1

STRUCTURE AND EVOLUTION OF NEARBY OB ASSOCIATIONS

1 1 1 1 1 1 1;2

P.T. de Zeeuw , A.G.A. Brown , J.H.J. de Bruijne , R. Ho ogerwerf , J. Lub , R.S. Le Po ole , A. Blaauw

1

Sterrewacht Leiden, P.O. Box 9513, 2300 RA Leiden, The Netherlands

2

Kapteyn Astronomical Institute, P.O. Box 800, 9700 AV Groningen, The Netherlands

The study of OB asso ciations is also imp ortant in ABSTRACT

the context of the evolution of the Galaxy. The kine-

matics of the nearest OB asso ciations provides in-

sightinto the prop erties and origin of the Gould Belt

We present the rst results of a comprehensive cen-

system. Furthermore, OB asso ciations are resp onsi-

sus of the stellar content of the nearby OB asso cia-

ble for large bubbles in the interstellar medium lled

tions based on Hipparcos p ositions, prop er motions

with hot gas e.g., Mac Low & McCray 1988. More

and parallaxes for 12842 candidate memb er stars dis-

detailed knowledge of their stellar content is imp or-

tributed over 21 elds on the sky. We use a new

tant for understanding the energetics and dynamical

metho d to identify moving groups in these elds see

evolution of the bubbles. Ultimately, establishing the

de Bruijne et al., these pro ceedings. Previously, as-

prop erties of the nearby asso ciations is a prerequisite

trometric memb ership in nearly all the nearby OB

for the interpretation of observations of extragalactic

asso ciations was known only for stars with sp ectral

star forming regions and starburst galaxies.

typ es earlier than B5. The Hipparcos measurements

now allow us to identify memb ers down to late F.

This census provides a rm basis for studies of galac-

The Solar neighb ourho o d contains a number of OB

tic and extragalactic star forming regions.

asso ciations, so that detailed studies are p ossible.

The most reliable memb ership determinations for OB

asso ciations are based on prop er motion studies. Al-

though these groups are unb ound, their expansion

Key words: OB asso ciations; HR diagram; luminos-

1

velo cities are only a few km s e.g., Mathieu 1986,

ity calibration; stars: early-typ e

so that the common space motion is p erceived as a

motion of the memb ers towards a convergent p oint

on the sky e.g., Blaauw 1946; Bertiau 1958. Mem-

b ership determinations have b een carried out byvar-

1. INTRODUCTION

ious investigators see Blaauw 1964, 1991 and refer-

ences therein. Due to their large extent on the sky,

OB asso ciations are not amenable to prop er motion

Ever since sp ectral classi cations for the bright stars

studies with photographic plates. Instead, one had

b ecame available it was evident that O and B stars

to rely on prop er motions from large scale surveys

are not distributed randomly on the sky, but in-

with meridian circles. This resulted in much uncer-

stead are concentrated in lo ose groups Blaauw 1964

taintyonmemb ership of stars of sp ectral typ e later

and references therein. Ambartsumian 1947 found

than B5. As a result, our knowledge of these young

that the stellar mass density in these groups, which

stellar groups remained rather limited.

were subsequently called OB asso ciations, is usu-

3

ally less than 0:1M pc . Bok 1934 had already

In order to remedy this problem the SPECTER con-

shown that suchlow-density stellar groups are unsta-

sortium was formed in Leiden in 1982. It success-

ble against Galactic tidal forces, so that the observed

fully prop osed the observation by Hipparcos of can-

OB asso ciations must be young, a conclusion sup-

didate memb ers of nearby OB asso ciations. An ex-

p orted by the ages derived from Hertzsprung{Russell

tensive program of ground-based observations was

diagrams. These groups are prime sites for the study

carried out in anticipation of the release of the Hip-

of star formation pro cesses and of the interaction of

parcos data see de Zeeuw et al. 1994. This in-

early-typ e stars with the interstellar medium see,

cluded Walraven V B LU W  photometry de Geus

e.g., Blaauw 1964, 1991 for reviews. Detailed knowl-

et al. 1989; Brown et al. 1994, mm and radio ob-

edge of the stellar content and structure of OB asso-

servations of the interstellar medium surrounding the

ciations allows us to address fundamental questions

asso ciations de Geus 1992; Brown et al. 1995, and

on the formation of stars in giant molecular clouds.

sp ectroscopy aimed at obtaining precise radial and

What is the initial mass function? What are the char-

rotational velo cities Verschueren et al. 1997; Brown

acteristics of the initial binary p opulation? What

&Verschueren 1997. Theoretical work included the

is the star formation eciency? Do all stars in a

distinguishing features of the formation of b ound and

group form at the same time? What pro cess causes

unb ound groups Verschueren & David 1989, and

the distinction b etween the formation of b ound op en

the reliability of the so-called kinematic ages of OB

clusters and unb ound asso ciations? How is angular

asso ciations Brown et al. 1997a.

momentum redistributed during star formation? oral05-06

2

Figure1. Proper motions for 532 members of the Sco OB2 association selectedfrom 4156 candidate stars in our Hipparcos

sample for the areabounded by the dashed lines. The dotted lines are the schematic boundaries between the three classical

subgroups Upper Scorpius, Upper Centaurus Lupus, and Lower Centaurus Crux.

prop er motion memb ers, and 3 of the 10 additional Here we present the preliminary results of the

photometric memb ers. We have added another 148 SPECTER census of the nearby asso ciations based on

memb ers, including one M giant. Figure 1 illustrates the Hipparcos prop er motions and trigonometric par-

the motions in the three elds, and shows that assign- allaxes. Wehave develop ed a new pro cedure to iden-

ing b oundaries to the subgroups is not trivial. The tify moving groups in the Hipparcos Catalog de

numb ers given ab ove are based on the preliminary Bruijne et al. 1997, and have applied it to 21 elds

b oundaries used by de Zeeuw et al. 1994. Analy- for which we have Hipparcos data. The elds are

sis of the entire Hipparcos Catalog, and inclusion of centered on 18 known asso ciations and 3 susp ected

radial velo city data, is needed to determine the full groups, all within 800 p c from the Sun. A list of the

extent of Sco OB2, and the division in subgroups. elds and their b oundaries, with the numberofob-

jects, can be found in de Zeeuw et al. 1994. The

dataset also includes 153 candidate runaway OB stars

Prior to the Hipparcos mission, we collected precise

as well as 49 stars from the disp ersed Cas{Tau as-

intermediate band Walraven V B LU W photometry

so ciation. Below we discuss a few examples of the

for 2243 of the 4156 Hipparcos stars in the three Sco

results obtained so far. Other examples are given in

OB2 elds de Geus et al. 1989, 1990. We have

de Bruijne et al. 1997 and Ho ogerwerf et al. 1997.

measurements for 276 of the total of 532 Hipparcos

memb ers. Figure 2 shows the Walraven [B U ]versus

[B L] diagram for the previously susp ected memb ers

top and the Hipparcos memb ers b ottom. Figure 2

2. SCORPIO{CENTAURUS

shows that the Hipparcos measurements allow identi-

cation of new memb ers to much later sp ectral typ es,

all the way to the mid F regime! The removal of inter-

The Sco OB2, or Scorpio{Centaurus, asso ciation con-

lop ers in the pre-Hipparcos memb ership lists results

tains three classical subgroups which di er in age

inamuch tighter correlation near the S-turn of the

Blaauw 1946, Blaauw 1964; de Geus et al. 1989:

main sequence around [B L]; [B U ]  0:2; 0:4.

Upp er Scorpius, Upp er Centaurus Lupus, and Lower

Stellar evolution moves stars towards the lower left:

Centaurus Crux. Our memb ership selection for Up-

this causes the slightdownward curve at the top of

p er Scorpius is describ ed in detail by de Bruijne et al.

the main sequence. The remaining spread is caused

1997. We have similarly analysed the other two

by i p eculiar sp ectra with, e.g., emission features

elds. For Upp er Centaurus Lupus 19 of the 29 clas-

cf. 48 Lib and Oph, ii undetected duplic-

sical prop er motion memb ers, and 30 of the 101 pre-

ity/multiplicity; and iii stellar rotation: high values

viously known photometric memb ers were selected.

of v sin i move stars towards the lower left. This e ect

Wehave discovered another 139 memb ers. For Lower

< <

is strongest for B7 { B9 stars 0:05 [B L] 0:13.

Centaurus Crux we con rm 15 of the classical 22

  oral05-06 3

-0.2

0

0.2 log Teff

0.4 log g 0.6

0

0.2

0.4

0.6

0 0.1 0.2 0.3 0 0.1 0.2 0.3 0 0.1 0.2 0.3

Figure 2. De-reddened Walraven colour-colour diagrams for the three classical subgroups of Sco OB2. a 84/91 pre-

Hipparcos `members' of Upper Scorpius top and 106/178 new members bottom for which we have Walraven photometry.

Speci c stars are indicated. b Idem for Upper Centaurus Lupus for 124/130 pre-Hipparcos `members' and 106/188 new

members, and c for Lower Centaurus Crux for 34/42 pre-Hipparcos `members' and 64/166 new members. The symbols

in the top panels have the fol lowing meaning: solid dots are the classical proper motion members, the open circles are

pre-Hipparcos probable proper motion members and the crosses indicate previously suggested photometric members. The

arrows in b indicate the approximate directions of increasing log T and log g .

e

The new memb ership lists can b e used to derive re- 3. VELA OB2

liable mass functions to much smaller masses than

p ossible previously. This requires multi-colour pho-

tometry for all memb ers, as well as a careful cor-

Brandt et al. 1971 noted the presence of 17 bright

rection for incompleteness in the Hipparcos Catalog

early-typ e stars within a few degrees of the Wolf{

2

at faint magnitudes. Application of our memb ership

Rayet binary system Velorum. They to ok the sim-

selection pro cedure to simulated moving groups su-

ilar distance mo duli for 10 of these stars including

p erimp osed on the Galactic eld p opulation shows

2

the Vel system as evidence for an OB asso ciation

that the exp ected numberofinterlop ers is insigni -

at  460 p c: Vela OB2. Straka 1973 investigated

cant for sp ectral typ es earlier than  F5. De Geus et

SAO prop er motions and Bright Star Catalog radial

al. 1989 derived ages of  5 Myr for Upp er Scorpius,

velo cities for these 10 stars. Only 5 of them, including

 10 Myr for Lower Centaurus Crux and  13 Myr

1 2

the visible multiple system and Vel, turned out

for Upp er Centaurus Lupus. These values are not

to share a common space motion. Photometric evi-

exp ected to change much, as most of the previously

dence for the existence of an asso ciation at a distance

known bright memb ers have b een con rmed by our

of ab out 450 p c was provided by several large-scale

selection pro cedure. The stars b eyond sp ectral typ e

studies e.g., Upton 1971; Straka 1973; Eggen 1986.

Fmay not yet have had time to reach the main se-

quence, and hence a detailed analysis of memb ership

Our Hipparcos sample for Vela OB2 contains 510

for these late-typ e stars will b e very interesting.

stars of sp ectral typ es O 9 and B 501 in the eld

   

255  `  270 and 20  b  5 . The magnitude

limits of our sample are V  9:0 for the O and B0{

B5 stars, and V  10:0 for B6{B9 stars. Only 4 of

the 10 Brandt et al. stars are con rmed as memb er.

1

Of the 6 remaining stars, Vel was not observed by

Hipparcos, and the other 5 are rejected as memb ers.

However, our memb ership selection pro cedure iden- oral05-06 4

100

50

00 ) 0

-50 Galactic latitude (

-100

-150

-200 2750 2700 2650 2600 2550 2500

Galactic longitude ( 0)

Figure3. Proper motions in the Vela OB2 association superimposedonagrey-scale representation of the IRAS 100 m

sky ux. The white dots show the Vela OB2 members and the black dots show the members of Trumpler 10 locatedata

distanceof362  20 pc. The core of the Vela OB2 association is surrounded by the IRAS Vela shel l.

ti es 112 new memb ers for Vela OB2, which brings iii Vela OB2 has an age of 20 Myr, and she showed

2

that then the observed kinetic energy of the IRAS

the total numb er to 116: 1 O9V star, Vel, and 114

Vela shell is of the same order of magnitude as the

B-typ e stars!

total amount of energy that the stars have injected

into the interstellar medium through the combined

We nd a mean distance of d = 415  10 p c for Vela

e ects of stellar winds and sup ernovae. Now that the

OB2. The intrinsic depth of the asso ciation cannot

present stellar content and the distance of Vela OB2

b e resolved from the parallax distribution. The new

have b een determined, a careful multi-colour photo-

memb ers of Vela OB2 are concentrated on the sky

metric study will allow a considerable re nementof

  

around `; b = 263 ; 7  within a radius of  5 .

Sahu's analysis. The Hipparcos photometry already

Sahu 1992 rep orted the detection of the so-called

indicates that manyof the Vela OB2 member stars

IRAS Vela shell in the IRAS Sky Survey Atlas maps.

are evolved.

This is an expanding shell, centered on Vela OB2,



with a pro jected radius of  8 Figure 3. Sahu as-

sumed that i the center of the IRAS Vela shell has The Vela OB2 eld contains a numb er of other mov-

a distance of 460 p c, ii Vela OB2 is a `standard as- ing groups. One example, Trumpler 10, is illustrated

so ciation' with a `normal' initial mass function, and in Figure 3. oral05-06

5

4. OTHER GROUPS b ership lists, esp ecially for the more distant asso ci-

ations. However, we susp ect that the calibration of

the upp er main sequence in the Hertzsprung{Russell

diagram may need revision.

Wehave also detected moving groups in the elds of

Per OB2 Blaauw 1950, Lac OB1 Blaauw & Morgan

1953, and Cep OB3 Garmany 1973, and have con-

Table 1. Distances to OB associations. The rst column

rmed the reality of the highly disp ersed asso ciation

lists the association and the second column the number

Cas{Tau Blaauw 1956. The asso ciations Persei

of members identi ed from our Hipparcos sample. The

Per OB3 and Collinder 121, as well as a new group

third column contains the pre-Hipparcos estimate of the

discovered in the eld of Cep OB2, are discussed by

distance to these associations and the last column lists

Ho ogerwerf et al. 1997. The Orion OB1 asso cia-

tion has b een studied extensively, mostly by photo-

the distance derived from the mean Hipparcos paral lax.

metric means e.g., Warren & Hesser 1978; Brown

The errors on the Hipparcos distances correspond to the

et al. 1994. It lies at a distance of  380 p c, near

errors in the mean paral lax.

the direction of the Solar antap ex. As a result, it is

dicult to distinguish its memb ers from the general

Name N D p c D p c

cl

Galactic disk p opulation based only on the Hippar-

cos prop er motions and parallaxes. We will rep ort on

Upp er Scorpius 178 160 145  2

this interesting asso ciation elsewhere.

Upp er Centaurus Lupus 188 145 140  2

Lower Centaurus Crux 166 120 118  2

Our Hipparcos sample also includes 153 candidate

Persei/Perseus OB3 87 170 176  5

OB runaway stars. The study of these stars is of in-

Perseus OB2 26 360 305  25

terest for settling the question of their origin; sup er-

Trumpler 10 21 400 362  20

nova explosions in high mass binaries Blaauw 1961

Vela OB2 116 450 415  10

or dynamical ejection e.g., Gies & Bolton 1986. We

Lacerta OB1a 45 530 382  25

are in the pro cess of retracing the paths of the run-

Lacerta OB1b 24 530 364  25

aways and the OB asso ciations in the Galactic po-

Collinder 121 105 760 546  30

tential in order to identify the parent asso ciations, as

Cepheus OB2 49 790 559  30

well as the age of the runaways. This is of consider-

able interest for studies of high mass binary evolution

e.g., van Rensb ergen et al. 1996 and of high mass

X-ray binaries e.g., Kap er et al. 1997.

5. DISTANCES

Wehave used the Hipparcos parallaxes for the mem-

b ers in the asso ciations to determine their mean dis-

tances. This requires some care, as the inverse of the

parallax is a biased distance indicator Brown et al.

1997b. As there is no evidence for a strong central

concentration of stars in our asso ciations, we assume

that to rst order the memb ers are distributed homo-

geneously in a sphere. In this case the exp ectation

value of the mean of the measured parallaxes is equal

to the true parallax, and corresp onds to the true dis-

tance of the asso ciation. The resulting distances to

the asso ciations are presented in Table 1. The er-

rors are derived from the errors in the mean paral-

laxes. Figure 4 shows that the new distances to the

asso ciations are systematically smaller than previous

estimates whichwere based mostly on photometry.

Due to the selection e ects in the Hipparcos Cata-

log, the observed distribution of parallaxes may not

Figure4. Distances derivedfrom the mean paral lax of the

b e representative of the true underlying parallax dis-

Hipparcos members of OB associations versus previous

tribution of an asso ciation. For example, magnitude

best estimates of the distance to these associations.

limits will bias the sample towards the stars closest

to us, and therefore could be resp onsible for part

of  the discrepancy evident in Figure 4. For this rea-

6. CONCLUDING REMARKS son wehave carried out Monte Carlo simulations of

the distance determination of spherical asso ciations,

which takeinto account the luminosity function, the

Application of our memb ership selection pro cedure to Hipparcos selection in a crude approximation and

Hipparcos parallaxes and prop er motions for 12842 the error on the parallax as a function of apparent

candidate stars in 21 elds with previously known magnitude. We nd that the magnitude limit bias

or susp ected OB asso ciations has signi cantly im- is small, and conclude that the results presented in

proved our knowledge of their stellar content. The Figure 4 are robust. Part of the di erence in dis-

numb er of rmly established kinematic memb ers has tances is no doubt due to the greatly improved mem- oral05-06

6

increased dramatically. The memb ership lists for the Brown A.G.A., de Geus E.J., de ZeeuwP.T., 1994,

early sp ectral typ es have b een re ned, and many new A&A 289, 101

memb ers have b een found of later sp ectral typ e, into

Brown A.G.A., Hartmann D., Burton W.B., 1995,

and in some cases b eyond the regime of the F stars,

A&A 300, 903

which is the range where the stars in these young

Brown A.G.A., Dekker G., de Zeeuw P.T., 1997a,

groups are still in their pre-main sequence phase.

MNRAS 285, 479

The measured distances for the asso ciations are sys-

Brown A.G.A., Arenou F., van Leeuwen F., Linde-

tematically smaller than indicated by previous pho-

gren L., Luri X., 1997b, these pro ceedings, ESA-

tometric determinations. Whereas part of this e ect

SP 402

must b e caused by the much improved memb ership

de Bruijne J.H.J., Ho ogerwerf R., Brown A.G.A.,

lists, esp ecially for the more distant groups, we sus-

Aguilar L.A., de ZeeuwP.T., 1997, these pro ceed-

p ect that a recalibration is needed of the upp er main

ings, ESA-SP 402 P05.10

sequence in the Hertzsprung{Russell diagram.

Eggen O.J., 1986, AJ 92, 1074

The next step is to re-examine our 21 elds, and the

Garmany C.D., 1973, AJ 78, 185

adjacent regions, using all data contained in the Hip-

de Geus E.J., 1992, A&A 262, 258

parcos Catalog. We exp ect to nd additional mem-

de Geus E.J., de ZeeuwP.T., Lub J., 1989, A&A 216,

b ers that were not included b ecause of our magnitude

44

limits or eld b oundaries. The resulting astromet-

ric memb ership lists can then be re ned further by

de Geus E.J., Lub J., van der Grift E., 1990, A&AS

radial velo city measurements. This is now feasible,

85, 915

as the astrometric memb ership selection has reduced

Gies D.R., Bolton C.T., 1986, ApJS 61, 419

the number of candidate asso ciation memb ers in a

Ho ogerwerf R., de Bruijne J.H.J., Brown A.G.A.,

eld typically by an order of magnitude. This also

Lub J., Blaauw A., de Zeeuw P.T., 1997, these

applies to the required completion of the multi-colour

pro ceedings, ESA-SP 402 P05.08

photometry. In the near future we will obtain radial

velo cities for the asso ciations Lac OB1, Cep OB2 and

Kap er L., et al., 1997, ApJ 475, L37 err. ApJ 479,

Per OB2 from approved observations at the McDon-

L153

ald and Haute-Provence observatories. We will study

Mac Low M.-M., McCray R., 1988, ApJ 324, 776

the binary p opulation in Sco OB2 by means of ra-

Mathieu R.D., 1986, in Highlights of Astronomy 7,

dial velo city measurements and near-infrared adap-

p. 481

tive optics searches for close companions to the B-

typ e stars. Subsequently we will extend these studies

van Rensb ergen W., van Beveren D., de Lo ore C.,

to the other asso ciations.

1996, A&A 305, 825

Sahu M., 1992, PhD Thesis, Groningen Univ.

Straka W.C., 1973, ApJ 180, 907

ACKNOWLEDGMENTS

Upton E.K.L., 1971, in The Gum Nebula and Re-

latedProblems, eds S.P. Maran, J.C. Brandt, T.P.

Stecher, NASA SP{332, p. 119

It is a pleasure to thank Eug ene de Geus and Michael

Verschueren W., David M., 1989, A&A 219, 105

Perryman for stimulating discussions, and Jan Brand

and Frank Israel for signi cant contributions to the

Verschueren W., Brown A.G.A., Hensb erge H., David

1982 SPECTER prop osal.

M., Le Po ole R.S., de Geus E.J., de ZeeuwP.T.,

1997, PASP, in press

Warren W.H., Hesser J.E., 1978, ApJS 36, 497

REFERENCES

de ZeeuwP.T., Brown A.G.A., Verschueren W., 1994,

in Galactic and Solar System Optical Astrometry:

Observation and Application, eds. L.V. Morrison

Ambartsumian V.A., 1947, in Stel lar Evolution and

& G.F. Gilmore, Cambridge Univ. Press, p. 215

Astrophysics, Armenian Acad. of Sci. German

transl., Abhandl. Sowjet. Astron. 1, 33 [1951]

Bertiau F.C., 1958, ApJ 128, 533

Blaauw A., 1946, PhD Thesis, Groningen Univ.

Blaauw A., 1950, BAN 11, 405

Blaauw A., Morgan W.W., 1953, ApJ 117, 256

Blaauw A., 1956, ApJ 123, 408

Blaauw A., 1961, BAN 15, 265

Blaauw A., 1964, ARA&A 2, 213

Blaauw A., 1991, in The Physics of Star Formation

and Early Stel lar Evolution, eds. C.J. Lada &

N.D. Kyla s, NATO ASI, Vol. 342, p. 125

Bok B.J., 1934, Harvard Coll. Obs., Circular 384, 1

Brandt J.C., Stecher T.P., Crawford D.L., Maran

S.P., 1971, ApJ 163, L99

Brown A.G.A., Verschueren W., 1997, A&A 319, 811 oral05-06