PEPSIOS PURELY INTERFEROMETRIC HIGH-RESOLUTION SCANNING SPECTROMETER IV. PERFORMANCE OF THE PEPSIOS SPECTROMETER

(1)

HAL Id: jpa-00213241

https://hal.archives-ouvertes.fr/jpa-00213241

Submitted on 1 Jan 1967

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

PEPSIOS PURELY INTERFEROMETRIC

HIGH-RESOLUTION SCANNING SPECTROMETER

IV. PERFORMANCE OF THE PEPSIOS

SPECTROMETER

F. Roesler, J. Mack

To cite this version:

F. Roesler, J. Mack. PEPSIOS PURELY INTERFEROMETRIC HIGH-RESOLUTION

(2)

JOURNAL DE PHYSIQUE Colloque C 2, supplkment nu no 3-4, Tome 28, mars-avril1967, page C 2

-

313

PEP SIO S PURELY INTERFEROMETRIC HIGH-RESOLUTION SCANNING

SPECTROMETER IV. PERFORMANCE OF THE PEPSIOS SPECTROMETER

F. L. ROESLER and J. E. MACK

Department of Physics, University of Wisconsin, Madison, Wisconsin 53 706 U. S. A.

The work reported in this paper was supported in part by grants from the National Science Foundation, and in part by the Research Committee of the Graduate School of the University of Wisconsin.

Abstract. - The PEPSIOS spectrometer is an association of three Fabry-Perot interferometers in series with an interference filter. The parameters of the optical elements are chosen such that isolation of a single common peak is achieved. This system offers the Fabry-Perot advantage to studies of absorption spectra, and the improved instrumental profile suggests application to problems requiring high contrast and suppressed wings. PEPSIOS has been applied to several problems of astronomical importance with a high measure of success. These are briefly described from the instrumental point of view.

Rksumk. - Le spectrombtre PEPSIOS est une association de trois interfkrombtres Fabry-Wrot en serie avec un filtre interfkrentiel. Les parambtres du systbme optique sont choisis de telle faqon qu'un pic commun unique soit isole. Ce systbme permet d'appliquer les avantages du Fabry-Perot B I'etude des spectres d'absorption et les amkliorations du profil instrumental suggkrent comme applications des problemes exigeant un contrale Bleve et l'ilimination des pieds. Le PEPSIOS a etk applique avec succb B plusieurs problemes astronomiques importants, qui sont rapidement dtcrits du point de vue instrumental.

The basic principles of the PEPSIOS [I] spectrometer have been previously described. Essentially it is a scanning spectrometer consisting of an association of three Fabry-Perot etalons in series with an interference filter. The spacings of the Fabry-Perots and pass-band of the filter are chosen so that a single spectral element can be isolated with a high degree of purity from a continuous spectrum. Since the first published description of the PEPSIOS spectrometer, understanding of the characteristics and operation of the instrument has been increased through its success- ful application to a number of spectroscopic problems. In this paper a more detailed description of certain aspects of the instrument is presented, and several problems to which it has been applied are described from the instrumental point of view.

Among the important characteristics of a spectrometer that are necessary to judge its quality and its suitability for particular spectroscopic problems are its resolution, its instrumental profile, its luminosity, and its parasitic light.

Resolution. - The resolution of the PEPSIOS spectrometer is roughly that of the widest spaced etalon

in the series. However one gains a factor which may be as high as about two due to the multiplication of Airy ordinates. This effect is shown in figure 1 where the profiles of the etalons are shown separately and multiplied. The illustration is for etalons of equal finesse and having spacing rations equal to those calculated by McNutt [2] to be optimum for the conditions which he considered. These ratios are

I, : 1, : l, = 1.000 0 : 0.883 1 : 0.724 4. In this case the resolving power is increased by a factor of 1.7 over the resolving power expected from a single etalon having a spacing equal to the widest of the series. For the problems so far attempted, resolving powers of about 4 x lo5 are typical.

Instrumental Profile.

-

It is also seen from figure 1 that the instrumental profile is considerably improved over the case of a single etalon, both because of the more rapid fall-off of the wings and because of the low value obtained for the minimum transmittance. These improvements are achieved experimentally as shown in figure 2, which is a PEPSIOS recording of the 5 890

a

line of Na from a hollow cathode. Also drawn on the figure are a resonance profile correspon-

(3)

C 2

-

314 F. L. ROESLER AND J. E. MACK

I " " l " " 1 " ' ~ structure, and the experimental profile obtained from

-

a double Fabry-Perot scanning spectrometer with

which we have had considerable experience [3]. In each case the advantage to the profile of adding eta- Ions is clear. These advantages are of great importance for studying faint components in the neighborhood of intense ones, for studying line profiles, and for reducing the parasitic light defined below.

The calculated curves of figure 1 assume that reflections between the etalons can be neglected. In practice reflections are suppressed by slightly tilting the etalons and by reflection losses from the windows between etalon chambers and the backs of the etalon plates. The experimental profile is intermediate between the case of no suppression and complete suppression of the reflections.

The problem of ghosts which arise from incomplete suppression of transmittance peaks of the individual etalons is one which seems a priori quite bothersome.

0 I 2 3 However both experimental tests and calculations

r r - r r show that the ghosts are or less everywhere ;

- 0 - P

moreover their positions may be accurately determi-

FIG. 1. - Product of three Airy functions. The curves A1,

A2, A3 show the Airy function for three interferometers with ned. Ghost suppression is clearly better for PEPSlOS spacers in the ratios 1.000 : 0.883 1 : 0.724 4. Their product than for the double etalon system.

AI A2 A3 has a half-maximum width smaller than the width of _Figure₂_{also shows that there is a small continuous}

the interferomcter with the largest spacer by the factor 1.7. The

product function is compared with Gaussian and Airy furctions background of the order of times the peak trans-

(G and A) with the same half-maximum widths. Wavenumbers mittance. This arises in part because the etalons trans-

are given in units of the half-maximum width of the widest spa- mit everywhere, and in part from scattered light. It

ced interferometer. _{is important to investigate the contribution of this}

background as well as the ghost contribution to the ding roughly to the Airy profile of a single etalon total light reaching the detector. This is especially

having the same half-width as that obtained for important for studies of continuous spectra where

PEPSIOS after taking account of the Na hyperfine even a small contribution can become an important

FIG. 2. - Experimentally recorded PEPSIOS instrumzntal profile. The 5 890 I% line of sodium was used t o obtain this

(4)

PEPSIOS PURELY INTERFEROMETRIC HIGH-RESOLUTION SCANNING C 2

-

315

fraction of the total light reaching the detector when integrated over an extended spectral range.

Parasitic light.

-

The quantity of light reaching the detector from a white light source due to transmittance

outside the main peak is called the parasitic light, and is expressed as a fraction of the light reaching the detector from the main passband. Theoretical calculations of the parasitic light have been performed by D. P. Mc Nutt [2] and J. 0. Stoner, Jr. [4] for sets of conditions which are thought to be realistic. The calculations, which were performed to find optimum spacing ratios using minimization of the parasitic light as a criterion are in qualitative agreement with each other and predict satisfactory values of the parasitic light. These calculations omit inter-reflections, and the actual experimental results appear to be intermediate between the case of no reflections and the case including all reflections [I]. A direct measurement

of the parasitic light as defined above is difficult, but an approximate measure is obtained by tracing a Na absorption cell fully blocking in a narrow region around each resonance line. Such a trace is shown in figure 3, and the value of 4

%

obtained is taken as quite satisfactory. The work by Daehler 151, to be

FIG. 3. - Parasitic light in the PEPSIOS spectrometer. This

figure shows the absorption of the 5 890 d Iine of Na from the continuum by a fully blocking sodium vapor cell recorded by PEPSIOS. The interference filter in this case was 8 d wide and had a resonance profile. The figure of 4 % can be expected to improve with a better interference filter.

mentioned later, tends to confirm the values for the parasitic light.

The value of the parasitic light can be expected to depend strongly on the instrumental parameters. This is illustrated in figure 4 which was prepared by

FABRY-PEROT FINESSE N

FIG. 4. - Dependance of parasitic light on filter and etalon

properties. The parasitic light for the three-interferometer PEPSIOS is shown as a function of Fabry-Perot finesse and the ratio of Fabry-Perot resolution to interference filter resolution. The points plotted are those calculated by Mc Nutt.

Daehler using the results of Mc Nutt's calculations. It is seen that the parasitic light is an important function of the etalon finesse and the ratio of the interference filter width to the PEPSIOS resolving limit. The parasitic light also depends upon the shape of the filter passband. The results of figure 4 assume a resonance shape, but filters with much better shape exist and give lower values for the parasitic light. Although the calculations cannot easily be confirmed experimentally, they serve as a guide to finding the optimum parameters for PEPSIOS.

(5)

C 2

-

316 F. L. ROESLER AND J. E. MACK

to measure directly the transmittance of PEPSIOS. However, for appropriate problems comparisons with other instruments have been made both directly and indirectly and in every case PEPSIOS has produced higher quality spectra. The comparison, of course, depends upon the conditions under with the comparison is made. If one makes the simple comparison between a PEPSIOS and a single etalon spectrometer looking at a source containing a single line, the result is that at a given resolution the transmittance of the polyetalon is lowered considerably by the product of the transmittances of the individual etalons. However, in a more complicated problem one must also consider the means of eliminating spectral elements which must be rejected before the desired elements may be studied unambiguously. The principal advantage of PEPSIOS arises when it is necessary to isolate a single spectral element from a very dense or continuous light source. It then offers the high luminosity-resolution product characteristic of Fabry-Perot spectrometers without the need of an excessively large premonochro- mator. For this reason our initial applications of PEPSIOS have been to astronomical problems requiring high precision measurements of absorption features. However, it should be pointed out that the high contrast and greatly suppressed wing of the instrumental profile make PEPSIOS attractive for profil measurements and studies of weak components in the neighborhood of intense ones.

A study of solar lithium.

-

The first problem to be considered illustrates the use of PEPSIOS for precise photometric spectroscopy. The problem studied by Daehler [5] involved the measurement of the profile of the solar lithium feature at the lithium resonance wavelength 6 708

A

in an attempt to determine the solar lithium abundance and isotope ratio. The problems required a relative precision of at least at the level of the solar continuum. The general problem is indicated in figure 5, which shows the supposed

FIG. 5.

-

Grating trace of the solar spectrum in the region aroung 6 708 4. The small box indicates the area included in figure 6, below.

lithium absorption as about 1

%

of the continuum. The instrumental arrangement used by Daehler for this problem is shown in figure 6 and is illustrative of the

RESOLUTlON MEASURING INTERFEROMETER To P R E S S U R ~

1

SCANNlllb SYSTEM INCANDESCENT TUNGSTEN LAMP CALIBRATION SPECTROMETER

"'

HaA8 ELECTRODELESS DISCHARGE TUBE

FIG. 6. - Instrumental arrangement for the solar lithium

experiment. The purpose of the various units is explained in the text. This basic arrangement, with modifications appropriate to a particular problem, has been used in all PEPSIOS studies.

(6)

PEPSIOS PURELY INTERFEROMETRIC HIGH-RESOLUTION SCANNING C 2 - 317

The alignment method used by Daehler is a variation of a method developed by Burnett 161, and consisted of carefully adjusting the forces on the etalon adjustment springs to achieve simultaneous parallelism and exact spacing of the etalons as verified by the equality of the scanning pressure at which the peak of the line from the Li lamp was found. In this case the three etalons scan together without pressure differences. The Li lamp was also used to fix the wavelength scale of the reference fringes.

One of the particularly troublesome problems which had to be oversome was the appearance of spurious fringes in traces of a continuous light source. An example of spurious fringes is shown in figure 7. It

FIG. 7. - An example of spurious fringes. This is a recording of the continuous spectrum from an incandescent tungsten lamp. The variations above the noise are generated by the PEPSIOS instrument under certain conditions. Their origin and elimina- tion are discussed in the text. The ampli tude of the fringes shown

is about 0.1 % of the continuum level.

was found that the most serious fringes, as high as 1

%,

arose from several plane-parallel elements in the optical path. When all windows and etalon plates were wedged there still remained the fringes of about 0.1

%

shown in the figure, and these were found to disappear when the etalons were slightly tilted off-axis.

The typicaI raw data obtained are shown in figure 8. A is a trace of the sun and B is a trace of an incan-

FIG. 8. - Typical data. A. Solar spectrum. B. Tungsten lamp

scan, giving interference filter profile. C. Spectrum of the lithium resonance lines from a hollow cathode lamp. D. Channel spectrum for measuring resolution. The uniform fringes at the bottom of each trace are the reference fringes.

descent tungsten light source verifiyng the absence of spurious fringes and giving the filter profile correction which was necessary because the filter was not scanned along with the etalons. One large scale division of A or B represents about 1

%

of the continuum. C and D illustrate fixing the wavelength scale of the reference fringes, and determining instrumental resolution, respectively. The data so obtained were corrected and averaged to provide a spectrum, shown in figure 9,

with an rrns deviation of less than 1.4 x Overall

accuracy was limited by amplifier gain drifts and by interference filter changes which are not fully under- stood. Careful comparison with the positions and sturcture of the Iithium resonance lines leads to the conclusion that the structure can not be simply attributed lithium.

(7)

E. L. ROESLER AND J. E. MACK WAVELENGTH

6707.0 67075 67080 67085

I . ' I I I I 1 I 4

FIG. 9.

-

Solar spectrumnear 6 708 A. This final spectrum is the average of seven PEPSIOS traces. The laboratory

positions of the lithium resonance lines are shown, with the terrestrial 6Li/7Li abundance ratio. The spectrum has been corrected for doppler and gravitational shifts. The rms deviation is less than 1.4 x

100 99

-

> 9 8 - V) z W I- 97

-

5 96

-

95

-

. ,

.

, t5 0 -5 -10

-

15 -20 -25

RECESSION APPROACH RADIAL V E L O C I T Y (km/se

94

FIG. 10. - Interstellar absorption in the light from cr Cygni in the neighborhood of the 5 890 A line of sodium. Ordinate :

intensity, arbitrary scale with true zero indicated crudely by the dashed line. This differs from the zero of the chart paper because of parasitic light and light contributed by the day sky. Abscissa : wavenumber intervals relative to the position of the laboratory sodium lamp reference at the time of the experiment. The scan time was about 25 min (From Ref. 7).

-

I I I I I I I I 1 1 1 1 1 1 1 1 1 l 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

149050 149040 149030 149020

(8)

PEPSIOS PURELY INTERFEROMETRIC HIGH-RESOLUTION SCANNING C 2

-

319

achieved with the PEPSIOS spectrometer were perfor- _0.10

med by L. M. Hobbs [7]. These studies were an investigation of interstellar sodium absorption lines, and a search for sodium absorption on Venus. The basic instrumental details were essentially those previously described. The principal differences were that the sources in each case had a very small angular diameter, roughly 12 seconds for Venus and a seeing- limited value of several seconds for a star, and were

much fainter than the sun. For these problems a large 0.05

telescope was required in order to properly match the etendue of the PEPSIOS. An example of the interstellar absorption results is shown in figure 10. Hobbs has recently obtained useful interstellar absorption results from fifth magnitude stars using the 120 inch telescope at Lick Observatory. An example of the venusian sodium measurements is shown in figure 11. It is

FIG. 11. - Spectrum of the day sky, left, and day sky plus Venus, right, in the neighborhood of the D2 line at 5 890 A. The interval scanned is about 2.7 K, and the scan time is about 10 min.

worth noting that the Venus measurements were taken during daylight, and scans of day sky and day sky- plus-Venus were necessary to obtain the Venus results. This also illustrates the potential application of PEPSIOS to day sky measurements.

Telluric sodium measurements.

-

As a final example, figure 12 shows the bottom few per cent of the solar

D 2

line at 5 890 with small dimples arising from absorption by terrestrial sodium [ 8 ] [9]. These measurements are being conducted as a continuing

MADISON, WISCONSIN, U.S.A. APRIL 8,1964 IT17 CST

I

-

RED I O O ~ K

FIG. 12. - The bottom of the Fraunhofer D1 absorption line of the solar spectrum. The terrestrial sodium is identified by its position relative to the solar line, which has a gravitational red shift plus a Doppler shift, and by resolution of the two hyperfine- structure blends which are observable by virtue of the low pressure and temperature of the upper atmosphere. Ordinate : intensity (continuum E 1.00) ; Abscissa : wavenumber (50 mk per division).

program for atmospheric sodium abundance measurements at Madison. Instrumental details are similar to those already discussed.

(9)

C 2

-

320 F. L. ROESLER AND J. E. MACK force of the adjustment springs, and the scanning

accomplished by pressure changes in a common chamber. For the well known method of polyetalon scanning with constant pressure differences we have sought to achieve a precision of a few hundredths of a torr. Although we have operated a conventional double-etalon scanning system for many years with a precision of 0.1 torr [3], the higher precision desired for PEPSIOS has just now been achieved, and the

scanning system is undergoing tests on the PEPSIOS

instrument for sodium absorption measurements. The sensing method is the same as described in the original

PEPSIOS paper, i. e. pressure difference mercury

manometers with a differential transformer core floa- ting in one arm to provide error signals, but now the pressure differences are maintained by a servo control- led pressure difference regulator.

Acknowledgements.

-

We are especially indebted t o C. R. Burnett, Mark Daehler, and L. M. Hobbs for their excellent work in applying the PEPSIOS

spectrometer. In addition, it is a pleasure to acknow- ledge the contributions of many students and depart- mental shop personnel t o various phases of work on this instrument. A word of thanks is due to the Re- search Committee of the University of Wisconsin for making it possible to present this report.

[I] MACK (J. E.), Mc NUTT (D. P.), ROESLER (F. L.) and CHABBAL (R.), AppI. Opt., 1963,2, 873.

[2] Mc NUTT (D. P.), J. Opt. SOC. Amer., 1965, 55, 288. [3] ROESLER (F. L.), PJz. D. Thesis, University of Wisconsin,

1962.

[4] STONER (J. O.), JY. J. Opt. SOC. Amer., 1966,56, 370.

151 DAEHLER Mark, Ph. D. Thesis, University of Wisconsin,

1966. To be published, in part, in the near future.

[6] BURNETT (C. R.), Private communication.

[7] HOBBS (L. M.), Ph. D. Thesis, University of Wisconsin,

1966 ; Ap. J., 1965, 142, 160. An additional publication is in preparation.

[8] BURNETT (C. R.), Science, 1965,147,736.

[9] Mc NUTT (D. P.) and MACK (J. E.), J. Geophys. Res.,