Light Sources in the 0.15-20-p, Spectral Range
`
`M. W. P. Cann
`
`Information
`The different kinds of light sources available for the 0.15-20-~ spectral range are surveyed.
`was obtained from the published literature, unpublished reports, light source manufacturers and also
`from individual persons. The aim has been to present sufficient information where availabl~ to show
`.
`,
`,
`t ere ative advantages of different sources-intensity, stability, and output uniformity were of prime
`h
`l
`interest. Continuum and line sources are included but lasers and pulsed sources are omitted. The
`sources are described under the main headings: Arc Discharge Sources, Glow Discharge Sources, and
`Incandescent Sources, with another section, Miscellaneous Sources to cover some which could not be
`included under the first three headings.
`'
`
`I.
`
`Introduction
`This paper is an extension of a survey which was
`undertaken in support of a continuing program at the
`George C. Marshall Space Flie:ht Center NASA in
`h. 1
`'
`'
`~
`w 1c 1 a remote sensing technique is used to study tur-
`bulent air flows. 1 The original survey was largely con(cid:173)
`fined to commercially available light sources and was
`reported earlier. 2
`lVIuch detail has been omitted in
`this paper and the subject matter has been extended
`to include all laboratory sources with the exception of
`pulsed sources, but including some repetitively pulsed
`sources for the uv, and lasers: the latter are such
`speci~liz~d devices, and there are now such a variety,
`that JUStice could not be done in a survey of this kind
`(several review articles on lasers have appeared in the
`last few years, some of them in this journal).
`The aim in this paper is to describe the different
`kinds of sources which exist or have been described
`in the literature, along with variations and improve(cid:173)
`ments reported later. Construction details are omitted
`but may be obtained from the literature cited.
`In the
`original survey, particular emphasis was placed on
`source performance in the form of the following charac(cid:173)
`teristics: (1) emission intensity-spectral steradiancy
`(W sr-1 em - 2 nm - 1);
`(2) uniformity-across the
`source as well as a function of angle; (3) drift-long
`term; (4) ac ripple in de sources; (5) random fluctua(cid:173)
`tions;
`(6) lifetime. When the above information is
`available this is indicated, but most literature sources
`~mit such data to emission intensities, often only rela(cid:173)
`tive.
`
`The author is with the ITT Research Institute, Chicago,
`Illinois 60616.
`Received 11 September 1968.
`
`Reviews of light sources have been made in the past
`and four of these complement the present one: W olfe 3
`(ir), Carlson and Clark4 (visible), Koller5 (uv), and
`Samson6 (vacuum uv). Another brief review is given
`by Kessler and Crosswhite.7 Use has been made of
`some of the material presented by these authors in
`order to achieve perspective, or to expand on a topic,
`but a minimum of duplication has been sought.
`The survey is divided into sections on arc discharges,
`glow discharges, and incandescent sources, followed by
`a miscellaneous section of sources not covered by these
`headings. The division between arc and glow dis(cid:173)
`charges is not always a clear one but this breakdown was
`found to be convenient.
`In the final section a summary
`is given of the general utility of the sources discussed as
`regards intensity and stability.
`(These quantities are
`summarized for several types of sources in Figs. 9, 10,
`and 11, which are referenced repeatedly throughout the
`paper.)
`
`II. Arc Discharge Sources
`Arcs as light sources owe their usefulness either to
`emission from the hot gas or to incandescence of the
`electrodes, and examples of each are given below.
`General descriptions of different kinds of arcs which
`have found application in the laboratory have been
`given by Finkelnburg and Maecker,8 lVIaecker, 9 and
`Lochte-Holtgreven. 10 Arcs are often found to exhibit
`fluctuations and instabilities which render them un(cid:173)
`suitable in many applications and it is only by exercisinO'
`considerable effort and ingenuity that highly stahl:
`arc source~ have been constructed.
`
`A. Compact Arcs (Short-Arc Lamps)
`The arc is contained within a fused )quartz envelope
`and is struck between two electrodes of tungsten or
`thor~ated tungsten for the higher power arcs; the' arc
`gap 1.s usually in the range 0.1-5 mm. The gas or vapor
`
`August 1959 I Vol. 8, No. 8 I APPLIED OPTICS 1645
`
`Energetiq Ex. 2072, page 1 - IPR2015-01377
`
`

`
`performance in the uv, it is possible to obtain these
`lamps with suprasil quartz envelopes.
`lamps are
`Capillary high-pressure mercury arc
`described in the commercial literature of PEK Labs
`Inc. and the General Electric Co. 21 The best known
`types are the A-H6 (1 kW, water-cooled) and the B-H6
`(900 vV, cooled by high-pressure air). Typical arc sizes
`are 2.5 em long and 0.1-0.2-cm diam; lifetimes vary
`from 5 h (for a 2-kvV arc) to 75 h for the A-H6.
`l\llore
`details of these lamps are given by Carlson and Clark4
`and by Koller. 5 A comparison between the spectral
`outputs in the uv of the A-H6 and other sources is given
`by Baum and Dunkelman, 22 which shows the A-H6 to
`be rather more intense than the xenon arc down to 240
`nm, but below this it is weaker.
`The shape of the electrodes has considerable influence
`on the radiative output characteristics. The brightest
`arcs are the small ones, operating with a very short
`gap, e.g., the Hanovia 959C has a gap of 0.6 mm.
`In
`general, the brightest lamps also have the greatest
`nonuniformity across the source and so are not neces(cid:173)
`sarily the most suitable for a given application, even
`though high brightness may be required. Apart from
`those arcs having three electrodes (a third electrode
`is sometimes used for starting), the radiative output is
`cylindrically symmetric, but variations are found in the
`planes which include the lamp axis. These variations
`are not symmetric, both because of the nonuniformities
`in the arc and also because radiant emission from the
`electrodes may be significant; see Fig. 1 and similar
`curves4 (others may be obtained from the manufac(cid:173)
`turers) which show the effect of electrode shape and
`spacing on the brightness contours. It seems that
`there is little available information on the wavelength
`dependence of these contours-Baum and Dunkel(cid:173)
`man22 show contours for a xenon arc at a wavelength of
`330 nm, but without any comparisons with other wave(cid:173)
`lengths.
`(Figures 9 and 10 show the spectral stera(cid:173)
`diancies of a mercury and a xenon arc compared with
`other sources-the emission is averaged over the visibly
`emitting region.) The spectral energy distribution is
`found to be fairly insensitive to operating power, 23
`Duncan, Hobbs, and Pai have made measurements on
`several xenon and mercury-xenon commercial arcs and
`give curves of spectral steradiancy for different op(cid:173)
`erating powers. 24
`For some applications there may be advantages in
`operating such lamps in a vacuum. This usually leads
`to early rupture due to overheating of the electrode
`seals. The manufacturers consulted had not had
`direct experience of this but believed that the lamps
`would operate in a vacuum if the seals were cooled, 18·23
`but with loss in lamp lifetime. Vacuum tests on a 2.2-
`kW Hanovia xenon lamp made at the Jet Propulsion
`Laboratory showed that the power had to be reduced. 19
`At the rated power the lamp life was only 59 hs: failure
`was catastrophic and was preceded by total envelope
`blackening (the lower seal attained a temperature which
`was 150 K higher than that found under normal opera(cid:173)
`tion).
`
`Fig. 1. Polar diagram of emission intensity for General Electric
`XE 5000 (5000 W lamp). Notice effect of anode emission.
`
`fill is at a high pressure during operation (20-40 atm)
`which results in a high luminous (visible) efficiency.
`There are also some mercury capillary arcs available
`which operate at pressures in excess of 100 atm, close
`to the limit set by the quartz envelope. The early
`history, theory, and construction has been described
`by Elenbaas, 11 for mercury lamps, and by Cumming12·13
`for xenon lamps; details of commercial lamps are
`available elsewhere. 2·4 These lamps operate up to
`high-power levels (e.g., PEK Labs Inc. cover the
`15 W-20 kvV range) and may require special cooling. 15-17
`The high brightness of the compact arc lamp is due
`to its high temperature. The mercury vapor arc
`operates in the 5000-7000 K range, 11 and xenon arcs
`have color temperatures around 6000 K, with true
`temperatures, in the hottest regions, of up to 9000 K
`(see Refs. 14 and 18). The lamps are normally avail(cid:173)
`able filled with xenon, mercury and xenon, or mercury;
`some mercury lamps also contain a small amount of
`argon to assist in starting. Compact arc lamps with
`special gas fills are available commercially on special
`order, and those studied include neon, argon, krypton,
`and combinations of xenon or krypton with mercury,
`cadmium, and zinc. 19 Some differences were noted in
`the spectral energy distribution: notably, increased ir
`emission with neon and enhanced blue and uv emission
`with a mercury-krypton-cadmium fill. Enhanced
`emission at 530 nm has also been obtained by adding
`thallium iodide to mercury arc lamps. 2° For improved
`
`1646 APPLIED OPTICS I Vol. 8, No.8 I August 1969
`
`Energetiq Ex. 2072, page 2 - IPR2015-01377
`
`

`
`Operation of short-arc lamps at higher than the rated
`power is possible but with reduced life. Laue 19 found
`that operation of a 5-kW lamp at 6 kW reduced the
`useful lamp life from 400 h to 100 h, and Osram18 report
`electrode deterioration and loss in arc stability when
`operated with short duration pulses at greatly increased
`current.
`In the selection of a light source, stability is an im(cid:173)
`portant criterion. The emission intensity should re(cid:173)
`main constant in time: the arc should not wander over
`the electrodes and it should restrike at the same loca(cid:173)
`tion after a period of extinction, in order to facilitate
`optical alignment. Power-supply ripple is not a serious
`factor since this can be reduced as required. Further(cid:173)
`more, the ripple in the light output may be appreciably
`less than that in the current (e.g., Howerton25 reports
`that a 4% ripple in the current gives rise to !% ripple
`in the emitted radiation, although others have found
`the two to be about the same 18 •23).
`Some types of arc are more stable than others, for
`example the Hanovia 901C is reported to be relatively
`stable with short-term fluctuations of less than 2%,
`but with the Hanovia 959C, which has a particularly
`small electrode spacing, the arc sometimes jumps from
`one location on the cathode to another, possibly 0.04
`em away 23-the arc is fairly stable initially and this
`effect starts after the first few hours of operation.
`(Wasserman26 reports similar arc wandering for the
`Hanovia 901C.) This phenomenon of arc wandering,
`and restrike at a different location, has been shown to
`be fairly general2 3 •26-28 and can contribute markedly
`to observed intensity fluctuations ( ± 10% in a few
`seconds, ± 40% in 10 min27); it has also been found
`that considerable improvement is obtained after 200 h
`to 300 h of operation (ageing). A pulsation of the arc
`has also been observed23
`: the arc remains in one spot
`but shows radial expansions and contractions, and there
`may be sudden changes in intensity.
`lVIeasurements
`do not appear to have been reported on mercury arcs,
`although it seems that these are more stable than the
`xenon arcs. 25
`Improved stability of either source can
`be obtained by using special power supplies, 29 or by
`employing feedback of the measured light output to
`control the lamp current. 28, 30
`
`B. Carbon Arc
`The high brightness of the carbon arc makes it an
`attractive light source, especially for the uv and ir.
`During operation, a crater depression is formed at the
`tip of the anode and this is viewed directly for the
`brightest source. The gases in the arc stream are
`hotter than the positive crater but normally have a
`much lower emissivity. At high-current densities the
`anode spot spreads over the entire tip of the anode and
`leads to rapid evaporation; with suitable additives to
`the electrodes, such an arc produces a very intense
`flame-this is the Beck arc. 8
`The temperature of the arc depends on the radiation
`conditions. A strongly radiating arc tends to be cooler,
`thus a 200 A Beck arc has a temperature of 6000-7000
`Kin the anode flame, while .a carbon arc at the same
`
`current with a scarcely luminous flame attains tempera(cid:173)
`tures of 11 000-12 000 K in the anode region. 8 The
`temperature of the carbon crater is lower. From the
`measured brightness, Finkelnburg and LatiP 1 deduced
`an equivalent blackbody temperature of 5400 K for the
`positive crater of a high-current arc, although the
`relative spectral energy distribution suggested a tem(cid:173)
`perature of 7000 K. This arc had a highly luminous
`flame which emitted more than 50% of the total
`radiative output of the arc. The theory of high(cid:173)
`intensity arcs is given in an earlier paper by Finkeln(cid:173)
`burg, 32 and by Harrington. 33 The latter discusses the
`effects of additives in the carbons.
`The spectral distribution of the emitted radiation
`depends on the region of the arc viewed, on power dis(cid:173)
`sipation, and on the nature of the carbons. Data on
`the spectral emission characteristics of the various
`types of carbons can be obtained from the National
`Carbon Company: Koller has summarized these in his
`book5 and reproduces several curves. The most
`noticeable feature is the strong emission from the CN
`violet bands, having a maximum intensity at a wave(cid:173)
`length of 388 nm, but also present are band systems of
`C2, N 2, and N2+. The different carbons cause most
`changes below 500 nm, but enhanced emission around
`700 nm is obtained with strontium-cored carbons.5
`33
`•
`In another paper Jayroe and Fowler34 report on the
`effect of the arc atmosphere on the appearance of the
`emission spectrum. They studied nitrogen, carbon
`dioxide, air, helium, argon, and mixtures of thes~. It
`was found that with helium and argon most of the band
`emission disappeared.
`In the presence of argon a
`strong clean continuum extended from 360 nm down to
`the 247.8-nm carbon line.
`In 1940 MacPherson published a paper 35 describing
`the properties of the low-current carbon arc and in(cid:173)
`dicating its suitability as a radiation standard, with an
`intensity much greater
`than
`the tungsten lamp.
`More recently other authors have published work on the
`same theme. 36- 38
`In this connection J affe39 has op(cid:173)
`erated a microscope illuminator arc at 7 A, with 6-mm
`electrodes, and claims properties as good as those ob(cid:173)
`tained by lVIacPherson with the larger arc. 35 Null and
`Lozier describe in detail the operation and properties of
`the low-current carbon arc, and give a set of conditions
`for optimum operation; an arc based on their work is
`now made by the Mole-Richardson Company. The
`positive crater of the carbon arc is a source of high in(cid:173)
`tensity, good uniformity36 •38 and good reproducibility(cid:173)
`measurements on twenty-one positive electrodes, out of
`six batches, gave brightness temperatures at 655.0 nm
`which all lay within 3797 ± 11 K. 38 Null and Lozier
`obtained an emissivity of 0.96-0.98 f.1. in the 400-nm to
`4.29-J.l. spectral range (these authors distrusted their
`results beyond 4.29 J.l.) and Shurer's results40 agree.
`The spectral steradiancy of the low-current carbon
`arc has been measured by Johnson, 41 using the same
`type of arc as MacPherson.
`Johnson's results are
`shown in Fig. 2: these show the sharp cutoff due to air
`below 190 nm, and strong carbon lines at 193 nm and
`247.8 nm.
`In the figure, the solid line represents
`
`August 1969 / Vol. 8, No.8 / APPLIED OPTICS 1647
`
`Energetiq Ex. 2072, page 3 - IPR2015-01377
`
`

`
`spectral steradiancy are given for the visible spectral
`region.
`
`C. Argon Arc
`Argon arcs are widely used in the laboratory and their
`properties have been described by Goldman, 44 Olsen45 •46
`and others. Olsen's arc was specially built to have the
`high stability required in quantitative spectroscopic
`studies. It operates at 400 A in 1.1 atm of argon
`between a thoriated tungsten rod cathode and a copper
`J\!Iass loss from the electrodes was negligi(cid:173)
`plate anode.
`ble during 100 h of operation. The measured arc volt(cid:173)
`age varied by less than 1% over day-long periods of
`operation, and it was possible to extinguish the arc,
`replace the cathode, and restrike with the measured
`voltage and radiation intensity within ±3% of the
`previous value. J\!leasured absolute spectral line in(cid:173)
`tensities at the beginning and end of a 2-day period of
`continuous operation were also reproduced within
`± 3% of their maxima. Olsen ascribes the good sta(cid:173)
`bility primarily to the high purity of the gas comprising
`the plasma and to the favorable volt-amp characteristics
`of the arc and power supply. In conjunction with the
`high-emission intensities obtained, these are attractive
`features for a light source. There is the usual drawback
`with an arc that the large-temperature gradients cause
`poor uniformity of emission across the effective emitting
`area.
`The very high temperatures encountered in the
`argon arc suggest that it might be a useful source for
`the uv and vacuum uv spectra] regions. Some approxi(cid:173)
`mate relative intensity calculations were made by the
`present author, and matched to the intensity given by
`
`:X:
`
`LOW CURRENT CARBON ARC
`3900 K
`
`6
`5
`
`1-
`<t
`a:
`<t
`III
`0
`.J
`(.!)
`<t
`0
`1-
`w
`::::
`1-
`<t
`.J
`w 3
`a:
`>-
`1-
`C/)
`z
`w
`1-z
`
`2
`
`10,000
`WAVELENGTH ( n m)
`
`15,000
`
`Fig. 3. Spectral steradiancies of the low-current carbon arc, 42
`zirconium arc62 and tungsten glower 174 relative to the globar;
`color temperatures are indicated.
`
`-I
`
` E
`
`
`c:
`>- 7 5
`u"C
`z 0
`~ :v 4
`o(cid:173)ct(/)
`a: N
`UJt
`f-E
`(/) u 2
`
`3
`
`0(
`~
`
`220 230 240 250 260
`WAVELENGTH
`(n m)
`
`Fig. 2. Spectral steradiancy of low-current carbon arc in the uv. 41
`
`radiation from the arc stream as well as from the
`incandescent anode. The broken line is for radiation
`from the arc stream only (viewed 1 mm in front of the
`anode) and shows that this is the major contributor at
`these wavelengths. Emission intensity data for wave(cid:173)
`lengths longer than 4.2 Jl. have been reported by Rupert
`and Strong, 42 who compared the relative intensities
`of the carbon arc and the globar, operated at 1175 K
`(see Fig. 3). The arc emission profile shows absorption
`due to H20 and C02 as well as a broad, shallow absorp(cid:173)
`tion feature around 10 JL which was not identified. The
`spectral steradiancy of the Mo1e-Richardson arc has
`been measured by Hattenburg for the spectral range
`210-850 nm (Ref. 37): his results show a continuum
`which follows the emission of a blackbody at a tem(cid:173)
`perature of 3792 K, within 2%-line and band emission
`are superposed on this continuum.
`Brightness distributions across the arc crater have
`been published for the high-current carbon arc31·33
`and for the low-current arc. 36·38 Stability measure(cid:173)
`ments have been reported by several authors. Jayroe
`and Fowler34 found that within a 5-sec period, very
`considerable
`intensity fluctuations
`(of millisecond
`duration) occurred with the carbon arc operating in air;
`operation in an argon atmosphere improved the steadi(cid:173)
`ness of the arc. Lee and Lewis38 report temperature
`variations of up to 10 K during the useful life of a car(cid:173)
`bon; short-term radiation fluctuations (measured with
`a time constant of Pi sec) showed an equivalent standard
`deviation of 3 K. The lVIole-Richardson Co. quotes
`long-term variations as equivalent to a temperature
`change of 12 K, measured at a wavelength of 650.0 nm,
`and short-term as 2 K: noise(~ 6Hz) as 3 K. Trans(cid:173)
`lated into intensity fluctuations, this is approximately
`a noise level of ± 1%, and a drift of ±2Y2% (for an
`unspecified time period).
`Commercial projection arcs have been investigated
`for use in solar simulation, these are high-current arcs.
`1\feasurements on the Genarco J\!Iodel No. J\!lE4CWlVI
`(205-235 A) arc were made by Alexander43 who gives
`curves showing uniformity of i.rradiance and stability.
`The Strong Electric Corporation arcs have been
`studied by Duncan, Hobbs, and Pai24-Jetarc, 145-A de;
`U-H-1 arc, 175-A de; J\!Iagnarc, 60 A de-and curves of
`
`1648 APPLIED OPTICS I Vol. '8, No. 8 I August 1969
`
`Energetiq Ex. 2072, page 4 - IPR2015-01377
`
`

`
`EXHAUST GAS
`
`Fig. 4. Giannini vortex stabilized radiation source.
`
`ARC
`
`Olsen for a wavelength of 553.5 nm. The result of this
`calculation is shown in Figs. 9, 10, and 11. Selfabsorp(cid:173)
`tion of radiation was neglected so that these curves
`represent upper limits only; however, they do suggest
`that the source may prove valuable in the uv and ir.
`The argon-arc spectrum shows many strong emission
`lines of Arl and Arii superimposed on a continuum.
`By introducing suitable powders into the gas feed to
`the arc, it is possible to excite emission lines of a variety
`of materials, 47 providing intense emission at new wave(cid:173)
`lengths.
`A commercially available argon arc is the "vortex
`stabilized radiation source" (VSRS), made by Giannini
`Scientific Corp. This source is available in various
`models operating in the 5-150 kW-range (electrical).
`A diagram of the source is shown in Fig. 4. The arc is
`contained within
`two fused-silica envelopes. The
`high-pressure gas flows between these envelopes, pro(cid:173)
`viding thermal insulation, and is introduced around
`the cathode where it forms the vortex which contains
`the arc. The exhaust gas is extracted through the
`hollow anode. Figure 5 shows the spectral energy
`distribution curve supplied by the manufacturer which
`appears to have been taken at low spectral resolution,
`and so is a little misleading. The VSRS has been
`studied extensively by Neuder and Mcintosh48 and
`their curves of spectral energy distribution show much
`more line structure. These authors also give stera(cid:173)
`diancy contours· for the VSRS. The angular distribu(cid:173)
`tion of the emitted radiation is approximately a cosine(cid:173)
`squared function. Giannini Scientific Corp. quotes the
`ripple in the output radiation as less than 1% for
`the 20-kW VSRS and indicates that an intensity drift
`of less than 1% in 5 min is attainable. Random
`fluctuations are also given as less than 1%. Neuder and
`Mcintosh found the stability over several hours to
`be around ±2%. The lifetime (of the 24-kW source)
`is in excess of 100 h without replacing the electrodes or
`silica tubes (which darken during operation).
`Several special arcs have been constructed for opera(cid:173)
`tion in the vacuum uv. Boldt describes a cascade arc
`which is equipped with interlocks and buffer gases to
`maintain transparency to the radiation. 49 •50 Another
`type is the magnetically confined vacuum arc described
`by Burns et al.,51 which is operated at very low pressures.
`These elaborate techniques may not offer significant
`gains for the spectral region down to 150 nm; if self(cid:173)
`absorption is not too great, a sapphire window on the
`
`Olsen-type arc should provide a source which is more
`intense than those normally used in the uv (see Fig. 9).
`
`D. Hydrogen Arc
`Several kinds of hydrogen arc are used as sources of
`uv radiation. Apart from the large, high-power devices
`such as those described by Finkelnburg and Maecker8
`or Lochte-Holtgreven, 10 there are the small encapsulated
`devices, usually operating at up to 200 W, which are
`available commercially. These low-power arcs are
`enclosed in Pyrex or fused-quartz envelopes, or may
`have a Pyrex envelope with quartz window, and are
`obtainable with either hydrogen or deuterium gas fills.
`Increases in output intensity by a factor of 2.5 have
`been claimed when deuterium is used instead of hydro(cid:173)
`gen.
`The commercial lamps, such as those marketed by
`Beckman, Sylvania, or Oriel Optics, operate at 30 W,
`50 W, or 60 W. Their construction includes a fila(cid:173)
`mentary electron source for starting, placed within the
`cathode which is a nickel cylinder. The anode is a
`half-cylinder of nickel and molybdenum and both the
`cathode and anode have small holes opposite one
`another. The hydrogen continuum generated by
`these lamps is useful in the upper regions of the vacuum
`uv, where the quartz window allows them to be used
`down to about 185 nm; with especially thin windows
`they may be used to about 160 nm. The lamps emit
`into a cone angle of approximately 60°. Lifetimes are
`in the 200-500-h range (to 50% of initial output).
`Ripple and noise figures of around 0.1% are quoted by
`the manufacturers and a circuit for improved stability
`is described by Simpson. 52
`A more powerful lamp, built by Nester, is described
`by Allen and Franklin53 and compared with other
`sources by Baum and Dunkelman22 ; a later version of
`the lamp54 is described in a report by Amicone, et
`al. 55 This lamp is filled with hydrogen to a pres(cid:173)
`sure of 0.55 torr and is water-cooled. As with the
`above arcs, a filament is used for starting. The dis(cid:173)
`charge operates at 1-1.3 A at 56 V.
`In the work re(cid:173)
`ported by Amicone et al., stability was fairly important,
`so this is probably quite good although no figures were
`
`~ e o.s
`
`~
`<1)
`
`:!: c
`~7 0.6
`1-
`~ ~ 0.4
`o!
`<t002
`a::~ .
`
`o.o~~~~~~~~~~~
`200
`500
`1000
`1400
`WAVELENGTH {nm)
`
`Fig. 5. Spectral intensity distribution for 17-kW VSRS viewed
`at 90° to the axis. Arc size 10 mm X 1.8 mm (Giannini litera(cid:173)
`hue).
`
`August 1969 I Vol. 8, No. 8 I APPLIED OPTICS 1649
`
`Energetiq Ex. 2072, page 5 - IPR2015-01377
`
`

`
`'
`
`>- 100
`(!)
`a::
`I.LI
`z 80
`I.LI
`
`I.LI
`> 60
`...
`<t
`_J
`I.LI
`a::
`....
`z
`I.LI
`u
`a::
`I.LI
`0..
`
`10
`8
`6
`4
`2
`WAVELENGTH {Microns)
`
`Fig. 6. Relative spectral energy distribution for Sylvania KlOOO
`concentrated arc lamp (zirconium arc with quartz window).
`
`given. A very similar lamp is supplied by Bausch &
`Lomb (the lamp is made by Glass instruments Inc.).
`Other arc lamps have been described in the literature
`by Johnson56 and Hartman.57 •58 The Hartman lamp
`has a quartz capillary 1 em in diam, with an orifice
`through the anode of about 0.33 em diam.
`In opera(cid:173)
`tion the lamp stability was reported to be good.
`From the geometry it is apparent that the emitted
`radiation will be confined to a comparatively narrow
`angle. Samson reports6 that the intensity is greater by
`a factor two or three than that produced by the cold(cid:173)
`cathode discharge (see Sec. III-A).
`Rendina59 has described a metal-cased arc lamp
`which incorporates features of both the Johnson and
`Hartman lamps, but which can be operated at higher
`power levels and which emits into a wide solid angle;
`the author claims ease of construction and operation
`as other advantages. Another design, similar
`to
`Rendina's, has also been published60 and may have
`some advantages: it is simpler in design but, unlike the
`Rendina lamp, is not water-cooled and so is limited to
`lower power levels.
`
`E. Zirconium Arc
`Concentrated arc lamps using zirconium oxide in
`the cathode were developed during World vVar II. A
`detailed discussion of the construction and performance
`of these small lamps has been given by Buckingham
`and Deibert.61 Although this paper is quite old now,
`it appears that little or no development has taken place
`since it appeared, so that the information is still
`accurate. The unique feature of these lamps is the
`cathode, which is made by packing zirconium oxide into
`a small cup at the end of a tungsten, molybdenum, or
`tantalum electrode, metals with high melting points.
`The anode is also a high melting-point material and,
`because of its large_ surface area, remains relatively
`cool. The anode is mounted close to the cathode tip
`and has a hole through which the cathode may be
`viewed. During operation, molten zirconium metal is
`formed and it is this which is the direct source of
`radiation.
`
`1650 APPLIED OPTICS I Vol. 8, No. 8 I August 1969
`
`These lamps have an argon fill and spectrograms
`show lines of Arl, Zrl, Arii, Zrii, and Zriii (lines of
`Arl and Zrii are the strongest) superimposed on a
`continuum which rises from about 400 nm and peaks at
`about 2.0 J.l. (see Fig. 6). The continuum radiation is
`emitted chiefly by the molten cathode surface and has
`a peak at around 2.0 J.l.. The lines are emitted by a
`cloud of excited gas and vapor near the cathode.
`The brightness distribution across the source is quite
`flat for the higher power lamps (25-100 vV). 61 Varia(cid:173)
`tions in the position of the arc stream and irregularities
`in the cathode surface may produce asymmetry in the
`brightness distribution and also some instability. The
`spot position may move slowly during operation, by an
`amount equal to a small fraction of its own diameter.
`These factors produce variations of about 10% in the
`emitted light. During the first few hours of operation,
`the intensity of the emitted radiation and the spot
`diameter decrease and the brightness increases. After
`about 100 h of operation, these characteristics are
`nearly stable with a brightness 1.4 times the initial
`brightness.
`For use in the ir, these arcs require modification to
`take a suitable window. Turrell62 and Hall and
`N ester63 have made such modifications. Turrell
`modified a Sylvania 300-W arc and compared the output
`with a globar operated at 1175 K. The arc was
`appreciably stronger, as can be seen in Fig. 3. Turrell
`reports that the arc was stable to 2% for periods of
`over 1 h (judging from Buckingham and Deibert's
`results, this was probably for an aged arc).
`
`F. Mercury Vapor Arcs
`Small mercury vapor lamps are manufactured by
`several companies. High-power
`lamps of special
`design are used in Raman spectroscopy and photo(cid:173)
`chemical studies, and other types are used for lighting
`purposes, usually in combination with a phosphor to
`enhance
`the
`luminous efficiency.
`In addition
`to
`mercury, these lamps nearly always contain a rare gas
`at low pressure for starting: in some types of lamp there
`is a mercury reservoir.
`Initiation of a discharge causes
`the mercury to evaporate and the arc to start: vaporiza(cid:173)
`tion of the mercury raises the pressure which increases
`the efficiency of the lamp.
`lVIost of the radiation is
`emitted in resonance line at 253.7 nm, with emission at
`the 184.9-nm resonance line smaller by a factor of ten
`to fifty; some improvement at the lower wavelength is
`possible by making the envelope of suprasil quartz.
`Barnes64 reports on a study of the absolute emission
`intensities of low-pressure mercury vapor lamps con(cid:173)
`taining different rare gases. He found that the 184.9-
`nm radiation was 12-34% of that at 253.7 nm. The
`absolute intensities varied with the bulb wall tempera(cid:173)
`ture, current, type and pressure of the gas fill. Other
`measurements have been published by Childs65 and
`Read. 66 The former investigated a commercial lamp
`and found that 92% of the radiation was emitted at
`253.7 nm; Childs also gives the relative emission in(cid:173)
`tensities of 45 mercury lines between 253.7 nm and
`579.0 nm for this source. It is clear that the relative
`
`Energetiq Ex. 2072, page 6 - IPR2015-01377
`
`

`
`em1sswn intensities at 253.7 nm and 184.9 nm (for
`example) can be varied quite appreciably by changing
`the operating conditions.
`Mercury arcs find many applications in Raman
`spectroscopy and in photochemistry, leading to the
`development of special forms such as the Toronto
`arc67- 69 and others. 70 Beckey, Groth, Okabe, and
`RommeF 1 have measured the output for several different
`types of lamp including one which was excited by
`microwaves.
`
`G. Other Metal Vapor Arcs
`The best known are probably the sodium arc lamps.
`In the visible part of the spectrum, these lamps typically
`emit all but 2% of the radiation in the intense D(cid:173)
`lines (589.0 nm and 589.6 nm). It is usual to have some
`rare gas present for starting the arc discharge, other(cid:173)
`wise it is necessary to provide some other way of vaporiz(cid:173)
`ing the metal. Small low-power light sources of this
`kind are available commercially (Osram, Philips, Hilger
`and Watts) filled with Cd, Cs, Hg, Hg-Cd, Hg-Cd-Z