News Briefs and Reports - NISTNews Briefs and Reports Developments GROUND BROKEN FOR NBS COLD...

Volume 93, Number 1, January-February 1988

Journal of Research of the National Bureau of Standards

In This Issue:DepartmentsNews Briefs and Reports

DEVELOPMENTSGround Broken for NBS Cold Neutron Research FacilityDetails Now Available on Superconducting ContactsNBS Testifies on Superconductivity LegislationNew Acoustic Technique Promises Better Way to Evaluate Installation of

Home Thermal InsulationNBS Reports to Congress on Structural Damage in Recent California QuakeNBS/GCA Joint Program on "Molecular Measurement Machine"Polymer Composite Processing Research Needs IdentifiedNBS Proposes Federal Standard for OSIMeasuring Faults in Parallel ProcessorsResearch Detects Nuclear Reactor Burn-Up of USSR CraftNBS Reports Cause of Fatal Building CollapseGood Atrium Design Can Save EnergyNew Way to Test Setting Time, Strength of Concrete

1

STANDARD REFERENCE MATERIALS 5Improving N-PAH Measurements Is Aim of New MaterialNBS, E. Fjeld Co. Program on Submicrometer Line Standards

NBS SERVICES 5NBS, NSA Setting Up Risk Management LaboratoryNBS Announces Accreditation Program for Laboratories

Testing Asbestos in Schools

Conferences/Events

Computer Security Conference Irene Isaac 71

CALENDAR 77



ArticlesThe NBS Scale of Spectral Irradiance James H. Walker, Robert D. Saunders, 7

John K. Jackson, and Donald A. McSparron

Radiometric Calibrations of Portable Jules Z. Klose, J. Mervin Bridges, 21Sources in the Vacuum Ultraviolet and William R. Ott

Grid Plate Calibration at the Theodore D. Doiron 41National Bureau of Standards

Phase Meter Calibration at NBS Raymond S. Turgel 53

The Application of Flame Spread J. G. Quintiere 61Theory to Predict MaterialPerformance



News Briefs and Reports

Developments

GROUND BROKEN FOR NBS COLDNEUTRON RESEARCH FACILITYGround was broken at NBS on November 20 forthe construction of one of the few cold neutronresearch facilities in the world. The new laboratorywill make cold neutron research easily available toU.S. scientists working in such fields as materialsscience, chemistry, and biology.

Typical areas of application include the design ofhigh-temperature, high-strength ceramics for en-gines; the study of improved semiconductordevices; the development of new magnetic alloysfor more efficient electric motors; the creation ofnew chemical catalysts; and the precise measure-ment of newly engineered biomolecules.

The $25-million facility will be housed in a spe-cially designed research hall to be built at the NBSresearch nuclear reactor. A recently installed "coldsource" supplies the low-energy neutrons.

Initial instrumentation will include: a high-reso-lution small-angle neutron scattering (SANS) spec-trometer, one of the fundamental tools of neutronanalysis, and an instrument which will rank withthe best of its kind in the world; the best neutrondepth-profiling instrument in the world; and twotemporary facilities, a second SANS instrumentand time-of-flight spectrometer. Ultimately, thecenter will include 15 experimental stations; 10 willbe instrumented and operated by NBS and five willbe instrumented and operated by outside groups.

These outside groups, Participating ResearchTeams (PRTs), will come from industrial, aca-demic, or government research institutions and willfully instrument and maintain their stations in re-turn for exclusive use of two-thirds of the available

research time. The high-resolution SANS spec-trometer is being constructed by a PRT consistingof NBS and Exxon Research and Engineering.

An external advisory committee will be set up tohandle the allocation of research time, which willbe made available to all U.S. users. The first exper-imental stations should become available late in1989.

For further information, contact Michael Roweat the National Bureau of Standards, Gaithersburg,MD 20899.

DETAILS NOW AVAILABLE ON SUPER-CONDUCTING CONTACTSThe high resistance that usually occurs where ex-ternal leads are attached to ceramic superconduc-tors is an obstacle to both measurements andpractical applications of the newly developed high-critical-temperature superconductors. Work byNBS and the Westinghouse Research and Devel-opment Center, Pittsburgh, Pennsylvania, has pro-duced a new method for making low-resistanceelectrical contacts on ceramic superconductors.

The researchers developed the new method afternoticing that there was a correlation between theages of many superconductor samples and theircontact resistivities. The new method involvesthree parts, which work together to produce lowcontact resistivity: minimizing the air exposuretime minimizes the degradation of the supercon-ductor surface that occurs before making the con-tacts; sputter etching the surface removes thedegraded layer; and depositing a thin layer of anoble metal-silver and gold were used-protectsthe surface and minimizes further degradation ofthe superconductor surface.

When the method was tried on old samples ex-posed to air for over 2 months, the contact resis-tivity was about 10 times higher than that for

1

Volume 93, Number 1, January-February 1988Journal of Research of the National Bureau of Standards

contacts formed on fresh samples only an hour old.A deeper sputter etch may be necessary for super-conductors having long exposure to air.

For the particular superconductor tested, bulkYBa2Cu3O7, the sample was etched to a depth of 20to 50 nanometers at 1.25 kV rms in 3-Pa argon forabout 3 minutes. A 1- to 6-micrometer-thick con-tact pad was immediately sputtered onto the sur-face (4.2 kV rms, no applied bias), also under anargon atmosphere. The superconductor tempera-ture was held to less than 1000C with a water-cooled sample holder.

For further information, contact: Jack Ekin,Electromagnetic Technology Division, NationalBureau of Standards, Boulder, CO 80303, or call303/497-5448.

NBS TESTIFIES ON SUPERCONDUCTIVITYLEGISLATIONOn October 7, NBS Deputy Director RaymondKammer testified before a joint meeting of twoHouse subcommittees considering pending legisla-tion on high-temperature superconductivity re-search.

Noting that much of the nation's expertise in su-perconductivity research lies in the national labora-tories like NBS, Kammer said, "It is our belief thatthese technical resources and those of other federallaboratories should be made available to our pri-vate industry without exerting the degree ofgovernment control that appears to be typical inmany other nations." Kammer opposed thecreation of additional levels of management tooversee the government's superconductivity re-search.

NEW ACOUSTIC TECHNIQUE PROMISESBETTER WAY TO EVALUATEINSTALLATION OF HOME THERMALINSULATIONNBS researchers are completing testing of a newmeasurement technique, based on acoustics, thatshows considerable promise for quality control inthe manufacture of insulation and could also be thebasis for efficient, low-cost tools for evaluatingloose-fill attic insulation after it is installed.

NBS researchers-fundedby the Mineral Insula-tion Manufacturers Association, the U.S. Depart-ment of Energy, and NBS-have developed ahigh-speed method for indirectly measuring eitherthe coverage or the R-value of certain types ofloose-fill insulation by passing sound wavesthrough a sample of material and measuring thesound insertion loss (i.e., the decrease in soundlevel) caused by the insulation at variousfrequencies.

In the laboratory system developed at NBS,pulses of sound containing energy in the frequencyrange of 2.5 to 25 kilohertz are projected through asample of insulating material. The samples of insu-lation are contained in individual baskets whichhave acoustically transparent bottoms and are sup-ported inside a specially designed test chamber.

The sound insertion loss of loose-fill insulationshas been found, for a given insulating material, tocorrelate quite well with the coverage (mass perunit area) of the material, particularly for fiberglassmaterials. Predicting coverage of rock wool insula-don is somewhat problematic, according to re-searchers, because of the presence of fused material(shot).

Researchers say that a version of the laboratoryapparatus used at NBS could be useful to insulationmanufacturers for quality control. The acoustic testapparatus collects and analyzes data in a few min-utes and yields results that would require severalhours to obtain using thermal measurement proce-dures. Moreover, they say, the principles could beapplied to field instrumentation for characterizinginsulation installed in attics by using acousticprobes either suspended over the insulation, or in-serted into the insulation through holes bored inthe ceiling. Two NBS researchers, Gerald V.Blessing and Daniel R. Flynn, recently weregranted a patent for various design concepts for afield instrument utilizing these ideas.

The results of the NBS experiments have beenpresented to ASTM Committee C16 on ThermalInsulation, which is interested in the acoustic tech-nique as a possible ASTM test method. In addition,ASTM has agreed to coordinate an effort to raisethe funds needed to develop a practical field instru-ment based on acoustic measurements of installedloose-fill insulation.

For further information, contact GeraldBlessing, A147 Sound Building, National Bureau ofStandards, Gaithersburg, MD 20899, or call 301/975-6627.

NBS REPORTS TO CONGRESS ONSTRUCTURAL DAMAGE IN RECENTCALIFORNIA QUAKEThe biggest share of the structural damage fromthe October earthquake in California was sustainedby older, unreinforced masonry buildings that werenot designed to absorb an earthquake's energy orwere not connected firmly to their foundations.That assessment was made by John W. Lyons, di-rector of the NBS National Engineering Labora-tory in testimony on November 10 before the U.S.House Subcommittee on Science, Research andTechnology.

2



Most of these observations reflect lessons al-ready learned from previous quakes and research,said Lyons, but they are important because in manyearthquake-susceptible parts of the country suchstructures are still being built.

In a report released at the hearings, H. S. Lew,head of the NBS structural evaluation program,also noted that light fixtures and suspended ceilingswhich were not properly secured fell during theearthquake. In addition, he said, walls with largewindows provided little stiffness to the structureand, in many cases, window glass fractured intolarge pieces and fell on sidewalks.

Lew, who was in California when the quakestruck, found only one highway bridge that sus-tained significant damage. Most bridges had beenrepaired or strengthened to increase their earth-quake resistance after the 1971 San Fernandoearthquake in which five bridges collapsed and 42others were damaged significantly.

As part of the National Earthquake Hazards Re-duction Program, NBS conducts research and pro-vides technical support to the private sector andgovernment agencies which are working to im-prove the performance of buildings and otherstructures subjected to earthquakes.

Last month, NBS researchers completed a seriesof laboratory tests on full-scale, reinforced con-crete bridge columns under simulated earthquakeconditions. This work, partially sponsored by theCalifornia Department of Transportation, will helpdetermine the effectiveness of current Californiadesign criteria. It also will be used to determinewhether small-scale tests and computer models canaccurately predict seismic performance of full-scale bridge structures.

NBS also is conducting laboratory studies on theperformance of reinforced concrete masonry underearthquake loadings. In addition, Lew and his staffare conducting research on strengthening methodsfor reinforced concrete structures.

NBS does not issue or enforce standards or reg-ulations. The results of Bureau research will go togroups that do set and enforce such standards andregulations and will provide the basis for safer,more economical design practices.

Lyons said that although this earthquake wasmodest, it cost three lives directly and more than$200 million in property damages. "It is a warningof the importance of the earthquake threat to theUnited States," he concluded.

For further information, contact H. S. Lew atthe National Bureau of Standards, Gaithersburg,MD 20899.

NBS/GCA JOINT PROGRAM ON"MOLECULAR MEASUREMENT MACHINE"NBS and GCA Corporation of Andover, Massa-chusetts, have begun a joint research program insupport of an NBS project to develop a prototypemeasuring machine that will be capable of measur-ing surface features over an area approximately 5centimeters square with a resolution of a singleatom. The "Molecular Measurement Machine,"under design by the NBS, will make use of ex-tremely high-resolution probes, such as a tunnelingmicroscope probe or atomic force probe, and willbe able to make surface roughness measurementsaccurate to a tenth of an angstrom. Stanley Stoneof GCA will work with NBS on the design anddevelopment of the machine structure, slideways,motor drives, interferometers, and associated con-trol system needed for this new generation of mea-surement machines.

For further information, contact Clayton Teagueat the National Bureau of Standards, Gaithersburg,MD 20899.

POLYMER COMPOSITE PROCESSINGRESEARCH NEEDS IDENTIFIEDTwo dozen companies met at NBS October 7 toaddress processing methods and barriers for poly-mer composites. These composites now are used indefense and specialty products, but other marketsare expected to open up if processing methods canbe improved.

The session at NBS, which was cochaired byrepresentatives from Ford and NBS, identified avariety of scientific and engineering needs for im-proved processing methods which are expected tobe of most interest to industry during the next 10 to15 years. This information will be used to developspecific projects at NBS to aid industry in develop-ing and applying these innovative materials.

NBS plans to pursue cooperative projects withindustry where appropriate. A report on the work-shop will be available in several months.

For further information, contact DonaldHunston at the National Bureau of Standards,Gaithersburg, MD 20899.

NBS PROPOSES FEDERAL STANDARDFOR OSINBS is asking for comments on a proposed FederalInformation Processing Standard (FIPS) whichadopts the Government Open Systems Intercon-nection Profile (GOSIP). (FIPS are developed byNBS for use by the Federal government.)

GOSIP defines a common set of data communi-cation protocols which enables computer systems

3



developed by different vendors to communicateand allows the users of different applications onthese systems to exchange information. It was de-veloped by the Government OSI Users Commit-tee, a group of Federal government agencies.GOSIP adopts international standards which en-able information processing devices to communi-cate with one another in a network. GOSIP isbased on agreements reached by vendors and com-puter users in NBS-sponsored workshops.

For a copy, contact Standards Processing Coor-dinator (ADP), Institute for Computer Sciencesand Technology, B64 Technology Building, Na-tional Bureau of Standards, Gaithersburg, MD20899, or call 301/975-2816.

Comments on the proposed Federal standardshould be sent to Director, Institute for ComputerSciences and Technology, Attn GOSIP, B154Technology Building, at the above address.

For technical information, contact Gerard F.Mulvenna, B217 Technology Building, NationalBureau of Standards, Gaithersburg, MD 20899, orcall 301/975-3631.

MEASURING FAULTS IN PARALLELPROCESSORSIn a recent study for the Defense Advanced Re-search Projects Agency, NBS researchers investi-gated a number of measurement techniques thatcan be used to detect and recover from hardwarefaults in parallel processors.

"Fault-tolerance" is particularly important incomputer systems used, for example, to solve largeproblems such as weather forecasting and to con-trol aircraft and spacecraft. The researchers lookedat ways to detect transmission and data storage er-rors and faults in processors, controllers, computercomponents, and input/output systems.

A report, On the Measurement of Fault-TolerantParallel Processors (NBSIR 87-3568), is availablefrom the National Technical Information Service,Springfield, VA 22161 for $11.95 prepaid. Orderby PB #87-208328/AS.

This work is part of an ongoing research effortto develop techniques and tools to measure the per-formance of parallel processors.

RESEARCH DETECTS NUCLEAR REACTORBURN-UP OF USSR CRAFTA study that examines the 1983 burn-up of a nu-clear reactor from the Soviet reconnaissance satel-lite Cosmos-1402 is the subject of a just-publishedreport. The satellite reactor, which normallywould have been boosted into a long-lived orbit,fell into the Earth's atmosphere and disintegratedbecause of a malfunction.

To examine this event more closely, researchersfrom the Department of Energy's EnvironmentalMeasurements Laboratory (EML) collaboratedwith NBS scientists to identify reactor particles byanalyzing stratospheric filter samples for their ratioof the radioactive isotopes uranium 235 and 238.The project was undertaken because officials fromEML-which investigates radioactive injectionsinto the stratosphere-suspected that the reactorburn-up had occurred but had no direct evidence.

The team analyzed atmospheric samples in theNBS inorganic mass spectrometry laboratory, acustom facility well-suited for determining isotopicratios. The samples, which were collected by spe-cial balloons at altitudes of 16 to 23 miles (27 to 39kilometers), indicated large excesses of U-235. Theresearchers say this appears to indicate that the re-actor did burn up in the stratosphere and that ra-dioactive particles from the event will continue tospread out and dilute, posing no danger on Earth.

The research team's report is in the October 23,1987, issue of Science.

NBS REPORTS CAUSE OF FATAL BUILDINGCOLLAPSEThe failure of a key component of the system forlifting concrete floor slabs in the L'Ambiance Plazaapartment building was the most probable causefor the building's April 23, 1987, collapse, saidCharles Culver, chief of the NBS Structures Divi-sion, at a press briefing on October 22.

Specifically, the collapse probably began when ajack rod supporting three concrete floor slabsslipped out of a U-shaped opening in a steelbracket. NBS researchers pinpointed the locationwhere the jacking system used to lift the slabs intofinal position most likely failed. They investigateda variety of possible causes of failure, but con-cluded that only the lifting system failure played asignificant role in initiating the collapse in which 28construction workers died.

NBS was asked by the U.S. Occupational Safetyand Health Administration to determine the causeof the collapse.

For information on ordering the NBS report, In-vestigation of LAmbiance Plaza Building Collapse inBridgeport, Connecticut (NBSIR 87-3640), contactthe Structures Division, B268 Building ResearchBuilding, National Bureau of Standards, Gaithers-burg, MD 20899, or call 301/975-6048.

GOOD ATRIUM DESIGN CAN SAVE ENERGYAtrium spaces using large amounts of glass to let indaylight are aesthetic. But are they energy effi-cient?

4



Using computer-simulation to evaluate atriumdesign in eight areas of the United States, NBS re-searchers found that savings ranging from aboutfive percent in the Madison, Wisconsin, area up to20 percent in Los Angeles are possible. However,total energy usage for areas like Madison weregreater, so the actual dollar savings may be similar.

The researchers looked at several atrium designsand the effects of a variety of options such as solarshading, heat storage, and single- and double-paneglazings. The best-performing options seem to bethe double-pane glazing combined with heat stor-age, but double-pane glazing alone also did well.Single-pane glazing is not effective, particularly incold climates, and, surprisingly, say the re-searchers, the solar shading was notbeneficial.

A report, Fenestration Design for Building Atria(NBSIR 87-3594), can be ordered for $13.95 pre-paid from the National Technical Information Ser-vice, Springfield, VA 22161. Order by title andPB #87-171427.

NEW WAY TO TEST SETTING TIME,STRENGTH OF CONCRETENBS researchers are experimenting with a new testmethod for measuring the setting time and strengthof concrete. The technique, called "impact-echo,"uses sound waves and originally was developed byNBS researchers to detect flaws in concrete.

Currently, no standard field or laboratory test isavailable for measuring the setting time of concretewhich may not accurately reflect environmentalconditions at the site.

With the impact-echo method, the increase inthe speed of sound through concrete is measuredand correlated with setting time and subsequentstrength development of the concrete.

A report on the method, Measurement of the Set-ting Time and Strength of Concrete by the Impact-Echo Method (NBSIR 87-3575), is available fromthe National Technical Information Service,Springfield, VA 22161, for $18.95 prepaid. Orderby PB #88-111851.

Standard Reference Materials

IMPROVING N-PAH MEASUREMENTS ISAIM OF NEW MATERIALAnalytical laboratories that measure environmentalsamples for their concentration of certain nitrated

polycyclic aromatic hydrocarbons (N-PAHs) willfind a new NBS standard primarily for use in cali-brating the chromatographic instruments used tomeasure these compounds. The new SRM, besidesfunctioning as a calibration tool, also can be used to"spike" laboratory samples with known amounts ofN-PAHs for research purposes.

The SRM has certified concentrations of four N-PAH compounds-1-nitropyrene, 1,3-dinitropy-rene, 1,6-dinitropyrene, and 1,8-dinitropyrene-ina solution of methylene chloride. Packaged in a kitcontaining five vials of the certified solution, thenew material, Dinitropyrene Isomers and 1-Ni-tropyrene in Methylene Chloride (SRM 1596), isavailable for $281.

NBS, E. FJELD CO. PROGRAM ONSUBMICROMETER LINE STANDARDSThe E. Fjeld Company of North Billerica, Massa-chusetts, a manufacturer of custom apparatus forscanning electron microscopes (SEMs), has starteda joint program with the NBS to aid in the devel-opment of the next generation of NBS mea-surement standards for integrated circuit (IC) man-ufacturers.

E. Fjeld will work with NBS on combining adiamond turning instrument with an SEM for in-process viewing of machining and on the designand development of a special SEM stage linked toa laser interferometer. These instruments will beused for the precision fabrication and measurementof typical IC structures with dimensions below onemicrometer on opaque substrates. Such structuresmay be the basis for a new series of NBS standardreference materials for the measurement of sub-micrometer IC structures.

NBS Services

NBS, NSA SETTING UP RISK MANAGEMENTLABORATORYNBS and the National Security Agency's NationalComputer Security Center are planning to establisha risk management research laboratory at NBSheadquarters in Gaithersburg, MD. Risk manage-ment involves analyzing information assets, threats,and vulnerabilities; determining a measure of risk;then selecting cost-effective safeguards for reduc-ing that risk.

The laboratory will be used to conduct researchand provide the tools, techniques, and guidanceneeded to conduct this process. Other planned uses

5



include helping Federal agencies select and usecommercial risk management software and provid-ing a clearinghouse for information on risk analysisand management for the Federal government.

For further information, contact DennisSteinnuer, B266 Technology Building, NationalBureau of Standards, Gaithersburg, MD 20899, orcall 301/975-3357.

NBS ANNOUNCES ACCREDITATIONPROGRAM FOR LABORATORIESTESTING ASBESTOS IN SCHOOLSAn accreditation program for laboratories that testfor asbestos in schools has been established by NBSunder the National Voluntary Laboratory Accredi-tation Program (NVLAP) to meet the require-ments of the Asbestos Hazard EmergencyResponse Act of 1986. The law requires the Envi-ronmental Protection Agency (EPA) to develop aprogram for the inspection, management, andabatement of asbestos in the nation's schools.

Under the law, NBS will accredit laboratoriesthat perform analysis for asbestos content in sam-ples of bulk insulation and building materials, andanalyze airborne particulates collected duringschool inspections and asbestos abatement projects.The evaluation of laboratories that apply for asbes-tos bulk testing will commence about October 1988.The first group of laboratories to be evaluated foraccreditation will be those applying by December28, 1987. Those applying after that date will beevaluated on a first-come first-served basis.

Accreditation for testing for airborne asbestoswill begin in mid-1989.

For information on the new asbestos program,contact: Harvey W. Berger, Manager, NationalVoluntary Laboratory Accreditation Program,A531 Administration Building, National Bureau ofStandards, Gaithersburg, MD 20899, or call301/975-4016.

6

Journal ofVolume 93, Number 1, January-February 1988

Research of the National Bureau of Standards

The NBS Scale of Spectral Irradiance

Volume 93 Number I January-February 1988

James H. Walker,Robert D. Saunders,John K. Jackson, andDonald A. MeSparron

National Bureau of StandardsGaithersburg, MD 20899

This paper describes the measurementmethods and the instrumentation used inthe realization and transfer of the NBSscale of spectral irradiance. The basicmeasurement equation for the irradiancerealization is derived. The spectral re-sponsivity function, linearity of re-sponse, and "size of source" effect ofthe spectroradiometer are described.The analysis of sources of error and theestimates of uncertainty are described.

The assigned uncertainties (3c, level) inspectral irradiance range from 2.2% at250 nm to 1.0% at 654.6 nm to 6.5% at2400 nm.

Key words: blackbody; calibrations; ra-diometry; response linearity; slit-scatter-ing function; spectral irradiance;standards.

Accepted: September 10, 1987

1. Introduction

Spectral irradiance, denoted Ex, is defined as theradiant flux of wavelength X incident on a surfaceper unit wavelength interval and per unit area onthe surface. Mathematically

Ex=d 2 .b/dX-dA, (1)

where d2 4' is the element of incident flux and dXand dA are the elements of wavelength and arearespectively.

The National Bureau of Standards (NBS)presently issues two types of spectral irradiancestandards. Type FEL (ANSI designation) lamps,modified to a medium bipost base, are calibrated asstandards of spectral irradiance at 31 wavelengthsover the spectral range 250 to 2400 nm. Deuteriumlamp standards of spectral irradiance are calibratedat 16 wavelengths over the spectral range 200 to350 nm and at a lower accuracy than the type FELlamps. Both these lamp standards are designated inNBS Special Publication 250 [1].

In 1963, NBS established a scale of spectral irra-diance [2]. In the early 1970's an improved scalewas developed [3] with uncertainties about 1/3those of the earlier scale. The detailed techniquesfor realizing this scale have undergone several evo-lutionary changes in the past decade. This paper isa description of the current process of realizationof the NBS spectral irradiance scale and of the cur-rent procedures for the routine spectral irradiancecalibrations.

Modified type FEL lamps are routinely cali-brated from 250 to 2400 nm. Deuterium lamps areroutinely calibrated from 200 to 350 nm. The spec-tral irradiance values transferred to the deuteriumlamps in the spectral range 200 to 250 nm are basedon the hydrogen and blackbody line arcs devel-oped primarily for use in the vacuum ultraviolet[4]. From 250 to 350 nm the reported spectral irra-diance values are transferred from the modifiedtype FEL lamps. The equipment used for the deu-terium lamp calibrations is identical to that used forthe modified type FEL lamp calibrations, and the

7



measurement procedures are very similar. Thebody of this paper will be limited to a descriptionof the calibration of the modified type FEL lamps.Detailed information on the calibration services ap-pears in a separate document [5].

2. Scale Derivation and Transfer

The NBS scale of spectral irradiance is derivedfrom the NBS scale of spectral radiance [6] whichis based on a realization of the International Practi-cal Temperature Scale (IPTS-68) starting with agold-point blackbody [7]. The average spectral ra-diance over the exit aperture of a special integrat-ing sphere source is determined and then the fluxfrom the sphere source which enters the receivingaperture of the spectroradiometer is calculated.This technique is used to determine the spectralirradiance at the detector receiving aperture andthus establishes a spectral irradiance scale. As amatter of convenience the scale is transferred to agroup of four, 1000 W, quartz-halogen lamp pri-mary working standards using an averagingsphere-monochromator combination designed forspectral irradiance measurements. These lamps areused to maintain the NBS scale of spectral irradi-ance. The lamps are recalibrated every 50 to 100burning hours.

Figure I shows the setup used to measure thespectral radiance of the special integrating spheresource. Figure 2 shows the setup used to transferthe spectral irradiance scale to a group of primaryworking standards.

The geometry used for the spectral irradiancedetermination is shown in figure 3. The followingmethod is used to determine the spectral irradi-ance at the receiving aperture of the spectrora-diometer produced by the integrating spheresource. The spectral irradiance, Ex, at the receivingaperture due to the spectral radiance, L, at anypoint on the source aperture is

E f=rLAd, (2)

where cX is the solid angle defined by the receivingaperture and a point on the source aperture. Tocalculate the flux at the receiving aperture due tothe entire source aperture, it is necessary to inte-grate over the entire projected area of the sourceaperture

(= f4 f L,-do-dA, (3)

where dA =dx-dy-cos0. LX is a function of 0,+, x,and y so that

(DX= tfA frL,(0, 4, x, y) .cos'.dco.dx.dy, (4)

where:

0 is the angle between the normal to the surfacesof the apertures and a line connecting a single pointon each aperture,

4r is the azimuthal angle,x is the horizontal location of a point on the

source aperture,y is the vertical location of a point on the source

aperture.

Figure 1. Spectral radiance measurement setup.

8



Figure 2. Spectral irradiance measurement setup.

Sou roeAperture Receiving

Aperture

radius = r, radius = r,

Figure 3. Irradiance calculation geometry.

Assuming that the source is Lambertian and nearlyuniform, L(0,A,, x, y) can be replaced by an aver-age radiance L4 to give

$'x=4.A fcos0.dco-dx-dy, (5)

where:

dd =cos0/r 2 -dx'dy',x',y' is a point on the receiving aperture,r is the slant distance from x', y' to x, y.

This gives

'Ik=L4- f IA cos2 9/r-dx '-dy'-dx -dy,

where:

ASA is the area of the source aperture,ARA is the area of the receiving aperture.

For circular, coaxial source and receiving aptures, this integral evaluates to

,=L,-7T /2-[R 2_ (R 4_ 4r,2.r2 2)1/2]

(6)

where:

R2 =d2 +r1 2 +r2 ,d is the normal distance between source and re-

ceiving apertures,r, is the radius of the source aperture,r2 is the radius of the receiving aperture.

A more convenient expression is

where 8=(r 2.r 22 )/R 4 . Finally,

The final step is to compare the spectroradiome-ter outputs produced by the integrating spheresource and each working standard.

Once the primary working standards have beencalibrated, they are used to measure the spectralirradiance of test lamps. Modified type FEL testlamps are calibrated in groups of twelve.

3. Measurement Apparatus

Spectral radiance and spectral irradiance calibra-tions are performed on the NBS Facility for Auto-mated Spectroradiometric Calibrations (FASCAL)[8]. Block diagrams of the measurement apparatusare shown in figures 1 and 2. The principal compo-nents are:

1. Variable-Temperature Blackbody(7) 2. Sources

9

t>2=Lx'(iTrjr2 )-(7r-r22)/R 2-[I +8+2.82+ 5-83+...], (8)

Ex=4)x/ARA=LX-(TrIr, 2)/R 2.Vo +8+ . (8a)

Volume 93, Number

Journal of Research of the1, January-February 1988

National Bureau of Standards

Pyrometer LampSpecial Integrating Sphere SourceSpectral Irradiance Primary Working Stan-dardsTest Lamps

3. SpectroradiometerFore-optics

Avenging SphereMirrors and Entrance Slit Masks

MonochromatorDetectors

4. Control and Data Acquisition System

3.1 Variable-Temperature Blackbody

The variable-temperature blackbody is used overa temperature range from about 800 'C to about2400 'C.

A schematic cross section of the variable-tem-perature blackbody is shown in figure 4. The black-body cavity is located in the central portion of ahigh density graphite tube, which is resistivelyheated in an argon atmosphere. Electric current issupplied to the graphite tube through water-cooledelectrical connections at each end of the tube. Thetube is surrounded by a double-walled graphite ra-diation shield, with carbon black fill between thewalls. This assembly is surrounded by a water-cooled metal housing, with an observation portwhich can be sealed during evacuation of the atmo-sphere within the housing prior to flushing withargon. A window is provided at the top of thehousing for visual pyrometer observation of thetemperatures along the tube interior. A secondwindow at the rear of the housing allows radiationfrom the rear wall of the graphite tube to fall on asilicon photodiode. The photodiode provides a sig-nal for automatic control of the saturable-reactorpower supply for the tube. A germanium photodi-ode, whose response extends further into the in-frared region, replaces the silicon cell for operationat temperatures below 1000 'C. The blackbodymounting provides adjustment in two angular andthree translational degrees of freedom, allowing forprecise positioning and radiometric scanning overthe target area and the beam solid angle.

The graphite tube is about 200 mm long, with aninner diameter of about 11 mm. The outer surfaceis tapered to improve temperature uniformity alongits length. The wall is about 4 mm thick at mid-length where a 2 mm diameter hole in the wallallows for observation of the emitted flux. The tubeis partitioned into small cylindrical sections by aseries of thin graphite disks separated by thingraphite cylinders located at intervals along the

bore. Holes in the graphite disks permit measure-ment of the temperatures in the middle and uppersections with a visual pyrometer. The holes vary indiameter from 6 mm for the uppermost disk to 0.75mm for the disk below the central section. The cen-tral cylindrical section, which provides the ob-served flux, is 9 mm high and 10 mm in diameter.The inside wall is threaded to reduce its partial re-flectivity [9,10]. Figure 5 shows a cross-sectionalview of the central section.

PYROMETER

DETECTOR

, GRADIENT OBSERVATIONWNDOW

X C-BARBON-BLACK

REMOVABLEWNDOW

GRAPHITE TJBE

Figure 4. Variable-temperature blackbody schematic.

The blackbody emissivity has been assessed bymeasurements of the solid angle subtended by thecavity opening, the partial reflectivity of thegraphite material [9], the temperature gradients,and the absorption by gases [10]. The solid anglesubtended at the rear wall of the cavity by the in-ner edge of the observation hole is about 0.03 sr.The measured partial reflectivity of the graphite is0.02 sr'. The measured temperature gradient overthe length of the viewing cavity is less than I K.Experimental investigations of possible absorptionof radiation by gases has disclosed only weak ab-sorption lines at 589 and 589.6 nm (Na) and at 766.5nm (K). The resulting estimate of emissivity is0.9990±0.0005.

3.2 Sources

3.2.1 Pyrometer Lamp This lamp is used as a sec-ondary standard for realizing the spectral radiance

10



scale. It is a highly stable vacuum tungsten-striplamp which is operated at a single current to pro-duce a spectral radiance of about eight times that ofa gold-point blackbody at 654.6 nm (about 1530 Kradiance temperature). The lamp drift rate is lessthan 0.02% per 100 hours when operated at a sin-gle current level.

Figure 5. Central section of variable-temperature blackbody.

3.2.2 Special Integrating Sphere Source Thissource has been specially constructed to be unpo-larized and to have high output in the IR part of itsspectrum. It consists of a heat-sinked, water-cooledintegrating sphere with a 1000 W quartz-halogen,modified type FEL lamp mounted next to the en-trance port. The integrating sphere is 5.0 cm in di-ameter with a 23 mm diameter entrance port and a20 mm diameter exit port located about 100' fromthe entrance port. The inside sphere wall is coatedwith pressed high purity polytetrafluoroethylene(PTFE) [18] to give high reflectivity in the IR. Amodified type FEL lamp is mounted with its envel-ope about 3 mm from the entrance port of thesphere and located so that it does not directly irra-diate the inside wall of the sphere opposite the exitport. The sphere itself is made of copper and ismounted in a heat-sinking copper plate. Coppertubing is soldered to the sphere and plate and theentire assembly is water-cooled to prevent thesource from overheating. A precision circularaperture whose area has been accurately measuredis attached at the exit port.

Because of the multiple reflections in the inte-grating sphere, entering radiation is randomized,producing a uniform, depolarized radiant flux atthe exit port. The uniformity is verified when theexit port aperture is mapped during the irradiancerealization procedure (fig. 6 shows a typical map-ping profile). Depolarization was tested at 650 nmusing an unpolarized source and a linear polarizerand found to be complete within the measurementprecision of 0.1% (3a- level).3.2.3 Spectral Irradiance Primary Working Stan-dards Four 1000 W quartz-halogen, modifiedtype FEL lamps were selected as primary workingstandards. This type lamp has a clear bulb and atungsten coiled-coil filament (CC-8) and has a ratedlife of 500 hours at 120 V. Before calibration, thelamp base is converted to a medium bipost base andthe base structure is encapsulated in an epoxy-ce-ramic compound. The posts that form the bipostbase are 6.35 mm (1/4 in) diameter cylindricalstainless steel rods that extend 20.64 mm (13/16 in)from the bottom of the epoxy-ceramic block. Theposts are spaced 22.23 mm (7/8 in) between cen-ters. A metal plate bearing the lamp identificationnumber and indicating the electrical polarity is at-tached to the rear surface (side away from thespectroradiometer) of the epoxy-ceramic block.3.2.4 Test Lamps The test lamps are also modi-fied type FEL lamps. A lamp screening process isused to select test lamps suitable for calibration.Lamps are annealed and then are checked for sta-bility, emission lines or absorption bands, and forvariations in goniometric output.

3.3 Spectroradiometer3.3.1 Fore-optics

Averaging Sphere The averaging sphere is 2.5 cmin diameter with a I cm2 area precision circularentrance port and a 3X 12 mm exit port locatedabout 100' from the entrance port. The insidesphere wall is coated with pressed high purityPTFE. This material has been found to fluoresce atcertain wavelengths under certain conditions [19],but when the sources being compared have ap-proximately the same spectral distribution, fluores-cence is not a problem. The radiation entering thesphere is randomized by multiple reflections in thesphere, thus producing uniform, depolarized radi-ant flux at the exit port. This uniformity was veri-fied to within the measurement precision of 0.1%(3a- level) by radiometrically scanning the exit portof the sphere. Depolarization was tested at 500 nmand 650 nm using an unpolarized source and a lin-ear polarizer and found to be complete within themeasurement precision of 0.1% (3cr level).

11



0.22 0.30 0.31

0.28 0.20 0.23 0.21 0.20 0.31 0.30 0.31 0.44

0.45 0.25 0.14 0.03 0.15 0.21 0.26 0.36 0.43 0.48 0.46

0.42 0.21 0.00 0.03 0.06 0.09 0.15 0.20 0.31 0.42 0.40 0.53 0.61

-0.10 0.25 0.06 -0.04 -0.04 0.03 0.12 0.16 0.10 0.24 0.37 0.43 0.55 0.67 0.75

0.32 0.11 -0.05 -0.11 -0.07 0.04 0.11 0.13 0.14 0.21 0.30 0.40 0.53 0.67 0.79

-0.44 0.21 -0.00 -0.15 -0.10 -0.13 -0.02 0.05 0.10 0.11 0.15 0.22 0.33 0.4? 0.66 0.20 0.86

0.46 0.15 -0.07 -0.21 -0.22 -0.14 -0.05 -0.01 0.06 0.00 0.03 0.10 0.25 0.42 0.61 0.00 0.80

0.40 0.00 -0.12 -0.28 -0.26 -0.21 -0.14 -0.00 0.00 0.05 0.00 0.13 0.24 0.3B 0.60 0.24 0.95

-0.20 -0.29 -0.24 -0.10 -0.12 -0.05 -0.00 0.06 0.14 0.26

-0.32 -0.34 -0.31 -0.25 -0.19 -0.14 -0.00 -0.00 0.12 0.28

0.44 0.67 0.93 I.to

0.46 0.68 0.94 1.09

0.12 -0.03 -0.36 -0.42 -0.40 -0.35 -0.30 -0.22 -0.16 -0.00 0.15 0.30 0.47 0.67 0.93

0.07 -0.12 -0.34 -0.43 -0.43 -0.37 -0.33 -0.27 -0.20 -0.02 0.15 0.32 0.50 0.69 0.90

-0.14 -0.35 -0.45 -0.47 -0.42 -0.37 -0.30 -0.20 -0.03 0.16 0.36 0.52 0.60

-0.29 -0.40 -0.44 -0.40 -0.34 -0.27 -0.16 0.00 0.17 0.35 0.50

-0.41 -0.41 -0.30 -0.29 -0.21 -0.10 0.05 0.23 0.29

-0.27 -0.19 -0.06

Values are % difference from central value

Wavelength= 654.6 rm

Target area=0.6 nm wide by 0.8 nm high

X Increment= 1.06 nm

Y Increment= 1.06 nm

Overall mapping correction = +0.14%

Figure 6. Mapping profile of integrating sphere aperture.

The difference in the solid angle of irradiationfor the irradiance lamp and the integrating spheresource is only a problem when the reflectance ofthe averaging sphere wall is not uniform. This highpurity, 3 mm thick PTFE sphere coating providesthis uniformity, and it was verified by determiningthe spectral irradiance of a lamp mirror-system [3]whose solid angle could be varied. Negligible dif-ference (



adjustable together as a unit from 0.01 to 3.0 mm,resulting in a nearly triangular-shaped spectralbandpass.3.3.3 Detectors Two interchangeable detectorsare used to cover the wavelength range of thespectroradiometer. For the 200 to 850 nm range, anend-on 11-stage photomultiplier with quartz win-dow and S-20 spectral response is placed behindthe exit slit. The detector is cooled to 258 K with athermoelectric cooler. The anode current is ampli-fied and converted to a 0 to 10 V signal by a pro-grammable dc amplifier. To ensure linearity ofresponse, the high voltage applied to the detector isnormally selected to restrict the detector current to500 nA or less.

A lead sulfide detector, cooled to 240 K by athermoelectric cooler, is used for the 800 to 2400nm range. The detector and the exit slit are placedat the foci of an ellipsoidal mirror, which imagesthe exit slit upon the detector with a demagnifica-tion of about 7. The detector output is amplifiedand converted to a 0 to I V signal by a phase-sensi-tive lock-in voltmeter, which is keyed to a 78 Hzsector disk placed just before the plane mirror inthe radiance mode or just after the exit port of theaveraging sphere in the irradiance mode.

The signal from either detector-amplifier combi-nation is fed to a 5 1/2 digit voltmeter, capable ofintegration times ranging from one second to sev-eral minutes. To facilitate alignment of optics orsources, a HeNe laser is placed at the detector posi-tion, so that its beam passes through the monochro-mator and fore-optics in the reverse direction.

3.4 Control and Data Acquisition System

After initial alignment, the FASCAL systempermits control of the entire measurement processfrom a remote operator console. Component posi-tions, instrument settings, sequence of operations,and data collection are effected by either storedcomputer programs, operator commands, or acombination of the two.

The system is directed by a microcomputerequipped with a CRT terminal and keyboard and ahigh-speed disk system for program and data stor-age. A modular interface controller [12] providesthe link between instruments and computer. Allmeasurement signals are multiplexed into the digi-tal voltmeter through the interface scanner, and theinstruments are remotely programmed and con-trolled through interface modules. All instrumentsettings and signal outputs are printed and storedon disk for later analysis.

The spectroradiometer (fore-optics, monochro-mator, and detectors), a closed-circuit TV camera,and a photoelectric pyrometer are mounted on acarriage. The carriage can be moved by remotecommand along a linear track, to position the spec-troradiometer in front of any of the sourcesmounted at fixed stations along the track. The av-erage move time between stations is a few seconds,and positions are repeatable to about 0. 1 mm. TheTV camera presents a highly magnified image ofthe monochromator entrance slit mask to video dis-plays at the spectroradiometer and at the operatorconsole for initial source alignment and subsequentmonitoring. The pyrometer is used for the initialsetting of the variable-temperature blackbody to itsapproximate temperature.

4. Measurement of Instrument and SourceParameters

4.1 Spectral Responsivity Function

The relative spectral responsivity function of thespectroradiometer is determined by an indirectmethod [13]. In this method, the relative responsiv-ity function is treated as the product of two terms,the responsivity factor and the slit-scattering func-tion, where the responsivity factor depends onlyupon the wavelength of the observed flux and theslit-scattering function depends only upon the dif-ference between the wavelength setting of themonochromator and the wavelength of the flux.This factorization of the spectral responsivity func-tion is valid if the instrument dispersion, aberra-tions, scattering, and diffraction are constant overthe wavelength region of interest. This assumptionis valid in the central portion of the relative respon-sivity function, but values for the distant wings aresubject to error due primarily to changes in scatter-ing and dispersion.

The responsivity factor is obtained by spectrallyscanning a continuous source standard of spectralradiance using narrow (0. I mm) slits. To determinethe slit-scattering function, an integrating sphere ir-radiated by a high-powered laser is spectrallyscanned by the spectroradiometer, with the slitwidths set at the 0.6 mm width used in the scalerealization and transfer. The plot of the output sig-nal versus wavelength is the mirror image of theplot of the slit-scattering function versus wave-length. For a 647 nm Kr laser, the function isnearly triangular in shape with a width at half-height of 2.5 nm. Relative to the peak value, themeasured values decrease to about IO-3 at 3 nm,10-4 at 15 rm, and tO- 7 at 70 mm from the central

13



wavelength. At 150 nm from the central wave-length, the value decreases to 10'- in theshort-wavelength wing and to 10-' in the long-wavelength wing. Scans with 488 nm (Ar), 514 nm(Ar), and 676 nm (Kr) yield similar results. Thesevalues were confirmed over the central and nearwing portions of the function by measurementswith the direct method, using a dye laser tunedthrough a series of wavelengths with the spectrora-diometer set at a fixed wavelength [14].

The measurement at 647 nm yielded the split-scattering function used for 654.6 nm, where thespectral distribution mismatch of a variable-tem-perature blackbody and a gold-point blackbody re-quires an accurate determination of the relativeresponsivity function. However, the measurementsin the visible cannot be applied with confidence tothe short-wavelength region, since the dispersionvaries by about a factor of 2.5. For this region, thecentral portion and near wings of the slit-scatteringfunction are determined by scans of a spectral linedischarge source, and values in the distantlong-wavelength wing are deduced from a mea-surement of the integrated spectrally-scattered ra-diation. With the wavelength set at a selected valuein the 200 to 250 nm region, the signal from a cali-brated lamp (radiance temperature 2475 K at 654.6nm) is recorded. A glass filter which blocks all ra-diation in the vicinity of the wavelength settingand passes about 90% of the radiation at longerwavelengths is inserted into the beam. The ratio ofsignals with and without filter is taken as the frac-tional contribution of spectrally scattered radiationto the signal. A second (identical) filter is added toinsure that only scattered light is being observed inthe filtered beam. Results with filters of differentcutoff wavelengths (Corning filters CS 0-56 andCS 0-52) both indicate an integrated scattered lightcontribution of less than 0.2% at 225 nm. The slitscattering function calculated from this result andthe known source distributions and responsivityfactor are less than 10'- at wavelengths greaterthan 200 nm from the central wavelength, in goodagreement with the values measured in the visible.

4.2 Linearity of Response

The degree of linearity of the spectroradiometerresponse is determined with an automated beamconjoiner [15,16]. A beam from a constant source issplit into two branches whose fluxes are indepen-dently attenuated or blocked before recombinationand further attenuation. The flux from bothbranches measured together should equal the sum

of the fluxes from each branch when measured sep-arately (additivity). The device provides 96 levelsof flux ranging over a factor of about 500. Thelevels are presented in random order to avoid sys-tematic errors and are interspersed with 29 zeroflux levels. A microcomputer controls the attenuat-ing filters and records the filter positions and ra-diometer signals. The data is least-squares fitted toa polynomial response function to determine a cor-rection factor by which the radiometer output sig-nal must be multiplied to obtain a quantityproportional to radiant flux.

The response function of the spectroradiometeris dependent upon the detector-amplifier em-ployed. With the photomultiplier tube in place(spectral range 200 to 850 nm), the instrument re-sponse at all wavelengths is linear to within 0.2%for a range of anode currents from 1 to 500 nA.Linearity measurements were performed at 900,600, 300, and 250 nm. For currents much less thanI nA, the signal is limited by noise. For currentsgreater than 1 ±A the correction increases rapidly,rising to 3% at 7 xtA. The anode current is re-stricted to less than 500 nA during measurementsby selection of appropriate photomultiplier tubevoltage. Correction factors for the amplifier rangesare determined from the measurement of a knownelectrical current and combined with the linearitycorrection factor.

Linearity tests of two PbS detectors resulted in acorrection factor which is a linear function of thesignal over the range I to 280 mV. The correctionvaries from 0.1% at 3 mV to about 9% at 300 mV.To avoid relying on large corrections, sources aretypically operated at near equality in the PbS spec-tral region.

4.3 Size of Source

The "size of source" effect (signal contributiondue to flux which originates outside the target areaand is scattered into the measured beam by thefore-optics) is determined by observing the changein signal from a 0.6 by 0.8 mm area of a uniformdiffuse source while placing various size masks onthe diffuse source. The masks expose source areaswhich closely approximate the radiant areas of thelamp, the blackbody and the integrating spheresource used in the scale realization. As a check, theeffect is also evaluated by observing changes in thenear-zero signal from a "black hole" (an absorbingcavity slightly larger than the 0.6 by 0.8 mm fieldstop) as the various surrounding area masks are po-sitioned. The observed differences are used to ap-ply a correction to the signals observed in source

14



comparisons. The effect is measured at wave-lengths of 654.6 and 350 nm, and values for otherwavelengths are estimated from the assumption ofan inverse wavelength dependence. The correctionvaries from 0.04% to 0.1% at 654.6 nm dependingupon the elapsed time since the last mirror recoating.

4.4 Polarization

The polarization properties of the spectrora-diometer and the sources do not play a significantrole in the spectral irradiance realization and willnot be discussed here. A discussion of polarizationproperties can be found in reference [6].

5. Process of Spectral IrradianceRealization

The spectral radiance of the special integratingsphere source is determined so that it can be usedas a transfer standard for determining spectral irra-diance. The spectral radiance output from the cen-ter point of the integrating sphere aperture iscompared to the spectral radiance output from avariable-temperature blackbody. The temperatureof the blackbody is determined by comparing it at654.6 nm to a high stability vacuum pyrometerlamp calibrated for a single temperature (about1530 K). The spectral radiance of the integratingsphere source is determined at 31 different wave-lengths from 250 to 2400 nm. The aperture of theintegrating sphere is mapped at 2000, 1050, 654.6and 300 nm and its average spectral radiance iscomputed for each wavelength. Figure 6 shows atypical mapping profile of the integrating sphereaperture. The mapping correction varied less than0.1% over the range of wavelengths measured.

The spectroradiometer is changed from the spec-tral radiance mode to the spectral irradiance mode(see figs. I and 2) and the spectral irradiances fromthe NBS primary working standards (PWS) arecompared to the spectral irradiance from the inte-grating sphere source (ISS). Appropriate partitionsand baffles are erected to reduce scattered light toless than 0.1%. The comparisons are done at thesame 31 wavelengths at which the integratingsphere source was calibrated for spectral radiance.Two separate determinations are performed oneach primary working standard. The spectral irra-diance of a primary working standard is deter-mined using the relationship

EX(PWS)=Lx.( 7rrl 2)/R 2 .SPWS/S1 SS

from the primary working standard to the irradi-ance signal from the integrating sphere source. Thefirst part of the expression comes from eq (8a)where 8=2-10-6.

The absolute output from the integrating spheresource is monitored at six wavelengths (2000, 1600,1050, 800, 600, and 400 nm) during the 30 to 40operating hours necessary to calibrate the primaryworking standards. Finally, the blackbody is usedagain to perform an abbreviated spectral radiancecalibration of the integrating sphere source. Spec-tral radiance drift corrections, linear with time, forthe integrating sphere source can then be made ifnecessary.

The measurement of the spectral radiance orspectral irradiance at a single wavelength takesfrom about 4 to 8 minutes, so it is only necessaryfor our detectors to have good short term stability.

6. Process of Spectral Irradiance Transfer

The four modified FEL primary working stan-dards are used to perform spectral irradiance cali-brations on test lamps. For a selected group of 12test lamps, each lamp is measured four times, oncein each of the four source positions and onceagainst each of the four primary working stan-dards. The screening and selection of test lampscan take several weeks and the calibration proce-dure for 12 test lamps takes from 2 to 3 weeks.Details of the routine spectral irradiance calibra-tions appears in a separate document [5].

7. Scale Realization Data Analysis

The spectral irradiance scale is generally realizedat the following 31 wavelengths:

250 nm260270280290

300310320330340

350400450500555

(9)

where Spws/Siss is the ratio of the irradiance signal

600 nm654.6700800900

10501150120013001540

16001700200021002300

2400

15



Certain wavelength regions in the IR are skipped(around 1400 nm and 1800 to 1980 nm) in order toavoid atmospheric absorption bands.

Since the total operating time for each primaryworking standard during a complete scale realiza-tion is relatively short (8 to 12 hours), no effort ismade to account for irradiance lamp drift. The finalassignment of spectral irradiance is simply at-tributed to the lamp as of the midpoint of the burn-ing time. Between scale realizations when thegroup of four primary working standards is beingused as a basis for calibrating additional lamps,their drifts are taken into account. Various empiri-cal drift models have been used [3]. The presentdrift equation is

Ex=A +B-t, (10)

where:

t is time in burning hours,A and B are constants determined by fitting.

The fitting is performed independently at eachwavelength.

Drift of the spectral radiance of the integratingsphere source is taken into account by simple linearinterpolation in time between the initial and finalspectral radiance values.

An interpolation equation was developed for cal-culating the spectral irradiance of tungsten halogenlamps at wavelengths between the 31 calibratedwavelengths. This equation is

Ex=(Ao+Al.?L+..+A,.X').?L-5exp(a+b/A). (11)

Setting the polynomial equal to 1, multiplying bothsides by X5, and taking the log of both sides givesln(Ex.X')=a+b/X, in which it will be recognizedthat exp(a) is an effective gray-body emissivity andb is closely related to the reciprocal of the distribu-tion temperature. A least squares fitting using aweighting of 1 is performed to determine a and b.With a and b thus fixed, eq (11) is least squaresfitted using a weighting of I/E, 2 (assuming con-stant percentage measurement error) to determineAo, At, ... An. In practice it has been found that thefinal fit is considerably improved if the spectrum isbroken into two spectral regions, 250 to 400 nmand 350 to 1600 nm, for separate fitting. See refer-ence [3] for examples of fitting eq (11) to lamp data.This method is only valid for the continuous spec-trum and does not predict emission lines and ab-sorption bands. Spectral irradiance valuespredicted using eq (11) have an uncertainty ofabout 0.5%.

8. Uncertainty Estimation

The spectral irradiance scale uncertainty analysisis broken down into three parts. First, the uncer-tainty in the spectral radiance of the integratingsphere source is determined. Second, the uncer-tainty in the transfer to the spectral irradiance pri-mary working standards is determined. Third, theuncertainty in the transfer from the primary work-ing standards to the irradiance test lamps is deter-mined. The overall uncertainty in the primaryworking standards is determined by combining inquadrature the first and second parts. The overalluncertainty in a group of test lamps is determinedby combining in quadrature all three parts. All un-certainties are estimated at the 3Ir level.

8.1 Integrating Sphere Source Spectral RadianceUncertainty

The uncertainties in the spectral radiance valuesassigned to the integrating sphere source are ob-tained from the observed precision of the measure-ments and the estimated systematic error in boththe measured and the provided quantities (e.g.,temperature of melting gold). Uncertainties ob-tained from the observed precision and from thepublished values of the physical constants arebased upon three standard deviations. Uncertain-ties of systematic errors are estimated at the equiv-alent of three standard deviations.

In order to examine the contributions of the var-ious errors to the uncertainty in the spectral radi-ance of the integrating sphere source, anapproximate equation for the complete measure-ment process was derived by using the Wien ap-proximation to the Planck relation. The details ofthe derivation are described in reference [6]. Theresulting equation is

Lxrs(sx-eB-d-Mx)[ci/[w-.X5 (er/x T~u)]]

*(SrfMr/eB)r , (12)

where, with VTBB denoting the variable-tempera-ture blackbody and GPBB denoting the gold-pointblackbody, the definitions of the quantities are:

Mx, signal ratio of the VTBB-integrating spheresource comparison,

Mr, signal ratio of the GPBB-VTBB comparison,Sx, size-of-source correction for the VTBB-inte-

grating sphere source comparison,

16



Eu, effective emissivity of the VTBB,d, correction for integrating sphere source drift

during calibration,sr, size-of-source correction for the GPBB-

VTBB comparison,fi, linearity-range factor correction,TAU, IPTS-68 temperature of melting gold,cl, first radiation constant,c2, second radiation constant,X, wavelength of the VTBB-integrating sphere

source comparison,X,, wavelength of the GPBB-VTBB comparison,

654.6 nm.

Spectral radiance uncertainties due to the factorsof eq (12) are obtained from the partial derivativewith respect to those factors and the estimated un-certainty in the factor. Differences between errorscalculated by eq (12) and those calculated by theexact Planck relation are negligible. Note that forthe wavelengths X and Xr this process yields theerror due to inserting the wrong wavelength in thespectral radiance calculation, not the error due toan incorrect wavelength setting.

In addition to the factors which appear explicitlyin eq (12), uncertainties in the ratios Mx and M,

arise from errors in the wavelength settings X (0.1nm) and Xr (0.05 nm), in the current measurementsof the vacuum pyrometer lamps (0.2 mA) and theintegrating sphere source lamp (0.3 mA), and in themeasured spectral responsivity function. The un-certainties in the ratios due to wavelength settingand electric current are assessed at a number ofwavelengths by measurement of the change in sig-nal ratio when varying these quantities. This tech-nique for determining the effect upon the signalratios due to the uncertainties in the measuredspectral responsivity function is derived in refer-ence [20]. The spectral radiance uncertainties dueto these factors are then deduced from the ratiouncertainties as before. The signal ratio, lamp cur-rent, and wavelength setting errors are consideredrandom; the remaining errors are systematic.

Table I lists the uncertainties obtained by thisprocess. The calculated uncertainties, in percent ofspectral radiance, are tabulated for a number ofwavelengths over the calibration range. The indi-vidual values are combined in quadrature to yieldthe combined uncertainty for each wavelength.These uncertainties apply to-the spectral radiancesvalues of the integrating sphere source.

Table 1. Integrating sphere source spectral radiance uncertainty (3o-) in percent

Wavelength (nm)

Source of error 250 350 654.6 900 1300 1600 2000 2400

TA. (S)M, (r)MA (r)Sr (s)SX (s)fr (s)d (s)E' (s)X, setting (r)A setting (r)c, (s)C2 (S)

Lamp currents:Quinn-Lee (r)1530 K (r)155 (r)

Spect. resp. (s)

Quadrature sum

1.290.160.250.260.100.260.100.160.150.030.000.13

0.920.110.180.190.100.190.100.090.080.040.000.10

0.490.080.080.100.100.100.100.000.040.030.000.05

0.360.200.200.070.100.070.100.030.020.010.000.04

0.250.170.220.050.100.050.100.050.010.020.000.03

0.200.120.330.040.100.040.100.060.000.010.000.02

0.11 0.08 0.04 0.03 0.02 0.020.05 0.04 0.02 0.02 0.01 0.010.08 0.06 0.03 0.02 0.02 0.010.08 0.06 0.03 0.02 0.02 0.02

1.41 1.01 0.55

0.160.090.660.030.100.030.100.070.010.010.000.02

0.130.361.080.030.100.030.100.070.010.010.000.01

0.01 0.010.01 0.010.01 0.010.01 0.01

0.49 0.41 0.44 0.70 1.16

Quadrature sumwithout TA,

0.58 0.42 0.25 0.34 0.33 0.39 0.69 1.15

Notes: Random errors denoted by (r), systematic errors by (s).Sources of error described in section 8.1.

17

i



8.2 Radiance to Irradiance Transfer Uncertainty

The uncertainty in the transfer from the integrat-ing sphere source to the spectral irradiance pri-mary working standards is obtained fromexamining the contributions of the various errors inthe following measurement equation,

Ex(PWS) =m -di f (Spws1Sjs) L%(ISS)

-[( -r12)/R 2], (13)

where:

Ex(PWS), spectral irradiance of a primary work-ing standard,

m, mapping correction for the average spectralradiance of the integrating sphere source,

di, integrating sphere source drift correction,f, linearity-range factor correction,SpwsSss, signal ratio of the primary working

standard-integrating sphere source comparison,L%(ISS), spectral radiance of the integrating

sphere source,(iT-r1

2)/R 2, geometric factor in the irradiance cal-culation [see eqs (7), (8a), and (9)]

In addition to the factors which appear explicitly ineq (13), uncertainties in the ratio Spws/Ss, arisefrom errors in the wavelength settings and in theelectrical current measurements of the sources.There are also uncertainties due to spectral scatter-ing, stray light, and averaging sphere responsivity.All these uncertainties have been evaluated and arelisted in table 2.

8.3 Test Lamp Irradiance Transfer Uncertainty

The uncertainty in the transfer from the spectralirradiance primary working standards to a group ofirradiance test lamps is obtained from examiningthe contributions of the various errors in the fol-lowing measurement equation,

Ex(TL) =f(STL/SPws)-EX(PWS), (14)

where:

Ex(TL), spectral irradiance of a test lamp,f, linearity-range factor correction,STL/SPWS, signal ratio of the test lamp-primary

working standard comparison,EX(PWS), spectral irradiance of a primary work-

ing standard.

Table 2. Radiance to irradiance transfer uncertainty (30-) in percent

Wavelength (nm)

Source of error 250 350 654.6 900 1300 1600 2000 2400

SpWs/Sjss (r) 0.42 0.08 0.06 0.84 0.86 1.46 2.60 5.73f(s) 0.26 0.19 0.10 0.07 0.05 0.04 0.03 0.03m (s) 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10di (s) 0.10 0.10 0.10 0.10 0.10 0.10 0.10 0.10X (r) 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02Lamp currents:

ISS (r) 0.08 0.06 0.03 0.02 0.02 0.01 0.01 0.01PWS (r) 0.08 0.06 0.03 0.02 0.02 0.01 0.01 0.01

Geom. Factor (s) 0.20 0.20 0.20 0.20 0.20 0.20 0.20 0.20Spec. Scat. (s) 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05Stray Light (s) 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02Av. Sph. Resp. (s) 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Systematic error 0.36 0.31 0.27 0.26 0.26 0.25 0.25 0.25

Random error 0.43 0.11 0.08 0.84 0.86 1.46 2.60 5.73

Quadrature sum 0.57 0.33 0.28 0.88 0.90 1.48 2.61 5.74

Notes: Random errors denoted by (r), systematic errors by (s).Sources of error described in section 8.2.

18



In addition to the factors which appear explicitly ineq (14), uncertainties in the ratio STL/SPWS arisefrom errors in the wavelength settings and in theelectrical current measurements of the sources. Allthese uncertainties have been evaluated and arelisted in table 3.

8.4 Overall Uncertainty of the Primary Working

Standards

Table 4 lists the overall uncertainties of the pri-mary working standards. It is made up bycombining the results of tables I and 2. The differ-ences between lines la and lb (and between 3a and3b) are caused by the systematic uncertainty intro-duced by an assumed uncertainty of 0.4 K in thegold-point temperature.

8.5 Overall Uncertainty of a Group of Test Lamps

Table 5 lists the overall uncertainties of a groupof test lamps. It is made up by combining the re-sults of tables 1, 2, and 3 and adding a model error.The model error is necessary because the primaryworking standards drift with time. A time driftmodel is applied for each of the primary workingstandards [see eq (10)] but the possibility that thisdrift may be wrong introduces an additional uncer-tainty in table 5, but not included in table 2 ortable 4. This uncertainty was obtained by compar-ing the calculated extrapolated spectral irradiancewith further scale realizations. When the primaryworking standards are used between scale realiza-tions, this additional uncertainty must be combinedin quadrature with the other uncertainties.

Table 3. Test lamp irradiance transfer uncertainty (30-) in percent

Wavelength (nm)

Source of error 250 350 654.6 900 1300 1600 2000 2400

ST,/Spws (r) 0.87 0.21 0.15 0.42 0.68 0.72 1.59 2.60

f (s) 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Lamp currents:PWS (r) 0.08 0.06 0.03 0.02 0.02 0.01 0.01 0.01

TL (r) 0.08 0.06 0.03 0.02 0.02 0.01 0.01 0.01

Systematic error 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01

Random error 0.88 0.22 0.16 0.42 0.68 0.72 1.59 2.60

Quadrature sum 0.88 0.22 0.16 0.42 0.68 0.72 1.59 2.60

Notes: Random errors denoted by (r), systematic errors by (s).

Sources of error described in section 8.3.

Table 4. 1986 spectral irradiance scale uncertainty (3a) in percent (derived from tables I and 2)

250 350 654.6 900 1300 1600 2000 2400nm nm nm nm nm nm nm nm

1. NBS spectral radiance scalea. Absolute error (with respect 1.41 1.01 0.55 0.49 0.41 0.44 0.70 1.16

to S1 units)b. NBS long term reproducibility 0.58 0.42 0.25 0.34 0.33 0.39 0.69 1.15

(without TAU, see table 1)

2. Radiance to irradiance transfera.' Systematic errors 0.36 0.31 0.27 0.26 0.26 0.25 0.25 0.25

b. Random errors (3o- precision) 0.43 0.11 0.08 0.84 0.86 1.46 2.60 5.73

3. Spectral irradiance scale uncertainty(quadrature sum)

a. With respect to SI units 1.52 1.06 0.62 1.01 0.99 1.55 2.71 5.85

b. NBS long term reproducibility 0.81 0.53 0.38 0.94 0.96 1.53 2.70 5.85

19



Table 5. 1986 spectral irradiance scale transfer uncertainty (3o-) in percent (derived from tables 1, 2, and 3)

250 350 654.6 900 1300 1600 2000 2400nm nm nm nm nm nm nm

1. NBS spectral radiance scalea. Absolute error (with respect 1.41 1.01 0.55 0.49 0.41 0.44 0.70 1.16

to SI units)b. NBS long term reproducibility 0.58 0.42 0.25 0.34 0.33 0.39 0.69 1.15

2. Radiance to irradiance transfera. Systematic errors 0.36 0.31 0.27 0.26 0.26 0.25 0.25 0.25b. Random errors (3o, precision) 0.43 0.11 0.08 0.84 0.86 1.46 2.60 5.73c. Model error 1.38 0.80 0.78 0.77 0.77 0.82 1.00 1.20

3. Test lamp irradiance transfera. Systematic errors 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01b. Random errors (3o- precision) 0.88 0.22 0.16 0.42 0.68 0.72 1.59 2.60

4. Uncertainty of reported values(quadrature sum)

a. With respect to SI units 2.23 1.35 1.01 1.34 1.42 1.89 3.29 6.51b. NBS long term reproducibility 1.83 0.99 0.88 1.29 1.40 1.88 3.29 6.51

The differences between lines la and lb (and be-tween 4a and 4b) are again caused by the system-atic uncertainty introduced by an assumeduncertainty of 0.4 K in the gold-point temperature.

About the authors: James H. Walker, Robert D.Saunders, John K Jackson, and Donald A.McSparron are members of the Radiometric PhysicsDivision of the NBS Center for Radiation Research.

References

[1] Uriano, G. A., Garner, E. L., Kirby, R. K., and Reed,W. P., eds., NBS Calibration Services Users Guide 1986-88, Natl. Bur. Stand. (U.S.) Spec. Publ. 250 (July 1986).

[2] Stair, R., Schneider, W. E., and Jackson, S. K., A NewStandard of Spectral Irradiance, Appl. Opt. 2, 1151-1154(1963).

[3] Saunders, R. D., and Shumaker, J. B., The 1973 NBS Scaleof Spectral Irradiance, NatI. Bur. Stand. (U.S.) Tech. Note594-13 (Apr. 1977).

[4] Klose, J. Z., and Bridges, J. M., NBS Measurement Ser-vices: Radiometric Standards in the Vacuum Ultraviolet,Natl. Bur. Stand. (U.S.) Spec. Publ. 250-3 (June 1987).

[5] Walker, J. H., Saunders, R. D., Jackson, J. K. andMcSparron, D. M., NBS Measurement Services: SpectralIrradiance Calibrations, NatI. Bur. Stand. (U.S.) Spec.Publ. 250-20 (Sept. 1987).

[6] Walker, J. H., Saunders, R. D., and Hattenburg, A. T.,NBS Measurement Services: Spectral Radiance Calibra-tions, NatI. Bur. Stand. (U.S.) Spec. Pub]. 250-1 (January1987). Also see by same authors, The NBS Scale of Spec-tral Radiance, Metrologia 24 (to be published 1987).

[7] Waters, W. R., Walker, J. H., and Hattenburg, A. T., NBSMeasurement Services: Radiance Temperature Calibra-tions, Natd. Bur. Stand. (U.S.) Spec. Publ. 250-7 (Oct.1987). Also see by same authors, The NBS Scale of Radi-

ance Temperature, J. Res. Natl. Bur, Stand., 92, 17 (1987).[81 NBS Optical Radiation News No. 18, November 1976.[91 De Vos, J. C., Evaluation of the Quality of a Blackbody,

Physica 20, 669-689 (1954).[10] Kostkowski, H. J., Erminy, D. E., and Hattenburg, A. T.,

High Accuracy Spectral Radiance Calibration of Tung-sten-Strip Lamps, Advances in Geophysics 14, New York,NY, Academic Press, Inc., 111-127 (1970).

[11] Wilkinson, F. J., Astigmatism Errors in Radiance Measure-ments, Metrologia 20, 11-18 (1984).

[12] Popenoe, C. H., and Campbell, M. S., MIDAS ModularInteractive Data Acquisition System-Description andSpecification, Natl. Bur. Stand. (U.S.) Tech. Note 790,(Aug. 1973).

[13] Kostkowski, H. J., The Relative Spectral Responsivity andSlit-Scattering Function of a Spectroradiometer, Chapter 7of Self-Study Manual on Optical Radiation Measurements:Part I-Concepts, Natl. Bur. Stand. (U.S.) Tech. Note910-4, 2-34 (June 1979).

[14] Saunders, R. D., and Shumaker, J. B., Apparatus Functionof a Prism-Grating Double Monochromator. (submitted toApplied Optics.)

[15] Saunders, R. D., and Shumaker, J. B., Automated Radio-metric Linearity Tester, Appl. Opt. 23, 3504-3506 (1984).

[16] Coslovi, L., and Righini, F., Fast Determination of theNonlinearity of Photodetectors, Appl. Opt. 19, 3200-3203(1980).

[17] The International Practical Temperature Scale of 1968,Metrologia 5, 35-44 (1969).

[18] Weidner, V. R., and Hsia, J. J., Reflectiop Properties ofPressed Polytetrafluoroethylene Powder, J. Opt. Soc.Amer. 71 (1981).

[191 Saunders, R. D., and Ott, W. R., Spectral Irradiance Mea-surements: Effect of UV Produced Fluorescence in Inte-grating Spheres, Appl. Opt. 15, 827 (1976).

[20] Kostkowski, H. J., and Nicodemus, F. E., An Introductionto the Measurement Equation, Chapter 5 of Self-StudyManual on Optical Radiation Measurements: Part I-Con-cepts, Natl. Bur. Stand. (U.S.) Tech. Note 910-2, 58-92(Feb. 1978).

20



Radiometric Calibrations of Portable Sourcesin the Vacuum Ultraviolet

Volume 93 Number 1 January-February 1988

Jules Z. Kiose, J. Mervin The radiometric calibration program their uncertainties. Finally, the calibra-Bridges, and William R. Ott carried out by the vacuum ultraviolet tion services are delineated in an ap-

radiometry group in the Atomic and pendix.

National Bureau of Standards Plasma Radiation Division of the Na-Gaithersburg, MD 20899 tional Bureau of Standards is presented

in brief. Descriptions are given of the Key words: arc (argCn); arc (blackbody);primary standards, which are the hydro- arc (hydrogen); irradiance; lamp (deu-gen arc and the blackbody line arc, and terium); radiance; radiometry; Standardsthe secondary standards, which are the (radiometric); ultraviolet; vacuum ultra-argon mini- and maxi-arcs and the deu- violet.terium arc lamp. The calibration meth-ods involving both spectral radiance andirradiance are then discussed along with Accepted: October 2, 1987

1. Introduction

The vacuum ultraviolet (VUV) region of thespectrum has become important in several areas ofresearch and development. These include space-based astronomy and astrophysics, thermonuclearfusion research, ultraviolet laser development, andgeneral atomic physics research. Applications ofVUV radiation in chemical, biophysical, and medi-cal fields are widespread. Many applications re-quire knowledge of not only the wavelength of theradiation involved but also the intensity or flux ofradiation. This implies a calibration of some type.The calibration may be based upon a standardsource, i.e., one whose output is known, or a stan-dard detector, i.e., one whose response to a givenradiation level is known. Two general cases can bedistinguished. In the first case one wishes to deter-mine how much radiation a source such as the sunor a plasma device is emitting at a given wave-length. Usually, the source is not monochromatic,so a monochromator must be used to select the

desired wavelength. In this case the most directprocedure is to employ a standard source. Thesource to be investigated as well as the standardsource are set up in turn so that radiation from eachsource passes through the same monochromatorand optical elements. The calibration is performedessentially by a direct substitution of the standardsource for the one to be calibrated.

The second case occurs when one wishes toknow the flux in a monochromatic beam of radia-tion, such as the flux emerging from the exit slit ofa monochromator. For this determination a stan-dard detector is more appropriate; the flux is deter-mined by simply measuring the signal when thedetector is irradiated with the beam to be cali-brated. If one were to attempt to perform the cali-bration in the first case above using a standarddetector or in the second case using a standardsource, one would in both cases need to know thespectral efficiency of the monochromator used inthe measurement. This would require an additionalmeasurement using a second monochromator and

21



would introduce additional uncertainties and com-plexities. Therefore, a need exists for both standardsources and standard detectors.

Standard sources may be divided into primarystandards and secondary or transfer standards. Pri-mary standards are ones whose output is knownfrom basic principles. The primary standards ofVUV radiation include plasma sources, especiallythe wall-stabilized hydrogen and blackbody linearcs, and electron storage rings emitting synchro-tron radiation.

There are storage rings at several laboratories,including NBS, which are used as primary VUVradiation standard sources. They produce highly-polarized continuum radiation for wavelengths

0.03 anm. Limited access to storage rings andproblems with the emitted polarized light, how-ever, make it desirable to have available other stan-dard VUV sources. The wall-stabilized hydrogenand blackbody line arcs were developed to providealternative primary standard sources. Thesesources, however, are also not able to be easilyused for calibrations. Hence, secondary or transferstandards which are relatively easy to apply havebeen developed. These include the deuterium lampand the argon mini-arc. These sources are morereadily available and make possible relatively inex-pensive and convenient calibrations. Also for re-searchers having access to a storage ring, thesecondary standards are useful in making possiblemore frequent calibrations. Finally, the secondarystandards possess some useful properties not char-acteristic of the available primary standards suchas, for example, emission over a relatively largesolid angle.

The two principal radiometric quantities whichare measured and calibrated are radiance and irra-diance. For an object or source which emits radia-tion, the radiance is the radiant power emitted perarea per solid angle, L =(W cm-2 sr-'). If thesource emits a continuum, i.e., emits radiation at allwavelengths near a particular wavelength, thespectral radiance is the radiance per wavelength in-terval or bandpass, LA=(W cm2- sr-1 nm-'). Theradiance will in general vary over the source area,the direction, and the wavelength. The definitionassumes that the area, solid angle, and wavelengthband are small enough that the radiance does notvary greatly within these quantities. Irradiance isthe radiant power incident upon a target per area,E = (W cm-2 ), and spectral irradiance is the radiantpower incident upon a target per area per wave-length band, Ex=(W cm-2 nm'). A source of radi-ation may serve as a standard source of irradiance

by operating it at a given distance from the targetarea. Some sources may be used as either standardradiance or standard irradiance sources. A separatecalibration must be performed, however, for eachquantity.

The services performed by the Atomic andPlasma Radiation Division of the National Bureauof Standards include tests and calibrations ofportable secondary VUV standard radiance and ir-radiance sources. These are usually rare-gas dimerlamps, which emit continuum radiation over lim-ited wavelength ranges, and hollow cathode lamps,which emit spectral lines in the wavelength rangefrom the VUV through the visible. All sources aregenerally supplied by customers. The main groupsof customers have included those in the fields ofspace-based astronomy and solar physics who haveused standard sources to calibrate satellite, rocket,or balloon-borne spectrometers. Other customershave needed calibrations in the 100-300 nm rangefor plasma radiation studies.

2. Apparatus2.1 Primary Standards

2.1.1 The Hydrogen Arc A high temperature wall-stabilized steady-state hydrogen arc has been de-veloped as our primary standard of spectral radi-ance [1]. This type of arc lends itself to such a usebecause at sufficiently high temperatures it yieldsabsolute intensities independent of other radiomet-ric standards or of the accuracy of any plasma di-agnostics. Previous efforts at lower powers werehindered by large uncertainties in plasma diagnos-tics, a difficulty that has been overcome in the hightemperature arc. Figure I illustrates the UV spec-trum of several of the more common standardsources, including that part of the hydrogen arcspectrum that is the subject of this paper.

The method depends upon the phenomenon thatthe continuum emission coefficient for a stronglyionized hydrogen plasma which is in or close to thecondition of local thermodynamic equilibrium(LTE) is calculable to within one percent [2], Thisfollows from the fact that the essential spectro-scopic constants, i.e., the continuum absorption co-efficients and transition probabilities, are exactlyknown for atomic hydrogen. The continuum inten-sities emitted from a typical pure hydrogen wall-stabilized arc discharge in the spectral regionabove 91.5 nm are optically thin and a function ofthe electron density and temperature [3,4]. In thelow power hydrogen arc the electron density

22

VoLurne 93, Number 1, Jantuary-February 1985


lD' 1-

10-14

go-2 -

lC -3 -

L

I B L AB - 12 5 Y00 K,Iff BLACKBODY

BLACKBODY LIMITEDFROM 12500K ARC

LINES

Lyax

1cm HYDROGEN ARC

CARBON ARC -

30 WATT . /D2 LAMP I/

ITUNG~STENSTRI

i LAMPI IIlL

100 200)

xfnml300

plete, and any further increase in arc temperatureresults only in a gradual decrease in intensity be-cause of the decrease in the number density. Theweak wavelength dependence of this maximumseen in figure 2 occurs because of the change in theenergy distribution of the equilibrated electrons asthe electron temperature is varied.

7E

z

di

U-

2

LS-

4.0-

3.0- _ x124 nm

2.0-

JoF,09

OS-37 _

0-

400

o l-6 c00

Figure 1. Cormpadamn9 of spectral radiances fEr far t:V soures.The ogtput of tie Lydroger arc is given fSo she temrersatre ofmairmu coninuum emnti on; the Obtputs of the other sources

are at typical operating conditions.

and temperature are determined from plasma diag-nostics in the visible spectral region using availableradiometric standards [3]. These quantities are thenused to calculate the continuum intensities in theVUV. There are known to be significant uncertain-ties in the plasma diagnostics, and as a result wehave adopted as our primary standard a higherpower hydrogen arc which operates such that a2-mm diameter wall-stabilized discharge reaches atemperature of about 20,000 K. For these conadi-tions the continuum emission coefficient as a func-ticn of temperature shows abroad maximum whichshifts with wavelength as is shown in figure 2 ThisoptimumV condition is brought about essentially bythe compensating effects of an increase in the ion-ization Fraction and a decrease in the total numberdensity as the temperature is increased n a constantI-atm pressure operating environment. Beyond thismaximum [5,6] the ionization is practically com-

l" am

eo000Ti K)

20000 22030

Figure 2. The emission coefficient for a I-atm hydrogen plasmain LTE as a function of tenperature for several wave-lengths: 250amr (-); 200nr, (Rae); 150am (----2; and124 nn (- - >.

Figure 3 shows the measured radial dependenceof the hydrogen continuum emission coefficient at190 run and the corresponding calculated LTEtemperature (2] for an arc current of 80A Becausethe maximum in the coefficient is broad and theabsolute magnitude of the peak intensity is not verysensitive to the electron temperature, the emissioncharacteristics of the plasma are nearly homoge-neous over its certral region which extends toabout 0.6 mm from the arc axis. This is especiallysignificant if only a small sample of the region isobserved as indicated in the figure. First, it meansthat the alignment precision. which was iesponsi-ble for much of the uncertainly in the low powerarc method, is not critical here. Second, it meansthat the arc current necessary to obtain the maxi-mum emission coefficient is also not critical.

21

0,

EaOr

E;

3:

I-4

102

Volume 93, Number i, 3anuary-February 1985


I

I

o0o00 K

0.4 0.2 2 0o 2 I OR nl I

Figare 3. Radial dependence of the continuum intensity and arctemperature for a 2-mm wall-stabilized hydrogen arc operatingat 80 A. The dashed lines define the dimensions of the plasmathat is observed by spectrometers.

Additional advantages of operating the hydro-gen arc under such conditions are (1) the HILyman band molecular emission, which limited thelow power hydrogen arc to wavelengths longerthan 165 nm, is negligible; (2) the assumption ofLTE appears to be very closely fulfilled as shownby experimental consistency checks and theoreticalvalidity criteri

Date post:	07-Feb-2021
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

News Briefs and Reports - NISTNews Briefs and Reports Developments GROUND BROKEN FOR NBS COLD...

Documents