Cluster Dynamics Largely Shapes Protoplanetary Disc Sizes: Kirsten Vincke and Susanne Pfalzner
Cluster Dynamics Largely Shapes Protoplanetary Disc Sizes: Kirsten Vincke and Susanne Pfalzner
Cluster Dynamics Largely Shapes Protoplanetary Disc Sizes: Kirsten Vincke and Susanne Pfalzner
ABSTRACT
It is still on open question to what degree the cluster environment influences the sizes of
protoplanetary discs surrounding young stars. Particularly so for the short-lived clusters typical
for the solar neighbourhood in which the stellar density and therefore the influence of the cluster
environment changes considerably over the first 10Myr. In previous studies often the effect of the
gas on the cluster dynamics has been neglected, this is remedied here. Using the code NBody6++
we study the stellar dynamics in different developmental phases - embedded, expulsion, expansion
- including the gas and quantify the effect of fly-bys on the disc size. We concentrate on massive
clusters (Mcl 103 6104 Msun ), which are representative for clusters like the ONC or NGC6611.
We find that not only the stellar density but also the duration of the embedded phase matters.
The densest clusters react fastest to the gas expulsion and drop quickly in density, here 98% of
relevant encounters happen before gas expulsion. By contrast, discs in sparser clusters are initially
less affected but as they expand slower 13% of discs are truncated after gas expulsion. For ONClike clusters we find that usually discs larger than 500AU are affected by the environment, which
corresponds to the observation that 200AU-sized discs are common. For NGC6611-like clusters
disc sizes are cut-down on average to roughly 100AU. A testable hypothesis would be that the
discs in the centre of NGC6611 should be on average 20AU and therefore considerably smaller
than in the ONC.
Subject headings: protoplanetary disks planetary systems galaxies: star clusters: general
1.
Introduction
Method
Mstars (1 SFE)
.
SFE
(1)
Cluster simulations
GMcl
3
rhm
1/2
,
(2)
The dynamical time scales for the cluster models E0-E52 are very short, between 0.8 0.14 Myr,
see Column 7 of Table 1. Therefore, and for better comparability of our cluster models, we assume
the gas expulsion process in all clusters to be instantaneous. This immediate removal of the gas
mass after t = temb leaves the cluster in a supervirial state, so that the cluster expands in order to
regain virial equilibrium. We will discuss the consequences of such an instantaneous gas-expulsion
on the results compared to a longer expulsion time
scale in Section 4. We follow the cluster expansion until 10 Myr have passed since the cluster
was fully formed.
(
0.32
0.28 rperi m12
, if rdisc < rprevious
rdisc =
rprevious ,
if rdisc rprevious ,
(3)
given by Breslau et al. (2014), where m12 =
m2 /m1 is the mass ratio between the disc-hosting
star (m1 ) and the perturber (m2 ), rperi the periastron distance in AU, and rprevious the disc size
previous to the fly-by in AU. This equation is valid
for coplanar, prograde, parabolic fly-bys. This
type of fly-by is more destructive than inclined,
retrograde or hyperbolic fly-bys (Clarke & Pringle
1993; Heller 1995; Hall 1997; Pfalzner et al. 2005b;
Bhandare & Pfalzner subm.). However, the effect
of inclined, retrograde and hyperbolic fly-bys is
much less investigated. First results by Bhandare
& Pfalzner indicate that non-coplanar encounters
have nevertheless a considerable effect on the disc
size. Thus, the here presented result has to be
regarded as lower limit of disc size, but will not
be considerably smaller than it would be in the
inclined case.
Viscous forces, which might lead to disc spread-
Results
The distinct clusters have very different influence on their protoplanetary discs, reflected for
example in the overall median disc size (Figure 4).
This median disc size is about thirteen times
smaller in model E52 (32 000 stars) than in
model E0 (1 000 stars) as not only the number
of fly-bys increases significantly with cluster density, but they are on average also closer or the
mass ratio is higher. For the densest clusters
most fly-bys happen at the beginning of the embedded phase, thus, the median disc sizes are
nearly the same at the end of the embedded
phase ( 108 AU, open squares) and at the end
of the simulations ( 104 AU, dots). However,
for model E0 the median disc size is significantly
larger ( 1 670 AU) at the end of the embedded
phase (2 Myr) than at the end of the simulations
( 1 350 AU) as roughly one seventh of the close
fly-bys occur in the expansion phase.
6
tration on the cluster centre is problematic, especially so for clusters after gas expulsion, which
span large areas. Taking our results as a guideline the observed median disc size in an ONC-like
cluster 10 Myr after cluster formation, for example, would be 50 AU in the cluster core (0.2 pc)
whereas the overall median disc size is more than
nine times as large ( 460 AU, dotted horizontal
line in Figure 4).
Choosing initially artificially large discs of
10 000 AU has the advantage that the obtained
results can be applied any smaller, real disc size.
Thus, Figure 7, tells us, for example, that if all
stars had an initial disc size of rinit 500 AU,
about half the stars had their discs severely truncated by fly-bys to disc sizes below 500 AU. An
initial disc size of more than 500 AU is a realistic scenario as surveys found discs in the ONC
with radii of 30 500 AU (McCaughrean & Odell
1996; Vicente & Alves 2005; Eisner et al. 2008;
Bally et al. 2015). Note, that at an age of approximately 1 Myr even those might already have been
reduced in size through photoevaporation and/or
fly-bys. In the case of more massive clusters like
NGC 6611 (E52 model) there are more and closer
interactions, so that independent of the initial disc
size (as long as rinit > 100 AU) the resulting median disc is 110 AU, see Fig. 4.
If observations of older clusters work like a microscope, what would an observed disc-size distribution in a cluster at different ages look like?
To answer this question, we investigate the ONCmodel cluster (E2) at different ages with an artificial fixed Field of View (FOV) (rF OV = 1 pc)
to mimic observations. Note that the FOV for
observations are usually squares whereas here we
present spheres with a radius of rFOV , centred on
the cluster origin. Figure 6 shows the resulting
disc-size statistics for an ONC-like cluster at 1 Myr
(white), 2 Myr (grey), and 10 Myr (black). The total number of small discs increases much stronger
than the number of discs with sizes of several hundreds of AU. The reason is that in the embedded
phase discs which are already influenced but still
a few hundreds of AU large still reduced in size by
follow-up encounters. In comparison, the shape of
the disc-size distribution barely changes between
the end of the embedded phase and the end of the
simulations.
Observations usually study only the central areas of a cluster, because their the stellar density
of cluster members is so high that member identification is relatively easy - basically the rate of
false-positives is very low. However, this concen-
In summary, observed disc sizes or disc-size distributions in massive clusters are a strong function
of the cluster age, its evolutionary stage, its initial conditions, and the FOV of the instrument.
One has to act with caution when comparing and
interpreting such results.
4.
Discussion
In this paper we studied how the cluster environment changes the sizes of discs surrounding
young stars. In contrast to previous work we took
the cluster development during the first 10 Myr
explicitly into account. Starting with initial conditions typical for young clusters at the end of
their formation phase in the solar neighbourhood,
we modelled the cluster dynamics from embedded
throughout the expansion phase and determined
the effect on the effect of gravitational interactions
between the stars on the disc sizes. These type of
simulations were performed for clusters of different
mass and density.
Our findings are the following:
1. It is essential to include the gas dynamics in
this kind of simulations, as the larger velocity dispersion leads to more encounters and
significantly smaller disc sizes than in a gasfree treatment.
2. The majority of disc-size changing fly-bys
always takes place in the embedded phase.
However, the slower expansion phase in
lower mass clusters means that here still
12% of disc-size changing fly-bys happen, in
comparison to just 2% for high-mass clusters.
3. For ONC-like clusters basically only discs
larger than 500 AU are affected by fly-bys,
whereas in NGC 6611-like clusters, cutting
discs below 100 AU happens for 50% of stars.
4. However, in all investigated cases the disc
sizes in the dense cluster centres are much
more affected than the average suggests. For
example, in the NGC 6611-like case the median disc size is 54 AU.
5. The duration of the embedded phase influences the final median disc size, but not as
strong as one would expect, because early
fly-bys reduce the disc size already, leading
to smaller cross sections for later fly-bys. In
the densest cluster the median disc size after
1 Myr is already 155 AU, at the end of the
embedded phase 108 AU, which is very close
to the final median disc size of 104 AU.
10
A.
The characteristics of stellar encounters change significantly with cluster density. For example, the relative
velocity between two encountering stars increases for denser clusters. Figure 8(a) depicts the average relative
encounter velocity - that is the velocity of the perturber relative to the host star at the time of periastron
passage - for three cluster models E0 (squares), E2 (dots), and E52 (triangles). This encounter velocity can
directly be correlated to the eccentricity of the perturbers orbit via:
The eccentricity e is directly related to the relative encounter velocity of the two stars at the time of
periastron passage:
s
G(m1 + m2 )
,
(A1)
venc = (1 + e)
rperi
where G is the gravitational constant, m1 is the mass of the host star, m2 the mass of the perturber, and
rperi the periastron distance, all in SI units. The eccentricity distribution for cluster models E0, E2, and E52
are shown in Fig. 8(b) for fly-bys leading to disc smaller than 500 AU.
In this study we assumed all fly-bys to be parabolic. This approximation only holds for the least dense
cluster model, as the encounter velocities and therefore the eccentricities clearly increases with cluster density
(see also Olczak et al. 2010). For the denser cluster models (especially E52) a detailed study of the influence
of hyperbolic fly-bys on disc sizes would be favourable. Previous studies suggest that their influence on
the discs (in these cases the disc mass and angular momentum) is much smaller than the one of parabolic
encounters (for detailed discussions see e.g. Pfalzner et al. 2005b; Olczak et al. 2010, 2012). Therefore, the
disc sizes presented here might be lower limits.
At very high densities, that is especially in cluster model E52, fly-bys are no longer 2-body encounters
but many-body interactions. This leads to the extreme eccentricities of e > 100. Especially for this type of
fly-by we expect the disc-size change to be smaller than for the here assumed prograde, coplanar, parabolic
case.
11
E0
E2
E51
E52
1000
100
10
0.1
0
Fig. 1.: Stellar mass density within 1.3 pc (initial half-mass radius) as a function of time for clusters of
different densities: E0 (dots), E2 (squares), E51 (asterisks, blue), and E52 (triangles). The duration of the
embedded phase is temb = 1 Myr for E51 and temb = 2 Myr for all other models.
160 000
120 000
100 000
80 000
2
60 000
40 000
140 000
20 000
0
1 000
2 000
4 000
8 000 16 000 32 000
number of stars in cluster
Fig. 2.: Number of fly-bys per cluster (grey boxes, left y-axis) and per star (black diamonds, right y-axis)
for the different cluster models. The black line only serves to guide the eye.
12
0.7
0.6
0.5
0.4
0.98
0.96
0.97
embedded phase
expansion phase
0.94
0.8
fraction of fly-bys
0.8
0.6
0.4
2
simulation time [Myr]
0.02
0.03
0.06
0.1
0.2
0.09
E0
E2
E51
E52
0.2
0.04
0.3
0.13
0.9
0.91
1.2
0.87
0
4
1 000
2 000
4 000
8 000
16 000
number of stars in cluster
32 000
Fig. 3.: (a) Cumulative fraction of fly-bys as a function of time for the cluster models E0 (squares),
E2 (dots), E51 (asterisks, blue), and E52 (triangles). The vertical lines depict the points in time of gas
expulsion for model E51 (1 Myr, dotted blue) and for all other models (2 Myr, solid black). (b) Fraction of
fly-bys as a function of the number of stars in the cluster for the embedded phase (black) and the expansion
phase (grey).
1800
end of simulation
end of embedded phase
end of embedded phase within 1.3 pc
1600
1400
1200
1000
800
600
400
200
0
1000
2000
4000
8000
16000
number of stars in cluster
32000
Fig. 4.: Overall median disc size for all stars for different cluster models at the end of the simulation
(10 Myr, dots), at the end of the embedded phase (2 Myr, squares), and at the end of the embedded phase
within a sphere of 1.3 pc (initial half-mass radius, triangles).
13
1800
1600
1800
embedded expansion
E0
E2
E52
1600
1400
median disc size [AU]
1400
this work
VBP15
1200
1000
800
600
1200
1000
800
600
400
400
200
200
0
0.1
1
distance to cluster centre [pc]
10
0.1
1
distance to cluster centre [pc]
Fig. 5.: Median disc size as a function of the distance to the cluster centre (a) for different cluster models
(E0 squares, E2 circles, and E52 triangles) at the end of the embedded phase (2 Myr, black symbols) and
at the end of the simulation covering the embedded, the gas expulsion, and the expansion phase (10 Myr
(red symbols); (b) for the ONC model (E2) in this work, i.e. with gas mass (circles, same as in (a)), and in
VBP15, i.e. without gas mass (squares). The lines in (a) and (b) only serve to guide the eye.
650
1 Myr
2 Myr
10 Myr
600
550
500
number of stars
450
400
350
300
250
200
150
100
50
0
0
100
200
300
400
500
600
700
800
900
1000
Fig. 6.: Disc-size distribution in the ONC-like cluster model (E2) for a fixed virtual FOV (1 pc) and
different time steps: 1 Myr (white), 2 Myr (grey), and 10 Myr (black).
14
500
>= 500 AU
300 AU
100 AU
400
300
200
100
0
0.1
1
distance to cluster centre [pc]
10
Fig. 7.: Median disc size at the end of the simulation (10 Myr) as a function of the distance to the
cluster centre for cluster model E2 for different initial disc sizes: 100 AU (circles), 300 AU (triangles), and
500 AU (dots).
35
E0
E2
E52
30
E0
E2
E52
4
25
20
15
10
5
0
0
10
15
20
25
relative encounter velocity [km/s]
30
35
10
100
eccentricity
1000
Fig. 8.: (a) relative encounter velocity distribution - that is the perturbers velocity relative the the hosts
velocity at the point of periastron passage - of all fly-bys and (b) eccentricity distribution of fly-bys leading
to a disc size < 500 AU as fractions of the total number of fly-bys for three cluster models: E0 (squares),
E2 (dots), and E52 (triangles).
15
Nstars
Nsim
E0
E1
E2
E3
E4
E52
E51
1 000
2 000
4 000
8 000
16 000
32 000
32 000
308
168
94
47
16
9
7
temb
[Myr]
2.0
2.0
2.0
2.0
2.0
2.0
1.0
rhm
[pc]
1.3
1.3
1.3
1.3
1.3
1.3
1.3
Mstars
[M ]
590.8
1 192.2
2 358.1
4 731.2
9 464.8
18 852.6
18 839.2
Mcl
[M ]
1 969.2
3 973.9
7 860.3
15 770.6
31 549.3
62 842.0
62 797.3
tdyn
[Myr]
0.67
0.47
0.33
0.24
0.17
0.12
0.12
Note.Column 1 indicates the model designation, followed by the initial number of stars in the cluster Nstars , the number
of simulations in campaign Nsim , the duration of the embedded phase temb , the initial half-mass radius rhm of the cluster, the
stellar mass of the cluster Mstars , the total cluster mass (including the gas mass) Mcl , and the resulting dynamical time scale
tdyn . For calculation of Mcl and tdyn , see text.
16
REFERENCES
Aarseth, S. J. 1973, Vistas in Astronomy, 15, 13
Aarseth, S. J. 2003, Gravitational N-Body Simulations
Adams, F. C. 2000, ApJ, 542, 964
Adams, F. C. 2010, ARA&A, 48, 47
Adams, F. C., Proszkow, E. M., Fatuzzo, M., & Myers, P. C. 2006, ApJ, 641, 504
Alexander, R. D., Clarke, C. J., & Pringle, J. E. 2006, MNRAS, 369, 229
Allison, R. J., Goodwin, S. P., Parker, R. J., Portegies Zwart, S. F., & de Grijs 2010, MNRAS, 407, 1098
Balbus, S. A. & Hawley, J. F. 2002, ApJ, 573, 749
Bally, J., Mann, R. K., Eisner, J., et al. 2015, ArXiv e-prints
Baumgardt, H. & Kroupa, P. 2007, MNRAS, 380, 1589
Bhandare, A. & Pfalzner, S. in prep.
Boily, C. M. & Kroupa, P. 2003a, MNRAS, 338, 665
Boily, C. M. & Kroupa, P. 2003b, MNRAS, 338, 673
Bonnell, I. A., Bate, M. R., Vine, S. G. 2003, MNRAS, 343, 413
Bonnell, I. A. & Davies, M. B. 1998, MNRAS, 295, 691
Breslau, A., Steinhausen, M., Vincke, K., & Pfalzner, S. 2014, A&A, 565, A130
Clarke, C. J. & Pringle, J. E. 1993, MNRAS, 261, 190
Dale, J. E., Ercolano, B., & Bonnell, I. A. 2012, MNRAS, 424, 377
Dale, J. E., Ercolano, B., & Bonnell, I. A. 2015, MNRAS, 451, 5506
de Juan Ovelar, M., Kruijssen, J. M. D., Bressert, E., et al. 2012, A&A, 546, L1
Drake, J. J., Ercolano, B., Flaccomio, E., & Micela, G. 2009, ApJ, 699, L35
Duchene, G. & Kraus, A. 2013, ARA&A, 51, 269
Dukes, D., & Krumholz, M. R. 2012, ApJ, 754, 56
Eisner, J. A., Plambeck, R. L., Carpenter, J. M., et al. 2008, ApJ, 683, 304
Ercolano, B., Drake, J. J., Raymond, J. C., & Clarke, C. C. 2008, ApJ, 688, 398
Fall, S. M., Chandar, R., & Whitmore, B. C. 2009, ApJ, 704, 453
Fellhauer, M. & Kroupa, P. 2005, ApJ, 630, 879
Garca, P. and Bronfman, L. and Nyman, L.-A. et al. 2014, ApJS, 212, 2
Geyer, M. P. & Burkert, A. 2001, MNRAS, 323, 988
Goodwin, S. P. 1997, MNRAS, 284, 785
17
18
This 2-column preprint was prepared with the AAS LATEX macros v5.2.
20