How to know there is a frequency shift?

[ShayanJ]

This is a question that I've had for years but because astrophysics and cosmology are down in the list of my interests, I constantly forget and remember the question.
My question is, how do we know there is shift(red or blue) in the spectrum of the light coming from a star? Because for understanding this, we should somehow have access to the spectrum of the light immediately emitted by the star, which I don't know how that is possible.
Thanks

 

[jedishrfu]

I would suspect it was an insight when noticing the hydrogen lines of various stars being shifted left or right and from a related observation that they are moving closer or further away.

From the wiki article there's more history to it with wave study in general as the basis:

https://en.wikipedia.org/wiki/Redshift

 

[e.bar.goum]

This is a question that I've had for years but because astrophysics and cosmology are down in the list of my interests, I constantly forget and remember the question.
My question is, how do we know there is shift(red or blue) in the spectrum of the light coming from a star? Because for understanding this, we should somehow have access to the spectrum of the light immediately emitted by the star, which I don't know how that is possible.
Thanks

This is where you need to know that the light from a star has many many absorption lines coming from the elements in the cooler stellar surface.

Here's an example showing the hydrogen lines, which will invariably be the strongest lines, but you can do the same thing to find how much iron (for instance) there is.

Then, since atomic physics is the same wherever you are, you know from laboratory experiments where you expect to see the absorption lines from whatever you are measuring. The difference between what you expect to see and what you measure is the red/blue shift.

 

[ShayanJ]

Yes, I know about the spectroscopy and this way of using it to find the redshift. But this will put us in a loop.
Because we should have a way, besides spectroscopy, to know the star's chemical composition. So that we can compare the spectrum of its light with its spectrum if we could measure it with a negligible amount of shift.
My problem is, I see no other means than spectroscopy, to find out a star's chemical composition.

 

[russ_watters]

Yes, I know about the spectroscopy and this way of using it to find the redshift. But this will put us in a loop.
Because we should have a way, besides spectroscopy, to know the star's chemical composition. So that we can compare the spectrum of its light with its spectrum if we could measure it with a negligible amount of shift.

I'm not understanding why you think this is a problem. Why should we have another way to find out a star's chemical composition if spectroscopy works fine? Why does that put us in a loop?

My problem is, I see no other means than spectroscopy, to find out a star's chemical composition.

Well, there is also a highly developed theory that describes how stars work.

 

[e.bar.goum]

Yes, I know about the spectroscopy and this way of using it to find the redshift. But this will put us in a loop.
Because we should have a way, besides spectroscopy, to know the star's chemical composition. So that we can compare the spectrum of its light with its spectrum if we could measure it with a negligible amount of shift.
My problem is, I see no other means than spectroscopy, to find out a star's chemical composition.

I don't understand how this puts us in a loop at all. Of course spectroscopy is the way you know the stars chemical composition. But you know from experimental studies in the lab on Earth very well what the absorption spectrum of hydrogen (and everything else) should be. Since we say that atomic physics works everywhere you are in the universe, if you know the expected spectrum, you can pick out very well what the chemical composition is, and thus the redshift.

ETA: Anyway, you know that the outer envelope of a star is going to be rich in hydrogen, almost no matter what else is going on. There's also going to be a decent amount of helium. That's a good start.

 

[ShayanJ]

Well, there is also a highly developed theory that describes how stars work.

Yes, but that should be tested experimentally first. So I think at least once in history, people should have used only observations.

Is it that the frequency shift can have no effect on the position of dark lines in the spectrum? It seems to me there should be a change. So if the position of dark lines change, how do we know this is for e.g. hydrogen because the frequency have changed!
I mean, we get the light, we observe a shifted spectrum. But because the spectrum is shifted, its necessarily different from the real spectrum of the star. So it won't show the real composition of the star. How can we use this both for finding out the composition and the amount of redshift?

 

[russ_watters]

Yes, but that should be tested experimentally first. So I think at least once in history, people should have used only observations.

I don't understand: astronomy/astrophysics is most certainly based on observation/experimentation. Really, all science is, by definition of the word "science"!

Is it that the frequency shift can have no effect on the position of dark lines in the spectrum? It seems to me there should be a change. So if the position of dark lines change, how do we know this is for e.g. hydrogen because the frequency have changed!
I mean, we get the light, we observe a shifted spectrum. But because the spectrum is shifted, its necessarily different from the real spectrum of the star. So it won't show the real composition of the star. How can we use this both for finding out the composition and the amount of redshift?

A spectrum isn't just a single line or random lines, they are in a pattern and the pattern is always the same for a given element. So if you see the same pattern (spacing, intensity), but the positions are shifted from what you expect, you can still recognize that that pattern corresponds to the same element. It's really not that much different from how you can still read text even if I change the font color/style or a color-blind person can still follow a traffic light.

And again, since the vast majority of a star's compisition is hydrogen, it is pretty easy/safe to make your first pass at identifying the star's spectrum from the expected position of the hydrogen lines.

 

[e.bar.goum]

Yes, but that should be tested experimentally first. So I think at least once in history, people should have used only observations.

Is it that the frequency shift can have no effect on the position of dark lines in the spectrum? It seems to me there should be a change. So if the position of dark lines change, how do we know this is for e.g. hydrogen because the frequency have changed!
I mean, we get the light, we observe a shifted spectrum. But because the spectrum is shifted, its necessarily different from the real spectrum of the star. So it won't show the real composition of the star. How can we use this both for finding out the composition and the amount of redshift?

No, of course the frequency shift changes the position of the dark lines in the spectrum. But, we not only know the expected frequencies of hydrogen etc, we also know their relative spacings. Absorption spectra are like fingerprints. Broadly speaking, you take your spectrograph, and your expected lines, and you shift the entire spectrum until the two line up. You can also have a pretty good idea of the relative intensities and shapes of the lines, but that's a bit more model dependent, and it's not really needed either. Let's construct a toy model:

Say I have three hydrogen lines. In my lab, I measure their positions to be at 4000 A, 5000 A, and 6500 A. I take a star, and measure the entire spectrum, and see four lines - at 4050 A, 5050 A, 6000 A and 6550 A. Immediately, it's obvious that the spacing of three of the lines are the same as hydrogen. - the 4050, the 5050 and the 6550. That must mean that I have a star with a frequency shift of 50A.

Schematically:

In real life, things are more complicated. Here's Arcturus.

http://prancer.physics.louisville.edu/classes/107/topics/stellar_spectra_examples/arcturus_high_resolution_spectrum_sm.jpg

 

[jedishrfu]

The spectral line pattern is known for elements on the Earth and was explained by Bohr model:

https://en.wikipedia.org/wiki/Bohr_model

so as astronomers looked at stars farther and farther away they began to see this shift and concluded it was due to redshift and the calculations for doppler effects coincided with what was measured.

 

[ShayanJ]

My question has been answered. Thanks all of you.

I don't understand: astronomy/astrophysics is most certainly based on observation/experimentation. Really, all science is, by definition of the word "science"!

This is only a misunderstanding of my comment. But I understand what you mean. So I guess its better to forget about it.
But to give it a last try, I think I can say I meant phenomenological astrophysics.