Star is 14.5 billion years old

Le 13/03/13 08:30, Martin Hardcastle a ecrit :
In article ,
jacob navia wrote:
Mr Hardcastle

(I spent a number of years in order no longer to have to be called
'Mr' in an academic context (-: )

Le 10/03/13 13:04, Martin Hardcastle a ecrit :
As far as I understand this, dark clouds that generate protostars must
cool at temperatures between 10-20 Kelvins. This is possible TODAY
because the CMB is at 2.75 Kelvin.

You postulate that at a temperature 100 degrees HIGHER those clouds can
form and moreover cool enough to reach those 10K without reaching
equilibrium with the CMB in 300 million years.


Excuse me, no offsense intended! I thought that addressing you as "Mr"
would let the discussion proceed in a calm tone but I see I got it 100%
wrong

Sorry Martin :-)

a) I don't 'postulate' any of this. I'm telling you what the standard
model for this stuff is. I didn't generate it.


Yes, I know, but the "you" was used in a general sense, not you in
particular. I am French and my english could be wrong but I was sure
you could use "you" meaning not somebody in particular but as a general
case.

b) No, that's not what I'm saying. The general idea appears to be that
the cooling in the early universe happens through lines of molecular
hydrogen. In the early universe, this can happen at much higher
temperatures than are associated with molecular hydrogen today,
because there is no ultraviolet light around to dissociate it. There
are some fairly classic papers about the details of cooling through
molecular hydrogen in the early universe, see e.g. Tegmark et al
http://adsabs.harvard.edu/abs/1997ApJ...474....1T). There is no
*intrinsic* requirement that stars form out of cold gas, it just
happens to work that way in the local universe, where dust provides
both a shield from the UV and a substrate on which molecular hydrogen
can form.


Interesting article. It proposes another way of creating stars in the
supposed "early" universe. That article could be a big blow for my line
of reasoning, but fortunately for me it speaks of HUGE gas clouds (more
than 1000 solar masses) that would create enormous stars. Here we are
speaking of a star smaller than the sun.

This mechanism is referenced in the article for the FIRST stars. One of
the points there is that those stars did NOT have the problem of UV
radiation since they should have been well... the first ones.

The star we are talking about however is NOT a "first" star since it
has some iron content, it is a nth generation star so it MUST be
shielded from UV radiation of the other stars by a COLD dark cloud as
stars in a current star factory.

In 300 million years can a galaxy (even protogalaxy) develop enough to
have star factories and all that?

Yes, in standard cosmology, they can. You can find this in pretty much
every paper about the early universe. Can you present a calculation to
show that they can't? Just saying 'it looks unlikely' whenever you see
something you don't like isn't science, I'm afraid.


Well, the initial density gradient must acquire enough matter from its
surroundings to form an object.

And let's calculate a bit, since you want some figures.

Suppose a big bang produced density gradient, and at its center some big
mass, atracting things in a radius of 5000 thousand light years. Our
galaxy has a radius of somewhere 50 000 light years, so a "proto"
galaxy (whatever that may be) should be a tenth of that.

A kilogram of hydrogen at 5000 light years has a fall time of

pi R ^(1.5)
--- x -------------
2 sqrt(2G(M+m))

where
R = 5000 light years = 9.4605284 x 10E15 x 5000 meters
M = 1E6 solar masses = 1.9891 x 10E30 kilograms x 10E6
m = 1 kg, let's forget that :-)
G = 6.67398 x 10E-11

I will print intermediate results to verify I did not make any mistake.
a = 9.46052*1000000000000000*5000
47302600000000000000
b = pow(a,1.5)
325332551124632433390390678348.06

M = M=1.9891E30
1989100000000000000000000000000

2*G*M = 265504272360000000000
c = sqrt(2*G*M)
16294301837.145
b/c = 19966031952531170380.610

Now we multiply by pi/2
31362569651705500132.87 seconds
993819861196.84 years

993.81 billion years

It would take our kg of hydrogen approx 1000 GIGA years to arrive to the
center...

OK, what happens if we do not have 1E6 solar masses but 1E9?

"b" above stays the same since it depends on the radius

2*G*M get's multiplied by 1000, the square root is now
515271066876.45, b/c is 631381368056973676.71
that multiplied by PI/2 is
991771533750630924.97 seconds
31427343452.94 years
31.427 Giga years.

OK?

A LOOOOOOOOOOOOOOOOOONG time :-)

And that with a density gradient having 1000 million masses of the sun!
Can those "impurities" appear in the aftermath of the big bang? Are they
compatible with the CMB smoothness?

Note that our own galaxy (not a "proto" galaxy) has a black hole at its
center of "only" 4 x 10E6 solar masses... For a mass of 1000 million
solar masses you would have to explain HOW that behemoth appears
immediately after a smooth big bang mass distribution, not an easy
task I presume.

BUT

Please correct me if I am wrong. You wanted calculations, I did some.
Are they correct?

Your move.


For example, a little googling turns up this review, relevant to this
whole thread: http://arxiv.org/abs/astro-ph/0409737 . Have a look at the
calculations of the formation redshift of protogalaxies in there. Do
you spot any errors?


Thanks for this reference. It is not at all bad for my point:

1) It says that the first protogalaxies will form at z = 30. This is the
age of this star!

2) Those galaxies at z=30 will form the first generation of stars:
quote
Taken together, these points strongly suggest that the first stars will
be very massive. Indeed, if this basic picture is correct, it is
difficult to see how accretion could be terminated early enough to
produce a solar mass star, since the predicted accretion rates
discussed earlier suggest that this mass of gas will build up in only
10-20 yr.
end quote

This star is smaller than the sun.

But I could have gotten something wrong of course. I will go in the next
days thorugh that paper again with more time. It is a very dense
paper and VERY long. But also it has some interesting points:

quote
We expect the first stars to form in small, H2-cooled protogalaxies,
with masses of 10^5-10^6 M_solar, at redshifts z = 30-40.
end quote

That is the redshift of this star. Yes, you can try to get it to 2 sigma
and bring it down to 6. But that is "cheating" really. And we ALL agree
at 3 sigma of course :-)

That paper is also interesting because of the openess with which the
authors discuss the myriads of parameters, assumptions (many of those
reasonable within the framework of a wrong BB theory) trying to figure
out "in silico" what happene after the supposed bang.

[Mod. note: yet again, non-ASCII characters fixed by hand... -- mjh]

In article ,
jacob navia writes:
My conviction is that it is not possible to postulate a
normal star forming just 100 million years or even 300 million years
after the supposed "bang".

You might have another look at that review article by Simon Glover:
http://xxx.lanl.gov/abs/astro-ph/0409737

On Wednesday, I went to a talk where the speaker commented that if it
weren't for "feedback" (something we don't really understand though
there are ideas about it), all the gas in the Universe would have
formed into stars by z=10. The theoretical problem is not forming
stars so soon, it's delaying most star formation until z=2 or so.

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA

In article ,
jacob navia writes:
You argue that there is a "general" tendency of misjudging the errors in
astrophysics measurements. This could be true, but would need some
data justifying your suppositions excuse me.

Eric's comment is relevant. My statement was based on quite a few
years of looking at (and generating) astronomical data. You are
welcome to do your own sampling.

The tendency doesn't mean anyone is doing anything incompetent or
evil. We authors estimate error bars based on every source of error
we can think of, but obviously we don't include the error sources we
haven't thought of. Those are sometimes well outside the limits we
expect, and the "wings" of the error distribution are distinctly
higher than Gaussian.

Note that the possibility that the observations are correct doesn't
even get mentioned!

I thought it was implied by putting the possibility they are wrong in
parentheses. To be clear, I expect the observations themselves are
correct, though there are stated uncertainties, real uncertainties,
and as always a remote possibility of a mistake. The oxygen
abundance is known to be uncertain, and other parts of the
theoretical framework for interpreting the observations may be wrong.

Yes, you can always try something but is it science really?

Science, at least a big part of it in my view, is keeping straight
what you know for sure, what you have reason to believe but aren't
sure of, and what is uncertain. (Obviously it's a matter of degree
of certainty, not discrete classes.) Where one puts a given
observation or theory is, to some extent, a matter of opinion, but in
general people who know about a given subject will have opinions
based on evidence and not too far apart. (There are some famous
exceptions to that last but not many.) An analogy is in professional
sports: fans can argue forever whether player A is/was better than B,
but everyone agrees that the Hall-of-Fame players are a lot better
than most others.

SW Do you remember a few years ago when claimed globular cluster ages
SW were 17 Gyr with error bars of only 1 or maybe 1.5 Gyr?

I would like the references concerning their "wrongness".

Google is helpful here. See
http://www.pnas.org/content/95/1/8.full.pdf
for an early review and
http://iopscience.iop.org/0004-637X/...6650.text.html
for a much later summary of the evidence. There are other references
in Google and in the second citation.

I see my memory of the history wasn't quite right; better stellar
physics, not only the Hipparcos distances, contributed to the
decrease in GC ages.

Last week astronomers discovered that the
THIRD nearest star from the sun was one of those, at only 6.5 light
years from this newsgroup...

Are you talking about the brown dwarf system discovered by WISE?
Brown dwarfs are quite different from G subdwarfs, and I know of no
evidence that the nearby system is old. G subdwarfs, despite the
name, are reasonably luminous and easily identified in color-
magnitude diagrams (e.g., with Hipparcos distances). I haven't
personally looked into details, but I'd be astonished if the local
measured density of G subdwarfs turns out to be wrong. (By the way,
G subdwarfs are not under-luminous for their mass. Instead they are
bluer than normal stars of the same mass because of the lack of
atomic absorption lines in their atmospheres.)

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA

On Fri, 15 Mar 13, Steve Willner wrote:
On Wednesday, I went to a talk where the speaker commented that if it
weren't for "feedback" (something we don't really understand though
there are ideas about it), all the gas in the Universe would have
formed into stars by z=10.

Uh oh, "dark feedback"... ;-)

On 3/15/2013 6:45 PM, jacob navia wrote:
...
Excuse me it CAN'T BE LOWER! That star is AT LEAST that age!

See Dan Riley's post on the meaning of error bars, and bear in mind
that stated error bars tend to be too small because some causes of
error are overlooked. (I'm not saying Bond et al. have done that,
only that it's a general tendency.)

I know what error bars are. They indicate the error that the scientist
that performs the observations thinks it is attached to the measurement.

Pleas reconsider. That is *not* what error bars mean.
The scientists do not think so at all. They are aware
that they do not know what the error is, but only know
a probability distribution for it.

Most importantly, this probability distribution is
wider than the error bars indicate. The scientists
know that the error can lie *outside* the error bars,
and for Gaussian distributions there is about 32%
that this is the case.

Please try to understand that it is not a case of
"stretching the error bars" if someone indicates that
errors can be larger than the error bars. On the
contrary, that is exactly what error bars mean. For
errors with a Gaussian distribution they tell you:
1) There's 68% chance the error is in this range
2) There's 32% chance it is outside this range
3) There's 16% chance the error is outside this
range on the *lower* side.

In particular, this last point seems to be something
you do not like!

...
Note that the possibility that the observations are correct doesn't
even get mentioned!

If errors are present with a continuous probability
distribution, then the chance would be zero that the
observation is "correct" (in the sense that the
measured value is exactly the true value!) Sorry for
being pedantic, but you are asking for it. :-)

...
OF COURSE those observations are wrong, if not, the whole big bang
theory is wrong.

How do you mean? It sounds more as if the precise
position in time of the big bang would be a few percent
wrong. How could you derive that the "whole" of the
theory is wrong?

.....
Getting back to the cosmology, the 2-sigma limit on the age
corresponds to z=6.8, 800 Myr after the Big Bang. My guess would be
that the star formed around then or a bit earlier. We see galaxies
at least as old as z=8, and there are claims for z10, so this epoch
isn't a problem.


If you stretch measurements and error bars that is correct, yes.

This was the remark I meant. You do not need to
stretch error bars. The given error bars tell you
that there is 5% chance that the true value is twice
or more times an error bar away from the measured
value. That's more probable than throwing 2 times
6 with a dice.

But it is definitely possible that the error bars
are not entirely correct. Suppose there is a large
uncertainty in the size of the error bars, then you
will get a much bigger chance that the measured
value is off by twice the (now given) error bar.
If the given bar is too large, the 5% will drop
to almost nothing, but if it is too small the 5%
will increase quickly, with a net result that you
get *more* chance overall (that the measured value
is off by at least twice the given error bar).

No stretching is needed at all! Just accepting what
the measurement results really mean in terms of
probabilities.

--
Jos

On Fri, 15 Mar 13, jacob navia wrote:
Le 13/03/13 08:30, Martin Hardcastle a ecrit :
(I spent a number of years in order no longer to have to be called
'Mr' in an academic context (-: )

Sorry Martin :-)

That's "Dr. Hardcastle", to you...

..... joking! :-))

[Mod. note: first names are fine on this forum, I suspect! -- mjh]

Le 15/03/13 22:15, jacob navia a écrit :
Suppose a big bang produced density gradient, and at its center some big
mass, atracting things in a radius of 5000 thousand light years. Our
galaxy has a radius of somewhere 50 000 light years, so a "proto"
galaxy (whatever that may be) should be a tenth of that.

A kilogram of hydrogen at 5000 light years has a fall time of

pi R ^(1.5)
--- x -------------
2 sqrt(2G(M+m))

where
R = 5000 light years = 9.4605284 x 10E15 x 5000 meters
M = 1E6 solar masses = 1.9891 x 10E30 kilograms x 10E6
m = 1 kg, let's forget that:-)
G = 6.67398 x 10E-11

I will print intermediate results to verify I did not make any mistake.
a = 9.46052*1000000000000000*5000
47302600000000000000
b = pow(a,1.5)
325332551124632433390390678348.06

M = M=1.9891E30
1989100000000000000000000000000



*** MISTAKE *** *** MISTAKE *** MISTAKE *** MISTAKE *** MISTAKE

1989100000000000000000000000000 Kg is the mass of 1 SUN

I was speaking of 1 MILLION suns!!!

This is a HORRIBLE mistake but the error is within a square root, so the
actual error is of a factor of 1000 since 1000 x 1000 = 1 million.


2*G*M = 265504272360000000000
c = sqrt(2*G*M)
16294301837.145
b/c = 19966031952531170380.610

Now we multiply by pi/2
31362569651705500132.87 seconds
993819861196.84 years

993.81 billion years


WRONG!

It is 993.81 MIILION years


It would take our kg of hydrogen approx 1000 GIGA years to arrive to the
center...


No, only 1 giga year.

OK, what happens if we do not have 1E6 solar masses but 1E9?

"b" above stays the same since it depends on the radius

2*G*M get's multiplied by 1000, the square root is now
515271066876.45, b/c is 631381368056973676.71
that multiplied by PI/2 is
991771533750630924.97 seconds
31427343452.94 years
31.427 Giga years.


NO. Only 31 million years.

OK?


No, not "OK" at all.

1) This looks now much more plausible for BB theory. A mass of 1E9 suns
would atract every kg of hydrogen at 5 000 LY in just 31 million years.

2) A mass of 1 million suns would need 1 GY.

So, it *is* possible according to this revised calculations to gather
matter quickly within the first hundred million years to form a proto
galaxy.

I apologize for this STUPID mistake again.


This will teach me not to do such kind of calculations past midnight.
This morning I spotted the error at first glance.

jacob

P.S. and I was speaking about error bars :-(

Op woensdag 13 maart 2013 08:30:26 UTC+1 schreef Martin Hardcastle het volgende:
In article ,

Yes, in standard cosmology, they can. You can find this in pretty much
every paper about the early universe. Can you present a calculation to
show that they can't? Just saying 'it looks unlikely' whenever you see
something you don't like isn't science, I'm afraid.

For example, a little googling turns up this review, relevant to this
whole thread: http://arxiv.org/abs/astro-ph/0409737 . Have a look at the
calculations of the formation redshift of protogalaxies in there. Do
you spot any errors?

The article you mention is a joy in reading, because I think it explains
in simple language the issues that are involved.
However it also raises certain question.
1. The article starts with the sentence:
"Astronomers have found themselves in the situation of knowing more about
the state of the Universe when it was only 380000 years old then
when it was 200 million years old"
2. Later on he writes: "The evolution of the dark matter component subsequent
to the epoch of last scattering etc."
3. And: "When it comes to understanding the behaviour of the baryonic
component we are on a much shakier ground"
4. At page 5 we read: "Given a mass function of this type, is there any way to
specify when the first halo of a given mass forms"
5. At little further: "Unlike the dark matter the baryons do not initially
form structures on very small scales, since pressure forces act to suppress
the growth of small-scale perturbations etc."

The evolution of the Universe can be divided into two parts:
the period before 380000 years after the Big Bang
and the period after 380000 years.
The most important components of the first period are
darkmatter, nonbaryonic matter and the CMBR.
The most important component of the second period is
baryonic matter and the evolution of stars and galaxies.
The picture emerges that we know the first period better
more accurate, than in the second period.
IMO this is tricky.
The problem is the dividing line of 380000 years.
It is easy possible that the ground work of star building
already started during the first period and that the time scale
of star building was much shorter compared with the present.
This same strict dividing line is also assumed he
http://background.uchicago.edu/~whu/.../angular4.html
"After recombination, the photons stream unimpeded"
IMO changes in physical processes at universal scale
are continuous in nature.
As such it is difficult to accept that first photons
are a local concept and all of a sudden become global.
The local concept means that they almost don't move.
The global concept means that they can move a distance
of 13.7 billion ly in a straight line.
This all of a sudden change is difficult to accept
partly also because photons interact with mass.
It is easy to accept that just after the Big Bang this
interaction was more severe than at present but
but this change should have happened more slowly
more continuous.

Nicolaas Vroom
http://users.pandora.be/nicvroom/

In article ,
Eric Flesch writes:
Uh oh, "dark feedback"... ;-)

Heh. To be fair, though, this is baryon physics, which we know for
sure we don't understand very well. The raw energy required to expel
gas and stop star formation seems to be available, but how (or when
or even whether) it couples to the gas is far from clear. Lots of
research still to be done.

Re arithmetic mistakes: we've all made them. We hope they get
corrected before making it into the refereed literature, but there's
a reason journals publish errata.

Re early star formation: the earliest galaxies are quite a bit
smaller than current ones, and smaller radii are more appropriate.
That makes the relevant free-fall time even less.

And finally: how about the Planck results? Papers are at
http://www.sciops.esa.int/index.php?...ished_Pap ers
but I confess I haven't looked at them yet.

--
Help keep our newsgroup healthy; please don't feed the trolls.
Steve Willner Phone 617-495-7123
Cambridge, MA 02138 USA

On Mar 23, 3:10 am, Steve Willner wrote:
[...]
And finally: how about the Planck results? Papers are athttp://www.sciops.esa.int/index.php?project=PLANCK&page=Planck_Publis...
but I confess I haven't looked at them yet.

http://arxiv.org/abs/1303.5076

As always data rules the day. The sterile neutrino hypothesis has
officially died due to an effective neutrino count of 3.