ORION AND THEORIES OF STAR FORMATION

Loading...

Taking too long?

Reload document

| Open in new tab

ORION AND THEORIES 01’ STAR FORMATION

Richard B. Larson

Astronomy Department
Yale University
New Haven, Connecticut 06.51 I

INTRODUCTION

Star formation and its various effects are evidently responsible for many of the
complex phenomena observed in the Orion Nebula region, and one would therefore like
to be able to explain the observations of this region on the basis of theories of star
formation. At present, however, any theoretical understanding of star formation is still
too rudimentary to bear any detailed comparison with the real world, so one can only
discuss some possible processes that may play a role in star formation and then appeal
to the observations to try to clarify their importance. In this paper, I shall mention
some of the small-scale processes that may be important in the formation of stars in
molecular clouds and then describe some data on young stars in Orion and other
regions that provide information about star formation in these regions.

As a caution against overly simplified theories of star formation, some large-scale
properties of the Orion region are worth keeping in mind. One is that the entire Orion
complex of molecular clouds and young stars, whose total mass probably exceeds 2 x
los Mo, is located more than 100 pc below the Galactic plane; thus, whatever process
formed the Orion clouds must have assembled them well out of the Galactic plane or,
equivalently, must have accelerated the material to a substantial velocity away from
the Galactic plane. Moreover, the CO maps that Thaddeus presents in this volume
reveal that the Orion clouds are very extensive and complex in structure, with several
major concentrations and filamentary extensions, and that they also have a complex
velocity field. These observations suggest that the Orion clouds were formed in a more
or less violent way by processes that left them in a turbulent state; if so, turbulent
motions may play a major role in the subsequent evolution of the clouds and the
processes that occur may be much more complex than either a simple contraction
process or a regular propagation of star formation through each cloud. In fact, the
available data show star formation proceeding simultaneously at several sites in several
different molecular concentrations in Orion and do not, when examined closely, clearly
follow the widely discussed picture of sequential star formation, but suggest a more
complex situation in which star formation can be initiated at many sites.

Ultimately, it is necessary to understand how, on small scales, the molecular cloud
material actually becomes condensed into stars and groups of stars. Related problems
of long-standing interest are understanding the mass spectrum with which stars form,
the time dependence of star formation, and the. efficiency of the conversion of gas into
stars. While complete theoretical answers cannot yet be given to these questions, it is at
least possible to discuss some of the processes that may be involved and some
observational constraints on their importance.

274

0077-8923/82/0395427451.75/1

0 1 9 8 2 , N Y A S

Larson: Orion & Theories of S t a r Formation

275

POSSIBLE PROCESSES OF STAR FORMATION

Gravity must clearly play a major role in compressing interstellar matter into stars,
and the classical picture of star formation is based on the concept of successive
gravitational fragmentation elaborated by Jeans, Hoyle, and others. If a cloud begins
to contract under gravity, regions of enhanced density containing more than the Jeans
mass can collapse separately to high densities and may fragment again into smaller and
smaller units as the density increases and the Jeans mass decreases. Most numerical
calculations of collapse and star formation have been motivated by this picture, but the
multidimensional calculations have not, so far, supported the concept of successive
fragmentation into objects much smaller than the initial Jeans mass. Accordingly, if
gravitational fragmentation is the only process that operates, one might expect
characteristic stellar masses to be comparable to the Jeans mass in the cloud from
which the stars form, being smallest in the regions of highest density.

On the other hand, purely hydrodynamic processes such as shock compression can
also play a role in compressing the gas in molecular clouds; shocks should occur often
because supersonic internal motions are observed in all but the smallest clouds. The
observed motions generally appear to be complex or turbulent, so the shock-
compressed regions will probably have an irregular filamentary or clumpy structure;
such supersonic turbulent motions may even be responsible for the formation of
protostars.’ If this is the primary mechanism of star formation, one might expect the
stars of lowest mass to form in the regions with the highest turbulent velocities.
Another purely hydrodynamic process that may become important after dense
protostellar clumps have formed, either by gravity or by turbulence, is collisional
coalescence and growth of protostellar objects.2

I n reality, gravity and hydrodynamics probably act together in the formation of
stars. For example, once a small dense core or embryonic star has begun to form in a
protostellar condensation, its mass can grow by a large factor by gravitationally
accreting the more diffuse surrounding gas. The accretion process may involve infall
through an accretion shock at the surface of the stellar core if the infalling material has
little angular momentum, or viscous inflow in an accretion disk if the material has a
large angular momentum. More complex processes involving accretion of lumps or
streams of gas may also occur. Numerical simulations’ and analytical theories4 of the
accretive growth of embryonic stars suggest that this process can account approxi-
mately for the form of the observed stellar mass spectrum. If such accumulation
processes are important in the formation of stars, it might be expected that the most
massive stars would form in the densest regions, i.e., in the dense cores of contracting
molecular clouds or forming star clusters where the gas density and accretion rate are
highest. Gravitational accretion is a runaway process, since the accretion rate increases
with the mass of the accreting object; therefore, stars of extremely large mass may
form, unless other effects, such as stellar winds, intervene to cut off accretion before all
the available gas has been consumed.

Once stars have begun to form in a molecular cloud, they may exert an important
influence on subsequent star formation. For example, the stars will begin to contribute
to the gravitational field in the cloud and may exert significant tidal forces on the gas.
I f the cloud contains a dense embedded cluster of stars, such as appears to exist in
Orion, the star cluster may produce an approximately symmetrical, centrally

276

Annals New York Academy of Sciences

condensed gravitational potential well that acts to concentrate the rcmaining gas to
very high densities at the center, possibly favoring the accretive build up of very
massive stars there. Since newly formed stars produce stellar winds with mass-loss
rates that increase strongly with stellar mass, the accretive build up of a massive star
may be cut off when its wind becomes strong enough to blow away the surrounding gas.
This would imply an upper mass limit that increases with increasing ambient gas
density, as was suggested previously on the basis of radiation pressure and ionization
effects.’

Given all the possible processes that have been mentioned, and others, such as
magnetohydrodynamic effects, that probably also occur, detailed predictions of how
stars form cannot yet be made; to understand more about the way in which stars
actually form, it is necessary to refer to observations of regions of star formation. Data
on the young stars in these regions are particularly useful because they provide a record
of the past history of star formation and can be used to study correlations between
stellar properties, such as the mass spectrum, and other characteristics of each region,
such as the spatial distribution of gas and young stars.

Y O U N G STARS I N ORION A N D O T H E R REGIONS

Cohen and Kuhi have published an extensive study of T Tauri stars in many regions
of star formation: providing quantitative data on temperatures and bolometric
luminosities from which the ages and masses of the stars can be estimated. These data
have been used to study the mass spectra, spatial distributions, and ages of the T Tauri
stars in the three most populous regions, namely, Taurus, Orion, and N G C 2264, and
to look both for possible differences in the stellar properties and for correlations with
other properties of the associated star-forming molecular complexes.’

The spatial distributions of the Cohen-Kuhi stars in Taurus, Orion, and N G C 2264
are plotted on similar linear scales in FIGURES 1-3, coded by stellar mass. In Taurus
(FIGURE I ) , the youngest stars are dispersed over a relatively large region about 40 pc
across and they are mostly in small groups containing from a few up to about ten stars.
These small groups are closely associated with a number of Barnard and Lynds dark
clouds whose positions are approximately indicated in FIGURE 1. Most of the T Tauri
stars have masses less than 1 M,; the most massive star in the region is a highly
obscured 8 2 star, located, as indicated, in the B18 cloud near the center of the region.
The distribution of the T Tauri stars in Orion is strikingly different (FIGURE 2),
being more compact and centrally concentrated than that in Taurus; most of the Orion
stars are in a single large cluster about 10 pc across, centered on the Trapezium. A
secondary concentration is associated with the reflection nebula N G C 1999, and the
remaining T Tauri stars are scattered along the length of the L1640 dark cloud. Most
of the known T Tauri stars in Orion have masses greater than 1 Mo, in contrast with
those in Taurus whose masses are mostly less than 1 M,. Another difference is that the
Orion stars are somewhat older, having a median age of 1 .O x 106 y, as compared with
6 x 10’ y for Taurus. Further evidence that the Orion region is older and more evolved
than Taurus is provided by the presence there of large numbers of flare stars, which are
generally fainter and older than the T Tauri stars; Orion has 325 known flare stars,
whereas Taurus has only 13.’

N G C 2264 (FIGURE 3) differs from Taurus in the same sense and perhaps to an

Larson: Orion & Theories of S t a r Formation

277

even more extreme degree than Orion; the spatial distribution of the T Tauri stars is
more compact and their median age is even greater, about 2 x 10′ y. As in Orion, most
of the known stars have masses greater than 1 M,.

The data therefore suggest that, as a star-forming region evolves, the gas becomes
more spatially condensed and the stars form in more massive and concentrated
clusters, with larger and larger masses. There is, in fact, direct evidence that, in some
young associations or clusters, the morc massive stars form later than the less massive

+ 320c

30′-

s,

TAURUS

02140 vI

0
07

I

;0209

10213

I

8 (1900)
–

+ 24O

~ l o p c

gey1

02 star

0 9

0

1LlS36

–

–

28′

26′

–

2 2 O

–

20°

–

IS0

T Tau0

M …

LlSSl
x %r.

1 0 . 2 5 – 0.5
0.5 – 1.0
o
1.0 -2.0
0 2.0 -4.0

l

t

e ~ 1 5 5 8

+ 1

6

°

4h52′”

44′”

36′”

4h28′” 20′”

12′”

4’004′”

a (1900)

FIGURE I .

The spatial distribution of the Cohen-Kuhi stars in Taurus, coded by mass, as
indicated in the lower right. The positions of the associated Barnard and Lynds dark clouds are
also indicated, as is the position of the star of earliest spectral type, a heavily obscured B2 star.

stars.’.” To detcrmine whether there is also evidence that the more massive stars in
Orion have formed with a more compact spatial distribution than the less massive ones,
we have shown, in FIGURE 4, only those T Tauri stars that have ages less than 106 y and
are still seen close to their places of formation; although the statistical significance of
the result is not large, the more massive T Tauri stars in FIGURE 4 are, i n fact,
considerably more spatially concentrated than the less massive ones. It is particularly

-3‘

1

I

I

ORION

1

1

8 (1900

– 6‘

[loPC

1

– 4‘

– 5’

-7

-8

* #

1 0

$ .

. I

O f .

10

0

,

P

o
0 0 I
NGC 1999

I

a

.

* @ 8 .

M
1 0.25 – 0.5
0 0.5 – 1.0
1.0 – 2.0
F) 2.0 – 4.0

5”am

34m

26m

5h22m

Sh 30m
u ( I 900)

F I G U R E 2. The spatial distribution of the Cohen-Kuhi stars in Orion, plotted on nearly the
same linear scale as that used in FIGURE 2. The Orion Nebula (NGC 1976) and the associated
young cluster are centered on the Trapezium, which contains the most massive star in the
region.

FIGCIR~. 3. The spatial dis-
tribution of the emission-line
stars in NGC 2264. The posi-
tion of the most massive star,
the 0 7 star S Mon, is also
indicated.

+10.5’

S Mon

NGC 2264

1 t
I 10.0”
+9.0 1

0.5 – 1.0
1.0 – 2 . 0
0
0 2 . 0 – 4.0

I 0.2s-0.s

M

..a

0

a (1900)

6’3em

36m

3 4m

6h 32m

a (1900)

Larson: Orion & Theories of S t a r Formation

279

striking that the most massive star in the Orion region, B’C Ori in the Trapezium, is
right at the center of the Orion cluster.

The formation of massive stars is continuing a t present in the dense core of the
Orion Molecular Cloud OMCI, whose compact cluster of luminous infrared sources
(the BN-KL cluster) is also centrally located and has a projected distance from the
Trapezium of only about 0.1 pc. When the stars in the infrared cluster blow away their
surrounding gas and become visible, as will probably soon happen, they will appear as a

ORION

I

age < 1 0 ~ y r I 0 0 . * . . 0 . - 3' - 4' - 5' -7 - 8 a (1900 - 6 ' 0 1 NQC I999 I M I 0.25 - 0.5 0.5 - 1.0 . 1.0 - 2.0 0 0 2.0 -4.0 I I 1 1 Sh 3em 34m 26m 5 5h30m 0 (1900) FIGURE 4. The spatial distribution of the emission-line stars in Orion that have ages less than 106y. new Trapezium-like subgroup of massive stars near the center of the Orion Nebula cluster. The infrared objects in OMC2 apparently represent a group of less massive stars forming further away from the center of the Orion cluster. MASS SPECTRA OF YOUNG STARS Because most of the T Tauri stars studied by Cohen and Kuhi lie on nearly vertical their estimated masses depend primarily on their Hayashi tracks in the H R diagram: 280 Annals New York Academy of Sciences spectral types and their derived mass spectra are not grossly affected by magnitude- dependent selection effects. There are striking differences in the distribution of spectral types of the T Tauri stars in Taurus, Orion, and NGC 2264. FIGURE 5 shows the ratio of the numbers of stars with spectra earlier or later than K6.5 (corresponding to masses larger or smaller than about 1 M,,) and brighter than various limiting apparent magnitudes, plotted as if all stars were placed at the distance of Orion. It is evident that, regardless of the limiting magnitude chosen, a large difference persists between Taurus and the other two regions. Even after corrections are made for the different median ages o f the stars in these regions,' the difference remains significant at about the 3a level. The mass spectra of the Cohen-Kuhi stars in Taurus and Orion are shown in FIGURE 6. Here the difference between Taurus and Orion appears as a turndown in the Orion mass spectrum at masses below 1 M,, This deficiency of low-mass stars may not apply to the entire population of young stars in Orion, which includes the numerous faint flare stars that are widely dispersed throughout the region, but it is nevertheless characteristic of the youngest stars that have just recently formed in the Orion Nebula cluster. It is noteworthy that the Taurus mass spectrum is very similar to the initial mass spectrum of field stars," while the Orion mass spectrum is similar to that of many open clusters, which often show a deficiency of low-mass stars." Thus, Taurus may be a typical site for the formation of field stars, while in Orion we see the formation of a typical open cluster. POSSIBLE INFERENCES FOR STAR FORMATION The data discussed above are all consistent with an evolutionary picture in which the properties of a large star-forming complex change systematically with time while stars continue to form for a period o f at least -10' y. At an early stage, a star-forming region may resemble the Taurus dark clouds, where the average density is relatively t FIGURE 5. The ratio of the number of emis- sion-line stars with spectral types of K6 and earlier to the number with spectral type K7 and later plotted against the limiting magnitude that the stars would have if all were placed at the distance of Orion. I I q -0.5 I 14 t I I5 B I 0 I 16 Vmax (at distance of Orion) 17 I ! 18 Larson: Orion & Theories of Star Formation 28 1 F I G U R ~ 6. The mass spectra of the emission-line stars in Taurus and Orion. Plotted are the numbers of stars in the equal logarithmic mass intervals 0.25-0.5, 0.5-1.0, 1.0 2.0, and 2.0-4.0 M,, normalized by the total number of stars in each region. The error bars give N ” 2 statistical errors. 2 -0.5 - ? - 5 2 - CI, Y -1.0- - 1.5 I low and small groups of low-mass stars are forming in scattered small molecular clumps and filaments. With time, the gas may become progressively more condensed into massive, dense clouds or cloud cores like OMC I , forming stars in larger and more centrally concentrated clusters, with a mass spectrum that increasingly favors massive stars. The most massive stars may form as a result of rapid accumulation processes in the dense central cores of these contracting clouds or forming star clusters. The culmination of this type of evolution may be represented by the present Orion Nebula region, where a dense cluster with very massive stars at its center has been forming for the past -10’ y and where the formation of massive stars is continuing near the center. The fact that the most massive stars appear to form only in a small region of very high density at the center of a forming star cluster is not easily explained by the traditional fragmentation theory of star formation, but suggests that accretive build up processes are involved. Moreover, the massive stars appear to form after, and perhaps as a result of, the previous formation of large groups of less massive stars; a possible explanation is that the gravitational field of a cluster of mostly low-mass stars helps to concentrate the remaining gas to the high densities required to form the more massive stars. The observations of gas flows associated with obscured young objects in Orion and elsewhere show, however, that no sooner does a massive star form than it begins to produce a strong stellar wind that blows away the gas in its vicinity, probably cutting off further accretion and star formation processes and exposing the newly formed stars to view. The formation of a massive star therefore requires that a large amount of gas be accumulated rapidly into a star before a wind can blow it away; this may explain why massive stars appear to form only in very dense regions, since only there can the required very high accretion rate be attained. Clouds or clumps of lower density can only produce lower accretion rates and, therefore, could only form less massive stars, in agreement with what is observed. REFERENCES 1. 2. LARSON, R. B. 1981. Mon. Not. R. Astron. Sac. 194 809. SILK, J. & T. TAKAHASHI. 1979. Astrophys. J. 2 2 9 242.