283 - nasa · pdf filestecher (1969) and by bless and savage (1970) ... in order to have a...

283

PRELIMINARY RESULTS ON INTERSTELLAR REDDENING

AS DEDUCED FROM FILTER PHOTOMETRY

Michel Laget*National Aeronautics and Space AdministrationGoddard Space Flight CenterGreenbelt, Maryland

ABSTRACT

Filter photometry has been used to derive theinterstellar reddening law from stars through thestudy of a single spectral type, BO. The deficiencyin the far ultraviolet flux of a supergiant relativeto a main sequence star is compared with the differ-ence in the flux distribution due to a change of onespectral class. Individual interstellar reddeningcurves show the general feature reported by Stecher(1969) and by Bless and Savage (1970). There is alarge amount of scatter in the far ultraviolet whichmay be partially due to a real difference in inter-stellar extinction and partially due to observationalinaccuracy. The mean ratio E(m^-V)/E(B-V) computedfrom the color-color diagrams using a least -squaresfit to the reddening line shows an increasing proba-ble error from long wavelength to short wavelengthwhich cannot be explained by the expected meaninstrumental error.

I. INTRODUCTION

The study of a single spectral type with the set of broad-band photometers on board the OAO-2 satellite is an appropri-ate means to observe the interstellar extinction from faint

National Academy of Science Research Associate on leave fromthe Laboratoire d'Astronomie Spatiale, Les Trois Lues 13,Marseille (12e), France.

https://ntrs.nasa.gov/search.jsp?R=19720024180 2018-05-25T15:27:34+00:00Z

284 SCIENTIFIC RESULTS OF OAO-2

stars which cannot be observed by the spectrum scanners.This type of observation provides additional data to recentresults reported from relatively high resolution scans byStecher (1969) and by Bless and Savage (1970) and allows usa statistical approach.

In this paper preliminary results are given about the in-terstellar reddening deduced from the study of twelve starsclassified on the MK system as type BO. The visual magnituderanges from 1.70 to 6.65; the maximum difference in B-V is0.38. The choice of type BO for this study is compatible withthe following considerations: (i) a sufficient number ofstars, including faint stars, was available to the satellite,(ii) the differential influence within the spectral type BO ofthe strongest lines such as C IV and Si IV could be expectedto be negligible in comparison to the FWHM of the band passes(about 250 A), and (iii) problems of identifying correctly theeffective wavelength are minimized. The effective wavelengthsof the photometers are defined in terms of a uniform spectralintensity distribution.

II. THE OBSERVATIONAL DATA

The original signal for each star, in terms of number ofcounts, has been reduced in the way described by Code et al.(1970). In those cases in which the star was observed on morethan one orbit, the mean value has been taken. Generally, thedeviation was a few hundredths of a magnitude, but this remarkdoes not refer to the filters ST4F1 (1550 A) and ST4F3 (1430 A)which have been corrected for a change in transmission withtime. These corrections were made using information publishedby Holm (1972) and values derived from those stars observedseveral times in the course of several months. It appearsthat for filter ST4F3, in particular, the correction is largeand uncertain. Filters ST4F5 (1680 A) and ST4F4 (1330 A) havenot been used for the present work because of unreliable re-sults which are suspected in ST4F5 and because ST4F4 coversthe region of the La interstellar absorption line.

The ultraviolet extinction results are sensitive to the UBVcharacteristics of the star as well as to the spectral classi-fication. In order to have a homogeneous set of stars we haveused classifications by Lesh (1968), when available, or byHiltner, Garrison and Schild (1969). Color excesses have beenderived using the intrinsic colors of Johnson (1963). Thedata for all the stars used are summarized in Table 1, includ-ing the sources used. The observations are given in Table 2.Three stars are classified as BO.5 in the adopted sources al-though previously they were classified as BO. We shall assumethat this difference has a negligible effect on the calcula-tions presented in the next section.

INTERSTELLAR REDDENING FROM FILTER PHOTOMETRY 285

S-l

(g0)

mQJ

OH

CQ

^

ftcn

0)

ojp^

QS3

r]1 CQ

•rl Q)

(0•H >,rH JOj3O n30) oft-P

O

mM co-P 0)O-rt

0) -Pftcn

^ ^ ~5 5 5J H! A

co f» cr»CM rH H

o o o1 1 1

rH rH 0VD ^1* t"*

•^ ^ rH

•>n3

> m H•

o o oCQ (X) ££)

-H•H M -rlM O Mo o

l-H

P -9- U)

CM CM COrH CM CMin oo I-HVD VD r^co co co

rH03

-Po cnx!<U Q) CQft-H 0)cn -P 1-3

•rl

CQ <d ,Q coO -rl d

a) 3 <D 3O O 4J njO 0) O -PG ft £ fiO <Ua u

_<j**^Hi ffi ffi W K

CM rH rH CO inO rH rH O CM

O O O O O1 + 1 + 1

VO rH o O\ CMCO <3« CO 00 OO

in VD CM i1 CM

H >H (0 <flH H in H >

•O O O O OCQ CQ Q3 CQ CQ

0 M 00 O Ocn s cn

<0 3- H

cn m oo co<a< r~ r^ co coco oo CM o ^r^3* CM co cr* o^CO CM ^ "vf "3*^< H rH rH rH

HII

e

Mtdc•HCQ

• •COd0)r]

(d ftrl (Uf< CJ

^<J**-

W J

CO OO rH

O O1 +

VO rrin vo

m m

(dH

j>in

o0 CQ

CO ^<0\ rH00 CNO CMin orH CM

CO

II

^<

* CQ0) dH 0)ft .fj

•H ftM 0)

EM U

__^<]

-"

^

O OH CM

O O1 +

^ COCT> O>

in VD

QJrQ ft

H >

O OPQ CQ

CnJ>i

C_)

o^VD

CM COr~ r^H r^^ VDO OCM CM

CMII

B

Midfi

•HCQ

• •

CO3

(Ur*ft0)U

^<J.

^

r-o

o+

mVD

VD

^>H

OCQ

O">coCO<T>OCM

•"-

VD

d•

rH

-P0)

MCU

4JrH/•Hffi

a

.13

OMH

§MU

••

5

•^~.00VD

H

COd)Hi

• •

^

•dUrl

CPCQ£>O

CO

(d-PcnQ)x!E-i

a)rH

•8£-«


COCM o

EH ^CO rH

rHCM o_|J | A

^^* LI I

EH mW rH

rHCM o

EH enCO i-l

CMCM o

EH •*CO CM

rHCM oCM roEH 0CO ^

inCM oCM enEH coCO CM

CMCM oCM rEH cn,CO CM

•CM orH COEH es>CO CM

r-tCM orH CMEH roCOCO

roCM orH mEH CMCO •*

QEC

rocn

oI

XTen

o1

inCM

rH1

ro

CM1

<£>CM•

ro1

•>f•

ro1

envo•

mI

ID

•

CO1

cnCM

• •

ro1

00rH

Q

<3>1

CMrHinVOro

VQro

o1

00ro

oI

•«*00

o1

rHrH

CM1

CMCO

•

CM1

rHrH

*roI

m•

COI

voro

4ro1

CM•

ro1

00CM

m

•**1

CNCM00VDro

encn

CM1

roen

CM1

rHCO

roI

rq-0\

•*r„ „

COCNrH

ro

rorH

rH

iorH

rH

iOrH

rH

iVfo

o1

CO*>f

•o1

oo

*rH

OCO

•rH

1

^1un«rHjI

v>in

•

rH1

cnvo

0

CM1

"=3*ro

00

**

r-H

ro+

-*en

CM+

cn

CM+

«*00

o

in**

•o+

ino

*o

^fo

•rH

1

CMCO

•

O1

•<*en

t

oI

mro

tt

CM1

Or-COCMCMrH

roo

CM1

inCM

CM1

rHVO

CM1

CMcn

ro1

inr~CMro^3*r-H

VD

"*

rH+

enro

rH+

00CD

O+

<*to,

O1'

CM0}

•O

'

rH10

•rH

1

I»m,•CMili

•«*'cn

m

01t

CMr•

CM1

in•

ro1

COroOen'S*H

CO<&CN

1

,in

CMi

inCO

CMf

Coo

^r'I

COro^J*cn^frH

ro•V

rH+

"*•rH

+

^^

0

"*"

CMiE^

O-1,

,

CMi

• '

iHi

''

CMjC~-

•*rH

1.

tn:(iO

••

CMi11

\£»O

*

0»1

no

•

CM1

roo

*ro1

00cnCOomrH

o-*CM+

inr-H

OJ+

CO

rH,

+

00ro

O+

r-o

•

o1

cn^3"

••O

''

cmyptt

m

rH!li

rH'

^*•

rH1

inin

•

rH1

vo00

•

CM

'

•*rHCMCMOCM

inO

CM

+

f^

cn

r-H+

CDa*

o"*"eoTCM

oli

00CO

•

01

mCO>

•>

rH'

It

t>

cn,*i

rH'1!

coto*

rHt

rHr^•

rHi «

00r^•

CM1

CMr~rH*a*oCM

orH

ro+

in^m+

'*'ia>

CMI

+

en,o

r-H4-

10in

••

o.-f

o>CMI

•t

O'.H>

&n>p»-•

ot

P^u>•

0t

mvo

*o1

COCO

•

rH1

CJ\COroenoCM

£UJ

4J

13

U)

CM

<U•H

•8EH


III. THE COLOR-COLOR DIAGRAMS

Diagrams of (m^-V) vs. (B-V) for representative effectivewavelengths are shown in Figure .1. Mean reddening lines have

(mx-V) -8.0

-7.8

-7.6

-7.4

-7.2

<m x -V)

-6.4

-6.0

-5.6

-5.2

-4.8

-4.4

-4.0

-3.6

STI, Fl. A,,, * 3320<A)

-0.3 -0.2 -0.1 0.0 O.I 0.2(B-V)

ST3, Fl. X,,,, = 1910(A)

-0.3 -0.2 -0.1 0.0 0.1 0.2

(B-V)

(mx-V)

<mx-V)

-8.0

-7.8

-7.6

-7.4

- 7 2

-70

-6.8

-6.6

-6.4

-6.2

-5.6

-5.2

-4.8

-4.4

-4.0

-3.6

-3.2

-2.8

ST2. F5. Xe f l = 2390(A)

-0.3 -0.2 -0.1 0.0 0.1 0.2

(B-V)

ST4, Fl, \e(, = 1550(A)LUMINOSITY CLASS

• Vo IV

» IIIla

-0.3 -0.2 -0.1 00 0.1 0.2(B-V)

F-cguie 7 . — C o t o f i - c - o t o f i d-iaQiamt, &oi fiz.pn.ni> zntatsive.wavttzYiQthb . Note, the. -inc.ste.cu> e. oft tkn Ac.atte.1 astound the.fLH.dde.n4.nQ L-im wj.th d<Lc,n.£.at>-inQ wave.te.ngth.


been obtained by a least squares fit of a straight line to theobserved points. The slopes of the reddening lines and theirprobable errors are listed in Table 3 together with the stand-ard deviations. The scatter of the points is due to instru-mental errors, possible errors in V and in (B-V), differencesin the interstellar extinction and in the intrinsic propertiesof the stars themselves. The slope of the straight line rep-resents a mean ratio A(m^-V)/A(B-V) which can be used to esti-mate the ultraviolet extinction when the E(B-V) is known. Thestar HD 206773 (BO Vpe) has been left out of the calculations.

Table 3. Statistical Properties of theObserved Reddening Lines

Xeff(i)

4250

3320

2980

2940

2460

2390

2030

1910

1550

1430

Stand.Dev.

0.03

0.05

0.05

0.06

0.08

0.09

0.10

0.14

0.28

0.30

A(mx-V)

A (B-V)

0.81

1.63

2.36

2.41

4.03

4.62

5.23

4.71

4.96

5.29

p.e.

0.06

0.08

0.09

0.10

0.12

0.15

0.18

0.21

0.42

0.44

x-1

(u-1)

2.35

3.01

3.36

3.40

4.07

4.18

4.93

5.24

6.45

6.99

The large standard deviations obtained for the filters witheffective wavelengths shorter than 2000 A may be due to in-trinsic differences in the far ultraviolet spectra of the pro-gram stars. Bless and Savage have shown that for stars of thesame spectral type but luminosity classes in the range II toV, the spectral distribution is closely the same. In order toinvestigate what differences may appear in the filter photo-metry, differences in magnitude between a supergiant and a


main sequence star (e Ori, u Ori) and between two main se-quence stars (T Sco, u Ori) all of type BO are shown in Figure2. The differences are normalized to zero at the effectivewavelength of the V pass band. The difference between typesBO.5 V and BO V is shown by the pair (X Lep, v Ori). Each ofthese stars has E(B-V) < 0.05, thus differential reddening isminimal.

0.2

0.10.0

1.00

0.8

0.6

0.4

0.2

0.0

(rSco - uOri)

• (<0ri -

x (XLep-wOri)

_i_

2. — tUcLu tfi-ib at^io n

1500 2000 2500 3000 3500X(A)

-in magnitude.* ofi the. iLit>La.\) <io Lzt&p<Lc.tfiat type.* to

The supergiant shows a large ultraviolet deficiency in com-parison to a main sequence star. A similar deficiency wasfound by Carruthers (1969) at 1115 A for e Ori. Mihalas (1970)has commented upon this in terms of a decrease in the surfacegravity and in effective temperature from main sequence tosupergiant stars.

IV. THE WAVELENGTH DEPENDENCE OF THE INTERSTELLAR EXTINCTION

By comparing the spectral distribution of a reddened starwith that of an unreddened star of the same spectral type and


luminosity class, one can deduce the wavelength dependence ofthe interstellar extinction in the different directions of thestars. This method assumes that stars of the same spectraltype and luminosity class have identical flux distributions.

Figure 3 illustrates five reddening curves obtained by com-paring four main sequence stars to the reference star u Oriand one supergiant to e Ori. The data are normalized to adifference in (B-V) = 1 mag for the pairs of stars consideredand for AV = 0. The observed values of (f)1 Ori and 5 Sco havebeen corrected from BO.5 to BO using values given in Figure 2for the pair (A Lep, u Ori). The shadowed area represents theprobable error around the mean values computed in § III.

On the whole, the points show a maximum between 4.18 and4.93 u"1 and a minimum between 5.24 and 6.45 y"1. The profileof the maximum has been given by Stecher (1965,1969) and Blessand Savage (1970). The first explanations that the maximum is

E(mx-V)

E(B-V)

• HD202214* SSco+ HD48434o <£| Oria HD122879

uOrivOr\uOrj«0ri

fi.gu.ie. 3.—lnte.ute.llai ie.dde.ni.ng ^oand £01 HV 209339that {,01 r, Opk (Gllia 7977).


due to graphite particles were given by Stecher and Donn (1965)and by Wickramasinghe and Guillaume (1965).

The interstellar extinction law is known to vary in the vi-sible and in the infrared (Johnson 1968). Figure 3 shows forrepresentative pairs of stars that in the range 4.93 to 7 y1

the interstellar extinction seems to differ significantly fromthe average value. It is therefore of importance to find outwhether or not this scatter is a real effect which can be at-tributed to an intrinsic property of the interstellar medium.

For such a purpose we must note first that a probable errorof 0.1 mag in the difference between two stars is on the aver-age magnified by a factor of 4 because <A(B-V)> = 0.25, andsecondly, that points lying within 2 or 3 probable errors can-not be interpreted as basically deviant from the mean curve.

It follows that the case obtained for the star HD 48434 issignificantly peculiar. However, it cannot be considered tobe representative of a low interstellar extinction because ofan uncertainty, difficult to evaluate, in the technical reduc-tion of the far ultraviolet data.

The star HD 202214 is slightly above the mean curve but itcannot be considered to be different. This star is doublewith a companion 1 mag fainter than the primary component.

The stars (J)1 Ori and HD 122879 present an intermediate case.The uncertainties in the data in the far ultraviolet fall inthe area corresponding to 2 or 3 p.e., while in the region4.93 to 5.24 y a real deviation seems to be present, indi-cating that for these two cases the values of the interstellarextinction may be below the average.

The curve obtained for the star HD 209339, shown in Figure4, has a peculiar profile which deserves additional comments.Indeed, the values obtained are within 2 p.e., indicating thatthis profile is not statistically different from the mean.But apparently there is no minimum in the range 4.93 to 7 y"1.The star HD 209339 is double with a companion 0.6 arc sec dis-tant and 2 mag fainter. In addition, since this star is astandard BO IV star in the Lesh classification system it isunlikely that the observed differences in .the reddening curveare due to misclassification. It cannot be excluded that thewavelength dependence of the interstellar extinction in thedirection of HD 209339 might be unusual, presenting some simi-larities with the results found for C Oph by Bless and Savage(1970) .

V. DISCUSSION

In the case considered here, there are two ways in which todeduce the wavelength dependence of the interstellar reddeningprofiles:

1. Individual profiles deduced from the comparison of


ftmx-V) 4

E(B-V)

3

2

1

oo

o HD 209339 - uOri

• £0ph

1A (/i-1)

4 . — T h e wave.te.ngth de.pe.nde.nc.e. o £ the. £nte.u tzttax.obta<ine.d fanom paisu o& Atau and ^n.om a

me.th.od.

pairs of stars. In this case each curve is sensitive to theaccuracy of V, (B-V) and m^. The largest uncertainty is inAm^ between the two stars. The stars used generally have beenobserved at different epochs and it is difficult at present togive an upper limit to the expected accuracy at each time.

2. A statistical approach, considering all the stars as awhole, without any discrimination of luminosity class. Theleast squares fit to the straight line

mx-V = a(B-V) + b

is used at each wavelength. We can evaluate the error in m^from the mean error found by considering the results for sev-eral stars observed several times at different epochs.

We have pointed out in § IV that the first method gives pe-culiar results significantly different from the mean, and thatamong the sample of stars used it is difficult to evaluatewhether or not these differences indicate real differences inthe reddening law.

An attempt to understand the constant increase from long


wavelength to short wavelength of the standard deviation andof the probable error around the mean value found in case 2can be made by estimating the statistical errors. We shalltake the probable error of 0.42 computed for the 1550 A filteras representative.

At this step we must briefly review the hypotheses used:(1) All the stars have the same intrinsic spectral distri-

bution.(2) There is a linear relationship between (m\-V) and (B-V).

The slope is the same for any luminosity class, and is charac-teristic of the interstellar medium.

(3) The method of least squares fit can be applied fortwelve stars and the probable error of the derived slope isindependent of intrinsic differences between main sequencestars and supergiants. This is confirmed by the fact that theprobable errors computed separately in the sample case are re-spectively ±0.48 and ±0.33 for main sequence stars and super-giants .

Under these hypotheses, mean errors of ±0.1 mag in m\,±0.04 on the correction added to allow for the variation ofthe sensitivity with time, and ±0.02 on V and (B-V) give anexpected total probable mean instrumental error of ±0.11 magin the derived mean slope.

It turns out that the instrumental error cannot explain thetotal probable error found. We conclude that a "natural"scatter of the data exists which can be attributed to a localchange in the physical properties of the interstellar medium.

I am grateful to Prof. A. D. Code of the University of Wis-consin for the opportunity to make these observations with theOAO-2 as part of the Goddard Space Flight Center Guest Observ-er Program.

Also, I would like to thank Dr. A. B. Underhill, Dr. A.Boggess and Mr. L. Dunkelman for helping me in the completionof this work.

My thanks are extended to the National Academy of Sciencefor the grant of a Research Associateship tenable at GoddardSpace Flight Center.

REFERENCES

B£e-6.6, R. C. and Savage., B. P. 1970, I.A.U. Symp. No. 36, 28.Catiiu.the.tu, G. R. 7969, Ap. and Space Scu.. 5., 387.Code., A. V., Hoack, T. E., McWa££, J. F., 8£e44, R. C. and

ie., C. F. 1970, Ap. J. 161. 377.V. ? . 1 9 7 1 , Nature- 2 2 9 . 2 3 7 .

H-Lltne.fi, W. A., 6 atLtvi* o n, R. F. and Schi.ld, R. E. 1969 , Ap. J.7 5 7 . 313.


Holm, A. ( / . 7 9 7 2 , tkli> volume..Johnson , H. L . 7 9 6 3 , Bcu>£c. kAthonomJ,cal Data, nd . K. Aa.

Strand (Chicago: UnîvÛ-ity o& Ch.-Lca.go PA.e^^) , p . 2 0 4 .L z & h , J . R. 79,55, Ap. J . Supp£. 77. 3 7 7 .

P. 7 9 7 0 , Ap. and Space Sex.. £, 50.T. P. 7965 , Ap. J . (Letê^) 7 4 2 . 1 6 B 3 .T. P . and Vonn, 8 . 7 9 6 5 , Ap. J . [Le^te/w) 7 4 2 . 7 6 ^ 7 .T. P. 7 9 6 9 , Ap. J . (Lvt tzJU] HI, L 7 2 5 .

, M. C. and Gtu.££aume, C. 7 9 6 5 , NatuA.e 2 0 7 , 3 6 6 ,

283 - nasa · pdf filestecher (1969) and by bless and savage (1970) ... in order to have a...

Documents