 There are many experts in the room, so I hope I don't bore you too much, because it will be a very simple and general introduction. I cannot touch too much in this short time. So now let's see. Okay, I think I stay better in this corner. When you do plasma spectroscopy, you have to think about the goals. And you know you have one goal could be you like to determine. No, I'm sorry. You know I have to learn. Oh, okay. I wanted to. Yes. I have to return. No. Okay. Sorry. You want to study radiative properties. Or you do spectroscopy because you want to study. You can do it. You want to study atoms in plasma environment influence on the atomic structure by electric and magnetic fields. And you want to test theoretical calculations and the modeling. We will hear a lot about it this week. Then you have the other area. You want to use plasmas as radiation sources. So you have to characterize your radiation source or you do plasma diagnostics for determining plasma parameters. Now, depending upon your task, you have to select the plasma and you have to select the diagnostic equipment, which usually is not the case. You have to take what is around in the lab. So how what is radiation? How does the radiation from a plasma look like? Well, I took an example here the radiation from hydrogen plasma of about a hundred is 70 EV. And you have the first continuum radiation. Then you have the recombination radiation when an electron recombines with an iron. You get the characteristic structure here with the edges and you have the line radiation. If you go to long wavelengths and your plasma becomes larger and higher densities, it becomes optically thick. And here the Bremstrahlung approaches first the black body limit. You cannot get out more of the plasma than the black body limit. But what can you do if you want optically thick radiation? You have to increase the density and the temperature then this point moves up. But it is difficult to get optically thick radiation, so the maximum out of a plasma by Bremstrahlung. But you can do it with line radiation. You put in more and more atomic species and they grow and finally can reach the black body limit. And you have a black body radiator at one specific wavelength. So the task here determines the choice of your spectroscopic equipment. But you should roughly know the temperature. If you want to do X-ray spectroscopy, you cannot use a 1 EV plasma because there won't be any radiation in that regime. So a rough knowledge of the temperature is helpful. Now what do you measure? When you have a plasma and you do spectroscopy, you measure only the radiation which comes out. And this radiation gives you the information inside the plasma. This radiation is called the spectral radiance. And nothing else what you can get is from plasma spectroscopy. But what you really want is the radiation from inside the plasma. And the radiance, spectral radiance is the integral along the line of sight of the local emission. But this doesn't help you very much. Only this is important if you want to use a source as a radiation source. But if you want to study properties inside the plasma, what can you do? You do observation at various directions from all directions. And you know it from the medical society people that you can do computer tomography. And from the computer tomography you get the local information. But if you know your plasma machines it's just not possible to do observation from many sides. So how does one help? Well there are exceptions. One is if you have cylindrical symmetric plasma because then you integrate along cords. You know the integration gives you here this radiance. And you do the able inversion and you get local emission coefficients. Equivalent to this is the configuration with a pinhole here because you can easily transform this to a cylindrical symmetry. You also can do it in tokamaks when you have a cylindrical plasma where the circles of equal densities are shifted by the chavronov shift. You can take this into account and you even can do it in cases where you have an elliptical cross-section. Another point people do mistakes is when you have small devices and have a thick wall. What happens in a thick wall? Here you know this would be the radiance here without the wall. Now you take a thick wall but luckily if you follow up the cords you see you get the same radiance as if this were singular. Only of course the cords are from another direction but they are identical. So you get correct results. Then comes the problem. Your radiation from the plasma goes through the air and air absorbs. We show to the next picture here you see air absorbs here below 200 nanometers. So no spectroscopy in air below 200 nanometers. Again in the X-ray region here below 0.2 nanometers or below 2, 3 angstroms. You can do X-ray spectroscopy. How can you help yourself? Now you don't have the right equipment which would be vacuum instruments for your spectrographs. You can fill the path from the plasma and the spectrograph. You can fill it with helium and you see with helium here you can go really further down and you have only a smaller region where you can not use vacuum UV equipment and you can also fill it with argon. So here are the examples for standard pressure and room temperature. But you have windows. So you know below the quartz window at 200 nanometers nothing is possible through quartz. But you could use sapphire and one interesting thing is if you do infrared spectroscopy you can go to about 10,000 here. You have longer wavelengths for 10 millimeter salt or calcium fluoride. Now it's lower. So in the lower region one uses beryllium foils here. They transmit as windows. They transmit here in this region. Here you have beryllium, lexan and captons as foils. You have to consider that but it can help you covering the spectral region. Okay now people use mirrors in spectrographs. And you won't believe it but there are publications in very excellent journals. They used spectrographs in normal incidence below 100 at below 10 nanometers which is complete nonsense because no mirror reflects. There have been publications on that. So here you see the reflectivity of mirror and you see going below 100 nanometers you can use gold and osmium and platinum are also possible here. But you have to consider that when you use a spectrograph what are the mirrors do? Or if you do an optics with mirrors. But you can help yourself a little bit by going to grazing instruments because the reflectivity goes up when you have radiation hitting at a small angle or large angle of incidence or small incidence angle hitting the mirror. And though the grazing incidence equipment can be used much lower. Well I come to this equipment in a moment. So you have to select your instrument according to your task and what you want to do. And now some design considerations. If you want to do survey spectra a low resolution instrument is needed. If you want to total line intensities then medium spectral resolution is recommended. And if you want to study line profiles you need high resolution instruments. And another thing is in many plasmas you don't have enough light. So you want to collect as much light as possible. And so you need instruments with a high throughput. And throughput was called also étendue. And this means high large entrance lid times solid angle which is recepted by the internal optics. Now you have to look at the spectral region. When you have a plasma and think we had inject atom then the atoms are being ionized. And if the temperature is high or temperature is given the ions go successively through the ionization stages here. And you see we start one ion stage after the other till they come to an equilibrium. And the equilibrium is reached when recombination is equal to ionization. And then you have a steady state situation. And in low at low density this equilibrium is reached in the corona equilibrium ionization is balanced by radiative recombination. Or at high densities you have a collisional equilibrium you have the sigh equilibrium. So when you reach a steady state I show you an example for neon. Here you see the steady state situation for neon is given here as a function of temperature. So for example if we have 100 EV neon plasma you see the ions are mostly in the helium like stage neon 9. And they stay a long time in this ionization stage till they are then fully ionized at very high temperature of 1000. But I will come to this later again when you know the ionization lines from an ionization stage in a plasma you can roughly estimate the temperature. So once you have no roughly the temperature and you have selected your spectral region you can go ahead and have another consideration. You have to do do you want a stigmatic system or an asthmatic system. Now I try to explain this here. You have an entrance slit you have a dispersing element in the spectrograph. And you know this point is imaged here another point here would be imaged up here. And you see the entrance slit gives an image in the exit plane. And in this direction you have a wavelength scale. So you can get the spectrum from an image in the plasma if you image your plasma on this slit. Now once you have selected set what you want we go to the types of spectrograph which is very standard and you have found an N in every book. You know the simplest one was this prism spectrograph and here now if I look such a picture you know and you would draw it I would immediately see if the guy who has done the drawing knows optics because in some books and even text books I have seen that these lens are the other way around. The curvature is in this side and here also on that side and this is very bad from an optics point of view. So you immediately can see if you build such an instrument if you thought about the optics of a lens. And it's wrong in many general text books. Now here you can use half of a prism but they are not used very much anymore but you can build it yourself very cheaply. Then you have the grating and you know everybody knows a reflection grating. I think I continue here and then you design your instrument and the most commonly used instrument is of because it's called the churny turn amount. And here it has good optical qualities with two mirrors and a plane grating. But now we are going to use this at lower wavelengths. We have done it by filling in helium but something happens. This has three reflecting surfaces and the reflection goes down of all three surfaces. So you can imagine if it's 50% reflectivity to the third power how far the reflective the throughput has gone down. So then one uses this type of grating which has a concave grating and the grating is mounted on the Rowland circle. Now you have only one reflective surface. So this helps quite a bit and you should consider it when you have a chance to select what you need. Now what I said is the reflectivity goes down too much. As a rule of thumb I always thought 330 nanometers below that you never can use this kind of instruments, normal instruments. Then you should go to the grating incident instrument. And the grating incident instrument here has also a curved grating sitting on the Rowland circle at the very narrow angle of incidence. A very large angle of incidence, a small angle of grating incidence. And it's focused on the Rowland circle. So you need for this kind of instruments, you need a detector with a flat with a curved detection plane. A very simple design and a cheap design is a so-called off Rowland design. You just put a plane detector here. You see it's focused at one point and it is not focused a little off the good focusing point. But you can live with it if you do the mount ride and you cannot do for all constructions. But in certain cases you can do it and these instruments are very cheap and they are on the market too. So with a grating incident instrument people have gone down to 0.54 nanometers for a grating incidence angle of two extra. If you go further down you of course limit your throughput, your angle of your acceptance angle of the instrument. And the problem is a grating incidence instrument requires great care in alignment. So but development has not stopped and the development was with the detectors. And at first with the grating one does not use a grating with equal spacing. One uses a grating with changing spacing and they have the characteristic that the plane of focusing is a plane. And so it's easier to align. It's most commonly used now but at disadvantage the grating incidence instruments are asthmatic. So if you have a system where you want to have a asthmatic image people have done this also. They have used various mirrors, toroidal mirrors to reimage two or three times the image to get a asthmatic image. Now what also have been used are toroidal gratings. They give two dimensional focusing and minimized the aberration. But you just have to look at the literature what is available. Now to very short wavelengths you go to crystals. And crystals I just mentioned you know you have reflection by the rag angle. But when you design your instruments you have to be aware of the following. If you have a point source you know each wavelength is reflected at a different angle and from a different point on the crystal. So this is different from mirror instruments where one wavelength is by focusing from the whole area. Here one wavelength in one direction is from focusing here on one point. So what you have from a point source are cones coming like that and you know this in all directions. And what one cuts out here is a small piece so reflection comes from this area and here are the lines. And depending how you set up your instruments these lines are parts of hyperbola and ellipse or parabola. When you cut a cone with a plane you know from. Now the focusing properties are then similar to that of a grating. I show you one typical mount which is used in large devices. Here you see this point on the Rowland circle is focused on to here. And this point this line is focused here from different points different wavelengths. So the radiation collected is from a large plasma volume. It's not from a point source. So this you have to be aware if you do use this alignment in for large focus large plasma devices like Tokomax. And it has advantages of course because you know if Tokomax you have rotation both lines have different Doppler shifts. So you can measure the Doppler shift from the line just to mention it. Okay you can also use another property of crystals. Crystals can polarize. So you can study the polarization if you use a crystal the reflection it's a blue stain which has been done also in the literature. So it's possible with crystals this kind of measurement. Another commonly used spectrograph is the Fanhamos mount. And you see here it's a cylindrical crystals and it's focused along the axis of the crystal mount. So you can have a plane detector here and they are also now to a cylindrically curved crystals. So excuse me spherical curved crystals. So you can also do imaging. So there's more in the literature than I can mention here at the moment. So the next point I want to mention are the detectors and there are many possibilities and they also depend upon what you want to do. And what you want to observe. And what a detector does it converts a flux into a signal. And though the spectral sensitivity is usually what characterizes the detector in respect to your applications. Now what it does you know you have to also think if you want to integrate or you want a time resource spectrum. So one has energy then you have besides a steady state you have slow varying plasmas and you have fast varying plasmas where you want detection in the picosecond range. So this gives you a device and so the general formula is the time response if I call it t and the frequency bandwidth of the response are given by the simple relation down at the bottom. Now another point is you have to look at the dark signal due to dark currents because this can destroy when you have very weak radiation. And then you have usually you help yourself by cooling the detector to down then the noise which is usually thermal noise, short noise can be reduced. Another point is seeing and these are mystics each all detect or most detectors have an internal time delay. And people correlate signals and forget the time delay and this is very important because and there are wrong things in the literature too that there were conclusions but it was just the time delay by different detectors. So you have to know the internal time delay. Time delay of cables is trivial but also there people have bad mystics by using different cables and neglecting the time in a cable for short time detections. Then you have to look at the linearity. It's called the dynamic range of the detectors and you have to look at the long time stability. Some very sensitive detectors age rather quickly and you know if you have things in the lab for some time it's not used. Suddenly it doesn't work properly anymore. Then you have area detectors like the charged CCDs and with CCDs they give you at the exit side you know from a point you get a two-dimensional image and here you have to look at the pixel size because it determines the spectral resolution. And from a point on the entrance of the detector you get on the output not a point but the point spread function which you have to identify for your analysis. Then important are gate times. If you want to look at specific times in the development or if you want to take frames, the number of frames and the repetition of frames is an important point to consider. Another device which is not so used, not so often used anymore but was commonly used are their photomultipliers. And here you have your photons are converted to electrons. They are multiplied at the dinodes and at the anode you get a signal. What you have to watch with photomultipliers. If the incident flux is too large you get too many electrons and you get space charge effects here. And some of these effects show up that suddenly the signal gets less and then it gets much larger. So are funny effects. And also the current, the anode current should not be too large and this gives also various problems. And one tries to help itself. You know you have in the photomultiplier to the dinode change you supply the voltage by a resistor chain. And if the amplification is too large the dinode cannot deliver the electrons anymore. So the amplification goes down. One can help this for a short time by putting in parallel to the resistor chain capacitors. So for short pulses these capacitors supply the electrons and keep the voltage constant. And the other case the voltage would drop and the amplification goes down and you get wrong results. So if you use a photomultiplier please look at the specifications not only of the photomultiplier but also of the internal construction, the dinode chain with the capacitors. Okay and then another point asked if you go to the vacuum UV and the X-ray region you can use scintillators in front of the photomultiplier. And the scintillators of course have also time constants. You have to look at you know they are nice scintillators with high efficiency. They have relatively long decay times. And one can do the following one add to the scintillators other elements which cut down the tracer elements which cut down the time constant but cut down even more the sensitivity. Now other photomultipliers are the channel photomultiplier or channel tron. Here you have a continuous diode and this is all a dinode and this then you have a continuous voltage along here. The dinode internally is also the voltage divider. Now if you put many of these small channel trons together you know you have the micro channel plates here. Many micro channel plates they are accelerated and finally hit the electrons hits a fourth four. Okay you can also gate. Now for the diodes you have a large number available even in the X-ray region you have many two-dimensional CCDs available. You can use ionization chambers. You also people use multi-wire proportional counters for the X-rays of course also for other particles. They are usually used in fusion devices. And here you can have them two-dimensional. You have a photon here hitting the cathode it makes an avalanche and when the avalanche reads the anode you measure the arrival time on both ends and you know at what positions. You do it two-dimensionally you know part of it goes through and you hit also the bottom one and you get also a coding in this direction and you know exactly where the photon hits. A very new development which is now continuously more and more investigated in supply are the gas electron multipliers. So what you do is you have in a foil you have this double cone little holes and electrons you can have a high electric field in this electron multiplier and you can have all the spatial resolution and they have the advantage a relatively cheap have an amplification of 1,000. They are very robust against damage and you have two-dimensional arrangement and you have if you can put two together you get a very large amplification. Okay then comes the main point is the calibration and really don't trust an instrument which you get delivered by a manufacturer. You see it's aligned but if they don't align it and calibrate it in your lab you can forget about it. I always like to tell a story. I made a student and she had done many many measurements and the temperature was an argon discharge. Thousands of elements analyzed and they obtained a temperature which sounds reasonable and she asked me and I tried to analyze the spectrum. I couldn't fit the spectrum. There was no strong lines. So then I was like there must be strong lines and by doing by hand shifting the spectrum around I found she was 80 angstrom off. So she had all lines were off by 80 angstrom. It was a low resolution instrument. So in a kind of argon plasma so you have so many lines that you always believe you are right. So I asked how did you do the calibration? I started from the factory. Well that's it. You have to do your calibration yourself with a line and at first the wavelength calibration. You must be sure you are measuring the right quantities. Okay now you see you have various reasons. You have UV if you want to calibration and wavelength in the visible in the X-ray region you have to use these charges which emits a proper radiation. But wavelength calibration is simple. The sensitivity calibration is more difficult and you have primary standards and you have secondary standards. And the best standard is of course a black body radiator and I know that the astronomers who do visible spectroscopy they also use they build their own black body radiator with an oven and they heat it to a definite temperature and have a small hole the black body radiation emanates. So this is given the whole round radiation is given by Planck's radiation law. Or it's just me maybe. Okay for the X-ray radiation electron storage rings are the radiometric standard source but they are difficult and the disadvantage is that they use the electron storage rings emit only in a small angle in one direction. And you really would like to have a radiator standard which emits into a large angle and hopefully also from a larger size. Now when you don't have the possibility of the X-ray calibration you cannot easily take your big spectrograph and go to an institute like NIST or another standard institute. It's nearly a PhD which is to do a calibration on these devices and how can you help yourself? And one possibility is the branching ratio method. Here you take, you put an iron in one plasma even that's what you want to study. And you take two lines, one is in the long wavelength region and one is in the short wavelength. Excuse me, this line is in the long wavelength region and this line is in the long wavelength region and you compare the radiance and it depends only on the ratio of the transition probability. So you look at, take a pair, one line is in, let's say in the visible where you can do easily the calibration and automatically you have the calibration in the UV, extreme UV. And you can do it with enough possibilities and enough very good transition probabilities available. Oh, then you have to watch, you know, when you do the experiments what can happen? Spectrograph can have varying intensity across the slit height because from the center of the slit and from a long slit it can be a reflection and have a different throughput, different cone inside the instrument. So you must calibrate this. You must calibrate the pixels. You cannot assume if you have a CC detector all the pixels have equal sensitivity. So you calibrate this effect which is called flat feeling. Then fast detectors see the iris effect. When you put a pile to a camera it takes a finite time to pass through the detector to open the detector for gating. So it's, you have to take into account that you may get wrong results. Okay, but you have many secondary standards. Okay, tungsten stripline, blue carbon arc, wall stabilized arcs, holocaust or discharge. In the X-ray region I don't know of any secondary standard. Some people have built themselves an own standard at home maybe. Okay, and most of all the material is from that book. Okay, thank you for your attention.