[BLANK_AUDIO]. So, let's take a look at how spectroscopy actually works. First thing we've got to do is to split the light up into its component parts. So the traditional way to do this is with a glass prism, like this one here. Now any transparent substance, like glass or water, bends light - that's refraction. So, we have all seen this lots of times - when we look at a drinking straw we put in a glass of lemonade, for instance. Now the thing is though that that bending of the light is different for different wavelengths. So as light goes through a prism it comes out the other side, the different wavelengths have bent by different amounts, and so the light is spread out into a rainbow. Now we can see that working here, especially if we turn the lights off - let's take a look. I've got a light source here, coming through this optical fibre. Bring it up to the prism at an angle here and then we will see the rainbow appear on my hand here. And so now, the light is hitting the prism at an angle, getting bent, and it's spread out into a rainbow, which you can see now on my hand - going from infrared through red to blue and ultraviolet which of course you can't see but it's all there. However, modern spectrographs always split the light up using a different technique, using a diffraction grating. I've got one here. Now, a diffraction grating has lots and lots of very fine grooves scratched into it. There are thousands per millimetre - I believe this one actually has 1200 per millimetre - so lots and lots of very fine grooves. So how does that help us make a spectrum? Now, let's imagine - this gets a little trickier than the prism so bear with me - imagine first of all the light of a single wavelength coming in towards the grating and hitting it. Then it gets reflected from each of those grooves independently - so you can imagine lots of light rays going off at this angle here... then over here somewhere, the different rays recombine and then because they are waves, they can interfere with each other. So if two peaks line up with each other, you get twice the amplitude, but if a peak lines up on top of a trough you get zero. Now that changes as you look at different angles. The result is that the light, having come in this way, as it bounces, there is an angle over here, a kind of a sweet spot, where you get a bright spot, okay. So the effect for light of a single wavelength then is the light is concentrated into a particular angle, you get a brighter spot there. Now you can imagine having light made up of a mixture of different wavelengths. That's coming in here - and you can probably guess where I'm going - that bright spot is at a different angle for different wavelengths. So the effect, when a mix of light comes in, is that you get a bright spot at each wavelength in different places, and just like with the prism, the light is then spread out into different wavelengths. If we look at the surface of a diffraction grating you can see the rainbow effect with, with your eye as you tilt it around. So that's how we split the light up into its component parts - but now we have to put this together with some optics on the back of a telescope to actually record the spectrum. So let's see how we do that. So what I have here is a lab bench, lashup - but you have to imagine this being on the back of a telescope. So the sky's over here, light is coming in from the universe this way, And then the mirror of the telescope is bringing the light to a focus here. Now at this point, with the light coming to a focus here - remember the telescope is over here - if we were taking a picture, if we were making an image, we'd put our CCD detector here. And then different stars and galaxies would produce spots all over the focal plane, and you get a picture where your CCD detector is. Now to get the spectrum, what we need to do is to take one of those stars or galaxies, isolate the light from that, and put it through our spectrograph system, which we have here. And you do that with some kind of metal plate with a hole or a slit in it. Now, I'm going to mimic that, by taking this light source in this box here and putting it through an optical fibre which is going precisely through the centre here. Now remember, at this point the light has been coming to a focus to make an image here if we'd just been taking a picture. And now the other side of this, it's now diverging again - so, we then need a lens to straighten out the light so it's going parallel, and that's known as a collimator. And then here is our grating. So now the parallel light comes down here, hits the grating, and then light bounces off this way. Light of different wavelengths bounces off in different directions. Then we need another lens to image the light again. And down over here, we place our CCD detector, and then we get an image of the spectrum. And so, now if we turn the lights off, we'll see the spectrum appear on that white sheet. And if we rotate the grating you can see the spectrum move in angle. So here is the spectrum here - in fact we, we get multiple copies of the spectrum - a complication that need not bother us here. But there is the spectrum and there is a second copy of the spectrum. So here in this little red box, we have a complete spectrograph. This fibre is feeding the light in here. Then inside the box we have a collimator lens and then a grating and a camera lens, and then a CCD detector which is recording the spectrum, and then the data from the CCD comes out here through this USB link. And it's fed around here, in to this laptop, where a piece of software will display the spectrum for us. So in this case, the lamp I'm going to use is a Helium lamp, so this will be glowing specifically in lines of helium. You can see where those are. So now if I take the optical fibre that's feeding into the spectrograph, I can hold this up to the lamp and you can see the helium lines appearing on the spectrum at very precise positions. And seeing those lines we know it's helium. So a spectrograph on a big telescope is really pretty much the same as our little box here except much bigger. So in this picture I'm showing you here, this is the ISIS spectrograph on the William Herschel Telescope on the island of La Palma. So it's fairly much the same except quite big. And in this little video you are seeing my graduate student changing the grating. And you can see that even the grating is quite big on this ... now the problem with a big spectrograph, is that even more than with imagining devices, as you tip the telescope around the sky, the whole setup is going to flex and that makes it very hard to keep all the components aligned and record the spectrum accurately. It's a very difficult engineering problem. One solution to that is to divert the light so it goes through the axis of the telescope, to spectrographs that stay in exactly the same place. And you can see that here in this image - it's showing the so-called Nasmyth platforms of the William Herschel Telescope. So here are a couple of problems with getting science out of spectrographs. The first one is that as we've seen repeatedly in this course, a lot of the science we want to do involves not regular visible light, but infrared light. And the problem with infrared, both in imaging and in spectroscopy is that everything in this room is glowing bright in the infrared, and that can swamp the astronomical signal we're trying to see, so you want to cool everything down as much as you can. Now if you're trying to cool down a little box like this that's easy enough, but if you've got a great big spectrograph, it's very hard, it's a serious engineering challenge. The other problem is fundamental to the way spectroscopy works because we spread the light out, we've made it fainter. If we have an image each star is a concentrated little spot of light - but we are taking that light from that spot and spreading it into a streak and then it is inevitably fainter, harder to detect. So, for example maybe a faint galaxy that might take 15 minutes of exposure, say, to get an image - that same object if we're trying to get a spectrum might take several hours. So, one practical consequence of that is that very often medium-sized telescopes concentrate on imaging, and the very biggest telescopes spend much more of their time doing spectroscopy because they can do it faster. But it also means, for instance, if we think about some of the very faintest objects we've ever seen with the Hubble Space Telescope, they're so faint, we simply can not get a spectrum at all. So there's one more important difference between imaging and spectroscopy. Traditional spectroscopy is very slow, because we're measuring one object at a time. This is very different from imaging. Let's illustrate that, here. Suppose we've made an image. On that image, we might have hundreds and thousands of individual stars and galaxies. So we measure a lot of objects in one go, with an image. Now, what I've described to you about having to do spectroscopy is that we might take one of those objects there, isolate the light from that object, put it through our spectrograph and then, on our detector, we'd get a streak corresponding to the spectrum of that one object. The rest of the CCD is wasted, we just had that one object. And then to measure all the rest, we have to move the telescope, do this one, this one, this one, this one, etcetera. It's a real pain, obviously. So a lot of modern astronomical spectroscopy is about working out how to do lots of objects in one go. Now the key to that you've already seen... You may remember when we looked at our lab lashup, we used an optical fibre like here. Now the cool thing about an optical fibre is that you can bend it so you can send the light from one place to another very easily. So, if we have a fibre on this object here, and it's in to the spectrograph, we can have another fibre, put it on this one, another fibre, put it on this one, another fibre, put it on this one. So then, the effect of that is that this one can make a spectrum here, this one makes another spectrum here, this one makes a spectrum here and so on. The net result is that we can get hundreds of objects and get their spectra in a single shot. It's an enormous saving. So another thing you can do... is suppose this object here is an extended object like a nice spiral galaxy... then what we can do is to take all our fibres, and place them in a close packed array, in a sort of grid, over the surface of that galaxy. And then each one of these fibres goes through the spectrograph and into our detector, and we get a complete set of spectra, one for each position on that galaxy. And so, for example, we can see what the velocity is separately for each spot on the galaxy - and that way, we can see how the galaxy is rotating. Very useful and interesting. So this technique of using a close packed array of fibres, is known as an integral field unit. So, here in this image, you are seeing a real fibre spectrograph in action. This is on the Anglo-Australian telescope. This was used to take spectra of hundreds of thousands of galaxies over a large area of sky, in order to do a complete survey of the local universe.
