A new model and tests of the JEM-EUSO Balloon pathfinders Fresnel optics
Abraham N. Díaz Damián
Space Engineering, master's level (120 credits) 2016
Luleå University of Technology
Department of Computer Science, Electrical and Space Engineering
Contents
1 Introduction 1
1.1 Cosmic rays . . . . 2
1.1.1 Extensive Air Showers . . . . 2
1.2 Air fluorescence detection technique . . . . 7
1.3 The JEM-EUSO Mission . . . . 9
1.3.1 EUSO-Balloon . . . . 10
1.3.2 EUSO-SPB . . . . 11
2 The EUSO-Balloon Optics 13 2.1 Fresnel optics . . . . 13
2.2 Optics configuration of the EUSO-Balloon flight . . . . 15
2.3 Modeling and Simulation of the EUSO Telescopes optics . . . . 15
2.3.1 EUSO-Balloon optical model . . . . 16
2.3.2 Ray tracing . . . . 17
2.3.3 System Evaluation . . . . 18
3 A new model of the EUSO-Balloon optics 21 3.1 Tool peak rounding . . . . 21
3.1.1 Simulation of rounded valleys . . . . 22
i
3.2 Lenses scratches . . . . 23
3.2.1 Scratch Model . . . . 24
3.3 Surface roughness . . . . 25
3.3.1 Rough Surface Model . . . . 25
3.4 Optical simulation results . . . . 27
3.4.1 Results of individual e↵ects . . . . 27
3.4.2 Results of combined e↵ects and comparison with measurements 28 3.5 Discussion of results . . . . 28
3.5.1 Tool peak rounding . . . . 28
3.5.2 Scratches . . . . 29
3.5.3 Surface Roughness . . . . 29
4 The EUSO-SPB Testing campaign 33 4.1 EUSO-SPB optical testbench . . . . 33
4.2 Testing procedure . . . . 34
4.3 Optical characterisation test results . . . . 36
5 Conclusion and perspectives 39
Abstract
EUSO-Balloon and EUSO-SPB are balloon borne pathfinder projects designed to val- idate the techniques of the JEM-EUSO space observatory. They are nadir pointing UV telescopes that use experimental experimental Fresnel optics to detect the ultravi- olet emission of Extensive Air Showers (EAS) induced by Ultra High Energy Cosmic Rays (UHECR) in the atmosphere. EUSO-Balloon was launched by the balloon di- vision of CNES (the french space agency) from Timmins, Ontario, Canada in 2014.
Despite the success of the mission the performance of the optics was lower than what it was originally modeled and led to many doubts regarding the understanding of the optics and fresnel lenses themselves. This thesis explores three parameters proposed to explain the reduction in efficiency of the system which were not simulated in the original characterization: the rounded valleys in the Fresnel lens grooves created by the tool peak radii, scratches on the surface of the lenses and the surface roughness of the fresnel lenses. These parameters were simulated and results show that they show a reduction in performance which approximates more the characterization measure- ments but still do not match exactly, leaving room for further analysis. EUSO-SPB1 is the successor of EUSO-Balloon with a launch planned in 2017 from Wanaka, New Zealand. The results of the first phase of the optics characterization campaign is pre- sented in this work. The results indicate that the optics performance is similar to that of EUSO-Balloon and require further understanding.
iii
R´ esum´ e
EUSO-Ballon et EUSO-SPB sont des t´elescopes d´emonstrateurs con¸cus pour valider les techniques de l’observatoire spatial JEM-EUSO. Ils sont des t´elescopes pointant au nadir qui utilisent des lentilles de Fresnel comme ´el´ements optiques et sont con¸cues pour d´etecter la lumi`ere ultraviolette cr´ee par les gerbes atmosph´eriques induites par les rayons cosmiques d’ultra haute ´energie. EUSO-Balloon a ´et´e lanc´e par la division ballon du CNES (Centre National d’´ Etudes Spatiales) depuis Timmins, Canada en 2014. Malgr´e le succ`es de la mission, le syst`eme optique a ´et´e moins performant par rapport au mod`ele num´erique de l’instrument. Cela a cr´e´e des doutes sur la compr´ehension du syst`eme optique et les lentilles de Fresnel. Ce rapport explore trois param`etres propos´es pour expliquer la r´eduction de efficacit´e du syst`eme qui n’ont pas ´et´e simul´ees dans la campagne de caract´erisation: les vall´ees rondes des zones de Fresnel cr´ees par le rayon l’outil d’usinage, les rayures dans la surface des lentilles et la rugosit´e de surface des lentilles de Fresnel. Ces param`etres ont ´et´e simul´ees et les r´esultats montrent une r´eduction de performance qui s’avoisine aux mesures de caract´erisation mais qui restent en divergence. Ces r´esultats laissent de la place pour une analyse ult´erieure des lentilles de Fresnel. EUSO-SPB1 est le successeur d’EUSO-Balloon avec un vol programme en 2017 depuis Wanaka, Nouvelle Z´elande.
Les r´esultats de la premi`ere phase de la campagne de caract´erisation optique sont aussi pr´esent´es. Ces r´esultats indiquent que la performance optique est similaire `a celle d’EUSO-Balloon et a besoin d’une meilleure compr´ehension.
v
Acknowledgements
This Master thesis is the culmination of a project which some time ago seemed only like a dream but that right now has become the best experience of my life so far, it has led me closer to a field which I’m very passionate about and it has become the gateway for future PhD work.
All of this wouldn’t have been possible without the work and help of many people whom i feel very grateful to. I would like to thank Victoria Barabash for taking the enormous task of coordinating and accepting me into the SPACEMASTER program, Peter Von Ballmoos for coordinating the Engineering track in Toulouse, for being a great tutor, always very supportive and of course for getting me involved in this great project. I would also like to thank the team at Colorado School of Mines (CSU) for giving me the great opportunity of working for a couple of weeks on their facilities and letting me experience, test and analyze firsthand the instrument which up to that moment i had only read about and simulated.
Being a SpaceMaster 10 roundie has been a great experience full of learning and growth. It wouldn’t be the same without my friends/colleagues who shared this expe- rience with me. I would like to thank Erick for helping me out when I just arrived to Germany and my colleagues and friends who were part of the Toulouse track: Aaron, Alberto, Beata, Ben, Ferdinand, George, Luca and Manuel. It will be an unforgettable adventure having shared the last ear of the program with you. Last but not least i would like to thank my mother Laura, and my whole family because without their un- conditional support i wouldn’t have been able to come this far, thank you for always supporting me in any idea I had!
Abraham D´ıaz Dami´an
vii
Chapter 1 Introduction
Everyday, the earth is constantly bombarded by visitors from outer space called cosmic rays. Cosmic rays are charged particles, that span a huge energy range comprising multiple orders of magnitude: from 10 6 to above 10 20 eV. In the lower energy range of the spectrum cosmic rays are very common and arrive by the seconds. In the high energy range, i.e., 10 18 eV we have the so called Ultra High Energy Cosmic Rays (UHECR). These are the most enigmatic and rarest cosmic messengers that visit the earth and despite having been observed for decades their astrophysical sources and acceleration mechanisms remain unknown.
The JEM-EUSO mission (Japanese Experiment Module - Extreme Universe Space Observatory) is a proposed space observatory with the goals of observing UHECRs at the highest energies (¿ 3 ⇥10 19 eV) with an unprecedented exposure and answering the fundamental questions about the UHECRs origins and acceleration mechanisms. This will be accomplished by observing the Ultra-Violet (UV) fluorescence emission induced from the UHECR nuclear interactions with the atmospheric molecules, the so called
”air fluorescence technique”. Although this technique is firmly established on ground, it is not yet validated from space and the project doesn’t meet the requirements to be a viable space mission. This validation will be accomplished through the development of experimental balloon pathfinders that will test the technology and techniques in working conditions similar to the proposed space telescope.
In this master thesis I will focus on the optics sub-system of the JEM-EUSO balloon pathfinders: EUSO-Balloon and EUSO-SPB (super pressure balloon). The telescopes are characterised by their experimental Fresnel optics. The optics is still not well understood and require further testing and study to achieve a better experimental performance. To understand better the context of this work I will start by reviewing the context of cosmic rays, their energy spectrum, the particle showers induced from cosmic rays and the fluorescence detection technique. I will then explain more in detail the context of the JEM-EUSO mission and its balloon pathfinders.
1
1.1 Cosmic rays
Cosmic rays were discovered in 1912 the Austrian scientist Victor Hess after he per- formed seven balloon flights to measure the ionization rate on the atmosphere [10].
Equipped with a set of electroscopes, the state of the art radiation detector, Hess per- formed di↵erent altitude incursions during the balloon flights and observed that the ionization rate increased as a function of altitude. To confirm this observation, Hess also performed flights during the night and a partial eclipse but this had no noticeable e↵ect on the ionization rate. From these experiments Hess correctly concluded that this mysterious radiation came from outer space and called it ”H¨ohenstrhalung” (high radiation). For this discovery he was awarded the Nobel prize in physics in 1936.
Figure 1.1 shows Victor Hess preparing for one of his iconic balloon flights.
Since the discovery of cosmic rays, intense experimental e↵orts in the field have allowed us to discover more about the nature of these particles and their unprecedented energies. In the left panel of figure 1.2 we can observe the full cosmic ray spectrum which spans over 11 orders of magnitude in energy and 24 in flux. A particular feature of the spectrum is that is described in large energy regions by a broken power law whose spectral index changes according to the energy region. The spectral index transitions are known as the ”knee” at about 3 ⇥ 10 15 eV and the ”ankle” at about 4 ⇥ 10 18 eV. Although we won’t discuss the astrophysical implications of the spectrum and its transitions, it is typically assumed that the knee region represents the end of the spectrum’s galactic cosmic accelerators and the ankle is associated with the emergence of extragalactic particles, however this is still a debated picture [11].
The right panel of figure 1.2 shows a close up of the high energy region of the cosmic ray spectrum. We can appreciate that above 10 20 eV we have the region of highest uncertainty in the spectrum. This is due to the low number of particles detected experimentally, a consequence of their extremely low flux of about 1 particle / km 2 century. In practice the direct detection of cosmic ray particles can only be done from space or stratospheric balloons since they interact with the atmospheric molecules and lose their energy in the process. More over direct detection is only feasible up to an energy of about 10 15 eV. Above this energy the flux of cosmic rays is so low that detection is only feasible through indirect means. To accomplish this we rely on the detection of the particle cascade produced by the primary cosmic ray, known as extensive air shower.
1.1.1 Extensive Air Showers
Extensive air (EAS) showers are produced when a cosmic ray or high energy photon
arrives to the earth and collides with an atmospheric molecule. This interaction pro-
duces a generation of secondary particles which in turn interact with more molecules
setting o↵ a chain reaction that creates a cascade of more generations of particles, all
from the energy of the primary particle [13]. In figure 1.3 we can see an schematic rep-
resentation of the EAS and its hadronic and electromagnetic constituents. Depending
1.1. COSMIC RAYS 3
Figure 1.1: Victor Hess during one of his iconic balloon flights
Figure 1.2: Cosmic ray spectrum obtained from results of multiple experiments. Left:
Spectrum from 10 8 to 10 21 eV. Right: Close up in the high energy range of the spectrum
from 10 18 to 10 21 eV. Plots from [12]
on the primary that produces the EAS (particle or photon) the shower development has di↵erent characteristics. A good and simple starting point for understanding EAS development is Heitler’s model. Initially developed for electromagnetic (EM) show- ers, this model has been extended to the scenario of hadronic showers by Matthews [matthews˙heitler˙2005]. Although the model is simple and does not take into ac- count all details of an EM shower it is good to show the physics involved and manages to predict accurately the most important characteristics of EM shower development.
Electromagnetic showers
In the original Heitler’s model (see fig 1.4a), an EM shower involves electrons (e ), positrons (e + ) and photons ( ) which experience a repeated two particle (e ± and ) multiplication after they travel a fixed distance (d) related to the radiation length in the medium ( r ) by d = r ln 2. The multiplication is done by two main processes:
one-photon Bremsstrahlung and e e + pair production. After n splitting lengths, with a distance of x = n r ln 2, the number of particles in the shower ((e and ) is N = 2 n = e x/
r. The multiplication process is halted when the individual e e + energies fall below the critical energy (⇠ e c ) threshold. At this point the particle energy is too low for pair production and bremsstrahlung due to the radiative losses being lower than the collision energy losses. In air the critical energy is 85 MeV.
If we consider a shower started by a photon of energy E 0 . The cascade reaches its maximum size N = N max when all particles have the critical energy ⇠ c e so that
E 0 = ⇠ e c N max (1.1)
The EM shower reaches its maximum size at the penetration depth X max , this is obtained by calculating the number of splitting lengths n c required so the energy per particle is reduced to ⇠ c e . Since N max = 2 n
cfrom eq. 1.1 we obtain that n c = ln[E 0 /⇠ c e ]/ ln 2 which gives
X max = n c r ln 2 = r ln[E 0 /⇠ c e ] (1.2)
The elongation rate ⇤, defined as the rate of X max change per decade of primary energy, is given by:
⇤ ⌘ dX max
d log 10 E 0
(1.3)
Using the X max from eq. 1.2 yields an elongation rate ⇤ = 2.3 r = 85 g/cm 2 for
EM showers in air.
1.1. COSMIC RAYS 5
Figure 1.3: Schematic representation of an extensive air shower showing its electromagnetic (blue), hadronic (red) and muonic (green) components
Figure 1.4: Heitler’s Model of electromagnetic shower
development.[matthews˙heitler˙2005]
The model overestimates the number of particles at the shower maximum by a factor of two to three. This happens because multiple photons are radiated during bremsstrahlung which leave the electron with less energy to maintain the multiplication process and also because the model does not treat the loss of particles as ionization ceases to occur.
Taking these e↵ects into account in simulations show that there is a photon to electron-positron number ratio of about six. This is one order of magnitude less elec- trons than what is predicted by Heitler’s model N max and holds true for higher energies and other type of media. If we wish to extract the number of electrons N e from Heitler’s shower size N we can adopt the correction factor:
N e = N/g (1.4)
Where g = 10 and is a simple order of magnitude estimate.
Hadronic showers
In the hadronic shower model, the atmosphere is assumed to consist of fixed thickness layers I ln 2. I is the interaction length of strongly interacting particles and it is assumed to be constant. For pions in air I ⇡ 1120g/cm 2 . After traversing one layer, the Hadrons interact producing N ch charged pions (⇡ ± ) and 1 2 N ch neutral pions (⇡ 0 ).
Neutral pions decay into photons, initiating splinter EM showers whereas charged pions continue interacting through subsequent layers. This process is repeated until the charged pions are below the critical energy (⇠ c ⇡ ) where they are assumed to decay into muons.
If we consider a primary cosmic ray proton that enters the atmosphere with an energy E 0 , then after traversing n layers of atmosphere there are N ⇡ = (N ch ) n charged pions. If we assume an equitable distribution of energy in the particle production process then these pions will carry a total energy of ( 2 3 ) n , with the remainder energy gone into splinter EM showers from neutral pion decay. Thus the energy per charged pion in the atmospheric layer is:
E ⇡ = E 0
( 3 2 N ch ) n (1.5)
knowing that one third of the primary energy is lost at each interaction stage due
to neutral pion decay which then initiates splinter EM showers. Then, to calculate the
primary energy we have to include the hadronic and EM shower component. This is
done by using the total number of pions N ⇡ and EM particles N max from the splinter
showers. Similar to eq. 1.1 the total energy is:
1.2. AIR FLUORESCENCE DETECTION TECHNIQUE 7
E 0 = ⇠ c e N max + ⇠ c ⇡ N µ
Scaling to the electron size N e = N max /g, then
E 0 = g⇠ c e
N e + ⇠ c ⇡ g⇠ c e N µ
⇡ 0.85 GeV (N e + 24N µ )
(1.6)
Overall, despite the shortcomings of these simplified models, they’re helpful to understand the air shower development and predict acurately two important features:
1. The maximum size of the shower is proportional to the primary energy E 0 2. The depth of shower maximum X max increases logarithmically at a rate of 85
g/cm 2 per decade of primary energy.
This means that if we can experimentally measure the size of the showers, then, we could estimate its most important characteristics even if we don’t directly detect the primary particle.
1.2 Air fluorescence detection technique
Current ground high energy cosmic ray observatories use a mix of ground arrays of par- ticle detectors and fluorescence telescopes to detect the hadronic and part of the elec- tromagnetic component of the shower (there is also a radio component and Cerenkov radiation produced in the shower). Since JEM-EUSO and its pathfinders exploit the fluorescence technique we will focus exclusively on it.
When an EAS is developing the energy deposits cause the excitation of atmospheric nitrogen molecules in the air. The spontaneus de-excitation produces an isotropic UV fluorescence light emission that is proportional to the energy deposit and can be used to perform a calorimetric measurement of the energy of the shower and thus of the primary particle. The fluorescence emission is produced in the wavelength range between 290 and 430 nm [19] and this is the window that the JEM-EUSO telescopes are optimized to detect. Figure ?? shows the measured fluorescence spectrum of air [8].
The technique works by measuring the air fluorescence light emitted during the
EAS. The fluorescence telescopes are composed of a primary mirror which focuses
light into a UV sensitive camera composed by an array of Photo-Multiplier Tubes
Wavelength (nm)
290 300 310 320 330 340 350 360 370 380 390 400 410 420 15
10
5
0
Relative intensity
Figure 1.5: Spectrum of the air fluorescence emission produced by Extensive Air Showers
(PMT). The fluorescence technique was pioneered by the ”fly’s eye” experiment which also detected experimentally one of the highest energy cosmic rays ever observed, with an energy of (3.2 ± 0.9) ⇥ 10 20 eV.
Currently the top UHECR observatories are the Pierre Auger Observatory (PAO) [7] and the Telescope Array (TA) [23]. Both observatories use multiple fluorescence telescopes in conjunction with particle detector arrays for the observation of EAS. In figure 1.6 we can see an schematic representation of the technique. In order to deal with the very low flux of cosmic rays at the highest energies (1 p/sr 2 km 2 century above 10 20 eV), the observatories have to cover very large surface areas. PAO covers an area of about 3000 km 2 whereas TA covers 700 km 2 , resulting in the monitoring of limited volume of the atmosphere.
It is clear that the limiting factor in the number of detections is the area covered
by the observatories. An increase in the size of the observatories would only linearly
increase the exposure to UHECRs. So what if instead of increasing the detection rate
by increasing the size of ground observatories we could instead observe a larger volume
of the atmosphere by looking further away?. This is what motivates the observation of
EAS from space and it’s where the JEM-EUSO mission comes into play. Although lim-
ited to the air fluorescence technique it is still a viable alternative since the technique
is already validated from ground. Validation from space is the next important stepand
its why it’s important to build prototypes to prove the feasibility of the mission.
1.3. THE JEM-EUSO MISSION 9
Figure 1.6: Schematic representation of the observation of an extensive air shower using a particle detector array and fluorescence telescopes. The shower axis of propagation is in the FOV of all telescopes and can be reconstructed from the analysis of the event
1.3 The JEM-EUSO Mission
The JEM-EUSO instrument [22] is a UV space telescope planned to be installed on- board the Japanese Experiment Module of the International Space Station (ISS). It is a wide field of view (FOV) nadir pointing telescope designed to observe the Extensive Air Showers (EAS) induced by Ultra High Energy Cosmic Rays by detecting their UV fluorescence tracks from the ISS orbit. The main objectives of the mission are the identification of the UHECR sources, the measurement of the energy spectra of individual sources and measuring the spectrum at the regions of highest uncertainty in the cosmic ray spectrum. in figure we can see an artist depiction of the instrument on-board the ISS.
To achieve its goals the instrument will be equipped with a wide FOV refractive- di↵ractive Fresnel optics and an ultra-fast camera consisting of Multi-anode Photo- multiplier tubes. The system will have a time resolution of 2.5 µs and will be capable of detecting individual UV photons from the fluorescence emission. From the ISS orbit at about 400 km altitude, JEM-EUSO will monitor in nadir mode a surface area of about 1.5 ⇥ 10 5 km 2 . The instrument can also be tilted increasing its surface cov- erage about six-fold. However, this will also increase the energy threshold necessary to detect a UHECR. This mode will be used to perform a study in the high energy limit of the cosmic ray spectrum with very good statistics. To achieve its objective the JEM-EUSO collaboration is developing ground, balloon and smaller scale tele- scopes to validate its technology. Our focus will be in the two balloon pathfinders:
EUSO-Balloon and EUSO-SPB.
1.3.1 EUSO-Balloon
EUSO-Balloon [1] is a balloon-borne nadir pointing UV telescope and its main objec- tives are: demonstrating the technology for JEM-EUSO, performing data acquisition, a UV background study and be pioneering mission for JEM-EUSO[2]. EUSO-Balloon used a scaled down version of the JEM-EUSO optics and photo-detector module to perform its light collection and data acquisition. Previous research work on EUSO- Balloon optics was done by C. Catalano[6] who performed the modeling and testing of the Fresnel optics as part of his PhD dissertation. The optics characterisation was done in two stages: one pre-flight campaign which was limited due to time constraints set by the balloon flight. Nevertheless, the work on the optics led to a successful first light of the instrument during its maiden balloon flight on Timmins, Canada. Dur- ing the flight EUSO-Balloon proved its capabilities as a technology demonstrator by successfully measuring the night sky UV background and detecting UV laser shots from kilometers away. This set up the groundwork for further improvements towards a space observatory. Following this first success a more thorough post-flight campaign which was performed to gain a deeper understanding of the Fresnel optics and its performance.
Figure 1.7: Measured fluorescence spectrum excited by 3 MeV electrons in dry air at 800 hPa and 293 K [8].
The present work takes o↵ from the work previously done by Catalano and attempts to better understand the optics performance of the instrument. Although the optical performance was above the minimal requirements, it was still lower than what was expected according to the performance prediced by the numerical model of the telesope.
Because of this Catalano concluded that there were various mechanical factors in the
fabrication and handling of the lenses that contributed to a di↵usion of light and
1.3. THE JEM-EUSO MISSION 11 thus a lower performance. The proposed factors were: rounded angles of the Fresnel lenses due to the dimensions of the cutting tool, scratched surfaces due to shipping and handling of the lenses, and the microscopic surface roughness of the lenses. These factors are explored and explained more in detail in chapter 3 as well as the results of the new simulations taking these parameters into account.
1.3.2 EUSO-SPB
EUSO-SPB is the next iteration of EUSO-Balloon, it will use the same observation
principle and will feature improved sub-systems. At the time of this work the mision is
in preparation for it’s flight which will take place around march 2017. The project will
be launched from Wanaka, New Zealand on-board the new Super Pressure Balloon
(SPB) currently in development by NASA. In chapter 4 I will describe part of the
testing campaign of EUSO-SPB performed with the objective of optimizing the optics
configuration for its future balloon flight.
Chapter 2
The EUSO-Balloon Optics
The optics of the EUSO-Balloon balloon and EUSO-SPB telescopes are a scaled down version of the larger JEM-EUSO optics. Their main characteristic is the use of Fresnel optics, an type of refractive optics still in an experimental stage for the EUSO projects and still requires further testing and understanding. The Fresnel lenses of EUSO- Balloon were the first of its kind to perform observations of a the UV night sky and EAS simulated with laser shots, proving their potential for the space observation of EAS. This section describes the functional principle of Fresnel lenses, the Fresnel optics configuration of EUSO-Balloon, its optical model and evaluation metrics.
2.1 Fresnel optics
Fresnel optics are a type of refractive optics whose design principle consists in reducing as much optical material as possible from a typical lens while retaining its surface profile and focusing power. This can be accomplished because light is refracted only at the boundaries of the lens with the surrounding medium. The intermediate bulk material contributes only to light absorption and extra mass without any performance benefit. Therefore, compared to a normal lens, a Fresnel lenses has an equivalent focusing power, it improves transmission by removing the bulk material and reduce the mass which results in flat lenses with a ring structure composed of many teeth equivalent to the surface profile of the equivalent baseline lens (see figure 2.2). These characteristics make Fresnel lenses attractive for large optics and space optics.
Although they may seem as rare devices, Fresnel lenses are quite ubiquitous and we can find them in various applications, mostly dealing with light concentration and illumination. For example we can find them in smartphone flashes and in the front lights of some cars which help in collimating the light towards the front of the devices.
Fresnel lenses are not a new technology, they were developed in the 18th century by Jean Augustine Fresnel and their initial application was to collimate of light from lighthouses to allow the light beam to propagate further and make it visible to ships further away from the coastline.
13
Figure 2.1: Left: EUSO-Balloon in it’s flight configuration with gondola structure and dampers. Right: Schematic view of the instrument box and optical bench.
Figure 2.2: Principle of Fresnel lens design. Starting from a base plano-convex lens (left),
the intermediate lens material is removed while preserving the surface profile. This results
in a thin lens (middle) with concentric rings (right) with similar focusing properties as the
base lens.
2.2. OPTICS CONFIGURATION OF THE EUSO-BALLOON FLIGHT 15 Fresnel lenses present several advantages as well as limitations inherent to their design. The main advantages include: reduced lens thickness, which results in better transmission efficiency since the light is absorbed by less optical material and also im- plies a reduction of cost in material and mounting; they are fast lenses, meaning they have shorter focal lengths; they have design freedom and can have large diameters. Dis- advantages include: chromatic aberration due to dispersion of di↵erent wavelengths, geometric shadowing and transmission loss due to the grooved structure, resolution limitations due to the flat prismatic type surface approximations and reduced contrast due to the vertical surfaces and the radii of peaks and valleys of the grooves [9].
2.2 Optics configuration of the EUSO-Balloon flight
The baseline configuration of the EUSO-Balloon optics consisted in a three lens system composed of two 1m 2 refractive Fresnel lenses known as L1 and L3 and one di↵ractive lens known as L2. The latter was placed between the refractive lenses and was de- signed to correct the chromatic aberration typical of refractive optics. The design of the EUSO-Balloon and EUSO-SPB lenses is completely. The lenses of were manufac- tured at the RIKEN institute in Japan, one of the main collaborators of JEM-EUSO.
They were designed and optimized with the industrial software CODE V and manu- factured in their laboratory already experienced in fabricating Fresnel Lenses for solar concentration applications.
The lenses are made in a 1 m x 1 m x 8 mm sheet of Polymethyl Methacrylate (PMMA) a thermoplastic with good optical properties. The L1 lens has Fresnel zones with a 1 mm width whereas the L3 lens has zones with a 2 mm width. The di↵ractive lens L2 was manufactured along L1 and L3 and was initially foreseen to take part in the characterization campaigns and flight of 2014. The lens was designed to correct the chromatic aberration caused by the refracting lenses in order to focus all the di↵erent UV wavelengths in the same position. Unfortunately after performing preliminary tests including this lens it was realized that its performance wasn’t good enough as it reduced the optical efficiency considerably and upon visual inspection it’s surface looked very hazy. Because of this the L2 was dropped and the characterization campaign and flight of EUSO-Balloon only made use of L1 and L3. Because of this, in this work only the two refractive lens configuration will be considered.
2.3 Modeling and Simulation of the EUSO Tele- scopes optics
The modeling and simulation of the EUSO telescopes is performed by an in-house
raytracing application developed by our collaborators at RIKEN Japan. The software
was developed initially for the JEM-EUSO Optics but it can be easily adapted for the
EUSO-Balloon or other pathfinder models. This program is run entirely in the bash
shell trough the use of commands and has no Graphic User Interface (GUI). To run a simulation a command in the shell is given specifying the wavelength, incidence angle, the number of rays traced and an XML configuration file containing the parameters of the telescope model.
2.3.1 EUSO-Balloon optical model
The geometry of the EUSO-Balloon optics is represented in a 3D cartesian system where the optical axis is the Z-axis and the direction of propagation of light is the positive direction. The system model is defined in several text files which can be modified to specify the characteristics of the system. In one main configuration file we can specify the Z axis position for each surface of the system, the optical medium after each surface interface (air or PMMA), the origin and injection radius of the light rays and the position of the focal surface. In this file we also specify the name of secondary text files which contain the description of each surface and the refractive index of various wavelengths for each medium. The main concepts and the physics behind ray tracing are explained in the section 2.3.2 and the evaluation of the optical system in section 2.3.3.
Ray tracing simulations normally make use of parametric lenses (spherical or as- pherical) to calculate the intersection points of light rays in the system. Since Fresnel lenses structure more complex, the profile of the surface is specified in secondary files which contain the coordinate points of the radial profile of the lens, the profiles of both sides of L1 and L3 are shown in figure 2.3.
Figure 2.3: Surface profile of the Fresnel lenses L1 and L3 from the center to a radius of 300 mm.Top: L1. Bottom: L3.
Surface Z position [mm] Medium
Source ( 1) -1 Air
L1 (flat) 0 PMMA (Ambient temp.)
L1 (Fresnel) 8 Air
L3 (Fresnel) 1123 PMMA (Ambient temp.)
L3 (Flat) 1131 Air
Focal Surface 1562
Table 2.1: Optical Configuration
2.3. MODELING AND SIMULATION OF THE EUSO TELESCOPES OPTICS 17
2.3.2 Ray tracing
Ray tracing is an technique derived from the geometrical optics approximation of light developed before the electromagnetic theory and is still relevant for the design and evaluation of optical systems. Geometrical optics assumes that light is discrete and infinitesimal model of light that propagates in straight lines along homogeneus media and changes direction when traversing di↵erent materials. These lines are called light rays and they are normal to the propagating wavefront of light [21]. This approx- imation is valid as long as the system is much larger than the wavelength of light as only refraction takes place. If the optical system is much smaller or has features with dimensions comparable to the wavelength of light being used then the e↵ects of di↵raction are non trivial and have to be taken into account.
Refraction index
The refraction index is a fundamental parameter in geometrical optics. It describes the ratio of the velocity of light in vacuum with respect to the velocity in an optical medium such as glass or plastic, it is denoted by the letter n.
n = Velocity in vacuum
velocity in medium (2.1)
The index of refraction of the PMMA material used to construct the lenses was measured at RIKEN and a table with it’s real values for several UV wavelengths is included in the raytracing software.
Snell’s law of Refraction
To calculate the trajectory of the light rays traced through the system we use of Snell’s law of refraction which tells us how light rays are refracted along two interfaces of di↵erent media:
n 1 sin I 1 = n 2 sin I 2 (2.2)
Where n 1 is the refractive index of the first medium, I 1 the incidence angle of the light ray with respect to the second medium, n 2 is the refractive index of the second medium and I 2 is the refracted angle in the second medium. This simple relationship is the basis for calculating the passage of light rays traced through the optical system.
This is represented schematically in figure 2.4.
θ 1
θ 2 n 1
n 2
n 1 θ 1 = n 1 θ 2
Figure 2.4: Schematic representation of the Snells law of refraction. The upper optical medium has a lower refractive index than the lower medium (in blue). The light ray changes direction at the interface of the two optical media.
2.3.3 System Evaluation
There are many significant parameters to evaluate the quality of the images produce by and optical system such as the Modulation Transfer Function (MTF), the Strehl ratio and the encircled energy distribution [21]. The figure used for the first campaign of characterization campaign of the optics was the encircled energy, for continuity reasons it will also be used for the analysis in this work.
Encircled Energy
The encircled energy (EE) is the measurement of the ratio of light in the Point Spread Function produced by the system with respect to the light entering the input pupil of the system as a function of the radius of the PSF. To calculate it experimentally first we need to measure the total energy of the incident light in the entrance pupil of the system and then measure the PSF. In practice the PSF is measured by scanning it in discrete intervals and making a grid of measurements. To calculate the encircled energy distribution as a function of the PSF radius we have to integrate the total energy at increasing radius intervals and then normalize it by the total incident energy as follows:
I encircled (r) = 1 n
X n 1
I n
✓ ⇡ ⇤ r 2 a
◆
(2.3)
Where n is the number of measurement points, I n is the integrated flux of all the
measurements in the radius interval and a is the area of the measurement point. A
2.3. MODELING AND SIMULATION OF THE EUSO TELESCOPES OPTICS 19 schematic represntation of this procedure and the resulting energy distribution can be seen in figure 2.5.
Figure 2.5: Left: Focal spot with radial section intervals. Right: Example of encircled energy plot with integrated energy percentages over each radial interval.
In the case of simulations, since light rays carry no energy information and they are infinitesimally small, the numerical calculation of the encircled is based on ray counting. The incident energy is the number of rays injected in the input pupil and the energy distribution is number of rays hitting the focal surface. For the numerical calculation we have:
I encircled (r) = 1 n
X n 1
I(R i r) (2.4)
Where n is the number incident of rays, I is the number of rays at the current radius interval, and R i is the current radius interval.
System Efficiency
In the original modelling of the system done in order to fabricate the lenses the RMS
diameter of the PSF for di↵erent wavelengths was estimated to be 9 mm. From this
number the efficiency of the system is defined as the percentage of encircled energy
contained in a 9 mm diameter (or a 4.5 mm radius) from the centroid of the PSF.
Chapter 3
A new model of the EUSO-Balloon optics
After the second characterization campaign done in 2015 after the Timmins balloon flight, the expected results as foreseen by the simulations didn’t match the results that were obtained after the measurements. This lead to a series of questions as to what could have caused these discrepancies. After a careful analysis Catalano [6] identified several causes, mostly dealing with mechanical features related to the fabrication and handling of the lenses. These causes were:
• Rounded fresnel valleys due to the finithe lenght of the diamond tool tips used to mill the lenses.
• Scratches caused during the shipping, handling and testing of the lenses.
• The microscopic surface roughness of the lenses.
These causes are not implemented on the ray tracing simulations, i.e. the lenses are perfectly flat in the original model. In this section I describe in detail the causes proposed by Catalano, I research existing models for these features and implement them into the digital model of the lenses, I analyse their e↵ect on the performance of the optics and compare it with the original characterization results of the 2015 testing campaing of EUSO-Balloon.
3.1 Tool peak rounding
The fabrication of the PMMA lenses is a highly technical and time extensive proce- dure. The lenses are manufactured in the fabrication laboratory at RIKEN in Japan and the procedure consists in cutting the surface of the lenses with a diamond tip tool to form the shape of the Fresnel profile in two steps:
21
1. First the surface of the raw PMMA sheet is cut with a coarse tool of a 0.5 mm diameter tip, which defines the Fresnel profile of the lens. Since the length of each teeth from top to bottom is 1 mm in L1 and 2 mm in L3, a second pass is needed to complete the profile.
2. A second pass with a fine tool of 0.05 mm diameter tip, finishes cutting the depth of the valleys to complete the fresnel profile.
(a) Lenses are first coarse cut with 0.5 mm diameter tool, then fine cut with 0.05 mm diameter tool.
(b) Due to the physical limitation owed to the dimension of the tool which cuts the Fresnel profile, part of the lens does not refract light as expected.
Figure 3.1: Tool peak rounding error
The fabrication steps can be seen in figure ??, this factor is important to consider because it can a↵ect performance significantly depending on the characteristics of a Fresnel lens i.e. diameter, number of Fresnel zones, width between each zones. As op- posed to a conventional lens which is a continuous surface, the rounded valleys will not refract light according to the intended design, therefore a small fraction of the lenses surface will not perform as intended (see fig. 3.1b)creating a loss in performance which becomes more significant for L3 because light is more concentrated by the previous lens.
3.1.1 Simulation of rounded valleys
In the original simulations, the profile of the lenses is considered to be ”perfect” i.e.
the vertex of the fresnel zone valleys is formed by 2 straight edges and the rounding
cause by the tool shape is not taken into account. Moreover L3 was cut with the fine
tool only up to a radius of 302 mm from the center, this was done in order to save
fabrication time since the simulations performed when designing the lenses determined
that no more light would pass through the surface of the lens starting from the afore-
mentioned radius. However as seen before, the Timmins configuration di↵ers from the
originally three lens configuration, so in order to verify if this a↵ected the performance
of the system, Catalano performed a ray tracing simulation and determined that for
larger incidence angles, the light rays do pass above the fine cut radius, hence passing
3.2. LENSES SCRATCHES 23 through the coarse Fresnel zones and a↵ecting the performance of the system.
Since the ray tracing simulation works by reading a text file with the point co- ordinates of the radial profile it was possible to simulate the tool peak rounding by modifying the original coordinate file which was possible due to the creation of an algorithm which reads the original coordinates, detects the corners in the lens profile and replaces the sharp valley coordinates with the coordinates of the radial section of a circle of 0.05 mm radius that fits into both sides of the valley. As mentioned before, L3 wasn’t cut with the fine tool along all of it’s radius so this is also taken into consideration.
The results of the simulation can be seen in figure ??, in the first subfigure a section of the rounded L1 profile can be seen with the original profile in a red line, in the second subfigure the rounded L3 profile at 302 mm away from the center, where the di↵erent simulated tool dimensions can be seen.
(a) L1 rounded profile (b) L3 rounded profile Figure 3.2: Tool peak rounding of L1 and L3
3.2 Lenses scratches
The PMMA lenses and in general any optical component are fragile devices. Having
an area of 1 ⇥1 m and a width of only 8 mm, they can bend very easily and although
PMMA is a resistant material the surface can get scratched easily as well. Therefore a
careful handling and storage is necessary to preserve the optical quality. Nevertheless
upon receiving and unpacking the set of lenses used for the Timmins flight delivered
from RIKEN, the team at IRAP noted that the lenses had some light scratches due to
the shipping which were noted and mapped. During the testing campaign, the bending
of the lenses mounted in a vertical position (in nadir mode) was tested. Sphere-Mounts
(Leica spherically mounted retro reflectors) were placed on various positions on each
lens to measure changes in the flatness profile with a laser tracker. Despite having a
tape layer to protect the lens, the surface was scratched unintentionally. In this section
I will attempt to implement a simple model of scratches and use it in the simulations.
3.2.1 Scratch Model
A model by Bosch et Al. [5] which was developed to simulate realistic scratches in computer 3D graphics was implemented to simulate scratches on the flat side of each lens in our ray tracing simulation. Bosch’s et al method consisted first in performing a scratch test on di↵erent metallic surfaces using a Rockwell diamond cone and ap- plying di↵erent forces for each test scratch. The hardness of each test material was measured on the vickers scale and afterwards the cross section of each scratch was analyzed with a profilometer to obtain the cross section dimensions of the scratch.
These values were used to relate the behavior of the scratches to the properties of the material in order to derive empirical formulas to determine the geometry of the scratch.
From these tests a scratch model was developed with the following characteristics:
a valley created by the penetrating scratching material and two adjacent crests, whose cross section area is equivalent to the cross section area of the valley, meaning that no material loss is assumed and the crests are created by the displaced material from the valley. It is important to note that this the of this model are related to the geometry of the tip of the indenter used in the tests and the force applied, therefore the scratches on our lenses might not share the same characteristics. Hence this is an approximation whose objective is to simulate the e↵ect of generic scratches on the system. The penetration depth of the valleys is derived by:
p = 0.182 r F N
H V
+ 0.55 0.014 (3.1)
where p is the penetration depth, F N is the force applied in kgf, H V is the vickers hardness (kg/mm 2 ) which is 19.9 for PMMA [20]. From the depth of the groove and the shape of the tool tip [17] the geometry of the peaks can be determined giving us the required parameters to simulate the scratch.
To implement this scratch model into the lenses I made a python script to simulate the scratches on the flat sides of the lenses. This works by first the reading the datafile with the original coordinates of the lenses, the user can choose the force applied to the surface and the percentage of the lens surface to scratch. The hardness of the PMMA material is hardcoded. The program calculates the dimensions of the scratch from the input values, interpolates the obtained coordinates into the x coordinates of the surface replacing the original flat profile with the scratched profile. Then the program calculates the radial rectuib area covered y scratch and repeats the process until the total scratched area reaches the percentage of lens area set in the input parameter.
An example of a simulated scratch is shown in figure and the simulation results are
shown in chapter ??.
3.3. SURFACE ROUGHNESS 25
Figure 3.3: Simulated scratch on lens surface
3.3 Surface roughness
All material surfaces wether natural or artificial have a degree of rugosity associated with them which could be visible or invisible to the eye. Since the advent of earlier telescopes it was known that smoother surfaces wether in lenses or mirrors resulted in better image quality, so polishing of optical surfaces became a critical part when creating lenses and mirrors for telescopes.
There are many parameters which could be used to describe a rough surface, the most commonly used ones are the arithmetic average roughness, Root-mean-square roughness, skewness and kurtosis and depending on the application of the surface one could be preferred over the other[15]. For telescope design it’s common to represent the surface roughness using RMS roughness parameter which is defined as the following:
R q = v u u t 1
N X N
i=1
y 2 i (3.2)
Where y i is the vertical distance from the ith point on the surface profile to the mean line, N is the number of points measured along the surface profile.
For the EUSO-Balloon optics the desired surface roughness is expected to be less than 20 nm RMS. The surface roughness of the lenses was measured with an atomic force microscope and determined to be satisfactory. Unfortunately I don’t have the measurement data and instead the approach is to randomly generate a rough surface as described in the next section.
3.3.1 Rough Surface Model
Surface profiles can be described deterministically or stochastically by their statistical
properties which is the case for random surfaces. For a random surface the stochastic
process describing it can take a wide variety of forms and behaviors. To simulate a
surface numerically di↵erent models have ben proposed however most studies make use
of Gaussian random process description to model the surfaces since they are simple
to implement and require only the height of the distribution function (HDF) and the autocovariance function (ACF)[18]. The model implemented for the simulations is a random gaussian rough surface outlined by Bergstrom, Powell and Kaplan [4].
A random rough surface, in this case one dimensional z = ⇠(x), can be character- ized two factors: its height distribution function which describes the mean deviation from the mean surface level (RMS roughness) and its autocovariance function which describes how the peaks and valleys of the surface are distributed along the surface length. Since the surface it’s assumed to have a Gaussian distribution then it’s HDF and ACF have the following properties:
p h (⇠) = 1
p 2⇡ 2 exp( ⇠ 2
2 2 ) (3.3)
C(x) = exp( x 2
⌧ 2 ) (3.4)
Where is the RMS roughness and ⌧ is the correlation length, since both the ACF and HDF the slope is also gaussian and is given by w = p
2 /⌧ [4].
To simulate the surface, first a one dimensional uncorrelated Gaussian surface distribution is generated with a Gaussian random number generator over an array of discrete points that equals the number of points that define the surface profile of each lens. To correlate the surface points, the distribution has to be convolved with the gaussian filter:
F (x) = 2
⌧ p
⇡ exp
✓ 2x 2
⌧ 2
◆
(3.5)
So we set:
⇣(x) = x 1
1
(F (x x 0 ) ⇥ ⇣ u (x 0 )dx 0 (3.6)
Which can be implemented via a Fast Fourier Transform algorithm. The resulting
simulated surfaces are shown in figure where both examples have a RMS roughness
of 20 nm and correlation lengths of 50µm and 20µm respectively. Results of the
simulations with this factor implemented are shown in chapter ??
3.4. OPTICAL SIMULATION RESULTS 27
Figure 3.4: Simulated gaussian random rough surfaces with di↵erent correlation lengths, Top: Surface with 20 nm RMS and 50 µm correlation length. Bottom: Surface with 20 nm RMS and 50 µm.
3.4 Optical simulation results
3.4.1 Results of individual e↵ects
The results of the encircled energies calculated from the simulations with the e↵ects cause by the tool peak rounding, scratches and surface roughness are shown in the figures 3.5a, 3.5b and 3.5c. For these simulations a wavelength of 337 nm was used since it’s the most representative one in the fluorescence spectrum and an incidence angle of 0 degree. These results are meant to show the individual e↵ect of each of the three parameters investigated.
(a) (b)
(c)
Figure 3.5: Encircled Energies of original lens profile and profile with: 3.5a: simulated tool peak rounding. 3.5b: Simulated scratches covering 1, 5 and 10% of L1 and L3 surface.
3.5c:Surface roughness with 20 nm RMS for all profiles and 0.10, 0.05 and 0.01 mm correlation length.
3.4.2 Results of combined e↵ects and comparison with mea- surements
The results of the simulations with the compounded e↵ects for the same wavelengths and incidence angles used in the post-flight characterization campaign of the optics are shown in the figures, the wavelength’s simulated are 313, 334, 365 and 405 nm.
All of the simulations take into account the rounded Fresnel valleys, 5% of scratched surface in L1 and L3 and a surface roughness of 20 nm RMS with a correlation length of 10µm.
3.5 Discussion of results
3.5.1 Tool peak rounding
Figure 3.5a show that the main contribution of the rounding of the fresnel groove valleys is a reduction in the overall encircled energy, this tells us that the grooves refract a portion of incoming light out of the focal plane, reducing the active surface of the lenses, however it does not explain the extreme amount of di↵use light which is characterized by mild slopes in the encircled energy plots as opposed to steep ones when light is better focused.
From figure 4.2 we can infer that the e↵ect of this parameter contributes to the
lower total encircled energy and the plateau in the new simulations is between the
error margin of the measurements, however the amount of di↵use light still does not
3.5. DISCUSSION OF RESULTS 29 match properly for all the wavelengths. This parameter is an inevitable constraint when fabricating a Fresnel and lens the only way to improve it would be to use smaller tools but this would imply a more costly and longer manufacturing process, so there is a tradeo↵ in the amount of quality to be expected from the lens and the complexity of the manufacturing process.
3.5.2 Scratches
Figure 3.5b shows that the e↵ect of scratches on the surface of the lenses contribute in reducing the overall energy in the focal surface just as the tool peak rounding but also fails to explain the di↵use light in the system. In the integrated simulation this parameter contributes to the lower total encircled energy along with the tool peak radii as seen in figure 4.2.
Unfortunately this parameter is not properly characterized since the lenses are not longer in Toulouse and the laboratory doesn’t have any surface characterization instrument for this task, therefore the simulation of the scratches on the lenses is for exploration purposes and gives a good idea of the reduction in performance with respect to the percentage of lens surface scratched. In the end this serves as a reminder that the lenses are to be handled with extreme care and any type of contact with their surface should be avoided.
3.5.3 Surface Roughness
Figure 3.5c shows that despite having a surface roughness of 20 nm RMS the encircled energies for all the configurations remains mostly the same, however it can be seen that that the plateau of the encircled energy starts to bend as the correlation length of the simulated surface decreases, this tells us that a small amount of light is di↵used at the plateau of the encircled energy. Figure 4.2 shows that for all the wavelengths the bending of the encircled energy profile approaches the measurements closer compared to the original simulations, but still in the middle range between 5 and 15 mm the plots for the 2 shorter wavelengths are farther away than the longer ones.
Only the height distribution is reported from the manufacturing facility but it is
shown that shorter correlation lengths also play an important role in the di↵usion of
light. Despite the e↵orts to simulate this parameter with the most precision there are
several limitations in the simulation that make it hard to do so, first of all the sam-
pling length of the lenses profile 10 µm so roughness profiles with correlation lengths
shorter than this are impossible to simulate at the moment. This is an important
factor because in order to simulate the scattering properly the spatial frequency of the
roughness has to be higher (shorter correlation lengths) to obtain steeper slopes and
get more accurate results.
Other important factors to consider further on and are not taken into account are the di↵raction losses of the fresnel lenses. Despite not being intended as di↵ractive elements, their design does cause some di↵raction losses due to their grated structure [14]. Because of this ray tracing might not be adequate enough for the analysis of Fresnel lenses and a wave optics simulation approach might be necessary to perform a more realistic simulation, specially if using the di↵ractive lens. Moreover, research done by Jiao and Cheng [16] shows that the machining parameters of CNC fabrication on PMMA play a crucial role in the resulting surface roughness and adequate pro- gramming of the process, i.e. spindle speed, angle of tool normal to the surface, feed rate, stepover, depth of cut and direction must be carefully controlled to guarantee the desired surface roughness of the element being processed. Since the angle of each Fresnel zone increases gradually it is a possibility that the angle of the tool normal to the surface changes gradually resulting in di↵erent surface roughness along the length of the lenses, however the manufacturing procedure is unknown at the moment.
(a) (b)
(c) (d)
3.5. DISCUSSION OF RESULTS 31
(e)
(f ) (g)
(h) (i)
(j) (k)
(l) (m)
(n) (o)
Figure 3.6: Results of new simulations against characterization measurements.
(a): 313 nm 0.1 °, (b): 313 nm 2.3°, (c): 313 nm 3.3°, (d): 313 nm 4.5°
(e): 334 nm 0.1 °, (f): 334 nm 3.3°, (g): 334 nm 4.5°
(h): 365 nm 0.1 °, (i): 365 nm 2.3°, (j): 365 nm 3.3°, (k): 365 nm 4.5°
(l): 405 nm 0.1 °, (m): 405 nm 2.3°, (n): 405 nm 3.3°, (o): 405 nm 4.5°
Chapter 4
The EUSO-SPB Testing campaign
EUSO-SPB is the second balloon pathfinder of JEM-EUSO which aims to detect for the first time an UHECR from above. Although it reuses most of the mechanical structure of EUSO-Balloon it also has significat improvements with respect to it’s predecessor.
The major improvements of EUSO-SPB include: an improved photo-detector module, read-out electronics, an on-board event trigger system. The instrument is planned to be launched on the still experimental NASA’s super pressure balloon in 2017. Compared to the Timmins flight in 2014 which only lasted one night, EUSO-SPB is expected to fly up to 100 days. Due to the expected long duration of this flight, the mechanical structure and power sub-system need to be significantly revamped to function properly in the harsh environment of the stratosphere and posses the power autonomy for the projected 100 days.
The funding of EUSO-SPB was provided by NASA and the responsibility of the optics characterization passed to the Colorado School of Mines (CSU) in Golden, USA. Thus, EUSO-Balloon was shipped from Toulouse to Golden to prepare it for its transformations into EUSO-SPB. Since this was the first time the team at CSU had direct contact with the optics, the Toulouse team traveled to Golden to help set up a test bench similar to the one of EUSO-Balloon flight and using the old set of Fresnel lenses prior to the arrival of the three new lenses of EUSO-SPB, this was done in order to first replicate the results obtained in Toulouse and set a starting point for the new characterization campaign which will also use the L2 di↵ractive lens.
4.1 EUSO-SPB optical testbench
To test the optics of EUSO-SPB we set up a testbench based on the one of EUSO- Balloon but with some modifications (see fig. 4.1). We tested the EUSO-SPB optics using a 2 and a 3 lens configuration to compare the performance of each configuration.
We used the following wavelenghts for the tests: 340, 350, 370 and 390 nm. To understand the angular dependence of the performance we tested two incidence angles:
0 °and 4°. As a light source we used a set of fiber coupled UV LEDs covering part of
33
the fluorescence emission spectrum of air. Each UV emission was narrowed to a single wavelenght by using a bandpass filter with a Full Width Half Maximum (FWHM) of 10 ±2 nm. The output of the fiber coupled LED was split in two ends. The output of one line was placed in front of a NIST traceable photodiode to monitor the flux from the LED; the output of the other line was placed at the focal point of a 1 m mirror. The divergent beam coming from the optical fiber was reflected on the mirror and collimated back to our optics, simulating a light source at infinity. To measure the input flux of the optics we used a polar scaner that measured various points in front of the input pupil of the optics. To measure the PSF performance we used an 3D scanner to take discrete measurements of the PSF. The testing procedure is described in the next section.
Figure 4.1: EUSO-SPB optics testbench set up. The components are the following: a) 1 m collimating mirror. b) Fiber coupled LEDs. c) Fiber coupled LED output. d) Input pupil polar scanner with photodiode. e) L1 Fresnel lens. f) L3 Fresnel lens. g) Photodiode.
e) PSF 3D Scanner. f) Alignment laser
4.2 Testing procedure
To characterize the performance of the optics at each wavelength first we measured the input flux. To do this we used the polar scanner to scan di↵erent points in front of L1. To determine the incident flux the measurement points were interpolated and integrated. In figure 4.2 we can see the plot of one incident flux scan.
To measure the PSF we performed two type of scans. First we performed what we call a ”line scan”, this was a scan performed using a 1x1 cm 2 pixel along the optical axis. It serves as a fast scan that allows to identify the highest flux concentration for each wavelength and identify the optimal scanning position for the following scan.
Next we performed what we call a ”cube scan”. In this scan we measured a 3D grid of scan points with a pixel size of 1x1 mm 2 , first we scanned one plane normal to the optical axis and then we scanned another planed along the optical axis forming a cube of 3D measurements. Both type of scans are shown schematially in figure 4.3.
After all the measurements where taken, the output of the tests was a csv file
containing the 3D coordinates and measurement value of each scan point. This file
4.2. TESTING PROCEDURE 35
X [cm]
Y [cm] Normalized intensity
