Université Paris XI - ORSAY Master 1 Nuclear Energy BONAMY Geoffrey
INTERNSHIP REPORT Detectors Calibration
Supervisor: Dr. Nicolas Delerue
Laboratoire de L’Accélérateur Linéaire Bât. 200 Université Paris Sud, 91405 Orsay Cedex France
Contents 1 Introduction 2 Theory and experiment 2.1 Coherent Smith-Purcell Radiation 2.2 CLIO . . . . . . . . . . . . . . . . 2.3 Experiment . . . . . . . . . . . . 2.3.1 Pyrodetector . . . . . . . 2.3.2 Acquisition module . . . .
2
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
3 Characterization of the aquisition module 3.1 Approach . . . . . . . . . . . . . . . . . . . . 3.1.1 Module response . . . . . . . . . . . . 3.1.2 Coding solution . . . . . . . . . . . . . 3.1.3 Uncertainties . . . . . . . . . . . . . . 3.2 Amplitude . . . . . . . . . . . . . . . . . . . . 3.2.1 Link between amplitude and frequency 3.2.2 Characterization of the amplitude . . . 3.3 Characterization of the Offset . . . . . . . . . 3.4 Characterization of the Sampling rate . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
. . . . . . . . .
. . . . .
3 3 4 6 6 7
. . . . . . . . .
9 9 9 10 11 12 12 13 14 15
4 Detector calibration 16 4.1 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 4.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 5 Conclusion
19
1
Chapter 1 Introduction This year at Nuclear Energy Master degree I have chosen to do my Master degree internship at the Laboratoire de l’Accélérateur Linéaire (LAL). The LAL was created in 1956 and it is currently under the direction of Achille Stocchi. It is a part of the IN2P3 (Institut National de Physiques Nucléaire et de Physique des Particules) institute of the CNRS. The laboratory research goes from the particles physics to the astrophysics and cosmology studies and accelerator physics. My 10 weeks internship takes place in the group ETALON (Emittance Transverse And LONgitudinal) leads by Nicolas DELERUE. The ETALON project at LAL is part of the accelerator department. Its aim is to develop advanced diagnostics for particle accelerators and especially to measure the longitudinal profile of electron bunches by using Coherent Smith Purcell-Radiation (CSPR). This Radiation will be explain in more details in the first part of this report. My work in ETALON was to study and find a way to characterize pyrodetectors used in the experiment for the detection of CSPR with the accelerator CLIO at L. This laser will also be presented in the first part of the report. My internship was more experimental and the goal was to find and create a reproductive characterisation of those detectors in order to study the results from the experimentations. First I will present the theoretical background of the Smith-Purcell Radiation and the CLIO laser experiment. In a second order I will explain all the work that I had to handle before the beginning of the calibration. Then on the final part of this report I will present you my ideas and steps for the calibration and also measurements.
2
Chapter 2 Theory and experiment 2.1
Coherent Smith-Purcell Radiation
Figure 2.1: Smith Purcell radiation [1] The Smith-Purcell radiation was discovered in 1953 by two physicists Smith and Purcell [5]. This radiation appears when an electron bunch passes near a grating, see Fig. 2.1. The electron bunches induces surface charges in the grating which will emit light. This phenomena is explain in detail in paper [2]. As it is shown on the figure 2.1 the light is spectrally dispersed. The radiation must be coherent, the coherency can be defined with the energy distribution an the electron bunch made of Ne electrons. dI dI = Ne Sinc + Ne2 Scoh (2.1) dΩ Ne dΩ 1 dI With dΩ the energy for one electron, Ω the solid angle. Sinc and Scoh represent the 1 incoherent and coherent part of the radiation. Thus the coherent emission begin when the coherent term is dominant, this append if the bunch size is in the order of, or less than, the wavelength. As explained in the article [2].
3
CHAPTER 2. THEORY AND EXPERIMENT The size of the bunch can be calculated with the length of the bunches delivered by the accelerator. If we make the approximation that c is the velocity of the bunch (ultra relativistic bunch), an accelerator that produce a pico-second length bunch will induce a limit of the coherent wavelength given by : l = time ∗ velocity ≈ 10−12 ∗ 3.108 = 300µmm
(2.2)
With l the length of the bunch. To see coherent Smith Purcell radiation, with this type of accelerator, the spectral range must be infrared and far infrared. This explain the use of the infrared detectors and the need to calibrate them. In the theory, the wavelength λ of the Smith Purcell radiation is defined as : l 1 − cos θ λ= n β With l the period of the grating, n the order of diffraction, β = the bunch and θ the angle of observation.
2.2
v c
(2.3)
with v the velocity of
CLIO
Free Electron Laser CLIO (Centre Laser Infrarouge d’Orsay) has been built in 1992, it is a Free-Electron Laser (FEL) that produce high peak power and tunable wavelength beam in the midinfrared spectral range. A FEL use a high-energy electron beam as an amplifying medium. The electron beam emits light as it wiggles through a periodic magnetic structure called undulator. The light is stored in in an optical cavity, and can interact back with the electrons. This interaction leads to a modulation of the electronic density, and a growth in intensity and coherence of the emitted light1 .
Figure 2.2: CLIO Free electron laser scheme http://clio.lcp.u-psud.fr/clio_eng/clio_eng.html
1
http://clio.lcp.u-psud.fr/clio_eng/clio_eng.html
M.Bonamy
page 4
CHAPTER 2. THEORY AND EXPERIMENT
Accelerator The ETALON group use CLIO for the research on Smith Purcell Radiation because it can produce single electron bunches of 8±1ps and an energy up to 50MeV.The CLIO accelerator consist of a thermionic gun, a subharmonic buncher (SHB), a fundamental buncher (FB) and an accelerating cavity (AC). The gun produce bunches about 1.5 ns long at an energy of 90 keV. Those bunches are then compressed by the sub-harmonic buncher to 200 ps or less to make it suitable for further compression with the fundamental buncher. This fundamental buncher further compresses the beam to a few ps and accelerates bunch to several MeV, making the electrons relativistic. The bunches are then further accelerated in the accelerating cavity to the operation energy (typically 15-50MeV) as explained in the article [4]. A full description of the FEL is made in the article [3].
Figure 2.3: CLIO accelerator scheme [3]
M.Bonamy
page 5
CHAPTER 2. THEORY AND EXPERIMENT
2.3 2.3.1
Experiment Pyrodetector
Figure 2.4: Mounted pyrodetector In order to detect the radiation emitted during the experiment, pyrodetectors Fig. 2.4 are used. The sensitive part, to the infrared modulation, of the detector is the grey cylinder electrode in the middle of the detector. The detector is mounted on the circuit with a 50GΩ load resistor. The signal can be observed using a BNC cable. When a radiation is absorbed by the ferromagnetic material (crystal) and converted into heat, it turn to increase the temperature of the crystal. The change in temperature alters the lattice spacings within the crystal producing a change in the polarization, below Curie temperature. When the temperature changes of ∆T , the surface charge Q will change by ∆Q following the equation given by : dP ∆T = AKP ∆T (2.4) ∆Q = A dT Where P is the electric polarisation, KP the pyroelectric coefficient and A the surface area of the detector element. The resulting current of the detector, when it is connected to an external circuit, can be written as : dT is = AKP (2.5) dt M.Bonamy
page 6
CHAPTER 2. THEORY AND EXPERIMENT
This resulting current is proportional to the rate of change of temperature. By connecting the circuit to an appropriate load the changing current will create a changing voltage that can be applied to an acquisition module to read the response. Those types of pyrodetectors are used because they have a good response in the far IR. As we seen in the Smith Purcell experimentation made on CLIO we need to be in IR to have a see coherent radiation and more in far IR. This The heat equation of a radiation of power P can be written as : Cth
dT + λ (T − T0 ) = P dt
(2.6)
Where T0 is the ambient temperature, P the radiative power, λ the thermal conductivity and Cth he thermal capacity
2.3.2
Acquisition module
The acquisition module have been built by Vitalii Khodnevych an Ukrainian student who did his internships with N.Delerue for the ETLON project. The module is shown on Fig. 2.5.
Figure 2.5: Aquisition module The part number 1 is the eight different channels inputs of the module. Those inputs are used to plugged the different pyrodetectors with BNC cables. The module is supply by the spot number 4. The part number two is made for connections. During my internship I always used the Ethernet connection. The BNC output number 3 is made for the trigger input. The acquisition module have two threads : • A DAQ thread to read from the ADC and to save in files. • An user interface which is designed to simplify user-board communication. M.Bonamy
page 7
CHAPTER 2. THEORY AND EXPERIMENT
The communication with the user is done by the use of sockets and require the knowledge of the IP address and port number. I have chosen to use Python language for the socket communication in order to use only one language for the reading part and the data processing. On the terminal, once connected with the module, there is the possibility to communicate with it. Different command are feasible: • -r send data from last trigger with EPOCH and trigger number. • -p(channel number)a 245 change the amplification of the channel to 245 (go from 255 to 0 ) • -p(channel number)o 253 change the offset of the channel to 253 (go from 255 to 0 ) The file received is a .txt with 8 columns for the eight channels. The the number of sample (line, acquisition) must be defined in the C code of the module. During my experiment I used 1800 samples per channels Here, on the Fig. 2.6, is an example of the discussion with the acquisition module on the terminal.
Figure 2.6: DAQ talk
M.Bonamy
page 8
Chapter 3 Characterization of the aquisition module 3.1 3.1.1
Approach Module response
My first approach at the beginning of this internship was to understand the module response in order to be able to understand the detector response (the detector is plugged into the acquisition module). My first task was to make sure that the output signal was not cut and well centred to have the maximum amplitude. To do that, I use a frequency pulse generator. This generator generates signals of different frequencies and amplitudes. The first measurements makes me realise that the bigger the frequency was the bigger the amplitude. As the experiment doesn’t go further the 100Hz, I made few acquisitions in order to have the better response I have chosen on the DAQ the best amplitude and offset parameters in the C code. I have settle the generator amplitude at 100mV PP and I have made acquisitions for 30, 50, 70 and 100 Hertz.
Figure 3.1: Module response for a sinusoidal signal (30 and 70 Hertz)
9
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
The Fig. 3.1 shows an acquisition made with a python script I wrote. The input signal was an sinusoid signal of 30Hz and 70Hz. We can observe the channels response to the signal input and directly observe that the amplitude depends on the frequency and the eight channels have not exactly the same response. Those difference in offset and amplitude are the reason of my work on the acquisition module.
3.1.2
Coding solution
First I tried to understand why the amplitude was linked to the frequency, and then to characterize the response of the eight channels, in amplitude offset and sampling rate. To characterize those parameters I have made two different codes in python. The first one is made for acquire the module results and to plot the figures. The second was made in order to fit the curves of the eight channels in the same time. The fit returns the best parameters of Amplitude, Frequency Phase and Offset using non-linear least squares. The Fig. 3.2 shows the plot of the channel 1 and its fit. and the Fig. 3.3 shows the plot and fit of the eight channels given by the full program.
Figure 3.2: Fit of the channel 1
Figure 3.3: Fit of the eight channels
M.Bonamy
page 10
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
3.1.3
Uncertainties
The acquisition module give an answer on 4.096 V coded in 8 bits over 256 values. So the precision in Volts on each sample is : 4.096/256 = 0.016V . Thus I implemented on the fit function this uncertainty. With the curve fit function I had access to the variance of the parameter estimated. To compute one standard deviation errors on the parameters I made the square of the standard deviation. Each time to do my fit the number of sample was 1800. The order of the uncertainties is 10−1 mV for the frequencies 30 and 50 Hertz and 1mV for 70Hz and 100Hz. Those uncertainties are small because I have made 1800 points acquisition. I’ll explain, on the following parts my results regarding the amplitude, offset and sampling rate of the module response to different frequency input. All the results shown have the uncertainty describe above. Regarding the sampling rate, the uncertainty is 1/1000 of its value.
M.Bonamy
page 11
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
3.2 3.2.1
Amplitude Link between amplitude and frequency
At the beginning I used the sinus mode of the generator, the input was a typical A sin(ωt+ Φ) and I was able to observe it with the oscilloscope. The module response was also a sinusoid of the same frequency but the amplitude kept changing every time with the frequency. My hint was to think about what can affect the amplitude when the frequency increase. During a derivation the amplitude is multiplied by ω and so, by the frequency. My guess was to say that the module was applying a derivative action on the input signal and I tired to prove it. Thanks to the different mode of the generator, I was able to validate my idea. Indeed by looking at the response for a square or triangle signal I was able to observe rectangle and Dirac response. See Fig. 3.4 to 3.7. The Input signal is the yellow one, the green one is the trigger.
Figure 3.4: Square input on the oscilloscope
Figure 3.5: Square output signal
Figure 3.6: Triangle input on the oscilloscope
Figure 3.7: Triangle output signal
M.Bonamy
page 12
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
3.2.2
Characterization of the amplitude
My second work on the amplitude was to characterize the eight channels. I have made 5 different measurements on 3 different days with different conditions. May 26th one acquisition, 27th and 30th two acquisitions, one the the morning and the other on the afternoon. The change between the 27th and 30th of may was that I didn’t unplugged the module during the two different acquisitions on the 27th . This was to see I there was to charge accumulation or heat issues for a long experiment. The same measurements are used for the offset and sampling rate study. Here are some plots to show the difference of amplitudes between the channels and also the moment of the acquisition.
Figure 3.8: Amplitude in function of the frequency
Figure 3.9: Amplitude in function of the frequency for the channel 1
By looking at the figure Fig. 3.9 we can see that there is a difference between each channels and also that this difference becomes bigger when the frequency is higher. We can observe in the figure Fig. 3.8 that the amplitude it is not perfectly constant even during one day a fortiori one day to another.
Conclusion about the amplitude We can see that the amplitude is directly correlated to the frequency with a linear dependency. That dependency is due to a derivative action. Also the amplitude is not exactly the same for each channels and the variation goes bigger with the frequency. That mean that, it will be needed to find the correlation in amplitude between each channels before any comparison. M.Bonamy
page 13
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
3.3
Characterization of the Offset
On the Fig 3.10 we can observe that the offset depends clearly on the channel used. Also this figures shows that for one acquisition, the offset is not linked to the frequency. On the other hand the Fig. 3.11 highlights the fact that the offset does depend on the measurement from one day to another. Also this difference is not the same for each channels.
Figure 3.10: Offset in function of the frequency
Figure 3.11: Offset in function of the frequency for the channel 1
Conclusion about the offset Regarding the results, it is shown that there is a multiple influence on the offset parameter. Either the channel or the acquisition time. However the offset calculation is the easiest to make. By implementing few code lines, like mean or median calculation, this problem can be easily manage. The idea will be to subtract the offset in order to have every channels at zero.
M.Bonamy
page 14
CHAPTER 3. CHARACTERIZATION OF THE AQUISITION MODULE
3.4
Characterization of the Sampling rate
As in the section Amplitude and offset, here are the results regarding the sampling rate. The first Fig. 3.12 represent the sampling rate in function of the frequency. As the number of points for the acquisition is always the same we have a linear dependency with the frequency. Also the second plot of the figure 2.10 shows the sampling rate per detector in function of the frequency. The Fig. 3.13 figure shows the sampling rate for the channel 1 for different acquisitions.
Figure 3.12: Sampling rate in function of the frequency
Figure 3.13: Sampling rate in function of the frequency for the channel 1
Conclusion about the sampling rate As we can see on the figures 3.12 and 3.13, my results shows that there was no dependency with either the channel or the moment of the acquisition were made.
M.Bonamy
page 15
Chapter 4 Detector calibration In order to do the detector calibration, I had to make the acquisition module calibration and this part took me a major part of my internship time. In this part I will present you the work I have made in order to begin the calibration and also my ideas for the continuation on this calibration. The calibration on the pyrodetectors need to be reproducible, in that matter I had to think of a way to do my measurement and also the assembly to make sure that the human uncertainty are the most negligible. In that matter, during the last part of the internship I have made a lot of experimental assembly. The following results come from the final assembly I think to be the most suitable to do the calibration for now.
4.1
Calibration
Assembly The last prototype of assembly is shown Fig. 4.1. In order to calibrate the detector, I put my source (light bulb) in the object focal point of a lens. Then the light will be focused on detectors using a parabolic mirror of 25mm OAP. The reason why it is suitable to make the image at the infinite is because, for a distance of few centimetres the parabolic mirror will get the same image with an intensity on average constant. This assumption will be discus after. Using this constant intensity, I have attached my mirror to a translation system in order to calibrate more than one detector at the time (the DAQ module have eight channels). On the picture 4.1 the light comes from the left. I made sure that my mirror and optical beam were well align with this "cage" linking every part of the assembly. The translation system can be easily piloted with a software on a computer and can move very precisely. The position is given with a number of turn from the starting position (0). However, because of hysteresis problem, a check-up are needed when multiples go and back have been performed. The detectors are mounted in two different ways. The position of the detector must be verified each day to be sure that the crystal is still on the focus point of the mirror. At the beginning of my work the modulation light was made with a chopper, but in order to have more short impulsion for the same frequency and intensity I have coupled an DC 16
CHAPTER 4. DETECTOR CALIBRATION
Figure 4.1: Optical Assembly generator with the pulse generator. The DC generator act like a follower. Characterization Before any acquisitions I made a test to see if the intensity was constant in an enough length. The acquisition is represented in Fig. 4.2. This acquisition have been made with a photo-diode placed on the support assembly, exactly like the mirror. For each measurement, a background acquisition have been made. We can see that the curve as two parts. During the first 4 centimetres, the intensity is clearly deceasing with a linear shape. Then the shape is constant with a mean value close to 250mV . In consequence the detectors must be placed in that constant area to be sure that the response will not be influenced.
Figure 4.2: Intensity in function of the distance to the lens
M.Bonamy
page 17
CHAPTER 4. DETECTOR CALIBRATION
The other acquisition that have been made is the response of a detector regarding the misalignment. The knowledge of the detector response to misalignment is very important because once the detector will be on the operating room and submitted to radiation, the operator need to be capable of realign the detector without entering after shooting. The figure 4.3 shows the detector response in function of the distance. The error bars on the x positions is one of my future work.
Figure 4.3: Response Amplitude in function of the distance to the lens The figure 4.3 has been made with the optical assembly describe above. I have made 40 steps of 2500 motor turns. In order to get the signal I had to implement in my fit code a part where I had each time the 50Hz noise. The amplitude on this plot is only the 30Hz contribution. In black are represented the error bars given by the fit function (same as in the acquisition module calibration). We can observe that the detector have a maximum constant response during 10 000 motor steps. Again, this studies will be done more precisely in the following days.
4.2
Future work
The future work to do will be to characterize the misalignment of the detector in the depth direction. Also to be sure that the detector response is the same during one day and furthermore from one day to another, I will take a lot of acquisition during one day and compare them together in amplitude, offset and also noise. The construction of a black-box to preserve the detector from the background can be conceivable. Once the results of the detector calibration will be good enough, this assembly can be use to continue the calibration but with only infra-red source of different wavelength.
M.Bonamy
page 18
Chapter 5 Conclusion During this internship I had to learn and discover a lot, as well on personal knowledge than on the research world. The goal of my internship was to calibrate and test the response of pyrodetector. However in order to calibrate them, I had to deal an manage some unknown problems. I had to calibrate the acquisition module. This part took me a lot of time. I have started my internship at ground-state knowledge about python coding, now I acquired abilities in this language. The acquisition module response is now understood. Also in order to keep track and to help the future person who wants to continue my work, I have created HTML pages to guide them. With the HTML knowledge this exercise needed an synthetic but clear presentation of my work and results. The second part of my internship was the detector calibration, but, as my internship ends the 6 of July and that I had a week free in the middle of may, this part will be continue after this report. Anyway, for now the beginning of this part helped me to really discover and understand the idea of reproducible in the research world. Now that I made my optical assembly, I’m looking at the response of different detectors to see if their response is the same during one day and one day to another. This work will permit to see there is a difference and in that case I’ll try to interpret it and see how it can be calibrated. On the personnel level, this internship helped me to see the research world, and the life in a laboratory with a team. I had to be also independent on my work and develop criticism on the results. I had the chance to visit and to take acquisitions in the CLIO control room for two days, this was very rewarding experience.
Thanks I would like to thank Nicolas for letting me work in his team. For his help in my project, his patience and explanations. I’m glad to have discovered and saw this world in is team. This, I’m sure, contributed to make this internship a wonderful experiment. Also, thanks to all the other interns for their help and kindness that made a friendly framework, AnneFleur, Vitalii, Maksym, Clément and Yu. Thanks to everyone, those 10 weeks internship were as well exciting as rewarding.
19
List of Figures 2.1 2.2 2.3 2.4 2.5 2.6
Smith Purcell radiation [1] . . . CLIO Free electron laser scheme CLIO accelerator scheme [3] . . Mounted pyrodetector . . . . . Aquisition module . . . . . . . . DAQ talk . . . . . . . . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
3 4 5 6 7 8
3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13
Module response for a sinusoidal signal (30 and 70 Hertz) . . Fit of the channel 1 . . . . . . . . . . . . . . . . . . . . . . Fit of the eight channels . . . . . . . . . . . . . . . . . . . . Square input on the oscilloscope . . . . . . . . . . . . . . . . Square output signal . . . . . . . . . . . . . . . . . . . . . . Triangle input on the oscilloscope . . . . . . . . . . . . . . . Triangle output signal . . . . . . . . . . . . . . . . . . . . . Amplitude in function of the frequency . . . . . . . . . . . . Amplitude in function of the frequency for the channel 1 . . Offset in function of the frequency . . . . . . . . . . . . . . Offset in function of the frequency for the channel 1 . . . . . Sampling rate in function of the frequency . . . . . . . . . . Sampling rate in function of the frequency for the channel 1
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
9 10 10 12 12 12 12 13 13 14 14 15 15
4.1 4.2 4.3
Optical Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Intensity in function of the distance to the lens . . . . . . . . . . . . . . . . 17 Response Amplitude in function of the distance to the lens . . . . . . . . . 18
20
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
Bibliography [1] H.L. Andrews, F. Bakkali Taheri, J. Barros, R. Bartolini, L. Cassinari, C. Clarke, S. Le Corre, N. Delerue, G. Doucas, N. Fuster-Martinez, I. Konoplev, M. Labat, C. Perry, A. Reichold, S. Stevenson, and M. Vieille Grosjean. Longitudinal profile monitors using coherent smith–purcell radiation. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 740:212 – 215, 2014. Proceedings of the first European Advanced Accelerator Concepts Workshop 2013. [2] J. H. Brownell, J. Walsh, and G. Doucas. Spontaneous smith-purcell radiation described through induced surface currents. Phys. Rev. E, 57:1075–1080, Jan 1998. [3] R Chaput, P Joly, B Kergosien, J Lesrel, and O Marcouillé. Optimisation of the FEL CLio Linear Accelerator. 1994. [4] Nicolas Delerue, Stéphane Jenzer, Vitalii Khodnevych, Jean-Paul Berthet, Francois Glotin, Jean-Michel Ortega, and Rui Prazeres. Study of Short Bunches at the Free Electron Laser CLIO. In Proceedings, 7th International Particle Accelerator Conference (IPAC 2016), page MOPMB005, 2016. [5] S. J. Smith and E. M. Purcell. Visible light from localized surface charges moving across a grating. Phys. Rev., 92:1069–1069, Nov 1953.
21
