Penning discharge as a photoelectric EUV spectroscopysource

E. S. Warden and H. W. Moos

A Penning discharge was investigated photoelectrically as a light source from 100 A to 300 A. The sourceshowed reasonable stability-10% over an hour's operation. Bright spectral lines observed were a mixtureof lines from the filling gas (Ne) and from sputtered cathode material (Al). The discharge was found to bean intense line source; an estimate of the brightness of the 160 A Al IV lines shows them to be comparablewith the highly ionized metal lines appearing in a pulsed tokamak discharge. Satellites of the Al IV reso-nance lines at 160 A appear clearly in the spectra.

1. Introduction

The measurement of plasma impurity radiationin the grazing incidence region between 100 A and 300A is of current interest, for example, controlled ther-monuclear research and solar studies. An electricallyquiet, cw, line source providing high fluxes of photonswould ease the problems involved with photometric andwavelength calibration of instruments designed to workin this region.

Extension of soft x-ray line sources' to longer wave-lengths is one possibility. These sources currently usethe K shell x rays of the second period elements and theM-shell x rays of the rare earth elements for wave-lengths shorter than 114 A, the K shell x ray of Be. TheL shell x rays of Al and Si at 170 A and 135 A have beenused in some calibration work, but these spectra aresomewhat broad. 2

A number of gas discharges, e.g., cold cathode, cap-illary discharge, have been developed that producewavelengths in the He spectrum down to 237 A.Shorter wavelengths can be produced by running dis-charges in Ne. Manson 3 reports usable low flux signalsdown to 130 A for a Ne hollow cathode discharge com-bined with polyvinyl chloride filters which eliminatedlong wavelength scattered light.

The Penning discharge or Philips ionization gauge(PIG) has not been extensively used as a spectral sourcedue to cathode erosion. It is, however, a standardsource of high charge state ions for accelerators. 4 Thebasic feature of the discharge, which typically operatesat pressures below 10-3Torr, is an external magneticfield oriented perpendicularly to the two cathode faces.Electrons emitted from the self-heated cathodes arethus forced to oscillate between the cathodes until they

The authors are with Johns Hopkins University, Physics Depart-ment, Baltimore, Maryland 21218.

Received 15 January 1977.

can diffuse to the anodes by virtue of collisions. Pre-vious work by Deslattes et al. 5 has demonstrated thatthe PIG source produced light from multiple ionizedspecies. We decided to investigate the spectral outputof a Penning discharge in order to determine whetherit was a useful high flux source of short wavelength lightfor use with photoelectric detectors. This paper reportsthe results of this study; the Penning discharge is indeeda useful source of bright emission lines with reasonablestability.

11. Experimental Details

The design (Fig. 1) was similar to that of Deslattesexcept for a few minor features. The body of the lampwas made of brass. The high voltage insulators wereconstructed of machinable ceramic. Viton 0-rings wereused to seal the insulators to the cathode and anode. Inview of the problem with cathode erosion the cathodeswere constructed in two pieces-for easy replacementand in order to minimize the construction time for newcathodes. A limiting diaphragm of brass, the samematerial as the anode, was found to be useful in ad-justing the diameter and the current-voltage charac-teristics of the discharge.

The power supply for the discharge was rated up to5 KV, 1 A. The discharge was ballasted by a 460-Q re-sistor and was normally run in a 3000-G magnetic fieldperpendicular to the cathode face. The bulk of themeasurements were made with Ne fill gas at pressuresranging from 10-4 Torr to 10-3 Torr. The previousmeasurements of Deslattes were made with 1000-Gfields, but it was found that higher magnetic field madethe discharge easier to operate.

The spectral output was measured with a 1-m grazingincidence monochromator which was capable of scan-ning the 0-600-A range and using a grating blazed at 300A. Light was detected with a channeltron operated inthe pulse counting mode. The measurements reportedhere were made with the grating in the monochromator

1902 APPLIED OPTICS / Vol. 16, No. 7 / July 1977

CATHODE

1.0-

V) 0.8-0

w 0.6-

, 0.4 -0C 0.2 -

0.0

Fig. 1. Schematic of the source. See text for details.

set for an angle of incidence of 800. The lamp outputwas checked at 85° angle of incidence and exhibited nosignificant differences in the characteristics of the ob-served spectra.

111. Results

The discharge operated in two modes, a low currentmode (Mode I) and a high current mode (Mode II).6The discharge was operated in most instances quiteclose to the transition between the two modes. A 10%change in gas pressure or applied voltage could oftenchange the discharge current and light output by morethan an order of magnitude by switching between thetwo modes. The choice of cathode material was foundto influence the current-voltage characteristics. Aftersome initial trails with other materials, Al was used asthe cathode material. During operation, the dischargecurrent could sometimes be observed to fluctuate at arate whose average value was about 3 Hz. The magni-tude of the fluctuations was approximately 5-10% of theaverage value. This fluctuation could be observed whenlooking directly at the discharge through the viewingport if the bright light was sufficiently attentuated.The fluctuation was also observable in the detectedEUV light. The discharge could be adjusted by varyingvoltage and gas pressure so as to eliminate these fluc-tuations with the limiting diaphragm open as wide as1 cm. A greater range of stable operating conditionscould be obtained by closing the diaphragm to sizes assmall as 3 mm. With a smaller anode channel, the lightintensity was reduced in a manner approximately pro-portional to the area of the channel.

After initial warmup, the discharge could be adjustedto a stable mode of operation. The intensity of aspectral line was found to be stable to within 10% forgiven values of current, voltage, and magnetic field overthe period of 1 h. However, oscillations with periodsof 15 sec to 1 min having an amplitude as great as 10%of the average signal could at times be observed.

The high current mode was used primarily to bringout the short wavelength features of the spectrum. Atypical spectrum for a high current Ne discharge is

0

0

I0 -

8 -

6 -

4 -

2-

ro 10 U)(D

Hz z

I

.' I

fr

.4r0N

z

.45

zII

.4

N N

I4

Li I

I I 120 140 160 180 200 220 240

WAVELENGTH (ANGSTROMS)

l1~ 2 3260 280 300

Fig. 2. Spectra from Ne discharges. Upper: hollow cathode run

at 0.5-A discharge current. Lower: Penning discharge run at 1.0-Adischarge current, 850 V. Note: upper trace is ten times more sen-

sitive than lower trace.

5-

-z

C)zWH

4-

3-

2 -

I -

0-

W

ro

tCL.Into0.

N

--LI

ON Icq )

U) CCu)In In01. C. Q.

Cq CQ NI I I

tn CO CO0(0 C C.

Nl N Nq

WAVELENGTH

Fig. 3. Detail of 160-A Al IV doublet (2p6 -2p 5 3s) and its satellites

in the 150-180-A wavelength region. The wavelengths of the two AlIV lines are 160.1 A and 161.7 A. The satellite lines, whose upperlevels are above the autoionization limit of Al III, are not fully resolved

in the figure.

July 1977 / Vol. 16, No. 7 / APPLIED OPTICS 1903

W

ItJ

""N. �

shown in Fig. 2. Note the two strong Al IV lines origi-nally reported by Deslattes at 160 A arising from Alsputtered off the cathodes. The intensity of these linesvaried linearly with current for a given discharge volt-age. At slightly longer wavelengths is a group of sat-ellites of 160-A lines possibly arising from processes suchas dielectronic recombination of Al IV, inner core exci-tation of Al III, or inner core ionization of Al II. Figure3 details the 160-A lines and their satellites. The upperlevels of the satellite transitions are identified as au-toionizing states of Al III; the presence of an additionalbound electron acts as a perturbation to the 2p6-2p53stransition. The perturbation of the 160-A transitionis too small to have been observed for dielectronic re-combination where the recombined electron has aprincipal quantum number greater than 3. Secondorder spectra reveal the group to consist of at leasteleven lines. A theoretical estimate by Lin7 of thewavelength splittings of the thirteen predicted satellitelines with respect to the main lines for the perturbingelectron in the n = 3 level confirms this identifica-tion.

For purposes of comparison, Fig. 2 also shows thespectra from a hollow cathode lamp similar to that de-scribed by Paresce et al. 8 operated at 0.5-A current. Atthese currents a large amount of sputtered material wasdeposited on the entrance slit and thus by inference onthe grating by the hollow cathode discharge. ThePenning discharge, although sputtering more material,deposited much less material on the entrance slit asthere was no direct path from cathode to entrance slit.The solid angle subtended by the two lamps was ap-proximately equal, the major limitation on the solidangle of hollow cathode lamp being the differentialpumping slit. The Penning discharge is not only richerbut much brighter in this wavelength region.

To reduce the saturation of channeltrons by highcounting rates, the PIG source was placed 50 cm fromthe entrance slit, and a system of masks placed betweensource and entrance slit was used to lower the intensity.A lower limit on the brightness of the 161.6-A Al IV linein Fig. 3 can be estimated using the published 9 quantumefficiency of channeltrons at 160 A (1%), an assumed10% reflectance in first order for the grating, the effec-tive solid angle viewed (4 X 10-5 sr), the size of the en-trance slit (6.5 X 10-4 cm2), and the count rate (46,000counts/sec). The resultant brightness of 2 X 1015photons/sec cm2 sr is comparable with the metal linebrightnesses in this spectral region reported for thepulsed tokamak discharges used in controlled thermo-nuclear research.10 11 The apparent brightness of the160 A Al IV doublet is about that of solar Ly a and is

fifty times greater than the solar He II 304.12 Thus, thePIG source does seem to be a high brightness source atshort wavelengths.

IV. Discussion

Modern magnet materials and design make feasiblethe construction of compact Penning discharges andtheir use as a routine tool in photoelectric spectroscopy.The major disadvantage of the Penning discharge iscathode erosion. This erosion occurs mainly when thedischarge is run at high currents, i.e., 0.5-1 A, where theerosion rate is on the order of 2 mm/h with Ne fill gas.However, complete refurbishing of the discharge re-quires only 1 manhour of work. Also, it is anticipatedthat the need for high current operation is limited tohigh flux measurements at short wavelengths and thatlamps of this type would be run at high currents for onlya small percentage of their total operation. (A com-mercial design for low current operation is presentlyavailable from Jobin-Yvon, but is not suitable for ob-taining the high ionization states reported here.)

Finally, it should be noted that the PIG dischargeneeds no differential pumping, offers less chance ofgrating contamination than the hollow cathode becausethe cathode is not in the direct line of sight, and is amore intense light source below 200 A. In addition, itoffers a strong line at 160 A midway between the 113-ABe x ray and the 237-A He line. Further investigationand development of this type of light source seem war-ranted.

This research was supported by the U. S. Energy Re-search and Development Administration under contractEY-76-S-02-2711.

References1. B. L. Henke and M A. Tester, Advances in X-Ray Analysis

(Plenum, New York, 1975), Vol. 18, p. 76.2. H. W. B. Skinner, Philos. Trans. R. Soc. London Ser. A: 239,95

(1940).3. J. E. Manson, Appl. Opt. 12, 1394 (1973).4. J. R. J. Bennett, IEEE Trans. Nucl. Sci. NS-18, 55 (1971).5. R. D. Deslattes, T. J. Peterson, and D. H. Tomboulian, J. Opt. Soc.

Am. 53, 302 (1963).6. E. Hinnov, Phys. Rev. A 14, 1533 (1976).7. D. L. Lin, Johns Hopkins University; private communication.8. F. Paresce, S. Kumar, and C. S. Bowyer, Appl. Opt. 10, 1904

(1971).9. J. E. Mack, F. Paresce, and S. Bowyer, Appl. Opt. 15, 861

(1976).10. E. Hinnov, Phys. Rev. A 14, 1533 (1976).11. Equipe TFR, Nucl. Fusion 15, 1053 (1975).12. L. A. Hall and H. E. Hinteregger, J. Geophys. Res. 75, 6959

(1970).

0

1904 APPLIED OPTICS / Vol. 16, No. 7 / July 1977