Friday, 21 October 2016

Star formation, filaments, and the initial distribution of young stars

I've talked before about molecular clouds, the birthplaces of stars, but only in very general terms. Today I want to go a little deeper to look at how stars are distributed within molecular clouds when they form, and how this can lead to the formation of star clusters.

Molecular clouds aren't just big fluffy clouds of gas that uniformly form stars, they have considerable structure. This is clear if you look at molecular clouds using an infrared telescope, particularly one that's tuned to the far-infrared, such as the Herschel Space Observatory (the further into the infrared you go, the cooler the material you can study because cooler things radiate light at longer wavelengths, i.e. the far infrared).

A far-infrared image of the Taurus molecular cloud, showing the filamentary structure of the gas
(Credit: Herschel Space Observatory)
This image from the Herschel Space Observatory shows the incredibly filamentary structure of the gas in the Taurus molecular cloud. This structure is thought to arise due to a combination of shock compression (due to collisions between material) and self-gravity (meaning the filaments can form gravitationally-stable structures by themselves). These filaments can be seen on all spatial scales, from the large filament spanning the entire image, through to small filaments emanating from the larger ones.

Many of these filaments are dense, containing many times the mass of our Sun in molecular gas. This high density means that they can become gravitationally unstable, which can lead them to collapse and potentially form stars. Simulations of filaments suggest that a single filament may actually fragment into multiple stars distributed along it's length.

The NGC 1333 molecular cloud. Contours
show the distribution of dense gas, while
 the red dots show the very young stars that
are still forming (Credit: Gutermuth et al. 2008)
One of the predictions of this model of filamentary star formation is that the densest gas in star forming regions should be distributed in filaments and that the youngest stars - those that are actually still forming - should be found within these filaments.

The image to the right shows that the distribution of very young stars (red dots) in the star forming molecular cloud NGC 1333 are almost entirely projected against (and therefore likely to be within) the filamentary dense gas shown by the contours.

This provides strong evidence that stars form directly out of the dense, filamentary gas in molecular clouds.

Not only do the young stars trace the spatial distribution of the dense gas, but they also appear to have very similar motions. Numerous studies have measured the radial velocities (the speed of an object along the line of sight - i.e. towards or away from us) of stars such as these and found that they are almost identical to the velocities of the gas (e.g., Walsh et al. 2004Hacar et al. 2016).

Figure showing the Orion A molecular cloud. The
background greyscale image show the dense gas
while the red dots show very young stars
(Credit: Megeath et al. 2012)
This picture where star formation occurs in dense gas is seen in all the star forming regions we've studied, and not just on small scales, but also on the much larger scales of giant molecular clouds such as the Orion A cloud, shown on the left.

This is important for understanding star formation because it not only tells us where stars will form, but also provides clues as to how stars form (i.e. the process requires very dense gas) and what must occur to set up the necessary conditions (i.e. the gas in molecular clouds must be compressed in dense filamentary structures that become self-gravitating).

It is also important for understanding star clusters because it tells us not just the initial spatial distribution of groups of stars that may go on to form clusters, but also their kinematics, i.e. how those stars are moving.

The next question to ask is how do these elongated and substructured groups of stars evolve into the classic centrally-concentrated images of star clusters we see so often? But that's a question for another day!