (e.g., alithium salt and a carbonate solvent)are dissolved in the polymer solution.A trifunctional molecule that reactswith the amine end caps at ambienttemperatures to form a gel is thenadded to the solution. The gelationtime and the properties of the result-ing film can be adjusted by changingthe length of the polymer chains. Theproperties of the film can also be ad-

Room-Temperature-Cured Copolymers for Lithium Battery GelElectrolytes Room-temperature curing offers an important advantage in room-temperature functionality.John H. Glenn Research Center, Cleveland, Ohio

A Polymer Gel Network based on a branched rod-coil polyimide-PEO copolymer is synthesized in a rel-atively simple process that can be adapted to tailoring the properties of the final product.

Imidization in Solution andAzeotropic Distillation

Rod-Coil Oligomer With Amine End Caps

X + 1 Coil Molecules

+

X Rod Molecules

H2N PEO NH2

O O

O O

YO O

RodCoil Rod

X – 1

Coil CoilH2N NH2

Lithium Salt + Liquid Additives

Polymer Gel Network

Trifunctional Molecules

N

N

C O

N

C

O

C O


justed through choice of the dianhy-dride, the length of the starting PEOoligomer molecules, and partial re-placement of the trifunctional mole-cule with a difunctional molecule.

4. The film can be packaged once gela-tion has occurred. Because the reac-tion solvent is inert toward all cell in-gredients, it is not necessary to

remove this solvent. Optionally, be-cause the reaction solvent boils at atemperature ≈100 C° lower than doesa typical cyclic carbonate solvent, thereaction solvent can be preferentiallyevaporated before packaging. This work was done by Mary Ann B.

Meador of Glenn Research Center and DeanM. Tigelaar of Ohio Aerospace Institute. Fur-

ther information is contained in a TSP (seepage 1).

Inquiries concerning rights for the com-mercial use of this invention should be ad-dressed to NASA Glenn Research Center, In-novative Partnerships Office, Attn: SteveFedor, Mail Stop 4–8, 21000 BrookparkRoad, Cleveland, Ohio 44135. Refer toLEW-18205-1.

Catalysts for Efficient Production of Carbon NanotubesSome alloys have been found to work at lower temperatures.Lyndon B. Johnson Space Center, Houston, Texas

Several metal alloys have shown prom-ise as improved catalysts for catalyticthermal decomposition of hydrocarbongases to produce carbon nanotubes(CNTs). Heretofore almost every experi-ment on the production of carbon nan-otubes by this method has involved theuse of iron, nickel, or cobalt as the cata-lyst. However, the catalytic-conversion ef-ficiencies of these metals have been ob-served to be limited. The identificationof better catalysts is part of a continuingprogram to develop means of mass pro-duction of high-quality carbon nan-otubes at costs lower than thoseachieved thus far (as much as $100/g forpurified multi-wall CNTs or $1,000/g forsingle-wall CNTs in year 2002).

The main effort thus far in this pro-gram has been the design and imple-mentation of a process tailored specifi-cally for high-throughput screening of

alloys for catalyzing the growth of CNTs.The process includes an integral combi-nation of (1) formulation of libraries ofcatalysts, (2) synthesis of CNTs from de-composition of ethylene on powders ofthe alloys in a pyrolytic chemical-vapor-decomposition reactor, and (3) scan-ning-electron-microscope screening ofthe CNTs thus synthesized to evaluatethe catalytic efficiencies of the alloys. In-formation gained in this process is putinto a database and analyzed to identifypromising alloy compositions, which areto be subjected to further evaluation in asubsequent round of testing.

The promising alloys identified thusfar have been the following (composi-tions in atomic percentages): 90 Co, 10Ti; 20 Co, 70 Ni, 5 Ti, 5 Ta; 90 Co, 10 Mo;20 Co, 75 Ni, 5 Mo; 80 Co, 10 Ti, 10 Al; 70Co, 15 Ni, 15 Ti; 80 Co, 10 Ni, 10 Ti; 70Co, 5 Ta, 5 Mo, 20 Mn; 80 Ni, 10 Mo, 10

A; 80 Co, 12 Ni, 8 Al; and 80 Co, 20 Cr. Some of these alloys have been found

to catalyze the formation of carbonnano tubes from ethylene at tempera-tures as low as 350 to 400 °C. In con-trast, the temperatures typically re-quired for prior catalysts range from550 to 750 °C.

This work was done by Ted X. Sun and YiDong of Intermatix Corp. for Johnson SpaceCenter. Further information is contained in aTSP (see page 1).


Intermatix Corp.351 Rheem Blvd.Moraga, CA 94556Refer to MSC-23477-1, volume and num-

ber of this NASA Tech Briefs issue, and thepage number.

Amorphous silk fibroin has shownpromise as a polymeric material deriv-able from natural sources for makingmembranes for use in removing CO2from mixed-gas streams. For most appli-cations of silk fibroin, for purposes otherthan gas separation, this material is usedin its highly crystalline, nearly naturalform because this form has uncom-monly high tensile strength. However,the crystalline phase of silk fibroin is im-permeable, making it necessary to con-vert the material to amorphous form toobtain the high permeability needed forgas separation.

Accordingly, one aspect of the pres-ent development is a process for gener-ating amorphous silk fibroin by treatingnative silk fibroin in an aqueousmethanol/salt solution. The resultingmaterial remains self-standing and canbe prepared as thin film suitable forpermeation testing. The permeabilityof this material by pure CO2 has beenfound to be highly improved, and itsmixed-gas permeability has been foundto exceed the mixed-gas permeabilitiesof several ultrahigh-CO2-permeablesynthetic polymers. Only one of the syn-thetic polymers — poly(trimethylsilyl-

propyne) [PTMSP] — may be morehighly permeable by CO2. PTMSP be-comes unstable with time, whereasamorphous silk should not, although atthe time of this reporting this has notbeen conclusively proven.

This work was done by Christopher M.Aberg, Anand K. Patel, Eun Seok Gil, andRichard J. Spontak of North Carolina StateUniversity and May-Britt Hagg of NorwegianUniversity of Science and Technology for John-son Space Center.

For further information, contact the JSCInnovation Partnerships Office at (281) 483-3809. MSC-24032-1

Amorphous Silk Fibroin Membranes for Separation of CO2Lyndon B. Johnson Space Center, Houston, Texas


Mechanics/Machinery

“Zero-Mass” Noninvasive Pressure Transducers High-performance strain gauges are formed in sputtered thin films. NASA’s Jet Propulsion Laboratory, Pasadena, California

Extremely lightweight, compact, non-invasive, rugged, relatively inexpensivestrain-gauge transducers have been de-veloped for use in measuring pressuresof fluids in tubes. These gauges wereoriginally intended for measuring pres-sures of spacecraft-propulsion fluids, butthey are also attractive for use in numer-ous terrestrial applications — especiallythose involving fluids that are extremelychemically reactive, fluids that must beisolated for hygienic purposes, fluidsthat must be allowed to flow without ob-struction, and fluid-containing tubes ex-posed to severe environments.

A basic pressure transducer of thistype comprises one or more pair(s) ofthin-film strain gauges integral with atube that contains the fluid of interest.Following established strain-gauge prac-tice, the gauges in each pair are con-nected into opposite arms of a Wheat-stone bridge (see figure). Typically, eachpressure transducer includes one pair(the active pair) of strain gauges formeasuring the hoop stress proportionalto the pressure of the fluid in the tubeand another pair (the dummy pair) ofstrain gauges that are nominally un-

strained: The dummy gauges aremounted on a substrate that is made ofthe same material as that of the tube.The substrate is welded to the tube atonly one spot so that stresses and strainsare not coupled from the tube into thesubstrate. The dummy strain gaugesmeasure neutral strains (basically, strainsassociated with thermal expansion), sothat the neutral-strain contribution canbe subtracted out of the final gaugereading.

The active strain gauges are orientedto obtain an adequate response to hoopstrain while minimizing the response totorsional and bending strains, which canbe induced by mechanical coupling be-tween the tube and other componentsof the fluid-handling system. The preciseoptimum orientation for this purposedepends on the Poisson’s ratios of thematerials used; a representative approxi-mate optimum orientation is specified asan angle of 61.3° between the gauge axisand the tube axis.

The strain gauges are formed by sput-ter deposition of a dielectric film di-rectly on the tube, followed by sputterdeposition of a film of a suitable piezore-

sistive material (typically, a nickel-chromium alloy), followed by laser cut-ting of strain-gauge grid lines into thepiezoresistive film. These gauges havebeen characterized by the term “zero-mass,” which is not entirely an exaggera-tion: the sputtered layers are so thin thatwhen one accounts for tube material re-moved by polishing in preparation forsputtering, one could find that the netmass of the tube plus pressure trans-ducer is equal to or less than that of theplain tube.

The connections between the straingauges and the external parts of theWheatstone bridge are made with thin-film flexible electrical leads. A thin filmof dielectric material can be sputteredover the strain gauges and bridge wiringto protect the gauge circuitry, to preventoutgassing, and/or to prevent chemicalreactions between the strain gauges andthe environment.

In addition to the advantages alreadymentioned, these pressure transducersoffer several advantages over prior pres-sure transducers, including those basedon strain gauges bonded to tubes by useof adhesives: • The intimate coupling between strain

gauges and tubes increases the magni-tudes and speeds of gauge responses,simplifies accounting for thermal coef-ficients, reduces thermal-responsetimes, and diminishes long-term driftsand zero shifts.

• A pressure transducer of this type is es-sentially part of the tube on which it ismounted, with little or no protuber-ances or additional mass, with high re-silience in the face of shock and vibra-tional loads.

• The high dissociation temperatures ofthe dielectric and piezoresistive filmsenables operation at high tempera-tures, while the thin-film, intimate-cou-pling nature of the gauge structuresextends the lower operating-tempera-ture limit down to the cryogenic range. This work was done by Frank T. Hartley of

Caltech for NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP (seepage 1). NPO-21194

The Active and Dummy Strain Gauges are used to measure hoop strain and neutral strain, respectively,in a tube that contains a pressurized fluid. The active gauges are oriented to minimize their responseto bending and torsional strains.

Spot Weld

Dummy Strain Gauge

Substrate for Dummy Strain Gauges (Larger- Diameter Segment of

Tube Material)

Excitation

Signal Out

1

2

4

3

+

+ –

–

61.3

Active Strain Gauge

Dummy Strain Gauge


In a recently invented class of piezo-electric diaphragm pumps, the elec-trode patterns on the piezoelectric di-aphragms are configured so that theelectric fields in the diaphragms havesymmetrical radial (along-the-surface)components in addition to through-the-thickness components. Previously, itwas accepted in the piezoelectric-trans-ducer art that in order to produce theout-of-plane bending displacement of adiaphragm needed for pumping, onemust make the electric field asymmetri-cal through the thickness, typically bymeans of electrodes placed on only oneside of the piezoelectric material. In thepresent invention, electrodes areplaced on both sides and patterned soas to produce substantial radial as wellas through-the-thickness components.Moreover, unlike in the prior art, theelectric field can be symmetricalthrough the thickness. Tests haveshown in a given diaphragm that anelectrode configuration according tothis invention produces more displace-ment than does a conventional one-sided electrode pattern.

The invention admits of numerousvariations characterized by various de-grees of complexity. Figure 1 is a simpli-fied depiction of a basic version. As inother piezoelectric diaphragm pumps ofsimilar basic design, the prime mover isa piezoelectric diaphragm. Applicationof a suitable voltage to the electrodes onthe diaphragm causes it to undergo out-of-plane bending. The bending displace-ment pushes a fluid out of, or pulls thefluid into, a chamber bounded partly bythe diaphragm. Also as in other di-aphragm pumps in general, check valvesensure that the fluid flows only inthrough one port and only out throughanother port.

Figure 2 shows the dia phragm inmore detail. In this case, the diaphragmis circular. The central region of the di-aphragm contains the piezoelectric ma-terial. There are two centrally located,intercirculating spiral electrodes on thetop side of the piezoelectric materialand two mirror-image replicas of themon the bottom side. The polarities ofthe voltages applied to the electrodesare chosen to produce a nearly symmet-rical, substantially radial electric field.The piezoelectric material and elec-

trodes are adhesively bonded togetherand sandwiched between adhesive-coated sheets of a flexible dielectric ma-terial, which extends radially outwardto form an outer annular region forsealing the diaphragm to the pumphousing.

This work was done by Robert G. Bryantand Dennis C. Working of Langley ResearchCenter and Karla Mossi, Nicholas D. Castro,and Poorna Mane of Virginia Common-wealth University. Further information is con-tained in a TSP (see page 1).LAR-16363-1

Radial-Electric-Field Piezoelectric Diaphragm PumpsDisplacements are increased in a departure from traditional electrode configurations.Langley Research Center, Hampton, Virginia

Figure 1. This Piezoelectric Diaphragm Pump is similar to other piezoelectric diaphragm pumps, exceptfor the advanced design of the diaphragm.

Central Region of Diaphragm Containing

Piezoelectric Material and Electrodes

Annular (Edge-Sealing) Outer Region of

Diaphragm

Housing

Outlet Check Valve Inlet Check Valve

+ – + – + – + – + – + –

+ – + – + – + – + – + –

A A

TOP AND BOTTOM VIEWS SHOWING MIRROR-IMAGE SPIRAL ELECTRODES

SECTION A-A

Annular (Edge-Sealing) Outer Region of

Diaphragm

Electric-FieldLines

PiezoelectricMaterial

Adhesive Bond

Electrodes

Figure 2. The Diaphragm features a central region containing a piezoelectric actuator and an annularouter region for sealing to the housing. The electrode pattern is chosen to ensure that the electricfield has a substantial radial component.


Ejector-Enhanced, Pulsed, Pressure-Gain CombustorSpecific fuel consumption of a gas turbine engine could be reduced by a few percent.John H. Glenn Research Center, Cleveland, Ohio

An experimental combination of anoff-the-shelf valved pulsejet combustorand an aerodynamically optimized ejec-tor has shown promise as a prototype ofimproved combustors for gas turbine en-gines. Despite their name, the “constantpressure” combustors heretofore used ingas turbine engines exhibit typical pres-sure losses ranging from 4 to 8 percentof the total pressures delivered by up-stream compressors. In contrast, thepresent ejector-enhanced pulsejet com-bustor exhibits a pressure rise of about3.5 percent at overall enthalpy and tem-perature ratios compatible with those ofmodern turbomachines. The modestpressure rise translates to a comparableincrease in overall engine efficiency and,consequently, a comparable decrease inspecific fuel consumption. The ejector-enhanced pulsejet combustor may alsooffer potential for reducing the emissionof harmful exhaust compounds by mak-ing it practical to employ a low-loss rich-burn/quench/lean-burn sequence.

Like all prior concepts for pressure-gain combustion, the present conceptinvolves an approximation of constant-volume combustion, which is inher-ently unsteady (in this case, morespecifically, cyclic). The consequent

unsteadiness in combustor exit flow isgenerally regarded as detrimental tothe performance of downstream turbo-machinery. Among other adverse ef-fects, this unsteadiness tends to detractfrom the thermodynamic benefits ofpressure gain. Therefore, it is desirablein any intermittent combustion processto minimize unsteadiness in the ex-haust path.

One of the easiest ways of approximat-ing constant-volume combustion is touse a process similar to that used in com-mercially available pulsejets. Unfortu-nately, the pulsed combustion effluent isfar too hot for any turbine and must,therefore, be mixed with some coolerbypass flow. The mixing process intro-duces losses that substantially degradeoverall engine performance.

The present concept of the ejector-en-hanced pulsejet combustor utilizes thedemonstrated ability of properly de-signed, unsteady-flow ejectors (augmen-tors) to efficiently entrain, mix, andsmooth secondary flows. The aerody-namically optimized ejector is placed aftof the pulsejet combustor. The pulsejetcombustor and the ejector are encasedin a shroud (see figure). The ejector isessentially an axisymmetric, tapered duct

with an aerodynamic inlet. The ejectorentrains the cool secondary flow to pro-duce an exhaust flow in a temperaturerange that can be tolerated by turboma-chinery. Acting in combination with theshroud, the ejector also reduces spatialand temporal nonuniformities in theflow, thereby mitigating the detrimentaleffects of the pulsed combustion process.Finally, the relatively high momentum-transfer efficiency of the ejector ensuresthat taken as a whole, the ejector-en-hanced pulsejet combustor yields a mod-est, time-averaged pressure rise.

In tests, the root-mean-square value ofexit pressure fluctuations was found to beabout 4.5 percent of the mean total pres-sure. This value is near the limit abovewhich pressure fluctuations are consid-ered to adversely affect the performanceof downstream turbomachinery.

This work was done by Daniel E. Paxson of GlennResearch Center and Kevin T. Dougherty of QSSGroup, Inc. Further information is contained ina TSP (see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressed toNASA Glenn Research Center, Innovative Part-nerships Office, Attn: Steve Fedor, Mail Stop4–8, 21000 Brookpark Road, Cleveland, Ohio44135. Refer to LEW-18096-1.

The Ejector-Enhanced Pulsejet Combustor mounted in a test rig is depicted here in cross section, approximately to scale.

Airflow

Perforated Liner

Total-Temperature Sensor Total-Temperature

Sensors

Total-Pressure Sensor

Total-Pressure Sensors

Differential-Pressure Sensor

Static-Pressure Sensor

Static-Pressure Sensor

Starting Air Line

Fuel-Tank Pressurization

Line Shroud

Struts Fuel Inlet

Combustor

Pulsejet

Ejector

Burst Disk

Gap


Physical Sciences

A modified scheme for electron-beam(E-beam) writing used in the fabricationof convex or concave diffraction grat-ings makes it possible to suppress theghost diffraction heretofore exhibitedby such gratings. Ghost diffraction is aspurious component of diffractioncaused by a spurious component of grat-ing periodicity as described below. Theghost diffraction orders appear betweenthe main diffraction orders and are typi-

cally more intense than is the diffusescattering from the grating. At such highintensity, ghost diffraction is the domi-nant source of degradation of gratingperformance.

The pattern of a convex or concavegrating is established by electron-beamwriting in a resist material coating a sub-strate that has the desired convex orconcave shape. Unfortunately, as a resultof the characteristics of electrostatic de-

flectors used to control the electronbeam, it is possible to expose only asmall field — typically between 0.5 and1.0 mm wide — at a given fixed positionof the electron gun relative to the sub-strate. To make a grating larger than thefield size, it is necessary to move the sub-strate to make it possible to write fieldscentered at different positions, so thatthe larger area is synthesized by “stitch-ing” the exposed fields.

Suppressing Ghost Diffraction in E-Beam-Written GratingsThe degree of periodicity of stitching errors is reduced.NASA’s Jet Propulsion Laboratory, Pasadena, California

The Second of Two Convex Gratings was divided into four annuli, within which the grating patterns were written with different field sizes. As a result, thesecond grating exhibited significantly less ghost diffraction.

Same Dual-Blaze Profile for Both Gratings

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.01E-4

1E-3

0.01

0.1

1

Ghost Orders(From First

Grating Only)

Main Order(From BothGratings)

FirstGrating

SecondGrating

No

rmal

ized

Eff

icie

ncy

Point Detector Position, Arbitrary Units

FIRST GRATING:MADE WITH ONE

FIELD SIZE

SECOND GRATING:MADE WITH FOUR

DIFFERENT FIELD SIZES

Equal-Height Slices Through Convex SubstrateCreate Annular E-Beam Pattern Sub-AreasE-Beam-Field Boundaries

Between Stage Moves


Even though the mechanical stageused to position the substrate can bevery accurate (positioning error of ≈ 20nm or less), field-stitching errors occur,causing underexposures or overexpo-sures that manifest themselves, after de-velopment of the resist, as increases ordecreases in grating thickness along thefield boundaries. Because all the fieldsare of the same size, the stitching errorsform another grating that has a periodequal to the field size. Hence, the lightscattered from the field boundaries addscoherently: this is ghost diffraction.

The modified scheme for electron-beam writing is based on the concept ofreducing the degree of periodicity of thestitching errors. In this scheme, the over-all grating area is divided into sub-areaswithin which the grating patterns arewritten in differently sized fields. For atypical convex or concave grating, the

sub-areas are most easily defined as an-nular areas that correspond to equal-height slices through the substrate (seefigure). Hence, the grating pattern ineach annulus is written with a differentfield size.

The ghost order intensities are pro-portional to the square of the scatteringamplitudes. Hence, if N different fieldsizes are used, the intensity of ghost dif-fraction can be expected to be reducedto approximately N –2 times the intensityobtained with a single field size.

To test this concept, two nominallyidentical gratings were fabricated. Thepattern of the first grating was writtenby stitching together fields of the samesize over its entire area, while the pat-tern of the second grating was estab-lished by use of four different field sizes.Whereas the ghost diffraction from thefirst grating was clearly noticeable, the

intensity of ghost diffraction from thesecond grating was so low as to be unde-tectable against the diffuse-scatteringbackground.

This work was done by Daniel Wilson andJohan Backlund of Caltech for NASA’s JetPropulsion Laboratory. Further informationis contained in a TSP (see page 1).


Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099(818) 354-2240E-mail: [email protected] to NPO-41302, volume and number


An analog electronic camera that ispart of a metrology system measures thevarying direction to a light-emittingdiode that serves as a bright point tar-get. In the original application forwhich the camera was developed, themetrological system is used to deter-mine the varying relative positions of ra-diating elements of an airborne syn-thetic-aperture-radar (SAR) antenna as

the airplane flexes during flight; preciseknowledge of the relative positions as afunction of time is needed for process-ing SAR readings.

It has been common metrology systempractice to measure the varying directionto a bright target by use of an electroniccamera of the charge-coupled-device oractive-pixel-sensor type. A major disad-vantage of this practice arises from the

necessity of reading out and digitizingthe outputs from a large number of pix-els and processing the resulting digitalvalues in a computer to determine thecentroid of a target: Because of the timetaken by the readout, digitization, andcomputation, the update rate is limitedto tens of hertz. In contrast, the analognature of the present camera makes itpossible to achieve an update rate of

Target-Tracking Camera for a Metrology SystemAngular measurements are updated at a rate of hundreds of hertz.NASA’s Jet Propulsion Laboratory, Pasadena, California

The Optics and Electronic Circuitry of the camera, shown here in simplified schematic form, provide two analog signals proportional to the horizontal andvertical angular displacements of the bright target.

X1

X =

X2

One-Dimensional PSD for Measuring Azimuth

Light From Bright Target

Beam Splitter

Cylindrical Lens

Cylindrical Lens

One-Dimensional PSD for Measuring Elevation

Δ

Σ

X1 – X2

X1 + X2


A method of polarimetric imaging,now undergoing development, involvesthe use of two photoelastic modulatorsin series, driven at equal amplitude butat different frequencies. The net effecton a beam of light is to cause (1) the di-rection of its polarization to rotate at theaverage of two excitation frequenciesand (2) the amplitude of its polarizationto be modulated at the beat frequency(the difference between the two excita-tion frequencies). The resulting modu-lated optical light beam is made to passthrough a polarizing filter and is de-tected at the beat frequency, which canbe chosen to equal the frame rate of anelectronic camera or the rate of sam-pling the outputs of photodetectors inan array.

The method was conceived to satisfy aneed to perform highly accurate polari-metric imaging, without cross-talk be-tween polarization channels, at framerates of the order of tens of hertz. The useof electro-optical modulators is necessi-tated by a need to obtain accuracy greaterthan that attainable by use of static polar-izing filters over separate fixed detectors.For imaging, photoelastic modulators arepreferable to such other electrio-opticalmodulators as Kerr cells and Pockels cellsin that photoelastic modulators operate atlower voltages, have greater angular ac-ceptances, and are easier to use. Prior tothe conception of the present method,polarimetric imaging at frame rates oftens of hertz using photoelastic modula-tors was not possible because the reso-nance frequencies of photoelastic modu-lators usually lie in the range from about20 to about 100 kHz.

It is conventional to characterize thepolarimetric state of incident light in

terms of the Stokes vector (I, Q, U, V),where I represents the total intensity; Qrepresents the excess of intensity of lightpolarized at an angle designated as 0°over that of light polarized at a relativeangle of 90°, U represents similarly theexcess of intensity at 45° over that 135°,and V represents the excess of intensityof right circular polarization over leftcircular polarization. It has been showntheoretically that in the present method,there should be no cross-talk betweenthe Q and U channels and that it shouldbe possible to obtain the ratio U/I fromtwo readings of a single photodetectortaken when the polarizer is in two orien-tations that differ by 45°.

The figure schematically depicts a lab-oratory setup that was used to demon-strate the feasibility of the method. Acollimated beam of white light was par-tially polarized by a glass plate at anoblique angle. The degree of polariza-tion could be changed by rotating theglass plate. The light then passedthrough a circular-polarization subsys-

tem that included (1) two photoelasticmodulators having their fast axes at anangle of 0°, sandwiched between (2) twoquarter-wave retarders oriented at an-gles of 45° and 135°, respectively. Thetwo photoelastic modulators had reso-nance frequencies of about 42 kHz, dif-fering by a beat frequency of about 9 Hz.The modulated light was then made topass through a 0° or 45° polarizer on theway to a photodetector. A band-pass fil-ter having a nominal pass wavelength of672 nm with 20-nm bandwidth wasmounted between the polarizer and thephotodetector. Results of several experi-ments at various degrees of linear polar-ization were found to agree substantiallywith theoretical predictions.

This work was done by Yu Wang,Thomas Cunningham, David Diner, EdgarDavis, Chao Sun, Bruce Hancock, GaryGutt, Jason Zan, and Nasrat Raouf of Cal-tech for NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP(see page 1).NPO-43806

Polarimetric Imaging Using Two Photoelastic ModulatorsThe frame rate is the difference between the resonance frequencies of the modulators.NASA’s Jet Propulsion Laboratory, Pasadena, California

White Light

GlassPlate Photoelastic

Modulators

Quarter-WaveRetarders Band-Pass

Filter

Polarizer at0º or 45º With Respect to FastOptical Axes of

Modulators

Photodetector

hundreds of hertz, and no computer isneeded to determine the centroid.

The camera is based on a position-sen-sitive detector (PSD), which is a rectan-gular photodiode with output contactsat opposite ends. PSDs are usually usedin triangulation for measuring small dis-tances. PSDs are manufactured in bothone- and two-dimensional versions.

Because it is very difficult to calibratetwo-dimensional PSDs accurately, thefocal-plane sensors used in this camera

are two orthogonally mounted one-di-mensional PSDs. The camera also in-cludes a beam splitter and two cylindri-cal lenses to focus line images of thetarget onto the PSDs — more specifi-cally, to form a horizontal line image onthe vertically oriented PSD and a verti-cal line image on the horizontally ori-ented PSD. The outputs from both endsof each PSD are processed by analog cir-cuitry (see figure) to obtain an analogsignal proportional to the displacement

of the image centroid from the mid-length position along the PSD. The di-rection-measuring error of the readouthas been found to be no more than1/2,700 of the angular width of the fieldof view.

This work was done by Carl Liebe, RandallBartman, Jacob Chapsky, AlexanderAbramovici, and David Brown of Caltech forNASA’s Jet Propulsion Laboratory. Further in-formation is contained in a TSP (see page 1).NPO-41466

Two Photoelastic Modulators driven at different frequencies were sandwiched between quarter-waveretarders, causing the polarization of the light to rotate at the average frequency at an amplitudethat oscillated at the difference frequency.

The figure depicts a proposed wide-angle lens that would be especially wellsuited for an electronic camera in whichthe focal plane is occupied by an imagesensor that has small pixels. The designof the lens is intended to satisfy require-ments for compactness, high imagequality, and reasonably low cost, whileaddressing issues peculiar to the opera-tion of small-pixel image sensors.Hence, this design is expected to enablethe development of a new generation ofcompact, high-performance electroniccameras. The lens example shown has a60° field of view and a relative aperture(f-number) of 3.2.

The main issues affecting the designare the following: • The response of a small-pixel image

sensor is sensitive to the angle of inci-dence of the light. At large angles ofincidence, the response includes ex-cessive crosstalk among pixels.

• When a lens of typical prior design im-ages a wide field, rays from the edge ofthe field are typically incident on theimage sensor at large angles. This ef-fect can be mitigated by use of a so-called image-space telecentric lens, forwhich the angle of incidence is con-stant. However, such a lens is typicallymuch larger than is a comparable non-telecentric lens.

• In the original intended application,in which the lens would be used tofocus light on a back-side-illumi-nated image sensor, there are re-quirements to minimize the size ofthe lens while making its optical be-havior nearly telecentric, to obtainnearly diffraction-limited imagequality while limiting distortion. Thefollowing are some key characteris-tics of the lens design:

• The lens would include an elementthat would function like an immersionlens. The image sensor would bemounted in direct contact with this el-ement. The incorporation of this ele-ment would enable maximization ofthe degree of telecentricity by bendingrays from the edge of the field propor-tionately more than those from themiddle, while otherwise exerting littleeffect on performance.

• A first doublet element, comprising twosubelements made of glasses character-ized by a large difference between theirindices of refraction, would be placedimmediately after an aperture stop.This doublet would control the fieldcurvature and the color correction.

• A second doublet element made fromtwo glasses that have similar, high in-dices of refraction but very differentdispersion values. This element would

control the chromatic correction andprovide most of the positive lens powernecessary for imaging.

• An “air lens” between a third doubletelement and a meniscus elementwould be used to balance the positivepower while affording some correctionfor aberrations.

• The aperture stop would be located atthe front of the lens.

• All of the lens elements and subele-ments are designed to have sphericalsurfaces and to be made of commonlyused glasses. Hence, the lens couldlikely be produced at lower cost thanwould be possible if aspherical shapesor unusual glasses were required. This work was done by Pantazis

Mouroulis and Edward Blazejewski of Cal-tech for NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP (seepage 1). NPO-44404

Aperture Stop

FirstDoubletElement

SecondDoubletElement

MeniscusElement

Immersion-Lenslike Element

ThirdDoublet Element


Miniature Wide-Angle Lens for Small-Pixel Electronic Camera The lens design addresses issues peculiar to small-pixel image sensors. NASA’s Jet Propulsion Laboratory, Pasadena, California

This Optical Layout shows the main features of the lens design. The total length of the lens from aper-ture stop to the image plane is less than two times its focal length.

vModal filters in the ≈10-µm spectralrange have been implemented as pla-nar dielectric waveguides in infrared in-terferometric applications such assearching for Earth-like planets. Whenlooking for a small, dim object

(“Earth”) in close proximity to a large,bright object (“Sun”), the interferomet-ric technique uses beams from two tele-scopes combined with a 180° phase shiftin order to cancel the light from abrighter object. The interferometer

baseline can be adjusted so that, at thesame time, the light from the dimmerobject arrives at the combiner in phase.This light can be detected and its in-frared (IR) optical spectra can be stud-ied. The cancellation of light from the

Modal Filters for Infrared InterferometryNASA's Jet Propulsion Laboratory, Pasadena, California


“Sun” to ≈106 is required; this is not pos-sible without special devices — modalfilters — that equalize the wavefronts ar-riving from the two telescopes.

Currently, modal filters in the ≈10-µmspectral range are implemented as sin-gle-mode fibers. Using semiconductortechnology, single-mode waveguides for

use as modal filters were fabricated. Twodesigns were implemented: one using anInGaAs waveguide layer matched to anInP substrate, and one using InAlAsmatched to an InP substrate. Photon De-sign software was used to design thewaveguides, with the main feature all de-signs being single-mode operation in the

10.5- to 17-µm spectral range. Prelimi-nary results show that the filter’s rejec-tion ratio is 26 dB

This work was done by Alexander Ksend-zov, Daniel R MacDonald, and AlexanderSoibel of Caltech for NASA’s Jet PropulsionLaboratory. For more information, [email protected]. NPO-44457

Mo3Sb7–xTex for Thermoelectric Power GenerationThese materials could be segmented with lower-temperature thermoelectric materials.NASA’s Jet Propulsion Laboratory, Pasadena, California

Compounds having compositions ofMo3Sb7–xTex (where x = 1.5 or 1.6) havebeen investigated as candidate thermo-electric materials. These compoundsare members of a class of semiconduc-tors that includes previously knownthermoelectric materials. All of thesecompounds have complex crystallineand electronic structures. Through se-lection of chemical compositions andprocessing conditions, it may be possi-ble to alter the structures to enhance oroptimize thermoelectric properties.

For the investigation, each specimen ofMo3Sb7–xTex was synthesized as follows:1. A mixture of specified proportions of

Mo, Sb, and Te powders was heatedfor 7 days at a temperature of 750 °Cin a covered boron nitride crucible inan evacuated, sealed fused silica am-pule.

2. The chemical reactions werequenched in cold water.

3. In an argon atmosphere, the cruciblewas opened and the reacted powdermixture was ground in an agate mor-tar.

4. The powder was annealed by againheating it as in step 1.

5, The powder was formed into a densecylindrical specimen by uniaxial press-ing in a high-density graphite die for 1hour at a pressure of about 20 kpsi(≈138 MPa) and temperature of 873°C in an argon atmosphere.The traditional thermoelectric figure

of merit, Z, is defined by the equation Z =α2/ρκ, where α is the Seebeck coeffi-cient, ρ is the electrical resistivity, and κ isthe thermal conductivity. Often, in cur-rent usage, the term “thermoelectric fig-ure of merit” signifies the dimensionlessproduct ZT, where T is the absolute tem-perature. The thermoelectric compatibil-ity factor, s, is defined by the equation s =[(1 + ZT)1/2 – 1]/αT. For maximum effi-ciency, s should not change with temper-

ature, both within a single material, andthroughout a segmented thermoelectric-generator leg, the segments of which aremade of different materials.

Values of ZT and s were calculatedfrom measurements of the pertinentphysical properties of Mo3Sb7–xTexspecimens as functions at various tem-

CeFe4Sb12

200

200 400 600 800 1,000 1,200 1,400

0

0.1

7

6

5

4

3

2

1

0

0.2

0.3

0.4

0.5ZT

0.6

0.7

0.8

0.9

1.0

300 400 500 600

Temperature, K

Temperature, K

Co

mp

atib

ility

Fac

tor,

V–1

700 800 900

.x = 1.6

x = 1.5

1,000 1,100

CC SbbCC bC b

Bi2Te3

Zn2Sb3 Te/Ag/Ge/Sb Alloy

Mo3Sb5.4Te1.6

SiGe

Ce0.9Fe3.5Co0.5Sb12

PbTe

Thermoelectric Figures of Merit and compatibility factors of Mo3Sb7–xTex and other compounds weredetermined as functions of temperature.


Two-Dimensional Quantum Model of a NanotransistorQuantum effects that become important at the nanoscale are taken into account.Ames Research Center, Moffett Field, California

A mathematical model, and softwareto implement the model, have been de-vised to enable numerical simulation ofthe transport of electric charge in, andthe resulting electrical performancecharacteristics of, a nanotransistor [inparticular, a metal oxide/semiconductorfield-effect transistor (MOSFET) havinga channel length of the order of tens ofnanometers] in which the overall devicegeometry, including the doping profilesand the injection of charge from thesource, gate, and drain contacts, are ap-proximated as being two-dimensional.The model and software constitute acom p utational framework for quantita-tively exploring such device-physics is-sues as those of source-drain and gateleakage currents, drain-induced barrierlowering, and threshold voltage shiftdue to quantization. The model andsoftware can also be used as means ofstudying the accuracy of quantum cor-rections to other semiclassical models.

The present model accounts for twoquantum effects that become increas-ingly important as channel length de-creases toward the nanometer range:quantization of the inversion layer andballistic transport of electrons across thechannel. Heretofore, some quantum ef-fects in nanotransistors have been ana-lyzed qualitatively by use of simple one-dimensional ballistic models, buttwo-dimensional models are necessaryfor obtaining quantitative results. Cen-tral to any quantum-mechanical ap-proach to modeling of charge transportis the self-consistent solution of a waveequation to describe the quantum-me-chanical aspect of the transport, Pois-son’s equation, and equations for statis-tics of the particle ensemble.

Non-equilibrium Green’s function(NEGF) formalisms have been success-ful in modeling steady-state transport ina variety of one-dimensional semicon-ductor structures. The present modelfor the two-dimensional case includesthe NEGF equations, which are solvedself-consistently with Poisson’s equa-tion. At the time of this work, this wasthe most accurate full quantum modelyet applied to simulation of two-dimen-sional semiconductor devices. Openboundary conditions (in which the nar-

row channel region opens into broadsource, gate, and drain regions) andtunneling through oxide are treated onan equal footing. Interactions betweenelectrons and phonons are taken intoaccount, causing the modeled transportto deviate from ballistic in a realisticmanner. Electrons in the wave-vector-space ellipsoids of the conduction bandare treated within the anisotropic-effec-tive-mass approximation.

Self-consistent solution of the Pois-son and NEGF equations is computa-

Drain Current Versus Gate Potential in a conceptual 25-nm-gate-length MOSFET commonly used as anexample for testing MOSFET-simulating software was calculated by means of a three-band and a one-band version of the present quantum-based model. For comparison, the plot also shows results froma classical diffusion-drift model and a quantum-corrected model embodied in a commercial MOSFET-simulation computer program. A drain bias of 1 V and a gate oxide thickness of 1.5 nm were used inthe simulations.

0 0.2 0.4 0.6 0.8 1.010–10

10–9

10–8

10–7

10–6

10–5

10–4

10–3

10–2

Dra

in C

urr

ent,

Am

per

es/M

icro

n

Gate Potential, Volts

Vd = 1V

Classical-Diffusion-Drift

Quantum(One-Band)

Quantum-Corrected

Quantum(Three-Band)

peratures, and the s values ofMo3Sb5.4Te1.6 were compared withthose of other state-of-the-art thermo-electric materials (see figure). The ZTvalues of both Mo3Sb5.5Te1.5 andMo3Sb5.4Te1.6 were found to increasewith temperature up to 1,050 K, whichis just below the decomposition temper-ature. The fact that ZT for x = 1.6 ex-ceeds that for x = 1.5 might be taken asa hint that one could increase ZT by in-creasing x, except for the observation

that attempts to synthesize Mo3Sb7–xTexhaving x > 1.6 resulted in specimensthat appeared to be multiphase. Hence,other approaches to doping may bemore promising.

The s value of Mo3Sb5.4Te1.6 wasfound to increase from about 1 V–1 at300 K (room temperature) to about 2V–1 at 1,000 K. This doubling of s indi-cates poor self-compatibility ofMo3Sb5.4Te1.6 over the affected tempera-ture range. However, Mo3Sb5.4Te1.6

could be suitable for segmentation: forexample, a Mo3Sb5.4Te1.6 segment(which can withstand a temperature>800 K) could be joined to a lower-tem-perature PbTe segment at an interfacetemperature at or slightly below 800 K,where their s values are equal.

This work was done by G. Jeffrey Snyder, FrankS. Gascoin, and Julia Rasmussen of Caltech forNASA’s Jet Propulsion Laboratory. For more in-formation, contact [email protected]


tion-intensive because of the number ofspatial and energy coordinates in-volved. Therefore, parallel distributedcomputing is imperative: the softwarethat implements the model distributesthe computations, energy-wise, to the

various processors. Initial simulationswere performed using, variously, be-tween 16 and 64 processors of an SGIOrigin multiprocessor computer. Thefigure presents an example of results ofone set of simulations.

This work was done by T. R. Govindanand B. Biegel of Ames Research Center andA. Svizhenko and M. P. Anantram of Com-puter Science Corp. Further information iscontained in a TSP (see page 1). ARC-15471-1

The figure schematically depicts somealternative designs of proposed com-pact, lightweight optoelectronic micro-scopes that would contain no lenses andwould generate magnified video imagesof specimens. Microscopes of this type

were described previously in “MiniatureMicroscope Without Lenses” (NPO -20218), NASA Tech Briefs, Vol. 22, No. 8(August 1998), page 43 and “ReflectiveVariants of Miniature Microscope With-out Lenses” (NPO 20610), NASA Tech

Briefs, Vol. 26, No. 9 (September 1999),page 6a. To recapitulate: In the designand construction of a microscope of thistype, the focusing optics of a conven-tional microscope are replaced by acombination of a microchannel filterand a charge-coupled-device (CCD)image detector. Elimination of focusingoptics reduces the size and weight of theinstrument and eliminates the need forthe time-consuming focusing operation.

The microscopes described in thecited prior articles contained two-dimen-sional CCDs registered with two-dimen-sional arrays of microchannels and, assuch, were designed to produce full two-dimensional images, without need forscanning. The microscopes of the pres-ent proposal would contain one-dimen-sional (line image) CCDs registered withlinear arrays of microchannels. In theoperation of such a microscope, onewould scan a specimen along a line per-pendicular to the array axis (in otherwords, one would scan in “pushbroom”fashion). One could then synthesize afull two-dimensional image of the speci-men from the line-image data acquiredat one-pixel increments of positionalong the scan.

In one of the proposed microscopes,a beam of unpolarized light for illumi-nating the specimen would enter fromthe side. This light would be reflecteddown onto the specimen by a nonpolar-izing beam splitter attached to the mi-crochannels at their lower ends. A por-tion of the light incident on thespecimen would be reflected upward,through the beam splitter and alongthe microchannels, to form an imageon the CCD.

If the nonpolarizing beam splitter werereplaced by a polarizing one, then thespecimen would be illuminated by s-polar-ized light. Upon reflection from the spec-imen, some of the s-polarized light wouldbecome p-polarized. Only the p-polarizedlight would contribute to the image on

Scanning Miniature Microscopes Without LensesPolarization-sensitive and multicolor versions should also be possible. NASA’s Jet Propulsion Laboratory, Pasadena, California

p-Polarized Light CCDs (Pixels Arranged Along a Line Perpendicular to This Page)

Microchannels Aligned With CCD Pixels

s-Polarized Light

Polarizing Beam Splitter

Nonpolarizing Beam Splitter

NONPOLARIZING MICROSCOPE POLARIZING MICROSCOPE

Specimen

B G R B G R

Blue LED Green LED Red LED Blue LED Green LED Red LED

Translation Stage

Specimen

MICROSCOPE THAT CAN TAKE THREE-COLOR PICTURES IN UNPOLARIZED AND POLARIZED LIGHT SIMULTANEOUSLY

Scanning Lensless Microscopes of various degrees of complexity and capability would be made fromline-imaging CCDs, the pixels of which would be aligned with microchannels.


the CCD; in other words, the image wouldcontain information on the polarization-rotating characteristic of the specimen.

The scanning microscopes describedabove could be used as building blocks fora multicolor, multipolarization micro-scope. In the example shown in the fig-ure, six scanning microscopes would be

assembled on a single translation stage.The microchannels would be inter-spersed with light sources that would com-prise light-emitting diodes (LEDs) cou-pled to the beam splitters via prismlikelight guides.

This work was done by Yu Wang of Cal-tech for NASA’s Jet Propulsion Laboratory.

Further information is contained in a TSP(see page 1).

This invention is owned by NASA, and apatent application has been filed. Inquiries con-cerning nonexclusive or exclusive license for itscommercial development should be addressed tothe Patent Counsel, NASA Management Of-fice–JPL. Refer to NPO-20821.

Manipulating Neutral Atoms in Chip-Based Magnetic Traps Magnetic-field gradients are used to accelerate and decelerate atoms. NASA’s Jet Propulsion Laboratory, Pasadena, California

Several techniques for manipulatingneutral atoms (more precisely, ultracoldclouds of neutral atoms) in chip-basedmagnetic traps and atomic waveguideshave been demonstrated. Such traps andwaveguides are promising componentsof future quantum sensors that wouldoffer sensitivities much greater thanthose of conventional sensors. Potentialapplications include gyroscopy and basicresearch in physical phenomena that in-volve gravitational and/or electromag-netic fields. The developed techniquesmake it possible to control atoms withgreater versatility and dexterity thanwere previously possible and, hence, canbe expected to contribute to the value ofchip-based magnetic traps and atomicwaveguides.

The basic principle of these tech-niques is to control gradient magneticfields with suitable timing so as to altera trap to exert position-, velocity-,and/or time-dependent forces onatoms in the trap to obtain desired ef-fects (see figure). The trap magneticfields are generated by controlled elec-tric currents flowing in both macro-scopic off-chip electromagnet coilsand microscopic wires on the surfaceof the chip.

The methods are best explained interms of examples. Rather than simplyallowing atoms to expand freely into anatomic waveguide, one can give them acontrollable push by switching on anexternally generated or a chip-basedgradient magnetic field. This push canincrease the speed of the atoms, typi-cally from about 5 to about 20 cm/s.Applying a non-linear magnetic-fieldgradient exerts different forces onatoms in different positions — a phe-nomenon that one can exploit by intro-ducing a delay between releasing atomsinto the waveguide and turning on themagnetic field.

Before the magnetic field is turnedon, the fastest atoms move away from theregion where the gradient will be thestrongest, while the slower atoms lag be-hind, remaining in that region for awhile. Hence, once the magnetic field isturned on, it can be expected to pushthe slower atoms harder than it will pushthe faster atoms. By controlling the am-plitude and delay of the gradient, onecan tailor the push so as to cause theslower atoms to catch up with the fasterones at a chosen location along thewaveguide, thereby effectively focusingthe atoms (in other words, greatly in-creasing the density of the cloud ofatoms) at that location. Of course, in ad-

dition, the acceleration of the sloweratoms effectively raises the temperatureof the cloud of atoms. In a proposedvariant of this accelerating-and-focusingtechnique, the gradient would be suit-ably repositioned along the waveguideand its amplitude and timing suitably al-tered, so as to preferentially deceleratethe faster atoms, thereby effectively cool-ing the cloud of atoms.

This work was done by David Aveline,Robert Thompson, Nathan Lundblad, LuteMaleki, Nan Yu, and James Kohel of Caltechfor NASA’s Jet Propulsion Laboratory. Fur-ther information is contained in a TSP (seepage 1).NPO-43015

A Cold Cloud of 87Rb Atoms is manipulated in an atomic waveguide by use of controlled gradientmagnetic fields. This sequence of images shows the cloud moving rightward toward a potential bar-rier, then splitting into a part that passes through the barrier and a part reflected leftward from thebarrier. The numbers on the axes are coordinates in units of 8-µm pixels.

100

150

150

150

50

100

100

200

200

200100 150 200 250 300 350 400 450 500


In a proposed alternative to previousapproaches to making hot-shoe contactsto the legs of thermoelectric devices, onerelies on differential thermal expansionto increase contact pressures for the pur-pose of reducing the electrical resistancesof contacts as temperatures increase. Theproposed approach is particularly applica-ble to thermoelectric devices containingp-type (positive-charge-carrier) legs madeof a Zintl compound (specifically,Yb14MnSb11) and n-type (negative-charge-carrier) legs made of SiGe.

This combination of thermoelectricmaterials has been selected for furtherdevelopment, primarily on the basis ofprojected thermoelectric performance.However, it is problematic to integrate,into a practical thermoelectric device,legs made of these materials along with ametal or semiconductor hot shoe that isrequired to be in thermal and electricalcontact with the legs. This is partly be-cause of the thermal-expansion mis-match of these materials: The coeffi-cient of thermal expansion (CTE) ofSiGe is 4.5 × 10–6 °C–1, while the CTE ofYb14MnSb11 is 20 × 10–6 °C–1. Simplyjoining a Yb14MnSb11 and a SiGe leg to acommon hot shoe could be expected toresult in significant thermal stresses ineither or both legs during operation.

Heretofore, such thermal stresses havebeen regarded as disadvantageous. Inthe proposed approach, stresses result-ing from the CTE mismatch would beturned to advantage.

The figure depicts a thermoelectricunicouple according to the proposedapproach. By use of established tech-niques, the n-type SiGe leg would bebonded to the hot shoe, which wouldbe made of Si, Mo, or graphite. How-ever, the Yb14MnSb11 leg would not bebonded to the hot shoe: instead, theYb14MnSb11 leg would be inserted in aprecisely fit hole in the hot shoe. Theprecision of the fit would be such thatupon assembly at room temperature,the contact pressure between the hotshoe and the Yb14MnSb11 leg would below. During heating up to a hot-shoeoperating temperature of 1,000 °C, thethermal expansion of the Yb14MnSb11leg would exceed that of the hole by anamount that would increase the con-tact pressure to >100 MPa. This pres-sure would suffice to keep the thermaland electrical contact resistances ac-ceptably low.

Optionally, if the hot shoe were madeof Si or Mo, the contact resistancescould be made even lower by adding athin, reactive layer of a metal at the in-

terface between the Yb14MnSb11 leg andthe hot shoe. Another option would beto taper the hole and the mating portionof the Yb14MnSb11 leg and to press-fitthe leg and the hot shoe together atroom temperature, thereby providingfor maintenance of at least some pres-sure and prevention of separation dur-ing thermal cycling.

This work was done by Jeffrey Sakamoto ofCaltech for NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP (seepage 1). NPO-44896

Expansion Compression Contacts for Thermoelectric LegsOrdinarily regarded as disadvantageous, thermal-expansion mismatch would be turned to advan-tage.NASA’s Jet Propulsion Laboratory, Pasadena, California

A A

TOP VIEW

Hot Shoe

Bond

n-TypeSiGe Leg

Y14MnSb11Leg (p-Type)

SECTION A-A

A Thermoelectric Unicouple as proposed wouldutilize thermal-expansion mismatch to obtainhigh contact pressure between the hot shoe andthe Yb14MnSb11 leg.


Information Sciences

A recently invented speech-recogni-tion method applies to words that are ar-ticulated by means of the tongue andthroat muscles but are otherwise notvoiced or, at most, are spoken sotto voce.This method could satisfy a need forspeech recognition under circumstancesin which normal audible speech is diffi-cult, poses a hazard, is disturbing to lis-teners, or compromises privacy. Themethod could also be used to augmenttraditional speech recognition by pro-viding an additional source of informa-tion about articulator activity. Themethod can be characterized as interme-diate between (1) conventional speechrecognition through processing of voicesounds and (2) a method, not yet devel-oped, of processing electroencephalo-graphic signals to extract unspokenwords directly from thoughts.

This method involves computationalprocessing of digitized electromyographic(EMG) signals from muscle innervationacquired by surface electrodes under asubject’s chin near the tongue and on theside of the subject’s throat near the larynx(see figure). After preprocessing, digitiza-tion, and feature extraction, EMG signalsare processed by a neural-network patternclassifier, implemented in software, thatperforms the bulk of the recognition taskas described below.

Before processing signals representingwords that one seeks to recognize, theneural network must be trained. Duringtraining, EMG signals representing knownwords and/or phrases are first sampledover specified time intervals (each typi-cally about 2 seconds long). The portionsof the signals recorded during each timeinterval are denoted sub-audible musclepatterns (SAMPs). Sequences of samplesof SAMPs for overlapping time intervalsare processed by a suitable signal-process-ing transform (SPT), which could be, forexample, a Fourier, Hartley, or wavelettransform. The SPT outputs are enteredinto a matrix of coefficients, which is thendecomposed into contiguous, non-over-lapping two-dimensional cells of entries,each cell corresponding to a feature. Neu-ral-network analysis is performed to esti-

Processing Electromyographic Signals To Recognize WordsThe speaker need not make any sound.Ames Research Center, Moffett Field, California

Sig

nal

Am

plit

ud

e, A

rbit

rary

Un

its

Time, Arbitrary Units

4

3

2

1

0

0

–1

–2

–3

–4

“Stop”4

3

2

1

0

0

–1

–2

–3

–4

“Go”43210

–1–2–3–4–5–6

0

“Left”

3

2

1

0

0

–1

–2

–3

–4

–5

“Alpha”1.5

1.0

0.5

0

–0.5

–1.0

–1.5

–2.00

“Omega”

0

“Right”2.5

2.0

1.5

1.0

0.5

0

–0.5

–1.0

–1.5

–2.0

SurfaceElectrodes

Surface Electrodes Under and Near the Chin acquire signals indicative of tongue and throat muscle ac-tivity during articulation of words, even when the words are unvoiced. The signals are then processedto recognize the words. The six signal samples shown here correspond to the noted words.


mate reference sets of weight coefficientsfor weighted sums of the SAMP featuresthat correspond to known words and/or phrases.

Once training has been done, a SAMPthat includes an unknown word is sam-pled and processed by the SPT, the SPToutputs are used to construct a matrix,the matrix is decomposed into cells, and

neural-network analysis is performed, allin the same manner as that of training.The weight coefficients computed dur-ing training are used to determinewhether there is a sufficiently closematch between an unknown word in theSAMP and a known word in the trainingdatabase. If such a match is found, theword is deemed to be recognized.

This work was done by C. C. Jorgensen andD. D. Lee of Ames Research Center. Further in-formation is contained in a TSP (see page 1).

This invention is owned by NASA and apatent application has been filed. Inquiriesconcerning rights for the commercial use ofthis invention should be addressed to theAmes Technology Partnerships Division at(650) 604-2954. Refer to ARC-15040-1.

A physical principle (more precisely, aprinciple that incorporates mathemati-cal models used in physics) has beenconceived as the basis of a method ofgenerating randomness in Monte Carlosimulations. The principle eliminatesthe need for conventional random-num-ber generators.

The Monte Carlo simulation methodis among the most powerful computa-tional methods for solving high-dimen-sional problems in physics, chemistry,economics, and information process-ing. The Monte Carlo simulationmethod is especially effective for solv-

ing problems in which computationalcomplexity increases exponentiallywith dimensionality. The main advan-tage of the Monte Carlo simulationmethod over other methods is that thedemand on computational resourcesbecomes independent of dimensional-ity. As augmented by the present prin-ciple, the Monte Carlo simulationmethod becomes an even more power-ful computational method that is espe-cially useful for solving problems asso-ciated with dynamics of fluids,planning, scheduling, and combinator-ial optimization.

The present principle is based oncoupling of dynamical equations withthe corresponding Liouville equation.The randomness is generated by non-Lipschitz instability of dynamics trig-gered and controlled by feedback fromthe Liouville equation. (In non-Lip-schitz dynamics, the derivatives of solu-tions of the dynamical equations are notrequired to be bounded.)

This work was done by Michail Zak of Cal-tech for NASA’s Jet Propulsion Laboratory.For further information, [email protected]

Physical Principle for Generation of RandomnessNASA’s Jet Propulsion Laboratory, Pasadena, California

DSN Beowulf Cluster-Based VLBI Correlator Software architecture is scalable to meet faster processing needs for future data processing. NASA’s Jet Propulsion Laboratory, Pasadena, California

The NASA Deep Space Network(DSN) requires a broadband VLBI (verylong baseline interferometry) correlatorto process data routinely taken as part ofthe VLBI source Catalogue Maintenanceand Enhancement task (CAT M&E) andthe Time and Earth Motion PrecisionObservations task (TEMPO). The dataprovided by these measurements are acrucial ingredient in the formation ofprecision deep-space navigation models.In addition, a VLBI correlator is neededto provide support for other VLBI re-lated activities for both internal and ex-ternal customers.

The JPL VLBI Correlator (JVC) wasdesigned, developed, and delivered tothe DSN as a successor to the legacyBlock II Correlator. The JVC is a full-ca-pability VLBI correlator that uses soft-ware processes running on multiplecomputers to cross-correlate two-an-tenna broadband noise data. Compo-nents of this new system (see Figure 1)

consist of Linux PCs integrated into aBeowulf Cluster, an existing Mark5 datastorage system, a RAID array, an exist-ing software correlator package (SoftC)originally developed for Delta DOR

Navigation processing, and various cus-tom-developed software processes andscripts.

Parallel processing on the JVC isachieved by assigning slave nodes of the

User Workstation

Post Process

MasterNode Station LAN

DSCC Station #2

DSCC Station #1

MARK5 #1

RAID #1

MARK5 #2

RAID #2

Node 1Node 2Node 3Node 4Node 5Node 6Node 7Node 8Node ...Node n

JVCDSN

Node 1Node 2Node 3Node 4Node 5Node 6Node 7Node 8Node ...Node n

MARK5 #2

MARK5 #1

RAID #1

RAID #2

User Workstation

Post Process

MasterNode

Figure 1. Components of the New Correlator are shown in this simplified block diagram.


A computational method and systembased on a hybrid of an artificial neuralnetwork (NN) and a support vector ma-chine (SVM) (see figure) has been con-ceived as a means of maximizing or min-imizing an objective function,optionally subject to one or more con-straints. Such maximization or mini-mization could be performed, for exam-ple, to optimize solve a data-regressionor data-classification problem or to opti-mize a design associated with a responsefunction. A response function can beconsidered as a subset of a response sur-face, which is a surface in a vector spaceof design and performance parameters.A typical example of a design problemthat the method and system can be usedto solve is that of an airfoil, for which aresponse function could be the spatialdistribution of pressure over the airfoil.In this example, the response surfacewould describe the pressure distribu-

tion as a function of the operating con-ditions and the geometric parameters ofthe airfoil.

The use of NNs to analyze physicalobjects in order to optimize their re-sponses under specified physical condi-tions is well known. NN analysis is suit-able for multidimensional in ter polationof data that lack structure and enablesthe representation and optimization ofa succession of numerical solutions ofincreasing complexity or increasing fi-delity to the real world. NN analysis isespecially useful in helping to satisfymultiple design objectives. FeedforwardNNs can be used to make estimatesbased on nonlinear mathematical mod-els. One difficulty associated with use ofa feedforward NN arises from the needfor nonlinear optimization to deter-mine connection weights among input,intermediate, and output variables. Itcan be very expensive to train an NN in

cases in which it is necessary to modellarge amounts of information.

Less widely known (in comparisonwith NNs) are support vector machines(SVMs), which were originally applied instatistical learning theory. In terms thatare necessarily oversimplified to fit thescope of this article, an SVM can be char-acterized as an algorithm that (1) effectsa nonlinear mapping of input vectorsinto a higher-dimensional feature spaceand (2) involves a dual formulation ofgoverning equations and constraints.One advantageous feature of the SVMapproach is that an objective function(which one seeks to minimize to obtaincoefficients that define an SVM mathe-matical model) is convex, so that unlikein the cases of many NN models, anylocal minimum of an SVM model is alsoa global minimum.

In the SVM approach as practicedheretofore, underlying feature-space co-

Hybrid NN/SVM Computational System for Optimizing DesignsThe NN and the SVM help each other “learn” in an iterative process.Ames Research Center, Moffett Field, California

Beowulf cluster to process separatescans in parallel until all scans havebeen processed. Due to the single-stream sequential playback of theMark5 data, some ramp-up time is re-quired before all nodes can have accessto required scan data. Core functions ofeach processing step are accomplishedusing optimized C programs. The coor-dination and execution of these pro-grams across the cluster is accom-plished using Pearl scripts, PostgreSQL

commands, and a handful of miscella-neous system utilities.

Mark5 data modules are loaded onMark5 Data systems playback units, oneper station. Data processing is startedwhen the operator scans the Mark5 sys-tems and runs a script that reads vari-ous configuration files and then cre-ates an experiment-dependent statusdatabase used to delegate parallel tasksbetween nodes and storage areas (seeFigure 2). This script forks into three

processes: extract, translate, and corre-late. Each of these processes iterates onavailable scan data and updates the sta-tus database as the work for each scanis completed.

The extract process coordinates andmonitors the transfer of data from eachof the Mark5s to the Beowulf RAID stor-age systems. The translate process mon-itors and executes the data conversionprocesses on available scan files, andwrites the translated files to the slavenodes. The correlate process monitorsthe execution of SoftC correlationprocesses on the slave nodes for scansthat have completed translation.

A comparison of the JVC and thelegacy Block II correlator outputs revealsthey are well within a formal error, andthat the data are comparable with re-spect to their use in flight navigation.The processing speed of the JVC is im-proved over the Block II correlator by afactor of 4, largely due to the elimina-tion of the reel-to-reel tape drives usedin the Block II correlator.

This work was done by Stephen P. Rogstad,Andre P. Jongeling, Susan G. Finley, Leslie A.White, Gabor E. Lanyi, John E. Clark, andCharles E. Goodhart of Caltech for NASA’s JetPropulsion Laboratory. For more information,contact [email protected]. NPO-46279

Master Node

Operator enters> corproc expnm -on

extract

translate

translate

cor records

logs

corproc

Mark5

Storage

Mark5

Storage

Extraction, Translation, and Correlation are parallel tasks

Mark5 Mark5

StorageStorage

Figure 2. Extraction, Translation, and Correlation are run as parallel tasks.


ordinates or functions must be specified.In the NN approach as practiced hereto-fore, resampling of data is needed to im-plement a process, known in the art asmodel hybridization, in which a superiorneural network is generated from thesynaptic-connection weight vectors ofmultiple neural networks that yield localminima with acceptably low errors. Whatis needed is a machine-learning algorithmthat combines the desirable features ofthe NN and SVM approaches and doesnot require intimate a priori familiaritywith operational details of the object to beoptimized. Preferably, the algorithmshould automatically provide a character-

ization of many or all of the aspects in fea-ture space needed for the analysis.

A hybrid NN/SVM system (see fig-ure) accepts inputs in the form of pa-rameter values, which are regarded asindependent coordinates in an inputvector space. In the construction of theSVM, the input coordinates aremapped into a feature space of appro-priately greater dimensionality, whereinthe coordinates include computedcombinations (e.g., powers and/orpolynomials) of the input space coordi-nates. The NN is initially programmedwith random synaptic-connectionweights and used to construct inner

products for the SVM. The inner prod-ucts are, in turn, used to compute La-grange multipliers. A training error as-sociated with the connection weightsand Lagrange multipliers is calculated.If the training error is too large, one ormore connection weights are changedand all of the foregoing (except the ini-tial programming with randomweights) steps are repeated. If the train-ing error is not too large, the connec-tion weights and the Lagrange multipli-ers are accepted as optimal.

An important advantage of this sys-tem over a conventional SVM is thatthe feature-space coordinates thatmust be specified a priori are deter-mined by the NN subsystem. More-over, the feature-space coordinates aregenerated by the NN subsystem to cor-respond to the data at hand; in otherwords, the feature space provided bythe NN subsystem evolves to match orcorrespond to the data. A featurespace that evolves in this manner is re-ferred to as “data-adaptive.” The fea-ture-space coordinates generated bythe NN subsystem can be easily aug-mented with additional feature-spacecoordinates (combinations of parame-ters) and kernel functions provided bythe user.

This work was done by Man MohanRai of Ames Research Center. Further infor-mation is contained in a TSP (see page 1).

This invention has been patented by NASA(U.S. Patent No. 6,961,719). Inquiries con-cerning rights for the commercial use of thisinvention should be addressed to the AmesTechnology Partnerships Division at (650)604-2954. Refer to ARC-14586.

A Hybrid NN/SVM System offers capabilities greater than those of an NN or a conventional SVM alone.

InputParameters

Connecting Weights

Hidden Layer ofNeural Network

Feature-SpaceCoordinates

SupportVector

Machine

Feature-Space Coordinatoror Kernel Functions Specified

by the User

Input Nodes of Neural Network

Bias


Books & Reports

Criteria for Modeling in LESof Multicomponent Fuel Flow

A report presents a study addressingthe question of which large-eddy simula-tion (LES) equations are appropriatefor modeling the flow of evaporatingdrops of a multicomponent liquid in agas (e.g., a spray of kerosene or dieselfuel in air). The LES equations are ob-tained from the direct numerical simula-tion (DNS) equations in which the solu-tion is computed at all flow length scales,by applying a spatial low-pass filter. Thus,in LES the small scales are removed andreplaced by terms that cannot be com-puted from the LES solution and insteadmust be modeled to retain the effect ofthe small scales into the equations. Themathematical form of these models is asubject of contemporary research.

For a single-component liquid, there isonly one LES formulation, but this studyrevealed that for a multicomponent liq-uid, there are two non-equivalent LES for-mulations for the conservation equationsdescribing the composition of the vapor.Criteria were proposed for selecting themulticomponent LES formulation thatgives the best accuracy and increasedcomputational efficiency. These criteriawere applied in examination of filteredDNS databases to compute the terms inthe LES equations. The DNS databasesare from mixing layers of diesel andkerosene fuels. The comparisons resultedin the selection of one of the multicompo-nent LES formulations as the most prom-ising with respect to all criteria.

This work was done by Josette Bellan andLaurent Selle of Caltech for NASA’s Jet Propul-sion Laboratory. Further information is con-tained in a TSP (see page 1). NPO-45065

Computerized Machine forCutting Space Shuttle Ther-mal Tiles

A report presents the concept of a ma-chine aboard the space shuttle that wouldcut oversized thermal-tile blanks to pre-cise sizes and shapes needed to replacetiles that were damaged or lost during as-cent to orbit. The machine would includea computer-controlled jigsaw enclosed ina clear acrylic shell that would prevent es-cape of cutting debris. A vacuum motorwould collect the debris into a reservoir

and would hold a tile blank securely inplace. A database stored in the computerwould contain the unique shape and di-mensions of every tile.

Once a broken or missing tile was iden-tified, its identification number would beentered into the computer, wherein thecutting pattern associated with that num-ber would be retrieved from the database.A tile blank would be locked into a crib inthe machine, the shell would be closed(proximity sensors would prevent activa-tion of the machine while the shell wasopen), and a “cut” command would besent from the computer. A blade would bemoved around the crib like a plotter, cut-ting the tile to the required size andshape. Once the tile was cut, an astronautwould take a space walk for installation.

This work was done by Luis E. Ramirezand Lisa A. Reuter of The Boeing Co. forJohnson Space Center. For further informa-tion, contact the JSC Innovation PartnershipsOffice at (281) 483-3809. MSC-24140-1

Orbiting Depot andReusable Lander for LunarTransportation

A document describes a conceptualtransportation system that would supportexploratory visits by humans to locationsdispersed across the surface of the Moonand provide transport of humans andcargo to sustain one or more permanentLunar outpost. The system architecturereflects requirements to (1) minimize theamount of vehicle hardware that must beexpended while maintaining high per-formance margins and (2) take advan-tage of emerging capabilities to producepropellants on the Moon while also en-abling efficient operation using propel-lants transported from Earth.

The system would include reusable sin-gle-stage lander spacecraft and a depot in alow orbit around the Moon. Each landerwould have descent, landing, and ascentcapabilities. A crew-taxi version of the lan-der would carry a pressurized crew mod-ule; a cargo version could carry a variety ofcargo containers. The depot would serve asa facility for storage and for refueling withpropellants delivered from Earth or pro-pellants produced on the Moon. Thedepot could receive propellants and cargosent from Earth on a variety of spacecraft.The depot could provide power and orbit

maintenance for crew vehicles from Earthand could serve as a safe haven for lunarcrews pending transport back to Earth.

This work was done by Andrew Petro ofJohnson Space Center. Further information iscontained in a TSP (see page 1). MSC-24231-1

FPGA-Based NetworkedPhasemeter for a Hetero-dyne Interferometer

A document discusses a component ofa laser metrology system designed tomeasure displacements along the line ofsight with precision on the order of atenth the diameter of an atom. This com-ponent, the phasemeter, measures therelative phase of two electrical signals andtransfers that information to a computer.

Because the metrology system measuresthe differences between two optical paths,the phasemeter has two inputs, calledmeasure and reference. The reference sig-nal is nominally a perfect square wave witha 50- percent duty cycle (though only ris-ing edges are used). As the metrology sys-tem detects motion, the difference be-tween the reference and measure signalphases is proportional to the displacementof the motion. The phasemeter, therefore,counts the elapsed time between risingedges in the two signals, and converts thetime into an estimate of phase delay.

The hardware consists of a circuit boardthat plugs into a COTS (commercial, off-the-shelf) Spartan-III FPGA (field-pro-grammable gate array) evaluation board. Ithas two BNC inputs, (reference and meas-ure), a CMOS logic chip to buffer the in-puts, and an Ethernet jack for transmittingreduceddata to a PC. Two extra BNC con-nectors can be attached for future expand-ability, such as external synchronization.Each phasemeter handles one metrologychannel. A bank of six phasemeters (andtwo zero-crossing detector cards) with anEthernet switch can monitor the rigid bodymotion of an object.

This device is smaller and cheaperthan existing zero-crossing phasemeters.Also, because it uses Ethernet for com-munication with a computer, instead of aVME bridge, it is much easier to use. Thephasemeter is a key part of the PrecisionDeployable Apertures and Structuresstrategic R&D effort to design large, de-ployable, segmented space telescopes.


This work was done by Shanti Rao of Cal-tech for NASA’s Jet Propulsion Laboratory.Further information is contained in a TSP (seepage 1). NPO-45504

Aquarius Digital ProcessingUnit

Three documents provide informationon a digital processing unit (DPU) forthe planned Aquarius mission, in whicha radiometer aboard a spacecraft orbit-ing Earth is to measure radiometric tem-peratures from which data on sea-surfacesalinity are to be deduced. The DPU isthe interface between the radiometer

and an instrument-command-and-datasystem aboard the spacecraft. The DPUcycles the radiometer through a pro-grammable sequence of states, collectsand processes all radiometric data, andcollects all housekeeping data pertainingto operation of the radiometer. The doc-uments summarize the DPU design, withemphasis on innovative aspects that in-clude mainly the following: • In the radiometer and the DPU, conver-

sion from analog voltages to digital datais effected by means of asynchronousvoltage-to-frequency converters in com-bination with a frequency-measure-ment scheme implemented in field-pro-grammable gate arrays (FPGAs).

• A scheme to compensate for aging andchanges in the temperature of theDPU in order to provide an overalltemperature-measurement accuracywithin 0.01 K includes a high-preci-sion, inexpensive DC temperature-measurement scheme and a drift-com-pensation scheme that was used on theCassini radar system.

• An interface among multiple FPGAsin the DPU guarantees setup andhold times. This work was done by Joshua Forgione,

George Winkert, and Norman Dobson ofGoddard Space Flight Center. Further infor-mation is contained in a TSP (see page 1).GSC-15413-1


Three-Dimensional Optical Coherence TomographyAdvantages over prior OCT systems include less bulk and greater speed.John H. Glenn Research Center, Cleveland, Ohio

Three-dimensional (3D) optical coher-ence tomography (OCT) is an advancedmethod of noninvasive infrared imagingof tissues in depth. Heretofore, commer-cial OCT systems for 3D imaging havebeen designed principally for external

ophthalmological examination. As ex-plained below, such systems have beenbased on a one-dimensional OCT princi-ple, and in the operation of such a sys-tem, 3D imaging is accomplished partlyby means of a combination of electronic

scanning along the optical (Z) axis andmechanical scanning along the two axes(X and Y) orthogonal to the optical axis.

In 3D OCT, 3D imaging involves aform of electronic scanning (without me-chanical scanning) along all three axes.Consequently, the need for mechanicaladjustment is minimal and the mecha-nism used to position the OCT probe canbe correspondingly more compact. A 3DOCT system also includes a probe of im-proved design and utilizes advanced sig-nal-processing techniques. Improve -ments in performance over prior OCTsystems include finer resolution, greaterspeed, and greater depth of field.

Figure 1 includes a simplified sche -matic representation of the optical sub-system of a typical prior OCT system. Inthis system, near-infrared light from anincandescent lamp or other low-coher-ence source is sent through opticalfibers and a fiber-optic coupler to a ref-erence mirror. Some of the light is alsosent through the fiber optics to a lensthat, in turn, focuses the light to apoint that lies at or near the depth ofinterest in a specimen. In the fiber-optic coupler, light reflected from thereference mirror is combined withlight scattered from a focal point in thespecimen and is then sent along an-other optical fiber to a photodetector.When the length of the optical pathfrom the light source to the mirrorequals or nearly equals the correspon-ding length to the focal point in thespecimen, the photodetector puts out a

Figure 1. In a Typical Prior OCT System, scanning in Z involves piezoelectric actuation of the referencemirror, and scanning in X and Y is involves mechanical motion of the probe.

Figure 2. In a 3D OCT System, scanning in all three dimensions involves a combination of amplitude modulation, nonlinear detection, and advanced signalprocessing.

SCHEMATIC OPTICAL LAYOUT

Lens

LensSpecimen

PiezoelectricTransducer

Reference Mirror

Light Source

Photodetector

Optical FiberFiber-Optic Coupler

INSTRUMENTAL RESPONSE (POINT-SPREAD FUNCTION)

Time During Transducer Actuation Cycle orCorresponding Mirror Displacement

Envelope

InstantaneousAmplitude

Inst

anta

neo

us

Am

plit

ud

eo

r En

velo

pe

Am

plit

ud

e

Single Mode Optical Fibers

Multimode Optical Fiber

MAGNIFIED VIEW OF PROBE TIP

Z Point-SpreadFunction

X or YPoint-Spread

Function

11

42

3

ConnectorPhase Shifter

Amplitude Modulator

Light Source

Photodetector

PersonalComputer

AM1

To Fiber TipsDisposable Probe

Optical Fibers

2

3

4

4312

AM2

AM3

AM4

1ϕx

ϕy

ϕz

Bio-Medical


signal representing a pixel at the focalpoint in the specimen. Scanning alongthe depth (Z) axis is accomplished byusing the piezoelectric transducer tomove the reference mirror closer to, orfarther from, the light source. Scan-ning along the X and Y axes is accom-plished by mechanical motion of theprobe along X and Y.

The lower part of Figure 1 depicts atypical instrumental response (point-spread function) in the photodetectoroutput obtained in scanning along theZ axis. The response includes oscilla-tions attributable to interference be-tween the light scattered from a pointin the specimen and light scatteredfrom the mirror. As the Z displacementincreases, the contrast of the interfer-ence pattern is reduced by the loss ofcoherence. Usually, the envelope of theoscillations (in contradistinction to theoscillations themselves) is what is meas-ured. In such a case, the width of theenvelope and, thus, the depth resolu-tion, is comparable to the coherencelength of the light source.

Figure 2 includes a simplified sche -matic representation of the optical sub-system of a 3D OCT system. This systemis based partly on the same principles asthose of the prior system. However,there are several important differences:• Light from the source is fed through a

more-complex fiber-optic subsystem,not only to a photodiode but to threesingle-mode optical fibers on a probe.Light emerging from the tips of thesethree fibers illuminates the specimenand creates a 3D interference patternin the specimen.

• Light scattered from the specimen iscollected and sent to the photodetec-tor by a wider, multimode optical fiber.The probe containing the illuminatingsingle-mode fibers and the light-col-lecting multimode optical fibers is sig-nificantly smaller and more rugged,relative to a lens-containing probe in aprior OCT system.

• Instead of utilizing lenses and a piezo-electric actuation of a reference mir-ror to effect scanning in Z and focus-ing in conjunction with mechanical

scanning in X and Y, the system uti-lizes a combination of (1) amplitudemodulation of the light in the three il-luminating optical fibers and of a por-tion of the source light sent directly tothe photodetector, (2) nonlinear de-tection, and (3) an advanced signal-processing technique that, amongother things, exploits the 3D nature ofthe interference pattern in order toobtain (4) a 3D point-spread functionthat affords localization in X, Y, and Z.In principle, because mechanicalscanning is no longer necessary, it ispossible to achieve scanning at a videoframe rate.This work was done by Mikhail Gutin, Xu-

Ming Wang, and Olga Gutin of Applied Sci-ence Innovations, Inc. for Glenn ResearchCenter. Further information is contained in aTSP (see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, InnovativePartnerships Office, Attn: Steve Fedor, MailStop 4–8, 21000 Brookpark Road, Cleve-land, Ohio 44135. Refer to LEW-18352-1.

Benchtop Antigen Detection Technique Using Nanofiltrationand Fluorescent DyesThis technique can help to monitor the quality of water by testing for contamination at restau-rants, water treatment plants, and food processing plants.John H. Glenn Research Center, Cleveland, Ohio

The designed benchtop technique isprimed to detect bacteria and virusesfrom antigenic surface marker proteinsin solutions, initially water. This inclusivebio-immunoassay uniquely combinesnanofiltration and near infrared (NIR)dyes conjugated to antibodies to isolateand distinguish microbial antigens,using laser excitation and spectrometricanalysis. The project goals include de-tecting microorganisms aboard the In-ternational Space Station, space shuttle,Crew Exploration Vehicle (CEV), andhuman habitats on future Moon andMars missions, ensuring astronaut safety.The technique is intended to improveand advance water contamination test-ing both commercially and environmen-tally as well. Lastly, this streamlined tech-nique poses to greatly simplify andexpedite testing of pathogens in com-plex matrices, such as blood, in hospitaland laboratory clinics.

The approach relies on NIR fluores-cent dyes derivatized to specific anti-

body sets that are selected to bind anddifferentiate microbial surface pro-teins, termed antigens. In a solutioncontaining an antigenic slurry, NIRconjugated antibodies are added to themixture, and will bind to the respectiveantigens if present. To eliminate anyfalse positives, excess antibodies, i.e.,those antibodies not bound to an anti-genic protein or those with no respec-tive antigen present, are removed viathe nanofiltration process using aportable, table-top centrifuge. The re-maining NIR dye/antibody and anti-genic protein pairs left on the nanofil-ter are transferred to cuvette, excitedby an NIR laser, and detected by spec-trometer. Using simple computer soft-ware, the results are easily interpretedas intensity peaks at the appropriateNIR offset wavelength emission.

Initial data reveal the assay sensitivelyidentified antigens at intensity countsof 100 IC or higher (or roughly 36 pW)with an accuracy of 85 percent for 2-

a.

b.

The Benchtop Analysis System (a) consists of amedium-power, class IV laser, a four-port cuvettesampler holder, a spectrometer, laptop and soft-ware, and fiber-optic cables. (b) The microcen-trifuge used in nanofiltration of the complexes.


A process for isolating tissue-specificprogenitor cells exploits solid fat tissueobtained as waste from such elective sur-gical procedures as abdominoplasties(“tummy tucks”) and breast reductions.Until now, a painful and risky process ofaspiration of bone marrow has beenused to obtain a limited number of tis-sue-specific progenitor cells.

The present process yields more tissue-specific progenitor cells and involves

much less pain and risk for the patient.This process includes separation of fatfrom skin, mincing of the fat into smallpieces, and forcing a fat saline mixturethrough a sieve. The mixture is then di-gested with collagenase type I in an incu-bator. After centrifugation tissue-specificprogenitor cells are recovered and placedin a tissue-culture medium in flasks orPetri dishes. The tissue-specific progeni-tor cells can be used for such purposes as

(1) generating three-dimensional tissueequivalent models for studying bone lossand muscle atrophy (among other defi-ciencies) and, ultimately, (2) generatingreplacements for tissues lost by the fatdonor because of injury or disease.

This work was done by Diane Byerly ofJohnson Space Center and Marguerite A. Sog-nier of Universities Space Research Associa-tion. Further information is contained in aTSP (see page 1). MSC-23775-1

Isolation of Precursor Cells From Waste Solid Fat TissueLyndon B. Johnson Space Center, Houston, Texas

hour incubations and 75 percent for 3-hour incubations. Interestingly, sam-ples incubated for less time (2 hours vs3 hours) produced an increased per-centage of antigen detection. Furthertesting at incubation times such as 1hour or lower could potentially in-crease positive predictability based on

the study’s results. Also encouragingwere negative control experiments withnonspecific antigens, beta galactosidaseand thyroglobulin, which showed re-sults of 100 percent accuracy, with nofalse positive detection.

This work was done by Maximilian C.Scardelletti and Vanessa Varaljay of Glenn Re-

search Center. Further information is con-tained in a TSP (see page 1).

Inquiries concerning rights for the commer-cial use of this invention should be addressedto NASA Glenn Research Center, InnovativePartnerships Office, Attn: Steve Fedor, MailStop 4–8, 21000 Brookpark Road, Cleve-land, Ohio 44135. Refer to LEW-18387-1.

The ultimate goal of planetary protec-tion research is to develop superiorstrategies for inactivating resistance-bearing micro-organisms like Rummeli -bacillus stabekisii. By first identifying theparticular physiologic pathway and/orstructural component of the cell/sporethat affords it such elevated tolerance,eradication regimes can then be de-signed to target these resistance-confer-ring moieties without jeopardizing thestructural integrity of spacecraft hard-ware. Fur thermore, hospitals and gov-ernment agencies frequently use biolog-ical indicators to ensure the efficacy of awide range of sterilization processes.The spores of Rummelibacillus stabekisii,which are far more resistant to many ofsuch perturbations, could likely serve asa more significant biological indicatorfor potential survival than those beingused currently.

Numerous surveys of the contami-nant microbial diversity housedwithin spacecraft assembly facilitiesover the past six years have resulted

in the recurrent isolation of spore-forming bacteria belonging to theBacillus genus. As Bacillus species arecapable of existing as metabolicallyinactive, extremely hardy spores,many lineages exhibit remarkable re-silience to varying modes of biore-duction/sterilization aimed at theireradication (UV and gamma radia-tion, oxidizing disinfectants, etc.).The microorganism Rummelibacillusstabekisii sp. nov. was isolated from thesurfaces of the cleanroom facility inwhich the Mars Exploration Rovers(MER) underwent assembly. This bac-terium has not been previously re-ported, and shows no close relationto any previously described species(as is assessed via 16S rRNA gene se-quence comparison). This uniqueisolate, and the Bacillus species mostgenetically similar to it, were sub-jected to a multitude of biochemicaltests in order to thoroughly charac-terize its taxonomic position basedon physiological and phylogenetic ev-

idence. The results clearly show thatthis bacterium is significantly differ-ent from its nearest relatives.

The microbial colonization of space-craft and cleanroom assembly facilitysurfaces is of major concern to NASAand others commissioning modern-dayspace exploration. The search for lifeelsewhere in the solar system will relyheavily on validated cleaning andsterility methods. It would be devastat-ing to the integrity of a mission di-rected at pristine environments suchas the Europa’s subsurface ocean orthe Martian polar caps to be compro-mised as a result of terrestrial micro-bial contamination. To this end, plane-tary protection policies are in place toensure the cleanliness and sterility ofmission-critical spacecraft componentsin order to prevent forward or back-ward contamination.

Spores of Bacillus subtilis, a modelspore-forming laboratory strain thatdemonstrates higher susceptibility toultraviolet and gamma radiation than

Identification of Bacteria and Determination of Biological Indicators Identifying mechanisms of micro-organisms can prevent forward contamination in space mis-sions and can help in developing new antibiotics and amino acids. NASA’s Jet Propulsion Laboratory, Pasadena, California


other wild type spore formers, havenevertheless been shown to survive upto six years under interstellar spaceconditions. Previously undescribedspore-forming species, such as Rum-melibacillus stabekisii, may exhibit even

greater resilience. It is in the best inter-ests of NASA to thoroughly understandthe physiological capabilities of eachand every novel micro-organism iso-lated from these spacecraft-associatedcleanrooms.

This work was done by KasthuriVenkateswaran, Myron T. La Duc, andParag A. Vaishampayan of Caltech forNASA’s Jet Propulsion Laboratory. For moreinformation, contact [email protected].

Further Development of Scaffolds for Regeneration of Nerves Scale-up toward clinically significant dimensions has been partially completed. NASA’s Jet Propulsion Laboratory, Pasadena, California

Progress has been made in continuingresearch on scaffolds for the guidedgrowth of nerves to replace damagedones. The scaffolds contain pores that areapproximately cylindrical and parallel,with nearly uniform widths ranging fromtens to hundreds of microns. At the ear-lier stage of development, experimentalscaffolds had been made from agarosehydrogel. Such a scaffold was made in amultistep process in which poly(methylmethacrylate) [PMMA] fibers were usedas templates for the pores. The processincluded placement of a bundle of thePMMA fibers in a tube, filling the inter-stices in the tube with a hot agarose solu-tion, cooling to turn the solution into agel, and then immersion in acetone to

dissolve the PMMA fibers. The scaffoldswere typically limited to about 25 poresper scaffold, square cross sections of nomore than about 1.5 by 1.5 mm, andlengths of no more than about 2 mm.

To be clinically relevant, the scaffoldsmust be scaled up: They are required tohave typical cross-sectional dimensionsof the order of 1 cm and to have lengthsin the approximate range of 2 to 2.5 cm.For repairs of the central nervous sys-tem, there is an additional requirementthat each scaffold contain betweenabout 100 and about 1,000 pores; for re-pairs of peripheral nerves, there is a re-quirement for sustained or timed re-lease of brain-derived neurotrophicfactor (BDNF) or another suitable

nerve-growth agent to enable growth tocontinue to the required lengths.

The work performed since the earlierstage has been oriented toward satisfyingthese and other requirements. The workhas included development of a more-complex version of the prior multistepprocess that has made it possible topartly satisfy the scaling-up requirementsin that scaffolds having cross sections ex-ceeding 1 cm2 in area and lengths up to1 cm have been fabricated. One notablefeature of the present version of theprocess is a multistep subprocess inwhich a template of polystyrene (PS)fibers is made from a composite of poly-styrene fibers surrounded by a continu-ous PMMA matrix. Another notable fea-

A Nerve-Growth Scaffold containing a reservoir of a nerve-growth agent would be attached to severed ends of a peripheral nerve.

PeripheralNerve

PeripheralNerve

Reservoir ofNerve-Growth Agent

>5 mm

AgaroseHydrogel

Reservoir

Channels

CROSS SECTION


ture of the present version of the processis the use of centrifugation to ensurecomplete permeation of the template bythe hot agarose solution.

To satisfy the requirement for sus-tained or timed release of nerve-growthagents, it has been proposed to incorpo-rate, into scaffolds, reservoirs containingsuch agents. In cases in which the agentis BDNF, the proposal encompasses an al-ternative approach in which the reser-voirs would be filled with genetically en-gineered cells that secrete BDNF. Thefigure illustrates the proposal as it mightbe implemented in a scaffold that would

be attached to the severed ends of a pe-ripheral nerve. Attached to the scaffoldwould be open-ended sleeves that wouldenable attachment to the severed nerveends. The pores in the scaffold wouldserve as channels to guide the growth ofthe nerve ends toward each other. Thereservoir containing the nerve-growthagent would be integrated into the outerwall of the scaffold. The nerve-growthagent would be delivered from the reser-voir to the channels by diffusion throughthe agarose hydrogel matrix.

This work was done by Jeffrey Sakamotoof Caltech and Mark Tuszynski of UC

San Diego for NASA’s Jet Propulsion Lab-oratory.




The purpose of this study was to deter-mine the effect of atomic oxygen (AO)exposure on the hydrophilicity of ninedifferent polymers for biomedical appli-cations. Atomic oxygen treatment canalter the chemistry and morphology ofpolymer surfaces, which may increasethe adhesion and spreading of cells onPetri dishes and enhance implantgrowth. Therefore, nine different poly-mers were exposed to atomic oxygenand water-contact angle, or hydrophilic-ity, was measured after exposure. To de-termine whether hydrophilicity remainsstatic after initial atomic oxygen expo-sure, or changes with higher fluence ex-

posures, the contact angles between thepolymer and water droplet placed onthe polymer’s surface were measuredversus AO fluence. The polymers wereexposed to atomic oxygen in a 100-W,13.56-MHz radio frequency (RF) plasmaasher, and the treatment was found tosignificantly alter the hydrophilicity ofnon-fluorinated polymers.

Pristine samples were compared withsamples that had been exposed to AOat various fluence levels. Minimum andmaximum fluences for the ashing trialswere set based on the effective AO ero-sion of a Kapton witness coupon in theasher. The time intervals for ashing

were determined by finding the loga-rithmic values of the minimum andmaximum fluences. The difference ofthese two values was divided by the de-sired number of intervals (ideally 10).The initial desired fluence was thenmultiplied by this result (2.37), as waseach subsequent desired fluence. Theflux in the asher was determined to beapproximately 3.0 × 1015 atoms/cm2

sec, and each polymer was exposed to amaximum fluence of 5.16 × 1020

atoms/cm2.It was determined that after the

shortest atomic oxygen exposure (flu-ence of 2.07 × 1018 atoms/cm2), non-

Chemically Assisted Photocatalytic Oxidation SystemLyndon B. Johnson Space Center, Houston, Texas

The chemically assisted photocat-alytic oxidation system (CAPOS) hasbeen proposed for destroying microor-ganisms and organic chemicals thatmay be suspended in the air or presenton surfaces of an air-handling systemthat ventilates an indoor environment.The CAPOS would comprise an up-stream and a downstream stage thatwould implement a tandem combina-tion of two partly redundant treat-ments. In the upstream stage, the airstream and, optionally, surfaces of theair-handling system would be treatedwith ozone, which would be generatedfrom oxygen in the air by means of an

electrical discharge or ultraviolet light.In the second stage, the air laden withozone and oxidation products fromthe first stage would be made to flow incontact with a silica-titania photocata-lyst exposed to ultraviolet light in thepresence of water vapor. Hydroxyl rad-icals generated by the photocatalyticaction would react with both carbon-containing chemicals and microorgan-isms to eventually produce water andcarbon dioxide, and ozone from thefirst stage would be photocatalyticallydegraded to O2. The net products ofthe two-stage treatment would be H2O,CO2, and O2.

This work was done by Jean Andino,Chang-Yu Wu, David Mazyck, and Arthur A.Teixeira of the University of Florida for JohnsonSpace Center. Further information is con-tained in a TSP (see page 1).


University of FloridaEnvironmental EngineeringWeil HallGainesville, FL 32611Refer to MSC-23828-1, volume and num-


Use of Atomic Oxygen for Increased Water Contact Angles ofVarious Polymers for Biomedical ApplicationsImproved polymer hydrophilicity is beneficial for cell culturing and implant growth.John H. Glenn Research Center, Cleveland, Ohio


fluorinated polymer samples becamesignificantly more hydrophilic thantheir pristine counterparts. This maybe due to either surface texturechanges or oxidation functionality sur-face changes. Despite long-term expo-sure (fluence of 5.16 × 1020

atoms/cm2), the water contact angles

remained relatively unchanged afterthe initial exposure. This implies thatincreasing the atomic oxygen fluenceafter an initial short exposure did notfurther affect the hydrophilicity of thepolymers. Rather, polymers were af-fected by a very short exposure (<1 ×1019 atoms/cm2). This indicates that

oxidation functionality is morelikely the contributor to increasedhydrophilicity than texture, as tex-ture continues to develop with flu-ence. The water contact angles offluorinated polymers were found tochange significantly less than non-fluorinated polymers for equivalentatomic oxygen exposures, and twoof the fluorinated polymers becamemore hydrophobic.

Significant decreases in the post-exposure water contact angle weremeasured for non-fluorinated poly-mers. The majority of change inwater contact angle was found tooccur with very low fluence expo-sures, indicating potential cell cultur-ing and other biomedical benefitswith very short treatment time.

This work was done by Kim de Grohof Glenn Research Center; LaurenBerger and Lily Roberts of HathawayBrown School; and Bruce Further infor-

mation is contained in a TSP (see page 1).Inquiries concerning rights for the commer-cial use of this invention should be ad-dressed to NASA Glenn Research Center, In-novative Partnerships Office, Attn: SteveFedor, Mail Stop 4–8, 21000 BrookparkRoad, Cleveland, Ohio 44135. Refer toLEW-18386-1.

Polymers Tested for atomic oxygen-altered hydrophilicity.

Abbreviation Polymer Name Trade Name Thickness

PE Polyethylene Alathon; Lupolen 2 mil

PET Polyethylene terephthalate Mylar A 2 mil

POM Polyoxymethylene Delrin; Celcon 4 mil

PS Polystyrene Lustrex; Polystyrol 2 mil

PP Polypropylene Profax; Propathene 20 mil

PMMA Polymethylmethacrylate Plexiglas; Lucite 2 mil

FEP Fluorinated ethylene propylene Teflon FEP 2 mil

PTFE Polytetrafluoroethylene Fluon; Teflon 2 mil

PCTFE Polychlorotrifluoroethylene Neoflon CTFE M-300 5 mil

Seats to prevent or limit crash injuriesto astronauts aboard the crew vehicle ofthe Orion spacecraft are undergoing de-velopment. The design of these seats in-corporates and goes beyond crash-protec-tion concepts embodied in priorspacecraft and racing-car seats to affordsuperior protection against impacts. Al-though the seats are designed to supportastronauts in a recumbent, quasi-fetal pos-ture that would likely not be suitable fornon-spacecraft applications, parts of thedesign could be adapted to military andsome civilian aircraft seats and to racing-car seats to increase levels of protection.

The main problem in designing anycrashworthy seat is to provide full sup-port of the occupant against anticipatedcrash and emergency-landing loads so asto safely limit motion, along any axis, ofany part of the occupant’s body relativeto (1) any other part of the occupant’sbody, (2) the spacecraft or other vehicle,

and (3) the seat itself. In the originalOrion spacecraft application and inother applications that could easily beenvisioned, the problem is complicatedby severe limits on space available forthe seat, a requirement to enable rapidegress by the occupant after a crash, anda requirement to provide for fitting ofthe seat to a wide range of sizes andshapes of a human body covered by acrash suit, space suit, or other protectivegarment. The problem is further com-plicated by other Orion-application-spe-cific require ments that must be omittedhere for the sake of brevity.

To accommodate the wide range ofcrewmember body lengths within thelimits on available space in the originalOrion application, the design providesfor taller crewmembers to pull their legsback closer toward their chests, whileshorter crewmembers can allow theirlegs to stretch out further. The range of

hip-support seat adjustments needed toeffect this accommodation, as derivedfrom NASA’s Human Systems Inte -gration Standard, was found to define aparabolic path along which the kneesmust be positioned. For a given occu-pant, the specific position along thepath depends on the distance from theheel to the back of the knee.

The application of the concept of par-abolic adjustment of the hip-supportstructure caused the seat pan to also takeon a parabolic shape, yielding the unan-ticipated additional benefit that the seatpan fits the occupant’s buttocks andthighs more nearly conformally than doseat pans of prior design. This morenearly conformal fit effectively elimi-nates a void between the occupant’sbody and the seat pan, thereby helpingto prevent what, in prior seat designs,was shifting of the occupant’s body intothat void during an impact.

Crashworthy Seats Would Afford Superior ProtectionAdjustments enable optimization of support for different body sizes and shapes.Lyndon B. Johnson Space Center, Houston, Texas


Open-Access, Low-Magnetic-Field MRI System for Lung ResearchLyndon B. Johnson Space Center, Houston, Texas

An open-access magnetic resonanceimaging (MRI) system is being developedfor use in research on orientational/grav-itational effects on lung physiology andfunction. The open-access geometry en-ables study of human subjects in diverseorientations. This system operates at amagnetic flux density, considerablysmaller than the flux densities of typicalother MRI systems, that can be generatedby resistive electromagnet coils (insteadof the more-expensive superconductingcoils of the other systems).

The human subject inhales air con-taining 3He or 129Xe atoms, the nu cle -ar spins of which have been polarizedby use of a laser beam to obtain a mag-

netic resonance that enables high-reso-lution gas space imaging at the low ap-plied magnetic field. The system in-cludes a bi-planar, constant-current,four-coil electromagnet assembly andassociated electronic circuitry to applya static magnetic field of 6.5 mTthroughout the lung volume; planarcoils and associated circuitry to apply apulsed magnetic-field-gradient foreach spatial dimension; a single, de-tachable radio-frequency coil and asso-ciated circuitry for inducing and de-tecting MRI signals; a table forsupporting a horizontal subject; andelectromagnetic shielding surroundingthe electromagnet coils.

This work was done by Ross W. Mair,Matthew S. Rosen, Leo L. Tsai, and RonaldL. Walsworth of the Harvard-SmithsonianCenter for Astrophysics; Mirko I. Hrovat ofMirtech, Inc.; Samuel Patz of Brigham andWomen’s Hospital; and Iullian C. Ruset andF. William Hersman of the University of NewHampshire for Johnson Space Center.


Harvard-Smithsonian Center for AstrophysicsCambridge, MA 02138Refer to MSC-24182-1/3-1, volume and

number of this NASA Tech Briefs issue, andthe page number.

Microfluidic Mixing Technology for a Universal Health SensorJohn H. Glenn Research Center, Cleveland, Ohio

A highly efficient means of microflu-idic mixing has been created for usewith the rHEALTH sensor — an ellipti-cal mixer and passive curvilinear mix-ing patterns. The rHEALTH sensor pro-vides rapid, handheld, complete bloodcount, cell differential counts, elec-trolyte measurements, and other labtests based on a reusable, flow-based mi-crofluidic platform.

These geometries allow for cleaningin a reusable manner, and also allow for

complete mixing of fluid streams. Themicrofluidic mixing is performed byflowing two streams of fluid into an ellip-tical or curvilinear design that allows thecombination of the flows into one chan-nel. The mixing is accomplished by ei-ther chaotic advection around micro -fluidic loops.

All components of the microfluidicchip are flow-through, meaning thatcleaning solution can be introducedinto the chip to flush out cells, plasma

proteins, and dye. Tests were performedon multiple chip geometries to showthat cleaning is efficient in any flow-through design. The conclusion fromthese experiments is that the chip canindeed be flushed out with microlitervolumes of solution and biological sam-ples are cleaned readily from the chipwith minimal effort.

The technology can be applied inreal-time health monitoring at patient’sbedside or in a doctor’s office, and real-

The seat includes a thigh-support struc-ture and a lower-leg support structure thatcan be adjusted for various heel-to-back-of-knee lengths. The occupant’s heels aresupported by a heel support pan andcould be affixed to the pan by clips similarto those of mountain biking shoes andpedals or by straps over the tops of thefeet. At the pivot between the thigh andlower-leg support structures there is a flatpanel that provides for strength in adjust-ment and provides lateral support of theknees. The combination of lateral sup-port of the knees and support and re-straint the of the feet is intended to pre-vent flailing or other movement of thelegs while the occupant is seated.

The seat includes lateral supports atthe hips that serve the dual purpose of

restraining the occupant from shiftinglaterally and providing structural sup-port to the rest of the seat by acting as agusset. To accommodate all hip sizes,the seat pan is designed to fit thelargest hip breadth allowable. For asmaller occupant, spacer pads can beinstalled to fill the voids. Shoulder sup-ports, which cover the shoulder jointsand extend short distances down thearms, are also sized for occupants hav-ing shoulders of maximum breadth andto be fitted to smaller occupants by useof spacer pads. The seat supports bonyprotrusions of the torso at the shoul-ders and hips only, leaving the mid-torso area free of supports to enablethe occupant to leave the seat by rollingthough the clear space.

The seat includes a head support. How-ever, head support on the prototype differsfrom the envisioned Orion head support:In Orion, the occupant would wear a spacehelmet and the head support would extendalong the right and left sides of the helmetto prevent lateral motion of the head.

Another prominent design feature is aload-distributing seven-point harnesssimilar to harnesses worn in off-road au-tomobile racing. The seven-point har-ness includes straps over the tops of theshoulders that act, in effect, as wrap-around extensions of the lateral shoul-der supports.

This work was done by Dustin Gohmert ofJohnson Space Center. Further information iscontained in a TSP (see page 1). MSC-24485-1


time clinical intervention in acute situa-tions. It also can be used for daily meas -urement of hematocrit for pa tients onanticoagulant drugs, or to detect acutemyocardial damage outside a hospital.

This work was done by Eugene Y Chanand Candice Bae of DNA Medicine Institutefor Glenn Research Center.

Inquiries concerning rights for the com-mercial use of this invention should be ad-

dressed to NASA Glenn Research Center, In-novative Partnerships Office, Attn: SteveFedor, Mail Stop 4–8, 21000 BrookparkRoad, Cleveland, Ohio 44135. Refer toLEW-18391-1.

Microfluidic Extraction of Biomarkers Using Water as Solvent Terahertz modulation of permittivity of water would enable solvation of molecules of interest.NASA’s Jet Propulsion Laboratory, Pasadena, California

A proposed device, denoted a minia-ture microfluidic biomarker extractor(µ-EX), would extract trace amounts ofchemicals of interest from samples, suchas soils and rocks. Traditionally, such ex-tractions are performed on a large scalewith hazardous organic solvents; eachsolvent capable of dissolving only thosemolecules lying within narrow ranges ofspecific chemical and physical charac-teristics that notably include volatility,electric charge, and polarity. In con-trast, in the µ-EX, extractions could beperformed by use of small amounts(typically between 0.1 and 100 µL) ofwater as a universal solvent.

As a rule of thumb, in order to enablesolvation and extraction of molecules, it isnecessary to use solvents that have polaritysufficiently close to the polarity of the tar-get molecules. The µ-EX would make se-lection of specific organic solvents unnec-essary, because µ-EX would exploit aunique property of liquid water: the possi-bility of tuning its polarity to match thepolarity of organic solvents appropriatefor extraction of molecules of interest.The change of the permittivity of waterwould be achieved by exploiting interac-tions between the translational states ofwater molecules and an imposed electro-magnetic field in the frequency range of

300 to 600 GHz. On a molecular level,these interactions would result in disrup-tion of the three-dimensional hydrogen-bonding network among liquid-watermolecules and subsequent solvation andhydrolysis of target molecules. The µ-EXis expected to be an efficient means of hy-drolyzing chemical bonds in complexmacromolecules as well and, thus, en-abling analysis of the building blocks ofthese complex chemical systems.

The µ-EX device would include a mi-crofluidic channel, part of which wouldlie within a waveguide coupled to an elec-tronically tuned source of broad-bandelectromagnetic radiation in the fre-

UnsolvatedMacromolecule

Hydrogen-BondingNetwork of

Water MoleculesDisrupted

Hydrogen-BondingNetwork of

Water Molecules

AppliedElectromagnetic

Field(300 to 600 GHz)

Waveguide Inlet Outlet

Flow

Microfluidic ChannelInteraction Volume

Solvated andHydrolized Molecules

Water Flowing in a Microfluidic Channel would be exposed to electromagnetic radiation, causing the solvation of biomarker macromolecules that wouldotherwise be insufficiently soluble in water.

quency range from 300 to 600 GHz (seefigure). The part of the microfluidicchannel lying in the waveguide wouldconstitute an interaction volume. The di-mensions of the interaction volumewould be chosen in accordance with theanticipated amount of solid sample mate-rial needed to ensure extraction of suffi-

cient amount of target molecules for de-tection and analysis. By means that werenot specified at the time of reporting theinformation for this article, the solid sam-ple material would be placed in the inter-action volume. Then the electromagneticfield would be imposed within the wave-guide and water would be pumped

through the interaction volume to effectthe extraction.

This work was done by Xenia Amashukeli,Harish Manohara, Goutam Chattopadhyay,and Imran Mehdi of Caltech for NASA’s JetPropulsion Laboratory. For more informa-tion, contact [email protected]. NPO-46150

A process for making monodis-perse liposomes having lipid bilayermembranes involves fewer, simplerprocess steps than do related priormethods. First, a microfluidic, cross-junction droplet generator is used toproduce vesicles comprising aqueous-solution droplets contained in single-layer lipid membranes. The vesiclesare collected in a lipid-solvent mixthat is at most partially soluble inwater and is less dense than is water.A layer of water is dispensed on top ofthe solvent. By virtue of the differ-ence in densities, the water sinks tothe bottom and the solvent floats to

the top. The vesicles, which have al-most the same density as that ofwater, become exchanged into thewater instead of floating to the top.As there are excess lipids in the sol-vent solution, in order for the vesiclesto remain in the water, the additionof a second lipid layer to each vesicleis energetically favored.

The resulting lipid bilayers presentthe hydrophilic ends of the lipid mole-cules to both the inner and outer mem-brane surfaces. If lipids of a second kindare dissolved in the solvent in sufficientexcess before use, then asymmetric lipo-somes may be formed.

This work was done by Donald E. Ackleyand Anita Forster of Nanotrope, Inc. for John-son Space Center. For further information,contact the JSC Innovation Partnerships Of-fice at (281) 483-3809.


Nanotrope Inc.3030 Bunker Hill StSan Diego, CA 92109Phone No.: (858) 270-7992Refer to MSC-24302-1, volume and num-


Droplet-Based Production of LiposomesLyndon B. Johnson Space Center, Houston, Texas


Microwell Arrays for Studying Many Individual CellsLyndon B. Johnson Space Center, Houston, Texas

“Laboratory-on-a-chip” devices that en-able the simultaneous culturing and in-terrogation of many individual livingcells have been invented. Each such de-vice includes a silicon nitride-coated sili-con chip containing an array of micro-machined wells sized so that each wellcan contain one cell in contact or prox-imity with a patch clamp or other suit-able single-cell-interrogating device. Atthe bottom of each well is a hole, typi-cally ≈ 0.5 µm wide, that connects thewell with one of many channels in a mi-

crofluidic network formed in a layer ofpoly(dimethylsiloxane) on the under-side of the chip. The microfluidic net-work makes it possible to address wells(and, thus, cells) individually to supplythem with selected biochemicals. Themicrofluidic channels also provide elec-trical contact to the bottoms of the wells.

This work was done by Albert Folch andTurgut Fettah Kosar of the University of Wash-ington for Johnson Space Center. For furtherinformation, contact the JSC Innovation Part-nerships Office at (281) 483-3809.

In accordance with Public Law 96-517,the contractor has elected to retain title tothis invention. Inquiries concerningrights for its commercial use should be ad-dressed to:

ROI CoordinatorOffice of Technology LicensingUniversity of Washington1107 NE 45th Street, Suite 200Seattle, WA 98105Refer to MSC-24046-1, volume and num-


Identifying and Inactivating Bacterial Spores NASA’s Jet Propulsion Laboratory, Pasadena, California

Problems associated with, and newstrategies for, inactivating resistant or-ganisms like Bacillus canaveralius(found at Kennedy Space Center dur-ing a survey of three NASA clean-rooms) have been defined. Identifying

the particular component of the sporethat allows its heightened resistancecan guide the development of steriliza-tion procedures that are targeted tothe specific molecules responsible forresistance, while avoiding using un-

duly harsh methods that jeopardizeequipment.

The key element of spore resistance isa multilayered protein shell that encasesthe spore called the spore coat. The coatof the best-studied spore-forming mi-


crobe, B. subtilis, consists of at least 45proteins, most of which are poorly char-acterized. Several protective roles for thecoat are well characterized including re-sistance to desiccation, large toxic mole-cules, ortho-phthalaldehyde, and ultravi-olet (UV) radiation.

One important long-term specific goalis an improved sterilization procedure thatwill enable NASA to meet planetary pro-tection requirements without a terminalheat sterilization step. This would support

the implementation of planetary protec-tion policies for life-detection missions.Typically, hospitals and government agen-cies use biological indicators to ensure thequality control of sterilization processes.The spores of B. canaveralius that are moreresistant to osmotic stress would serve as abetter biological indicator for potentialsurvival than those in use currently.

This work was done by David Newcombe,Anne Dekas, and Kasthuri Venkateswaran ofCaltech for NASA’s Jet Propulsion Laboratory.


Innovative Technology Assets ManagementJPLMail Stop 202-2334800 Oak Grove DrivePasadena, CA 91109-8099E-mail: [email protected] to NPO-45182, volume and number of this

NASA Tech Briefs issue, and the page number.

National Aeronautics andSpace Administration

Technology Focus Electronics/Computers - NASA€¦ · 27 Suppressing Ghost Diffraction in...

Documents

Transcript of Technology Focus Electronics/Computers - NASA€¦ · 27 Suppressing Ghost Diffraction in...