what creates the "star points" in telescope pictures of stars?

rbj

it's a basic question. e.g. see this hi-rez Hubble Deep Space pic:

http://upload.wikimedia.org/wikipedi...full_jpgNR.jpg

i count about three bright objects with pointy cross-like projections from the bright object. i take it that these three objects are stars in the foreground (the Milky Way) and all of the other objects are, i assume distant galaxies well beyond distances in the Milky Way.

so what causes those pointy projections? why four points? why are they all oriented in the same direction? (i presume it has something to do with the orientation of the lens and if the Hubble were rotated a little, the points would move relative to the background.)

since the stars and the galaxies are all, pretty much, at a focus of infinity, why are they different? why shouldn't a bright, dot-like galaxy, have similar pointy projections?

Nabeshin

http://en.wikipedia.org/wiki/Diffraction_spike

turbo

Here is a very simple explanation. You can search on "diffraction spikes" for further explanations.

http://en.wikipedia.org/wiki/Diffraction_spike

Edit: Nabeshin beat me to the Wiki.

rbj

so, in HST Deep Space, is it solely because those three objects (now i count four objects with spikes) are so much brighter that they have diffraction spikes? they don't appear to be any orders of magnitude brighter than the galaxies behind them, yet their spikes are pronounced. and none of the other objects have those spikes at all that i can see.

turbo

Stars are essentially point-sources of light, while the background galaxies are extended objects. You can Google "diffraction spikes" as I suggested to find out why this is true.

rbj

another, sorta engineering question, is since the distant (and weak) light must pass through *some* glass or transparent medium in the reflecting mirrors, why not use a plate of very clean glass to hold the secondary mirror, rather than opaque struts that cause these diffraction spikes?

also, why not locate the secondary mirror (or even the eyepiece or camera) off to the side of the main telescope cylinder? it would mean grinding the main reflecting mirror a little differently, but it seems to me that grinding and polishing that mirror is a big deal requiring ultra high precision and computer control of the shape anyway. it's just that a different less-symmetrical shape would be the target to grind and polish toward. or maybe, just like the HST flaws were corrected with intervening lens, maybe the main reflecting mirror can be ground to a perfect parabolic shape as it is (or should be) now, but, because the secondary mirror is off to the side, it would be shaped slightly differently than a flat plate to correct for that asymmetry.

another solution is to turn or rotate the telescope 45 degrees and retake all of the images (maybe even more than one retake) and, with a really good computer program, merge the two (or more) digital photos and where they differ greatly (because of the spikes), use the non-spiked pixels. can't be much worse than the processing we had to do to get the Voyager images of the outer planets so nice and clean. and even if it takes an extra week or two of HST and as much time outa some supercomputer, i would think that, for all of humanity and the science community, that price is worth it to get these unparalleled images with no artifacts (other than pixelization when you zoom way in) due to the human technology. aren't we the least bit curious what might live behind those diffraction spikes?

rbj

that difference seems not so clear with the HST. there are certainly stars (that might be closer than those in the Deep Space pic) that HST sees as tiny-little balls, no?

turbo

HST cannot "see" stars as balls. They are all point-sources of light. If the stars are sufficiently bright, they will create diffraction effects, like halos and spikes.

I have a 6" APO refractor, and I have used home-made aperture masks to create such spikes in order to be able to rotate the mask and view a fainter companion in a double-star system. Diffraction effects are not always a bad thing in astronomy.

twofish-quant

Because every photon is precious, and you can just ignore the spikes, whereas even the tiny bit of light that gets blocked by glass means that you can't see dim objects.

The other thing with the space telescope is that it does a lot of UV observations so that if you use glass, it has to be expensive special glass, because ordinary glass is opaque to UV.

Been tried.

http://en.wikipedia.org/wiki/Reflect...pe#Herschelian

The trouble is that it turns out to be a less stable design. You end up with less support for the main mirror. Also the way that telescope grinding works is that that you start off with a flat piece of glass, and you can grind the mirror in a rotating lathe. If the mirror is off axis, then you can no longer use circular motion to grind the mirror.

Or you can just use computer processing to get rid of the spikes, if they are annoying.

One thing to point out is that astronomers hardly ever look at "images". You can think of a telescope as a light bucket that concentrates light that goes into various instruments that measure color or brightness but really don't care about what the image looks like.

It's trivial to program a computer to mathematically remove the spikes, but what you end up with is

a point of light that's a really, really small fraction of a pixel.

The thing about stars is that they are so small that you can't see anything all all. Everything about the image of a star is an artifact that is generated by the telescope, and *none* of that is "real."

What you are fighting is the fact that light is a wave and waves "spread out". The amount of "spread out" by the light is far, far more than the size of stars, so when looking at stars, all you are seeing is light getting spread out by the telescope.

rbj

really? even when viewing exoplanets around reasonably close stars?

http://en.wikipedia.org/wiki/File:44...414-a-full.jpg

there was a star blanked out in that photo. was it a point source or was there some visible photosphere? what about really big stars like Betelgeuse? isn't the photosphere of this red giant as big as a solar system? how is it that we can see exoplanets, but stars as big as the orbits of these exoplanets are seen as point sources?

my question was motivated by seeing similarly bright objects in the same photo, where some had diffaction spikes and others had no sign of any diffraction spike. this is still a curiosity.

twofish-quant

Nope. All stars (even red supergiants) end up as point sources of light.

Now it is possible to get images of the larger stars but this involves a lot of specialized processing that Hubble isn't set up for.

http://en.wikipedia.org/wiki/Astrono...interferometer

twofish-quant

Something else about Hubble. The important thing about Hubble isn't that it can see without atmospheric distortion. There turn out to be a number of ways you can subtract atmospheric distortion mathematically from an image. The really important thing is that Hubble can see in the ultraviolet which you can't see from earth.

The computing power you need to remove difraction is tiny. An iPhone could probably do it without breaking a sweat.

turbo

I can't wait for the JWST! There is so much information in the infrared... Particularly in very distant objects whose light has been heavily redshifted.

russ_watters

I string wires across the front of my telescope if I want to create the effect:

http://www.russsscope.net/images/Horsehead-HaRGB.jpg

hipparchos

This is done in a number of telescope designs, such as the Schmidt-Cassegrain and Maksutov. But the diffraction spikes are useful in some applications, specifically astrometry (measurement of stellar positions) because it confirms that the object is a point source (star) rather than a galaxy, and the spikes aid in determining the star's position.

This is also done in a number of designs, notably the Shiefspiegler. The problems aren't so much the grinding (it can be done by hand using appropriate techniques) but that a telescope built this way has strong aberrations unless the f-ratio is quite large.

There are much easier ways to answer that question: for example, you can rotate the telescope (or the spider) around the optical axis to put the spikes in a different position. Or you can build a circular spider that suppresses the spikes entirely. (A typical spider looks like an "X" when viewed along the axis; a spike-suppression spider looks like an "8" viewed along the axis.)

SHISHKABOB

can't we see Betelgeuse as a disk?

turbo

Not with a space telescope. You need widely-space optical telescopes and interferometry to pull that off.