PDA

View Full Version : Free-floating Planets Number



whimsyfree
2012-Jan-12, 12:08 PM
Post edit by admin: I have extracted this from the "Fun Papers ..." thread to start a new thread with whimsyfree's post.

http://arxiv.org/abs/1201.2175 Planet-Planet Scattering Alone Cannot Explain the Free-Floating Planet Population.
Dimitri Veras, Sean N. Raymond

This paper gets off to a bad start with Eq. 1, which is wrong.

Nfree
------- = fgiant funstable nejec ,
Nstars

where the three factors on the RHS are per star rates, Nfree is the number of unbound planets and Nstars is the number of main sequence stars. This cannot be correct. Consider periods in which the galaxy is in approximate steady state, i.e. the numbers and masses of stars entering the main sequence is approximately equal to the numbers and (formation) masses of those leaving. Over such a period Nstars is constant but Nfree is always increasing and the rates on the RHS are presumably constant. To correct this one has to put the number of stars which have ever formed in the galaxy in the denominator of the LHS.

They also misrepresent Sumi et al. (2011). Their per star figure of 1.8 refers to the number of unbound and distant Jupiter mass planets. Sumi et al. state that up to 0.4 planets per star may be bound in orbits a<500AU. Correcting these errors considerably reduces the required nejec.

kzb
2012-Jan-12, 05:54 PM
On the face of it, that seems a fair point about the equation.
Maybe the majority of stars that have ever formed are still around today, so the error caused is small? In fact if you include stellar remnants in Nstars, it's probably not far off correct.

IsaacKuo
2012-Jan-12, 09:17 PM
The majority of stars are red and orange dwarfs, and the universe isn't old enough for them to have left the main sequence yet.

But I suspect things could be highly skewed by short-lived stars--especially a hypothesized early burst of Population III stars. Even though short-lived stars would only represent a tiny fraction of the observed stars, they could have resulted in oodles of rogue planets. Besides potentially ejecting planets during initial planet formation, short-lived stars tend to die in ways the involve serious reductions in mass. Planets which were previously in orbit could suddenly find themselves with escape velocity due to the reduced mass of the stellar remnant.

neilzero
2012-Jan-12, 10:46 PM
Yes there are a bunch of minor factors. Suppose we add half of the number of compact stars to the number in the denominator? That assumes some compact stars and "planets" were formed before there were any main sequence stars, and that a few mergers, and framentations in all classes have occured. Did we decide Mercury mass is the minimum that can be called a planet = It may be forever before we can determine if most bodies have cleared their orbit of debris?
Also it is probable that some "planets" were ejected from brown dwarfs. Former moons of giant planets, are likely out there, and might be counted as free planets if they have more mass than Mercury. Neil

IsaacKuo
2012-Jan-12, 11:08 PM
Yes there are a bunch of minor factors. Suppose we add half of the number of compact stars to the number in the denominator? That assumes some compact stars and "planets" were formed before there were any main sequence stars, and that a few mergers, and framentations in all classes have occured. Did we decide Mercury mass is the minimum that can be called a planet = It may be forever before we can determine if most bodies have cleared their orbit of debris?
Also it is probable that some "planets" were ejected from brown dwarfs. Former moons of giant planets, are likely out there, and might be countedas free planets if they have more mass than Mercury. Neil

This paper was only concerned with large gas giants, because that is what could be detected by microlensing.

whimsyfree
2012-Jan-13, 02:02 AM
The majority of stars are red and orange dwarfs, and the universe isn't old enough for them to have left the main sequence yet.


This is true but two other factors need to be taken into account. Partly they are so common because they are so long-lived. It is necessary to consider the distribution of masses at formation rather than the current distribution. The former distribution will still show a majority of small stars but will not so skewed as the latter. The second factor you allude to below.


But I suspect things could be highly skewed by short-lived stars--especially a hypothesized early burst of Population III stars. Even though short-lived stars would only represent a tiny fraction of the observed stars, they could have resulted in oodles of rogue planets. Besides potentially ejecting planets during initial planet formation, short-lived stars tend to die in ways the involve serious reductions in mass. Planets which were previously in orbit could suddenly find themselves with escape velocity due to the reduced mass of the stellar remnant.

This is also important. Red dwarfs are numerous but jovian mass planets around them are rare, so their contribution to the population of unbound Jovians may not be great. Little is known about the planetary systems of high mass stars but it seems plausible that they could be more populated than those of sun-like and smaller stars. The current population of K-M dwarfs is as many as there have ever been, but the past population of O-F dwarfs is much greater than the present population, and it is these stars which have the potential to have contributed disproportionately to the unbound Jovian population, both by planet-planet scattering and end-of-life loss.

So it would be necessary to take into account both the initial mass function of stars and the stellar formation rate in estimating Nstars. This is a daunting task complicated by the fact that neither quantity is constant. Perhaps NZ's idea of adding in all (I can't see any reason to use half) the stellar remnants would work. Are there reliable estimates for their number?

kzb
2012-Jan-13, 12:35 PM
Whimseyfree wrote:
...past population of O-F dwarfs is much greater than the present population, and it is these stars which have the potential to have contributed disproportionately to the unbound Jovian population...

That's a good point, it's the initial mass function you need, not the currently observed mass function.

But acting against this factor is the metallicity factor. You need at least solar metallicity before Jovian planets occur. Lower metallicity stars have Neptunes and super-earths.

So the early generations of O-F stars may not have had sufficient metallicity to have Jovian planets. The level of metallicity required would only have been reached later in galactic history.

whimsyfree
2012-Jan-14, 11:24 AM
You need at least solar metallicity before Jovian planets occur. Lower metallicity stars have Neptunes and super-earths.


I don't believe that. Do you have a reference?

kzb
2012-Jan-14, 05:41 PM
I don't believe that. Do you have a reference?

I'm generalising, of course. I got the idea from the HARPS papers of a few months back. I may not be 100% correct but I still think the general idea is sound -a minimum degree of metallicity would have to be arrived at before you got a significant number of Jovians being formed in the galaxy.

Kullat Nunu
2012-Jan-14, 08:26 PM
You need at least solar metallicity before Jovian planets occur.

That is not true. Jovian planets are rarer among metal-deficient stars, but certainly not uncommon.

Correlation diagrams from the Extrasolar Planets Encyclopaedia: stellar metallicity vs. planet mass:
16151

Close-up (M >= 0.1 MJ, metallicity < 10^0.0 solar):
16152

As you can see, there are plenty of Jovian planets there, even one anomalously metal-poor star, HIP 13044, which may be from a satellite galaxy consumed by the Milky Way.

whimsyfree
2012-Jan-15, 05:11 AM
A recent arXiv paper (http://arxiv.org/abs/1201.2412) is relevant to Isaac's idea. At least for solar mass stars it seems that planet escape due to stellar mass loss is unlikely. Larger stars that go supernova would be a different proposition.

IsaacKuo
2012-Jan-15, 08:37 AM
A recent arXiv paper (http://arxiv.org/abs/1201.2412) is relevant to Isaac's idea. At least for solar mass stars it seems that planet escape due to stellar mass loss is unlikely. Larger stars that go supernova would be a different proposition.
Not relevant, since I was specifically talking about short-lived stars. Solar mass stars are long-lived.

Short-lived stars tend to go supernova, except at the extreme high end and maybe on the low end, depending on your definition of "short-lived".

whimsyfree
2012-Jan-16, 08:50 AM
Another paper from the arXiv thread is apropos.


http://arxiv.org/abs/1201.2687 How many things smaller than stars are out there orbiting in the galaxy, but not bound to a specific star? I wonder about this from time to time. This paper estimates that there could be 100,000 such objects greater than about a quarter the mass of our Moon in the galaxy for every star in the galaxy. The paper makes predictions of how many such objects will be observed by Gaia (basically only Jupiter mass and greater), and suggests a dedicated orbiting mission for the next decade to stare at a patch of the Southern Sky and make more precise and shorter time interval gravitational lensing measurements than we have done so far... to look for these things and get more concrete numbers.

However the methodology used in this paper is hard to accept. Based on some counts of objects of mass >1e-3 Msun they fit a power law and then extrapolate this to bodies of mass 1e-8 Msun. This is analogous to fitting a power law to the counts-height relationship of Americans over 200kg and then extrapolating this to conclude there must be trillions of Americans between 2 and 20 milligrams in weight.

There are some other issues with the paper which I may address later, if I have time.

kzb
2012-Jan-16, 12:37 PM
That is not true. Jovian planets are rarer among metal-deficient stars, but certainly not uncommon.

Correlation diagrams from the Extrasolar Planets Encyclopaedia: stellar metallicity vs. planet mass:
16151

Close-up (M >= 0.1 MJ, metallicity < 10^0.0 solar):
16152

As you can see, there are plenty of Jovian planets there, even one anomalously metal-poor star, HIP 13044, which may be from a satellite galaxy consumed by the Milky Way.

Those two plots absolutely back up the main point I was trying to make. That is, you cannot just integrate the number of the O-F stars formed in the galaxy up to the current epoch and expect that to be directly propotional to the number of Jovians, based on current populations. You have to take into account that metallicity has increased over time, and that means (on average), earlier O-F stars would've contributed fewer Jovian planets.

IsaacKuo
2012-Jan-16, 03:30 PM
Another paper from the arXiv thread is apropos.



However the methodology used in this paper is hard to accept. Based on some counts of objects of mass >1e-3 Msun they fit a power law and then extrapolate this to bodies of mass 1e-8 Msun. This is analogous to fitting a power law to the counts-height relationship of Americans over 200kg and then extrapolating this to conclude there must be trillions of Americans between 2 and 20 milligrams in weight.

They looked at the power law that applies to KBOs. This seems entirely reasonable to me.

Your accusation is off base. It would be more like figuring out the counts-mass relationship of Americans in Baltimore and extrapolating the demographics of the continental USA from counts of those over 200kg in mass. Sure, there will be some error if the distribution in Baltimore is skewed compared to the rest of the country. But it's a reasonable starting point.

kzb
2012-Jan-16, 06:42 PM
They looked at the power law that applies to KBOs. This seems entirely reasonable to me.

Your accusation is off base. It would be more like figuring out the counts-mass relationship of Americans in Baltimore and extrapolating the demographics of the continental USA from counts of those over 200kg in mass. Sure, there will be some error if the distribution in Baltimore is skewed compared to the rest of the country. But it's a reasonable starting point.


Quote from the paper:

Objects with mass < 10−2M⊙ are believed
to originate from two distinct processes. Between the
Jupiter mass and the deuterium-burning mass, many
of these objects may form similar to stars by gravita-
tional fragmentation. Below Jupiter masses, they likely
are born in protoplanetary disks and dynamically-ejected
during the evolution of the system. It is unknown from
a theoretical perspective whether there is a smooth con-
tinuation of the mass function at the dividing mass that
separates these populations.

I've not read the whole paper yet. BUT, observations of star-forming regions show that planetary-mass objects are formed less frequently than stars. I think the population fraction is something like 14% (from memory).

If we are saying there are X2 as many free-floating Jupiters as there are visible stars, the majority can only have come from ejections. I think their smooth power law can only be very approximate !

chornedsnorkack
2012-Jan-16, 09:43 PM
Good populations of young massive low metallicity stars can be found in Magellanic clouds.

How do the planet abundances in Clouds compare with Milky Way?

Drunk Vegan
2012-Jan-17, 12:22 AM
Good populations of young massive low metallicity stars can be found in Magellanic clouds.

How do the planet abundances in Clouds compare with Milky Way?

We don't know yet because no planets discovered there yet - too distant for current detection techniques.

whimsyfree
2012-Jan-17, 03:55 AM
Those two plots absolutely back up the main point I was trying to make. That is, you cannot just integrate the number of the O-F stars formed in the galaxy up to the current epoch and expect that to be directly propotional to the number of Jovians, based on current populations.

I don't know what point you were trying to make but that is not what you said. You claimed that sub-solar metallicity stars had no Jovians at all. That's false.


They looked at the power law that applies to KBOs. This seems entirely reasonable to me.

Your accusation is off base. It would be more like figuring out the counts-mass relationship of Americans in Baltimore and extrapolating the demographics of the continental USA from counts of those over 200kg in mass. Sure, there will be some error if the distribution in Baltimore is skewed compared to the rest of the country. But it's a reasonable starting point.

No, you are wrong. They may have looked at the power that applies to KBOs but that was not the data to which they fitted their model. Where does it say otherwise? In fact their "1e5" extrapolation is based on a model in Sumi et al. (2011) that is in turn based on observations of micro-lensing events triggered by objects that are probably at least Jovian.

The fact that their sample was even more unrepresentative than my analogy is not a point in their favor.


Quote from the paper:

...
I've not read the whole paper yet. BUT, observations of star-forming regions show that planetary-mass objects are formed less frequently than stars. I think the population fraction is something like 14% (from memory).

If we are saying there are X2 as many free-floating Jupiters as there are visible stars, the majority can only have come from ejections. I think their smooth power law can only be very approximate !

I think it can only be speculation. Note that they cut off their extrapolation at 1e-8. Why they do this? The reason is that as you continue the extrapolation to smaller masses the number of objects and their total mass approaches infinity. They cut it off before it became obviously ridiculous.

It is very likely that small unbound bodies are very numerous. Migrating gas giants in protoplanetary disks doubtless eject planetesimals. Solar mass stars will lose their Oort cloud in late life. Their methodology is still wrong.

kzb
2012-Jan-18, 12:51 PM
Whimseyfree wrote:
I don't know what point you were trying to make but that is not what you said. You claimed that sub-solar metallicity stars had no Jovians at all. That's false.

It's called hyperbole. Anyhow let's leave that, I think my point about metallicity is clear.

About the mass function, it's a little unclear to me what they have used at the lower mass range. They discuss the distribution of KBO masses, but it's not clear to me whether this is actually used in their model or not.

However, I don't think what they are doing is that unreasonable. They've constrained their numbers relative to MACHO searches at >10^-7 solar mass, and also relative to the mass density of the galactic disk. They are saying these bodies might well exist in these numbers, so please can we have some resources to find them? I for one would be interested in the result and I bet others are too.

Another point I have thought about: the Ejected Planet Mass Function (EPMS) will not be the same as the PIMF (Planet Initial Mass Function). This is because larger members of the planet population are the ones doing the ejecting, and it is smaller bodies that get ejected, on average. So, relative to the PIMF, the EPMS will be sparser for Jovians, but relatively rich in smaller bodies.

This means that the 2 free-floating Jovians per star is the tip of the iceberg. There should be lots more FF Neptunes, Superearths etc still not detected.

That then begs the question, on statistical grounds, there really ought to be several of these objects closer to us than the Alpha Centauri system. So why have we not detected them yet?

kzb
2012-Jan-19, 01:04 PM
Going back to the original paper under discussion here:

http://arxiv.org/abs/1201.2175 Planet-Planet Scattering Alone Cannot Explain the Free-Floating Planet Population.
Dimitri Veras, Sean N. Raymond

This is my basic understanding of what they are saying in this paper:

We have models of planetary system formation and evolution. These models predict something in the region of 30% (maybe 50% tops) of giant planets are ejected from their home system. Since we observe that only a minority of stellar systems currently have even one giant planet, there is an unexplained discrepancy with the Sumi et al observation of 1.8 free-floating giant planets. In fact when you go through the models in detail, you need something like 25 giant planets per star to account for the observation. This is untenable, (a) because exoplanet detections don't support anything like this number and, (b) the formation models don't produce this number.

I'm not sure what point they are making though. Are they saying the Sumi et al number of free-floaters must be wrong? Or are they saying the models must be wrong? I think the former, they are saying there is something wrong with the estimate of 1.8 free floaters per star.

However, I am not too sure. How do we know that a large proportion (actually the majority) of planets formed are not ejected early on? How well constrained by actual observation is this?

neilzero
2012-Jan-19, 02:55 PM
I'm guessing, a free floating Jupiter with same cloud top temperature should be detectable in infrared if it has as much proper motion as Barnard's Star in seconds of arc per century up to 2 light years away. 1/3 light year for a simular Neptune object at Neptune cloud top temperature? Colder, smaller or with less proper motion, we likely need a more sensitive seach. The density (closer than Centarii Proxima) of free floating planets Jupiter to Neptune size could be 1/10 th the average (this century) for our galaxy, but 1% is very improbable.
Almost as sensitive, is finding them when they ocult = transit a distant star. Is there a third method? Doubling the funding will likely double the sensitivity of any method? Neil

IsaacKuo
2012-Jan-19, 03:08 PM
http://arxiv.org/abs/1201.2175 Planet-Planet Scattering Alone Cannot Explain the Free-Floating Planet Population.
Dimitri Veras, Sean N. Raymond
[...]
I'm not sure what point they are making though. Are they saying the Sumi et al number of free-floaters must be wrong? Or are they saying the models must be wrong? I think the former, they are saying there is something wrong with the estimate of 1.8 free floaters per star.
No, they seem to accept the observed estimate of 1.8 free floaters per star.

Call me a crazy mind reader, but I think they're saying that planet-planet scattering alone cannot explain the free-floating planet population. In other words, they accept both the observed number of free floating planets as well as the models for planet-planet scattering. They suggest that something else accounts for the extra number of free floating planets.

The key word is "alone". This suggests that other formation mechanisms/planetary history may account for the observed free-floating planet population.

If they doubted the planetary models, the title would be something like "Planet-Planet Scattering models invalidated by Free-Floating Planet Population".

If they doubted the free-floating planet observations, the title would be something like "Free-Floating Planet Observations inconsistent with Planet-Planet Scattering."

But the wording of the title, and in particular the inclusion of the word "alone" implies that both the observations and models are accepted, but that there needs to be further investigation into alternate mechanisms by which free-floating planets come into being.

Certainly there are other possible means by which free-floating planets come into being. The most obvious would be if they could form in interstellar space independently of a full blown star system.

IsaacKuo
2012-Jan-19, 03:16 PM
This means that the 2 free-floating Jovians per star is the tip of the iceberg. There should be lots more FF Neptunes, Superearths etc still not detected.

That then begs the question, on statistical grounds, there really ought to be several of these objects closer to us than the Alpha Centauri system. So why have we not detected them yet?
They would be awesomely difficult to detect. It is a mainstream scientific hypothesis that the orbit of Sedna is due to an as yet undiscovered gas giant or other large planet within the Oort cloud. Obviously, this hypothesis would not be mainstream if such a thing were easy to detect.

WISE was expected to be able to detect a nearby brown dwarf, if it were close enough. But even WISE wouldn't have been able to detect a planet, which would be colder.

As far as I know, there isn't really a good way to detect such a planet yet. Microlensing only works for far away planets, due to the extremely long "focal distance". My guess is that the best chance to detect such planets would be with some sort of occultation sensing experiment. This is a bit like microlensing, but it's based on the principle of diffraction as the planet passes in front of the other star.

kzb
2012-Jan-19, 06:35 PM
Isaackuo wrote:
Certainly there are other possible means by which free-floating planets come into being. The most obvious would be if they could form in interstellar space independently of a full blown star system.

But this is contradicted by observations of star-forming regions. There are apparently fewer planetary mass objects formed than stars.

To be fair, their close-out section does include other sources

(quote)
Other potential sources of free-floating planets exist.
One such source is dynamical ejection from multiple-star
systems; only about two-thirds of all stars are single stars
(Lada 2006) so multiple-star systems may provide a com-
parable contribution to free-floaters. Another potential
source arises from external forces such as passing stars,
galactic tides, or – most relevant for close (. 100 AU)
planets – perturbations while stars are still in their birth
clusters. [continues...]

So now I am beginning to get it. The models only account for ejections caused by intra-system effects, and in single-star systems to boot.

The conclusion then is that more planets are formed than we currently see. A larger number are formed, and a large proportion are ejected or at least caused to escape by external influences.

IsaacKuo
2012-Jan-19, 07:30 PM
Isaackuo wrote:
Certainly there are other possible means by which free-floating planets come into being. The most obvious would be if they could form in interstellar space independently of a full blown star system.

But this is contradicted by observations of star-forming regions. There are apparently fewer planetary mass objects formed than stars.
I'll admit I'm not familiar with the science of these observations. How could they even detect planetary mass objects being formed?

Also, I don't assume that star-forming regions are the only places where free-floating planets might form in interstellar space.

whimsyfree
2012-Jan-20, 05:47 AM
I'll admit I'm not familiar with the science of these observations. How could they even detect planetary mass objects being formed?


Young gas giants are hot. They've looked for isolated hot young gas giants in star forming regions and not found many.


Also, I don't assume that star-forming regions are the only places where free-floating planets might form in interstellar space.

Well I don't know where else would be likely.

The formation of isolated Jupiter or greater mass planets from collapsing gas clouds is conceivable, but the idea that galactic clouds could condense leaving behind a solid object less massive than the moon seems far fetched. A different mechanism would be required to form such small bodies. That is one of the reasons it is inappropriate to extrapolate a mass distribution for the first kind of object to the second.

IsaacKuo
2012-Jan-20, 07:48 AM
Young gas giants are hot. They've looked for isolated hot young gas giants in star forming regions and not found many.
I would greatly appreciate an example reference for this (that isn't behind a paywall). Gas giants would cool off more quickly than brown dwarfs, so they could be hard to detect by the time they leave a Bok globule or the opaque region of dust where they formed.

Well I don't know where else would be likely.
Only a fraction of Bok globules are observed to be star forming. It could be that Bok globules are generally capable of forming smaller mass objects but only a fraction of them have the conditions necessary to form stars.

The formation of isolated Jupiter or greater mass planets from collapsing gas clouds is conceivable, but the idea that galactic clouds could condense leaving behind a solid object less massive than the moon seems far fetched. A different mechanism would be required to form such small bodies. That is one of the reasons it is inappropriate to extrapolate a mass distribution for the first kind of object to the second.
There's also the bottom-up theory of formation where knots of dust coalesce to form "kernels" for further formation. In that case, there should be many small bodies compared to the larger ones.

Pure gas collapse, if it occurs, would occur in dust free regions where it's more readily visible. In contrast, dust knots would tend to occur where dust is dense, where the process would be less visible or completely obscured.

kzb
2012-Jan-20, 12:41 PM
Isaackuo wrote
I would greatly appreciate an example reference for this

Here's a reference I found quickly just now. There are others, and more recent ones too.

Substellar Objects in Nearby Young Clusters (SONYC): The bottom of the Initial Mass Function in NGC1333
A Sholz et al

http://arxiv.org/abs/0907.2243

Quote:

There are two alternative estimates for the frequency
of planemos in the ONC in the literature: According to
Lucas et al. (2006), 7.5% (1-14%) of the total population
in the cluster have a mass of 0.003-0.015M⊙, where the
wide range of possible values is mostly a reflection for
age uncertainty. Lucas et al. (2005) find a drop by factor
two in the mass function at the Deuterium burning limit,
a deviation from a flat mass function in the substellar
regime. Their upper limit for the fraction of planemos in
the total cluster population is 10-13%.

The consensus of these types of study is apparently that the mass function declines in frequency at the low end. There are fewer PMOs in star-forming regions than stars. Clearly there are massive observational challenges in these studies, but the authors appear careful to account for these in their estimates.

IsaacKuo
2012-Jan-20, 01:58 PM
Isaackuo wrote
I would greatly appreciate an example reference for this

Here's a reference I found quickly just now. There are others, and more recent ones too.

Substellar Objects in Nearby Young Clusters (SONYC): The bottom of the Initial Mass Function in NGC1333
A Sholz et al

http://arxiv.org/abs/0907.2243
Thanks! The keyword I was failing to search for was "substellar objects". I'll read this reference, and related references when I get the chance.


The consensus of these types of study is apparently that the mass function declines in frequency at the low end. There are fewer PMOs in star-forming regions than stars. Clearly there are massive observational challenges in these studies, but the authors appear careful to account for these in their estimates.

If we assume star forming regions are the only place that planetary mass objects form, and we assume that this frequency distribution is representative across the entire history of the galaxy, then these constraints would apply to both independently formed free-floating planets and also early ejected free-floating planets. In other words, these surveys should also see any free floating planets which were ejected early on, as well as ones which formed themselves independently.

Unless...

There would be a bias against ejected planets or independently formed free-floating planets with high velocities. Such high velocities might be a result of, say, being pushed by gas jets or by being initially formed by knots of dust within gas jets. In either case, they'd be underrepresented in cluster surveys simply because they'd leave the clusters so quickly.

kzb
2012-Jan-20, 04:31 PM
Isaackuo wrote:
If we assume star forming regions are the only place that planetary mass objects form, and we assume that this frequency distribution is representative across the entire history of the galaxy, then these constraints would apply to both independently formed free-floating planets and also early ejected free-floating planets. In other words, these surveys should also see any free floating planets which were ejected early on, as well as ones which formed themselves independently.

I'm not sure about this. The PMOs detected in star forming regions are formed by fragmentation an gravitational collapse in the cloud, in the same way as stars. The young stars themselves have their own circumstellar discs which are in an early stage of development. My take on it was that any planets forming in these discs (by the planet formation mechanism) would not be resolved separately to their host stars, and so would not be counted in the surveys. Although if any escape from orbit early on, then I guess they could be counted.

But it won't be a complete census of PMOs, because the planetary systems are still in a state of development in these areas. Somehow a lot of planets must escape from their star after this early stage, to account for the 1.8 PMOs per star result.

In either case, they'd be underrepresented in cluster surveys simply because they'd leave the clusters so quickly.

Somewhere I have a paper where they analyse the object density against distance from the cloud centre, and continue for some distance outside the cloud, precisely to detect these postulated escapees. They didn't find any. But when you think about it, lower mass objects should end up with a higher dispersion velocity after several interractions with more massive objects, so there is reason to think they would be preferentially ejected from the cloud.

IsaacKuo
2012-Jan-20, 08:39 PM
Isaackuo wrote:
If we assume star forming regions are the only place that planetary mass objects form, and we assume that this frequency distribution is representative across the entire history of the galaxy, then these constraints would apply to both independently formed free-floating planets and also early ejected free-floating planets. In other words, these surveys should also see any free floating planets which were ejected early on, as well as ones which formed themselves independently.

I'm not sure about this. The PMOs detected in star forming regions are formed by fragmentation an gravitational collapse in the cloud, in the same way as stars. The young stars themselves have their own circumstellar discs which are in an early stage of development. My take on it was that any planets forming in these discs (by the planet formation mechanism) would not be resolved separately to their host stars, and so would not be counted in the surveys. Although if any escape from orbit early on, then I guess they could be counted.
That is what I meant by early ejected planets.

But it won't be a complete census of PMOs, because the planetary systems are still in a state of development in these areas. Somehow a lot of planets must escape from their star after this early stage, to account for the 1.8 PMOs per star result.
With the caveat that we are assuming that planetary mass objects aren't also formed elsewhere, and we are assuming that the current formation fraction has always been the case for the entire history of the galaxy (including the galaxy's early history).

I highlighted those two assumptions because I'd question both.

kzb
2012-Jan-21, 05:59 PM
IsaacKuo: Yes I think this is an interesting point. The stated purpose of these surveys of star forming regions is precisely to pin down the low end of the initial mass function. That is why they are doing it. Maybe I'd better re-read the papers but I don't recall planetary ejection being mentioned. I guess that would confound these studies, because we don't know how complete the planetary ejection is at the time of the survey.

As to the stellar IMF, from what I've read, the constancy of this over time and in all environments is almost a central dogma of modern astronomy. Yes you can find differences between individual clouds, but averaged over a big enough volume it is a constant. I mean you could explain galactic rotation curves without recourse to non-baryonic matter, if you allowed a systematic change in the IMF with galactic radius.

IsaacKuo
2012-Jan-21, 06:12 PM
My main concern is the currently mysterious Population III stars. We know the mass function was different back then, but there's little agreement on what it must have been. While it would seem implausible that there would be planets so early on, we have found planets as far back as we could (and earlier in the universe than we expected). Our extreme lack of data on population III stars leaves a wide range of possibilities.

antoniseb
2012-Jan-21, 07:30 PM
My main concern is the currently mysterious Population III stars. ...

One thing we don't know is how many Jupiter-sized planets were formed by terrestrial planet-forming mechanism (in a disk), and how many were formed as small star-cores that never got big enough to be stars. Population III stars could have as many of that type of planet as modern high-metal stars.

whimsyfree
2012-Jan-23, 01:02 AM
I would greatly appreciate an example reference for this (that isn't behind a paywall). Gas giants would cool off more quickly than brown dwarfs, so they could be hard to detect by the time they leave a Bok globule or the opaque region of dust where they formed.


kzb already.



Only a fraction of Bok globules are observed to be star forming. It could be that Bok globules are generally capable of forming smaller mass objects but only a fraction of them have the conditions necessary to form stars.

There's also the bottom-up theory of formation where knots of dust coalesce to form "kernels" for further formation. In that case, there should be many small bodies compared to the larger ones.


Do you have any references that explain this theory? If it is correct then it would be inappropriate to extend the power law derived from objects with a different formation mechanism to this new hypothesized population.



If we assume star forming regions are the only place that planetary mass objects form, and we assume that this frequency distribution is representative across the entire history of the galaxy, then these constraints would apply to both independently formed free-floating planets and also early ejected free-floating planets. In other words, these surveys should also see any free floating planets which were ejected early on, as well as ones which formed themselves independently.


So far I don't have a problem.


Unless...

There would be a bias against ejected planets or independently formed free-floating planets with high velocities. Such high velocities might be a result of, say, being pushed by gas jets or by being initially formed by knots of dust within gas jets. In either case, they'd be underrepresented in cluster surveys simply because they'd leave the clusters so quickly.

Seems highly speculative. Do you have any numbers that show this is plausible? i.e. veclocities, ages, distances travelled? The idea of Jovians in star forming regions being accelerated to high velocities by gas jets in (astronomically) short periods of times seems unlikely.



I'm not sure about this. The PMOs detected in star forming regions are formed by fragmentation an gravitational collapse in the cloud, in the same way as stars. The young stars themselves have their own circumstellar discs which are in an early stage of development. My take on it was that any planets forming in these discs (by the planet formation mechanism) would not be resolved separately to their host stars, and so would not be counted in the surveys. Although if any escape from orbit early on, then I guess they could be counted.

But it won't be a complete census of PMOs, because the planetary systems are still in a state of development in these areas. Somehow a lot of planets must escape from their star after this early stage, to account for the 1.8 PMOs per star result.

In either case, they'd be underrepresented in cluster surveys simply because they'd leave the clusters so quickly.

Somewhere I have a paper where they analyse the object density against distance from the cloud centre, and continue for some distance outside the cloud, precisely to detect these postulated escapees. They didn't find any. But when you think about it, lower mass objects should end up with a higher dispersion velocity after several interractions with more massive objects, so there is reason to think they would be preferentially ejected from the cloud.

I'm not sure that's right. Lower mass objects are preferentially ejected but is there evidence that their ejection velocities are greater, and, if so, by how much? Planet-planet scattering is a form of gravity assist, and so the delta-v should be limited to twice the orbital velocity. A tiny object gets no more out of it than a planetary mass object.

kzb
2012-Jan-23, 12:35 PM
Whimseyfree wrote:
I'm not sure that's right. Lower mass objects are preferentially ejected but is there evidence that their ejection velocities are greater, and, if so, by how much? Planet-planet scattering is a form of gravity assist, and so the delta-v should be limited to twice the orbital velocity. A tiny object gets no more out of it than a planetary mass object.

I was thinking more about the situation post-ejection from the planetary system, but still within the stellar cluster. The velocity and path of a free PMO in a stellar cluster will be affected by gravitational interactions with the stars it passes nearby. The velocities of the stars themselves will be much less affected, simply because of the difference in mass.

I must admit I have no idea of the size of this effect of this in relation to what we are discussing. It may be there are just not enough close interactions, even in a dense cluster, to cause preferential ejections of PMOs relative to stars.

kzb
2012-Jan-23, 12:42 PM
One thing we don't know is how many Jupiter-sized planets were formed by terrestrial planet-forming mechanism (in a disk), and how many were formed as small star-cores that never got big enough to be stars. Population III stars could have as many of that type of planet as modern high-metal stars.

I thought Pop III stars couldn't form planets, at least following the core accretion model?

IsaacKuo
2012-Jan-24, 03:08 PM
I thought Pop III stars couldn't form planets, at least following the core accretion model?
The discovery of PSR B1620-26 b didn't cause any crisis in the theory of planetary formation. If gas giants can form by pure gas collapse, then it's a natural explanation for PSR B1620-26 b. On the other hand, if the mass of BSR B1620-26 b is too small for pure gas collapse, then core accretion is the obvious candidate--suggesting that there may have actually been sufficient dust early on.

A.DIM
2012-Jan-24, 03:46 PM
By whatever mechanism they're formed, if free floaters are so common they provide a possible mode of interstellar panspermia.
Fascinating!

kzb
2012-Jan-24, 06:41 PM
The discovery of PSR B1620-26 b didn't cause any crisis in the theory of planetary formation. If gas giants can form by pure gas collapse, then it's a natural explanation for PSR B1620-26 b. On the other hand, if the mass of BSR B1620-26 b is too small for pure gas collapse, then core accretion is the obvious candidate--suggesting that there may have actually been sufficient dust early on.

But has it been said that the primary of this planet was a bona fide Pop III star?

Is it plausible that the PMO in this case could've formed independently (as a star), and be a member of the low-mass end of the IMF. It got captured, sometime after formation, in the dense stellar environment of a globular cluster.

Whatever, I don't think this planet is proof that Pop III stars formed planets in a circumstellar disc like current stars. Could the disc even last long enough?